diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,129623 @@ +{ + "best_metric": 1.60741866, + "best_model_checkpoint": "/home/ubuntu/s2/swift/output_qwen0_5B_2epochs/qwen2_5-0_5b/v0-20241211-015110/checkpoint-62000", + "epoch": 1.5728056823947236, + "eval_steps": 100, + "global_step": 62000, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "acc": 0.61795592, + "epoch": 2.536783358701167e-05, + "grad_norm": 6.84375, + "learning_rate": 2.536783358701167e-09, + "loss": 1.89075112, + "memory(GiB)": 14.88, + "step": 1, + "train_speed(iter/s)": 0.054125 + }, + { + "acc": 0.57646704, + "epoch": 0.00012683916793505834, + "grad_norm": 7.09375, + "learning_rate": 1.2683916793505834e-08, + "loss": 2.04062319, + "memory(GiB)": 18.73, + "step": 5, + "train_speed(iter/s)": 0.220218 + }, + { + "acc": 0.56165848, + "epoch": 0.0002536783358701167, + "grad_norm": 6.15625, + "learning_rate": 2.536783358701167e-08, + "loss": 2.15522804, + "memory(GiB)": 18.73, + "step": 10, + "train_speed(iter/s)": 0.3585 + }, + { + "acc": 0.55539145, + "epoch": 0.000380517503805175, + "grad_norm": 6.78125, + "learning_rate": 3.80517503805175e-08, + "loss": 2.24601555, + "memory(GiB)": 29.37, + "step": 15, + "train_speed(iter/s)": 0.453561 + }, + { + "acc": 0.564114, + "epoch": 0.0005073566717402334, + "grad_norm": 8.125, + "learning_rate": 5.073566717402334e-08, + "loss": 2.15130463, + "memory(GiB)": 45.31, + "step": 20, + "train_speed(iter/s)": 0.522365 + }, + { + "acc": 0.55909257, + "epoch": 0.0006341958396752917, + "grad_norm": 6.78125, + "learning_rate": 6.341958396752917e-08, + "loss": 2.18794022, + "memory(GiB)": 45.31, + "step": 25, + "train_speed(iter/s)": 0.575347 + }, + { + "acc": 0.57406425, + "epoch": 0.00076103500761035, + "grad_norm": 9.625, + "learning_rate": 7.6103500761035e-08, + "loss": 2.13946915, + "memory(GiB)": 45.31, + "step": 30, + "train_speed(iter/s)": 0.616678 + }, + { + "acc": 0.56211061, + "epoch": 0.0008878741755454084, + "grad_norm": 6.9375, + "learning_rate": 8.878741755454084e-08, + "loss": 2.19021778, + "memory(GiB)": 45.31, + "step": 35, + "train_speed(iter/s)": 0.650402 + }, + { + "acc": 0.58360443, + "epoch": 0.0010147133434804667, + "grad_norm": 7.28125, + "learning_rate": 1.0147133434804667e-07, + "loss": 2.13348656, + "memory(GiB)": 45.31, + "step": 40, + "train_speed(iter/s)": 0.677568 + }, + { + "acc": 0.58169861, + "epoch": 0.001141552511415525, + "grad_norm": 7.78125, + "learning_rate": 1.1415525114155251e-07, + "loss": 2.12300892, + "memory(GiB)": 45.31, + "step": 45, + "train_speed(iter/s)": 0.700417 + }, + { + "acc": 0.582936, + "epoch": 0.0012683916793505834, + "grad_norm": 7.21875, + "learning_rate": 1.2683916793505834e-07, + "loss": 2.10119209, + "memory(GiB)": 45.31, + "step": 50, + "train_speed(iter/s)": 0.720064 + }, + { + "acc": 0.57512264, + "epoch": 0.0013952308472856417, + "grad_norm": 7.375, + "learning_rate": 1.3952308472856418e-07, + "loss": 2.25024033, + "memory(GiB)": 45.31, + "step": 55, + "train_speed(iter/s)": 0.737398 + }, + { + "acc": 0.5534358, + "epoch": 0.0015220700152207, + "grad_norm": 7.6875, + "learning_rate": 1.5220700152207e-07, + "loss": 2.18388481, + "memory(GiB)": 45.31, + "step": 60, + "train_speed(iter/s)": 0.752232 + }, + { + "acc": 0.57148142, + "epoch": 0.0016489091831557584, + "grad_norm": 7.5, + "learning_rate": 1.6489091831557585e-07, + "loss": 2.14568672, + "memory(GiB)": 45.31, + "step": 65, + "train_speed(iter/s)": 0.765564 + }, + { + "acc": 0.57567549, + "epoch": 0.0017757483510908167, + "grad_norm": 6.5625, + "learning_rate": 1.7757483510908168e-07, + "loss": 2.14711494, + "memory(GiB)": 45.31, + "step": 70, + "train_speed(iter/s)": 0.778158 + }, + { + "acc": 0.5580235, + "epoch": 0.001902587519025875, + "grad_norm": 7.84375, + "learning_rate": 1.9025875190258752e-07, + "loss": 2.1960022, + "memory(GiB)": 45.31, + "step": 75, + "train_speed(iter/s)": 0.788237 + }, + { + "acc": 0.58497405, + "epoch": 0.0020294266869609334, + "grad_norm": 6.84375, + "learning_rate": 2.0294266869609335e-07, + "loss": 2.0636013, + "memory(GiB)": 45.31, + "step": 80, + "train_speed(iter/s)": 0.79697 + }, + { + "acc": 0.55256033, + "epoch": 0.0021562658548959918, + "grad_norm": 7.28125, + "learning_rate": 2.1562658548959918e-07, + "loss": 2.26351585, + "memory(GiB)": 45.31, + "step": 85, + "train_speed(iter/s)": 0.805073 + }, + { + "acc": 0.57233224, + "epoch": 0.00228310502283105, + "grad_norm": 7.0625, + "learning_rate": 2.2831050228310502e-07, + "loss": 2.12445126, + "memory(GiB)": 45.31, + "step": 90, + "train_speed(iter/s)": 0.812156 + }, + { + "acc": 0.55544052, + "epoch": 0.0024099441907661084, + "grad_norm": 6.96875, + "learning_rate": 2.409944190766109e-07, + "loss": 2.25526085, + "memory(GiB)": 45.31, + "step": 95, + "train_speed(iter/s)": 0.818579 + }, + { + "acc": 0.57460895, + "epoch": 0.0025367833587011668, + "grad_norm": 7.0, + "learning_rate": 2.536783358701167e-07, + "loss": 2.11610985, + "memory(GiB)": 45.31, + "step": 100, + "train_speed(iter/s)": 0.824498 + }, + { + "epoch": 0.0025367833587011668, + "eval_acc": 0.5655295179759007, + "eval_loss": 2.1939127445220947, + "eval_runtime": 112.5369, + "eval_samples_per_second": 56.604, + "eval_steps_per_second": 28.302, + "step": 100 + }, + { + "acc": 0.55362706, + "epoch": 0.002663622526636225, + "grad_norm": 8.1875, + "learning_rate": 2.6636225266362255e-07, + "loss": 2.19062538, + "memory(GiB)": 58.09, + "step": 105, + "train_speed(iter/s)": 0.43182 + }, + { + "acc": 0.58142099, + "epoch": 0.0027904616945712835, + "grad_norm": 6.25, + "learning_rate": 2.7904616945712836e-07, + "loss": 2.09046326, + "memory(GiB)": 58.09, + "step": 110, + "train_speed(iter/s)": 0.442608 + }, + { + "acc": 0.57731471, + "epoch": 0.002917300862506342, + "grad_norm": 6.90625, + "learning_rate": 2.917300862506342e-07, + "loss": 2.19999924, + "memory(GiB)": 58.09, + "step": 115, + "train_speed(iter/s)": 0.452974 + }, + { + "acc": 0.5735929, + "epoch": 0.0030441400304414, + "grad_norm": 6.40625, + "learning_rate": 3.0441400304414e-07, + "loss": 2.14763947, + "memory(GiB)": 58.09, + "step": 120, + "train_speed(iter/s)": 0.462852 + }, + { + "acc": 0.57480755, + "epoch": 0.0031709791983764585, + "grad_norm": 7.0625, + "learning_rate": 3.170979198376459e-07, + "loss": 2.04066162, + "memory(GiB)": 58.09, + "step": 125, + "train_speed(iter/s)": 0.472398 + }, + { + "acc": 0.5666563, + "epoch": 0.003297818366311517, + "grad_norm": 7.25, + "learning_rate": 3.297818366311517e-07, + "loss": 2.14853497, + "memory(GiB)": 58.09, + "step": 130, + "train_speed(iter/s)": 0.481577 + }, + { + "acc": 0.58718462, + "epoch": 0.003424657534246575, + "grad_norm": 7.3125, + "learning_rate": 3.4246575342465755e-07, + "loss": 2.08722267, + "memory(GiB)": 58.09, + "step": 135, + "train_speed(iter/s)": 0.490346 + }, + { + "acc": 0.57080498, + "epoch": 0.0035514967021816335, + "grad_norm": 7.71875, + "learning_rate": 3.5514967021816336e-07, + "loss": 2.11941071, + "memory(GiB)": 58.09, + "step": 140, + "train_speed(iter/s)": 0.49891 + }, + { + "acc": 0.57152734, + "epoch": 0.003678335870116692, + "grad_norm": 8.4375, + "learning_rate": 3.678335870116692e-07, + "loss": 2.13112221, + "memory(GiB)": 58.09, + "step": 145, + "train_speed(iter/s)": 0.507157 + }, + { + "acc": 0.57918911, + "epoch": 0.00380517503805175, + "grad_norm": 7.9375, + "learning_rate": 3.8051750380517503e-07, + "loss": 2.08793926, + "memory(GiB)": 58.09, + "step": 150, + "train_speed(iter/s)": 0.515128 + }, + { + "acc": 0.57022357, + "epoch": 0.0039320142059868085, + "grad_norm": 9.0, + "learning_rate": 3.932014205986809e-07, + "loss": 2.11384754, + "memory(GiB)": 58.09, + "step": 155, + "train_speed(iter/s)": 0.522904 + }, + { + "acc": 0.5806479, + "epoch": 0.004058853373921867, + "grad_norm": 7.125, + "learning_rate": 4.058853373921867e-07, + "loss": 2.15086823, + "memory(GiB)": 58.09, + "step": 160, + "train_speed(iter/s)": 0.530302 + }, + { + "acc": 0.55636721, + "epoch": 0.004185692541856925, + "grad_norm": 8.9375, + "learning_rate": 4.1856925418569256e-07, + "loss": 2.20608959, + "memory(GiB)": 58.09, + "step": 165, + "train_speed(iter/s)": 0.537458 + }, + { + "acc": 0.55536098, + "epoch": 0.0043125317097919835, + "grad_norm": 7.25, + "learning_rate": 4.3125317097919837e-07, + "loss": 2.19889297, + "memory(GiB)": 58.09, + "step": 170, + "train_speed(iter/s)": 0.544371 + }, + { + "acc": 0.5666749, + "epoch": 0.004439370877727042, + "grad_norm": 8.375, + "learning_rate": 4.4393708777270423e-07, + "loss": 2.21741619, + "memory(GiB)": 58.09, + "step": 175, + "train_speed(iter/s)": 0.551026 + }, + { + "acc": 0.56853361, + "epoch": 0.0045662100456621, + "grad_norm": 7.5625, + "learning_rate": 4.5662100456621004e-07, + "loss": 2.23747597, + "memory(GiB)": 58.09, + "step": 180, + "train_speed(iter/s)": 0.557568 + }, + { + "acc": 0.58063903, + "epoch": 0.0046930492135971585, + "grad_norm": 8.1875, + "learning_rate": 4.693049213597159e-07, + "loss": 2.13855095, + "memory(GiB)": 58.09, + "step": 185, + "train_speed(iter/s)": 0.563928 + }, + { + "acc": 0.55975227, + "epoch": 0.004819888381532217, + "grad_norm": 6.09375, + "learning_rate": 4.819888381532218e-07, + "loss": 2.14818268, + "memory(GiB)": 58.09, + "step": 190, + "train_speed(iter/s)": 0.569907 + }, + { + "acc": 0.57561865, + "epoch": 0.004946727549467275, + "grad_norm": 6.78125, + "learning_rate": 4.946727549467275e-07, + "loss": 2.18602638, + "memory(GiB)": 58.09, + "step": 195, + "train_speed(iter/s)": 0.575746 + }, + { + "acc": 0.56912994, + "epoch": 0.0050735667174023336, + "grad_norm": 7.0, + "learning_rate": 5.073566717402334e-07, + "loss": 2.13915958, + "memory(GiB)": 58.09, + "step": 200, + "train_speed(iter/s)": 0.581414 + }, + { + "epoch": 0.0050735667174023336, + "eval_acc": 0.5658240439290682, + "eval_loss": 2.1915087699890137, + "eval_runtime": 113.8731, + "eval_samples_per_second": 55.939, + "eval_steps_per_second": 27.97, + "step": 200 + }, + { + "acc": 0.56568356, + "epoch": 0.005200405885337392, + "grad_norm": 7.53125, + "learning_rate": 5.200405885337392e-07, + "loss": 2.18363876, + "memory(GiB)": 58.09, + "step": 205, + "train_speed(iter/s)": 0.438877 + }, + { + "acc": 0.57015295, + "epoch": 0.00532724505327245, + "grad_norm": 6.90625, + "learning_rate": 5.327245053272451e-07, + "loss": 2.17856483, + "memory(GiB)": 58.09, + "step": 210, + "train_speed(iter/s)": 0.444622 + }, + { + "acc": 0.55709519, + "epoch": 0.005454084221207509, + "grad_norm": 6.75, + "learning_rate": 5.454084221207509e-07, + "loss": 2.1720253, + "memory(GiB)": 58.09, + "step": 215, + "train_speed(iter/s)": 0.450229 + }, + { + "acc": 0.58103299, + "epoch": 0.005580923389142567, + "grad_norm": 6.40625, + "learning_rate": 5.580923389142567e-07, + "loss": 2.08644981, + "memory(GiB)": 58.09, + "step": 220, + "train_speed(iter/s)": 0.455712 + }, + { + "acc": 0.57829776, + "epoch": 0.005707762557077625, + "grad_norm": 6.0625, + "learning_rate": 5.707762557077626e-07, + "loss": 2.05979195, + "memory(GiB)": 58.09, + "step": 225, + "train_speed(iter/s)": 0.461088 + }, + { + "acc": 0.56890464, + "epoch": 0.005834601725012684, + "grad_norm": 6.15625, + "learning_rate": 5.834601725012684e-07, + "loss": 2.18327713, + "memory(GiB)": 58.09, + "step": 230, + "train_speed(iter/s)": 0.466375 + }, + { + "acc": 0.58338404, + "epoch": 0.005961440892947742, + "grad_norm": 6.09375, + "learning_rate": 5.961440892947743e-07, + "loss": 2.05939064, + "memory(GiB)": 58.09, + "step": 235, + "train_speed(iter/s)": 0.471526 + }, + { + "acc": 0.55905323, + "epoch": 0.0060882800608828, + "grad_norm": 6.0625, + "learning_rate": 6.0882800608828e-07, + "loss": 2.21649303, + "memory(GiB)": 58.09, + "step": 240, + "train_speed(iter/s)": 0.476553 + }, + { + "acc": 0.57836189, + "epoch": 0.006215119228817859, + "grad_norm": 7.40625, + "learning_rate": 6.215119228817859e-07, + "loss": 2.10696831, + "memory(GiB)": 58.09, + "step": 245, + "train_speed(iter/s)": 0.481501 + }, + { + "acc": 0.56371498, + "epoch": 0.006341958396752917, + "grad_norm": 7.21875, + "learning_rate": 6.341958396752918e-07, + "loss": 2.2602911, + "memory(GiB)": 58.09, + "step": 250, + "train_speed(iter/s)": 0.486384 + }, + { + "acc": 0.55713639, + "epoch": 0.006468797564687975, + "grad_norm": 6.9375, + "learning_rate": 6.468797564687976e-07, + "loss": 2.21758518, + "memory(GiB)": 58.09, + "step": 255, + "train_speed(iter/s)": 0.491167 + }, + { + "acc": 0.56888924, + "epoch": 0.006595636732623034, + "grad_norm": 7.78125, + "learning_rate": 6.595636732623034e-07, + "loss": 2.1437027, + "memory(GiB)": 58.09, + "step": 260, + "train_speed(iter/s)": 0.495839 + }, + { + "acc": 0.574404, + "epoch": 0.006722475900558092, + "grad_norm": 6.40625, + "learning_rate": 6.722475900558092e-07, + "loss": 2.1231205, + "memory(GiB)": 58.09, + "step": 265, + "train_speed(iter/s)": 0.500459 + }, + { + "acc": 0.59161148, + "epoch": 0.00684931506849315, + "grad_norm": 8.1875, + "learning_rate": 6.849315068493151e-07, + "loss": 2.10280876, + "memory(GiB)": 58.09, + "step": 270, + "train_speed(iter/s)": 0.505001 + }, + { + "acc": 0.56359673, + "epoch": 0.006976154236428209, + "grad_norm": 7.09375, + "learning_rate": 6.97615423642821e-07, + "loss": 2.14115467, + "memory(GiB)": 58.09, + "step": 275, + "train_speed(iter/s)": 0.509406 + }, + { + "acc": 0.56990967, + "epoch": 0.007102993404363267, + "grad_norm": 7.4375, + "learning_rate": 7.102993404363267e-07, + "loss": 2.17339745, + "memory(GiB)": 58.09, + "step": 280, + "train_speed(iter/s)": 0.513734 + }, + { + "acc": 0.58299518, + "epoch": 0.007229832572298325, + "grad_norm": 7.03125, + "learning_rate": 7.229832572298326e-07, + "loss": 2.07805138, + "memory(GiB)": 58.09, + "step": 285, + "train_speed(iter/s)": 0.517985 + }, + { + "acc": 0.59587684, + "epoch": 0.007356671740233384, + "grad_norm": 7.1875, + "learning_rate": 7.356671740233384e-07, + "loss": 2.08956184, + "memory(GiB)": 58.09, + "step": 290, + "train_speed(iter/s)": 0.522149 + }, + { + "acc": 0.58444786, + "epoch": 0.007483510908168442, + "grad_norm": 7.46875, + "learning_rate": 7.483510908168443e-07, + "loss": 2.1169714, + "memory(GiB)": 58.09, + "step": 295, + "train_speed(iter/s)": 0.526238 + }, + { + "acc": 0.58079667, + "epoch": 0.0076103500761035, + "grad_norm": 6.625, + "learning_rate": 7.610350076103501e-07, + "loss": 2.07222099, + "memory(GiB)": 58.09, + "step": 300, + "train_speed(iter/s)": 0.530201 + }, + { + "epoch": 0.0076103500761035, + "eval_acc": 0.5665927357784697, + "eval_loss": 2.185087203979492, + "eval_runtime": 114.331, + "eval_samples_per_second": 55.715, + "eval_steps_per_second": 27.858, + "step": 300 + }, + { + "acc": 0.57507534, + "epoch": 0.007737189244038559, + "grad_norm": 8.1875, + "learning_rate": 7.737189244038559e-07, + "loss": 2.17152939, + "memory(GiB)": 58.09, + "step": 305, + "train_speed(iter/s)": 0.442107 + }, + { + "acc": 0.57872977, + "epoch": 0.007864028411973617, + "grad_norm": 6.125, + "learning_rate": 7.864028411973618e-07, + "loss": 2.04192314, + "memory(GiB)": 58.09, + "step": 310, + "train_speed(iter/s)": 0.445958 + }, + { + "acc": 0.57205267, + "epoch": 0.007990867579908675, + "grad_norm": 6.6875, + "learning_rate": 7.990867579908676e-07, + "loss": 2.21588326, + "memory(GiB)": 58.09, + "step": 315, + "train_speed(iter/s)": 0.449772 + }, + { + "acc": 0.58382926, + "epoch": 0.008117706747843734, + "grad_norm": 6.5, + "learning_rate": 8.117706747843734e-07, + "loss": 2.1064106, + "memory(GiB)": 58.09, + "step": 320, + "train_speed(iter/s)": 0.453521 + }, + { + "acc": 0.54901762, + "epoch": 0.008244545915778792, + "grad_norm": 6.6875, + "learning_rate": 8.244545915778793e-07, + "loss": 2.19961834, + "memory(GiB)": 58.09, + "step": 325, + "train_speed(iter/s)": 0.457234 + }, + { + "acc": 0.55888462, + "epoch": 0.00837138508371385, + "grad_norm": 7.40625, + "learning_rate": 8.371385083713851e-07, + "loss": 2.21889915, + "memory(GiB)": 58.09, + "step": 330, + "train_speed(iter/s)": 0.460883 + }, + { + "acc": 0.5599515, + "epoch": 0.008498224251648909, + "grad_norm": 7.84375, + "learning_rate": 8.49822425164891e-07, + "loss": 2.18161201, + "memory(GiB)": 58.09, + "step": 335, + "train_speed(iter/s)": 0.464513 + }, + { + "acc": 0.57747803, + "epoch": 0.008625063419583967, + "grad_norm": 7.0, + "learning_rate": 8.625063419583967e-07, + "loss": 2.13993759, + "memory(GiB)": 58.09, + "step": 340, + "train_speed(iter/s)": 0.468058 + }, + { + "acc": 0.57594738, + "epoch": 0.008751902587519025, + "grad_norm": 6.125, + "learning_rate": 8.751902587519026e-07, + "loss": 2.12446156, + "memory(GiB)": 58.09, + "step": 345, + "train_speed(iter/s)": 0.471557 + }, + { + "acc": 0.56923542, + "epoch": 0.008878741755454084, + "grad_norm": 7.09375, + "learning_rate": 8.878741755454085e-07, + "loss": 2.11939297, + "memory(GiB)": 58.09, + "step": 350, + "train_speed(iter/s)": 0.475001 + }, + { + "acc": 0.57172971, + "epoch": 0.009005580923389142, + "grad_norm": 7.09375, + "learning_rate": 9.005580923389143e-07, + "loss": 2.16535454, + "memory(GiB)": 58.09, + "step": 355, + "train_speed(iter/s)": 0.478405 + }, + { + "acc": 0.5720365, + "epoch": 0.0091324200913242, + "grad_norm": 7.53125, + "learning_rate": 9.132420091324201e-07, + "loss": 2.02883244, + "memory(GiB)": 58.09, + "step": 360, + "train_speed(iter/s)": 0.481751 + }, + { + "acc": 0.58192215, + "epoch": 0.009259259259259259, + "grad_norm": 7.9375, + "learning_rate": 9.259259259259259e-07, + "loss": 2.11514702, + "memory(GiB)": 58.09, + "step": 365, + "train_speed(iter/s)": 0.485059 + }, + { + "acc": 0.56529708, + "epoch": 0.009386098427194317, + "grad_norm": 7.71875, + "learning_rate": 9.386098427194318e-07, + "loss": 2.17183685, + "memory(GiB)": 58.09, + "step": 370, + "train_speed(iter/s)": 0.48834 + }, + { + "acc": 0.56914883, + "epoch": 0.009512937595129375, + "grad_norm": 6.5625, + "learning_rate": 9.512937595129377e-07, + "loss": 2.23746357, + "memory(GiB)": 58.09, + "step": 375, + "train_speed(iter/s)": 0.491552 + }, + { + "acc": 0.58338146, + "epoch": 0.009639776763064434, + "grad_norm": 7.9375, + "learning_rate": 9.639776763064435e-07, + "loss": 2.08708191, + "memory(GiB)": 58.09, + "step": 380, + "train_speed(iter/s)": 0.494699 + }, + { + "acc": 0.56976299, + "epoch": 0.009766615930999492, + "grad_norm": 8.3125, + "learning_rate": 9.766615930999493e-07, + "loss": 2.1305254, + "memory(GiB)": 58.09, + "step": 385, + "train_speed(iter/s)": 0.497826 + }, + { + "acc": 0.56847897, + "epoch": 0.00989345509893455, + "grad_norm": 6.75, + "learning_rate": 9.89345509893455e-07, + "loss": 2.1421936, + "memory(GiB)": 58.09, + "step": 390, + "train_speed(iter/s)": 0.500925 + }, + { + "acc": 0.57755151, + "epoch": 0.010020294266869609, + "grad_norm": 6.65625, + "learning_rate": 1.002029426686961e-06, + "loss": 2.1770319, + "memory(GiB)": 58.09, + "step": 395, + "train_speed(iter/s)": 0.503969 + }, + { + "acc": 0.55601387, + "epoch": 0.010147133434804667, + "grad_norm": 7.5, + "learning_rate": 1.0147133434804667e-06, + "loss": 2.20693245, + "memory(GiB)": 58.09, + "step": 400, + "train_speed(iter/s)": 0.506988 + }, + { + "epoch": 0.010147133434804667, + "eval_acc": 0.5676501048387063, + "eval_loss": 2.175985813140869, + "eval_runtime": 113.1084, + "eval_samples_per_second": 56.318, + "eval_steps_per_second": 28.159, + "step": 400 + }, + { + "acc": 0.57561436, + "epoch": 0.010273972602739725, + "grad_norm": 6.65625, + "learning_rate": 1.0273972602739727e-06, + "loss": 2.14138947, + "memory(GiB)": 58.09, + "step": 405, + "train_speed(iter/s)": 0.44419 + }, + { + "acc": 0.56715174, + "epoch": 0.010400811770674784, + "grad_norm": 7.28125, + "learning_rate": 1.0400811770674785e-06, + "loss": 2.16634846, + "memory(GiB)": 58.09, + "step": 410, + "train_speed(iter/s)": 0.447112 + }, + { + "acc": 0.58196907, + "epoch": 0.010527650938609842, + "grad_norm": 6.875, + "learning_rate": 1.0527650938609842e-06, + "loss": 2.06305351, + "memory(GiB)": 58.09, + "step": 415, + "train_speed(iter/s)": 0.450008 + }, + { + "acc": 0.57910528, + "epoch": 0.0106544901065449, + "grad_norm": 7.75, + "learning_rate": 1.0654490106544902e-06, + "loss": 2.12572136, + "memory(GiB)": 58.09, + "step": 420, + "train_speed(iter/s)": 0.452859 + }, + { + "acc": 0.56836276, + "epoch": 0.010781329274479959, + "grad_norm": 6.625, + "learning_rate": 1.078132927447996e-06, + "loss": 2.12446976, + "memory(GiB)": 58.09, + "step": 425, + "train_speed(iter/s)": 0.455683 + }, + { + "acc": 0.5694973, + "epoch": 0.010908168442415017, + "grad_norm": 7.03125, + "learning_rate": 1.0908168442415017e-06, + "loss": 2.12779865, + "memory(GiB)": 58.09, + "step": 430, + "train_speed(iter/s)": 0.458499 + }, + { + "acc": 0.56835136, + "epoch": 0.011035007610350075, + "grad_norm": 6.6875, + "learning_rate": 1.1035007610350077e-06, + "loss": 2.16080875, + "memory(GiB)": 58.09, + "step": 435, + "train_speed(iter/s)": 0.461239 + }, + { + "acc": 0.57685852, + "epoch": 0.011161846778285134, + "grad_norm": 6.625, + "learning_rate": 1.1161846778285134e-06, + "loss": 2.09854145, + "memory(GiB)": 58.09, + "step": 440, + "train_speed(iter/s)": 0.463976 + }, + { + "acc": 0.58293648, + "epoch": 0.011288685946220192, + "grad_norm": 6.125, + "learning_rate": 1.1288685946220194e-06, + "loss": 2.09977322, + "memory(GiB)": 58.09, + "step": 445, + "train_speed(iter/s)": 0.466706 + }, + { + "acc": 0.57144241, + "epoch": 0.01141552511415525, + "grad_norm": 6.65625, + "learning_rate": 1.1415525114155251e-06, + "loss": 2.14928741, + "memory(GiB)": 58.09, + "step": 450, + "train_speed(iter/s)": 0.469369 + }, + { + "acc": 0.57457638, + "epoch": 0.011542364282090309, + "grad_norm": 6.75, + "learning_rate": 1.154236428209031e-06, + "loss": 2.15114937, + "memory(GiB)": 58.09, + "step": 455, + "train_speed(iter/s)": 0.472007 + }, + { + "acc": 0.57902718, + "epoch": 0.011669203450025367, + "grad_norm": 5.8125, + "learning_rate": 1.1669203450025369e-06, + "loss": 2.09303741, + "memory(GiB)": 58.09, + "step": 460, + "train_speed(iter/s)": 0.474606 + }, + { + "acc": 0.57912216, + "epoch": 0.011796042617960426, + "grad_norm": 7.03125, + "learning_rate": 1.1796042617960426e-06, + "loss": 2.16181335, + "memory(GiB)": 58.09, + "step": 465, + "train_speed(iter/s)": 0.47717 + }, + { + "acc": 0.560989, + "epoch": 0.011922881785895484, + "grad_norm": 6.84375, + "learning_rate": 1.1922881785895486e-06, + "loss": 2.25429764, + "memory(GiB)": 58.09, + "step": 470, + "train_speed(iter/s)": 0.479735 + }, + { + "acc": 0.59846892, + "epoch": 0.012049720953830542, + "grad_norm": 5.8125, + "learning_rate": 1.2049720953830543e-06, + "loss": 2.02553291, + "memory(GiB)": 58.09, + "step": 475, + "train_speed(iter/s)": 0.482266 + }, + { + "acc": 0.57369366, + "epoch": 0.0121765601217656, + "grad_norm": 6.96875, + "learning_rate": 1.21765601217656e-06, + "loss": 2.15166779, + "memory(GiB)": 58.09, + "step": 480, + "train_speed(iter/s)": 0.484751 + }, + { + "acc": 0.57437801, + "epoch": 0.012303399289700659, + "grad_norm": 5.0625, + "learning_rate": 1.230339928970066e-06, + "loss": 2.12149315, + "memory(GiB)": 58.09, + "step": 485, + "train_speed(iter/s)": 0.487242 + }, + { + "acc": 0.58520784, + "epoch": 0.012430238457635717, + "grad_norm": 6.0625, + "learning_rate": 1.2430238457635718e-06, + "loss": 2.08534393, + "memory(GiB)": 58.09, + "step": 490, + "train_speed(iter/s)": 0.48969 + }, + { + "acc": 0.5787189, + "epoch": 0.012557077625570776, + "grad_norm": 8.0, + "learning_rate": 1.2557077625570776e-06, + "loss": 2.17281342, + "memory(GiB)": 58.09, + "step": 495, + "train_speed(iter/s)": 0.492143 + }, + { + "acc": 0.57244587, + "epoch": 0.012683916793505834, + "grad_norm": 7.34375, + "learning_rate": 1.2683916793505835e-06, + "loss": 2.08293724, + "memory(GiB)": 58.09, + "step": 500, + "train_speed(iter/s)": 0.49456 + }, + { + "epoch": 0.012683916793505834, + "eval_acc": 0.5691482184104204, + "eval_loss": 2.162841796875, + "eval_runtime": 112.8026, + "eval_samples_per_second": 56.47, + "eval_steps_per_second": 28.235, + "step": 500 + }, + { + "acc": 0.55795674, + "epoch": 0.012810755961440892, + "grad_norm": 7.1875, + "learning_rate": 1.2810755961440893e-06, + "loss": 2.15731926, + "memory(GiB)": 58.09, + "step": 505, + "train_speed(iter/s)": 0.445536 + }, + { + "acc": 0.57902861, + "epoch": 0.01293759512937595, + "grad_norm": 6.46875, + "learning_rate": 1.2937595129375953e-06, + "loss": 2.15045204, + "memory(GiB)": 58.09, + "step": 510, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.57203355, + "epoch": 0.013064434297311009, + "grad_norm": 6.78125, + "learning_rate": 1.306443429731101e-06, + "loss": 2.19813042, + "memory(GiB)": 58.09, + "step": 515, + "train_speed(iter/s)": 0.450192 + }, + { + "acc": 0.5899065, + "epoch": 0.013191273465246067, + "grad_norm": 5.40625, + "learning_rate": 1.3191273465246068e-06, + "loss": 2.06854572, + "memory(GiB)": 58.09, + "step": 520, + "train_speed(iter/s)": 0.452477 + }, + { + "acc": 0.57648339, + "epoch": 0.013318112633181126, + "grad_norm": 6.84375, + "learning_rate": 1.3318112633181127e-06, + "loss": 2.07820473, + "memory(GiB)": 58.09, + "step": 525, + "train_speed(iter/s)": 0.454756 + }, + { + "acc": 0.59465828, + "epoch": 0.013444951801116184, + "grad_norm": 6.71875, + "learning_rate": 1.3444951801116185e-06, + "loss": 2.03313255, + "memory(GiB)": 58.09, + "step": 530, + "train_speed(iter/s)": 0.457003 + }, + { + "acc": 0.58114901, + "epoch": 0.013571790969051242, + "grad_norm": 6.5625, + "learning_rate": 1.3571790969051243e-06, + "loss": 2.05926666, + "memory(GiB)": 58.09, + "step": 535, + "train_speed(iter/s)": 0.459243 + }, + { + "acc": 0.58064313, + "epoch": 0.0136986301369863, + "grad_norm": 6.625, + "learning_rate": 1.3698630136986302e-06, + "loss": 2.15930367, + "memory(GiB)": 58.09, + "step": 540, + "train_speed(iter/s)": 0.46145 + }, + { + "acc": 0.56529388, + "epoch": 0.013825469304921359, + "grad_norm": 6.5, + "learning_rate": 1.382546930492136e-06, + "loss": 2.17727165, + "memory(GiB)": 58.09, + "step": 545, + "train_speed(iter/s)": 0.463659 + }, + { + "acc": 0.56990838, + "epoch": 0.013952308472856417, + "grad_norm": 7.0625, + "learning_rate": 1.395230847285642e-06, + "loss": 2.1872509, + "memory(GiB)": 58.09, + "step": 550, + "train_speed(iter/s)": 0.465837 + }, + { + "acc": 0.57715831, + "epoch": 0.014079147640791476, + "grad_norm": 6.4375, + "learning_rate": 1.4079147640791477e-06, + "loss": 2.12239723, + "memory(GiB)": 58.09, + "step": 555, + "train_speed(iter/s)": 0.467989 + }, + { + "acc": 0.58138347, + "epoch": 0.014205986808726534, + "grad_norm": 6.53125, + "learning_rate": 1.4205986808726534e-06, + "loss": 2.06029472, + "memory(GiB)": 58.09, + "step": 560, + "train_speed(iter/s)": 0.470119 + }, + { + "acc": 0.57148027, + "epoch": 0.014332825976661592, + "grad_norm": 7.875, + "learning_rate": 1.4332825976661594e-06, + "loss": 2.17127838, + "memory(GiB)": 58.09, + "step": 565, + "train_speed(iter/s)": 0.472268 + }, + { + "acc": 0.5769032, + "epoch": 0.01445966514459665, + "grad_norm": 8.6875, + "learning_rate": 1.4459665144596652e-06, + "loss": 2.07768211, + "memory(GiB)": 58.09, + "step": 570, + "train_speed(iter/s)": 0.474374 + }, + { + "acc": 0.59854722, + "epoch": 0.014586504312531709, + "grad_norm": 6.28125, + "learning_rate": 1.458650431253171e-06, + "loss": 2.02412167, + "memory(GiB)": 58.09, + "step": 575, + "train_speed(iter/s)": 0.476438 + }, + { + "acc": 0.57099662, + "epoch": 0.014713343480466767, + "grad_norm": 6.90625, + "learning_rate": 1.4713343480466769e-06, + "loss": 2.21544037, + "memory(GiB)": 58.09, + "step": 580, + "train_speed(iter/s)": 0.478514 + }, + { + "acc": 0.57785525, + "epoch": 0.014840182648401826, + "grad_norm": 7.125, + "learning_rate": 1.4840182648401826e-06, + "loss": 2.17163811, + "memory(GiB)": 58.09, + "step": 585, + "train_speed(iter/s)": 0.480592 + }, + { + "acc": 0.55674849, + "epoch": 0.014967021816336884, + "grad_norm": 5.59375, + "learning_rate": 1.4967021816336886e-06, + "loss": 2.1069252, + "memory(GiB)": 58.09, + "step": 590, + "train_speed(iter/s)": 0.482607 + }, + { + "acc": 0.56835804, + "epoch": 0.015093860984271942, + "grad_norm": 5.75, + "learning_rate": 1.5093860984271944e-06, + "loss": 2.12497196, + "memory(GiB)": 58.09, + "step": 595, + "train_speed(iter/s)": 0.484623 + }, + { + "acc": 0.59031286, + "epoch": 0.015220700152207, + "grad_norm": 5.6875, + "learning_rate": 1.5220700152207001e-06, + "loss": 2.06097107, + "memory(GiB)": 58.09, + "step": 600, + "train_speed(iter/s)": 0.486626 + }, + { + "epoch": 0.015220700152207, + "eval_acc": 0.5708752684677172, + "eval_loss": 2.146491527557373, + "eval_runtime": 113.4805, + "eval_samples_per_second": 56.133, + "eval_steps_per_second": 28.066, + "step": 600 + }, + { + "acc": 0.5907094, + "epoch": 0.015347539320142059, + "grad_norm": 6.09375, + "learning_rate": 1.534753932014206e-06, + "loss": 2.03423805, + "memory(GiB)": 58.09, + "step": 605, + "train_speed(iter/s)": 0.446108 + }, + { + "acc": 0.57663136, + "epoch": 0.015474378488077117, + "grad_norm": 5.875, + "learning_rate": 1.5474378488077118e-06, + "loss": 2.08584633, + "memory(GiB)": 58.09, + "step": 610, + "train_speed(iter/s)": 0.448062 + }, + { + "acc": 0.57774086, + "epoch": 0.015601217656012176, + "grad_norm": 6.21875, + "learning_rate": 1.5601217656012176e-06, + "loss": 2.09043198, + "memory(GiB)": 58.09, + "step": 615, + "train_speed(iter/s)": 0.449998 + }, + { + "acc": 0.58956842, + "epoch": 0.015728056823947234, + "grad_norm": 5.96875, + "learning_rate": 1.5728056823947236e-06, + "loss": 2.00049877, + "memory(GiB)": 58.09, + "step": 620, + "train_speed(iter/s)": 0.451934 + }, + { + "acc": 0.57578068, + "epoch": 0.015854895991882292, + "grad_norm": 5.6875, + "learning_rate": 1.5854895991882293e-06, + "loss": 2.10935745, + "memory(GiB)": 58.09, + "step": 625, + "train_speed(iter/s)": 0.453847 + }, + { + "acc": 0.58070493, + "epoch": 0.01598173515981735, + "grad_norm": 6.1875, + "learning_rate": 1.5981735159817353e-06, + "loss": 2.12280045, + "memory(GiB)": 58.09, + "step": 630, + "train_speed(iter/s)": 0.455742 + }, + { + "acc": 0.58392453, + "epoch": 0.01610857432775241, + "grad_norm": 6.78125, + "learning_rate": 1.610857432775241e-06, + "loss": 2.10500374, + "memory(GiB)": 58.09, + "step": 635, + "train_speed(iter/s)": 0.457624 + }, + { + "acc": 0.56682339, + "epoch": 0.016235413495687467, + "grad_norm": 6.6875, + "learning_rate": 1.6235413495687468e-06, + "loss": 2.12986374, + "memory(GiB)": 58.09, + "step": 640, + "train_speed(iter/s)": 0.459504 + }, + { + "acc": 0.58058124, + "epoch": 0.016362252663622526, + "grad_norm": 5.9375, + "learning_rate": 1.6362252663622528e-06, + "loss": 2.14026375, + "memory(GiB)": 58.09, + "step": 645, + "train_speed(iter/s)": 0.461349 + }, + { + "acc": 0.56498117, + "epoch": 0.016489091831557584, + "grad_norm": 5.71875, + "learning_rate": 1.6489091831557585e-06, + "loss": 2.10380497, + "memory(GiB)": 58.09, + "step": 650, + "train_speed(iter/s)": 0.463194 + }, + { + "acc": 0.57264342, + "epoch": 0.016615930999492642, + "grad_norm": 5.4375, + "learning_rate": 1.6615930999492643e-06, + "loss": 2.15320625, + "memory(GiB)": 58.09, + "step": 655, + "train_speed(iter/s)": 0.465021 + }, + { + "acc": 0.57095966, + "epoch": 0.0167427701674277, + "grad_norm": 6.0, + "learning_rate": 1.6742770167427702e-06, + "loss": 2.06447887, + "memory(GiB)": 58.09, + "step": 660, + "train_speed(iter/s)": 0.466831 + }, + { + "acc": 0.58545179, + "epoch": 0.01686960933536276, + "grad_norm": 5.46875, + "learning_rate": 1.686960933536276e-06, + "loss": 2.01951256, + "memory(GiB)": 58.09, + "step": 665, + "train_speed(iter/s)": 0.468608 + }, + { + "acc": 0.58088417, + "epoch": 0.016996448503297817, + "grad_norm": 5.84375, + "learning_rate": 1.699644850329782e-06, + "loss": 2.08067741, + "memory(GiB)": 58.1, + "step": 670, + "train_speed(iter/s)": 0.470386 + }, + { + "acc": 0.58742523, + "epoch": 0.017123287671232876, + "grad_norm": 6.65625, + "learning_rate": 1.7123287671232877e-06, + "loss": 2.06641312, + "memory(GiB)": 58.1, + "step": 675, + "train_speed(iter/s)": 0.47215 + }, + { + "acc": 0.5650301, + "epoch": 0.017250126839167934, + "grad_norm": 5.96875, + "learning_rate": 1.7250126839167935e-06, + "loss": 2.19955997, + "memory(GiB)": 58.1, + "step": 680, + "train_speed(iter/s)": 0.47392 + }, + { + "acc": 0.58013334, + "epoch": 0.017376966007102992, + "grad_norm": 5.84375, + "learning_rate": 1.7376966007102994e-06, + "loss": 2.05740891, + "memory(GiB)": 58.1, + "step": 685, + "train_speed(iter/s)": 0.475649 + }, + { + "acc": 0.58957648, + "epoch": 0.01750380517503805, + "grad_norm": 5.9375, + "learning_rate": 1.7503805175038052e-06, + "loss": 2.12545013, + "memory(GiB)": 58.1, + "step": 690, + "train_speed(iter/s)": 0.477391 + }, + { + "acc": 0.57258396, + "epoch": 0.01763064434297311, + "grad_norm": 7.0625, + "learning_rate": 1.7630644342973112e-06, + "loss": 2.08827343, + "memory(GiB)": 58.1, + "step": 695, + "train_speed(iter/s)": 0.479123 + }, + { + "acc": 0.57630472, + "epoch": 0.017757483510908167, + "grad_norm": 5.46875, + "learning_rate": 1.775748351090817e-06, + "loss": 2.08242302, + "memory(GiB)": 58.1, + "step": 700, + "train_speed(iter/s)": 0.480823 + }, + { + "epoch": 0.017757483510908167, + "eval_acc": 0.572730990856327, + "eval_loss": 2.130608320236206, + "eval_runtime": 114.9834, + "eval_samples_per_second": 55.399, + "eval_steps_per_second": 27.7, + "step": 700 + }, + { + "acc": 0.56871805, + "epoch": 0.017884322678843226, + "grad_norm": 5.8125, + "learning_rate": 1.7884322678843227e-06, + "loss": 2.08843346, + "memory(GiB)": 58.1, + "step": 705, + "train_speed(iter/s)": 0.446048 + }, + { + "acc": 0.56865363, + "epoch": 0.018011161846778284, + "grad_norm": 5.6875, + "learning_rate": 1.8011161846778286e-06, + "loss": 2.12287178, + "memory(GiB)": 58.1, + "step": 710, + "train_speed(iter/s)": 0.447716 + }, + { + "acc": 0.5653419, + "epoch": 0.018138001014713342, + "grad_norm": 5.375, + "learning_rate": 1.8138001014713344e-06, + "loss": 2.1662447, + "memory(GiB)": 58.1, + "step": 715, + "train_speed(iter/s)": 0.44938 + }, + { + "acc": 0.58151999, + "epoch": 0.0182648401826484, + "grad_norm": 6.4375, + "learning_rate": 1.8264840182648401e-06, + "loss": 2.08704548, + "memory(GiB)": 58.1, + "step": 720, + "train_speed(iter/s)": 0.451045 + }, + { + "acc": 0.565307, + "epoch": 0.01839167935058346, + "grad_norm": 6.0, + "learning_rate": 1.8391679350583461e-06, + "loss": 2.13599167, + "memory(GiB)": 58.1, + "step": 725, + "train_speed(iter/s)": 0.452697 + }, + { + "acc": 0.578827, + "epoch": 0.018518518518518517, + "grad_norm": 6.125, + "learning_rate": 1.8518518518518519e-06, + "loss": 2.00157909, + "memory(GiB)": 58.1, + "step": 730, + "train_speed(iter/s)": 0.454335 + }, + { + "acc": 0.56656232, + "epoch": 0.018645357686453576, + "grad_norm": 6.1875, + "learning_rate": 1.8645357686453578e-06, + "loss": 2.1325592, + "memory(GiB)": 58.1, + "step": 735, + "train_speed(iter/s)": 0.455969 + }, + { + "acc": 0.57751007, + "epoch": 0.018772196854388634, + "grad_norm": 7.09375, + "learning_rate": 1.8772196854388636e-06, + "loss": 2.12488174, + "memory(GiB)": 58.1, + "step": 740, + "train_speed(iter/s)": 0.45759 + }, + { + "acc": 0.584062, + "epoch": 0.018899036022323693, + "grad_norm": 5.4375, + "learning_rate": 1.8899036022323693e-06, + "loss": 2.02887325, + "memory(GiB)": 58.1, + "step": 745, + "train_speed(iter/s)": 0.459196 + }, + { + "acc": 0.57995129, + "epoch": 0.01902587519025875, + "grad_norm": 5.78125, + "learning_rate": 1.9025875190258753e-06, + "loss": 2.07564201, + "memory(GiB)": 58.1, + "step": 750, + "train_speed(iter/s)": 0.460792 + }, + { + "acc": 0.59045911, + "epoch": 0.01915271435819381, + "grad_norm": 5.75, + "learning_rate": 1.915271435819381e-06, + "loss": 2.05422363, + "memory(GiB)": 58.1, + "step": 755, + "train_speed(iter/s)": 0.462376 + }, + { + "acc": 0.56149321, + "epoch": 0.019279553526128868, + "grad_norm": 5.3125, + "learning_rate": 1.927955352612887e-06, + "loss": 2.08703136, + "memory(GiB)": 58.1, + "step": 760, + "train_speed(iter/s)": 0.463931 + }, + { + "acc": 0.59644594, + "epoch": 0.019406392694063926, + "grad_norm": 6.25, + "learning_rate": 1.9406392694063926e-06, + "loss": 2.12244968, + "memory(GiB)": 58.1, + "step": 765, + "train_speed(iter/s)": 0.465481 + }, + { + "acc": 0.57373838, + "epoch": 0.019533231861998984, + "grad_norm": 6.4375, + "learning_rate": 1.9533231861998985e-06, + "loss": 2.09988823, + "memory(GiB)": 58.1, + "step": 770, + "train_speed(iter/s)": 0.467046 + }, + { + "acc": 0.58926535, + "epoch": 0.019660071029934043, + "grad_norm": 5.65625, + "learning_rate": 1.9660071029934045e-06, + "loss": 2.0737978, + "memory(GiB)": 58.1, + "step": 775, + "train_speed(iter/s)": 0.468568 + }, + { + "acc": 0.55917034, + "epoch": 0.0197869101978691, + "grad_norm": 6.15625, + "learning_rate": 1.97869101978691e-06, + "loss": 2.16007786, + "memory(GiB)": 58.1, + "step": 780, + "train_speed(iter/s)": 0.470109 + }, + { + "acc": 0.59510069, + "epoch": 0.01991374936580416, + "grad_norm": 6.125, + "learning_rate": 1.991374936580416e-06, + "loss": 1.95591621, + "memory(GiB)": 58.1, + "step": 785, + "train_speed(iter/s)": 0.471617 + }, + { + "acc": 0.58288422, + "epoch": 0.020040588533739218, + "grad_norm": 6.40625, + "learning_rate": 2.004058853373922e-06, + "loss": 2.06537952, + "memory(GiB)": 58.1, + "step": 790, + "train_speed(iter/s)": 0.473123 + }, + { + "acc": 0.58917351, + "epoch": 0.020167427701674276, + "grad_norm": 7.40625, + "learning_rate": 2.016742770167428e-06, + "loss": 2.00254402, + "memory(GiB)": 58.1, + "step": 795, + "train_speed(iter/s)": 0.474627 + }, + { + "acc": 0.59073582, + "epoch": 0.020294266869609334, + "grad_norm": 5.84375, + "learning_rate": 2.0294266869609335e-06, + "loss": 1.95720024, + "memory(GiB)": 58.1, + "step": 800, + "train_speed(iter/s)": 0.476108 + }, + { + "epoch": 0.020294266869609334, + "eval_acc": 0.5743732341498038, + "eval_loss": 2.1150147914886475, + "eval_runtime": 111.9223, + "eval_samples_per_second": 56.914, + "eval_steps_per_second": 28.457, + "step": 800 + }, + { + "acc": 0.57163839, + "epoch": 0.020421106037544393, + "grad_norm": 6.34375, + "learning_rate": 2.0421106037544395e-06, + "loss": 2.08774834, + "memory(GiB)": 58.1, + "step": 805, + "train_speed(iter/s)": 0.446744 + }, + { + "acc": 0.58671284, + "epoch": 0.02054794520547945, + "grad_norm": 6.46875, + "learning_rate": 2.0547945205479454e-06, + "loss": 2.00996399, + "memory(GiB)": 58.1, + "step": 810, + "train_speed(iter/s)": 0.448222 + }, + { + "acc": 0.59209747, + "epoch": 0.02067478437341451, + "grad_norm": 6.625, + "learning_rate": 2.067478437341451e-06, + "loss": 2.11006374, + "memory(GiB)": 58.1, + "step": 815, + "train_speed(iter/s)": 0.449683 + }, + { + "acc": 0.58112731, + "epoch": 0.020801623541349568, + "grad_norm": 6.53125, + "learning_rate": 2.080162354134957e-06, + "loss": 2.04138203, + "memory(GiB)": 58.1, + "step": 820, + "train_speed(iter/s)": 0.451127 + }, + { + "acc": 0.56349602, + "epoch": 0.020928462709284626, + "grad_norm": 5.4375, + "learning_rate": 2.092846270928463e-06, + "loss": 2.10374203, + "memory(GiB)": 58.1, + "step": 825, + "train_speed(iter/s)": 0.452565 + }, + { + "acc": 0.55964966, + "epoch": 0.021055301877219684, + "grad_norm": 6.28125, + "learning_rate": 2.1055301877219685e-06, + "loss": 2.1757328, + "memory(GiB)": 58.1, + "step": 830, + "train_speed(iter/s)": 0.454006 + }, + { + "acc": 0.58940411, + "epoch": 0.021182141045154743, + "grad_norm": 4.78125, + "learning_rate": 2.1182141045154744e-06, + "loss": 1.98896866, + "memory(GiB)": 58.1, + "step": 835, + "train_speed(iter/s)": 0.455428 + }, + { + "acc": 0.58555031, + "epoch": 0.0213089802130898, + "grad_norm": 7.0625, + "learning_rate": 2.1308980213089804e-06, + "loss": 2.03969383, + "memory(GiB)": 58.1, + "step": 840, + "train_speed(iter/s)": 0.456852 + }, + { + "acc": 0.57818289, + "epoch": 0.02143581938102486, + "grad_norm": 5.84375, + "learning_rate": 2.143581938102486e-06, + "loss": 2.02093258, + "memory(GiB)": 58.1, + "step": 845, + "train_speed(iter/s)": 0.458259 + }, + { + "acc": 0.60250626, + "epoch": 0.021562658548959918, + "grad_norm": 5.8125, + "learning_rate": 2.156265854895992e-06, + "loss": 2.05443401, + "memory(GiB)": 58.1, + "step": 850, + "train_speed(iter/s)": 0.459671 + }, + { + "acc": 0.56007051, + "epoch": 0.021689497716894976, + "grad_norm": 6.09375, + "learning_rate": 2.168949771689498e-06, + "loss": 2.12000847, + "memory(GiB)": 58.1, + "step": 855, + "train_speed(iter/s)": 0.461072 + }, + { + "acc": 0.58610115, + "epoch": 0.021816336884830034, + "grad_norm": 5.8125, + "learning_rate": 2.1816336884830034e-06, + "loss": 2.08215179, + "memory(GiB)": 58.1, + "step": 860, + "train_speed(iter/s)": 0.462449 + }, + { + "acc": 0.55657902, + "epoch": 0.021943176052765093, + "grad_norm": 7.34375, + "learning_rate": 2.1943176052765094e-06, + "loss": 2.07644501, + "memory(GiB)": 58.1, + "step": 865, + "train_speed(iter/s)": 0.463819 + }, + { + "acc": 0.58080263, + "epoch": 0.02207001522070015, + "grad_norm": 6.71875, + "learning_rate": 2.2070015220700153e-06, + "loss": 2.1232357, + "memory(GiB)": 58.1, + "step": 870, + "train_speed(iter/s)": 0.465168 + }, + { + "acc": 0.57364683, + "epoch": 0.02219685438863521, + "grad_norm": 6.53125, + "learning_rate": 2.2196854388635213e-06, + "loss": 2.08617592, + "memory(GiB)": 58.1, + "step": 875, + "train_speed(iter/s)": 0.466524 + }, + { + "acc": 0.59218321, + "epoch": 0.022323693556570268, + "grad_norm": 7.15625, + "learning_rate": 2.232369355657027e-06, + "loss": 2.02167473, + "memory(GiB)": 58.1, + "step": 880, + "train_speed(iter/s)": 0.467857 + }, + { + "acc": 0.57566214, + "epoch": 0.022450532724505326, + "grad_norm": 6.78125, + "learning_rate": 2.245053272450533e-06, + "loss": 2.10848694, + "memory(GiB)": 58.1, + "step": 885, + "train_speed(iter/s)": 0.4692 + }, + { + "acc": 0.57690935, + "epoch": 0.022577371892440384, + "grad_norm": 6.34375, + "learning_rate": 2.2577371892440388e-06, + "loss": 2.10077057, + "memory(GiB)": 58.1, + "step": 890, + "train_speed(iter/s)": 0.470556 + }, + { + "acc": 0.59474163, + "epoch": 0.022704211060375443, + "grad_norm": 4.875, + "learning_rate": 2.2704211060375443e-06, + "loss": 2.05679932, + "memory(GiB)": 58.1, + "step": 895, + "train_speed(iter/s)": 0.471877 + }, + { + "acc": 0.59358177, + "epoch": 0.0228310502283105, + "grad_norm": 6.375, + "learning_rate": 2.2831050228310503e-06, + "loss": 2.07524223, + "memory(GiB)": 58.1, + "step": 900, + "train_speed(iter/s)": 0.473181 + }, + { + "epoch": 0.0228310502283105, + "eval_acc": 0.5761036263455763, + "eval_loss": 2.0988223552703857, + "eval_runtime": 113.9568, + "eval_samples_per_second": 55.898, + "eval_steps_per_second": 27.949, + "step": 900 + }, + { + "acc": 0.57166119, + "epoch": 0.02295788939624556, + "grad_norm": 5.625, + "learning_rate": 2.2957889396245563e-06, + "loss": 2.08735542, + "memory(GiB)": 58.1, + "step": 905, + "train_speed(iter/s)": 0.446743 + }, + { + "acc": 0.57747388, + "epoch": 0.023084728564180618, + "grad_norm": 6.03125, + "learning_rate": 2.308472856418062e-06, + "loss": 2.07981911, + "memory(GiB)": 58.1, + "step": 910, + "train_speed(iter/s)": 0.448036 + }, + { + "acc": 0.58042278, + "epoch": 0.023211567732115676, + "grad_norm": 5.0625, + "learning_rate": 2.3211567732115678e-06, + "loss": 2.02468204, + "memory(GiB)": 58.1, + "step": 915, + "train_speed(iter/s)": 0.449314 + }, + { + "acc": 0.61442428, + "epoch": 0.023338406900050734, + "grad_norm": 4.875, + "learning_rate": 2.3338406900050737e-06, + "loss": 1.94776154, + "memory(GiB)": 58.1, + "step": 920, + "train_speed(iter/s)": 0.450587 + }, + { + "acc": 0.57519131, + "epoch": 0.023465246067985793, + "grad_norm": 5.21875, + "learning_rate": 2.3465246067985793e-06, + "loss": 2.1024704, + "memory(GiB)": 58.1, + "step": 925, + "train_speed(iter/s)": 0.451863 + }, + { + "acc": 0.58424702, + "epoch": 0.02359208523592085, + "grad_norm": 8.25, + "learning_rate": 2.3592085235920852e-06, + "loss": 2.06277103, + "memory(GiB)": 58.1, + "step": 930, + "train_speed(iter/s)": 0.453137 + }, + { + "acc": 0.58619661, + "epoch": 0.02371892440385591, + "grad_norm": 6.71875, + "learning_rate": 2.371892440385591e-06, + "loss": 2.02461262, + "memory(GiB)": 58.1, + "step": 935, + "train_speed(iter/s)": 0.454405 + }, + { + "acc": 0.59015274, + "epoch": 0.023845763571790968, + "grad_norm": 5.9375, + "learning_rate": 2.384576357179097e-06, + "loss": 2.0354763, + "memory(GiB)": 58.1, + "step": 940, + "train_speed(iter/s)": 0.455673 + }, + { + "acc": 0.58388462, + "epoch": 0.023972602739726026, + "grad_norm": 6.625, + "learning_rate": 2.3972602739726027e-06, + "loss": 2.02807159, + "memory(GiB)": 58.1, + "step": 945, + "train_speed(iter/s)": 0.456935 + }, + { + "acc": 0.57431579, + "epoch": 0.024099441907661084, + "grad_norm": 7.125, + "learning_rate": 2.4099441907661087e-06, + "loss": 2.1198658, + "memory(GiB)": 58.1, + "step": 950, + "train_speed(iter/s)": 0.458186 + }, + { + "acc": 0.60085821, + "epoch": 0.024226281075596143, + "grad_norm": 7.03125, + "learning_rate": 2.4226281075596147e-06, + "loss": 2.03528137, + "memory(GiB)": 58.1, + "step": 955, + "train_speed(iter/s)": 0.459424 + }, + { + "acc": 0.5809402, + "epoch": 0.0243531202435312, + "grad_norm": 6.96875, + "learning_rate": 2.43531202435312e-06, + "loss": 2.10258236, + "memory(GiB)": 58.1, + "step": 960, + "train_speed(iter/s)": 0.460649 + }, + { + "acc": 0.57716618, + "epoch": 0.02447995941146626, + "grad_norm": 5.84375, + "learning_rate": 2.447995941146626e-06, + "loss": 2.13247681, + "memory(GiB)": 58.1, + "step": 965, + "train_speed(iter/s)": 0.461872 + }, + { + "acc": 0.58431964, + "epoch": 0.024606798579401318, + "grad_norm": 5.46875, + "learning_rate": 2.460679857940132e-06, + "loss": 2.08380051, + "memory(GiB)": 58.1, + "step": 970, + "train_speed(iter/s)": 0.463086 + }, + { + "acc": 0.57374287, + "epoch": 0.024733637747336376, + "grad_norm": 5.84375, + "learning_rate": 2.4733637747336377e-06, + "loss": 2.13242855, + "memory(GiB)": 58.1, + "step": 975, + "train_speed(iter/s)": 0.464288 + }, + { + "acc": 0.60974894, + "epoch": 0.024860476915271434, + "grad_norm": 5.625, + "learning_rate": 2.4860476915271436e-06, + "loss": 2.00461884, + "memory(GiB)": 58.1, + "step": 980, + "train_speed(iter/s)": 0.465491 + }, + { + "acc": 0.5824585, + "epoch": 0.024987316083206493, + "grad_norm": 5.78125, + "learning_rate": 2.4987316083206496e-06, + "loss": 2.07370567, + "memory(GiB)": 58.1, + "step": 985, + "train_speed(iter/s)": 0.466688 + }, + { + "acc": 0.57614508, + "epoch": 0.02511415525114155, + "grad_norm": 5.84375, + "learning_rate": 2.511415525114155e-06, + "loss": 2.12399292, + "memory(GiB)": 58.1, + "step": 990, + "train_speed(iter/s)": 0.467883 + }, + { + "acc": 0.57358203, + "epoch": 0.02524099441907661, + "grad_norm": 5.53125, + "learning_rate": 2.5240994419076615e-06, + "loss": 2.0714653, + "memory(GiB)": 58.1, + "step": 995, + "train_speed(iter/s)": 0.469076 + }, + { + "acc": 0.59364719, + "epoch": 0.025367833587011668, + "grad_norm": 7.4375, + "learning_rate": 2.536783358701167e-06, + "loss": 1.98752384, + "memory(GiB)": 58.1, + "step": 1000, + "train_speed(iter/s)": 0.470267 + }, + { + "epoch": 0.025367833587011668, + "eval_acc": 0.5777345899216979, + "eval_loss": 2.080941915512085, + "eval_runtime": 113.5409, + "eval_samples_per_second": 56.103, + "eval_steps_per_second": 28.052, + "step": 1000 + }, + { + "acc": 0.57457829, + "epoch": 0.025494672754946726, + "grad_norm": 6.28125, + "learning_rate": 2.549467275494673e-06, + "loss": 2.09337349, + "memory(GiB)": 58.1, + "step": 1005, + "train_speed(iter/s)": 0.446718 + }, + { + "acc": 0.5866128, + "epoch": 0.025621511922881784, + "grad_norm": 6.15625, + "learning_rate": 2.5621511922881786e-06, + "loss": 2.13685551, + "memory(GiB)": 58.1, + "step": 1010, + "train_speed(iter/s)": 0.447886 + }, + { + "acc": 0.57901697, + "epoch": 0.025748351090816843, + "grad_norm": 6.34375, + "learning_rate": 2.5748351090816846e-06, + "loss": 2.09885941, + "memory(GiB)": 58.1, + "step": 1015, + "train_speed(iter/s)": 0.449043 + }, + { + "acc": 0.58588743, + "epoch": 0.0258751902587519, + "grad_norm": 6.90625, + "learning_rate": 2.5875190258751905e-06, + "loss": 2.06828918, + "memory(GiB)": 58.1, + "step": 1020, + "train_speed(iter/s)": 0.450205 + }, + { + "acc": 0.59164329, + "epoch": 0.02600202942668696, + "grad_norm": 5.71875, + "learning_rate": 2.6002029426686965e-06, + "loss": 2.04384689, + "memory(GiB)": 58.1, + "step": 1025, + "train_speed(iter/s)": 0.451354 + }, + { + "acc": 0.58611956, + "epoch": 0.026128868594622018, + "grad_norm": 5.03125, + "learning_rate": 2.612886859462202e-06, + "loss": 2.01312637, + "memory(GiB)": 58.1, + "step": 1030, + "train_speed(iter/s)": 0.452493 + }, + { + "acc": 0.59721603, + "epoch": 0.026255707762557076, + "grad_norm": 5.5625, + "learning_rate": 2.625570776255708e-06, + "loss": 2.07603264, + "memory(GiB)": 58.1, + "step": 1035, + "train_speed(iter/s)": 0.453636 + }, + { + "acc": 0.58024497, + "epoch": 0.026382546930492135, + "grad_norm": 7.65625, + "learning_rate": 2.6382546930492135e-06, + "loss": 2.05065346, + "memory(GiB)": 58.1, + "step": 1040, + "train_speed(iter/s)": 0.454773 + }, + { + "acc": 0.59431887, + "epoch": 0.026509386098427193, + "grad_norm": 5.6875, + "learning_rate": 2.65093860984272e-06, + "loss": 2.10402222, + "memory(GiB)": 58.1, + "step": 1045, + "train_speed(iter/s)": 0.455887 + }, + { + "acc": 0.58888636, + "epoch": 0.02663622526636225, + "grad_norm": 6.34375, + "learning_rate": 2.6636225266362255e-06, + "loss": 2.07406387, + "memory(GiB)": 58.1, + "step": 1050, + "train_speed(iter/s)": 0.457006 + }, + { + "acc": 0.58189731, + "epoch": 0.02676306443429731, + "grad_norm": 6.375, + "learning_rate": 2.6763064434297314e-06, + "loss": 2.02127075, + "memory(GiB)": 58.1, + "step": 1055, + "train_speed(iter/s)": 0.45812 + }, + { + "acc": 0.58191009, + "epoch": 0.026889903602232368, + "grad_norm": 5.125, + "learning_rate": 2.688990360223237e-06, + "loss": 2.05131874, + "memory(GiB)": 58.1, + "step": 1060, + "train_speed(iter/s)": 0.459224 + }, + { + "acc": 0.59165268, + "epoch": 0.027016742770167426, + "grad_norm": 5.21875, + "learning_rate": 2.701674277016743e-06, + "loss": 2.0028265, + "memory(GiB)": 58.1, + "step": 1065, + "train_speed(iter/s)": 0.460323 + }, + { + "acc": 0.58109169, + "epoch": 0.027143581938102485, + "grad_norm": 6.375, + "learning_rate": 2.7143581938102485e-06, + "loss": 2.07121277, + "memory(GiB)": 58.1, + "step": 1070, + "train_speed(iter/s)": 0.461418 + }, + { + "acc": 0.58962994, + "epoch": 0.027270421106037543, + "grad_norm": 5.78125, + "learning_rate": 2.727042110603755e-06, + "loss": 2.03678818, + "memory(GiB)": 58.1, + "step": 1075, + "train_speed(iter/s)": 0.462512 + }, + { + "acc": 0.56551304, + "epoch": 0.0273972602739726, + "grad_norm": 5.5, + "learning_rate": 2.7397260273972604e-06, + "loss": 2.16976528, + "memory(GiB)": 58.1, + "step": 1080, + "train_speed(iter/s)": 0.463596 + }, + { + "acc": 0.60050197, + "epoch": 0.02752409944190766, + "grad_norm": 5.8125, + "learning_rate": 2.7524099441907664e-06, + "loss": 2.01963692, + "memory(GiB)": 58.1, + "step": 1085, + "train_speed(iter/s)": 0.464681 + }, + { + "acc": 0.60033646, + "epoch": 0.027650938609842718, + "grad_norm": 5.15625, + "learning_rate": 2.765093860984272e-06, + "loss": 1.98261986, + "memory(GiB)": 58.1, + "step": 1090, + "train_speed(iter/s)": 0.465758 + }, + { + "acc": 0.57258253, + "epoch": 0.027777777777777776, + "grad_norm": 6.15625, + "learning_rate": 2.7777777777777783e-06, + "loss": 2.06698036, + "memory(GiB)": 58.1, + "step": 1095, + "train_speed(iter/s)": 0.466842 + }, + { + "acc": 0.58307362, + "epoch": 0.027904616945712835, + "grad_norm": 6.5625, + "learning_rate": 2.790461694571284e-06, + "loss": 2.00427818, + "memory(GiB)": 58.1, + "step": 1100, + "train_speed(iter/s)": 0.467911 + }, + { + "epoch": 0.027904616945712835, + "eval_acc": 0.579531407119674, + "eval_loss": 2.0623281002044678, + "eval_runtime": 112.269, + "eval_samples_per_second": 56.739, + "eval_steps_per_second": 28.369, + "step": 1100 + }, + { + "acc": 0.58611879, + "epoch": 0.028031456113647893, + "grad_norm": 6.0625, + "learning_rate": 2.80314561136479e-06, + "loss": 2.02119904, + "memory(GiB)": 58.1, + "step": 1105, + "train_speed(iter/s)": 0.44684 + }, + { + "acc": 0.59812632, + "epoch": 0.02815829528158295, + "grad_norm": 6.5, + "learning_rate": 2.8158295281582954e-06, + "loss": 2.00622635, + "memory(GiB)": 58.1, + "step": 1110, + "train_speed(iter/s)": 0.447909 + }, + { + "acc": 0.5802361, + "epoch": 0.02828513444951801, + "grad_norm": 5.78125, + "learning_rate": 2.8285134449518014e-06, + "loss": 2.09116268, + "memory(GiB)": 58.1, + "step": 1115, + "train_speed(iter/s)": 0.448971 + }, + { + "acc": 0.57350545, + "epoch": 0.028411973617453068, + "grad_norm": 5.6875, + "learning_rate": 2.841197361745307e-06, + "loss": 2.09282684, + "memory(GiB)": 58.1, + "step": 1120, + "train_speed(iter/s)": 0.450021 + }, + { + "acc": 0.57460332, + "epoch": 0.028538812785388126, + "grad_norm": 5.5, + "learning_rate": 2.8538812785388133e-06, + "loss": 2.04692535, + "memory(GiB)": 58.1, + "step": 1125, + "train_speed(iter/s)": 0.45108 + }, + { + "acc": 0.58731594, + "epoch": 0.028665651953323185, + "grad_norm": 6.3125, + "learning_rate": 2.866565195332319e-06, + "loss": 2.01711597, + "memory(GiB)": 58.1, + "step": 1130, + "train_speed(iter/s)": 0.452136 + }, + { + "acc": 0.59110603, + "epoch": 0.028792491121258243, + "grad_norm": 5.03125, + "learning_rate": 2.879249112125825e-06, + "loss": 2.01370506, + "memory(GiB)": 58.1, + "step": 1135, + "train_speed(iter/s)": 0.453184 + }, + { + "acc": 0.58467035, + "epoch": 0.0289193302891933, + "grad_norm": 5.78125, + "learning_rate": 2.8919330289193303e-06, + "loss": 2.07677879, + "memory(GiB)": 58.1, + "step": 1140, + "train_speed(iter/s)": 0.454223 + }, + { + "acc": 0.58224087, + "epoch": 0.02904616945712836, + "grad_norm": 5.09375, + "learning_rate": 2.9046169457128363e-06, + "loss": 2.03106861, + "memory(GiB)": 58.1, + "step": 1145, + "train_speed(iter/s)": 0.455258 + }, + { + "acc": 0.58103757, + "epoch": 0.029173008625063418, + "grad_norm": 5.90625, + "learning_rate": 2.917300862506342e-06, + "loss": 2.06879482, + "memory(GiB)": 58.1, + "step": 1150, + "train_speed(iter/s)": 0.456296 + }, + { + "acc": 0.57492504, + "epoch": 0.029299847792998476, + "grad_norm": 5.84375, + "learning_rate": 2.9299847792998482e-06, + "loss": 2.07535973, + "memory(GiB)": 58.1, + "step": 1155, + "train_speed(iter/s)": 0.457343 + }, + { + "acc": 0.60014844, + "epoch": 0.029426686960933535, + "grad_norm": 6.25, + "learning_rate": 2.9426686960933538e-06, + "loss": 2.0338295, + "memory(GiB)": 58.1, + "step": 1160, + "train_speed(iter/s)": 0.458369 + }, + { + "acc": 0.58941851, + "epoch": 0.029553526128868593, + "grad_norm": 6.4375, + "learning_rate": 2.9553526128868598e-06, + "loss": 2.01237602, + "memory(GiB)": 58.1, + "step": 1165, + "train_speed(iter/s)": 0.459381 + }, + { + "acc": 0.57980275, + "epoch": 0.02968036529680365, + "grad_norm": 5.625, + "learning_rate": 2.9680365296803653e-06, + "loss": 2.09141636, + "memory(GiB)": 58.1, + "step": 1170, + "train_speed(iter/s)": 0.460407 + }, + { + "acc": 0.59483747, + "epoch": 0.02980720446473871, + "grad_norm": 5.6875, + "learning_rate": 2.9807204464738717e-06, + "loss": 2.02767811, + "memory(GiB)": 58.1, + "step": 1175, + "train_speed(iter/s)": 0.461425 + }, + { + "acc": 0.60195074, + "epoch": 0.029934043632673768, + "grad_norm": 5.4375, + "learning_rate": 2.9934043632673772e-06, + "loss": 1.94009094, + "memory(GiB)": 58.1, + "step": 1180, + "train_speed(iter/s)": 0.462439 + }, + { + "acc": 0.57435117, + "epoch": 0.030060882800608826, + "grad_norm": 5.375, + "learning_rate": 3.006088280060883e-06, + "loss": 2.04282379, + "memory(GiB)": 58.1, + "step": 1185, + "train_speed(iter/s)": 0.463452 + }, + { + "acc": 0.577742, + "epoch": 0.030187721968543885, + "grad_norm": 6.15625, + "learning_rate": 3.0187721968543887e-06, + "loss": 2.03994942, + "memory(GiB)": 58.1, + "step": 1190, + "train_speed(iter/s)": 0.464464 + }, + { + "acc": 0.57857523, + "epoch": 0.030314561136478943, + "grad_norm": 5.9375, + "learning_rate": 3.0314561136478947e-06, + "loss": 2.04877777, + "memory(GiB)": 58.1, + "step": 1195, + "train_speed(iter/s)": 0.465455 + }, + { + "acc": 0.61877623, + "epoch": 0.030441400304414, + "grad_norm": 6.25, + "learning_rate": 3.0441400304414002e-06, + "loss": 1.97868118, + "memory(GiB)": 58.1, + "step": 1200, + "train_speed(iter/s)": 0.466454 + }, + { + "epoch": 0.030441400304414, + "eval_acc": 0.5814076001064471, + "eval_loss": 2.042628765106201, + "eval_runtime": 112.1687, + "eval_samples_per_second": 56.789, + "eval_steps_per_second": 28.395, + "step": 1200 + }, + { + "acc": 0.5757021, + "epoch": 0.03056823947234906, + "grad_norm": 6.46875, + "learning_rate": 3.0568239472349066e-06, + "loss": 2.13078766, + "memory(GiB)": 58.1, + "step": 1205, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.57486238, + "epoch": 0.030695078640284118, + "grad_norm": 6.59375, + "learning_rate": 3.069507864028412e-06, + "loss": 1.99296398, + "memory(GiB)": 58.1, + "step": 1210, + "train_speed(iter/s)": 0.448188 + }, + { + "acc": 0.5830471, + "epoch": 0.030821917808219176, + "grad_norm": 5.34375, + "learning_rate": 3.082191780821918e-06, + "loss": 2.03044128, + "memory(GiB)": 58.1, + "step": 1215, + "train_speed(iter/s)": 0.449164 + }, + { + "acc": 0.58052006, + "epoch": 0.030948756976154235, + "grad_norm": 6.15625, + "learning_rate": 3.0948756976154237e-06, + "loss": 2.0468399, + "memory(GiB)": 58.14, + "step": 1220, + "train_speed(iter/s)": 0.450147 + }, + { + "acc": 0.59650674, + "epoch": 0.031075596144089293, + "grad_norm": 5.53125, + "learning_rate": 3.1075596144089297e-06, + "loss": 2.04331741, + "memory(GiB)": 58.14, + "step": 1225, + "train_speed(iter/s)": 0.451123 + }, + { + "acc": 0.59364333, + "epoch": 0.03120243531202435, + "grad_norm": 5.8125, + "learning_rate": 3.120243531202435e-06, + "loss": 2.03374901, + "memory(GiB)": 58.14, + "step": 1230, + "train_speed(iter/s)": 0.45209 + }, + { + "acc": 0.58367114, + "epoch": 0.03132927447995941, + "grad_norm": 5.78125, + "learning_rate": 3.1329274479959416e-06, + "loss": 2.0957283, + "memory(GiB)": 58.14, + "step": 1235, + "train_speed(iter/s)": 0.453056 + }, + { + "acc": 0.5884491, + "epoch": 0.03145611364789447, + "grad_norm": 5.875, + "learning_rate": 3.145611364789447e-06, + "loss": 2.0294363, + "memory(GiB)": 58.14, + "step": 1240, + "train_speed(iter/s)": 0.454021 + }, + { + "acc": 0.58413467, + "epoch": 0.031582952815829526, + "grad_norm": 5.34375, + "learning_rate": 3.158295281582953e-06, + "loss": 2.0904974, + "memory(GiB)": 58.14, + "step": 1245, + "train_speed(iter/s)": 0.45498 + }, + { + "acc": 0.58544159, + "epoch": 0.031709791983764585, + "grad_norm": 6.5625, + "learning_rate": 3.1709791983764586e-06, + "loss": 2.03991356, + "memory(GiB)": 58.14, + "step": 1250, + "train_speed(iter/s)": 0.455933 + }, + { + "acc": 0.58787956, + "epoch": 0.03183663115169964, + "grad_norm": 5.1875, + "learning_rate": 3.183663115169965e-06, + "loss": 2.05537472, + "memory(GiB)": 58.14, + "step": 1255, + "train_speed(iter/s)": 0.456889 + }, + { + "acc": 0.59399815, + "epoch": 0.0319634703196347, + "grad_norm": 4.71875, + "learning_rate": 3.1963470319634706e-06, + "loss": 2.06450615, + "memory(GiB)": 58.14, + "step": 1260, + "train_speed(iter/s)": 0.457837 + }, + { + "acc": 0.58911357, + "epoch": 0.03209030948756976, + "grad_norm": 5.28125, + "learning_rate": 3.2090309487569765e-06, + "loss": 1.98663044, + "memory(GiB)": 58.14, + "step": 1265, + "train_speed(iter/s)": 0.458785 + }, + { + "acc": 0.58299508, + "epoch": 0.03221714865550482, + "grad_norm": 5.96875, + "learning_rate": 3.221714865550482e-06, + "loss": 2.05958347, + "memory(GiB)": 58.14, + "step": 1270, + "train_speed(iter/s)": 0.45973 + }, + { + "acc": 0.59095502, + "epoch": 0.032343987823439876, + "grad_norm": 5.3125, + "learning_rate": 3.234398782343988e-06, + "loss": 2.08648758, + "memory(GiB)": 58.14, + "step": 1275, + "train_speed(iter/s)": 0.460652 + }, + { + "acc": 0.58892069, + "epoch": 0.032470826991374935, + "grad_norm": 4.96875, + "learning_rate": 3.2470826991374936e-06, + "loss": 2.01748333, + "memory(GiB)": 58.14, + "step": 1280, + "train_speed(iter/s)": 0.461593 + }, + { + "acc": 0.58466053, + "epoch": 0.03259766615930999, + "grad_norm": 5.5, + "learning_rate": 3.259766615931e-06, + "loss": 2.04656467, + "memory(GiB)": 58.14, + "step": 1285, + "train_speed(iter/s)": 0.462518 + }, + { + "acc": 0.604704, + "epoch": 0.03272450532724505, + "grad_norm": 5.84375, + "learning_rate": 3.2724505327245055e-06, + "loss": 1.99920521, + "memory(GiB)": 58.14, + "step": 1290, + "train_speed(iter/s)": 0.463446 + }, + { + "acc": 0.58460693, + "epoch": 0.03285134449518011, + "grad_norm": 5.78125, + "learning_rate": 3.2851344495180115e-06, + "loss": 2.04618454, + "memory(GiB)": 58.14, + "step": 1295, + "train_speed(iter/s)": 0.46436 + }, + { + "acc": 0.58162394, + "epoch": 0.03297818366311517, + "grad_norm": 5.5, + "learning_rate": 3.297818366311517e-06, + "loss": 2.08075142, + "memory(GiB)": 58.14, + "step": 1300, + "train_speed(iter/s)": 0.465273 + }, + { + "epoch": 0.03297818366311517, + "eval_acc": 0.5832720956085554, + "eval_loss": 2.023688316345215, + "eval_runtime": 112.6353, + "eval_samples_per_second": 56.554, + "eval_steps_per_second": 28.277, + "step": 1300 + }, + { + "acc": 0.6012743, + "epoch": 0.033105022831050226, + "grad_norm": 7.9375, + "learning_rate": 3.310502283105023e-06, + "loss": 1.98675652, + "memory(GiB)": 58.14, + "step": 1305, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.59395251, + "epoch": 0.033231861998985285, + "grad_norm": 5.46875, + "learning_rate": 3.3231861998985286e-06, + "loss": 2.00059795, + "memory(GiB)": 58.14, + "step": 1310, + "train_speed(iter/s)": 0.448358 + }, + { + "acc": 0.61236, + "epoch": 0.03335870116692034, + "grad_norm": 5.5, + "learning_rate": 3.335870116692035e-06, + "loss": 1.90900192, + "memory(GiB)": 58.14, + "step": 1315, + "train_speed(iter/s)": 0.449254 + }, + { + "acc": 0.57951021, + "epoch": 0.0334855403348554, + "grad_norm": 5.875, + "learning_rate": 3.3485540334855405e-06, + "loss": 2.04443588, + "memory(GiB)": 58.14, + "step": 1320, + "train_speed(iter/s)": 0.450154 + }, + { + "acc": 0.59460707, + "epoch": 0.03361237950279046, + "grad_norm": 6.4375, + "learning_rate": 3.3612379502790465e-06, + "loss": 1.96191101, + "memory(GiB)": 58.14, + "step": 1325, + "train_speed(iter/s)": 0.451041 + }, + { + "acc": 0.60482206, + "epoch": 0.03373921867072552, + "grad_norm": 7.0, + "learning_rate": 3.373921867072552e-06, + "loss": 2.02402935, + "memory(GiB)": 58.14, + "step": 1330, + "train_speed(iter/s)": 0.451932 + }, + { + "acc": 0.59342928, + "epoch": 0.033866057838660577, + "grad_norm": 5.53125, + "learning_rate": 3.3866057838660584e-06, + "loss": 2.0785635, + "memory(GiB)": 58.14, + "step": 1335, + "train_speed(iter/s)": 0.452815 + }, + { + "acc": 0.60314054, + "epoch": 0.033992897006595635, + "grad_norm": 5.46875, + "learning_rate": 3.399289700659564e-06, + "loss": 1.97551765, + "memory(GiB)": 58.14, + "step": 1340, + "train_speed(iter/s)": 0.453698 + }, + { + "acc": 0.60864201, + "epoch": 0.03411973617453069, + "grad_norm": 5.125, + "learning_rate": 3.41197361745307e-06, + "loss": 2.00481873, + "memory(GiB)": 58.14, + "step": 1345, + "train_speed(iter/s)": 0.454573 + }, + { + "acc": 0.58488483, + "epoch": 0.03424657534246575, + "grad_norm": 5.15625, + "learning_rate": 3.4246575342465754e-06, + "loss": 2.00058441, + "memory(GiB)": 58.14, + "step": 1350, + "train_speed(iter/s)": 0.455438 + }, + { + "acc": 0.60826778, + "epoch": 0.03437341451040081, + "grad_norm": 6.40625, + "learning_rate": 3.4373414510400814e-06, + "loss": 1.93858299, + "memory(GiB)": 58.14, + "step": 1355, + "train_speed(iter/s)": 0.456308 + }, + { + "acc": 0.59984961, + "epoch": 0.03450025367833587, + "grad_norm": 6.3125, + "learning_rate": 3.450025367833587e-06, + "loss": 1.98000469, + "memory(GiB)": 58.14, + "step": 1360, + "train_speed(iter/s)": 0.457168 + }, + { + "acc": 0.59185562, + "epoch": 0.03462709284627093, + "grad_norm": 5.6875, + "learning_rate": 3.4627092846270933e-06, + "loss": 1.99953384, + "memory(GiB)": 58.14, + "step": 1365, + "train_speed(iter/s)": 0.458033 + }, + { + "acc": 0.58549786, + "epoch": 0.034753932014205985, + "grad_norm": 5.5625, + "learning_rate": 3.475393201420599e-06, + "loss": 2.00680447, + "memory(GiB)": 58.14, + "step": 1370, + "train_speed(iter/s)": 0.458898 + }, + { + "acc": 0.59121161, + "epoch": 0.03488077118214104, + "grad_norm": 5.96875, + "learning_rate": 3.488077118214105e-06, + "loss": 2.00466576, + "memory(GiB)": 58.14, + "step": 1375, + "train_speed(iter/s)": 0.45975 + }, + { + "acc": 0.60226107, + "epoch": 0.0350076103500761, + "grad_norm": 5.53125, + "learning_rate": 3.5007610350076104e-06, + "loss": 2.01248531, + "memory(GiB)": 58.14, + "step": 1380, + "train_speed(iter/s)": 0.460599 + }, + { + "acc": 0.60449491, + "epoch": 0.03513444951801116, + "grad_norm": 5.6875, + "learning_rate": 3.5134449518011164e-06, + "loss": 1.97046242, + "memory(GiB)": 58.14, + "step": 1385, + "train_speed(iter/s)": 0.461452 + }, + { + "acc": 0.58740883, + "epoch": 0.03526128868594622, + "grad_norm": 5.6875, + "learning_rate": 3.5261288685946223e-06, + "loss": 2.056567, + "memory(GiB)": 58.14, + "step": 1390, + "train_speed(iter/s)": 0.4623 + }, + { + "acc": 0.57723293, + "epoch": 0.03538812785388128, + "grad_norm": 5.59375, + "learning_rate": 3.5388127853881283e-06, + "loss": 2.11377258, + "memory(GiB)": 58.14, + "step": 1395, + "train_speed(iter/s)": 0.463138 + }, + { + "acc": 0.57502775, + "epoch": 0.035514967021816335, + "grad_norm": 5.53125, + "learning_rate": 3.551496702181634e-06, + "loss": 1.97540092, + "memory(GiB)": 58.14, + "step": 1400, + "train_speed(iter/s)": 0.463979 + }, + { + "epoch": 0.035514967021816335, + "eval_acc": 0.5851236403240705, + "eval_loss": 2.00679874420166, + "eval_runtime": 113.8968, + "eval_samples_per_second": 55.928, + "eval_steps_per_second": 27.964, + "step": 1400 + }, + { + "acc": 0.57469149, + "epoch": 0.03564180618975139, + "grad_norm": 6.375, + "learning_rate": 3.56418061897514e-06, + "loss": 2.07700272, + "memory(GiB)": 58.14, + "step": 1405, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.57950678, + "epoch": 0.03576864535768645, + "grad_norm": 5.84375, + "learning_rate": 3.5768645357686453e-06, + "loss": 2.08248138, + "memory(GiB)": 58.14, + "step": 1410, + "train_speed(iter/s)": 0.448139 + }, + { + "acc": 0.59639082, + "epoch": 0.03589548452562151, + "grad_norm": 5.125, + "learning_rate": 3.5895484525621517e-06, + "loss": 2.01292553, + "memory(GiB)": 58.14, + "step": 1415, + "train_speed(iter/s)": 0.448971 + }, + { + "acc": 0.60409479, + "epoch": 0.03602232369355657, + "grad_norm": 6.0, + "learning_rate": 3.6022323693556573e-06, + "loss": 1.96956444, + "memory(GiB)": 58.14, + "step": 1420, + "train_speed(iter/s)": 0.4498 + }, + { + "acc": 0.59116673, + "epoch": 0.03614916286149163, + "grad_norm": 6.15625, + "learning_rate": 3.6149162861491632e-06, + "loss": 2.03165779, + "memory(GiB)": 58.14, + "step": 1425, + "train_speed(iter/s)": 0.450624 + }, + { + "acc": 0.60644951, + "epoch": 0.036276002029426685, + "grad_norm": 5.34375, + "learning_rate": 3.6276002029426688e-06, + "loss": 1.94401093, + "memory(GiB)": 58.14, + "step": 1430, + "train_speed(iter/s)": 0.451444 + }, + { + "acc": 0.57770939, + "epoch": 0.03640284119736174, + "grad_norm": 5.84375, + "learning_rate": 3.6402841197361748e-06, + "loss": 2.02814827, + "memory(GiB)": 58.14, + "step": 1435, + "train_speed(iter/s)": 0.452262 + }, + { + "acc": 0.59192715, + "epoch": 0.0365296803652968, + "grad_norm": 7.4375, + "learning_rate": 3.6529680365296803e-06, + "loss": 2.05361481, + "memory(GiB)": 58.14, + "step": 1440, + "train_speed(iter/s)": 0.453091 + }, + { + "acc": 0.61672406, + "epoch": 0.03665651953323186, + "grad_norm": 5.21875, + "learning_rate": 3.6656519533231867e-06, + "loss": 1.98095169, + "memory(GiB)": 58.14, + "step": 1445, + "train_speed(iter/s)": 0.453905 + }, + { + "acc": 0.58148623, + "epoch": 0.03678335870116692, + "grad_norm": 5.21875, + "learning_rate": 3.6783358701166922e-06, + "loss": 2.10712395, + "memory(GiB)": 58.14, + "step": 1450, + "train_speed(iter/s)": 0.454719 + }, + { + "acc": 0.60510659, + "epoch": 0.03691019786910198, + "grad_norm": 5.4375, + "learning_rate": 3.691019786910198e-06, + "loss": 2.03598137, + "memory(GiB)": 58.14, + "step": 1455, + "train_speed(iter/s)": 0.455537 + }, + { + "acc": 0.59899988, + "epoch": 0.037037037037037035, + "grad_norm": 6.15625, + "learning_rate": 3.7037037037037037e-06, + "loss": 1.91426125, + "memory(GiB)": 58.14, + "step": 1460, + "train_speed(iter/s)": 0.456358 + }, + { + "acc": 0.59026952, + "epoch": 0.03716387620497209, + "grad_norm": 5.75, + "learning_rate": 3.7163876204972097e-06, + "loss": 2.01717033, + "memory(GiB)": 58.14, + "step": 1465, + "train_speed(iter/s)": 0.457166 + }, + { + "acc": 0.5948554, + "epoch": 0.03729071537290715, + "grad_norm": 5.15625, + "learning_rate": 3.7290715372907157e-06, + "loss": 1.99462662, + "memory(GiB)": 58.14, + "step": 1470, + "train_speed(iter/s)": 0.457973 + }, + { + "acc": 0.5895648, + "epoch": 0.03741755454084221, + "grad_norm": 6.21875, + "learning_rate": 3.7417554540842216e-06, + "loss": 1.99374218, + "memory(GiB)": 58.14, + "step": 1475, + "train_speed(iter/s)": 0.458784 + }, + { + "acc": 0.61564498, + "epoch": 0.03754439370877727, + "grad_norm": 5.34375, + "learning_rate": 3.754439370877727e-06, + "loss": 1.95908012, + "memory(GiB)": 58.14, + "step": 1480, + "train_speed(iter/s)": 0.459587 + }, + { + "acc": 0.58806829, + "epoch": 0.03767123287671233, + "grad_norm": 5.875, + "learning_rate": 3.767123287671233e-06, + "loss": 1.93563194, + "memory(GiB)": 58.14, + "step": 1485, + "train_speed(iter/s)": 0.460381 + }, + { + "acc": 0.58792725, + "epoch": 0.037798072044647385, + "grad_norm": 6.375, + "learning_rate": 3.7798072044647387e-06, + "loss": 2.0438406, + "memory(GiB)": 58.14, + "step": 1490, + "train_speed(iter/s)": 0.461175 + }, + { + "acc": 0.57838364, + "epoch": 0.03792491121258244, + "grad_norm": 6.40625, + "learning_rate": 3.792491121258245e-06, + "loss": 2.11203232, + "memory(GiB)": 58.14, + "step": 1495, + "train_speed(iter/s)": 0.461973 + }, + { + "acc": 0.58842378, + "epoch": 0.0380517503805175, + "grad_norm": 5.6875, + "learning_rate": 3.8051750380517506e-06, + "loss": 2.02233601, + "memory(GiB)": 58.14, + "step": 1500, + "train_speed(iter/s)": 0.462758 + }, + { + "epoch": 0.0380517503805175, + "eval_acc": 0.5869334083086398, + "eval_loss": 1.9907699823379517, + "eval_runtime": 112.6647, + "eval_samples_per_second": 56.539, + "eval_steps_per_second": 28.27, + "step": 1500 + }, + { + "acc": 0.59982653, + "epoch": 0.03817858954845256, + "grad_norm": 5.21875, + "learning_rate": 3.817858954845256e-06, + "loss": 1.98179169, + "memory(GiB)": 58.14, + "step": 1505, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.58439264, + "epoch": 0.03830542871638762, + "grad_norm": 5.96875, + "learning_rate": 3.830542871638762e-06, + "loss": 2.00547409, + "memory(GiB)": 58.14, + "step": 1510, + "train_speed(iter/s)": 0.448185 + }, + { + "acc": 0.59878612, + "epoch": 0.03843226788432268, + "grad_norm": 5.53125, + "learning_rate": 3.843226788432268e-06, + "loss": 1.97039299, + "memory(GiB)": 58.14, + "step": 1515, + "train_speed(iter/s)": 0.448958 + }, + { + "acc": 0.58891411, + "epoch": 0.038559107052257735, + "grad_norm": 6.375, + "learning_rate": 3.855910705225774e-06, + "loss": 1.98719673, + "memory(GiB)": 58.14, + "step": 1520, + "train_speed(iter/s)": 0.449735 + }, + { + "acc": 0.58927441, + "epoch": 0.03868594622019279, + "grad_norm": 5.34375, + "learning_rate": 3.86859462201928e-06, + "loss": 1.9856411, + "memory(GiB)": 58.14, + "step": 1525, + "train_speed(iter/s)": 0.450511 + }, + { + "acc": 0.58314414, + "epoch": 0.03881278538812785, + "grad_norm": 6.25, + "learning_rate": 3.881278538812785e-06, + "loss": 2.02250271, + "memory(GiB)": 58.14, + "step": 1530, + "train_speed(iter/s)": 0.451284 + }, + { + "acc": 0.60146899, + "epoch": 0.03893962455606291, + "grad_norm": 5.84375, + "learning_rate": 3.893962455606292e-06, + "loss": 1.99159927, + "memory(GiB)": 58.14, + "step": 1535, + "train_speed(iter/s)": 0.452057 + }, + { + "acc": 0.60928993, + "epoch": 0.03906646372399797, + "grad_norm": 6.40625, + "learning_rate": 3.906646372399797e-06, + "loss": 1.91212196, + "memory(GiB)": 58.14, + "step": 1540, + "train_speed(iter/s)": 0.452824 + }, + { + "acc": 0.56870441, + "epoch": 0.03919330289193303, + "grad_norm": 5.71875, + "learning_rate": 3.919330289193303e-06, + "loss": 2.03145256, + "memory(GiB)": 58.14, + "step": 1545, + "train_speed(iter/s)": 0.453589 + }, + { + "acc": 0.57842994, + "epoch": 0.039320142059868085, + "grad_norm": 6.21875, + "learning_rate": 3.932014205986809e-06, + "loss": 2.00378914, + "memory(GiB)": 58.14, + "step": 1550, + "train_speed(iter/s)": 0.454362 + }, + { + "acc": 0.61088352, + "epoch": 0.03944698122780314, + "grad_norm": 6.3125, + "learning_rate": 3.944698122780315e-06, + "loss": 1.96507072, + "memory(GiB)": 58.14, + "step": 1555, + "train_speed(iter/s)": 0.45513 + }, + { + "acc": 0.5916419, + "epoch": 0.0395738203957382, + "grad_norm": 8.8125, + "learning_rate": 3.95738203957382e-06, + "loss": 2.03187447, + "memory(GiB)": 58.14, + "step": 1560, + "train_speed(iter/s)": 0.455892 + }, + { + "acc": 0.59330702, + "epoch": 0.03970065956367326, + "grad_norm": 6.15625, + "learning_rate": 3.970065956367327e-06, + "loss": 1.97511959, + "memory(GiB)": 58.14, + "step": 1565, + "train_speed(iter/s)": 0.456649 + }, + { + "acc": 0.59780612, + "epoch": 0.03982749873160832, + "grad_norm": 5.5625, + "learning_rate": 3.982749873160832e-06, + "loss": 2.00135841, + "memory(GiB)": 58.14, + "step": 1570, + "train_speed(iter/s)": 0.457411 + }, + { + "acc": 0.57598472, + "epoch": 0.03995433789954338, + "grad_norm": 7.5, + "learning_rate": 3.995433789954338e-06, + "loss": 2.070998, + "memory(GiB)": 58.14, + "step": 1575, + "train_speed(iter/s)": 0.458166 + }, + { + "acc": 0.60508847, + "epoch": 0.040081177067478435, + "grad_norm": 4.78125, + "learning_rate": 4.008117706747844e-06, + "loss": 2.01071415, + "memory(GiB)": 58.14, + "step": 1580, + "train_speed(iter/s)": 0.458913 + }, + { + "acc": 0.59113226, + "epoch": 0.04020801623541349, + "grad_norm": 7.0, + "learning_rate": 4.02080162354135e-06, + "loss": 1.9657011, + "memory(GiB)": 58.14, + "step": 1585, + "train_speed(iter/s)": 0.45967 + }, + { + "acc": 0.59855337, + "epoch": 0.04033485540334855, + "grad_norm": 6.0625, + "learning_rate": 4.033485540334856e-06, + "loss": 1.9439846, + "memory(GiB)": 58.14, + "step": 1590, + "train_speed(iter/s)": 0.460421 + }, + { + "acc": 0.59298334, + "epoch": 0.04046169457128361, + "grad_norm": 5.25, + "learning_rate": 4.046169457128362e-06, + "loss": 2.02098694, + "memory(GiB)": 58.14, + "step": 1595, + "train_speed(iter/s)": 0.46116 + }, + { + "acc": 0.57667694, + "epoch": 0.04058853373921867, + "grad_norm": 5.9375, + "learning_rate": 4.058853373921867e-06, + "loss": 2.0582428, + "memory(GiB)": 58.14, + "step": 1600, + "train_speed(iter/s)": 0.461905 + }, + { + "epoch": 0.04058853373921867, + "eval_acc": 0.5886140861945868, + "eval_loss": 1.9762248992919922, + "eval_runtime": 112.9974, + "eval_samples_per_second": 56.373, + "eval_steps_per_second": 28.186, + "step": 1600 + }, + { + "acc": 0.60419512, + "epoch": 0.04071537290715373, + "grad_norm": 5.46875, + "learning_rate": 4.071537290715373e-06, + "loss": 1.88787327, + "memory(GiB)": 58.14, + "step": 1605, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.59176865, + "epoch": 0.040842212075088785, + "grad_norm": 6.875, + "learning_rate": 4.084221207508879e-06, + "loss": 2.00019836, + "memory(GiB)": 58.14, + "step": 1610, + "train_speed(iter/s)": 0.448232 + }, + { + "acc": 0.58779345, + "epoch": 0.040969051243023843, + "grad_norm": 5.96875, + "learning_rate": 4.096905124302385e-06, + "loss": 2.00668335, + "memory(GiB)": 58.14, + "step": 1615, + "train_speed(iter/s)": 0.448964 + }, + { + "acc": 0.59657793, + "epoch": 0.0410958904109589, + "grad_norm": 5.28125, + "learning_rate": 4.109589041095891e-06, + "loss": 1.95897408, + "memory(GiB)": 58.14, + "step": 1620, + "train_speed(iter/s)": 0.449693 + }, + { + "acc": 0.58437681, + "epoch": 0.04122272957889396, + "grad_norm": 6.28125, + "learning_rate": 4.122272957889397e-06, + "loss": 2.11465492, + "memory(GiB)": 58.14, + "step": 1625, + "train_speed(iter/s)": 0.450418 + }, + { + "acc": 0.58726759, + "epoch": 0.04134956874682902, + "grad_norm": 7.21875, + "learning_rate": 4.134956874682902e-06, + "loss": 1.99460888, + "memory(GiB)": 58.14, + "step": 1630, + "train_speed(iter/s)": 0.451146 + }, + { + "acc": 0.62223482, + "epoch": 0.04147640791476408, + "grad_norm": 5.15625, + "learning_rate": 4.147640791476408e-06, + "loss": 1.88918247, + "memory(GiB)": 58.14, + "step": 1635, + "train_speed(iter/s)": 0.451864 + }, + { + "acc": 0.59386044, + "epoch": 0.041603247082699135, + "grad_norm": 5.375, + "learning_rate": 4.160324708269914e-06, + "loss": 2.0271759, + "memory(GiB)": 58.14, + "step": 1640, + "train_speed(iter/s)": 0.452587 + }, + { + "acc": 0.60839391, + "epoch": 0.041730086250634194, + "grad_norm": 4.8125, + "learning_rate": 4.17300862506342e-06, + "loss": 1.88440781, + "memory(GiB)": 58.14, + "step": 1645, + "train_speed(iter/s)": 0.453311 + }, + { + "acc": 0.59974365, + "epoch": 0.04185692541856925, + "grad_norm": 6.125, + "learning_rate": 4.185692541856926e-06, + "loss": 1.87821274, + "memory(GiB)": 58.14, + "step": 1650, + "train_speed(iter/s)": 0.454027 + }, + { + "acc": 0.59998851, + "epoch": 0.04198376458650431, + "grad_norm": 5.875, + "learning_rate": 4.198376458650432e-06, + "loss": 1.93571701, + "memory(GiB)": 58.14, + "step": 1655, + "train_speed(iter/s)": 0.454745 + }, + { + "acc": 0.60226812, + "epoch": 0.04211060375443937, + "grad_norm": 6.75, + "learning_rate": 4.211060375443937e-06, + "loss": 1.89863091, + "memory(GiB)": 58.14, + "step": 1660, + "train_speed(iter/s)": 0.455459 + }, + { + "acc": 0.58071132, + "epoch": 0.04223744292237443, + "grad_norm": 5.78125, + "learning_rate": 4.223744292237444e-06, + "loss": 2.11747818, + "memory(GiB)": 58.14, + "step": 1665, + "train_speed(iter/s)": 0.456169 + }, + { + "acc": 0.59682236, + "epoch": 0.042364282090309485, + "grad_norm": 5.6875, + "learning_rate": 4.236428209030949e-06, + "loss": 1.9822731, + "memory(GiB)": 58.14, + "step": 1670, + "train_speed(iter/s)": 0.456876 + }, + { + "acc": 0.5858675, + "epoch": 0.042491121258244544, + "grad_norm": 6.21875, + "learning_rate": 4.249112125824455e-06, + "loss": 2.00417633, + "memory(GiB)": 58.14, + "step": 1675, + "train_speed(iter/s)": 0.457586 + }, + { + "acc": 0.59361582, + "epoch": 0.0426179604261796, + "grad_norm": 6.90625, + "learning_rate": 4.261796042617961e-06, + "loss": 1.97185211, + "memory(GiB)": 58.14, + "step": 1680, + "train_speed(iter/s)": 0.458285 + }, + { + "acc": 0.62019005, + "epoch": 0.04274479959411466, + "grad_norm": 6.34375, + "learning_rate": 4.274479959411467e-06, + "loss": 1.87001457, + "memory(GiB)": 58.14, + "step": 1685, + "train_speed(iter/s)": 0.458983 + }, + { + "acc": 0.59541726, + "epoch": 0.04287163876204972, + "grad_norm": 5.28125, + "learning_rate": 4.287163876204972e-06, + "loss": 1.90830975, + "memory(GiB)": 58.14, + "step": 1690, + "train_speed(iter/s)": 0.459687 + }, + { + "acc": 0.59603844, + "epoch": 0.04299847792998478, + "grad_norm": 5.53125, + "learning_rate": 4.299847792998479e-06, + "loss": 1.91053753, + "memory(GiB)": 58.14, + "step": 1695, + "train_speed(iter/s)": 0.460386 + }, + { + "acc": 0.60193319, + "epoch": 0.043125317097919835, + "grad_norm": 7.375, + "learning_rate": 4.312531709791984e-06, + "loss": 1.97927856, + "memory(GiB)": 58.14, + "step": 1700, + "train_speed(iter/s)": 0.461086 + }, + { + "epoch": 0.043125317097919835, + "eval_acc": 0.5902596716265395, + "eval_loss": 1.9628418684005737, + "eval_runtime": 114.517, + "eval_samples_per_second": 55.625, + "eval_steps_per_second": 27.812, + "step": 1700 + }, + { + "acc": 0.57796144, + "epoch": 0.043252156265854894, + "grad_norm": 6.96875, + "learning_rate": 4.32521562658549e-06, + "loss": 2.00307999, + "memory(GiB)": 58.14, + "step": 1705, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.59091091, + "epoch": 0.04337899543378995, + "grad_norm": 6.0, + "learning_rate": 4.337899543378996e-06, + "loss": 2.01779633, + "memory(GiB)": 58.14, + "step": 1710, + "train_speed(iter/s)": 0.44806 + }, + { + "acc": 0.57111516, + "epoch": 0.04350583460172501, + "grad_norm": 5.0, + "learning_rate": 4.350583460172502e-06, + "loss": 2.06058617, + "memory(GiB)": 58.14, + "step": 1715, + "train_speed(iter/s)": 0.448753 + }, + { + "acc": 0.61192236, + "epoch": 0.04363267376966007, + "grad_norm": 6.25, + "learning_rate": 4.363267376966007e-06, + "loss": 1.92471466, + "memory(GiB)": 58.14, + "step": 1720, + "train_speed(iter/s)": 0.449443 + }, + { + "acc": 0.59001474, + "epoch": 0.04375951293759513, + "grad_norm": 5.59375, + "learning_rate": 4.375951293759514e-06, + "loss": 2.00295353, + "memory(GiB)": 58.14, + "step": 1725, + "train_speed(iter/s)": 0.450137 + }, + { + "acc": 0.59872637, + "epoch": 0.043886352105530185, + "grad_norm": 6.15625, + "learning_rate": 4.388635210553019e-06, + "loss": 1.95385628, + "memory(GiB)": 58.14, + "step": 1730, + "train_speed(iter/s)": 0.45083 + }, + { + "acc": 0.6004015, + "epoch": 0.044013191273465244, + "grad_norm": 5.375, + "learning_rate": 4.401319127346525e-06, + "loss": 1.95184097, + "memory(GiB)": 58.14, + "step": 1735, + "train_speed(iter/s)": 0.451519 + }, + { + "acc": 0.60364981, + "epoch": 0.0441400304414003, + "grad_norm": 7.625, + "learning_rate": 4.414003044140031e-06, + "loss": 1.95455666, + "memory(GiB)": 58.14, + "step": 1740, + "train_speed(iter/s)": 0.452202 + }, + { + "acc": 0.60542698, + "epoch": 0.04426686960933536, + "grad_norm": 5.90625, + "learning_rate": 4.426686960933537e-06, + "loss": 1.9811676, + "memory(GiB)": 58.14, + "step": 1745, + "train_speed(iter/s)": 0.45288 + }, + { + "acc": 0.5867034, + "epoch": 0.04439370877727042, + "grad_norm": 5.59375, + "learning_rate": 4.439370877727043e-06, + "loss": 2.04107285, + "memory(GiB)": 58.14, + "step": 1750, + "train_speed(iter/s)": 0.453562 + }, + { + "acc": 0.58724022, + "epoch": 0.04452054794520548, + "grad_norm": 6.71875, + "learning_rate": 4.4520547945205486e-06, + "loss": 1.96066132, + "memory(GiB)": 58.14, + "step": 1755, + "train_speed(iter/s)": 0.454236 + }, + { + "acc": 0.61269112, + "epoch": 0.044647387113140535, + "grad_norm": 6.03125, + "learning_rate": 4.464738711314054e-06, + "loss": 1.87263718, + "memory(GiB)": 58.14, + "step": 1760, + "train_speed(iter/s)": 0.454911 + }, + { + "acc": 0.59345675, + "epoch": 0.044774226281075594, + "grad_norm": 5.5625, + "learning_rate": 4.47742262810756e-06, + "loss": 2.01479778, + "memory(GiB)": 58.14, + "step": 1765, + "train_speed(iter/s)": 0.455588 + }, + { + "acc": 0.60399456, + "epoch": 0.04490106544901065, + "grad_norm": 4.9375, + "learning_rate": 4.490106544901066e-06, + "loss": 1.89943275, + "memory(GiB)": 58.14, + "step": 1770, + "train_speed(iter/s)": 0.45626 + }, + { + "acc": 0.60044622, + "epoch": 0.04502790461694571, + "grad_norm": 6.9375, + "learning_rate": 4.502790461694572e-06, + "loss": 2.01143799, + "memory(GiB)": 58.14, + "step": 1775, + "train_speed(iter/s)": 0.456924 + }, + { + "acc": 0.59995494, + "epoch": 0.04515474378488077, + "grad_norm": 5.78125, + "learning_rate": 4.5154743784880776e-06, + "loss": 2.03026352, + "memory(GiB)": 58.14, + "step": 1780, + "train_speed(iter/s)": 0.457601 + }, + { + "acc": 0.60096474, + "epoch": 0.04528158295281583, + "grad_norm": 5.09375, + "learning_rate": 4.5281582952815835e-06, + "loss": 1.95925446, + "memory(GiB)": 58.14, + "step": 1785, + "train_speed(iter/s)": 0.458262 + }, + { + "acc": 0.60372648, + "epoch": 0.045408422120750885, + "grad_norm": 5.53125, + "learning_rate": 4.540842212075089e-06, + "loss": 1.92858753, + "memory(GiB)": 58.14, + "step": 1790, + "train_speed(iter/s)": 0.458921 + }, + { + "acc": 0.59696364, + "epoch": 0.045535261288685944, + "grad_norm": 5.875, + "learning_rate": 4.553526128868595e-06, + "loss": 1.98838577, + "memory(GiB)": 58.14, + "step": 1795, + "train_speed(iter/s)": 0.459586 + }, + { + "acc": 0.59612103, + "epoch": 0.045662100456621, + "grad_norm": 4.65625, + "learning_rate": 4.566210045662101e-06, + "loss": 1.98218117, + "memory(GiB)": 58.14, + "step": 1800, + "train_speed(iter/s)": 0.460246 + }, + { + "epoch": 0.045662100456621, + "eval_acc": 0.5918338188485748, + "eval_loss": 1.950043797492981, + "eval_runtime": 112.4711, + "eval_samples_per_second": 56.637, + "eval_steps_per_second": 28.318, + "step": 1800 + }, + { + "acc": 0.62371159, + "epoch": 0.04578893962455606, + "grad_norm": 5.875, + "learning_rate": 4.5788939624556065e-06, + "loss": 1.82332497, + "memory(GiB)": 58.14, + "step": 1805, + "train_speed(iter/s)": 0.447535 + }, + { + "acc": 0.59204745, + "epoch": 0.04591577879249112, + "grad_norm": 5.0, + "learning_rate": 4.5915778792491125e-06, + "loss": 1.9619072, + "memory(GiB)": 58.14, + "step": 1810, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.57713361, + "epoch": 0.04604261796042618, + "grad_norm": 7.53125, + "learning_rate": 4.6042617960426185e-06, + "loss": 2.09043503, + "memory(GiB)": 58.14, + "step": 1815, + "train_speed(iter/s)": 0.448828 + }, + { + "acc": 0.59929843, + "epoch": 0.046169457128361235, + "grad_norm": 5.90625, + "learning_rate": 4.616945712836124e-06, + "loss": 1.95694923, + "memory(GiB)": 58.14, + "step": 1820, + "train_speed(iter/s)": 0.449477 + }, + { + "acc": 0.60815802, + "epoch": 0.046296296296296294, + "grad_norm": 6.21875, + "learning_rate": 4.62962962962963e-06, + "loss": 1.95484276, + "memory(GiB)": 58.14, + "step": 1825, + "train_speed(iter/s)": 0.450129 + }, + { + "acc": 0.60430002, + "epoch": 0.04642313546423135, + "grad_norm": 6.78125, + "learning_rate": 4.6423135464231355e-06, + "loss": 1.99801254, + "memory(GiB)": 58.14, + "step": 1830, + "train_speed(iter/s)": 0.450773 + }, + { + "acc": 0.6199368, + "epoch": 0.04654997463216641, + "grad_norm": 7.21875, + "learning_rate": 4.6549974632166415e-06, + "loss": 1.89275017, + "memory(GiB)": 58.14, + "step": 1835, + "train_speed(iter/s)": 0.451421 + }, + { + "acc": 0.62170529, + "epoch": 0.04667681380010147, + "grad_norm": 5.4375, + "learning_rate": 4.6676813800101475e-06, + "loss": 1.88402481, + "memory(GiB)": 58.14, + "step": 1840, + "train_speed(iter/s)": 0.452058 + }, + { + "acc": 0.59287224, + "epoch": 0.04680365296803653, + "grad_norm": 5.96875, + "learning_rate": 4.6803652968036534e-06, + "loss": 1.98575096, + "memory(GiB)": 58.14, + "step": 1845, + "train_speed(iter/s)": 0.452699 + }, + { + "acc": 0.61426272, + "epoch": 0.046930492135971585, + "grad_norm": 5.28125, + "learning_rate": 4.6930492135971586e-06, + "loss": 1.92058563, + "memory(GiB)": 58.14, + "step": 1850, + "train_speed(iter/s)": 0.453329 + }, + { + "acc": 0.59839396, + "epoch": 0.047057331303906644, + "grad_norm": 5.875, + "learning_rate": 4.705733130390665e-06, + "loss": 1.89336357, + "memory(GiB)": 58.14, + "step": 1855, + "train_speed(iter/s)": 0.453971 + }, + { + "acc": 0.58242912, + "epoch": 0.0471841704718417, + "grad_norm": 4.4375, + "learning_rate": 4.7184170471841705e-06, + "loss": 1.99090919, + "memory(GiB)": 58.14, + "step": 1860, + "train_speed(iter/s)": 0.454607 + }, + { + "acc": 0.59446974, + "epoch": 0.04731100963977676, + "grad_norm": 5.625, + "learning_rate": 4.7311009639776765e-06, + "loss": 1.98187103, + "memory(GiB)": 58.14, + "step": 1865, + "train_speed(iter/s)": 0.455234 + }, + { + "acc": 0.5853014, + "epoch": 0.04743784880771182, + "grad_norm": 5.84375, + "learning_rate": 4.743784880771182e-06, + "loss": 1.9931736, + "memory(GiB)": 58.14, + "step": 1870, + "train_speed(iter/s)": 0.455868 + }, + { + "acc": 0.58511429, + "epoch": 0.04756468797564688, + "grad_norm": 5.75, + "learning_rate": 4.756468797564688e-06, + "loss": 1.96111069, + "memory(GiB)": 58.14, + "step": 1875, + "train_speed(iter/s)": 0.456498 + }, + { + "acc": 0.61168108, + "epoch": 0.047691527143581935, + "grad_norm": 5.53125, + "learning_rate": 4.769152714358194e-06, + "loss": 2.00027313, + "memory(GiB)": 58.14, + "step": 1880, + "train_speed(iter/s)": 0.457124 + }, + { + "acc": 0.61667805, + "epoch": 0.047818366311516994, + "grad_norm": 4.6875, + "learning_rate": 4.7818366311517e-06, + "loss": 1.93731632, + "memory(GiB)": 58.14, + "step": 1885, + "train_speed(iter/s)": 0.457748 + }, + { + "acc": 0.59898443, + "epoch": 0.04794520547945205, + "grad_norm": 5.25, + "learning_rate": 4.7945205479452054e-06, + "loss": 1.96002235, + "memory(GiB)": 58.14, + "step": 1890, + "train_speed(iter/s)": 0.458381 + }, + { + "acc": 0.60575113, + "epoch": 0.04807204464738711, + "grad_norm": 6.0, + "learning_rate": 4.807204464738711e-06, + "loss": 1.91183472, + "memory(GiB)": 58.14, + "step": 1895, + "train_speed(iter/s)": 0.459008 + }, + { + "acc": 0.60391388, + "epoch": 0.04819888381532217, + "grad_norm": 9.25, + "learning_rate": 4.819888381532217e-06, + "loss": 1.93079605, + "memory(GiB)": 58.14, + "step": 1900, + "train_speed(iter/s)": 0.459633 + }, + { + "epoch": 0.04819888381532217, + "eval_acc": 0.5934915195325017, + "eval_loss": 1.9372066259384155, + "eval_runtime": 113.4783, + "eval_samples_per_second": 56.134, + "eval_steps_per_second": 28.067, + "step": 1900 + }, + { + "acc": 0.61051693, + "epoch": 0.04832572298325723, + "grad_norm": 7.21875, + "learning_rate": 4.832572298325723e-06, + "loss": 1.88222332, + "memory(GiB)": 58.14, + "step": 1905, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.60433345, + "epoch": 0.048452562151192285, + "grad_norm": 6.46875, + "learning_rate": 4.845256215119229e-06, + "loss": 1.91181335, + "memory(GiB)": 58.14, + "step": 1910, + "train_speed(iter/s)": 0.448113 + }, + { + "acc": 0.60619183, + "epoch": 0.048579401319127344, + "grad_norm": 5.0625, + "learning_rate": 4.857940131912735e-06, + "loss": 1.91536102, + "memory(GiB)": 58.14, + "step": 1915, + "train_speed(iter/s)": 0.448733 + }, + { + "acc": 0.61726627, + "epoch": 0.0487062404870624, + "grad_norm": 5.6875, + "learning_rate": 4.87062404870624e-06, + "loss": 1.82464237, + "memory(GiB)": 58.14, + "step": 1920, + "train_speed(iter/s)": 0.44935 + }, + { + "acc": 0.59062819, + "epoch": 0.04883307965499746, + "grad_norm": 5.34375, + "learning_rate": 4.883307965499746e-06, + "loss": 1.96244125, + "memory(GiB)": 58.14, + "step": 1925, + "train_speed(iter/s)": 0.44997 + }, + { + "acc": 0.58106537, + "epoch": 0.04895991882293252, + "grad_norm": 5.4375, + "learning_rate": 4.895991882293252e-06, + "loss": 2.04310741, + "memory(GiB)": 58.14, + "step": 1930, + "train_speed(iter/s)": 0.450593 + }, + { + "acc": 0.60148048, + "epoch": 0.04908675799086758, + "grad_norm": 5.6875, + "learning_rate": 4.908675799086758e-06, + "loss": 1.90917358, + "memory(GiB)": 58.14, + "step": 1935, + "train_speed(iter/s)": 0.451201 + }, + { + "acc": 0.60195508, + "epoch": 0.049213597158802636, + "grad_norm": 5.5, + "learning_rate": 4.921359715880264e-06, + "loss": 1.97665138, + "memory(GiB)": 58.14, + "step": 1940, + "train_speed(iter/s)": 0.451816 + }, + { + "acc": 0.60252924, + "epoch": 0.049340436326737694, + "grad_norm": 5.625, + "learning_rate": 4.93404363267377e-06, + "loss": 1.93755741, + "memory(GiB)": 58.14, + "step": 1945, + "train_speed(iter/s)": 0.452425 + }, + { + "acc": 0.62167377, + "epoch": 0.04946727549467275, + "grad_norm": 4.78125, + "learning_rate": 4.946727549467275e-06, + "loss": 1.8892662, + "memory(GiB)": 58.14, + "step": 1950, + "train_speed(iter/s)": 0.453039 + }, + { + "acc": 0.60844717, + "epoch": 0.04959411466260781, + "grad_norm": 5.375, + "learning_rate": 4.959411466260781e-06, + "loss": 1.85567608, + "memory(GiB)": 58.14, + "step": 1955, + "train_speed(iter/s)": 0.453652 + }, + { + "acc": 0.60361886, + "epoch": 0.04972095383054287, + "grad_norm": 5.90625, + "learning_rate": 4.972095383054287e-06, + "loss": 1.98617973, + "memory(GiB)": 58.14, + "step": 1960, + "train_speed(iter/s)": 0.45427 + }, + { + "acc": 0.59965658, + "epoch": 0.04984779299847793, + "grad_norm": 5.625, + "learning_rate": 4.984779299847793e-06, + "loss": 1.91477356, + "memory(GiB)": 58.14, + "step": 1965, + "train_speed(iter/s)": 0.454882 + }, + { + "acc": 0.60240784, + "epoch": 0.049974632166412986, + "grad_norm": 7.5, + "learning_rate": 4.997463216641299e-06, + "loss": 1.9810585, + "memory(GiB)": 58.14, + "step": 1970, + "train_speed(iter/s)": 0.45549 + }, + { + "acc": 0.61513596, + "epoch": 0.050101471334348044, + "grad_norm": 6.125, + "learning_rate": 5.010147133434805e-06, + "loss": 1.89760685, + "memory(GiB)": 58.14, + "step": 1975, + "train_speed(iter/s)": 0.456093 + }, + { + "acc": 0.61477098, + "epoch": 0.0502283105022831, + "grad_norm": 7.3125, + "learning_rate": 5.02283105022831e-06, + "loss": 1.87652206, + "memory(GiB)": 58.14, + "step": 1980, + "train_speed(iter/s)": 0.456694 + }, + { + "acc": 0.5905941, + "epoch": 0.05035514967021816, + "grad_norm": 6.875, + "learning_rate": 5.035514967021817e-06, + "loss": 2.02313042, + "memory(GiB)": 58.14, + "step": 1985, + "train_speed(iter/s)": 0.457293 + }, + { + "acc": 0.60349789, + "epoch": 0.05048198883815322, + "grad_norm": 4.96875, + "learning_rate": 5.048198883815323e-06, + "loss": 1.91415157, + "memory(GiB)": 58.14, + "step": 1990, + "train_speed(iter/s)": 0.457895 + }, + { + "acc": 0.58522735, + "epoch": 0.05060882800608828, + "grad_norm": 6.3125, + "learning_rate": 5.060882800608828e-06, + "loss": 1.99870968, + "memory(GiB)": 58.14, + "step": 1995, + "train_speed(iter/s)": 0.458498 + }, + { + "acc": 0.60471926, + "epoch": 0.050735667174023336, + "grad_norm": 5.78125, + "learning_rate": 5.073566717402334e-06, + "loss": 1.91932545, + "memory(GiB)": 58.14, + "step": 2000, + "train_speed(iter/s)": 0.459101 + }, + { + "epoch": 0.050735667174023336, + "eval_acc": 0.5950685911257032, + "eval_loss": 1.9254412651062012, + "eval_runtime": 111.8276, + "eval_samples_per_second": 56.963, + "eval_steps_per_second": 28.481, + "step": 2000 + }, + { + "acc": 0.59650655, + "epoch": 0.050862506341958394, + "grad_norm": 7.03125, + "learning_rate": 5.086250634195841e-06, + "loss": 1.93570957, + "memory(GiB)": 58.14, + "step": 2005, + "train_speed(iter/s)": 0.447765 + }, + { + "acc": 0.5935904, + "epoch": 0.05098934550989345, + "grad_norm": 5.28125, + "learning_rate": 5.098934550989346e-06, + "loss": 1.91012802, + "memory(GiB)": 58.14, + "step": 2010, + "train_speed(iter/s)": 0.448364 + }, + { + "acc": 0.61528053, + "epoch": 0.05111618467782851, + "grad_norm": 7.03125, + "learning_rate": 5.111618467782852e-06, + "loss": 1.90204659, + "memory(GiB)": 58.14, + "step": 2015, + "train_speed(iter/s)": 0.448956 + }, + { + "acc": 0.61206756, + "epoch": 0.05124302384576357, + "grad_norm": 5.0625, + "learning_rate": 5.124302384576357e-06, + "loss": 1.87378044, + "memory(GiB)": 58.14, + "step": 2020, + "train_speed(iter/s)": 0.449552 + }, + { + "acc": 0.60812511, + "epoch": 0.05136986301369863, + "grad_norm": 5.625, + "learning_rate": 5.136986301369864e-06, + "loss": 1.90723839, + "memory(GiB)": 58.14, + "step": 2025, + "train_speed(iter/s)": 0.450141 + }, + { + "acc": 0.61099358, + "epoch": 0.051496702181633686, + "grad_norm": 5.9375, + "learning_rate": 5.149670218163369e-06, + "loss": 1.87865791, + "memory(GiB)": 58.14, + "step": 2030, + "train_speed(iter/s)": 0.450735 + }, + { + "acc": 0.6056509, + "epoch": 0.051623541349568744, + "grad_norm": 6.78125, + "learning_rate": 5.162354134956875e-06, + "loss": 1.97122879, + "memory(GiB)": 58.14, + "step": 2035, + "train_speed(iter/s)": 0.451324 + }, + { + "acc": 0.61655397, + "epoch": 0.0517503805175038, + "grad_norm": 5.96875, + "learning_rate": 5.175038051750381e-06, + "loss": 1.89492035, + "memory(GiB)": 58.14, + "step": 2040, + "train_speed(iter/s)": 0.451912 + }, + { + "acc": 0.6011075, + "epoch": 0.05187721968543886, + "grad_norm": 8.25, + "learning_rate": 5.187721968543887e-06, + "loss": 1.96894283, + "memory(GiB)": 58.14, + "step": 2045, + "train_speed(iter/s)": 0.452504 + }, + { + "acc": 0.60051799, + "epoch": 0.05200405885337392, + "grad_norm": 5.6875, + "learning_rate": 5.200405885337393e-06, + "loss": 1.87004318, + "memory(GiB)": 58.14, + "step": 2050, + "train_speed(iter/s)": 0.45309 + }, + { + "acc": 0.60961027, + "epoch": 0.05213089802130898, + "grad_norm": 5.875, + "learning_rate": 5.213089802130898e-06, + "loss": 1.87247982, + "memory(GiB)": 58.14, + "step": 2055, + "train_speed(iter/s)": 0.453675 + }, + { + "acc": 0.611905, + "epoch": 0.052257737189244036, + "grad_norm": 4.8125, + "learning_rate": 5.225773718924404e-06, + "loss": 1.88813839, + "memory(GiB)": 58.14, + "step": 2060, + "train_speed(iter/s)": 0.454255 + }, + { + "acc": 0.57968302, + "epoch": 0.052384576357179094, + "grad_norm": 6.4375, + "learning_rate": 5.238457635717911e-06, + "loss": 2.07924652, + "memory(GiB)": 58.14, + "step": 2065, + "train_speed(iter/s)": 0.454835 + }, + { + "acc": 0.60606308, + "epoch": 0.05251141552511415, + "grad_norm": 6.03125, + "learning_rate": 5.251141552511416e-06, + "loss": 1.96282997, + "memory(GiB)": 58.14, + "step": 2070, + "train_speed(iter/s)": 0.455412 + }, + { + "acc": 0.58872027, + "epoch": 0.05263825469304921, + "grad_norm": 4.9375, + "learning_rate": 5.263825469304922e-06, + "loss": 2.00989532, + "memory(GiB)": 58.14, + "step": 2075, + "train_speed(iter/s)": 0.455987 + }, + { + "acc": 0.59235725, + "epoch": 0.05276509386098427, + "grad_norm": 5.53125, + "learning_rate": 5.276509386098427e-06, + "loss": 1.98008156, + "memory(GiB)": 58.14, + "step": 2080, + "train_speed(iter/s)": 0.456567 + }, + { + "acc": 0.60046291, + "epoch": 0.05289193302891933, + "grad_norm": 6.09375, + "learning_rate": 5.289193302891934e-06, + "loss": 1.91245632, + "memory(GiB)": 58.14, + "step": 2085, + "train_speed(iter/s)": 0.457137 + }, + { + "acc": 0.60955739, + "epoch": 0.053018772196854386, + "grad_norm": 4.96875, + "learning_rate": 5.30187721968544e-06, + "loss": 1.88564758, + "memory(GiB)": 58.14, + "step": 2090, + "train_speed(iter/s)": 0.457711 + }, + { + "acc": 0.59646101, + "epoch": 0.053145611364789444, + "grad_norm": 5.0625, + "learning_rate": 5.314561136478945e-06, + "loss": 1.95583382, + "memory(GiB)": 58.14, + "step": 2095, + "train_speed(iter/s)": 0.458281 + }, + { + "acc": 0.60126219, + "epoch": 0.0532724505327245, + "grad_norm": 5.5625, + "learning_rate": 5.327245053272451e-06, + "loss": 1.92181625, + "memory(GiB)": 58.14, + "step": 2100, + "train_speed(iter/s)": 0.458853 + }, + { + "epoch": 0.0532724505327245, + "eval_acc": 0.5962700899077027, + "eval_loss": 1.9146267175674438, + "eval_runtime": 112.2936, + "eval_samples_per_second": 56.726, + "eval_steps_per_second": 28.363, + "step": 2100 + }, + { + "acc": 0.62901831, + "epoch": 0.05339928970065956, + "grad_norm": 5.78125, + "learning_rate": 5.339928970065957e-06, + "loss": 1.81416721, + "memory(GiB)": 58.14, + "step": 2105, + "train_speed(iter/s)": 0.448005 + }, + { + "acc": 0.60744023, + "epoch": 0.05352612886859462, + "grad_norm": 6.96875, + "learning_rate": 5.352612886859463e-06, + "loss": 1.95803032, + "memory(GiB)": 58.14, + "step": 2110, + "train_speed(iter/s)": 0.448573 + }, + { + "acc": 0.60780368, + "epoch": 0.05365296803652968, + "grad_norm": 7.0, + "learning_rate": 5.365296803652969e-06, + "loss": 1.90680275, + "memory(GiB)": 58.14, + "step": 2115, + "train_speed(iter/s)": 0.44914 + }, + { + "acc": 0.60624113, + "epoch": 0.053779807204464736, + "grad_norm": 5.78125, + "learning_rate": 5.377980720446474e-06, + "loss": 1.84182854, + "memory(GiB)": 58.14, + "step": 2120, + "train_speed(iter/s)": 0.449702 + }, + { + "acc": 0.58699579, + "epoch": 0.053906646372399794, + "grad_norm": 7.6875, + "learning_rate": 5.390664637239981e-06, + "loss": 1.93875313, + "memory(GiB)": 58.14, + "step": 2125, + "train_speed(iter/s)": 0.450266 + }, + { + "acc": 0.60300746, + "epoch": 0.05403348554033485, + "grad_norm": 5.71875, + "learning_rate": 5.403348554033486e-06, + "loss": 1.94216881, + "memory(GiB)": 58.14, + "step": 2130, + "train_speed(iter/s)": 0.450828 + }, + { + "acc": 0.61000161, + "epoch": 0.05416032470826991, + "grad_norm": 5.28125, + "learning_rate": 5.416032470826992e-06, + "loss": 1.86070023, + "memory(GiB)": 58.14, + "step": 2135, + "train_speed(iter/s)": 0.451388 + }, + { + "acc": 0.61669793, + "epoch": 0.05428716387620497, + "grad_norm": 5.71875, + "learning_rate": 5.428716387620497e-06, + "loss": 1.87709007, + "memory(GiB)": 58.14, + "step": 2140, + "train_speed(iter/s)": 0.451949 + }, + { + "acc": 0.6074769, + "epoch": 0.05441400304414003, + "grad_norm": 5.84375, + "learning_rate": 5.441400304414004e-06, + "loss": 1.92715111, + "memory(GiB)": 58.14, + "step": 2145, + "train_speed(iter/s)": 0.45251 + }, + { + "acc": 0.60644889, + "epoch": 0.054540842212075086, + "grad_norm": 5.0625, + "learning_rate": 5.45408422120751e-06, + "loss": 1.91768532, + "memory(GiB)": 58.14, + "step": 2150, + "train_speed(iter/s)": 0.453062 + }, + { + "acc": 0.60205259, + "epoch": 0.054667681380010144, + "grad_norm": 5.21875, + "learning_rate": 5.466768138001015e-06, + "loss": 2.00214577, + "memory(GiB)": 58.14, + "step": 2155, + "train_speed(iter/s)": 0.45362 + }, + { + "acc": 0.61421919, + "epoch": 0.0547945205479452, + "grad_norm": 4.84375, + "learning_rate": 5.479452054794521e-06, + "loss": 1.96235409, + "memory(GiB)": 58.14, + "step": 2160, + "train_speed(iter/s)": 0.454171 + }, + { + "acc": 0.59525928, + "epoch": 0.05492135971588026, + "grad_norm": 7.75, + "learning_rate": 5.492135971588028e-06, + "loss": 1.90372467, + "memory(GiB)": 58.14, + "step": 2165, + "train_speed(iter/s)": 0.454724 + }, + { + "acc": 0.58474002, + "epoch": 0.05504819888381532, + "grad_norm": 4.875, + "learning_rate": 5.504819888381533e-06, + "loss": 2.01237469, + "memory(GiB)": 58.14, + "step": 2170, + "train_speed(iter/s)": 0.455278 + }, + { + "acc": 0.61578698, + "epoch": 0.05517503805175038, + "grad_norm": 5.34375, + "learning_rate": 5.517503805175039e-06, + "loss": 1.89675865, + "memory(GiB)": 58.14, + "step": 2175, + "train_speed(iter/s)": 0.455823 + }, + { + "acc": 0.59921474, + "epoch": 0.055301877219685436, + "grad_norm": 5.75, + "learning_rate": 5.530187721968544e-06, + "loss": 1.91646137, + "memory(GiB)": 58.14, + "step": 2180, + "train_speed(iter/s)": 0.456365 + }, + { + "acc": 0.60379057, + "epoch": 0.055428716387620494, + "grad_norm": 6.46875, + "learning_rate": 5.542871638762051e-06, + "loss": 1.95771179, + "memory(GiB)": 58.14, + "step": 2185, + "train_speed(iter/s)": 0.456916 + }, + { + "acc": 0.60007744, + "epoch": 0.05555555555555555, + "grad_norm": 5.46875, + "learning_rate": 5.555555555555557e-06, + "loss": 1.96841164, + "memory(GiB)": 58.14, + "step": 2190, + "train_speed(iter/s)": 0.457461 + }, + { + "acc": 0.61251578, + "epoch": 0.05568239472349061, + "grad_norm": 4.28125, + "learning_rate": 5.568239472349062e-06, + "loss": 1.8744194, + "memory(GiB)": 58.14, + "step": 2195, + "train_speed(iter/s)": 0.458006 + }, + { + "acc": 0.60381517, + "epoch": 0.05580923389142567, + "grad_norm": 7.25, + "learning_rate": 5.580923389142568e-06, + "loss": 1.94256439, + "memory(GiB)": 58.14, + "step": 2200, + "train_speed(iter/s)": 0.458549 + }, + { + "epoch": 0.05580923389142567, + "eval_acc": 0.5976173894807028, + "eval_loss": 1.904833436012268, + "eval_runtime": 112.2781, + "eval_samples_per_second": 56.734, + "eval_steps_per_second": 28.367, + "step": 2200 + }, + { + "acc": 0.60379534, + "epoch": 0.05593607305936073, + "grad_norm": 5.9375, + "learning_rate": 5.593607305936074e-06, + "loss": 1.91352329, + "memory(GiB)": 58.14, + "step": 2205, + "train_speed(iter/s)": 0.448179 + }, + { + "acc": 0.60819073, + "epoch": 0.056062912227295786, + "grad_norm": 4.90625, + "learning_rate": 5.60629122272958e-06, + "loss": 1.88056679, + "memory(GiB)": 58.14, + "step": 2210, + "train_speed(iter/s)": 0.44871 + }, + { + "acc": 0.60504451, + "epoch": 0.056189751395230844, + "grad_norm": 6.09375, + "learning_rate": 5.618975139523085e-06, + "loss": 1.96973114, + "memory(GiB)": 58.14, + "step": 2215, + "train_speed(iter/s)": 0.449243 + }, + { + "acc": 0.61302967, + "epoch": 0.0563165905631659, + "grad_norm": 5.3125, + "learning_rate": 5.631659056316591e-06, + "loss": 1.85189438, + "memory(GiB)": 58.14, + "step": 2220, + "train_speed(iter/s)": 0.449775 + }, + { + "acc": 0.60943823, + "epoch": 0.05644342973110096, + "grad_norm": 6.25, + "learning_rate": 5.644342973110098e-06, + "loss": 1.88922539, + "memory(GiB)": 58.14, + "step": 2225, + "train_speed(iter/s)": 0.450303 + }, + { + "acc": 0.60862675, + "epoch": 0.05657026889903602, + "grad_norm": 6.15625, + "learning_rate": 5.657026889903603e-06, + "loss": 1.90148239, + "memory(GiB)": 58.14, + "step": 2230, + "train_speed(iter/s)": 0.450825 + }, + { + "acc": 0.60232863, + "epoch": 0.05669710806697108, + "grad_norm": 4.8125, + "learning_rate": 5.669710806697109e-06, + "loss": 1.97099495, + "memory(GiB)": 58.14, + "step": 2235, + "train_speed(iter/s)": 0.451352 + }, + { + "acc": 0.6039731, + "epoch": 0.056823947234906136, + "grad_norm": 5.0625, + "learning_rate": 5.682394723490614e-06, + "loss": 1.91322823, + "memory(GiB)": 58.14, + "step": 2240, + "train_speed(iter/s)": 0.451878 + }, + { + "acc": 0.59522576, + "epoch": 0.056950786402841194, + "grad_norm": 5.40625, + "learning_rate": 5.695078640284121e-06, + "loss": 1.94261074, + "memory(GiB)": 58.14, + "step": 2245, + "train_speed(iter/s)": 0.452405 + }, + { + "acc": 0.59672103, + "epoch": 0.05707762557077625, + "grad_norm": 5.4375, + "learning_rate": 5.7077625570776266e-06, + "loss": 1.87163544, + "memory(GiB)": 58.14, + "step": 2250, + "train_speed(iter/s)": 0.452937 + }, + { + "acc": 0.60106378, + "epoch": 0.05720446473871131, + "grad_norm": 5.125, + "learning_rate": 5.720446473871132e-06, + "loss": 1.95121593, + "memory(GiB)": 58.14, + "step": 2255, + "train_speed(iter/s)": 0.453462 + }, + { + "acc": 0.61815538, + "epoch": 0.05733130390664637, + "grad_norm": 6.21875, + "learning_rate": 5.733130390664638e-06, + "loss": 1.80842323, + "memory(GiB)": 58.14, + "step": 2260, + "train_speed(iter/s)": 0.453988 + }, + { + "acc": 0.59947915, + "epoch": 0.05745814307458143, + "grad_norm": 6.15625, + "learning_rate": 5.7458143074581445e-06, + "loss": 1.91878815, + "memory(GiB)": 58.14, + "step": 2265, + "train_speed(iter/s)": 0.45451 + }, + { + "acc": 0.58676567, + "epoch": 0.057584982242516486, + "grad_norm": 5.5, + "learning_rate": 5.75849822425165e-06, + "loss": 1.9787714, + "memory(GiB)": 58.14, + "step": 2270, + "train_speed(iter/s)": 0.455032 + }, + { + "acc": 0.59482555, + "epoch": 0.057711821410451544, + "grad_norm": 6.125, + "learning_rate": 5.7711821410451556e-06, + "loss": 1.97928467, + "memory(GiB)": 58.14, + "step": 2275, + "train_speed(iter/s)": 0.455551 + }, + { + "acc": 0.60976601, + "epoch": 0.0578386605783866, + "grad_norm": 5.46875, + "learning_rate": 5.783866057838661e-06, + "loss": 1.87940559, + "memory(GiB)": 58.14, + "step": 2280, + "train_speed(iter/s)": 0.456071 + }, + { + "acc": 0.58562603, + "epoch": 0.05796549974632166, + "grad_norm": 5.78125, + "learning_rate": 5.7965499746321675e-06, + "loss": 2.00219612, + "memory(GiB)": 58.14, + "step": 2285, + "train_speed(iter/s)": 0.456593 + }, + { + "acc": 0.61238904, + "epoch": 0.05809233891425672, + "grad_norm": 6.03125, + "learning_rate": 5.809233891425673e-06, + "loss": 1.95115395, + "memory(GiB)": 58.14, + "step": 2290, + "train_speed(iter/s)": 0.457111 + }, + { + "acc": 0.60475931, + "epoch": 0.05821917808219178, + "grad_norm": 5.5625, + "learning_rate": 5.821917808219179e-06, + "loss": 1.91742439, + "memory(GiB)": 58.14, + "step": 2295, + "train_speed(iter/s)": 0.457632 + }, + { + "acc": 0.60946064, + "epoch": 0.058346017250126836, + "grad_norm": 5.28125, + "learning_rate": 5.834601725012684e-06, + "loss": 1.85216446, + "memory(GiB)": 58.14, + "step": 2300, + "train_speed(iter/s)": 0.458151 + }, + { + "epoch": 0.058346017250126836, + "eval_acc": 0.5988907442399288, + "eval_loss": 1.895876169204712, + "eval_runtime": 112.8747, + "eval_samples_per_second": 56.434, + "eval_steps_per_second": 28.217, + "step": 2300 + }, + { + "acc": 0.59721847, + "epoch": 0.058472856418061894, + "grad_norm": 6.8125, + "learning_rate": 5.8472856418061905e-06, + "loss": 1.93088341, + "memory(GiB)": 58.14, + "step": 2305, + "train_speed(iter/s)": 0.448201 + }, + { + "acc": 0.60022249, + "epoch": 0.05859969558599695, + "grad_norm": 5.9375, + "learning_rate": 5.8599695585996965e-06, + "loss": 1.91122932, + "memory(GiB)": 58.14, + "step": 2310, + "train_speed(iter/s)": 0.448712 + }, + { + "acc": 0.60493164, + "epoch": 0.05872653475393201, + "grad_norm": 4.875, + "learning_rate": 5.872653475393202e-06, + "loss": 1.86035271, + "memory(GiB)": 58.14, + "step": 2315, + "train_speed(iter/s)": 0.449222 + }, + { + "acc": 0.61038189, + "epoch": 0.05885337392186707, + "grad_norm": 5.5, + "learning_rate": 5.8853373921867076e-06, + "loss": 1.96235199, + "memory(GiB)": 58.14, + "step": 2320, + "train_speed(iter/s)": 0.449733 + }, + { + "acc": 0.61910405, + "epoch": 0.05898021308980213, + "grad_norm": 6.03125, + "learning_rate": 5.898021308980214e-06, + "loss": 1.84855499, + "memory(GiB)": 58.14, + "step": 2325, + "train_speed(iter/s)": 0.450243 + }, + { + "acc": 0.59141207, + "epoch": 0.059107052257737186, + "grad_norm": 5.96875, + "learning_rate": 5.9107052257737195e-06, + "loss": 2.02407112, + "memory(GiB)": 58.14, + "step": 2330, + "train_speed(iter/s)": 0.450747 + }, + { + "acc": 0.5989542, + "epoch": 0.059233891425672244, + "grad_norm": 5.09375, + "learning_rate": 5.9233891425672255e-06, + "loss": 1.97891426, + "memory(GiB)": 58.14, + "step": 2335, + "train_speed(iter/s)": 0.451251 + }, + { + "acc": 0.60870714, + "epoch": 0.0593607305936073, + "grad_norm": 5.34375, + "learning_rate": 5.936073059360731e-06, + "loss": 1.88683434, + "memory(GiB)": 58.14, + "step": 2340, + "train_speed(iter/s)": 0.451761 + }, + { + "acc": 0.60967364, + "epoch": 0.05948756976154236, + "grad_norm": 5.46875, + "learning_rate": 5.948756976154237e-06, + "loss": 1.92153244, + "memory(GiB)": 58.14, + "step": 2345, + "train_speed(iter/s)": 0.452268 + }, + { + "acc": 0.6127944, + "epoch": 0.05961440892947742, + "grad_norm": 5.46875, + "learning_rate": 5.961440892947743e-06, + "loss": 1.90284595, + "memory(GiB)": 58.14, + "step": 2350, + "train_speed(iter/s)": 0.45277 + }, + { + "acc": 0.62133894, + "epoch": 0.05974124809741248, + "grad_norm": 5.65625, + "learning_rate": 5.9741248097412485e-06, + "loss": 1.8314642, + "memory(GiB)": 58.14, + "step": 2355, + "train_speed(iter/s)": 0.45327 + }, + { + "acc": 0.60800447, + "epoch": 0.059868087265347536, + "grad_norm": 5.09375, + "learning_rate": 5.9868087265347545e-06, + "loss": 1.87077923, + "memory(GiB)": 58.14, + "step": 2360, + "train_speed(iter/s)": 0.453773 + }, + { + "acc": 0.61515021, + "epoch": 0.059994926433282594, + "grad_norm": 5.625, + "learning_rate": 5.99949264332826e-06, + "loss": 1.84514027, + "memory(GiB)": 58.14, + "step": 2365, + "train_speed(iter/s)": 0.454271 + }, + { + "acc": 0.60447173, + "epoch": 0.06012176560121765, + "grad_norm": 4.75, + "learning_rate": 6.012176560121766e-06, + "loss": 1.89534531, + "memory(GiB)": 58.14, + "step": 2370, + "train_speed(iter/s)": 0.454769 + }, + { + "acc": 0.61261654, + "epoch": 0.06024860476915271, + "grad_norm": 6.375, + "learning_rate": 6.0248604769152715e-06, + "loss": 1.88374252, + "memory(GiB)": 58.14, + "step": 2375, + "train_speed(iter/s)": 0.455266 + }, + { + "acc": 0.62044411, + "epoch": 0.06037544393708777, + "grad_norm": 6.34375, + "learning_rate": 6.0375443937087775e-06, + "loss": 1.80461407, + "memory(GiB)": 58.14, + "step": 2380, + "train_speed(iter/s)": 0.455765 + }, + { + "acc": 0.60993481, + "epoch": 0.06050228310502283, + "grad_norm": 5.125, + "learning_rate": 6.050228310502284e-06, + "loss": 1.96366043, + "memory(GiB)": 58.14, + "step": 2385, + "train_speed(iter/s)": 0.456265 + }, + { + "acc": 0.61328502, + "epoch": 0.060629122272957886, + "grad_norm": 5.125, + "learning_rate": 6.062912227295789e-06, + "loss": 1.8827301, + "memory(GiB)": 58.14, + "step": 2390, + "train_speed(iter/s)": 0.45676 + }, + { + "acc": 0.61052256, + "epoch": 0.060755961440892944, + "grad_norm": 5.3125, + "learning_rate": 6.075596144089295e-06, + "loss": 1.82140503, + "memory(GiB)": 58.14, + "step": 2395, + "train_speed(iter/s)": 0.457258 + }, + { + "acc": 0.60249681, + "epoch": 0.060882800608828, + "grad_norm": 6.125, + "learning_rate": 6.0882800608828005e-06, + "loss": 1.98190269, + "memory(GiB)": 58.14, + "step": 2400, + "train_speed(iter/s)": 0.457756 + }, + { + "epoch": 0.060882800608828, + "eval_acc": 0.6000379332716987, + "eval_loss": 1.8866535425186157, + "eval_runtime": 112.4845, + "eval_samples_per_second": 56.63, + "eval_steps_per_second": 28.315, + "step": 2400 + }, + { + "acc": 0.59910173, + "epoch": 0.06100963977676306, + "grad_norm": 5.03125, + "learning_rate": 6.100963977676307e-06, + "loss": 1.95506439, + "memory(GiB)": 58.14, + "step": 2405, + "train_speed(iter/s)": 0.448254 + }, + { + "acc": 0.59968038, + "epoch": 0.06113647894469812, + "grad_norm": 5.25, + "learning_rate": 6.113647894469813e-06, + "loss": 1.93774776, + "memory(GiB)": 58.14, + "step": 2410, + "train_speed(iter/s)": 0.448743 + }, + { + "acc": 0.61921587, + "epoch": 0.06126331811263318, + "grad_norm": 5.75, + "learning_rate": 6.126331811263318e-06, + "loss": 1.8587225, + "memory(GiB)": 58.14, + "step": 2415, + "train_speed(iter/s)": 0.449227 + }, + { + "acc": 0.63117027, + "epoch": 0.061390157280568236, + "grad_norm": 5.34375, + "learning_rate": 6.139015728056824e-06, + "loss": 1.8523159, + "memory(GiB)": 58.14, + "step": 2420, + "train_speed(iter/s)": 0.449711 + }, + { + "acc": 0.60653915, + "epoch": 0.061516996448503294, + "grad_norm": 5.625, + "learning_rate": 6.151699644850331e-06, + "loss": 1.89694252, + "memory(GiB)": 58.14, + "step": 2425, + "train_speed(iter/s)": 0.450196 + }, + { + "acc": 0.62044978, + "epoch": 0.06164383561643835, + "grad_norm": 4.6875, + "learning_rate": 6.164383561643836e-06, + "loss": 1.83068085, + "memory(GiB)": 58.14, + "step": 2430, + "train_speed(iter/s)": 0.45068 + }, + { + "acc": 0.59551821, + "epoch": 0.06177067478437341, + "grad_norm": 5.3125, + "learning_rate": 6.177067478437342e-06, + "loss": 1.92752857, + "memory(GiB)": 58.14, + "step": 2435, + "train_speed(iter/s)": 0.451163 + }, + { + "acc": 0.60254188, + "epoch": 0.06189751395230847, + "grad_norm": 5.4375, + "learning_rate": 6.189751395230847e-06, + "loss": 1.91536694, + "memory(GiB)": 58.14, + "step": 2440, + "train_speed(iter/s)": 0.451641 + }, + { + "acc": 0.60145302, + "epoch": 0.06202435312024353, + "grad_norm": 7.375, + "learning_rate": 6.202435312024354e-06, + "loss": 1.97209625, + "memory(GiB)": 58.14, + "step": 2445, + "train_speed(iter/s)": 0.452119 + }, + { + "acc": 0.62152462, + "epoch": 0.062151192288178586, + "grad_norm": 6.3125, + "learning_rate": 6.215119228817859e-06, + "loss": 1.79224625, + "memory(GiB)": 58.14, + "step": 2450, + "train_speed(iter/s)": 0.452598 + }, + { + "acc": 0.59656768, + "epoch": 0.062278031456113644, + "grad_norm": 5.21875, + "learning_rate": 6.227803145611365e-06, + "loss": 1.92446404, + "memory(GiB)": 58.14, + "step": 2455, + "train_speed(iter/s)": 0.453077 + }, + { + "acc": 0.60080857, + "epoch": 0.0624048706240487, + "grad_norm": 5.78125, + "learning_rate": 6.24048706240487e-06, + "loss": 1.92613869, + "memory(GiB)": 58.14, + "step": 2460, + "train_speed(iter/s)": 0.453559 + }, + { + "acc": 0.62344441, + "epoch": 0.06253170979198376, + "grad_norm": 6.71875, + "learning_rate": 6.253170979198377e-06, + "loss": 1.87104588, + "memory(GiB)": 58.14, + "step": 2465, + "train_speed(iter/s)": 0.45403 + }, + { + "acc": 0.59803681, + "epoch": 0.06265854895991882, + "grad_norm": 5.09375, + "learning_rate": 6.265854895991883e-06, + "loss": 1.9206789, + "memory(GiB)": 58.14, + "step": 2470, + "train_speed(iter/s)": 0.454501 + }, + { + "acc": 0.61187105, + "epoch": 0.06278538812785388, + "grad_norm": 5.875, + "learning_rate": 6.278538812785388e-06, + "loss": 1.92976303, + "memory(GiB)": 58.14, + "step": 2475, + "train_speed(iter/s)": 0.454975 + }, + { + "acc": 0.61190963, + "epoch": 0.06291222729578894, + "grad_norm": 5.34375, + "learning_rate": 6.291222729578894e-06, + "loss": 1.86850891, + "memory(GiB)": 58.14, + "step": 2480, + "train_speed(iter/s)": 0.455445 + }, + { + "acc": 0.61170764, + "epoch": 0.063039066463724, + "grad_norm": 5.1875, + "learning_rate": 6.303906646372401e-06, + "loss": 1.94937782, + "memory(GiB)": 58.14, + "step": 2485, + "train_speed(iter/s)": 0.455915 + }, + { + "acc": 0.6186388, + "epoch": 0.06316590563165905, + "grad_norm": 5.5625, + "learning_rate": 6.316590563165906e-06, + "loss": 1.83750896, + "memory(GiB)": 58.14, + "step": 2490, + "train_speed(iter/s)": 0.456387 + }, + { + "acc": 0.6037365, + "epoch": 0.06329274479959411, + "grad_norm": 6.0, + "learning_rate": 6.329274479959412e-06, + "loss": 1.88511868, + "memory(GiB)": 58.14, + "step": 2495, + "train_speed(iter/s)": 0.456861 + }, + { + "acc": 0.59893236, + "epoch": 0.06341958396752917, + "grad_norm": 5.125, + "learning_rate": 6.341958396752917e-06, + "loss": 1.95585556, + "memory(GiB)": 58.14, + "step": 2500, + "train_speed(iter/s)": 0.457332 + }, + { + "epoch": 0.06341958396752917, + "eval_acc": 0.6012294056382712, + "eval_loss": 1.8778204917907715, + "eval_runtime": 113.2563, + "eval_samples_per_second": 56.244, + "eval_steps_per_second": 28.122, + "step": 2500 + }, + { + "acc": 0.61564579, + "epoch": 0.06354642313546423, + "grad_norm": 7.46875, + "learning_rate": 6.354642313546424e-06, + "loss": 1.92390862, + "memory(GiB)": 58.14, + "step": 2505, + "train_speed(iter/s)": 0.448167 + }, + { + "acc": 0.59960327, + "epoch": 0.06367326230339929, + "grad_norm": 4.8125, + "learning_rate": 6.36732623033993e-06, + "loss": 1.93815136, + "memory(GiB)": 58.14, + "step": 2510, + "train_speed(iter/s)": 0.448632 + }, + { + "acc": 0.61533532, + "epoch": 0.06380010147133434, + "grad_norm": 5.75, + "learning_rate": 6.380010147133435e-06, + "loss": 1.80193939, + "memory(GiB)": 58.14, + "step": 2515, + "train_speed(iter/s)": 0.4491 + }, + { + "acc": 0.61330819, + "epoch": 0.0639269406392694, + "grad_norm": 5.125, + "learning_rate": 6.392694063926941e-06, + "loss": 1.90321312, + "memory(GiB)": 58.14, + "step": 2520, + "train_speed(iter/s)": 0.449566 + }, + { + "acc": 0.59923172, + "epoch": 0.06405377980720446, + "grad_norm": 5.3125, + "learning_rate": 6.405377980720447e-06, + "loss": 1.9303793, + "memory(GiB)": 58.14, + "step": 2525, + "train_speed(iter/s)": 0.450034 + }, + { + "acc": 0.62410908, + "epoch": 0.06418061897513952, + "grad_norm": 6.4375, + "learning_rate": 6.418061897513953e-06, + "loss": 1.76170559, + "memory(GiB)": 58.14, + "step": 2530, + "train_speed(iter/s)": 0.450506 + }, + { + "acc": 0.60147963, + "epoch": 0.06430745814307458, + "grad_norm": 6.6875, + "learning_rate": 6.430745814307458e-06, + "loss": 1.850107, + "memory(GiB)": 58.14, + "step": 2535, + "train_speed(iter/s)": 0.450976 + }, + { + "acc": 0.60691957, + "epoch": 0.06443429731100964, + "grad_norm": 5.21875, + "learning_rate": 6.443429731100964e-06, + "loss": 1.90494766, + "memory(GiB)": 58.14, + "step": 2540, + "train_speed(iter/s)": 0.451446 + }, + { + "acc": 0.63060875, + "epoch": 0.0645611364789447, + "grad_norm": 5.84375, + "learning_rate": 6.456113647894471e-06, + "loss": 1.83058548, + "memory(GiB)": 58.14, + "step": 2545, + "train_speed(iter/s)": 0.451916 + }, + { + "acc": 0.61419878, + "epoch": 0.06468797564687975, + "grad_norm": 4.34375, + "learning_rate": 6.468797564687976e-06, + "loss": 1.8514328, + "memory(GiB)": 58.14, + "step": 2550, + "train_speed(iter/s)": 0.452382 + }, + { + "acc": 0.62784376, + "epoch": 0.06481481481481481, + "grad_norm": 6.5, + "learning_rate": 6.481481481481482e-06, + "loss": 1.78377953, + "memory(GiB)": 58.14, + "step": 2555, + "train_speed(iter/s)": 0.452853 + }, + { + "acc": 0.6120739, + "epoch": 0.06494165398274987, + "grad_norm": 5.21875, + "learning_rate": 6.494165398274987e-06, + "loss": 1.87635803, + "memory(GiB)": 58.14, + "step": 2560, + "train_speed(iter/s)": 0.453324 + }, + { + "acc": 0.59721899, + "epoch": 0.06506849315068493, + "grad_norm": 5.5625, + "learning_rate": 6.506849315068494e-06, + "loss": 1.90338287, + "memory(GiB)": 58.14, + "step": 2565, + "train_speed(iter/s)": 0.453793 + }, + { + "acc": 0.60396547, + "epoch": 0.06519533231861999, + "grad_norm": 5.46875, + "learning_rate": 6.519533231862e-06, + "loss": 1.90700645, + "memory(GiB)": 58.14, + "step": 2570, + "train_speed(iter/s)": 0.454265 + }, + { + "acc": 0.62037644, + "epoch": 0.06532217148655504, + "grad_norm": 5.125, + "learning_rate": 6.532217148655505e-06, + "loss": 1.80530777, + "memory(GiB)": 58.14, + "step": 2575, + "train_speed(iter/s)": 0.454731 + }, + { + "acc": 0.60401158, + "epoch": 0.0654490106544901, + "grad_norm": 4.84375, + "learning_rate": 6.544901065449011e-06, + "loss": 2.01804085, + "memory(GiB)": 58.14, + "step": 2580, + "train_speed(iter/s)": 0.455199 + }, + { + "acc": 0.59663267, + "epoch": 0.06557584982242516, + "grad_norm": 7.4375, + "learning_rate": 6.557584982242518e-06, + "loss": 1.97994652, + "memory(GiB)": 58.14, + "step": 2585, + "train_speed(iter/s)": 0.455666 + }, + { + "acc": 0.60921412, + "epoch": 0.06570268899036022, + "grad_norm": 5.46875, + "learning_rate": 6.570268899036023e-06, + "loss": 1.87197304, + "memory(GiB)": 58.14, + "step": 2590, + "train_speed(iter/s)": 0.456131 + }, + { + "acc": 0.61398516, + "epoch": 0.06582952815829528, + "grad_norm": 5.3125, + "learning_rate": 6.582952815829529e-06, + "loss": 1.9007225, + "memory(GiB)": 58.14, + "step": 2595, + "train_speed(iter/s)": 0.456598 + }, + { + "acc": 0.61256599, + "epoch": 0.06595636732623034, + "grad_norm": 5.90625, + "learning_rate": 6.595636732623034e-06, + "loss": 1.96448936, + "memory(GiB)": 58.14, + "step": 2600, + "train_speed(iter/s)": 0.45706 + }, + { + "epoch": 0.06595636732623034, + "eval_acc": 0.6025194710898756, + "eval_loss": 1.8688422441482544, + "eval_runtime": 113.5003, + "eval_samples_per_second": 56.123, + "eval_steps_per_second": 28.062, + "step": 2600 + }, + { + "acc": 0.61291656, + "epoch": 0.0660832064941654, + "grad_norm": 6.84375, + "learning_rate": 6.608320649416541e-06, + "loss": 1.83021393, + "memory(GiB)": 58.14, + "step": 2605, + "train_speed(iter/s)": 0.448228 + }, + { + "acc": 0.60811567, + "epoch": 0.06621004566210045, + "grad_norm": 6.59375, + "learning_rate": 6.621004566210046e-06, + "loss": 1.9022419, + "memory(GiB)": 58.14, + "step": 2610, + "train_speed(iter/s)": 0.448683 + }, + { + "acc": 0.61229768, + "epoch": 0.06633688483003551, + "grad_norm": 6.0625, + "learning_rate": 6.633688483003552e-06, + "loss": 1.83395748, + "memory(GiB)": 58.14, + "step": 2615, + "train_speed(iter/s)": 0.449138 + }, + { + "acc": 0.60656805, + "epoch": 0.06646372399797057, + "grad_norm": 5.75, + "learning_rate": 6.646372399797057e-06, + "loss": 1.85824127, + "memory(GiB)": 58.14, + "step": 2620, + "train_speed(iter/s)": 0.449595 + }, + { + "acc": 0.60877857, + "epoch": 0.06659056316590563, + "grad_norm": 6.65625, + "learning_rate": 6.659056316590564e-06, + "loss": 1.88488064, + "memory(GiB)": 58.14, + "step": 2625, + "train_speed(iter/s)": 0.450046 + }, + { + "acc": 0.61313748, + "epoch": 0.06671740233384069, + "grad_norm": 6.46875, + "learning_rate": 6.67174023338407e-06, + "loss": 1.90876999, + "memory(GiB)": 58.14, + "step": 2630, + "train_speed(iter/s)": 0.450502 + }, + { + "acc": 0.61640196, + "epoch": 0.06684424150177574, + "grad_norm": 7.0625, + "learning_rate": 6.684424150177575e-06, + "loss": 1.88373146, + "memory(GiB)": 58.14, + "step": 2635, + "train_speed(iter/s)": 0.450952 + }, + { + "acc": 0.6129179, + "epoch": 0.0669710806697108, + "grad_norm": 6.21875, + "learning_rate": 6.697108066971081e-06, + "loss": 1.94215336, + "memory(GiB)": 58.14, + "step": 2640, + "train_speed(iter/s)": 0.451405 + }, + { + "acc": 0.61619987, + "epoch": 0.06709791983764586, + "grad_norm": 5.40625, + "learning_rate": 6.709791983764588e-06, + "loss": 1.83145981, + "memory(GiB)": 58.14, + "step": 2645, + "train_speed(iter/s)": 0.451855 + }, + { + "acc": 0.61294098, + "epoch": 0.06722475900558092, + "grad_norm": 6.59375, + "learning_rate": 6.722475900558093e-06, + "loss": 1.88779945, + "memory(GiB)": 58.14, + "step": 2650, + "train_speed(iter/s)": 0.452301 + }, + { + "acc": 0.6304718, + "epoch": 0.06735159817351598, + "grad_norm": 6.65625, + "learning_rate": 6.735159817351599e-06, + "loss": 1.83628044, + "memory(GiB)": 58.14, + "step": 2655, + "train_speed(iter/s)": 0.452744 + }, + { + "acc": 0.61725063, + "epoch": 0.06747843734145104, + "grad_norm": 6.625, + "learning_rate": 6.747843734145104e-06, + "loss": 1.85713902, + "memory(GiB)": 58.14, + "step": 2660, + "train_speed(iter/s)": 0.45319 + }, + { + "acc": 0.60948086, + "epoch": 0.0676052765093861, + "grad_norm": 4.96875, + "learning_rate": 6.760527650938611e-06, + "loss": 1.84074936, + "memory(GiB)": 58.14, + "step": 2665, + "train_speed(iter/s)": 0.453631 + }, + { + "acc": 0.60286841, + "epoch": 0.06773211567732115, + "grad_norm": 5.15625, + "learning_rate": 6.773211567732117e-06, + "loss": 1.91347103, + "memory(GiB)": 58.14, + "step": 2670, + "train_speed(iter/s)": 0.454071 + }, + { + "acc": 0.61806631, + "epoch": 0.06785895484525621, + "grad_norm": 5.0, + "learning_rate": 6.785895484525622e-06, + "loss": 1.9417408, + "memory(GiB)": 58.14, + "step": 2675, + "train_speed(iter/s)": 0.454506 + }, + { + "acc": 0.62566824, + "epoch": 0.06798579401319127, + "grad_norm": 5.625, + "learning_rate": 6.798579401319128e-06, + "loss": 1.82996368, + "memory(GiB)": 58.14, + "step": 2680, + "train_speed(iter/s)": 0.454948 + }, + { + "acc": 0.61032419, + "epoch": 0.06811263318112633, + "grad_norm": 5.15625, + "learning_rate": 6.811263318112634e-06, + "loss": 1.92929764, + "memory(GiB)": 58.14, + "step": 2685, + "train_speed(iter/s)": 0.45539 + }, + { + "acc": 0.60257921, + "epoch": 0.06823947234906139, + "grad_norm": 5.53125, + "learning_rate": 6.82394723490614e-06, + "loss": 1.8732523, + "memory(GiB)": 58.14, + "step": 2690, + "train_speed(iter/s)": 0.455831 + }, + { + "acc": 0.60895119, + "epoch": 0.06836631151699644, + "grad_norm": 5.09375, + "learning_rate": 6.836631151699645e-06, + "loss": 1.89161205, + "memory(GiB)": 58.14, + "step": 2695, + "train_speed(iter/s)": 0.456268 + }, + { + "acc": 0.5987586, + "epoch": 0.0684931506849315, + "grad_norm": 5.375, + "learning_rate": 6.849315068493151e-06, + "loss": 1.94489403, + "memory(GiB)": 58.14, + "step": 2700, + "train_speed(iter/s)": 0.456705 + }, + { + "epoch": 0.0684931506849315, + "eval_acc": 0.6035881198674675, + "eval_loss": 1.860156536102295, + "eval_runtime": 111.9733, + "eval_samples_per_second": 56.889, + "eval_steps_per_second": 28.444, + "step": 2700 + }, + { + "acc": 0.60492058, + "epoch": 0.06861998985286656, + "grad_norm": 5.4375, + "learning_rate": 6.861998985286658e-06, + "loss": 1.90669708, + "memory(GiB)": 58.14, + "step": 2705, + "train_speed(iter/s)": 0.448321 + }, + { + "acc": 0.60875154, + "epoch": 0.06874682902080162, + "grad_norm": 6.125, + "learning_rate": 6.874682902080163e-06, + "loss": 1.98989906, + "memory(GiB)": 58.14, + "step": 2710, + "train_speed(iter/s)": 0.448757 + }, + { + "acc": 0.61828809, + "epoch": 0.06887366818873668, + "grad_norm": 5.625, + "learning_rate": 6.887366818873669e-06, + "loss": 1.82437973, + "memory(GiB)": 58.14, + "step": 2715, + "train_speed(iter/s)": 0.449187 + }, + { + "acc": 0.61650009, + "epoch": 0.06900050735667174, + "grad_norm": 5.90625, + "learning_rate": 6.900050735667174e-06, + "loss": 1.8771843, + "memory(GiB)": 58.14, + "step": 2720, + "train_speed(iter/s)": 0.449621 + }, + { + "acc": 0.6184906, + "epoch": 0.0691273465246068, + "grad_norm": 4.71875, + "learning_rate": 6.912734652460681e-06, + "loss": 1.87500553, + "memory(GiB)": 58.14, + "step": 2725, + "train_speed(iter/s)": 0.450054 + }, + { + "acc": 0.60677981, + "epoch": 0.06925418569254185, + "grad_norm": 7.71875, + "learning_rate": 6.925418569254187e-06, + "loss": 1.84384727, + "memory(GiB)": 58.14, + "step": 2730, + "train_speed(iter/s)": 0.450484 + }, + { + "acc": 0.60466657, + "epoch": 0.06938102486047691, + "grad_norm": 5.28125, + "learning_rate": 6.938102486047692e-06, + "loss": 1.83755951, + "memory(GiB)": 58.14, + "step": 2735, + "train_speed(iter/s)": 0.450913 + }, + { + "acc": 0.60684156, + "epoch": 0.06950786402841197, + "grad_norm": 5.3125, + "learning_rate": 6.950786402841198e-06, + "loss": 1.8586935, + "memory(GiB)": 58.14, + "step": 2740, + "train_speed(iter/s)": 0.451341 + }, + { + "acc": 0.62611828, + "epoch": 0.06963470319634703, + "grad_norm": 5.59375, + "learning_rate": 6.9634703196347046e-06, + "loss": 1.82037735, + "memory(GiB)": 58.14, + "step": 2745, + "train_speed(iter/s)": 0.451772 + }, + { + "acc": 0.61206102, + "epoch": 0.06976154236428209, + "grad_norm": 5.3125, + "learning_rate": 6.97615423642821e-06, + "loss": 1.90325394, + "memory(GiB)": 58.14, + "step": 2750, + "train_speed(iter/s)": 0.452204 + }, + { + "acc": 0.60792751, + "epoch": 0.06988838153221714, + "grad_norm": 5.75, + "learning_rate": 6.988838153221716e-06, + "loss": 1.9160717, + "memory(GiB)": 58.14, + "step": 2755, + "train_speed(iter/s)": 0.452632 + }, + { + "acc": 0.6066947, + "epoch": 0.0700152207001522, + "grad_norm": 4.84375, + "learning_rate": 7.001522070015221e-06, + "loss": 1.88898773, + "memory(GiB)": 58.14, + "step": 2760, + "train_speed(iter/s)": 0.453052 + }, + { + "acc": 0.61785488, + "epoch": 0.07014205986808726, + "grad_norm": 6.09375, + "learning_rate": 7.014205986808728e-06, + "loss": 1.93794117, + "memory(GiB)": 58.14, + "step": 2765, + "train_speed(iter/s)": 0.453475 + }, + { + "acc": 0.61566505, + "epoch": 0.07026889903602232, + "grad_norm": 5.34375, + "learning_rate": 7.026889903602233e-06, + "loss": 1.86988449, + "memory(GiB)": 58.14, + "step": 2770, + "train_speed(iter/s)": 0.453899 + }, + { + "acc": 0.61492205, + "epoch": 0.07039573820395738, + "grad_norm": 7.53125, + "learning_rate": 7.039573820395739e-06, + "loss": 1.90968246, + "memory(GiB)": 58.14, + "step": 2775, + "train_speed(iter/s)": 0.454324 + }, + { + "acc": 0.59384127, + "epoch": 0.07052257737189244, + "grad_norm": 5.53125, + "learning_rate": 7.052257737189245e-06, + "loss": 1.92777691, + "memory(GiB)": 58.14, + "step": 2780, + "train_speed(iter/s)": 0.454749 + }, + { + "acc": 0.62231426, + "epoch": 0.0706494165398275, + "grad_norm": 6.0625, + "learning_rate": 7.064941653982751e-06, + "loss": 1.89345779, + "memory(GiB)": 58.14, + "step": 2785, + "train_speed(iter/s)": 0.455167 + }, + { + "acc": 0.6072772, + "epoch": 0.07077625570776255, + "grad_norm": 5.40625, + "learning_rate": 7.077625570776257e-06, + "loss": 1.89978943, + "memory(GiB)": 58.14, + "step": 2790, + "train_speed(iter/s)": 0.455588 + }, + { + "acc": 0.61941037, + "epoch": 0.07090309487569761, + "grad_norm": 6.5625, + "learning_rate": 7.090309487569762e-06, + "loss": 1.80426388, + "memory(GiB)": 58.14, + "step": 2795, + "train_speed(iter/s)": 0.456008 + }, + { + "acc": 0.63117952, + "epoch": 0.07102993404363267, + "grad_norm": 6.125, + "learning_rate": 7.102993404363268e-06, + "loss": 1.86595287, + "memory(GiB)": 58.14, + "step": 2800, + "train_speed(iter/s)": 0.456427 + }, + { + "epoch": 0.07102993404363267, + "eval_acc": 0.6048084181783925, + "eval_loss": 1.8509750366210938, + "eval_runtime": 112.5821, + "eval_samples_per_second": 56.581, + "eval_steps_per_second": 28.29, + "step": 2800 + }, + { + "acc": 0.6240252, + "epoch": 0.07115677321156773, + "grad_norm": 5.90625, + "learning_rate": 7.1156773211567745e-06, + "loss": 1.84464493, + "memory(GiB)": 58.14, + "step": 2805, + "train_speed(iter/s)": 0.4483 + }, + { + "acc": 0.62661514, + "epoch": 0.07128361237950279, + "grad_norm": 5.75, + "learning_rate": 7.12836123795028e-06, + "loss": 1.89510536, + "memory(GiB)": 58.14, + "step": 2810, + "train_speed(iter/s)": 0.448716 + }, + { + "acc": 0.59872684, + "epoch": 0.07141045154743784, + "grad_norm": 5.0625, + "learning_rate": 7.1410451547437856e-06, + "loss": 1.89223671, + "memory(GiB)": 58.14, + "step": 2815, + "train_speed(iter/s)": 0.449133 + }, + { + "acc": 0.60908093, + "epoch": 0.0715372907153729, + "grad_norm": 5.46875, + "learning_rate": 7.153729071537291e-06, + "loss": 1.81616497, + "memory(GiB)": 58.14, + "step": 2820, + "train_speed(iter/s)": 0.449549 + }, + { + "acc": 0.62107067, + "epoch": 0.07166412988330796, + "grad_norm": 5.03125, + "learning_rate": 7.1664129883307975e-06, + "loss": 1.83826218, + "memory(GiB)": 58.14, + "step": 2825, + "train_speed(iter/s)": 0.449966 + }, + { + "acc": 0.63013668, + "epoch": 0.07179096905124302, + "grad_norm": 5.34375, + "learning_rate": 7.1790969051243035e-06, + "loss": 1.81878738, + "memory(GiB)": 58.14, + "step": 2830, + "train_speed(iter/s)": 0.450382 + }, + { + "acc": 0.59435673, + "epoch": 0.07191780821917808, + "grad_norm": 6.90625, + "learning_rate": 7.191780821917809e-06, + "loss": 1.97414513, + "memory(GiB)": 58.14, + "step": 2835, + "train_speed(iter/s)": 0.450797 + }, + { + "acc": 0.60926948, + "epoch": 0.07204464738711314, + "grad_norm": 5.9375, + "learning_rate": 7.2044647387113146e-06, + "loss": 1.91032143, + "memory(GiB)": 58.14, + "step": 2840, + "train_speed(iter/s)": 0.451211 + }, + { + "acc": 0.62160091, + "epoch": 0.0721714865550482, + "grad_norm": 4.8125, + "learning_rate": 7.2171486555048205e-06, + "loss": 1.81906433, + "memory(GiB)": 58.14, + "step": 2845, + "train_speed(iter/s)": 0.451621 + }, + { + "acc": 0.60461559, + "epoch": 0.07229832572298325, + "grad_norm": 5.21875, + "learning_rate": 7.2298325722983265e-06, + "loss": 1.88771286, + "memory(GiB)": 58.14, + "step": 2850, + "train_speed(iter/s)": 0.452034 + }, + { + "acc": 0.63199983, + "epoch": 0.07242516489091831, + "grad_norm": 5.90625, + "learning_rate": 7.242516489091832e-06, + "loss": 1.86336899, + "memory(GiB)": 58.14, + "step": 2855, + "train_speed(iter/s)": 0.452444 + }, + { + "acc": 0.62258148, + "epoch": 0.07255200405885337, + "grad_norm": 6.21875, + "learning_rate": 7.2552004058853376e-06, + "loss": 1.79663887, + "memory(GiB)": 58.14, + "step": 2860, + "train_speed(iter/s)": 0.452852 + }, + { + "acc": 0.62211504, + "epoch": 0.07267884322678843, + "grad_norm": 5.625, + "learning_rate": 7.267884322678844e-06, + "loss": 1.92335587, + "memory(GiB)": 58.14, + "step": 2865, + "train_speed(iter/s)": 0.45326 + }, + { + "acc": 0.6139039, + "epoch": 0.07280568239472349, + "grad_norm": 5.9375, + "learning_rate": 7.2805682394723495e-06, + "loss": 1.90454826, + "memory(GiB)": 58.14, + "step": 2870, + "train_speed(iter/s)": 0.453666 + }, + { + "acc": 0.60160236, + "epoch": 0.07293252156265854, + "grad_norm": 5.59375, + "learning_rate": 7.2932521562658555e-06, + "loss": 1.87880173, + "memory(GiB)": 58.14, + "step": 2875, + "train_speed(iter/s)": 0.454079 + }, + { + "acc": 0.60643549, + "epoch": 0.0730593607305936, + "grad_norm": 4.1875, + "learning_rate": 7.305936073059361e-06, + "loss": 1.90190887, + "memory(GiB)": 58.14, + "step": 2880, + "train_speed(iter/s)": 0.454485 + }, + { + "acc": 0.62963877, + "epoch": 0.07318619989852866, + "grad_norm": 4.96875, + "learning_rate": 7.318619989852867e-06, + "loss": 1.89132309, + "memory(GiB)": 58.14, + "step": 2885, + "train_speed(iter/s)": 0.454893 + }, + { + "acc": 0.60651903, + "epoch": 0.07331303906646372, + "grad_norm": 6.5, + "learning_rate": 7.331303906646373e-06, + "loss": 1.88069534, + "memory(GiB)": 58.14, + "step": 2890, + "train_speed(iter/s)": 0.455302 + }, + { + "acc": 0.62535725, + "epoch": 0.07343987823439878, + "grad_norm": 5.4375, + "learning_rate": 7.3439878234398785e-06, + "loss": 1.84180565, + "memory(GiB)": 58.14, + "step": 2895, + "train_speed(iter/s)": 0.455707 + }, + { + "acc": 0.60730028, + "epoch": 0.07356671740233384, + "grad_norm": 5.75, + "learning_rate": 7.3566717402333845e-06, + "loss": 1.98328476, + "memory(GiB)": 58.14, + "step": 2900, + "train_speed(iter/s)": 0.45611 + }, + { + "epoch": 0.07356671740233384, + "eval_acc": 0.6059835976198961, + "eval_loss": 1.8424972295761108, + "eval_runtime": 113.2782, + "eval_samples_per_second": 56.233, + "eval_steps_per_second": 28.117, + "step": 2900 + }, + { + "acc": 0.61364303, + "epoch": 0.0736935565702689, + "grad_norm": 5.96875, + "learning_rate": 7.369355657026891e-06, + "loss": 1.87470016, + "memory(GiB)": 58.14, + "step": 2905, + "train_speed(iter/s)": 0.448226 + }, + { + "acc": 0.60001698, + "epoch": 0.07382039573820395, + "grad_norm": 4.5, + "learning_rate": 7.382039573820396e-06, + "loss": 1.96950932, + "memory(GiB)": 58.14, + "step": 2910, + "train_speed(iter/s)": 0.448632 + }, + { + "acc": 0.61663094, + "epoch": 0.07394723490613901, + "grad_norm": 6.75, + "learning_rate": 7.394723490613902e-06, + "loss": 1.90482388, + "memory(GiB)": 58.14, + "step": 2915, + "train_speed(iter/s)": 0.449038 + }, + { + "acc": 0.60103149, + "epoch": 0.07407407407407407, + "grad_norm": 5.09375, + "learning_rate": 7.4074074074074075e-06, + "loss": 1.92143154, + "memory(GiB)": 58.14, + "step": 2920, + "train_speed(iter/s)": 0.449441 + }, + { + "acc": 0.60360765, + "epoch": 0.07420091324200913, + "grad_norm": 5.59375, + "learning_rate": 7.420091324200914e-06, + "loss": 1.89524937, + "memory(GiB)": 58.14, + "step": 2925, + "train_speed(iter/s)": 0.449843 + }, + { + "acc": 0.60471892, + "epoch": 0.07432775240994419, + "grad_norm": 6.75, + "learning_rate": 7.432775240994419e-06, + "loss": 1.90937576, + "memory(GiB)": 58.14, + "step": 2930, + "train_speed(iter/s)": 0.450246 + }, + { + "acc": 0.62133865, + "epoch": 0.07445459157787924, + "grad_norm": 6.0, + "learning_rate": 7.445459157787925e-06, + "loss": 1.789254, + "memory(GiB)": 58.14, + "step": 2935, + "train_speed(iter/s)": 0.450648 + }, + { + "acc": 0.62306175, + "epoch": 0.0745814307458143, + "grad_norm": 5.5, + "learning_rate": 7.458143074581431e-06, + "loss": 1.8388485, + "memory(GiB)": 58.14, + "step": 2940, + "train_speed(iter/s)": 0.451049 + }, + { + "acc": 0.64608455, + "epoch": 0.07470826991374936, + "grad_norm": 5.90625, + "learning_rate": 7.470826991374937e-06, + "loss": 1.72617111, + "memory(GiB)": 58.14, + "step": 2945, + "train_speed(iter/s)": 0.451444 + }, + { + "acc": 0.59843822, + "epoch": 0.07483510908168442, + "grad_norm": 6.53125, + "learning_rate": 7.483510908168443e-06, + "loss": 1.95106297, + "memory(GiB)": 58.14, + "step": 2950, + "train_speed(iter/s)": 0.451842 + }, + { + "acc": 0.62830148, + "epoch": 0.07496194824961948, + "grad_norm": 4.8125, + "learning_rate": 7.496194824961948e-06, + "loss": 1.75840912, + "memory(GiB)": 58.14, + "step": 2955, + "train_speed(iter/s)": 0.452236 + }, + { + "acc": 0.62830124, + "epoch": 0.07508878741755454, + "grad_norm": 5.03125, + "learning_rate": 7.508878741755454e-06, + "loss": 1.7966526, + "memory(GiB)": 58.14, + "step": 2960, + "train_speed(iter/s)": 0.452634 + }, + { + "acc": 0.61973314, + "epoch": 0.0752156265854896, + "grad_norm": 5.71875, + "learning_rate": 7.521562658548961e-06, + "loss": 1.82047462, + "memory(GiB)": 58.14, + "step": 2965, + "train_speed(iter/s)": 0.453035 + }, + { + "acc": 0.61963325, + "epoch": 0.07534246575342465, + "grad_norm": 5.71875, + "learning_rate": 7.534246575342466e-06, + "loss": 1.82842522, + "memory(GiB)": 58.14, + "step": 2970, + "train_speed(iter/s)": 0.453435 + }, + { + "acc": 0.62633748, + "epoch": 0.07546930492135971, + "grad_norm": 4.625, + "learning_rate": 7.546930492135972e-06, + "loss": 1.80132027, + "memory(GiB)": 58.14, + "step": 2975, + "train_speed(iter/s)": 0.453833 + }, + { + "acc": 0.61981435, + "epoch": 0.07559614408929477, + "grad_norm": 6.53125, + "learning_rate": 7.559614408929477e-06, + "loss": 1.78781509, + "memory(GiB)": 58.14, + "step": 2980, + "train_speed(iter/s)": 0.454228 + }, + { + "acc": 0.60884252, + "epoch": 0.07572298325722983, + "grad_norm": 5.78125, + "learning_rate": 7.572298325722984e-06, + "loss": 1.86665077, + "memory(GiB)": 58.14, + "step": 2985, + "train_speed(iter/s)": 0.454628 + }, + { + "acc": 0.62612743, + "epoch": 0.07584982242516489, + "grad_norm": 5.28125, + "learning_rate": 7.58498224251649e-06, + "loss": 1.82060127, + "memory(GiB)": 58.14, + "step": 2990, + "train_speed(iter/s)": 0.455021 + }, + { + "acc": 0.61098433, + "epoch": 0.07597666159309995, + "grad_norm": 5.0625, + "learning_rate": 7.597666159309995e-06, + "loss": 1.89073334, + "memory(GiB)": 58.14, + "step": 2995, + "train_speed(iter/s)": 0.455416 + }, + { + "acc": 0.6229835, + "epoch": 0.076103500761035, + "grad_norm": 5.5625, + "learning_rate": 7.610350076103501e-06, + "loss": 1.83051796, + "memory(GiB)": 58.14, + "step": 3000, + "train_speed(iter/s)": 0.455811 + }, + { + "epoch": 0.076103500761035, + "eval_acc": 0.6071863497038239, + "eval_loss": 1.834290623664856, + "eval_runtime": 112.1211, + "eval_samples_per_second": 56.814, + "eval_steps_per_second": 28.407, + "step": 3000 + }, + { + "acc": 0.62700224, + "epoch": 0.07623033992897006, + "grad_norm": 4.65625, + "learning_rate": 7.623033992897007e-06, + "loss": 1.77658653, + "memory(GiB)": 58.14, + "step": 3005, + "train_speed(iter/s)": 0.448269 + }, + { + "acc": 0.60678177, + "epoch": 0.07635717909690512, + "grad_norm": 5.71875, + "learning_rate": 7.635717909690512e-06, + "loss": 1.87845783, + "memory(GiB)": 58.14, + "step": 3010, + "train_speed(iter/s)": 0.448662 + }, + { + "acc": 0.61887217, + "epoch": 0.07648401826484018, + "grad_norm": 5.9375, + "learning_rate": 7.648401826484018e-06, + "loss": 1.86014633, + "memory(GiB)": 58.14, + "step": 3015, + "train_speed(iter/s)": 0.449054 + }, + { + "acc": 0.63407774, + "epoch": 0.07661085743277524, + "grad_norm": 5.375, + "learning_rate": 7.661085743277524e-06, + "loss": 1.78521271, + "memory(GiB)": 58.14, + "step": 3020, + "train_speed(iter/s)": 0.449444 + }, + { + "acc": 0.6252676, + "epoch": 0.0767376966007103, + "grad_norm": 4.71875, + "learning_rate": 7.67376966007103e-06, + "loss": 1.81846733, + "memory(GiB)": 58.14, + "step": 3025, + "train_speed(iter/s)": 0.449833 + }, + { + "acc": 0.62281542, + "epoch": 0.07686453576864535, + "grad_norm": 5.375, + "learning_rate": 7.686453576864536e-06, + "loss": 1.79361, + "memory(GiB)": 58.14, + "step": 3030, + "train_speed(iter/s)": 0.450219 + }, + { + "acc": 0.60511341, + "epoch": 0.07699137493658041, + "grad_norm": 5.65625, + "learning_rate": 7.699137493658042e-06, + "loss": 1.8661087, + "memory(GiB)": 58.14, + "step": 3035, + "train_speed(iter/s)": 0.450606 + }, + { + "acc": 0.62472086, + "epoch": 0.07711821410451547, + "grad_norm": 4.9375, + "learning_rate": 7.711821410451548e-06, + "loss": 1.74990273, + "memory(GiB)": 58.14, + "step": 3040, + "train_speed(iter/s)": 0.45099 + }, + { + "acc": 0.61164966, + "epoch": 0.07724505327245053, + "grad_norm": 5.75, + "learning_rate": 7.724505327245054e-06, + "loss": 1.82849693, + "memory(GiB)": 58.14, + "step": 3045, + "train_speed(iter/s)": 0.451382 + }, + { + "acc": 0.61925054, + "epoch": 0.07737189244038559, + "grad_norm": 6.125, + "learning_rate": 7.73718924403856e-06, + "loss": 1.82250595, + "memory(GiB)": 58.14, + "step": 3050, + "train_speed(iter/s)": 0.451769 + }, + { + "acc": 0.62500401, + "epoch": 0.07749873160832065, + "grad_norm": 4.90625, + "learning_rate": 7.749873160832066e-06, + "loss": 1.84241638, + "memory(GiB)": 58.14, + "step": 3055, + "train_speed(iter/s)": 0.452156 + }, + { + "acc": 0.61435013, + "epoch": 0.0776255707762557, + "grad_norm": 4.6875, + "learning_rate": 7.76255707762557e-06, + "loss": 1.81692009, + "memory(GiB)": 58.14, + "step": 3060, + "train_speed(iter/s)": 0.452543 + }, + { + "acc": 0.63202238, + "epoch": 0.07775240994419076, + "grad_norm": 5.75, + "learning_rate": 7.775240994419078e-06, + "loss": 1.8022438, + "memory(GiB)": 58.14, + "step": 3065, + "train_speed(iter/s)": 0.452925 + }, + { + "acc": 0.61397858, + "epoch": 0.07787924911212582, + "grad_norm": 5.75, + "learning_rate": 7.787924911212584e-06, + "loss": 1.85419598, + "memory(GiB)": 58.14, + "step": 3070, + "train_speed(iter/s)": 0.453306 + }, + { + "acc": 0.61774077, + "epoch": 0.07800608828006088, + "grad_norm": 5.625, + "learning_rate": 7.800608828006088e-06, + "loss": 1.84221916, + "memory(GiB)": 58.14, + "step": 3075, + "train_speed(iter/s)": 0.453691 + }, + { + "acc": 0.6142952, + "epoch": 0.07813292744799594, + "grad_norm": 4.875, + "learning_rate": 7.813292744799594e-06, + "loss": 1.82860069, + "memory(GiB)": 58.14, + "step": 3080, + "train_speed(iter/s)": 0.454075 + }, + { + "acc": 0.60901523, + "epoch": 0.078259766615931, + "grad_norm": 5.46875, + "learning_rate": 7.8259766615931e-06, + "loss": 1.81847115, + "memory(GiB)": 58.14, + "step": 3085, + "train_speed(iter/s)": 0.454457 + }, + { + "acc": 0.60886526, + "epoch": 0.07838660578386605, + "grad_norm": 5.84375, + "learning_rate": 7.838660578386606e-06, + "loss": 1.89605694, + "memory(GiB)": 58.14, + "step": 3090, + "train_speed(iter/s)": 0.454838 + }, + { + "acc": 0.61970091, + "epoch": 0.07851344495180111, + "grad_norm": 4.625, + "learning_rate": 7.851344495180112e-06, + "loss": 1.84955826, + "memory(GiB)": 58.14, + "step": 3095, + "train_speed(iter/s)": 0.45522 + }, + { + "acc": 0.6121417, + "epoch": 0.07864028411973617, + "grad_norm": 4.9375, + "learning_rate": 7.864028411973618e-06, + "loss": 1.86377239, + "memory(GiB)": 58.14, + "step": 3100, + "train_speed(iter/s)": 0.455601 + }, + { + "epoch": 0.07864028411973617, + "eval_acc": 0.608449678047623, + "eval_loss": 1.8263449668884277, + "eval_runtime": 112.6212, + "eval_samples_per_second": 56.561, + "eval_steps_per_second": 28.281, + "step": 3100 + }, + { + "acc": 0.61682882, + "epoch": 0.07876712328767123, + "grad_norm": 5.4375, + "learning_rate": 7.876712328767124e-06, + "loss": 1.84462891, + "memory(GiB)": 58.14, + "step": 3105, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.61170282, + "epoch": 0.07889396245560629, + "grad_norm": 5.375, + "learning_rate": 7.88939624556063e-06, + "loss": 1.87310658, + "memory(GiB)": 58.14, + "step": 3110, + "train_speed(iter/s)": 0.448661 + }, + { + "acc": 0.6301981, + "epoch": 0.07902080162354135, + "grad_norm": 5.4375, + "learning_rate": 7.902080162354136e-06, + "loss": 1.75871811, + "memory(GiB)": 58.14, + "step": 3115, + "train_speed(iter/s)": 0.449049 + }, + { + "acc": 0.64542499, + "epoch": 0.0791476407914764, + "grad_norm": 5.9375, + "learning_rate": 7.91476407914764e-06, + "loss": 1.74482307, + "memory(GiB)": 58.14, + "step": 3120, + "train_speed(iter/s)": 0.449435 + }, + { + "acc": 0.62039104, + "epoch": 0.07927447995941146, + "grad_norm": 5.46875, + "learning_rate": 7.927447995941148e-06, + "loss": 1.81275291, + "memory(GiB)": 58.14, + "step": 3125, + "train_speed(iter/s)": 0.44982 + }, + { + "acc": 0.61505303, + "epoch": 0.07940131912734652, + "grad_norm": 5.4375, + "learning_rate": 7.940131912734654e-06, + "loss": 1.82998772, + "memory(GiB)": 58.14, + "step": 3130, + "train_speed(iter/s)": 0.450205 + }, + { + "acc": 0.63439422, + "epoch": 0.07952815829528158, + "grad_norm": 5.0, + "learning_rate": 7.952815829528158e-06, + "loss": 1.77613087, + "memory(GiB)": 58.14, + "step": 3135, + "train_speed(iter/s)": 0.450591 + }, + { + "acc": 0.62651486, + "epoch": 0.07965499746321664, + "grad_norm": 5.71875, + "learning_rate": 7.965499746321664e-06, + "loss": 1.82649231, + "memory(GiB)": 58.14, + "step": 3140, + "train_speed(iter/s)": 0.450972 + }, + { + "acc": 0.61029425, + "epoch": 0.0797818366311517, + "grad_norm": 5.1875, + "learning_rate": 7.978183663115172e-06, + "loss": 1.94225922, + "memory(GiB)": 58.14, + "step": 3145, + "train_speed(iter/s)": 0.451357 + }, + { + "acc": 0.6186645, + "epoch": 0.07990867579908675, + "grad_norm": 4.125, + "learning_rate": 7.990867579908676e-06, + "loss": 1.84380264, + "memory(GiB)": 58.14, + "step": 3150, + "train_speed(iter/s)": 0.451738 + }, + { + "acc": 0.60702405, + "epoch": 0.08003551496702181, + "grad_norm": 4.9375, + "learning_rate": 8.003551496702182e-06, + "loss": 1.88564758, + "memory(GiB)": 58.14, + "step": 3155, + "train_speed(iter/s)": 0.452119 + }, + { + "acc": 0.61457863, + "epoch": 0.08016235413495687, + "grad_norm": 5.15625, + "learning_rate": 8.016235413495688e-06, + "loss": 1.81395073, + "memory(GiB)": 58.14, + "step": 3160, + "train_speed(iter/s)": 0.452501 + }, + { + "acc": 0.60833502, + "epoch": 0.08028919330289193, + "grad_norm": 6.84375, + "learning_rate": 8.028919330289194e-06, + "loss": 1.92483559, + "memory(GiB)": 58.14, + "step": 3165, + "train_speed(iter/s)": 0.45288 + }, + { + "acc": 0.62526989, + "epoch": 0.08041603247082699, + "grad_norm": 4.6875, + "learning_rate": 8.0416032470827e-06, + "loss": 1.80583191, + "memory(GiB)": 58.14, + "step": 3170, + "train_speed(iter/s)": 0.453261 + }, + { + "acc": 0.61353626, + "epoch": 0.08054287163876205, + "grad_norm": 5.25, + "learning_rate": 8.054287163876206e-06, + "loss": 1.82396183, + "memory(GiB)": 58.14, + "step": 3175, + "train_speed(iter/s)": 0.453641 + }, + { + "acc": 0.61734285, + "epoch": 0.0806697108066971, + "grad_norm": 5.25, + "learning_rate": 8.066971080669712e-06, + "loss": 1.83472328, + "memory(GiB)": 58.14, + "step": 3180, + "train_speed(iter/s)": 0.45402 + }, + { + "acc": 0.60642791, + "epoch": 0.08079654997463216, + "grad_norm": 5.90625, + "learning_rate": 8.079654997463218e-06, + "loss": 1.91571655, + "memory(GiB)": 58.14, + "step": 3185, + "train_speed(iter/s)": 0.454401 + }, + { + "acc": 0.61522112, + "epoch": 0.08092338914256722, + "grad_norm": 5.75, + "learning_rate": 8.092338914256724e-06, + "loss": 1.8248661, + "memory(GiB)": 58.14, + "step": 3190, + "train_speed(iter/s)": 0.454778 + }, + { + "acc": 0.60942669, + "epoch": 0.08105022831050228, + "grad_norm": 6.15625, + "learning_rate": 8.105022831050228e-06, + "loss": 1.79037781, + "memory(GiB)": 58.14, + "step": 3195, + "train_speed(iter/s)": 0.455156 + }, + { + "acc": 0.62582579, + "epoch": 0.08117706747843734, + "grad_norm": 5.125, + "learning_rate": 8.117706747843734e-06, + "loss": 1.77895699, + "memory(GiB)": 58.14, + "step": 3200, + "train_speed(iter/s)": 0.45553 + }, + { + "epoch": 0.08117706747843734, + "eval_acc": 0.6095091359444069, + "eval_loss": 1.8175314664840698, + "eval_runtime": 112.5563, + "eval_samples_per_second": 56.594, + "eval_steps_per_second": 28.297, + "step": 3200 + }, + { + "acc": 0.61690254, + "epoch": 0.0813039066463724, + "grad_norm": 5.3125, + "learning_rate": 8.130390664637242e-06, + "loss": 1.8259449, + "memory(GiB)": 58.14, + "step": 3205, + "train_speed(iter/s)": 0.448439 + }, + { + "acc": 0.62860312, + "epoch": 0.08143074581430745, + "grad_norm": 5.28125, + "learning_rate": 8.143074581430746e-06, + "loss": 1.78546467, + "memory(GiB)": 58.14, + "step": 3210, + "train_speed(iter/s)": 0.448811 + }, + { + "acc": 0.6273778, + "epoch": 0.08155758498224251, + "grad_norm": 5.3125, + "learning_rate": 8.155758498224252e-06, + "loss": 1.82557812, + "memory(GiB)": 58.14, + "step": 3215, + "train_speed(iter/s)": 0.449186 + }, + { + "acc": 0.62210217, + "epoch": 0.08168442415017757, + "grad_norm": 6.15625, + "learning_rate": 8.168442415017758e-06, + "loss": 1.81140614, + "memory(GiB)": 58.14, + "step": 3220, + "train_speed(iter/s)": 0.449559 + }, + { + "acc": 0.60306826, + "epoch": 0.08181126331811263, + "grad_norm": 6.3125, + "learning_rate": 8.181126331811264e-06, + "loss": 1.9109745, + "memory(GiB)": 58.14, + "step": 3225, + "train_speed(iter/s)": 0.449934 + }, + { + "acc": 0.61548305, + "epoch": 0.08193810248604769, + "grad_norm": 4.75, + "learning_rate": 8.19381024860477e-06, + "loss": 1.80882435, + "memory(GiB)": 58.14, + "step": 3230, + "train_speed(iter/s)": 0.45031 + }, + { + "acc": 0.62371755, + "epoch": 0.08206494165398275, + "grad_norm": 5.15625, + "learning_rate": 8.206494165398276e-06, + "loss": 1.77641735, + "memory(GiB)": 58.14, + "step": 3235, + "train_speed(iter/s)": 0.450685 + }, + { + "acc": 0.61926298, + "epoch": 0.0821917808219178, + "grad_norm": 6.09375, + "learning_rate": 8.219178082191782e-06, + "loss": 1.84104881, + "memory(GiB)": 58.14, + "step": 3240, + "train_speed(iter/s)": 0.451057 + }, + { + "acc": 0.61810226, + "epoch": 0.08231861998985286, + "grad_norm": 5.125, + "learning_rate": 8.231861998985288e-06, + "loss": 1.79076538, + "memory(GiB)": 58.14, + "step": 3245, + "train_speed(iter/s)": 0.451426 + }, + { + "acc": 0.63153791, + "epoch": 0.08244545915778792, + "grad_norm": 5.09375, + "learning_rate": 8.244545915778794e-06, + "loss": 1.74150486, + "memory(GiB)": 58.14, + "step": 3250, + "train_speed(iter/s)": 0.451798 + }, + { + "acc": 0.61119537, + "epoch": 0.08257229832572298, + "grad_norm": 4.625, + "learning_rate": 8.2572298325723e-06, + "loss": 1.85210724, + "memory(GiB)": 58.14, + "step": 3255, + "train_speed(iter/s)": 0.452167 + }, + { + "acc": 0.62431269, + "epoch": 0.08269913749365804, + "grad_norm": 5.1875, + "learning_rate": 8.269913749365804e-06, + "loss": 1.76628704, + "memory(GiB)": 58.14, + "step": 3260, + "train_speed(iter/s)": 0.452535 + }, + { + "acc": 0.61237116, + "epoch": 0.0828259766615931, + "grad_norm": 5.625, + "learning_rate": 8.282597666159312e-06, + "loss": 1.79230824, + "memory(GiB)": 58.14, + "step": 3265, + "train_speed(iter/s)": 0.452907 + }, + { + "acc": 0.61545811, + "epoch": 0.08295281582952815, + "grad_norm": 5.71875, + "learning_rate": 8.295281582952816e-06, + "loss": 1.85704842, + "memory(GiB)": 58.14, + "step": 3270, + "train_speed(iter/s)": 0.453274 + }, + { + "acc": 0.61269817, + "epoch": 0.08307965499746321, + "grad_norm": 5.96875, + "learning_rate": 8.307965499746322e-06, + "loss": 1.81740627, + "memory(GiB)": 58.14, + "step": 3275, + "train_speed(iter/s)": 0.453639 + }, + { + "acc": 0.60666189, + "epoch": 0.08320649416539827, + "grad_norm": 4.78125, + "learning_rate": 8.320649416539828e-06, + "loss": 1.91259232, + "memory(GiB)": 58.14, + "step": 3280, + "train_speed(iter/s)": 0.454009 + }, + { + "acc": 0.62171803, + "epoch": 0.08333333333333333, + "grad_norm": 5.3125, + "learning_rate": 8.333333333333334e-06, + "loss": 1.76733837, + "memory(GiB)": 58.14, + "step": 3285, + "train_speed(iter/s)": 0.454377 + }, + { + "acc": 0.62376614, + "epoch": 0.08346017250126839, + "grad_norm": 6.03125, + "learning_rate": 8.34601725012684e-06, + "loss": 1.73643665, + "memory(GiB)": 58.14, + "step": 3290, + "train_speed(iter/s)": 0.454746 + }, + { + "acc": 0.61811113, + "epoch": 0.08358701166920345, + "grad_norm": 7.78125, + "learning_rate": 8.358701166920346e-06, + "loss": 1.86945801, + "memory(GiB)": 58.14, + "step": 3295, + "train_speed(iter/s)": 0.455114 + }, + { + "acc": 0.60208273, + "epoch": 0.0837138508371385, + "grad_norm": 4.96875, + "learning_rate": 8.371385083713852e-06, + "loss": 1.93450298, + "memory(GiB)": 58.14, + "step": 3300, + "train_speed(iter/s)": 0.45548 + }, + { + "epoch": 0.0837138508371385, + "eval_acc": 0.6107256743495467, + "eval_loss": 1.810210108757019, + "eval_runtime": 113.0769, + "eval_samples_per_second": 56.333, + "eval_steps_per_second": 28.167, + "step": 3300 + }, + { + "acc": 0.60897026, + "epoch": 0.08384069000507356, + "grad_norm": 5.84375, + "learning_rate": 8.384069000507358e-06, + "loss": 1.88332233, + "memory(GiB)": 58.14, + "step": 3305, + "train_speed(iter/s)": 0.448568 + }, + { + "acc": 0.6155386, + "epoch": 0.08396752917300862, + "grad_norm": 4.625, + "learning_rate": 8.396752917300864e-06, + "loss": 1.83776817, + "memory(GiB)": 58.14, + "step": 3310, + "train_speed(iter/s)": 0.448927 + }, + { + "acc": 0.60938406, + "epoch": 0.08409436834094368, + "grad_norm": 5.15625, + "learning_rate": 8.40943683409437e-06, + "loss": 1.83270493, + "memory(GiB)": 58.14, + "step": 3315, + "train_speed(iter/s)": 0.449285 + }, + { + "acc": 0.61576252, + "epoch": 0.08422120750887874, + "grad_norm": 5.375, + "learning_rate": 8.422120750887874e-06, + "loss": 1.86774483, + "memory(GiB)": 58.14, + "step": 3320, + "train_speed(iter/s)": 0.449638 + }, + { + "acc": 0.61662235, + "epoch": 0.0843480466768138, + "grad_norm": 4.96875, + "learning_rate": 8.434804667681381e-06, + "loss": 1.75339928, + "memory(GiB)": 58.14, + "step": 3325, + "train_speed(iter/s)": 0.449996 + }, + { + "acc": 0.62029095, + "epoch": 0.08447488584474885, + "grad_norm": 5.375, + "learning_rate": 8.447488584474887e-06, + "loss": 1.83997078, + "memory(GiB)": 58.14, + "step": 3330, + "train_speed(iter/s)": 0.450348 + }, + { + "acc": 0.61939445, + "epoch": 0.08460172501268391, + "grad_norm": 6.03125, + "learning_rate": 8.460172501268392e-06, + "loss": 1.85392971, + "memory(GiB)": 58.14, + "step": 3335, + "train_speed(iter/s)": 0.450698 + }, + { + "acc": 0.61963158, + "epoch": 0.08472856418061897, + "grad_norm": 4.75, + "learning_rate": 8.472856418061898e-06, + "loss": 1.80520058, + "memory(GiB)": 58.14, + "step": 3340, + "train_speed(iter/s)": 0.451046 + }, + { + "acc": 0.6401906, + "epoch": 0.08485540334855403, + "grad_norm": 5.25, + "learning_rate": 8.485540334855404e-06, + "loss": 1.76707306, + "memory(GiB)": 58.14, + "step": 3345, + "train_speed(iter/s)": 0.451396 + }, + { + "acc": 0.64308271, + "epoch": 0.08498224251648909, + "grad_norm": 4.21875, + "learning_rate": 8.49822425164891e-06, + "loss": 1.74596481, + "memory(GiB)": 58.14, + "step": 3350, + "train_speed(iter/s)": 0.451744 + }, + { + "acc": 0.61765594, + "epoch": 0.08510908168442415, + "grad_norm": 4.875, + "learning_rate": 8.510908168442416e-06, + "loss": 1.84585762, + "memory(GiB)": 58.14, + "step": 3355, + "train_speed(iter/s)": 0.452094 + }, + { + "acc": 0.60037155, + "epoch": 0.0852359208523592, + "grad_norm": 5.71875, + "learning_rate": 8.523592085235922e-06, + "loss": 1.88099785, + "memory(GiB)": 58.14, + "step": 3360, + "train_speed(iter/s)": 0.452447 + }, + { + "acc": 0.61336927, + "epoch": 0.08536276002029426, + "grad_norm": 5.3125, + "learning_rate": 8.536276002029428e-06, + "loss": 1.80283871, + "memory(GiB)": 58.14, + "step": 3365, + "train_speed(iter/s)": 0.452798 + }, + { + "acc": 0.61765428, + "epoch": 0.08548959918822932, + "grad_norm": 6.25, + "learning_rate": 8.548959918822933e-06, + "loss": 1.89463348, + "memory(GiB)": 58.14, + "step": 3370, + "train_speed(iter/s)": 0.453148 + }, + { + "acc": 0.61489816, + "epoch": 0.08561643835616438, + "grad_norm": 5.375, + "learning_rate": 8.56164383561644e-06, + "loss": 1.84810829, + "memory(GiB)": 58.14, + "step": 3375, + "train_speed(iter/s)": 0.4535 + }, + { + "acc": 0.63380384, + "epoch": 0.08574327752409944, + "grad_norm": 6.25, + "learning_rate": 8.574327752409944e-06, + "loss": 1.76909733, + "memory(GiB)": 58.14, + "step": 3380, + "train_speed(iter/s)": 0.453845 + }, + { + "acc": 0.61153407, + "epoch": 0.0858701166920345, + "grad_norm": 5.53125, + "learning_rate": 8.587011669203451e-06, + "loss": 1.83102589, + "memory(GiB)": 58.14, + "step": 3385, + "train_speed(iter/s)": 0.454192 + }, + { + "acc": 0.63582916, + "epoch": 0.08599695585996955, + "grad_norm": 6.15625, + "learning_rate": 8.599695585996957e-06, + "loss": 1.78380661, + "memory(GiB)": 58.14, + "step": 3390, + "train_speed(iter/s)": 0.454542 + }, + { + "acc": 0.61066709, + "epoch": 0.08612379502790461, + "grad_norm": 5.84375, + "learning_rate": 8.612379502790462e-06, + "loss": 1.89545879, + "memory(GiB)": 58.14, + "step": 3395, + "train_speed(iter/s)": 0.454887 + }, + { + "acc": 0.61927481, + "epoch": 0.08625063419583967, + "grad_norm": 5.5, + "learning_rate": 8.625063419583968e-06, + "loss": 1.82426186, + "memory(GiB)": 58.14, + "step": 3400, + "train_speed(iter/s)": 0.455232 + }, + { + "epoch": 0.08625063419583967, + "eval_acc": 0.6118954228160274, + "eval_loss": 1.8020211458206177, + "eval_runtime": 112.001, + "eval_samples_per_second": 56.875, + "eval_steps_per_second": 28.437, + "step": 3400 + }, + { + "acc": 0.62710462, + "epoch": 0.08637747336377473, + "grad_norm": 5.59375, + "learning_rate": 8.637747336377475e-06, + "loss": 1.81724014, + "memory(GiB)": 58.14, + "step": 3405, + "train_speed(iter/s)": 0.448591 + }, + { + "acc": 0.6170856, + "epoch": 0.08650431253170979, + "grad_norm": 6.03125, + "learning_rate": 8.65043125317098e-06, + "loss": 1.88748283, + "memory(GiB)": 58.14, + "step": 3410, + "train_speed(iter/s)": 0.448942 + }, + { + "acc": 0.60297766, + "epoch": 0.08663115169964485, + "grad_norm": 5.78125, + "learning_rate": 8.663115169964485e-06, + "loss": 1.81512585, + "memory(GiB)": 58.14, + "step": 3415, + "train_speed(iter/s)": 0.449292 + }, + { + "acc": 0.5959734, + "epoch": 0.0867579908675799, + "grad_norm": 6.34375, + "learning_rate": 8.675799086757991e-06, + "loss": 1.89495277, + "memory(GiB)": 58.14, + "step": 3420, + "train_speed(iter/s)": 0.449639 + }, + { + "acc": 0.62047443, + "epoch": 0.08688483003551496, + "grad_norm": 6.1875, + "learning_rate": 8.688483003551497e-06, + "loss": 1.8824234, + "memory(GiB)": 58.14, + "step": 3425, + "train_speed(iter/s)": 0.449985 + }, + { + "acc": 0.63176394, + "epoch": 0.08701166920345002, + "grad_norm": 5.15625, + "learning_rate": 8.701166920345003e-06, + "loss": 1.79829102, + "memory(GiB)": 58.14, + "step": 3430, + "train_speed(iter/s)": 0.450328 + }, + { + "acc": 0.61930971, + "epoch": 0.08713850837138508, + "grad_norm": 5.40625, + "learning_rate": 8.71385083713851e-06, + "loss": 1.84827385, + "memory(GiB)": 58.14, + "step": 3435, + "train_speed(iter/s)": 0.450677 + }, + { + "acc": 0.60883965, + "epoch": 0.08726534753932014, + "grad_norm": 5.8125, + "learning_rate": 8.726534753932014e-06, + "loss": 1.94455986, + "memory(GiB)": 58.14, + "step": 3440, + "train_speed(iter/s)": 0.451024 + }, + { + "acc": 0.61605301, + "epoch": 0.0873921867072552, + "grad_norm": 6.125, + "learning_rate": 8.739218670725521e-06, + "loss": 1.79549141, + "memory(GiB)": 58.14, + "step": 3445, + "train_speed(iter/s)": 0.451369 + }, + { + "acc": 0.61880531, + "epoch": 0.08751902587519025, + "grad_norm": 5.15625, + "learning_rate": 8.751902587519027e-06, + "loss": 1.80067139, + "memory(GiB)": 58.14, + "step": 3450, + "train_speed(iter/s)": 0.451715 + }, + { + "acc": 0.61545162, + "epoch": 0.08764586504312531, + "grad_norm": 5.875, + "learning_rate": 8.764586504312532e-06, + "loss": 1.89304962, + "memory(GiB)": 58.14, + "step": 3455, + "train_speed(iter/s)": 0.452057 + }, + { + "acc": 0.6373147, + "epoch": 0.08777270421106037, + "grad_norm": 5.0, + "learning_rate": 8.777270421106037e-06, + "loss": 1.84837685, + "memory(GiB)": 58.14, + "step": 3460, + "train_speed(iter/s)": 0.452396 + }, + { + "acc": 0.63052659, + "epoch": 0.08789954337899543, + "grad_norm": 5.15625, + "learning_rate": 8.789954337899545e-06, + "loss": 1.8142704, + "memory(GiB)": 58.14, + "step": 3465, + "train_speed(iter/s)": 0.452739 + }, + { + "acc": 0.61962991, + "epoch": 0.08802638254693049, + "grad_norm": 5.75, + "learning_rate": 8.80263825469305e-06, + "loss": 1.83392391, + "memory(GiB)": 58.14, + "step": 3470, + "train_speed(iter/s)": 0.453084 + }, + { + "acc": 0.63001647, + "epoch": 0.08815322171486555, + "grad_norm": 4.84375, + "learning_rate": 8.815322171486555e-06, + "loss": 1.74116135, + "memory(GiB)": 58.14, + "step": 3475, + "train_speed(iter/s)": 0.453425 + }, + { + "acc": 0.62627082, + "epoch": 0.0882800608828006, + "grad_norm": 5.9375, + "learning_rate": 8.828006088280061e-06, + "loss": 1.83807144, + "memory(GiB)": 58.14, + "step": 3480, + "train_speed(iter/s)": 0.453769 + }, + { + "acc": 0.624786, + "epoch": 0.08840690005073566, + "grad_norm": 5.90625, + "learning_rate": 8.840690005073567e-06, + "loss": 1.79691505, + "memory(GiB)": 58.14, + "step": 3485, + "train_speed(iter/s)": 0.45411 + }, + { + "acc": 0.61368771, + "epoch": 0.08853373921867072, + "grad_norm": 5.53125, + "learning_rate": 8.853373921867073e-06, + "loss": 1.81415997, + "memory(GiB)": 58.14, + "step": 3490, + "train_speed(iter/s)": 0.454451 + }, + { + "acc": 0.62553358, + "epoch": 0.08866057838660578, + "grad_norm": 5.0, + "learning_rate": 8.86605783866058e-06, + "loss": 1.83932266, + "memory(GiB)": 58.14, + "step": 3495, + "train_speed(iter/s)": 0.454793 + }, + { + "acc": 0.62242126, + "epoch": 0.08878741755454084, + "grad_norm": 4.4375, + "learning_rate": 8.878741755454085e-06, + "loss": 1.75874329, + "memory(GiB)": 58.14, + "step": 3500, + "train_speed(iter/s)": 0.455133 + }, + { + "epoch": 0.08878741755454084, + "eval_acc": 0.6133417332413689, + "eval_loss": 1.7945423126220703, + "eval_runtime": 112.4994, + "eval_samples_per_second": 56.623, + "eval_steps_per_second": 28.311, + "step": 3500 + }, + { + "acc": 0.61936059, + "epoch": 0.0889142567224759, + "grad_norm": 4.625, + "learning_rate": 8.891425672247591e-06, + "loss": 1.76236572, + "memory(GiB)": 58.14, + "step": 3505, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.62370062, + "epoch": 0.08904109589041095, + "grad_norm": 4.78125, + "learning_rate": 8.904109589041097e-06, + "loss": 1.82231445, + "memory(GiB)": 58.14, + "step": 3510, + "train_speed(iter/s)": 0.448988 + }, + { + "acc": 0.61198301, + "epoch": 0.08916793505834601, + "grad_norm": 6.9375, + "learning_rate": 8.916793505834601e-06, + "loss": 1.85827618, + "memory(GiB)": 58.14, + "step": 3515, + "train_speed(iter/s)": 0.449327 + }, + { + "acc": 0.62784176, + "epoch": 0.08929477422628107, + "grad_norm": 4.96875, + "learning_rate": 8.929477422628107e-06, + "loss": 1.83762989, + "memory(GiB)": 58.14, + "step": 3520, + "train_speed(iter/s)": 0.449666 + }, + { + "acc": 0.62667394, + "epoch": 0.08942161339421613, + "grad_norm": 6.28125, + "learning_rate": 8.942161339421615e-06, + "loss": 1.81751804, + "memory(GiB)": 58.14, + "step": 3525, + "train_speed(iter/s)": 0.450001 + }, + { + "acc": 0.62553697, + "epoch": 0.08954845256215119, + "grad_norm": 6.90625, + "learning_rate": 8.95484525621512e-06, + "loss": 1.82806931, + "memory(GiB)": 58.14, + "step": 3530, + "train_speed(iter/s)": 0.450336 + }, + { + "acc": 0.61312857, + "epoch": 0.08967529173008625, + "grad_norm": 5.21875, + "learning_rate": 8.967529173008625e-06, + "loss": 1.8523407, + "memory(GiB)": 58.14, + "step": 3535, + "train_speed(iter/s)": 0.450671 + }, + { + "acc": 0.62588091, + "epoch": 0.0898021308980213, + "grad_norm": 4.59375, + "learning_rate": 8.980213089802131e-06, + "loss": 1.75777435, + "memory(GiB)": 58.14, + "step": 3540, + "train_speed(iter/s)": 0.451008 + }, + { + "acc": 0.62678781, + "epoch": 0.08992897006595636, + "grad_norm": 5.28125, + "learning_rate": 8.992897006595637e-06, + "loss": 1.73664169, + "memory(GiB)": 58.14, + "step": 3545, + "train_speed(iter/s)": 0.451343 + }, + { + "acc": 0.62679572, + "epoch": 0.09005580923389142, + "grad_norm": 5.0, + "learning_rate": 9.005580923389143e-06, + "loss": 1.77318115, + "memory(GiB)": 58.14, + "step": 3550, + "train_speed(iter/s)": 0.451676 + }, + { + "acc": 0.60809536, + "epoch": 0.09018264840182648, + "grad_norm": 4.625, + "learning_rate": 9.01826484018265e-06, + "loss": 1.81697884, + "memory(GiB)": 58.14, + "step": 3555, + "train_speed(iter/s)": 0.452009 + }, + { + "acc": 0.61045055, + "epoch": 0.09030948756976154, + "grad_norm": 6.40625, + "learning_rate": 9.030948756976155e-06, + "loss": 1.90286446, + "memory(GiB)": 58.14, + "step": 3560, + "train_speed(iter/s)": 0.452341 + }, + { + "acc": 0.62537527, + "epoch": 0.0904363267376966, + "grad_norm": 4.8125, + "learning_rate": 9.043632673769661e-06, + "loss": 1.82007408, + "memory(GiB)": 58.14, + "step": 3565, + "train_speed(iter/s)": 0.452674 + }, + { + "acc": 0.62571921, + "epoch": 0.09056316590563165, + "grad_norm": 5.40625, + "learning_rate": 9.056316590563167e-06, + "loss": 1.77356339, + "memory(GiB)": 58.14, + "step": 3570, + "train_speed(iter/s)": 0.453005 + }, + { + "acc": 0.62236099, + "epoch": 0.09069000507356671, + "grad_norm": 4.78125, + "learning_rate": 9.069000507356673e-06, + "loss": 1.77000313, + "memory(GiB)": 58.14, + "step": 3575, + "train_speed(iter/s)": 0.453336 + }, + { + "acc": 0.62174726, + "epoch": 0.09081684424150177, + "grad_norm": 5.3125, + "learning_rate": 9.081684424150177e-06, + "loss": 1.79823151, + "memory(GiB)": 58.14, + "step": 3580, + "train_speed(iter/s)": 0.45367 + }, + { + "acc": 0.61999578, + "epoch": 0.09094368340943683, + "grad_norm": 5.0, + "learning_rate": 9.094368340943685e-06, + "loss": 1.76183739, + "memory(GiB)": 58.14, + "step": 3585, + "train_speed(iter/s)": 0.454002 + }, + { + "acc": 0.62854447, + "epoch": 0.09107052257737189, + "grad_norm": 6.0, + "learning_rate": 9.10705225773719e-06, + "loss": 1.72111435, + "memory(GiB)": 58.14, + "step": 3590, + "train_speed(iter/s)": 0.454334 + }, + { + "acc": 0.62514219, + "epoch": 0.09119736174530695, + "grad_norm": 5.4375, + "learning_rate": 9.119736174530695e-06, + "loss": 1.85423927, + "memory(GiB)": 58.14, + "step": 3595, + "train_speed(iter/s)": 0.454667 + }, + { + "acc": 0.630972, + "epoch": 0.091324200913242, + "grad_norm": 5.59375, + "learning_rate": 9.132420091324201e-06, + "loss": 1.78647766, + "memory(GiB)": 58.14, + "step": 3600, + "train_speed(iter/s)": 0.454995 + }, + { + "epoch": 0.091324200913242, + "eval_acc": 0.6143890758861784, + "eval_loss": 1.7867289781570435, + "eval_runtime": 112.9582, + "eval_samples_per_second": 56.393, + "eval_steps_per_second": 28.196, + "step": 3600 + }, + { + "acc": 0.63963957, + "epoch": 0.09145104008117706, + "grad_norm": 5.1875, + "learning_rate": 9.145104008117707e-06, + "loss": 1.71173744, + "memory(GiB)": 58.14, + "step": 3605, + "train_speed(iter/s)": 0.448666 + }, + { + "acc": 0.6186388, + "epoch": 0.09157787924911212, + "grad_norm": 4.34375, + "learning_rate": 9.157787924911213e-06, + "loss": 1.83856735, + "memory(GiB)": 58.14, + "step": 3610, + "train_speed(iter/s)": 0.448993 + }, + { + "acc": 0.6158443, + "epoch": 0.09170471841704718, + "grad_norm": 4.65625, + "learning_rate": 9.170471841704719e-06, + "loss": 1.84079876, + "memory(GiB)": 58.14, + "step": 3615, + "train_speed(iter/s)": 0.449316 + }, + { + "acc": 0.60763006, + "epoch": 0.09183155758498224, + "grad_norm": 6.40625, + "learning_rate": 9.183155758498225e-06, + "loss": 1.85199814, + "memory(GiB)": 58.14, + "step": 3620, + "train_speed(iter/s)": 0.449642 + }, + { + "acc": 0.61157689, + "epoch": 0.0919583967529173, + "grad_norm": 4.8125, + "learning_rate": 9.195839675291731e-06, + "loss": 1.84833813, + "memory(GiB)": 58.14, + "step": 3625, + "train_speed(iter/s)": 0.449965 + }, + { + "acc": 0.62855687, + "epoch": 0.09208523592085235, + "grad_norm": 5.46875, + "learning_rate": 9.208523592085237e-06, + "loss": 1.79317932, + "memory(GiB)": 58.14, + "step": 3630, + "train_speed(iter/s)": 0.45029 + }, + { + "acc": 0.63908348, + "epoch": 0.09221207508878741, + "grad_norm": 5.9375, + "learning_rate": 9.221207508878743e-06, + "loss": 1.76104279, + "memory(GiB)": 58.14, + "step": 3635, + "train_speed(iter/s)": 0.450614 + }, + { + "acc": 0.6186429, + "epoch": 0.09233891425672247, + "grad_norm": 5.71875, + "learning_rate": 9.233891425672247e-06, + "loss": 1.82644615, + "memory(GiB)": 58.14, + "step": 3640, + "train_speed(iter/s)": 0.450938 + }, + { + "acc": 0.62932568, + "epoch": 0.09246575342465753, + "grad_norm": 6.0, + "learning_rate": 9.246575342465755e-06, + "loss": 1.80244827, + "memory(GiB)": 58.14, + "step": 3645, + "train_speed(iter/s)": 0.45126 + }, + { + "acc": 0.6201941, + "epoch": 0.09259259259259259, + "grad_norm": 5.75, + "learning_rate": 9.25925925925926e-06, + "loss": 1.85186386, + "memory(GiB)": 58.14, + "step": 3650, + "train_speed(iter/s)": 0.451579 + }, + { + "acc": 0.62032866, + "epoch": 0.09271943176052765, + "grad_norm": 5.34375, + "learning_rate": 9.271943176052765e-06, + "loss": 1.78758621, + "memory(GiB)": 58.14, + "step": 3655, + "train_speed(iter/s)": 0.4519 + }, + { + "acc": 0.64322929, + "epoch": 0.0928462709284627, + "grad_norm": 5.90625, + "learning_rate": 9.284627092846271e-06, + "loss": 1.77593269, + "memory(GiB)": 58.14, + "step": 3660, + "train_speed(iter/s)": 0.452226 + }, + { + "acc": 0.63530312, + "epoch": 0.09297311009639776, + "grad_norm": 5.5625, + "learning_rate": 9.297311009639777e-06, + "loss": 1.83622093, + "memory(GiB)": 58.14, + "step": 3665, + "train_speed(iter/s)": 0.45255 + }, + { + "acc": 0.61935596, + "epoch": 0.09309994926433282, + "grad_norm": 5.625, + "learning_rate": 9.309994926433283e-06, + "loss": 1.88022728, + "memory(GiB)": 58.14, + "step": 3670, + "train_speed(iter/s)": 0.452872 + }, + { + "acc": 0.60928097, + "epoch": 0.09322678843226788, + "grad_norm": 6.0625, + "learning_rate": 9.322678843226789e-06, + "loss": 1.8983181, + "memory(GiB)": 58.14, + "step": 3675, + "train_speed(iter/s)": 0.453194 + }, + { + "acc": 0.61711197, + "epoch": 0.09335362760020294, + "grad_norm": 7.65625, + "learning_rate": 9.335362760020295e-06, + "loss": 1.86148987, + "memory(GiB)": 58.14, + "step": 3680, + "train_speed(iter/s)": 0.453516 + }, + { + "acc": 0.62589025, + "epoch": 0.093480466768138, + "grad_norm": 5.625, + "learning_rate": 9.348046676813801e-06, + "loss": 1.75282478, + "memory(GiB)": 58.14, + "step": 3685, + "train_speed(iter/s)": 0.453835 + }, + { + "acc": 0.61987696, + "epoch": 0.09360730593607305, + "grad_norm": 4.28125, + "learning_rate": 9.360730593607307e-06, + "loss": 1.81348324, + "memory(GiB)": 58.14, + "step": 3690, + "train_speed(iter/s)": 0.454153 + }, + { + "acc": 0.6352385, + "epoch": 0.09373414510400811, + "grad_norm": 5.5625, + "learning_rate": 9.373414510400813e-06, + "loss": 1.74209366, + "memory(GiB)": 58.14, + "step": 3695, + "train_speed(iter/s)": 0.45447 + }, + { + "acc": 0.60674095, + "epoch": 0.09386098427194317, + "grad_norm": 7.28125, + "learning_rate": 9.386098427194317e-06, + "loss": 1.90820541, + "memory(GiB)": 58.14, + "step": 3700, + "train_speed(iter/s)": 0.454791 + }, + { + "epoch": 0.09386098427194317, + "eval_acc": 0.6154209611405381, + "eval_loss": 1.7795366048812866, + "eval_runtime": 113.7704, + "eval_samples_per_second": 55.99, + "eval_steps_per_second": 27.995, + "step": 3700 + }, + { + "acc": 0.61215563, + "epoch": 0.09398782343987823, + "grad_norm": 5.15625, + "learning_rate": 9.398782343987825e-06, + "loss": 1.88767815, + "memory(GiB)": 58.14, + "step": 3705, + "train_speed(iter/s)": 0.448589 + }, + { + "acc": 0.61937075, + "epoch": 0.09411466260781329, + "grad_norm": 5.71875, + "learning_rate": 9.41146626078133e-06, + "loss": 1.75435829, + "memory(GiB)": 58.14, + "step": 3710, + "train_speed(iter/s)": 0.448905 + }, + { + "acc": 0.63805113, + "epoch": 0.09424150177574835, + "grad_norm": 6.5, + "learning_rate": 9.424150177574835e-06, + "loss": 1.80361481, + "memory(GiB)": 58.14, + "step": 3715, + "train_speed(iter/s)": 0.449224 + }, + { + "acc": 0.61011209, + "epoch": 0.0943683409436834, + "grad_norm": 6.875, + "learning_rate": 9.436834094368341e-06, + "loss": 1.84064255, + "memory(GiB)": 58.14, + "step": 3720, + "train_speed(iter/s)": 0.449537 + }, + { + "acc": 0.63848524, + "epoch": 0.09449518011161846, + "grad_norm": 6.28125, + "learning_rate": 9.449518011161849e-06, + "loss": 1.72705383, + "memory(GiB)": 58.14, + "step": 3725, + "train_speed(iter/s)": 0.449849 + }, + { + "acc": 0.63795338, + "epoch": 0.09462201927955352, + "grad_norm": 5.4375, + "learning_rate": 9.462201927955353e-06, + "loss": 1.72041512, + "memory(GiB)": 58.14, + "step": 3730, + "train_speed(iter/s)": 0.450166 + }, + { + "acc": 0.61618023, + "epoch": 0.09474885844748858, + "grad_norm": 4.8125, + "learning_rate": 9.474885844748859e-06, + "loss": 1.80995827, + "memory(GiB)": 58.14, + "step": 3735, + "train_speed(iter/s)": 0.450481 + }, + { + "acc": 0.61782494, + "epoch": 0.09487569761542364, + "grad_norm": 4.875, + "learning_rate": 9.487569761542365e-06, + "loss": 1.78222961, + "memory(GiB)": 58.14, + "step": 3740, + "train_speed(iter/s)": 0.450794 + }, + { + "acc": 0.63476105, + "epoch": 0.0950025367833587, + "grad_norm": 5.09375, + "learning_rate": 9.50025367833587e-06, + "loss": 1.77090759, + "memory(GiB)": 58.14, + "step": 3745, + "train_speed(iter/s)": 0.451104 + }, + { + "acc": 0.61616001, + "epoch": 0.09512937595129375, + "grad_norm": 4.875, + "learning_rate": 9.512937595129377e-06, + "loss": 1.83497086, + "memory(GiB)": 58.14, + "step": 3750, + "train_speed(iter/s)": 0.451417 + }, + { + "acc": 0.62591643, + "epoch": 0.09525621511922881, + "grad_norm": 6.9375, + "learning_rate": 9.525621511922883e-06, + "loss": 1.85297222, + "memory(GiB)": 58.14, + "step": 3755, + "train_speed(iter/s)": 0.451727 + }, + { + "acc": 0.61138306, + "epoch": 0.09538305428716387, + "grad_norm": 5.09375, + "learning_rate": 9.538305428716389e-06, + "loss": 1.83282948, + "memory(GiB)": 58.14, + "step": 3760, + "train_speed(iter/s)": 0.45204 + }, + { + "acc": 0.62786837, + "epoch": 0.09550989345509893, + "grad_norm": 5.71875, + "learning_rate": 9.550989345509895e-06, + "loss": 1.83867035, + "memory(GiB)": 58.14, + "step": 3765, + "train_speed(iter/s)": 0.452352 + }, + { + "acc": 0.61680865, + "epoch": 0.09563673262303399, + "grad_norm": 5.4375, + "learning_rate": 9.5636732623034e-06, + "loss": 1.82690926, + "memory(GiB)": 58.14, + "step": 3770, + "train_speed(iter/s)": 0.452663 + }, + { + "acc": 0.62666397, + "epoch": 0.09576357179096905, + "grad_norm": 5.375, + "learning_rate": 9.576357179096905e-06, + "loss": 1.76569366, + "memory(GiB)": 58.14, + "step": 3775, + "train_speed(iter/s)": 0.45297 + }, + { + "acc": 0.61075048, + "epoch": 0.0958904109589041, + "grad_norm": 5.34375, + "learning_rate": 9.589041095890411e-06, + "loss": 1.83794975, + "memory(GiB)": 58.14, + "step": 3780, + "train_speed(iter/s)": 0.453277 + }, + { + "acc": 0.60944796, + "epoch": 0.09601725012683916, + "grad_norm": 5.0625, + "learning_rate": 9.601725012683919e-06, + "loss": 1.8591301, + "memory(GiB)": 58.14, + "step": 3785, + "train_speed(iter/s)": 0.453588 + }, + { + "acc": 0.62628341, + "epoch": 0.09614408929477422, + "grad_norm": 5.125, + "learning_rate": 9.614408929477423e-06, + "loss": 1.82615585, + "memory(GiB)": 58.14, + "step": 3790, + "train_speed(iter/s)": 0.453897 + }, + { + "acc": 0.61339693, + "epoch": 0.09627092846270928, + "grad_norm": 4.3125, + "learning_rate": 9.627092846270929e-06, + "loss": 1.8399025, + "memory(GiB)": 58.14, + "step": 3795, + "train_speed(iter/s)": 0.454203 + }, + { + "acc": 0.60255346, + "epoch": 0.09639776763064434, + "grad_norm": 5.78125, + "learning_rate": 9.639776763064435e-06, + "loss": 1.85498562, + "memory(GiB)": 58.14, + "step": 3800, + "train_speed(iter/s)": 0.454516 + }, + { + "epoch": 0.09639776763064434, + "eval_acc": 0.6165802654242823, + "eval_loss": 1.772341251373291, + "eval_runtime": 112.1189, + "eval_samples_per_second": 56.815, + "eval_steps_per_second": 28.407, + "step": 3800 + }, + { + "acc": 0.62302971, + "epoch": 0.0965246067985794, + "grad_norm": 4.3125, + "learning_rate": 9.65246067985794e-06, + "loss": 1.82755051, + "memory(GiB)": 58.14, + "step": 3805, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.62357435, + "epoch": 0.09665144596651445, + "grad_norm": 5.09375, + "learning_rate": 9.665144596651447e-06, + "loss": 1.77882996, + "memory(GiB)": 58.14, + "step": 3810, + "train_speed(iter/s)": 0.448886 + }, + { + "acc": 0.58213558, + "epoch": 0.09677828513444951, + "grad_norm": 7.25, + "learning_rate": 9.677828513444953e-06, + "loss": 1.87296791, + "memory(GiB)": 58.14, + "step": 3815, + "train_speed(iter/s)": 0.449197 + }, + { + "acc": 0.60783415, + "epoch": 0.09690512430238457, + "grad_norm": 6.1875, + "learning_rate": 9.690512430238459e-06, + "loss": 1.90851517, + "memory(GiB)": 58.14, + "step": 3820, + "train_speed(iter/s)": 0.449503 + }, + { + "acc": 0.64452748, + "epoch": 0.09703196347031963, + "grad_norm": 4.78125, + "learning_rate": 9.703196347031965e-06, + "loss": 1.70827713, + "memory(GiB)": 58.14, + "step": 3825, + "train_speed(iter/s)": 0.449808 + }, + { + "acc": 0.63953352, + "epoch": 0.09715880263825469, + "grad_norm": 5.21875, + "learning_rate": 9.71588026382547e-06, + "loss": 1.73099289, + "memory(GiB)": 58.14, + "step": 3830, + "train_speed(iter/s)": 0.450112 + }, + { + "acc": 0.60430994, + "epoch": 0.09728564180618975, + "grad_norm": 5.25, + "learning_rate": 9.728564180618977e-06, + "loss": 1.86597404, + "memory(GiB)": 58.14, + "step": 3835, + "train_speed(iter/s)": 0.45042 + }, + { + "acc": 0.61606503, + "epoch": 0.0974124809741248, + "grad_norm": 5.125, + "learning_rate": 9.74124809741248e-06, + "loss": 1.85424938, + "memory(GiB)": 58.14, + "step": 3840, + "train_speed(iter/s)": 0.450728 + }, + { + "acc": 0.61780777, + "epoch": 0.09753932014205986, + "grad_norm": 5.09375, + "learning_rate": 9.753932014205988e-06, + "loss": 1.8107111, + "memory(GiB)": 58.14, + "step": 3845, + "train_speed(iter/s)": 0.451034 + }, + { + "acc": 0.63499184, + "epoch": 0.09766615930999492, + "grad_norm": 4.8125, + "learning_rate": 9.766615930999493e-06, + "loss": 1.7432745, + "memory(GiB)": 58.14, + "step": 3850, + "train_speed(iter/s)": 0.45134 + }, + { + "acc": 0.63544383, + "epoch": 0.09779299847792998, + "grad_norm": 5.0625, + "learning_rate": 9.779299847792999e-06, + "loss": 1.72224121, + "memory(GiB)": 58.14, + "step": 3855, + "train_speed(iter/s)": 0.451643 + }, + { + "acc": 0.62918215, + "epoch": 0.09791983764586504, + "grad_norm": 5.59375, + "learning_rate": 9.791983764586505e-06, + "loss": 1.72505569, + "memory(GiB)": 58.14, + "step": 3860, + "train_speed(iter/s)": 0.451946 + }, + { + "acc": 0.62854624, + "epoch": 0.0980466768138001, + "grad_norm": 5.03125, + "learning_rate": 9.80466768138001e-06, + "loss": 1.82730312, + "memory(GiB)": 58.14, + "step": 3865, + "train_speed(iter/s)": 0.452251 + }, + { + "acc": 0.61791906, + "epoch": 0.09817351598173515, + "grad_norm": 5.40625, + "learning_rate": 9.817351598173517e-06, + "loss": 1.85250893, + "memory(GiB)": 58.14, + "step": 3870, + "train_speed(iter/s)": 0.452555 + }, + { + "acc": 0.62267046, + "epoch": 0.09830035514967021, + "grad_norm": 4.59375, + "learning_rate": 9.830035514967023e-06, + "loss": 1.83560543, + "memory(GiB)": 58.14, + "step": 3875, + "train_speed(iter/s)": 0.452858 + }, + { + "acc": 0.61110196, + "epoch": 0.09842719431760527, + "grad_norm": 5.28125, + "learning_rate": 9.842719431760529e-06, + "loss": 1.89259396, + "memory(GiB)": 58.14, + "step": 3880, + "train_speed(iter/s)": 0.453162 + }, + { + "acc": 0.62783184, + "epoch": 0.09855403348554033, + "grad_norm": 6.59375, + "learning_rate": 9.855403348554034e-06, + "loss": 1.77797241, + "memory(GiB)": 58.14, + "step": 3885, + "train_speed(iter/s)": 0.453466 + }, + { + "acc": 0.59533892, + "epoch": 0.09868087265347539, + "grad_norm": 5.40625, + "learning_rate": 9.86808726534754e-06, + "loss": 1.84901543, + "memory(GiB)": 58.14, + "step": 3890, + "train_speed(iter/s)": 0.453772 + }, + { + "acc": 0.62068543, + "epoch": 0.09880771182141045, + "grad_norm": 5.1875, + "learning_rate": 9.880771182141046e-06, + "loss": 1.88023758, + "memory(GiB)": 58.14, + "step": 3895, + "train_speed(iter/s)": 0.454073 + }, + { + "acc": 0.63935137, + "epoch": 0.0989345509893455, + "grad_norm": 4.90625, + "learning_rate": 9.89345509893455e-06, + "loss": 1.77827415, + "memory(GiB)": 58.14, + "step": 3900, + "train_speed(iter/s)": 0.454376 + }, + { + "epoch": 0.0989345509893455, + "eval_acc": 0.617520659637871, + "eval_loss": 1.765673041343689, + "eval_runtime": 113.4872, + "eval_samples_per_second": 56.13, + "eval_steps_per_second": 28.065, + "step": 3900 + }, + { + "acc": 0.63380542, + "epoch": 0.09906139015728056, + "grad_norm": 5.6875, + "learning_rate": 9.906139015728058e-06, + "loss": 1.78636627, + "memory(GiB)": 58.14, + "step": 3905, + "train_speed(iter/s)": 0.448521 + }, + { + "acc": 0.61196938, + "epoch": 0.09918822932521562, + "grad_norm": 4.71875, + "learning_rate": 9.918822932521563e-06, + "loss": 1.84834785, + "memory(GiB)": 58.14, + "step": 3910, + "train_speed(iter/s)": 0.44882 + }, + { + "acc": 0.62668991, + "epoch": 0.09931506849315068, + "grad_norm": 4.78125, + "learning_rate": 9.931506849315069e-06, + "loss": 1.82334385, + "memory(GiB)": 58.14, + "step": 3915, + "train_speed(iter/s)": 0.449121 + }, + { + "acc": 0.62209506, + "epoch": 0.09944190766108574, + "grad_norm": 4.96875, + "learning_rate": 9.944190766108575e-06, + "loss": 1.8567482, + "memory(GiB)": 58.14, + "step": 3920, + "train_speed(iter/s)": 0.449421 + }, + { + "acc": 0.63411107, + "epoch": 0.0995687468290208, + "grad_norm": 5.53125, + "learning_rate": 9.95687468290208e-06, + "loss": 1.70525742, + "memory(GiB)": 58.14, + "step": 3925, + "train_speed(iter/s)": 0.449723 + }, + { + "acc": 0.6268219, + "epoch": 0.09969558599695585, + "grad_norm": 5.90625, + "learning_rate": 9.969558599695586e-06, + "loss": 1.79684181, + "memory(GiB)": 58.14, + "step": 3930, + "train_speed(iter/s)": 0.450022 + }, + { + "acc": 0.61677246, + "epoch": 0.09982242516489091, + "grad_norm": 5.5625, + "learning_rate": 9.982242516489092e-06, + "loss": 1.82491226, + "memory(GiB)": 58.14, + "step": 3935, + "train_speed(iter/s)": 0.450321 + }, + { + "acc": 0.63249168, + "epoch": 0.09994926433282597, + "grad_norm": 5.21875, + "learning_rate": 9.994926433282598e-06, + "loss": 1.74161472, + "memory(GiB)": 58.14, + "step": 3940, + "train_speed(iter/s)": 0.450622 + }, + { + "acc": 0.6476191, + "epoch": 0.10007610350076103, + "grad_norm": 4.8125, + "learning_rate": 9.999999960413982e-06, + "loss": 1.72309265, + "memory(GiB)": 58.14, + "step": 3945, + "train_speed(iter/s)": 0.450921 + }, + { + "acc": 0.61407857, + "epoch": 0.10020294266869609, + "grad_norm": 9.1875, + "learning_rate": 9.99999971849943e-06, + "loss": 1.84395828, + "memory(GiB)": 58.14, + "step": 3950, + "train_speed(iter/s)": 0.451221 + }, + { + "acc": 0.62937422, + "epoch": 0.10032978183663115, + "grad_norm": 5.15625, + "learning_rate": 9.999999256662563e-06, + "loss": 1.8419838, + "memory(GiB)": 58.14, + "step": 3955, + "train_speed(iter/s)": 0.45152 + }, + { + "acc": 0.64251051, + "epoch": 0.1004566210045662, + "grad_norm": 5.59375, + "learning_rate": 9.999998574903408e-06, + "loss": 1.70181484, + "memory(GiB)": 58.14, + "step": 3960, + "train_speed(iter/s)": 0.451819 + }, + { + "acc": 0.62504053, + "epoch": 0.10058346017250126, + "grad_norm": 5.875, + "learning_rate": 9.99999767322199e-06, + "loss": 1.79243317, + "memory(GiB)": 58.14, + "step": 3965, + "train_speed(iter/s)": 0.452117 + }, + { + "acc": 0.62244406, + "epoch": 0.10071029934043632, + "grad_norm": 4.90625, + "learning_rate": 9.999996551618353e-06, + "loss": 1.75773678, + "memory(GiB)": 58.14, + "step": 3970, + "train_speed(iter/s)": 0.452414 + }, + { + "acc": 0.60930867, + "epoch": 0.10083713850837138, + "grad_norm": 5.84375, + "learning_rate": 9.999995210092545e-06, + "loss": 1.82529449, + "memory(GiB)": 58.14, + "step": 3975, + "train_speed(iter/s)": 0.452709 + }, + { + "acc": 0.61774755, + "epoch": 0.10096397767630644, + "grad_norm": 6.625, + "learning_rate": 9.999993648644622e-06, + "loss": 1.78912544, + "memory(GiB)": 58.14, + "step": 3980, + "train_speed(iter/s)": 0.453007 + }, + { + "acc": 0.62640629, + "epoch": 0.1010908168442415, + "grad_norm": 6.0625, + "learning_rate": 9.999991867274656e-06, + "loss": 1.80796242, + "memory(GiB)": 58.14, + "step": 3985, + "train_speed(iter/s)": 0.453304 + }, + { + "acc": 0.62726254, + "epoch": 0.10121765601217655, + "grad_norm": 6.09375, + "learning_rate": 9.999989865982725e-06, + "loss": 1.75446987, + "memory(GiB)": 58.14, + "step": 3990, + "train_speed(iter/s)": 0.453602 + }, + { + "acc": 0.60661621, + "epoch": 0.10134449518011161, + "grad_norm": 5.09375, + "learning_rate": 9.999987644768917e-06, + "loss": 1.75037861, + "memory(GiB)": 58.14, + "step": 3995, + "train_speed(iter/s)": 0.453895 + }, + { + "acc": 0.62552071, + "epoch": 0.10147133434804667, + "grad_norm": 5.53125, + "learning_rate": 9.999985203633327e-06, + "loss": 1.80774097, + "memory(GiB)": 58.14, + "step": 4000, + "train_speed(iter/s)": 0.454188 + }, + { + "epoch": 0.10147133434804667, + "eval_acc": 0.6185696733519184, + "eval_loss": 1.7597835063934326, + "eval_runtime": 112.8153, + "eval_samples_per_second": 56.464, + "eval_steps_per_second": 28.232, + "step": 4000 + }, + { + "acc": 0.6291091, + "epoch": 0.10159817351598173, + "grad_norm": 6.84375, + "learning_rate": 9.999982542576065e-06, + "loss": 1.76779251, + "memory(GiB)": 58.14, + "step": 4005, + "train_speed(iter/s)": 0.448516 + }, + { + "acc": 0.6216218, + "epoch": 0.10172501268391679, + "grad_norm": 6.03125, + "learning_rate": 9.999979661597247e-06, + "loss": 1.82360649, + "memory(GiB)": 58.14, + "step": 4010, + "train_speed(iter/s)": 0.44881 + }, + { + "acc": 0.63545084, + "epoch": 0.10185185185185185, + "grad_norm": 5.3125, + "learning_rate": 9.999976560697002e-06, + "loss": 1.74934006, + "memory(GiB)": 58.14, + "step": 4015, + "train_speed(iter/s)": 0.449103 + }, + { + "acc": 0.64277163, + "epoch": 0.1019786910197869, + "grad_norm": 5.84375, + "learning_rate": 9.999973239875462e-06, + "loss": 1.77237892, + "memory(GiB)": 58.14, + "step": 4020, + "train_speed(iter/s)": 0.449396 + }, + { + "acc": 0.63172255, + "epoch": 0.10210553018772196, + "grad_norm": 5.8125, + "learning_rate": 9.999969699132776e-06, + "loss": 1.76154366, + "memory(GiB)": 58.14, + "step": 4025, + "train_speed(iter/s)": 0.449691 + }, + { + "acc": 0.61161127, + "epoch": 0.10223236935565702, + "grad_norm": 4.625, + "learning_rate": 9.999965938469102e-06, + "loss": 1.8439537, + "memory(GiB)": 58.14, + "step": 4030, + "train_speed(iter/s)": 0.449985 + }, + { + "acc": 0.62062168, + "epoch": 0.10235920852359208, + "grad_norm": 5.75, + "learning_rate": 9.9999619578846e-06, + "loss": 1.81092777, + "memory(GiB)": 58.14, + "step": 4035, + "train_speed(iter/s)": 0.450279 + }, + { + "acc": 0.62890124, + "epoch": 0.10248604769152714, + "grad_norm": 6.4375, + "learning_rate": 9.999957757379451e-06, + "loss": 1.81078758, + "memory(GiB)": 58.14, + "step": 4040, + "train_speed(iter/s)": 0.450568 + }, + { + "acc": 0.63805218, + "epoch": 0.1026128868594622, + "grad_norm": 7.3125, + "learning_rate": 9.999953336953834e-06, + "loss": 1.71456261, + "memory(GiB)": 58.14, + "step": 4045, + "train_speed(iter/s)": 0.450857 + }, + { + "acc": 0.61768446, + "epoch": 0.10273972602739725, + "grad_norm": 5.28125, + "learning_rate": 9.999948696607946e-06, + "loss": 1.78419456, + "memory(GiB)": 58.14, + "step": 4050, + "train_speed(iter/s)": 0.451146 + }, + { + "acc": 0.62755127, + "epoch": 0.10286656519533231, + "grad_norm": 4.84375, + "learning_rate": 9.999943836341992e-06, + "loss": 1.80323219, + "memory(GiB)": 58.14, + "step": 4055, + "train_speed(iter/s)": 0.451438 + }, + { + "acc": 0.65122805, + "epoch": 0.10299340436326737, + "grad_norm": 5.34375, + "learning_rate": 9.999938756156185e-06, + "loss": 1.69991837, + "memory(GiB)": 58.14, + "step": 4060, + "train_speed(iter/s)": 0.451729 + }, + { + "acc": 0.62248263, + "epoch": 0.10312024353120243, + "grad_norm": 5.4375, + "learning_rate": 9.999933456050747e-06, + "loss": 1.81368866, + "memory(GiB)": 58.14, + "step": 4065, + "train_speed(iter/s)": 0.452019 + }, + { + "acc": 0.61825347, + "epoch": 0.10324708269913749, + "grad_norm": 5.0, + "learning_rate": 9.999927936025914e-06, + "loss": 1.84223785, + "memory(GiB)": 58.14, + "step": 4070, + "train_speed(iter/s)": 0.452312 + }, + { + "acc": 0.63171749, + "epoch": 0.10337392186707255, + "grad_norm": 5.4375, + "learning_rate": 9.999922196081928e-06, + "loss": 1.73060284, + "memory(GiB)": 58.14, + "step": 4075, + "train_speed(iter/s)": 0.452605 + }, + { + "acc": 0.62899675, + "epoch": 0.1035007610350076, + "grad_norm": 5.0625, + "learning_rate": 9.99991623621904e-06, + "loss": 1.74504051, + "memory(GiB)": 58.14, + "step": 4080, + "train_speed(iter/s)": 0.452894 + }, + { + "acc": 0.6265625, + "epoch": 0.10362760020294266, + "grad_norm": 5.625, + "learning_rate": 9.999910056437512e-06, + "loss": 1.79673958, + "memory(GiB)": 58.14, + "step": 4085, + "train_speed(iter/s)": 0.453187 + }, + { + "acc": 0.62976837, + "epoch": 0.10375443937087772, + "grad_norm": 4.4375, + "learning_rate": 9.999903656737618e-06, + "loss": 1.79148903, + "memory(GiB)": 58.14, + "step": 4090, + "train_speed(iter/s)": 0.453479 + }, + { + "acc": 0.62661695, + "epoch": 0.10388127853881278, + "grad_norm": 5.21875, + "learning_rate": 9.999897037119637e-06, + "loss": 1.74700851, + "memory(GiB)": 58.14, + "step": 4095, + "train_speed(iter/s)": 0.45377 + }, + { + "acc": 0.62962437, + "epoch": 0.10400811770674784, + "grad_norm": 5.28125, + "learning_rate": 9.999890197583862e-06, + "loss": 1.82050076, + "memory(GiB)": 58.14, + "step": 4100, + "train_speed(iter/s)": 0.45406 + }, + { + "epoch": 0.10400811770674784, + "eval_acc": 0.6191967420834139, + "eval_loss": 1.7540366649627686, + "eval_runtime": 112.4618, + "eval_samples_per_second": 56.641, + "eval_steps_per_second": 28.321, + "step": 4100 + }, + { + "acc": 0.61344967, + "epoch": 0.1041349568746829, + "grad_norm": 6.46875, + "learning_rate": 9.999883138130593e-06, + "loss": 1.83070374, + "memory(GiB)": 58.14, + "step": 4105, + "train_speed(iter/s)": 0.448541 + }, + { + "acc": 0.63450255, + "epoch": 0.10426179604261795, + "grad_norm": 4.90625, + "learning_rate": 9.999875858760143e-06, + "loss": 1.77065277, + "memory(GiB)": 58.14, + "step": 4110, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.62327752, + "epoch": 0.10438863521055301, + "grad_norm": 5.53125, + "learning_rate": 9.999868359472826e-06, + "loss": 1.86088276, + "memory(GiB)": 58.14, + "step": 4115, + "train_speed(iter/s)": 0.449113 + }, + { + "acc": 0.62800627, + "epoch": 0.10451547437848807, + "grad_norm": 4.8125, + "learning_rate": 9.999860640268977e-06, + "loss": 1.79443016, + "memory(GiB)": 58.14, + "step": 4120, + "train_speed(iter/s)": 0.449397 + }, + { + "acc": 0.63783755, + "epoch": 0.10464231354642313, + "grad_norm": 5.5625, + "learning_rate": 9.999852701148935e-06, + "loss": 1.77015991, + "memory(GiB)": 58.14, + "step": 4125, + "train_speed(iter/s)": 0.44968 + }, + { + "acc": 0.64586406, + "epoch": 0.10476915271435819, + "grad_norm": 5.90625, + "learning_rate": 9.999844542113049e-06, + "loss": 1.78799438, + "memory(GiB)": 58.14, + "step": 4130, + "train_speed(iter/s)": 0.449967 + }, + { + "acc": 0.61216378, + "epoch": 0.10489599188229325, + "grad_norm": 5.09375, + "learning_rate": 9.999836163161675e-06, + "loss": 1.85857239, + "memory(GiB)": 58.14, + "step": 4135, + "train_speed(iter/s)": 0.450253 + }, + { + "acc": 0.6562192, + "epoch": 0.1050228310502283, + "grad_norm": 5.65625, + "learning_rate": 9.999827564295187e-06, + "loss": 1.67194862, + "memory(GiB)": 58.14, + "step": 4140, + "train_speed(iter/s)": 0.450539 + }, + { + "acc": 0.6264616, + "epoch": 0.10514967021816336, + "grad_norm": 5.21875, + "learning_rate": 9.999818745513958e-06, + "loss": 1.75949821, + "memory(GiB)": 58.14, + "step": 4145, + "train_speed(iter/s)": 0.450822 + }, + { + "acc": 0.6279695, + "epoch": 0.10527650938609842, + "grad_norm": 5.28125, + "learning_rate": 9.99980970681838e-06, + "loss": 1.7826004, + "memory(GiB)": 58.14, + "step": 4150, + "train_speed(iter/s)": 0.451106 + }, + { + "acc": 0.6429636, + "epoch": 0.10540334855403348, + "grad_norm": 6.03125, + "learning_rate": 9.999800448208846e-06, + "loss": 1.7293129, + "memory(GiB)": 58.14, + "step": 4155, + "train_speed(iter/s)": 0.451392 + }, + { + "acc": 0.62088828, + "epoch": 0.10553018772196854, + "grad_norm": 5.03125, + "learning_rate": 9.999790969685767e-06, + "loss": 1.76002274, + "memory(GiB)": 58.14, + "step": 4160, + "train_speed(iter/s)": 0.451673 + }, + { + "acc": 0.61855145, + "epoch": 0.1056570268899036, + "grad_norm": 7.0625, + "learning_rate": 9.999781271249559e-06, + "loss": 1.78885899, + "memory(GiB)": 58.14, + "step": 4165, + "train_speed(iter/s)": 0.451956 + }, + { + "acc": 0.63343887, + "epoch": 0.10578386605783865, + "grad_norm": 5.09375, + "learning_rate": 9.999771352900647e-06, + "loss": 1.74517441, + "memory(GiB)": 58.14, + "step": 4170, + "train_speed(iter/s)": 0.452238 + }, + { + "acc": 0.63536668, + "epoch": 0.10591070522577371, + "grad_norm": 5.1875, + "learning_rate": 9.999761214639469e-06, + "loss": 1.61596909, + "memory(GiB)": 58.14, + "step": 4175, + "train_speed(iter/s)": 0.452519 + }, + { + "acc": 0.6453671, + "epoch": 0.10603754439370877, + "grad_norm": 4.4375, + "learning_rate": 9.999750856466472e-06, + "loss": 1.68646221, + "memory(GiB)": 58.14, + "step": 4180, + "train_speed(iter/s)": 0.4528 + }, + { + "acc": 0.63252106, + "epoch": 0.10616438356164383, + "grad_norm": 4.96875, + "learning_rate": 9.99974027838211e-06, + "loss": 1.70660515, + "memory(GiB)": 58.14, + "step": 4185, + "train_speed(iter/s)": 0.453082 + }, + { + "acc": 0.63777542, + "epoch": 0.10629122272957889, + "grad_norm": 4.09375, + "learning_rate": 9.999729480386846e-06, + "loss": 1.72413597, + "memory(GiB)": 58.14, + "step": 4190, + "train_speed(iter/s)": 0.453364 + }, + { + "acc": 0.63738689, + "epoch": 0.10641806189751395, + "grad_norm": 5.40625, + "learning_rate": 9.999718462481157e-06, + "loss": 1.72675591, + "memory(GiB)": 58.14, + "step": 4195, + "train_speed(iter/s)": 0.453644 + }, + { + "acc": 0.64150829, + "epoch": 0.106544901065449, + "grad_norm": 4.75, + "learning_rate": 9.99970722466553e-06, + "loss": 1.73381214, + "memory(GiB)": 58.14, + "step": 4200, + "train_speed(iter/s)": 0.453925 + }, + { + "epoch": 0.106544901065449, + "eval_acc": 0.6201346296931457, + "eval_loss": 1.7487373352050781, + "eval_runtime": 113.0521, + "eval_samples_per_second": 56.346, + "eval_steps_per_second": 28.173, + "step": 4200 + }, + { + "acc": 0.63454752, + "epoch": 0.10667174023338406, + "grad_norm": 6.09375, + "learning_rate": 9.999695766940458e-06, + "loss": 1.72905521, + "memory(GiB)": 58.14, + "step": 4205, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.6222558, + "epoch": 0.10679857940131912, + "grad_norm": 5.0, + "learning_rate": 9.999684089306442e-06, + "loss": 1.76112251, + "memory(GiB)": 58.14, + "step": 4210, + "train_speed(iter/s)": 0.448794 + }, + { + "acc": 0.63157911, + "epoch": 0.10692541856925418, + "grad_norm": 4.875, + "learning_rate": 9.999672191763999e-06, + "loss": 1.75762348, + "memory(GiB)": 58.14, + "step": 4215, + "train_speed(iter/s)": 0.449073 + }, + { + "acc": 0.64144144, + "epoch": 0.10705225773718924, + "grad_norm": 5.53125, + "learning_rate": 9.99966007431365e-06, + "loss": 1.64606743, + "memory(GiB)": 58.14, + "step": 4220, + "train_speed(iter/s)": 0.449351 + }, + { + "acc": 0.63285217, + "epoch": 0.1071790969051243, + "grad_norm": 4.125, + "learning_rate": 9.99964773695593e-06, + "loss": 1.75802803, + "memory(GiB)": 58.14, + "step": 4225, + "train_speed(iter/s)": 0.44963 + }, + { + "acc": 0.6388093, + "epoch": 0.10730593607305935, + "grad_norm": 5.40625, + "learning_rate": 9.999635179691381e-06, + "loss": 1.76360455, + "memory(GiB)": 58.14, + "step": 4230, + "train_speed(iter/s)": 0.44991 + }, + { + "acc": 0.62783031, + "epoch": 0.10743277524099441, + "grad_norm": 4.96875, + "learning_rate": 9.999622402520553e-06, + "loss": 1.81153564, + "memory(GiB)": 58.14, + "step": 4235, + "train_speed(iter/s)": 0.450188 + }, + { + "acc": 0.61198092, + "epoch": 0.10755961440892947, + "grad_norm": 6.4375, + "learning_rate": 9.999609405444012e-06, + "loss": 1.80095291, + "memory(GiB)": 58.14, + "step": 4240, + "train_speed(iter/s)": 0.450463 + }, + { + "acc": 0.63580613, + "epoch": 0.10768645357686453, + "grad_norm": 5.40625, + "learning_rate": 9.999596188462328e-06, + "loss": 1.69204712, + "memory(GiB)": 58.14, + "step": 4245, + "train_speed(iter/s)": 0.450746 + }, + { + "acc": 0.63853812, + "epoch": 0.10781329274479959, + "grad_norm": 4.34375, + "learning_rate": 9.99958275157608e-06, + "loss": 1.71433411, + "memory(GiB)": 58.14, + "step": 4250, + "train_speed(iter/s)": 0.451026 + }, + { + "acc": 0.63923063, + "epoch": 0.10794013191273465, + "grad_norm": 5.375, + "learning_rate": 9.999569094785862e-06, + "loss": 1.81869488, + "memory(GiB)": 58.14, + "step": 4255, + "train_speed(iter/s)": 0.45131 + }, + { + "acc": 0.6371985, + "epoch": 0.1080669710806697, + "grad_norm": 5.625, + "learning_rate": 9.999555218092273e-06, + "loss": 1.74946365, + "memory(GiB)": 58.14, + "step": 4260, + "train_speed(iter/s)": 0.451592 + }, + { + "acc": 0.63811803, + "epoch": 0.10819381024860476, + "grad_norm": 5.65625, + "learning_rate": 9.999541121495926e-06, + "loss": 1.75966206, + "memory(GiB)": 58.14, + "step": 4265, + "train_speed(iter/s)": 0.451874 + }, + { + "acc": 0.62519026, + "epoch": 0.10832064941653982, + "grad_norm": 6.03125, + "learning_rate": 9.999526804997439e-06, + "loss": 1.81126308, + "memory(GiB)": 58.14, + "step": 4270, + "train_speed(iter/s)": 0.452156 + }, + { + "acc": 0.6206193, + "epoch": 0.10844748858447488, + "grad_norm": 5.0, + "learning_rate": 9.99951226859744e-06, + "loss": 1.8414917, + "memory(GiB)": 58.14, + "step": 4275, + "train_speed(iter/s)": 0.452435 + }, + { + "acc": 0.63537245, + "epoch": 0.10857432775240994, + "grad_norm": 5.21875, + "learning_rate": 9.999497512296572e-06, + "loss": 1.76503983, + "memory(GiB)": 58.14, + "step": 4280, + "train_speed(iter/s)": 0.452716 + }, + { + "acc": 0.64053645, + "epoch": 0.108701166920345, + "grad_norm": 7.5, + "learning_rate": 9.999482536095483e-06, + "loss": 1.75181522, + "memory(GiB)": 58.14, + "step": 4285, + "train_speed(iter/s)": 0.452997 + }, + { + "acc": 0.62255592, + "epoch": 0.10882800608828005, + "grad_norm": 4.59375, + "learning_rate": 9.999467339994827e-06, + "loss": 1.74321728, + "memory(GiB)": 58.14, + "step": 4290, + "train_speed(iter/s)": 0.453276 + }, + { + "acc": 0.6438282, + "epoch": 0.10895484525621511, + "grad_norm": 5.40625, + "learning_rate": 9.99945192399528e-06, + "loss": 1.74759979, + "memory(GiB)": 58.14, + "step": 4295, + "train_speed(iter/s)": 0.453557 + }, + { + "acc": 0.62297983, + "epoch": 0.10908168442415017, + "grad_norm": 5.90625, + "learning_rate": 9.999436288097515e-06, + "loss": 1.76541634, + "memory(GiB)": 58.14, + "step": 4300, + "train_speed(iter/s)": 0.453833 + }, + { + "epoch": 0.10908168442415017, + "eval_acc": 0.6208949661963582, + "eval_loss": 1.7438240051269531, + "eval_runtime": 113.7389, + "eval_samples_per_second": 56.005, + "eval_steps_per_second": 28.003, + "step": 4300 + }, + { + "acc": 0.63015323, + "epoch": 0.10920852359208523, + "grad_norm": 6.1875, + "learning_rate": 9.99942043230222e-06, + "loss": 1.80679054, + "memory(GiB)": 58.14, + "step": 4305, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.63322983, + "epoch": 0.10933536276002029, + "grad_norm": 5.40625, + "learning_rate": 9.999404356610095e-06, + "loss": 1.69326534, + "memory(GiB)": 58.14, + "step": 4310, + "train_speed(iter/s)": 0.448786 + }, + { + "acc": 0.64163394, + "epoch": 0.10946220192795535, + "grad_norm": 6.15625, + "learning_rate": 9.999388061021846e-06, + "loss": 1.79806824, + "memory(GiB)": 58.14, + "step": 4315, + "train_speed(iter/s)": 0.44906 + }, + { + "acc": 0.62432628, + "epoch": 0.1095890410958904, + "grad_norm": 5.3125, + "learning_rate": 9.99937154553819e-06, + "loss": 1.79620895, + "memory(GiB)": 58.14, + "step": 4320, + "train_speed(iter/s)": 0.449332 + }, + { + "acc": 0.61566505, + "epoch": 0.10971588026382546, + "grad_norm": 5.6875, + "learning_rate": 9.999354810159852e-06, + "loss": 1.76232872, + "memory(GiB)": 58.14, + "step": 4325, + "train_speed(iter/s)": 0.449602 + }, + { + "acc": 0.62264271, + "epoch": 0.10984271943176052, + "grad_norm": 5.34375, + "learning_rate": 9.999337854887567e-06, + "loss": 1.73234653, + "memory(GiB)": 58.14, + "step": 4330, + "train_speed(iter/s)": 0.449872 + }, + { + "acc": 0.63787823, + "epoch": 0.10996955859969558, + "grad_norm": 5.21875, + "learning_rate": 9.999320679722086e-06, + "loss": 1.74618568, + "memory(GiB)": 58.14, + "step": 4335, + "train_speed(iter/s)": 0.450143 + }, + { + "acc": 0.62766075, + "epoch": 0.11009639776763064, + "grad_norm": 4.84375, + "learning_rate": 9.999303284664159e-06, + "loss": 1.71300373, + "memory(GiB)": 58.14, + "step": 4340, + "train_speed(iter/s)": 0.450413 + }, + { + "acc": 0.64673505, + "epoch": 0.1102232369355657, + "grad_norm": 4.78125, + "learning_rate": 9.999285669714555e-06, + "loss": 1.71865788, + "memory(GiB)": 58.14, + "step": 4345, + "train_speed(iter/s)": 0.450683 + }, + { + "acc": 0.64527864, + "epoch": 0.11035007610350075, + "grad_norm": 5.03125, + "learning_rate": 9.999267834874044e-06, + "loss": 1.65094757, + "memory(GiB)": 58.14, + "step": 4350, + "train_speed(iter/s)": 0.450956 + }, + { + "acc": 0.63289995, + "epoch": 0.11047691527143581, + "grad_norm": 5.96875, + "learning_rate": 9.999249780143416e-06, + "loss": 1.72538681, + "memory(GiB)": 58.14, + "step": 4355, + "train_speed(iter/s)": 0.451225 + }, + { + "acc": 0.6310523, + "epoch": 0.11060375443937087, + "grad_norm": 4.78125, + "learning_rate": 9.999231505523463e-06, + "loss": 1.7701992, + "memory(GiB)": 58.14, + "step": 4360, + "train_speed(iter/s)": 0.451497 + }, + { + "acc": 0.60786505, + "epoch": 0.11073059360730593, + "grad_norm": 5.03125, + "learning_rate": 9.999213011014987e-06, + "loss": 1.81345711, + "memory(GiB)": 58.14, + "step": 4365, + "train_speed(iter/s)": 0.451769 + }, + { + "acc": 0.62757382, + "epoch": 0.11085743277524099, + "grad_norm": 4.71875, + "learning_rate": 9.999194296618805e-06, + "loss": 1.76174946, + "memory(GiB)": 58.14, + "step": 4370, + "train_speed(iter/s)": 0.452036 + }, + { + "acc": 0.63591194, + "epoch": 0.11098427194317605, + "grad_norm": 6.8125, + "learning_rate": 9.999175362335735e-06, + "loss": 1.74051361, + "memory(GiB)": 58.14, + "step": 4375, + "train_speed(iter/s)": 0.452308 + }, + { + "acc": 0.63282638, + "epoch": 0.1111111111111111, + "grad_norm": 5.28125, + "learning_rate": 9.999156208166614e-06, + "loss": 1.78131981, + "memory(GiB)": 58.14, + "step": 4380, + "train_speed(iter/s)": 0.452581 + }, + { + "acc": 0.62626019, + "epoch": 0.11123795027904616, + "grad_norm": 5.5625, + "learning_rate": 9.999136834112284e-06, + "loss": 1.74566879, + "memory(GiB)": 58.14, + "step": 4385, + "train_speed(iter/s)": 0.45285 + }, + { + "acc": 0.63161502, + "epoch": 0.11136478944698122, + "grad_norm": 4.34375, + "learning_rate": 9.999117240173597e-06, + "loss": 1.79278793, + "memory(GiB)": 58.14, + "step": 4390, + "train_speed(iter/s)": 0.453119 + }, + { + "acc": 0.62327871, + "epoch": 0.11149162861491628, + "grad_norm": 5.65625, + "learning_rate": 9.999097426351412e-06, + "loss": 1.84273453, + "memory(GiB)": 58.14, + "step": 4395, + "train_speed(iter/s)": 0.453391 + }, + { + "acc": 0.62689791, + "epoch": 0.11161846778285134, + "grad_norm": 5.40625, + "learning_rate": 9.999077392646606e-06, + "loss": 1.75429306, + "memory(GiB)": 58.14, + "step": 4400, + "train_speed(iter/s)": 0.45366 + }, + { + "epoch": 0.11161846778285134, + "eval_acc": 0.6214852714046214, + "eval_loss": 1.739027500152588, + "eval_runtime": 114.3722, + "eval_samples_per_second": 55.695, + "eval_steps_per_second": 27.848, + "step": 4400 + }, + { + "acc": 0.63877306, + "epoch": 0.1117453069507864, + "grad_norm": 5.5625, + "learning_rate": 9.999057139060055e-06, + "loss": 1.75726471, + "memory(GiB)": 58.14, + "step": 4405, + "train_speed(iter/s)": 0.448437 + }, + { + "acc": 0.6433804, + "epoch": 0.11187214611872145, + "grad_norm": 4.84375, + "learning_rate": 9.999036665592653e-06, + "loss": 1.71312466, + "memory(GiB)": 58.14, + "step": 4410, + "train_speed(iter/s)": 0.448704 + }, + { + "acc": 0.64489183, + "epoch": 0.11199898528665651, + "grad_norm": 4.96875, + "learning_rate": 9.999015972245298e-06, + "loss": 1.73755589, + "memory(GiB)": 58.14, + "step": 4415, + "train_speed(iter/s)": 0.448971 + }, + { + "acc": 0.63046446, + "epoch": 0.11212582445459157, + "grad_norm": 5.46875, + "learning_rate": 9.998995059018901e-06, + "loss": 1.78997326, + "memory(GiB)": 58.14, + "step": 4420, + "train_speed(iter/s)": 0.449235 + }, + { + "acc": 0.62649784, + "epoch": 0.11225266362252663, + "grad_norm": 5.65625, + "learning_rate": 9.998973925914384e-06, + "loss": 1.73370438, + "memory(GiB)": 58.14, + "step": 4425, + "train_speed(iter/s)": 0.449499 + }, + { + "acc": 0.62724686, + "epoch": 0.11237950279046169, + "grad_norm": 5.0625, + "learning_rate": 9.998952572932675e-06, + "loss": 1.77701378, + "memory(GiB)": 58.14, + "step": 4430, + "train_speed(iter/s)": 0.449763 + }, + { + "acc": 0.61130695, + "epoch": 0.11250634195839675, + "grad_norm": 5.34375, + "learning_rate": 9.998931000074712e-06, + "loss": 1.85735683, + "memory(GiB)": 58.14, + "step": 4435, + "train_speed(iter/s)": 0.450032 + }, + { + "acc": 0.62680559, + "epoch": 0.1126331811263318, + "grad_norm": 5.34375, + "learning_rate": 9.998909207341446e-06, + "loss": 1.77192841, + "memory(GiB)": 58.14, + "step": 4440, + "train_speed(iter/s)": 0.450298 + }, + { + "acc": 0.61906872, + "epoch": 0.11276002029426686, + "grad_norm": 5.46875, + "learning_rate": 9.998887194733833e-06, + "loss": 1.85887356, + "memory(GiB)": 58.14, + "step": 4445, + "train_speed(iter/s)": 0.450565 + }, + { + "acc": 0.64016647, + "epoch": 0.11288685946220192, + "grad_norm": 5.125, + "learning_rate": 9.998864962252843e-06, + "loss": 1.71196823, + "memory(GiB)": 58.14, + "step": 4450, + "train_speed(iter/s)": 0.450827 + }, + { + "acc": 0.62125816, + "epoch": 0.11301369863013698, + "grad_norm": 5.1875, + "learning_rate": 9.998842509899456e-06, + "loss": 1.79626656, + "memory(GiB)": 58.14, + "step": 4455, + "train_speed(iter/s)": 0.451094 + }, + { + "acc": 0.63397903, + "epoch": 0.11314053779807204, + "grad_norm": 6.03125, + "learning_rate": 9.998819837674655e-06, + "loss": 1.77970695, + "memory(GiB)": 58.14, + "step": 4460, + "train_speed(iter/s)": 0.451358 + }, + { + "acc": 0.64185648, + "epoch": 0.1132673769660071, + "grad_norm": 7.5, + "learning_rate": 9.99879694557944e-06, + "loss": 1.74106216, + "memory(GiB)": 58.14, + "step": 4465, + "train_speed(iter/s)": 0.451621 + }, + { + "acc": 0.64832673, + "epoch": 0.11339421613394216, + "grad_norm": 5.25, + "learning_rate": 9.998773833614816e-06, + "loss": 1.72003975, + "memory(GiB)": 58.14, + "step": 4470, + "train_speed(iter/s)": 0.451885 + }, + { + "acc": 0.64814358, + "epoch": 0.11352105530187721, + "grad_norm": 5.21875, + "learning_rate": 9.998750501781803e-06, + "loss": 1.65770454, + "memory(GiB)": 58.14, + "step": 4475, + "train_speed(iter/s)": 0.452148 + }, + { + "acc": 0.64199409, + "epoch": 0.11364789446981227, + "grad_norm": 5.71875, + "learning_rate": 9.998726950081425e-06, + "loss": 1.7796257, + "memory(GiB)": 58.14, + "step": 4480, + "train_speed(iter/s)": 0.45241 + }, + { + "acc": 0.63876882, + "epoch": 0.11377473363774733, + "grad_norm": 6.15625, + "learning_rate": 9.998703178514717e-06, + "loss": 1.81871719, + "memory(GiB)": 58.14, + "step": 4485, + "train_speed(iter/s)": 0.452677 + }, + { + "acc": 0.61971855, + "epoch": 0.11390157280568239, + "grad_norm": 6.3125, + "learning_rate": 9.998679187082724e-06, + "loss": 1.74147835, + "memory(GiB)": 58.14, + "step": 4490, + "train_speed(iter/s)": 0.45294 + }, + { + "acc": 0.62179823, + "epoch": 0.11402841197361745, + "grad_norm": 4.46875, + "learning_rate": 9.998654975786506e-06, + "loss": 1.76019974, + "memory(GiB)": 58.14, + "step": 4495, + "train_speed(iter/s)": 0.453205 + }, + { + "acc": 0.63448839, + "epoch": 0.1141552511415525, + "grad_norm": 5.78125, + "learning_rate": 9.998630544627123e-06, + "loss": 1.72916336, + "memory(GiB)": 58.14, + "step": 4500, + "train_speed(iter/s)": 0.453468 + }, + { + "epoch": 0.1141552511415525, + "eval_acc": 0.6224657712799179, + "eval_loss": 1.734817385673523, + "eval_runtime": 114.6258, + "eval_samples_per_second": 55.572, + "eval_steps_per_second": 27.786, + "step": 4500 + }, + { + "acc": 0.64289494, + "epoch": 0.11428209030948756, + "grad_norm": 4.71875, + "learning_rate": 9.998605893605653e-06, + "loss": 1.70340538, + "memory(GiB)": 58.14, + "step": 4505, + "train_speed(iter/s)": 0.448352 + }, + { + "acc": 0.63723078, + "epoch": 0.11440892947742262, + "grad_norm": 5.125, + "learning_rate": 9.998581022723178e-06, + "loss": 1.73938599, + "memory(GiB)": 58.14, + "step": 4510, + "train_speed(iter/s)": 0.448613 + }, + { + "acc": 0.61551995, + "epoch": 0.11453576864535768, + "grad_norm": 5.3125, + "learning_rate": 9.998555931980792e-06, + "loss": 1.81156197, + "memory(GiB)": 58.14, + "step": 4515, + "train_speed(iter/s)": 0.448875 + }, + { + "acc": 0.61945362, + "epoch": 0.11466260781329274, + "grad_norm": 4.71875, + "learning_rate": 9.998530621379599e-06, + "loss": 1.75532379, + "memory(GiB)": 58.14, + "step": 4520, + "train_speed(iter/s)": 0.449139 + }, + { + "acc": 0.64460568, + "epoch": 0.1147894469812278, + "grad_norm": 4.375, + "learning_rate": 9.998505090920713e-06, + "loss": 1.72848473, + "memory(GiB)": 58.14, + "step": 4525, + "train_speed(iter/s)": 0.449405 + }, + { + "acc": 0.63762555, + "epoch": 0.11491628614916286, + "grad_norm": 5.375, + "learning_rate": 9.998479340605257e-06, + "loss": 1.73102989, + "memory(GiB)": 58.14, + "step": 4530, + "train_speed(iter/s)": 0.449667 + }, + { + "acc": 0.63705587, + "epoch": 0.11504312531709791, + "grad_norm": 4.5, + "learning_rate": 9.99845337043436e-06, + "loss": 1.7236248, + "memory(GiB)": 58.14, + "step": 4535, + "train_speed(iter/s)": 0.449931 + }, + { + "acc": 0.61686687, + "epoch": 0.11516996448503297, + "grad_norm": 6.375, + "learning_rate": 9.998427180409171e-06, + "loss": 1.83120613, + "memory(GiB)": 58.14, + "step": 4540, + "train_speed(iter/s)": 0.450192 + }, + { + "acc": 0.62705369, + "epoch": 0.11529680365296803, + "grad_norm": 6.8125, + "learning_rate": 9.998400770530836e-06, + "loss": 1.78143158, + "memory(GiB)": 58.14, + "step": 4545, + "train_speed(iter/s)": 0.450456 + }, + { + "acc": 0.62718897, + "epoch": 0.11542364282090309, + "grad_norm": 5.0, + "learning_rate": 9.99837414080052e-06, + "loss": 1.74348793, + "memory(GiB)": 58.14, + "step": 4550, + "train_speed(iter/s)": 0.45072 + }, + { + "acc": 0.62926569, + "epoch": 0.11555048198883815, + "grad_norm": 5.0625, + "learning_rate": 9.998347291219393e-06, + "loss": 1.76136131, + "memory(GiB)": 58.14, + "step": 4555, + "train_speed(iter/s)": 0.450983 + }, + { + "acc": 0.63724442, + "epoch": 0.1156773211567732, + "grad_norm": 5.3125, + "learning_rate": 9.998320221788635e-06, + "loss": 1.76352921, + "memory(GiB)": 58.14, + "step": 4560, + "train_speed(iter/s)": 0.451246 + }, + { + "acc": 0.60729675, + "epoch": 0.11580416032470826, + "grad_norm": 6.65625, + "learning_rate": 9.998292932509438e-06, + "loss": 1.9108429, + "memory(GiB)": 58.14, + "step": 4565, + "train_speed(iter/s)": 0.451507 + }, + { + "acc": 0.62277074, + "epoch": 0.11593099949264332, + "grad_norm": 4.90625, + "learning_rate": 9.998265423383003e-06, + "loss": 1.80425949, + "memory(GiB)": 58.14, + "step": 4570, + "train_speed(iter/s)": 0.451768 + }, + { + "acc": 0.63655186, + "epoch": 0.11605783866057838, + "grad_norm": 6.28125, + "learning_rate": 9.998237694410537e-06, + "loss": 1.76361046, + "memory(GiB)": 58.14, + "step": 4575, + "train_speed(iter/s)": 0.452029 + }, + { + "acc": 0.64174104, + "epoch": 0.11618467782851344, + "grad_norm": 6.3125, + "learning_rate": 9.998209745593264e-06, + "loss": 1.7118351, + "memory(GiB)": 58.14, + "step": 4580, + "train_speed(iter/s)": 0.452292 + }, + { + "acc": 0.6311769, + "epoch": 0.1163115169964485, + "grad_norm": 5.15625, + "learning_rate": 9.99818157693241e-06, + "loss": 1.77057285, + "memory(GiB)": 58.14, + "step": 4585, + "train_speed(iter/s)": 0.452551 + }, + { + "acc": 0.64642916, + "epoch": 0.11643835616438356, + "grad_norm": 5.1875, + "learning_rate": 9.998153188429216e-06, + "loss": 1.62410469, + "memory(GiB)": 58.14, + "step": 4590, + "train_speed(iter/s)": 0.452813 + }, + { + "acc": 0.63776679, + "epoch": 0.11656519533231861, + "grad_norm": 5.03125, + "learning_rate": 9.99812458008493e-06, + "loss": 1.70000572, + "memory(GiB)": 58.14, + "step": 4595, + "train_speed(iter/s)": 0.453073 + }, + { + "acc": 0.62350159, + "epoch": 0.11669203450025367, + "grad_norm": 4.25, + "learning_rate": 9.998095751900806e-06, + "loss": 1.77704659, + "memory(GiB)": 58.14, + "step": 4600, + "train_speed(iter/s)": 0.453334 + }, + { + "epoch": 0.11669203450025367, + "eval_acc": 0.6230740404824878, + "eval_loss": 1.7313827276229858, + "eval_runtime": 113.2999, + "eval_samples_per_second": 56.222, + "eval_steps_per_second": 28.111, + "step": 4600 + }, + { + "acc": 0.62139163, + "epoch": 0.11681887366818873, + "grad_norm": 4.21875, + "learning_rate": 9.99806670387812e-06, + "loss": 1.75947609, + "memory(GiB)": 58.14, + "step": 4605, + "train_speed(iter/s)": 0.44839 + }, + { + "acc": 0.62973518, + "epoch": 0.11694571283612379, + "grad_norm": 5.625, + "learning_rate": 9.998037436018144e-06, + "loss": 1.76128922, + "memory(GiB)": 58.14, + "step": 4610, + "train_speed(iter/s)": 0.44865 + }, + { + "acc": 0.63088779, + "epoch": 0.11707255200405885, + "grad_norm": 5.40625, + "learning_rate": 9.998007948322168e-06, + "loss": 1.69484825, + "memory(GiB)": 58.14, + "step": 4615, + "train_speed(iter/s)": 0.448905 + }, + { + "acc": 0.62563944, + "epoch": 0.1171993911719939, + "grad_norm": 5.09375, + "learning_rate": 9.997978240791487e-06, + "loss": 1.77591267, + "memory(GiB)": 58.14, + "step": 4620, + "train_speed(iter/s)": 0.449163 + }, + { + "acc": 0.63080463, + "epoch": 0.11732623033992896, + "grad_norm": 6.0625, + "learning_rate": 9.99794831342741e-06, + "loss": 1.77028713, + "memory(GiB)": 58.14, + "step": 4625, + "train_speed(iter/s)": 0.449417 + }, + { + "acc": 0.63224735, + "epoch": 0.11745306950786402, + "grad_norm": 5.25, + "learning_rate": 9.99791816623125e-06, + "loss": 1.78392754, + "memory(GiB)": 58.14, + "step": 4630, + "train_speed(iter/s)": 0.449675 + }, + { + "acc": 0.64080734, + "epoch": 0.11757990867579908, + "grad_norm": 4.625, + "learning_rate": 9.997887799204335e-06, + "loss": 1.69095192, + "memory(GiB)": 58.14, + "step": 4635, + "train_speed(iter/s)": 0.44993 + }, + { + "acc": 0.62064586, + "epoch": 0.11770674784373414, + "grad_norm": 4.625, + "learning_rate": 9.997857212348e-06, + "loss": 1.77523727, + "memory(GiB)": 58.14, + "step": 4640, + "train_speed(iter/s)": 0.450185 + }, + { + "acc": 0.63761692, + "epoch": 0.1178335870116692, + "grad_norm": 4.46875, + "learning_rate": 9.997826405663593e-06, + "loss": 1.74796543, + "memory(GiB)": 58.14, + "step": 4645, + "train_speed(iter/s)": 0.450441 + }, + { + "acc": 0.65436001, + "epoch": 0.11796042617960426, + "grad_norm": 5.0625, + "learning_rate": 9.997795379152468e-06, + "loss": 1.58777275, + "memory(GiB)": 58.14, + "step": 4650, + "train_speed(iter/s)": 0.450695 + }, + { + "acc": 0.60329132, + "epoch": 0.11808726534753931, + "grad_norm": 5.375, + "learning_rate": 9.997764132815985e-06, + "loss": 1.84420433, + "memory(GiB)": 58.14, + "step": 4655, + "train_speed(iter/s)": 0.450951 + }, + { + "acc": 0.62300425, + "epoch": 0.11821410451547437, + "grad_norm": 5.5625, + "learning_rate": 9.997732666655524e-06, + "loss": 1.81390247, + "memory(GiB)": 58.14, + "step": 4660, + "train_speed(iter/s)": 0.451207 + }, + { + "acc": 0.61180248, + "epoch": 0.11834094368340943, + "grad_norm": 5.09375, + "learning_rate": 9.997700980672469e-06, + "loss": 1.80420074, + "memory(GiB)": 58.14, + "step": 4665, + "train_speed(iter/s)": 0.451463 + }, + { + "acc": 0.63818851, + "epoch": 0.11846778285134449, + "grad_norm": 5.21875, + "learning_rate": 9.997669074868208e-06, + "loss": 1.68481178, + "memory(GiB)": 58.14, + "step": 4670, + "train_speed(iter/s)": 0.451718 + }, + { + "acc": 0.6398519, + "epoch": 0.11859462201927955, + "grad_norm": 5.40625, + "learning_rate": 9.997636949244151e-06, + "loss": 1.74873466, + "memory(GiB)": 58.14, + "step": 4675, + "train_speed(iter/s)": 0.451972 + }, + { + "acc": 0.62650094, + "epoch": 0.1187214611872146, + "grad_norm": 5.03125, + "learning_rate": 9.997604603801707e-06, + "loss": 1.78236198, + "memory(GiB)": 58.14, + "step": 4680, + "train_speed(iter/s)": 0.452228 + }, + { + "acc": 0.62283201, + "epoch": 0.11884830035514966, + "grad_norm": 5.09375, + "learning_rate": 9.9975720385423e-06, + "loss": 1.84657478, + "memory(GiB)": 58.14, + "step": 4685, + "train_speed(iter/s)": 0.452481 + }, + { + "acc": 0.64016471, + "epoch": 0.11897513952308472, + "grad_norm": 5.59375, + "learning_rate": 9.997539253467361e-06, + "loss": 1.76661339, + "memory(GiB)": 58.14, + "step": 4690, + "train_speed(iter/s)": 0.452735 + }, + { + "acc": 0.629918, + "epoch": 0.11910197869101978, + "grad_norm": 5.21875, + "learning_rate": 9.997506248578334e-06, + "loss": 1.73544044, + "memory(GiB)": 58.14, + "step": 4695, + "train_speed(iter/s)": 0.452988 + }, + { + "acc": 0.64385133, + "epoch": 0.11922881785895484, + "grad_norm": 5.6875, + "learning_rate": 9.997473023876671e-06, + "loss": 1.78379498, + "memory(GiB)": 58.14, + "step": 4700, + "train_speed(iter/s)": 0.453241 + }, + { + "epoch": 0.11922881785895484, + "eval_acc": 0.6235164560632032, + "eval_loss": 1.727088212966919, + "eval_runtime": 112.7884, + "eval_samples_per_second": 56.477, + "eval_steps_per_second": 28.239, + "step": 4700 + }, + { + "acc": 0.63448782, + "epoch": 0.1193556570268899, + "grad_norm": 4.75, + "learning_rate": 9.997439579363831e-06, + "loss": 1.78977203, + "memory(GiB)": 58.14, + "step": 4705, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.62294664, + "epoch": 0.11948249619482496, + "grad_norm": 5.0, + "learning_rate": 9.997405915041288e-06, + "loss": 1.77439079, + "memory(GiB)": 58.14, + "step": 4710, + "train_speed(iter/s)": 0.448674 + }, + { + "acc": 0.64314108, + "epoch": 0.11960933536276001, + "grad_norm": 5.71875, + "learning_rate": 9.99737203091052e-06, + "loss": 1.70816956, + "memory(GiB)": 58.14, + "step": 4715, + "train_speed(iter/s)": 0.448921 + }, + { + "acc": 0.64581261, + "epoch": 0.11973617453069507, + "grad_norm": 5.03125, + "learning_rate": 9.997337926973018e-06, + "loss": 1.71403198, + "memory(GiB)": 58.14, + "step": 4720, + "train_speed(iter/s)": 0.449169 + }, + { + "acc": 0.63275752, + "epoch": 0.11986301369863013, + "grad_norm": 5.46875, + "learning_rate": 9.997303603230282e-06, + "loss": 1.7627388, + "memory(GiB)": 58.14, + "step": 4725, + "train_speed(iter/s)": 0.449415 + }, + { + "acc": 0.64267807, + "epoch": 0.11998985286656519, + "grad_norm": 5.25, + "learning_rate": 9.997269059683822e-06, + "loss": 1.70204964, + "memory(GiB)": 58.14, + "step": 4730, + "train_speed(iter/s)": 0.449662 + }, + { + "acc": 0.61263685, + "epoch": 0.12011669203450025, + "grad_norm": 5.875, + "learning_rate": 9.997234296335159e-06, + "loss": 1.78822365, + "memory(GiB)": 58.14, + "step": 4735, + "train_speed(iter/s)": 0.449912 + }, + { + "acc": 0.63470836, + "epoch": 0.1202435312024353, + "grad_norm": 4.875, + "learning_rate": 9.997199313185821e-06, + "loss": 1.74361115, + "memory(GiB)": 58.14, + "step": 4740, + "train_speed(iter/s)": 0.450159 + }, + { + "acc": 0.64303627, + "epoch": 0.12037037037037036, + "grad_norm": 5.15625, + "learning_rate": 9.997164110237345e-06, + "loss": 1.64466343, + "memory(GiB)": 58.14, + "step": 4745, + "train_speed(iter/s)": 0.450407 + }, + { + "acc": 0.63831496, + "epoch": 0.12049720953830542, + "grad_norm": 6.28125, + "learning_rate": 9.99712868749128e-06, + "loss": 1.7146286, + "memory(GiB)": 58.14, + "step": 4750, + "train_speed(iter/s)": 0.450655 + }, + { + "acc": 0.65628538, + "epoch": 0.12062404870624048, + "grad_norm": 5.6875, + "learning_rate": 9.997093044949186e-06, + "loss": 1.72340527, + "memory(GiB)": 58.14, + "step": 4755, + "train_speed(iter/s)": 0.450903 + }, + { + "acc": 0.63663931, + "epoch": 0.12075088787417554, + "grad_norm": 5.8125, + "learning_rate": 9.997057182612631e-06, + "loss": 1.69316998, + "memory(GiB)": 58.14, + "step": 4760, + "train_speed(iter/s)": 0.451147 + }, + { + "acc": 0.62728376, + "epoch": 0.1208777270421106, + "grad_norm": 5.8125, + "learning_rate": 9.997021100483188e-06, + "loss": 1.73591309, + "memory(GiB)": 58.14, + "step": 4765, + "train_speed(iter/s)": 0.451394 + }, + { + "acc": 0.64358292, + "epoch": 0.12100456621004566, + "grad_norm": 6.125, + "learning_rate": 9.996984798562448e-06, + "loss": 1.66552658, + "memory(GiB)": 58.14, + "step": 4770, + "train_speed(iter/s)": 0.45164 + }, + { + "acc": 0.6273417, + "epoch": 0.12113140537798071, + "grad_norm": 5.84375, + "learning_rate": 9.996948276852008e-06, + "loss": 1.73432961, + "memory(GiB)": 58.14, + "step": 4775, + "train_speed(iter/s)": 0.451889 + }, + { + "acc": 0.63287511, + "epoch": 0.12125824454591577, + "grad_norm": 5.21875, + "learning_rate": 9.99691153535347e-06, + "loss": 1.81104622, + "memory(GiB)": 58.14, + "step": 4780, + "train_speed(iter/s)": 0.452135 + }, + { + "acc": 0.64076881, + "epoch": 0.12138508371385083, + "grad_norm": 5.0625, + "learning_rate": 9.996874574068457e-06, + "loss": 1.74649353, + "memory(GiB)": 58.14, + "step": 4785, + "train_speed(iter/s)": 0.452382 + }, + { + "acc": 0.63243017, + "epoch": 0.12151192288178589, + "grad_norm": 4.25, + "learning_rate": 9.996837392998586e-06, + "loss": 1.75946579, + "memory(GiB)": 58.14, + "step": 4790, + "train_speed(iter/s)": 0.45263 + }, + { + "acc": 0.63010359, + "epoch": 0.12163876204972095, + "grad_norm": 5.90625, + "learning_rate": 9.996799992145501e-06, + "loss": 1.75592041, + "memory(GiB)": 58.14, + "step": 4795, + "train_speed(iter/s)": 0.452875 + }, + { + "acc": 0.62746639, + "epoch": 0.121765601217656, + "grad_norm": 5.0625, + "learning_rate": 9.996762371510843e-06, + "loss": 1.77874947, + "memory(GiB)": 58.14, + "step": 4800, + "train_speed(iter/s)": 0.453123 + }, + { + "epoch": 0.121765601217656, + "eval_acc": 0.6240428428731195, + "eval_loss": 1.723838210105896, + "eval_runtime": 113.0662, + "eval_samples_per_second": 56.339, + "eval_steps_per_second": 28.169, + "step": 4800 + }, + { + "acc": 0.63768234, + "epoch": 0.12189244038559106, + "grad_norm": 5.15625, + "learning_rate": 9.996724531096264e-06, + "loss": 1.74425735, + "memory(GiB)": 58.14, + "step": 4805, + "train_speed(iter/s)": 0.448399 + }, + { + "acc": 0.63574476, + "epoch": 0.12201927955352612, + "grad_norm": 5.71875, + "learning_rate": 9.996686470903434e-06, + "loss": 1.74980125, + "memory(GiB)": 58.14, + "step": 4810, + "train_speed(iter/s)": 0.448644 + }, + { + "acc": 0.631991, + "epoch": 0.12214611872146118, + "grad_norm": 4.9375, + "learning_rate": 9.996648190934025e-06, + "loss": 1.70170517, + "memory(GiB)": 58.14, + "step": 4815, + "train_speed(iter/s)": 0.448889 + }, + { + "acc": 0.63558216, + "epoch": 0.12227295788939624, + "grad_norm": 5.875, + "learning_rate": 9.996609691189718e-06, + "loss": 1.73584061, + "memory(GiB)": 58.14, + "step": 4820, + "train_speed(iter/s)": 0.449134 + }, + { + "acc": 0.63971071, + "epoch": 0.1223997970573313, + "grad_norm": 5.0, + "learning_rate": 9.996570971672209e-06, + "loss": 1.72283211, + "memory(GiB)": 58.14, + "step": 4825, + "train_speed(iter/s)": 0.449376 + }, + { + "acc": 0.61943631, + "epoch": 0.12252663622526636, + "grad_norm": 5.9375, + "learning_rate": 9.996532032383202e-06, + "loss": 1.81644154, + "memory(GiB)": 58.14, + "step": 4830, + "train_speed(iter/s)": 0.449617 + }, + { + "acc": 0.63256316, + "epoch": 0.12265347539320141, + "grad_norm": 5.6875, + "learning_rate": 9.996492873324406e-06, + "loss": 1.80444565, + "memory(GiB)": 58.14, + "step": 4835, + "train_speed(iter/s)": 0.449864 + }, + { + "acc": 0.63840075, + "epoch": 0.12278031456113647, + "grad_norm": 5.4375, + "learning_rate": 9.996453494497546e-06, + "loss": 1.75831375, + "memory(GiB)": 58.14, + "step": 4840, + "train_speed(iter/s)": 0.450107 + }, + { + "acc": 0.63326774, + "epoch": 0.12290715372907153, + "grad_norm": 6.46875, + "learning_rate": 9.996413895904355e-06, + "loss": 1.76342812, + "memory(GiB)": 58.14, + "step": 4845, + "train_speed(iter/s)": 0.450354 + }, + { + "acc": 0.63627186, + "epoch": 0.12303399289700659, + "grad_norm": 5.5625, + "learning_rate": 9.996374077546573e-06, + "loss": 1.71198921, + "memory(GiB)": 58.14, + "step": 4850, + "train_speed(iter/s)": 0.450598 + }, + { + "acc": 0.63379359, + "epoch": 0.12316083206494165, + "grad_norm": 5.53125, + "learning_rate": 9.996334039425952e-06, + "loss": 1.71940689, + "memory(GiB)": 58.14, + "step": 4855, + "train_speed(iter/s)": 0.450842 + }, + { + "acc": 0.63757434, + "epoch": 0.1232876712328767, + "grad_norm": 6.90625, + "learning_rate": 9.996293781544255e-06, + "loss": 1.69967937, + "memory(GiB)": 58.14, + "step": 4860, + "train_speed(iter/s)": 0.451087 + }, + { + "acc": 0.63274517, + "epoch": 0.12341451040081176, + "grad_norm": 5.0, + "learning_rate": 9.996253303903247e-06, + "loss": 1.7288847, + "memory(GiB)": 58.14, + "step": 4865, + "train_speed(iter/s)": 0.451332 + }, + { + "acc": 0.63049917, + "epoch": 0.12354134956874682, + "grad_norm": 4.5625, + "learning_rate": 9.996212606504713e-06, + "loss": 1.72965984, + "memory(GiB)": 58.14, + "step": 4870, + "train_speed(iter/s)": 0.451576 + }, + { + "acc": 0.65030603, + "epoch": 0.12366818873668188, + "grad_norm": 5.78125, + "learning_rate": 9.996171689350444e-06, + "loss": 1.75297108, + "memory(GiB)": 58.14, + "step": 4875, + "train_speed(iter/s)": 0.451819 + }, + { + "acc": 0.6482976, + "epoch": 0.12379502790461694, + "grad_norm": 4.8125, + "learning_rate": 9.996130552442237e-06, + "loss": 1.67681484, + "memory(GiB)": 58.14, + "step": 4880, + "train_speed(iter/s)": 0.452064 + }, + { + "acc": 0.63315368, + "epoch": 0.123921867072552, + "grad_norm": 4.96875, + "learning_rate": 9.996089195781902e-06, + "loss": 1.65789948, + "memory(GiB)": 58.14, + "step": 4885, + "train_speed(iter/s)": 0.452305 + }, + { + "acc": 0.63042555, + "epoch": 0.12404870624048706, + "grad_norm": 5.5, + "learning_rate": 9.996047619371256e-06, + "loss": 1.76520863, + "memory(GiB)": 58.14, + "step": 4890, + "train_speed(iter/s)": 0.452548 + }, + { + "acc": 0.64517727, + "epoch": 0.12417554540842211, + "grad_norm": 6.40625, + "learning_rate": 9.996005823212132e-06, + "loss": 1.73255177, + "memory(GiB)": 58.14, + "step": 4895, + "train_speed(iter/s)": 0.45279 + }, + { + "acc": 0.63175397, + "epoch": 0.12430238457635717, + "grad_norm": 6.75, + "learning_rate": 9.995963807306368e-06, + "loss": 1.74234314, + "memory(GiB)": 58.14, + "step": 4900, + "train_speed(iter/s)": 0.453034 + }, + { + "epoch": 0.12430238457635717, + "eval_acc": 0.6244764853403362, + "eval_loss": 1.720935583114624, + "eval_runtime": 113.2657, + "eval_samples_per_second": 56.239, + "eval_steps_per_second": 28.12, + "step": 4900 + }, + { + "acc": 0.63740139, + "epoch": 0.12442922374429223, + "grad_norm": 4.71875, + "learning_rate": 9.995921571655808e-06, + "loss": 1.71653137, + "memory(GiB)": 58.14, + "step": 4905, + "train_speed(iter/s)": 0.448397 + }, + { + "acc": 0.62355433, + "epoch": 0.12455606291222729, + "grad_norm": 6.3125, + "learning_rate": 9.995879116262312e-06, + "loss": 1.77159882, + "memory(GiB)": 58.14, + "step": 4910, + "train_speed(iter/s)": 0.448638 + }, + { + "acc": 0.64844608, + "epoch": 0.12468290208016235, + "grad_norm": 5.09375, + "learning_rate": 9.995836441127749e-06, + "loss": 1.65455017, + "memory(GiB)": 58.14, + "step": 4915, + "train_speed(iter/s)": 0.448877 + }, + { + "acc": 0.64269943, + "epoch": 0.1248097412480974, + "grad_norm": 5.21875, + "learning_rate": 9.995793546253993e-06, + "loss": 1.63613682, + "memory(GiB)": 58.14, + "step": 4920, + "train_speed(iter/s)": 0.449117 + }, + { + "acc": 0.64152327, + "epoch": 0.12493658041603246, + "grad_norm": 5.75, + "learning_rate": 9.995750431642933e-06, + "loss": 1.74731522, + "memory(GiB)": 58.14, + "step": 4925, + "train_speed(iter/s)": 0.449357 + }, + { + "acc": 0.63020616, + "epoch": 0.12506341958396752, + "grad_norm": 6.0625, + "learning_rate": 9.995707097296465e-06, + "loss": 1.76956387, + "memory(GiB)": 58.14, + "step": 4930, + "train_speed(iter/s)": 0.449597 + }, + { + "acc": 0.6298851, + "epoch": 0.1251902587519026, + "grad_norm": 5.1875, + "learning_rate": 9.995663543216493e-06, + "loss": 1.75992546, + "memory(GiB)": 58.14, + "step": 4935, + "train_speed(iter/s)": 0.449835 + }, + { + "acc": 0.64089832, + "epoch": 0.12531709791983764, + "grad_norm": 5.15625, + "learning_rate": 9.995619769404936e-06, + "loss": 1.72555828, + "memory(GiB)": 58.14, + "step": 4940, + "train_speed(iter/s)": 0.450073 + }, + { + "acc": 0.61417646, + "epoch": 0.1254439370877727, + "grad_norm": 6.28125, + "learning_rate": 9.995575775863717e-06, + "loss": 1.82039833, + "memory(GiB)": 58.14, + "step": 4945, + "train_speed(iter/s)": 0.450309 + }, + { + "acc": 0.65074611, + "epoch": 0.12557077625570776, + "grad_norm": 5.65625, + "learning_rate": 9.995531562594773e-06, + "loss": 1.72393265, + "memory(GiB)": 58.14, + "step": 4950, + "train_speed(iter/s)": 0.450546 + }, + { + "acc": 0.64919972, + "epoch": 0.12569761542364283, + "grad_norm": 5.8125, + "learning_rate": 9.995487129600046e-06, + "loss": 1.67064629, + "memory(GiB)": 58.14, + "step": 4955, + "train_speed(iter/s)": 0.450784 + }, + { + "acc": 0.6427999, + "epoch": 0.12582445459157787, + "grad_norm": 4.875, + "learning_rate": 9.995442476881491e-06, + "loss": 1.7232132, + "memory(GiB)": 58.14, + "step": 4960, + "train_speed(iter/s)": 0.451019 + }, + { + "acc": 0.63472452, + "epoch": 0.12595129375951294, + "grad_norm": 4.625, + "learning_rate": 9.995397604441076e-06, + "loss": 1.71817131, + "memory(GiB)": 58.14, + "step": 4965, + "train_speed(iter/s)": 0.45126 + }, + { + "acc": 0.64442329, + "epoch": 0.126078132927448, + "grad_norm": 5.34375, + "learning_rate": 9.995352512280767e-06, + "loss": 1.75287018, + "memory(GiB)": 58.14, + "step": 4970, + "train_speed(iter/s)": 0.451498 + }, + { + "acc": 0.64105768, + "epoch": 0.12620497209538306, + "grad_norm": 6.1875, + "learning_rate": 9.995307200402555e-06, + "loss": 1.74288979, + "memory(GiB)": 58.14, + "step": 4975, + "train_speed(iter/s)": 0.45174 + }, + { + "acc": 0.63702526, + "epoch": 0.1263318112633181, + "grad_norm": 5.8125, + "learning_rate": 9.995261668808429e-06, + "loss": 1.6472868, + "memory(GiB)": 58.14, + "step": 4980, + "train_speed(iter/s)": 0.45198 + }, + { + "acc": 0.65537586, + "epoch": 0.12645865043125318, + "grad_norm": 5.3125, + "learning_rate": 9.995215917500395e-06, + "loss": 1.65482178, + "memory(GiB)": 58.14, + "step": 4985, + "train_speed(iter/s)": 0.45222 + }, + { + "acc": 0.61985517, + "epoch": 0.12658548959918822, + "grad_norm": 4.84375, + "learning_rate": 9.995169946480459e-06, + "loss": 1.72397575, + "memory(GiB)": 58.14, + "step": 4990, + "train_speed(iter/s)": 0.452459 + }, + { + "acc": 0.62642651, + "epoch": 0.1267123287671233, + "grad_norm": 5.875, + "learning_rate": 9.99512375575065e-06, + "loss": 1.73385048, + "memory(GiB)": 58.14, + "step": 4995, + "train_speed(iter/s)": 0.452697 + }, + { + "acc": 0.6396203, + "epoch": 0.12683916793505834, + "grad_norm": 5.90625, + "learning_rate": 9.995077345312994e-06, + "loss": 1.75152912, + "memory(GiB)": 58.14, + "step": 5000, + "train_speed(iter/s)": 0.452936 + }, + { + "epoch": 0.12683916793505834, + "eval_acc": 0.6249945168040634, + "eval_loss": 1.71718168258667, + "eval_runtime": 112.8946, + "eval_samples_per_second": 56.424, + "eval_steps_per_second": 28.212, + "step": 5000 + }, + { + "acc": 0.62750554, + "epoch": 0.1269660071029934, + "grad_norm": 5.59375, + "learning_rate": 9.995030715169535e-06, + "loss": 1.77503719, + "memory(GiB)": 58.14, + "step": 5005, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.62299132, + "epoch": 0.12709284627092846, + "grad_norm": 4.8125, + "learning_rate": 9.994983865322327e-06, + "loss": 1.785462, + "memory(GiB)": 58.14, + "step": 5010, + "train_speed(iter/s)": 0.44865 + }, + { + "acc": 0.63477411, + "epoch": 0.12721968543886353, + "grad_norm": 6.5, + "learning_rate": 9.994936795773424e-06, + "loss": 1.72480125, + "memory(GiB)": 58.14, + "step": 5015, + "train_speed(iter/s)": 0.448886 + }, + { + "acc": 0.6355526, + "epoch": 0.12734652460679857, + "grad_norm": 4.5, + "learning_rate": 9.994889506524903e-06, + "loss": 1.7173584, + "memory(GiB)": 58.14, + "step": 5020, + "train_speed(iter/s)": 0.449124 + }, + { + "acc": 0.63450956, + "epoch": 0.12747336377473364, + "grad_norm": 5.90625, + "learning_rate": 9.994841997578839e-06, + "loss": 1.77025185, + "memory(GiB)": 58.14, + "step": 5025, + "train_speed(iter/s)": 0.449362 + }, + { + "acc": 0.64327793, + "epoch": 0.1276002029426687, + "grad_norm": 4.28125, + "learning_rate": 9.994794268937325e-06, + "loss": 1.73540573, + "memory(GiB)": 58.14, + "step": 5030, + "train_speed(iter/s)": 0.4496 + }, + { + "acc": 0.62260504, + "epoch": 0.12772704211060376, + "grad_norm": 6.90625, + "learning_rate": 9.994746320602457e-06, + "loss": 1.80532665, + "memory(GiB)": 58.14, + "step": 5035, + "train_speed(iter/s)": 0.44984 + }, + { + "acc": 0.63432403, + "epoch": 0.1278538812785388, + "grad_norm": 4.59375, + "learning_rate": 9.994698152576347e-06, + "loss": 1.74718781, + "memory(GiB)": 58.14, + "step": 5040, + "train_speed(iter/s)": 0.450081 + }, + { + "acc": 0.63752632, + "epoch": 0.12798072044647388, + "grad_norm": 5.6875, + "learning_rate": 9.994649764861114e-06, + "loss": 1.72059212, + "memory(GiB)": 58.14, + "step": 5045, + "train_speed(iter/s)": 0.45032 + }, + { + "acc": 0.62703476, + "epoch": 0.12810755961440892, + "grad_norm": 5.28125, + "learning_rate": 9.994601157458882e-06, + "loss": 1.74051857, + "memory(GiB)": 58.14, + "step": 5050, + "train_speed(iter/s)": 0.45056 + }, + { + "acc": 0.61078701, + "epoch": 0.128234398782344, + "grad_norm": 4.78125, + "learning_rate": 9.994552330371792e-06, + "loss": 1.78131847, + "memory(GiB)": 58.14, + "step": 5055, + "train_speed(iter/s)": 0.450798 + }, + { + "acc": 0.62378693, + "epoch": 0.12836123795027904, + "grad_norm": 4.5, + "learning_rate": 9.994503283601993e-06, + "loss": 1.7459137, + "memory(GiB)": 58.14, + "step": 5060, + "train_speed(iter/s)": 0.451038 + }, + { + "acc": 0.62276478, + "epoch": 0.1284880771182141, + "grad_norm": 4.46875, + "learning_rate": 9.99445401715164e-06, + "loss": 1.80179634, + "memory(GiB)": 58.14, + "step": 5065, + "train_speed(iter/s)": 0.451277 + }, + { + "acc": 0.63484154, + "epoch": 0.12861491628614916, + "grad_norm": 4.78125, + "learning_rate": 9.994404531022901e-06, + "loss": 1.68935127, + "memory(GiB)": 58.14, + "step": 5070, + "train_speed(iter/s)": 0.451512 + }, + { + "acc": 0.64244261, + "epoch": 0.12874175545408423, + "grad_norm": 6.75, + "learning_rate": 9.994354825217954e-06, + "loss": 1.68698025, + "memory(GiB)": 58.14, + "step": 5075, + "train_speed(iter/s)": 0.451749 + }, + { + "acc": 0.62662172, + "epoch": 0.12886859462201927, + "grad_norm": 5.625, + "learning_rate": 9.99430489973898e-06, + "loss": 1.75492172, + "memory(GiB)": 58.14, + "step": 5080, + "train_speed(iter/s)": 0.451985 + }, + { + "acc": 0.62019196, + "epoch": 0.12899543378995434, + "grad_norm": 7.71875, + "learning_rate": 9.994254754588182e-06, + "loss": 1.81921692, + "memory(GiB)": 58.14, + "step": 5085, + "train_speed(iter/s)": 0.452223 + }, + { + "acc": 0.63716955, + "epoch": 0.1291222729578894, + "grad_norm": 4.875, + "learning_rate": 9.99420438976776e-06, + "loss": 1.68088188, + "memory(GiB)": 58.14, + "step": 5090, + "train_speed(iter/s)": 0.45246 + }, + { + "acc": 0.62650213, + "epoch": 0.12924911212582446, + "grad_norm": 5.875, + "learning_rate": 9.994153805279932e-06, + "loss": 1.74250259, + "memory(GiB)": 58.14, + "step": 5095, + "train_speed(iter/s)": 0.452699 + }, + { + "acc": 0.63830347, + "epoch": 0.1293759512937595, + "grad_norm": 5.46875, + "learning_rate": 9.994103001126923e-06, + "loss": 1.67629318, + "memory(GiB)": 58.14, + "step": 5100, + "train_speed(iter/s)": 0.452934 + }, + { + "epoch": 0.1293759512937595, + "eval_acc": 0.6254711893041542, + "eval_loss": 1.714188814163208, + "eval_runtime": 112.5712, + "eval_samples_per_second": 56.586, + "eval_steps_per_second": 28.293, + "step": 5100 + }, + { + "acc": 0.63393269, + "epoch": 0.12950279046169458, + "grad_norm": 5.96875, + "learning_rate": 9.994051977310966e-06, + "loss": 1.72539673, + "memory(GiB)": 58.14, + "step": 5105, + "train_speed(iter/s)": 0.448509 + }, + { + "acc": 0.61500268, + "epoch": 0.12962962962962962, + "grad_norm": 5.15625, + "learning_rate": 9.994000733834307e-06, + "loss": 1.84072914, + "memory(GiB)": 58.14, + "step": 5110, + "train_speed(iter/s)": 0.448741 + }, + { + "acc": 0.64216518, + "epoch": 0.1297564687975647, + "grad_norm": 4.4375, + "learning_rate": 9.993949270699197e-06, + "loss": 1.71509991, + "memory(GiB)": 58.14, + "step": 5115, + "train_speed(iter/s)": 0.448973 + }, + { + "acc": 0.65311861, + "epoch": 0.12988330796549974, + "grad_norm": 5.03125, + "learning_rate": 9.993897587907904e-06, + "loss": 1.65904999, + "memory(GiB)": 58.14, + "step": 5120, + "train_speed(iter/s)": 0.449206 + }, + { + "acc": 0.63206854, + "epoch": 0.1300101471334348, + "grad_norm": 5.4375, + "learning_rate": 9.993845685462697e-06, + "loss": 1.72628212, + "memory(GiB)": 58.14, + "step": 5125, + "train_speed(iter/s)": 0.44944 + }, + { + "acc": 0.610113, + "epoch": 0.13013698630136986, + "grad_norm": 6.53125, + "learning_rate": 9.993793563365864e-06, + "loss": 1.78667793, + "memory(GiB)": 58.14, + "step": 5130, + "train_speed(iter/s)": 0.449673 + }, + { + "acc": 0.62896633, + "epoch": 0.13026382546930493, + "grad_norm": 5.0625, + "learning_rate": 9.993741221619692e-06, + "loss": 1.71539001, + "memory(GiB)": 58.14, + "step": 5135, + "train_speed(iter/s)": 0.449906 + }, + { + "acc": 0.63802094, + "epoch": 0.13039066463723997, + "grad_norm": 5.28125, + "learning_rate": 9.993688660226486e-06, + "loss": 1.72593117, + "memory(GiB)": 58.14, + "step": 5140, + "train_speed(iter/s)": 0.450138 + }, + { + "acc": 0.65150294, + "epoch": 0.13051750380517504, + "grad_norm": 6.9375, + "learning_rate": 9.993635879188557e-06, + "loss": 1.65838318, + "memory(GiB)": 58.14, + "step": 5145, + "train_speed(iter/s)": 0.45037 + }, + { + "acc": 0.63751578, + "epoch": 0.1306443429731101, + "grad_norm": 5.625, + "learning_rate": 9.993582878508229e-06, + "loss": 1.72323799, + "memory(GiB)": 58.14, + "step": 5150, + "train_speed(iter/s)": 0.450604 + }, + { + "acc": 0.62342672, + "epoch": 0.13077118214104516, + "grad_norm": 5.0625, + "learning_rate": 9.993529658187829e-06, + "loss": 1.74900551, + "memory(GiB)": 58.14, + "step": 5155, + "train_speed(iter/s)": 0.450838 + }, + { + "acc": 0.64327831, + "epoch": 0.1308980213089802, + "grad_norm": 5.40625, + "learning_rate": 9.9934762182297e-06, + "loss": 1.73193378, + "memory(GiB)": 58.14, + "step": 5160, + "train_speed(iter/s)": 0.451073 + }, + { + "acc": 0.64361877, + "epoch": 0.13102486047691528, + "grad_norm": 5.25, + "learning_rate": 9.993422558636194e-06, + "loss": 1.69143028, + "memory(GiB)": 58.14, + "step": 5165, + "train_speed(iter/s)": 0.451307 + }, + { + "acc": 0.63712921, + "epoch": 0.13115169964485032, + "grad_norm": 4.6875, + "learning_rate": 9.99336867940967e-06, + "loss": 1.70390816, + "memory(GiB)": 58.14, + "step": 5170, + "train_speed(iter/s)": 0.451539 + }, + { + "acc": 0.64330087, + "epoch": 0.1312785388127854, + "grad_norm": 5.15625, + "learning_rate": 9.993314580552497e-06, + "loss": 1.66401463, + "memory(GiB)": 58.14, + "step": 5175, + "train_speed(iter/s)": 0.45177 + }, + { + "acc": 0.63401985, + "epoch": 0.13140537798072044, + "grad_norm": 6.875, + "learning_rate": 9.993260262067054e-06, + "loss": 1.69700222, + "memory(GiB)": 58.14, + "step": 5180, + "train_speed(iter/s)": 0.452001 + }, + { + "acc": 0.62357326, + "epoch": 0.1315322171486555, + "grad_norm": 6.1875, + "learning_rate": 9.993205723955734e-06, + "loss": 1.77450199, + "memory(GiB)": 58.14, + "step": 5185, + "train_speed(iter/s)": 0.452233 + }, + { + "acc": 0.63847804, + "epoch": 0.13165905631659056, + "grad_norm": 6.84375, + "learning_rate": 9.993150966220933e-06, + "loss": 1.70064011, + "memory(GiB)": 58.14, + "step": 5190, + "train_speed(iter/s)": 0.452466 + }, + { + "acc": 0.62693038, + "epoch": 0.13178589548452563, + "grad_norm": 5.1875, + "learning_rate": 9.993095988865057e-06, + "loss": 1.74888954, + "memory(GiB)": 58.14, + "step": 5195, + "train_speed(iter/s)": 0.452697 + }, + { + "acc": 0.62766085, + "epoch": 0.13191273465246067, + "grad_norm": 4.625, + "learning_rate": 9.99304079189053e-06, + "loss": 1.76516953, + "memory(GiB)": 58.14, + "step": 5200, + "train_speed(iter/s)": 0.452929 + }, + { + "epoch": 0.13191273465246067, + "eval_acc": 0.6260268198257325, + "eval_loss": 1.7115122079849243, + "eval_runtime": 112.8977, + "eval_samples_per_second": 56.423, + "eval_steps_per_second": 28.211, + "step": 5200 + }, + { + "acc": 0.64038615, + "epoch": 0.13203957382039574, + "grad_norm": 6.96875, + "learning_rate": 9.992985375299775e-06, + "loss": 1.71148281, + "memory(GiB)": 58.14, + "step": 5205, + "train_speed(iter/s)": 0.448574 + }, + { + "acc": 0.61575127, + "epoch": 0.1321664129883308, + "grad_norm": 6.15625, + "learning_rate": 9.992929739095232e-06, + "loss": 1.74047546, + "memory(GiB)": 58.14, + "step": 5210, + "train_speed(iter/s)": 0.448801 + }, + { + "acc": 0.62180467, + "epoch": 0.13229325215626586, + "grad_norm": 5.78125, + "learning_rate": 9.992873883279345e-06, + "loss": 1.80330601, + "memory(GiB)": 58.14, + "step": 5215, + "train_speed(iter/s)": 0.449026 + }, + { + "acc": 0.6183321, + "epoch": 0.1324200913242009, + "grad_norm": 4.3125, + "learning_rate": 9.992817807854575e-06, + "loss": 1.72780037, + "memory(GiB)": 58.14, + "step": 5220, + "train_speed(iter/s)": 0.449252 + }, + { + "acc": 0.63860044, + "epoch": 0.13254693049213598, + "grad_norm": 5.15625, + "learning_rate": 9.992761512823386e-06, + "loss": 1.74751244, + "memory(GiB)": 58.14, + "step": 5225, + "train_speed(iter/s)": 0.449478 + }, + { + "acc": 0.63357019, + "epoch": 0.13267376966007102, + "grad_norm": 6.34375, + "learning_rate": 9.992704998188255e-06, + "loss": 1.73373642, + "memory(GiB)": 58.14, + "step": 5230, + "train_speed(iter/s)": 0.449701 + }, + { + "acc": 0.61008377, + "epoch": 0.1328006088280061, + "grad_norm": 5.875, + "learning_rate": 9.992648263951668e-06, + "loss": 1.84115257, + "memory(GiB)": 58.14, + "step": 5235, + "train_speed(iter/s)": 0.449927 + }, + { + "acc": 0.6349618, + "epoch": 0.13292744799594114, + "grad_norm": 4.8125, + "learning_rate": 9.992591310116118e-06, + "loss": 1.75060005, + "memory(GiB)": 58.14, + "step": 5240, + "train_speed(iter/s)": 0.450151 + }, + { + "acc": 0.63464537, + "epoch": 0.1330542871638762, + "grad_norm": 5.0625, + "learning_rate": 9.992534136684112e-06, + "loss": 1.69572353, + "memory(GiB)": 58.14, + "step": 5245, + "train_speed(iter/s)": 0.450376 + }, + { + "acc": 0.63878732, + "epoch": 0.13318112633181126, + "grad_norm": 5.375, + "learning_rate": 9.992476743658165e-06, + "loss": 1.71288414, + "memory(GiB)": 58.14, + "step": 5250, + "train_speed(iter/s)": 0.4506 + }, + { + "acc": 0.6463376, + "epoch": 0.13330796549974633, + "grad_norm": 4.4375, + "learning_rate": 9.992419131040803e-06, + "loss": 1.65821724, + "memory(GiB)": 58.14, + "step": 5255, + "train_speed(iter/s)": 0.450822 + }, + { + "acc": 0.65013642, + "epoch": 0.13343480466768137, + "grad_norm": 5.375, + "learning_rate": 9.992361298834555e-06, + "loss": 1.63699226, + "memory(GiB)": 58.14, + "step": 5260, + "train_speed(iter/s)": 0.451048 + }, + { + "acc": 0.63408928, + "epoch": 0.13356164383561644, + "grad_norm": 5.21875, + "learning_rate": 9.99230324704197e-06, + "loss": 1.77814865, + "memory(GiB)": 58.14, + "step": 5265, + "train_speed(iter/s)": 0.451274 + }, + { + "acc": 0.64577188, + "epoch": 0.1336884830035515, + "grad_norm": 5.625, + "learning_rate": 9.992244975665598e-06, + "loss": 1.73048325, + "memory(GiB)": 58.14, + "step": 5270, + "train_speed(iter/s)": 0.451499 + }, + { + "acc": 0.6382637, + "epoch": 0.13381532217148656, + "grad_norm": 4.75, + "learning_rate": 9.992186484708003e-06, + "loss": 1.73220997, + "memory(GiB)": 58.14, + "step": 5275, + "train_speed(iter/s)": 0.451725 + }, + { + "acc": 0.63703451, + "epoch": 0.1339421613394216, + "grad_norm": 5.75, + "learning_rate": 9.992127774171759e-06, + "loss": 1.67782555, + "memory(GiB)": 58.14, + "step": 5280, + "train_speed(iter/s)": 0.45195 + }, + { + "acc": 0.62347593, + "epoch": 0.13406900050735668, + "grad_norm": 5.46875, + "learning_rate": 9.992068844059446e-06, + "loss": 1.86431198, + "memory(GiB)": 58.14, + "step": 5285, + "train_speed(iter/s)": 0.452172 + }, + { + "acc": 0.62769251, + "epoch": 0.13419583967529172, + "grad_norm": 6.78125, + "learning_rate": 9.992009694373658e-06, + "loss": 1.73803234, + "memory(GiB)": 58.14, + "step": 5290, + "train_speed(iter/s)": 0.452393 + }, + { + "acc": 0.63466682, + "epoch": 0.1343226788432268, + "grad_norm": 7.5, + "learning_rate": 9.991950325116995e-06, + "loss": 1.7753149, + "memory(GiB)": 58.14, + "step": 5295, + "train_speed(iter/s)": 0.452614 + }, + { + "acc": 0.62895641, + "epoch": 0.13444951801116184, + "grad_norm": 5.03125, + "learning_rate": 9.99189073629207e-06, + "loss": 1.69495907, + "memory(GiB)": 58.14, + "step": 5300, + "train_speed(iter/s)": 0.452836 + }, + { + "epoch": 0.13444951801116184, + "eval_acc": 0.6264725775449236, + "eval_loss": 1.708350419998169, + "eval_runtime": 112.8832, + "eval_samples_per_second": 56.43, + "eval_steps_per_second": 28.215, + "step": 5300 + }, + { + "acc": 0.61796231, + "epoch": 0.1345763571790969, + "grad_norm": 6.125, + "learning_rate": 9.991830927901505e-06, + "loss": 1.83663063, + "memory(GiB)": 58.14, + "step": 5305, + "train_speed(iter/s)": 0.448562 + }, + { + "acc": 0.63969641, + "epoch": 0.13470319634703196, + "grad_norm": 5.3125, + "learning_rate": 9.991770899947925e-06, + "loss": 1.66208706, + "memory(GiB)": 58.14, + "step": 5310, + "train_speed(iter/s)": 0.448778 + }, + { + "acc": 0.63490686, + "epoch": 0.13483003551496703, + "grad_norm": 6.90625, + "learning_rate": 9.991710652433977e-06, + "loss": 1.72326355, + "memory(GiB)": 58.14, + "step": 5315, + "train_speed(iter/s)": 0.449001 + }, + { + "acc": 0.6203721, + "epoch": 0.13495687468290207, + "grad_norm": 5.34375, + "learning_rate": 9.991650185362308e-06, + "loss": 1.74798393, + "memory(GiB)": 58.14, + "step": 5320, + "train_speed(iter/s)": 0.449222 + }, + { + "acc": 0.63403864, + "epoch": 0.13508371385083714, + "grad_norm": 5.71875, + "learning_rate": 9.991589498735577e-06, + "loss": 1.68383026, + "memory(GiB)": 58.14, + "step": 5325, + "train_speed(iter/s)": 0.449443 + }, + { + "acc": 0.63904982, + "epoch": 0.1352105530187722, + "grad_norm": 6.28125, + "learning_rate": 9.991528592556454e-06, + "loss": 1.70016575, + "memory(GiB)": 58.14, + "step": 5330, + "train_speed(iter/s)": 0.449667 + }, + { + "acc": 0.61987438, + "epoch": 0.13533739218670726, + "grad_norm": 6.15625, + "learning_rate": 9.991467466827618e-06, + "loss": 1.79643078, + "memory(GiB)": 58.14, + "step": 5335, + "train_speed(iter/s)": 0.449889 + }, + { + "acc": 0.63877478, + "epoch": 0.1354642313546423, + "grad_norm": 6.125, + "learning_rate": 9.99140612155176e-06, + "loss": 1.75025101, + "memory(GiB)": 58.14, + "step": 5340, + "train_speed(iter/s)": 0.450109 + }, + { + "acc": 0.61600952, + "epoch": 0.13559107052257738, + "grad_norm": 5.78125, + "learning_rate": 9.991344556731572e-06, + "loss": 1.7739048, + "memory(GiB)": 58.14, + "step": 5345, + "train_speed(iter/s)": 0.45033 + }, + { + "acc": 0.64483056, + "epoch": 0.13571790969051242, + "grad_norm": 5.15625, + "learning_rate": 9.991282772369766e-06, + "loss": 1.68408813, + "memory(GiB)": 58.14, + "step": 5350, + "train_speed(iter/s)": 0.450552 + }, + { + "acc": 0.63678727, + "epoch": 0.1358447488584475, + "grad_norm": 5.53125, + "learning_rate": 9.99122076846906e-06, + "loss": 1.67467785, + "memory(GiB)": 58.14, + "step": 5355, + "train_speed(iter/s)": 0.450771 + }, + { + "acc": 0.64348612, + "epoch": 0.13597158802638254, + "grad_norm": 5.0, + "learning_rate": 9.991158545032181e-06, + "loss": 1.72426262, + "memory(GiB)": 58.14, + "step": 5360, + "train_speed(iter/s)": 0.450991 + }, + { + "acc": 0.65286303, + "epoch": 0.1360984271943176, + "grad_norm": 5.1875, + "learning_rate": 9.991096102061865e-06, + "loss": 1.68236923, + "memory(GiB)": 58.14, + "step": 5365, + "train_speed(iter/s)": 0.451212 + }, + { + "acc": 0.63559065, + "epoch": 0.13622526636225266, + "grad_norm": 6.53125, + "learning_rate": 9.991033439560858e-06, + "loss": 1.71892796, + "memory(GiB)": 58.14, + "step": 5370, + "train_speed(iter/s)": 0.451433 + }, + { + "acc": 0.62383866, + "epoch": 0.13635210553018773, + "grad_norm": 4.65625, + "learning_rate": 9.990970557531918e-06, + "loss": 1.79070187, + "memory(GiB)": 58.14, + "step": 5375, + "train_speed(iter/s)": 0.451653 + }, + { + "acc": 0.65104985, + "epoch": 0.13647894469812277, + "grad_norm": 5.59375, + "learning_rate": 9.990907455977809e-06, + "loss": 1.69440556, + "memory(GiB)": 58.14, + "step": 5380, + "train_speed(iter/s)": 0.451872 + }, + { + "acc": 0.624965, + "epoch": 0.13660578386605784, + "grad_norm": 4.53125, + "learning_rate": 9.990844134901308e-06, + "loss": 1.75557594, + "memory(GiB)": 58.14, + "step": 5385, + "train_speed(iter/s)": 0.452092 + }, + { + "acc": 0.65195866, + "epoch": 0.1367326230339929, + "grad_norm": 4.9375, + "learning_rate": 9.9907805943052e-06, + "loss": 1.66792908, + "memory(GiB)": 58.14, + "step": 5390, + "train_speed(iter/s)": 0.45231 + }, + { + "acc": 0.63221412, + "epoch": 0.13685946220192796, + "grad_norm": 4.59375, + "learning_rate": 9.990716834192278e-06, + "loss": 1.66729813, + "memory(GiB)": 58.14, + "step": 5395, + "train_speed(iter/s)": 0.452528 + }, + { + "acc": 0.62673078, + "epoch": 0.136986301369863, + "grad_norm": 5.15625, + "learning_rate": 9.990652854565348e-06, + "loss": 1.7620472, + "memory(GiB)": 58.14, + "step": 5400, + "train_speed(iter/s)": 0.452746 + }, + { + "epoch": 0.136986301369863, + "eval_acc": 0.6268168178079164, + "eval_loss": 1.7059340476989746, + "eval_runtime": 112.4627, + "eval_samples_per_second": 56.641, + "eval_steps_per_second": 28.321, + "step": 5400 + }, + { + "acc": 0.62790203, + "epoch": 0.13711314053779808, + "grad_norm": 5.28125, + "learning_rate": 9.990588655427225e-06, + "loss": 1.77883072, + "memory(GiB)": 58.14, + "step": 5405, + "train_speed(iter/s)": 0.448567 + }, + { + "acc": 0.64419107, + "epoch": 0.13723997970573312, + "grad_norm": 5.65625, + "learning_rate": 9.99052423678073e-06, + "loss": 1.71225319, + "memory(GiB)": 58.14, + "step": 5410, + "train_speed(iter/s)": 0.448784 + }, + { + "acc": 0.63921566, + "epoch": 0.1373668188736682, + "grad_norm": 5.40625, + "learning_rate": 9.990459598628697e-06, + "loss": 1.66631508, + "memory(GiB)": 58.14, + "step": 5415, + "train_speed(iter/s)": 0.449002 + }, + { + "acc": 0.63895082, + "epoch": 0.13749365804160324, + "grad_norm": 4.4375, + "learning_rate": 9.990394740973972e-06, + "loss": 1.6214489, + "memory(GiB)": 58.14, + "step": 5420, + "train_speed(iter/s)": 0.449215 + }, + { + "acc": 0.64224005, + "epoch": 0.1376204972095383, + "grad_norm": 5.75, + "learning_rate": 9.990329663819405e-06, + "loss": 1.74676895, + "memory(GiB)": 58.14, + "step": 5425, + "train_speed(iter/s)": 0.449432 + }, + { + "acc": 0.63956337, + "epoch": 0.13774733637747336, + "grad_norm": 5.40625, + "learning_rate": 9.99026436716786e-06, + "loss": 1.69299278, + "memory(GiB)": 58.14, + "step": 5430, + "train_speed(iter/s)": 0.449646 + }, + { + "acc": 0.63139544, + "epoch": 0.13787417554540843, + "grad_norm": 4.28125, + "learning_rate": 9.990198851022207e-06, + "loss": 1.7079216, + "memory(GiB)": 58.14, + "step": 5435, + "train_speed(iter/s)": 0.449861 + }, + { + "acc": 0.6418067, + "epoch": 0.13800101471334347, + "grad_norm": 5.15625, + "learning_rate": 9.99013311538533e-06, + "loss": 1.69390507, + "memory(GiB)": 58.14, + "step": 5440, + "train_speed(iter/s)": 0.450076 + }, + { + "acc": 0.64191446, + "epoch": 0.13812785388127855, + "grad_norm": 5.875, + "learning_rate": 9.99006716026012e-06, + "loss": 1.72927017, + "memory(GiB)": 58.14, + "step": 5445, + "train_speed(iter/s)": 0.450292 + }, + { + "acc": 0.63658485, + "epoch": 0.1382546930492136, + "grad_norm": 4.90625, + "learning_rate": 9.990000985649475e-06, + "loss": 1.7440485, + "memory(GiB)": 58.14, + "step": 5450, + "train_speed(iter/s)": 0.450506 + }, + { + "acc": 0.63153868, + "epoch": 0.13838153221714866, + "grad_norm": 5.125, + "learning_rate": 9.989934591556308e-06, + "loss": 1.7365303, + "memory(GiB)": 58.14, + "step": 5455, + "train_speed(iter/s)": 0.45072 + }, + { + "acc": 0.6239574, + "epoch": 0.1385083713850837, + "grad_norm": 6.1875, + "learning_rate": 9.98986797798354e-06, + "loss": 1.76665497, + "memory(GiB)": 58.14, + "step": 5460, + "train_speed(iter/s)": 0.450933 + }, + { + "acc": 0.63665466, + "epoch": 0.13863521055301878, + "grad_norm": 5.84375, + "learning_rate": 9.989801144934102e-06, + "loss": 1.71201077, + "memory(GiB)": 58.14, + "step": 5465, + "train_speed(iter/s)": 0.451148 + }, + { + "acc": 0.6433466, + "epoch": 0.13876204972095382, + "grad_norm": 6.03125, + "learning_rate": 9.98973409241093e-06, + "loss": 1.7140377, + "memory(GiB)": 58.14, + "step": 5470, + "train_speed(iter/s)": 0.451363 + }, + { + "acc": 0.63071918, + "epoch": 0.1388888888888889, + "grad_norm": 5.125, + "learning_rate": 9.989666820416974e-06, + "loss": 1.70875015, + "memory(GiB)": 58.14, + "step": 5475, + "train_speed(iter/s)": 0.451577 + }, + { + "acc": 0.63630266, + "epoch": 0.13901572805682394, + "grad_norm": 7.40625, + "learning_rate": 9.989599328955195e-06, + "loss": 1.76143684, + "memory(GiB)": 58.14, + "step": 5480, + "train_speed(iter/s)": 0.451793 + }, + { + "acc": 0.63964424, + "epoch": 0.139142567224759, + "grad_norm": 4.5625, + "learning_rate": 9.98953161802856e-06, + "loss": 1.70436306, + "memory(GiB)": 58.14, + "step": 5485, + "train_speed(iter/s)": 0.452004 + }, + { + "acc": 0.65204549, + "epoch": 0.13926940639269406, + "grad_norm": 6.4375, + "learning_rate": 9.98946368764005e-06, + "loss": 1.66906738, + "memory(GiB)": 58.14, + "step": 5490, + "train_speed(iter/s)": 0.452217 + }, + { + "acc": 0.66132069, + "epoch": 0.13939624556062913, + "grad_norm": 6.84375, + "learning_rate": 9.989395537792647e-06, + "loss": 1.68920097, + "memory(GiB)": 58.14, + "step": 5495, + "train_speed(iter/s)": 0.452431 + }, + { + "acc": 0.62196002, + "epoch": 0.13952308472856417, + "grad_norm": 5.03125, + "learning_rate": 9.989327168489356e-06, + "loss": 1.73628559, + "memory(GiB)": 58.14, + "step": 5500, + "train_speed(iter/s)": 0.452643 + }, + { + "epoch": 0.13952308472856417, + "eval_acc": 0.6272792862194857, + "eval_loss": 1.7042996883392334, + "eval_runtime": 113.2184, + "eval_samples_per_second": 56.263, + "eval_steps_per_second": 28.131, + "step": 5500 + }, + { + "acc": 0.63909187, + "epoch": 0.13964992389649925, + "grad_norm": 5.03125, + "learning_rate": 9.989258579733179e-06, + "loss": 1.74666195, + "memory(GiB)": 58.14, + "step": 5505, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.63207836, + "epoch": 0.1397767630644343, + "grad_norm": 5.03125, + "learning_rate": 9.989189771527133e-06, + "loss": 1.65514946, + "memory(GiB)": 58.14, + "step": 5510, + "train_speed(iter/s)": 0.448724 + }, + { + "acc": 0.6344851, + "epoch": 0.13990360223236936, + "grad_norm": 5.03125, + "learning_rate": 9.989120743874248e-06, + "loss": 1.73748684, + "memory(GiB)": 58.14, + "step": 5515, + "train_speed(iter/s)": 0.448935 + }, + { + "acc": 0.62594395, + "epoch": 0.1400304414003044, + "grad_norm": 7.375, + "learning_rate": 9.989051496777556e-06, + "loss": 1.80770874, + "memory(GiB)": 58.14, + "step": 5520, + "train_speed(iter/s)": 0.449149 + }, + { + "acc": 0.63148661, + "epoch": 0.14015728056823948, + "grad_norm": 5.40625, + "learning_rate": 9.988982030240104e-06, + "loss": 1.74213238, + "memory(GiB)": 58.14, + "step": 5525, + "train_speed(iter/s)": 0.449357 + }, + { + "acc": 0.63529158, + "epoch": 0.14028411973617452, + "grad_norm": 5.90625, + "learning_rate": 9.988912344264949e-06, + "loss": 1.79423065, + "memory(GiB)": 58.14, + "step": 5530, + "train_speed(iter/s)": 0.449569 + }, + { + "acc": 0.63416309, + "epoch": 0.1404109589041096, + "grad_norm": 4.5625, + "learning_rate": 9.988842438855156e-06, + "loss": 1.6672617, + "memory(GiB)": 58.14, + "step": 5535, + "train_speed(iter/s)": 0.449779 + }, + { + "acc": 0.62745543, + "epoch": 0.14053779807204464, + "grad_norm": 5.90625, + "learning_rate": 9.988772314013799e-06, + "loss": 1.73622322, + "memory(GiB)": 58.14, + "step": 5540, + "train_speed(iter/s)": 0.449991 + }, + { + "acc": 0.63059883, + "epoch": 0.1406646372399797, + "grad_norm": 5.6875, + "learning_rate": 9.988701969743961e-06, + "loss": 1.72747498, + "memory(GiB)": 58.14, + "step": 5545, + "train_speed(iter/s)": 0.450204 + }, + { + "acc": 0.6283535, + "epoch": 0.14079147640791476, + "grad_norm": 4.40625, + "learning_rate": 9.98863140604874e-06, + "loss": 1.75545044, + "memory(GiB)": 58.14, + "step": 5550, + "train_speed(iter/s)": 0.450416 + }, + { + "acc": 0.63903456, + "epoch": 0.14091831557584983, + "grad_norm": 5.25, + "learning_rate": 9.988560622931233e-06, + "loss": 1.68769913, + "memory(GiB)": 58.14, + "step": 5555, + "train_speed(iter/s)": 0.450626 + }, + { + "acc": 0.65626926, + "epoch": 0.14104515474378487, + "grad_norm": 4.875, + "learning_rate": 9.988489620394562e-06, + "loss": 1.65706215, + "memory(GiB)": 58.14, + "step": 5560, + "train_speed(iter/s)": 0.450836 + }, + { + "acc": 0.6474865, + "epoch": 0.14117199391171995, + "grad_norm": 4.875, + "learning_rate": 9.988418398441842e-06, + "loss": 1.67927246, + "memory(GiB)": 58.14, + "step": 5565, + "train_speed(iter/s)": 0.451046 + }, + { + "acc": 0.61654778, + "epoch": 0.141298833079655, + "grad_norm": 5.5625, + "learning_rate": 9.98834695707621e-06, + "loss": 1.83884811, + "memory(GiB)": 58.14, + "step": 5570, + "train_speed(iter/s)": 0.451257 + }, + { + "acc": 0.62592435, + "epoch": 0.14142567224759006, + "grad_norm": 4.78125, + "learning_rate": 9.98827529630081e-06, + "loss": 1.73308716, + "memory(GiB)": 58.14, + "step": 5575, + "train_speed(iter/s)": 0.451468 + }, + { + "acc": 0.63740864, + "epoch": 0.1415525114155251, + "grad_norm": 4.59375, + "learning_rate": 9.988203416118788e-06, + "loss": 1.646661, + "memory(GiB)": 58.14, + "step": 5580, + "train_speed(iter/s)": 0.451676 + }, + { + "acc": 0.64474587, + "epoch": 0.14167935058346018, + "grad_norm": 5.0625, + "learning_rate": 9.98813131653331e-06, + "loss": 1.68919258, + "memory(GiB)": 58.14, + "step": 5585, + "train_speed(iter/s)": 0.451885 + }, + { + "acc": 0.62830696, + "epoch": 0.14180618975139522, + "grad_norm": 4.78125, + "learning_rate": 9.988058997547548e-06, + "loss": 1.7219902, + "memory(GiB)": 58.14, + "step": 5590, + "train_speed(iter/s)": 0.452092 + }, + { + "acc": 0.62381487, + "epoch": 0.1419330289193303, + "grad_norm": 6.15625, + "learning_rate": 9.987986459164678e-06, + "loss": 1.81512146, + "memory(GiB)": 58.14, + "step": 5595, + "train_speed(iter/s)": 0.452304 + }, + { + "acc": 0.63764067, + "epoch": 0.14205986808726534, + "grad_norm": 6.21875, + "learning_rate": 9.987913701387897e-06, + "loss": 1.73536224, + "memory(GiB)": 58.14, + "step": 5600, + "train_speed(iter/s)": 0.452514 + }, + { + "epoch": 0.14205986808726534, + "eval_acc": 0.6277634785311469, + "eval_loss": 1.7015951871871948, + "eval_runtime": 114.8653, + "eval_samples_per_second": 55.456, + "eval_steps_per_second": 27.728, + "step": 5600 + }, + { + "acc": 0.62242794, + "epoch": 0.1421867072552004, + "grad_norm": 4.875, + "learning_rate": 9.9878407242204e-06, + "loss": 1.7536972, + "memory(GiB)": 58.14, + "step": 5605, + "train_speed(iter/s)": 0.448403 + }, + { + "acc": 0.62527184, + "epoch": 0.14231354642313546, + "grad_norm": 5.59375, + "learning_rate": 9.9877675276654e-06, + "loss": 1.72032719, + "memory(GiB)": 58.14, + "step": 5610, + "train_speed(iter/s)": 0.448612 + }, + { + "acc": 0.64107642, + "epoch": 0.14244038559107053, + "grad_norm": 4.8125, + "learning_rate": 9.987694111726114e-06, + "loss": 1.69206257, + "memory(GiB)": 58.14, + "step": 5615, + "train_speed(iter/s)": 0.448819 + }, + { + "acc": 0.64274669, + "epoch": 0.14256722475900557, + "grad_norm": 4.84375, + "learning_rate": 9.987620476405774e-06, + "loss": 1.68389683, + "memory(GiB)": 58.14, + "step": 5620, + "train_speed(iter/s)": 0.449027 + }, + { + "acc": 0.65106792, + "epoch": 0.14269406392694065, + "grad_norm": 5.0625, + "learning_rate": 9.987546621707616e-06, + "loss": 1.67289543, + "memory(GiB)": 58.14, + "step": 5625, + "train_speed(iter/s)": 0.449234 + }, + { + "acc": 0.62957878, + "epoch": 0.1428209030948757, + "grad_norm": 5.59375, + "learning_rate": 9.98747254763489e-06, + "loss": 1.74797707, + "memory(GiB)": 58.14, + "step": 5630, + "train_speed(iter/s)": 0.449441 + }, + { + "acc": 0.61637611, + "epoch": 0.14294774226281076, + "grad_norm": 5.75, + "learning_rate": 9.987398254190855e-06, + "loss": 1.78369484, + "memory(GiB)": 58.14, + "step": 5635, + "train_speed(iter/s)": 0.44965 + }, + { + "acc": 0.63721752, + "epoch": 0.1430745814307458, + "grad_norm": 5.40625, + "learning_rate": 9.987323741378777e-06, + "loss": 1.73150635, + "memory(GiB)": 58.14, + "step": 5640, + "train_speed(iter/s)": 0.449858 + }, + { + "acc": 0.66023273, + "epoch": 0.14320142059868088, + "grad_norm": 5.96875, + "learning_rate": 9.987249009201934e-06, + "loss": 1.63697414, + "memory(GiB)": 58.14, + "step": 5645, + "train_speed(iter/s)": 0.450065 + }, + { + "acc": 0.64306326, + "epoch": 0.14332825976661592, + "grad_norm": 6.53125, + "learning_rate": 9.987174057663613e-06, + "loss": 1.67622528, + "memory(GiB)": 58.14, + "step": 5650, + "train_speed(iter/s)": 0.450275 + }, + { + "acc": 0.62775774, + "epoch": 0.143455098934551, + "grad_norm": 5.6875, + "learning_rate": 9.987098886767111e-06, + "loss": 1.70162601, + "memory(GiB)": 58.14, + "step": 5655, + "train_speed(iter/s)": 0.450482 + }, + { + "acc": 0.61782589, + "epoch": 0.14358193810248604, + "grad_norm": 5.53125, + "learning_rate": 9.987023496515734e-06, + "loss": 1.81647758, + "memory(GiB)": 58.14, + "step": 5660, + "train_speed(iter/s)": 0.450688 + }, + { + "acc": 0.63810525, + "epoch": 0.1437087772704211, + "grad_norm": 6.65625, + "learning_rate": 9.9869478869128e-06, + "loss": 1.76525631, + "memory(GiB)": 58.14, + "step": 5665, + "train_speed(iter/s)": 0.450896 + }, + { + "acc": 0.62699556, + "epoch": 0.14383561643835616, + "grad_norm": 5.09375, + "learning_rate": 9.98687205796163e-06, + "loss": 1.75273781, + "memory(GiB)": 58.14, + "step": 5670, + "train_speed(iter/s)": 0.451105 + }, + { + "acc": 0.63688421, + "epoch": 0.14396245560629123, + "grad_norm": 5.78125, + "learning_rate": 9.986796009665562e-06, + "loss": 1.66628647, + "memory(GiB)": 58.14, + "step": 5675, + "train_speed(iter/s)": 0.451313 + }, + { + "acc": 0.63535285, + "epoch": 0.14408929477422627, + "grad_norm": 4.5, + "learning_rate": 9.986719742027944e-06, + "loss": 1.64191208, + "memory(GiB)": 58.14, + "step": 5680, + "train_speed(iter/s)": 0.451518 + }, + { + "acc": 0.6325182, + "epoch": 0.14421613394216135, + "grad_norm": 5.375, + "learning_rate": 9.986643255052125e-06, + "loss": 1.70225487, + "memory(GiB)": 58.14, + "step": 5685, + "train_speed(iter/s)": 0.451727 + }, + { + "acc": 0.62961588, + "epoch": 0.1443429731100964, + "grad_norm": 4.78125, + "learning_rate": 9.986566548741473e-06, + "loss": 1.73679352, + "memory(GiB)": 58.14, + "step": 5690, + "train_speed(iter/s)": 0.451934 + }, + { + "acc": 0.64627624, + "epoch": 0.14446981227803146, + "grad_norm": 5.5, + "learning_rate": 9.98648962309936e-06, + "loss": 1.69186974, + "memory(GiB)": 58.14, + "step": 5695, + "train_speed(iter/s)": 0.452137 + }, + { + "acc": 0.64414415, + "epoch": 0.1445966514459665, + "grad_norm": 5.78125, + "learning_rate": 9.986412478129171e-06, + "loss": 1.65069256, + "memory(GiB)": 58.14, + "step": 5700, + "train_speed(iter/s)": 0.452345 + }, + { + "epoch": 0.1445966514459665, + "eval_acc": 0.6280116323129645, + "eval_loss": 1.7001053094863892, + "eval_runtime": 113.0183, + "eval_samples_per_second": 56.363, + "eval_steps_per_second": 28.181, + "step": 5700 + }, + { + "acc": 0.63531656, + "epoch": 0.14472349061390158, + "grad_norm": 7.5625, + "learning_rate": 9.9863351138343e-06, + "loss": 1.72501373, + "memory(GiB)": 58.14, + "step": 5705, + "train_speed(iter/s)": 0.448374 + }, + { + "acc": 0.65072451, + "epoch": 0.14485032978183662, + "grad_norm": 5.09375, + "learning_rate": 9.986257530218146e-06, + "loss": 1.67599182, + "memory(GiB)": 58.14, + "step": 5710, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.63838205, + "epoch": 0.1449771689497717, + "grad_norm": 5.03125, + "learning_rate": 9.986179727284124e-06, + "loss": 1.650177, + "memory(GiB)": 58.14, + "step": 5715, + "train_speed(iter/s)": 0.448782 + }, + { + "acc": 0.62168951, + "epoch": 0.14510400811770674, + "grad_norm": 6.625, + "learning_rate": 9.986101705035656e-06, + "loss": 1.78340683, + "memory(GiB)": 58.14, + "step": 5720, + "train_speed(iter/s)": 0.448986 + }, + { + "acc": 0.62519422, + "epoch": 0.1452308472856418, + "grad_norm": 5.25, + "learning_rate": 9.986023463476175e-06, + "loss": 1.782201, + "memory(GiB)": 58.14, + "step": 5725, + "train_speed(iter/s)": 0.44919 + }, + { + "acc": 0.62733116, + "epoch": 0.14535768645357686, + "grad_norm": 5.25, + "learning_rate": 9.985945002609119e-06, + "loss": 1.75075874, + "memory(GiB)": 58.14, + "step": 5730, + "train_speed(iter/s)": 0.449396 + }, + { + "acc": 0.63228254, + "epoch": 0.14548452562151193, + "grad_norm": 6.71875, + "learning_rate": 9.985866322437942e-06, + "loss": 1.78584824, + "memory(GiB)": 58.14, + "step": 5735, + "train_speed(iter/s)": 0.449601 + }, + { + "acc": 0.63814306, + "epoch": 0.14561136478944697, + "grad_norm": 7.03125, + "learning_rate": 9.985787422966105e-06, + "loss": 1.7019392, + "memory(GiB)": 58.14, + "step": 5740, + "train_speed(iter/s)": 0.449809 + }, + { + "acc": 0.63795519, + "epoch": 0.14573820395738205, + "grad_norm": 6.4375, + "learning_rate": 9.985708304197075e-06, + "loss": 1.79319534, + "memory(GiB)": 58.14, + "step": 5745, + "train_speed(iter/s)": 0.450014 + }, + { + "acc": 0.63502231, + "epoch": 0.1458650431253171, + "grad_norm": 5.125, + "learning_rate": 9.985628966134336e-06, + "loss": 1.71769161, + "memory(GiB)": 58.14, + "step": 5750, + "train_speed(iter/s)": 0.45022 + }, + { + "acc": 0.63223782, + "epoch": 0.14599188229325216, + "grad_norm": 6.53125, + "learning_rate": 9.985549408781377e-06, + "loss": 1.7181015, + "memory(GiB)": 58.14, + "step": 5755, + "train_speed(iter/s)": 0.450426 + }, + { + "acc": 0.62663708, + "epoch": 0.1461187214611872, + "grad_norm": 4.84375, + "learning_rate": 9.985469632141693e-06, + "loss": 1.78344822, + "memory(GiB)": 58.14, + "step": 5760, + "train_speed(iter/s)": 0.450631 + }, + { + "acc": 0.63541446, + "epoch": 0.14624556062912228, + "grad_norm": 4.90625, + "learning_rate": 9.985389636218797e-06, + "loss": 1.71079865, + "memory(GiB)": 58.14, + "step": 5765, + "train_speed(iter/s)": 0.450836 + }, + { + "acc": 0.63279858, + "epoch": 0.14637239979705732, + "grad_norm": 5.3125, + "learning_rate": 9.985309421016207e-06, + "loss": 1.68989525, + "memory(GiB)": 58.14, + "step": 5770, + "train_speed(iter/s)": 0.451043 + }, + { + "acc": 0.63986549, + "epoch": 0.1464992389649924, + "grad_norm": 5.375, + "learning_rate": 9.985228986537451e-06, + "loss": 1.71737595, + "memory(GiB)": 58.14, + "step": 5775, + "train_speed(iter/s)": 0.45125 + }, + { + "acc": 0.63557129, + "epoch": 0.14662607813292744, + "grad_norm": 5.3125, + "learning_rate": 9.985148332786068e-06, + "loss": 1.68898048, + "memory(GiB)": 58.14, + "step": 5780, + "train_speed(iter/s)": 0.451455 + }, + { + "acc": 0.63252234, + "epoch": 0.1467529173008625, + "grad_norm": 5.75, + "learning_rate": 9.985067459765603e-06, + "loss": 1.73890591, + "memory(GiB)": 58.14, + "step": 5785, + "train_speed(iter/s)": 0.451662 + }, + { + "acc": 0.6361464, + "epoch": 0.14687975646879756, + "grad_norm": 5.15625, + "learning_rate": 9.984986367479615e-06, + "loss": 1.74285889, + "memory(GiB)": 58.14, + "step": 5790, + "train_speed(iter/s)": 0.45187 + }, + { + "acc": 0.64615755, + "epoch": 0.14700659563673263, + "grad_norm": 5.09375, + "learning_rate": 9.984905055931668e-06, + "loss": 1.66937733, + "memory(GiB)": 58.14, + "step": 5795, + "train_speed(iter/s)": 0.452077 + }, + { + "acc": 0.61166282, + "epoch": 0.14713343480466767, + "grad_norm": 5.09375, + "learning_rate": 9.984823525125342e-06, + "loss": 1.78655033, + "memory(GiB)": 58.14, + "step": 5800, + "train_speed(iter/s)": 0.452284 + }, + { + "epoch": 0.14713343480466767, + "eval_acc": 0.6284841271399608, + "eval_loss": 1.6969976425170898, + "eval_runtime": 113.8122, + "eval_samples_per_second": 55.969, + "eval_steps_per_second": 27.985, + "step": 5800 + }, + { + "acc": 0.63178768, + "epoch": 0.14726027397260275, + "grad_norm": 4.8125, + "learning_rate": 9.984741775064222e-06, + "loss": 1.75447617, + "memory(GiB)": 58.14, + "step": 5805, + "train_speed(iter/s)": 0.448355 + }, + { + "acc": 0.62364249, + "epoch": 0.1473871131405378, + "grad_norm": 4.84375, + "learning_rate": 9.984659805751904e-06, + "loss": 1.77045193, + "memory(GiB)": 58.14, + "step": 5810, + "train_speed(iter/s)": 0.44856 + }, + { + "acc": 0.64836936, + "epoch": 0.14751395230847286, + "grad_norm": 5.0625, + "learning_rate": 9.984577617191993e-06, + "loss": 1.71013165, + "memory(GiB)": 58.14, + "step": 5815, + "train_speed(iter/s)": 0.448765 + }, + { + "acc": 0.62805309, + "epoch": 0.1476407914764079, + "grad_norm": 5.625, + "learning_rate": 9.984495209388102e-06, + "loss": 1.76644897, + "memory(GiB)": 58.14, + "step": 5820, + "train_speed(iter/s)": 0.448972 + }, + { + "acc": 0.62435865, + "epoch": 0.14776763064434298, + "grad_norm": 5.875, + "learning_rate": 9.984412582343859e-06, + "loss": 1.77659149, + "memory(GiB)": 58.14, + "step": 5825, + "train_speed(iter/s)": 0.449179 + }, + { + "acc": 0.63662348, + "epoch": 0.14789446981227802, + "grad_norm": 5.96875, + "learning_rate": 9.984329736062896e-06, + "loss": 1.64894829, + "memory(GiB)": 58.14, + "step": 5830, + "train_speed(iter/s)": 0.449385 + }, + { + "acc": 0.6193377, + "epoch": 0.1480213089802131, + "grad_norm": 5.25, + "learning_rate": 9.984246670548858e-06, + "loss": 1.74502544, + "memory(GiB)": 58.14, + "step": 5835, + "train_speed(iter/s)": 0.449588 + }, + { + "acc": 0.63835092, + "epoch": 0.14814814814814814, + "grad_norm": 5.03125, + "learning_rate": 9.984163385805398e-06, + "loss": 1.76451836, + "memory(GiB)": 58.14, + "step": 5840, + "train_speed(iter/s)": 0.449793 + }, + { + "acc": 0.63051615, + "epoch": 0.1482749873160832, + "grad_norm": 6.6875, + "learning_rate": 9.984079881836182e-06, + "loss": 1.72792664, + "memory(GiB)": 58.14, + "step": 5845, + "train_speed(iter/s)": 0.449997 + }, + { + "acc": 0.6393959, + "epoch": 0.14840182648401826, + "grad_norm": 6.09375, + "learning_rate": 9.983996158644877e-06, + "loss": 1.687257, + "memory(GiB)": 58.14, + "step": 5850, + "train_speed(iter/s)": 0.450201 + }, + { + "acc": 0.64106522, + "epoch": 0.14852866565195333, + "grad_norm": 4.28125, + "learning_rate": 9.983912216235172e-06, + "loss": 1.62324905, + "memory(GiB)": 58.14, + "step": 5855, + "train_speed(iter/s)": 0.450406 + }, + { + "acc": 0.64793434, + "epoch": 0.14865550481988837, + "grad_norm": 5.90625, + "learning_rate": 9.983828054610754e-06, + "loss": 1.76565208, + "memory(GiB)": 58.14, + "step": 5860, + "train_speed(iter/s)": 0.450611 + }, + { + "acc": 0.63874741, + "epoch": 0.14878234398782345, + "grad_norm": 5.8125, + "learning_rate": 9.983743673775328e-06, + "loss": 1.75894775, + "memory(GiB)": 58.14, + "step": 5865, + "train_speed(iter/s)": 0.450815 + }, + { + "acc": 0.63766284, + "epoch": 0.1489091831557585, + "grad_norm": 6.71875, + "learning_rate": 9.983659073732604e-06, + "loss": 1.71356277, + "memory(GiB)": 58.14, + "step": 5870, + "train_speed(iter/s)": 0.451021 + }, + { + "acc": 0.63759117, + "epoch": 0.14903602232369356, + "grad_norm": 6.1875, + "learning_rate": 9.983574254486303e-06, + "loss": 1.72629471, + "memory(GiB)": 58.14, + "step": 5875, + "train_speed(iter/s)": 0.451224 + }, + { + "acc": 0.65143981, + "epoch": 0.1491628614916286, + "grad_norm": 4.6875, + "learning_rate": 9.983489216040158e-06, + "loss": 1.71128082, + "memory(GiB)": 58.14, + "step": 5880, + "train_speed(iter/s)": 0.451429 + }, + { + "acc": 0.64596019, + "epoch": 0.14928970065956368, + "grad_norm": 4.90625, + "learning_rate": 9.983403958397907e-06, + "loss": 1.72300262, + "memory(GiB)": 58.14, + "step": 5885, + "train_speed(iter/s)": 0.451634 + }, + { + "acc": 0.62979126, + "epoch": 0.14941653982749872, + "grad_norm": 5.53125, + "learning_rate": 9.9833184815633e-06, + "loss": 1.86155815, + "memory(GiB)": 58.14, + "step": 5890, + "train_speed(iter/s)": 0.451837 + }, + { + "acc": 0.6280376, + "epoch": 0.1495433789954338, + "grad_norm": 4.8125, + "learning_rate": 9.983232785540097e-06, + "loss": 1.7970787, + "memory(GiB)": 58.14, + "step": 5895, + "train_speed(iter/s)": 0.45204 + }, + { + "acc": 0.62411852, + "epoch": 0.14967021816336884, + "grad_norm": 5.71875, + "learning_rate": 9.983146870332068e-06, + "loss": 1.74064713, + "memory(GiB)": 58.14, + "step": 5900, + "train_speed(iter/s)": 0.452243 + }, + { + "epoch": 0.14967021816336884, + "eval_acc": 0.6287222545063516, + "eval_loss": 1.6949149370193481, + "eval_runtime": 113.9019, + "eval_samples_per_second": 55.925, + "eval_steps_per_second": 27.963, + "step": 5900 + }, + { + "acc": 0.62579064, + "epoch": 0.1497970573313039, + "grad_norm": 7.375, + "learning_rate": 9.98306073594299e-06, + "loss": 1.77762947, + "memory(GiB)": 58.14, + "step": 5905, + "train_speed(iter/s)": 0.448379 + }, + { + "acc": 0.64695463, + "epoch": 0.14992389649923896, + "grad_norm": 6.53125, + "learning_rate": 9.982974382376656e-06, + "loss": 1.70982094, + "memory(GiB)": 58.14, + "step": 5910, + "train_speed(iter/s)": 0.448582 + }, + { + "acc": 0.65326896, + "epoch": 0.15005073566717403, + "grad_norm": 5.3125, + "learning_rate": 9.98288780963686e-06, + "loss": 1.67041759, + "memory(GiB)": 58.14, + "step": 5915, + "train_speed(iter/s)": 0.448785 + }, + { + "acc": 0.63187575, + "epoch": 0.15017757483510907, + "grad_norm": 4.59375, + "learning_rate": 9.98280101772741e-06, + "loss": 1.77075977, + "memory(GiB)": 58.14, + "step": 5920, + "train_speed(iter/s)": 0.448989 + }, + { + "acc": 0.63178511, + "epoch": 0.15030441400304415, + "grad_norm": 4.96875, + "learning_rate": 9.982714006652126e-06, + "loss": 1.73691139, + "memory(GiB)": 58.14, + "step": 5925, + "train_speed(iter/s)": 0.44919 + }, + { + "acc": 0.63479156, + "epoch": 0.1504312531709792, + "grad_norm": 4.75, + "learning_rate": 9.982626776414834e-06, + "loss": 1.70704155, + "memory(GiB)": 58.14, + "step": 5930, + "train_speed(iter/s)": 0.449391 + }, + { + "acc": 0.62522068, + "epoch": 0.15055809233891426, + "grad_norm": 4.84375, + "learning_rate": 9.98253932701937e-06, + "loss": 1.73893242, + "memory(GiB)": 58.14, + "step": 5935, + "train_speed(iter/s)": 0.449594 + }, + { + "acc": 0.64340014, + "epoch": 0.1506849315068493, + "grad_norm": 6.4375, + "learning_rate": 9.98245165846958e-06, + "loss": 1.7726963, + "memory(GiB)": 58.14, + "step": 5940, + "train_speed(iter/s)": 0.449796 + }, + { + "acc": 0.64046078, + "epoch": 0.15081177067478438, + "grad_norm": 5.96875, + "learning_rate": 9.982363770769323e-06, + "loss": 1.7123291, + "memory(GiB)": 58.14, + "step": 5945, + "train_speed(iter/s)": 0.449999 + }, + { + "acc": 0.63388672, + "epoch": 0.15093860984271942, + "grad_norm": 5.71875, + "learning_rate": 9.98227566392246e-06, + "loss": 1.73712234, + "memory(GiB)": 58.14, + "step": 5950, + "train_speed(iter/s)": 0.450202 + }, + { + "acc": 0.63900013, + "epoch": 0.1510654490106545, + "grad_norm": 6.375, + "learning_rate": 9.982187337932871e-06, + "loss": 1.71467361, + "memory(GiB)": 58.14, + "step": 5955, + "train_speed(iter/s)": 0.450404 + }, + { + "acc": 0.6155602, + "epoch": 0.15119228817858954, + "grad_norm": 4.71875, + "learning_rate": 9.98209879280444e-06, + "loss": 1.77856636, + "memory(GiB)": 58.14, + "step": 5960, + "train_speed(iter/s)": 0.450607 + }, + { + "acc": 0.6283668, + "epoch": 0.1513191273465246, + "grad_norm": 6.0, + "learning_rate": 9.982010028541057e-06, + "loss": 1.752075, + "memory(GiB)": 58.14, + "step": 5965, + "train_speed(iter/s)": 0.450808 + }, + { + "acc": 0.64005599, + "epoch": 0.15144596651445966, + "grad_norm": 6.1875, + "learning_rate": 9.981921045146633e-06, + "loss": 1.68224182, + "memory(GiB)": 58.14, + "step": 5970, + "train_speed(iter/s)": 0.45101 + }, + { + "acc": 0.64318318, + "epoch": 0.15157280568239473, + "grad_norm": 5.09375, + "learning_rate": 9.981831842625079e-06, + "loss": 1.67549915, + "memory(GiB)": 58.14, + "step": 5975, + "train_speed(iter/s)": 0.451212 + }, + { + "acc": 0.6340692, + "epoch": 0.15169964485032977, + "grad_norm": 5.375, + "learning_rate": 9.981742420980316e-06, + "loss": 1.74056034, + "memory(GiB)": 58.14, + "step": 5980, + "train_speed(iter/s)": 0.451414 + }, + { + "acc": 0.63649821, + "epoch": 0.15182648401826485, + "grad_norm": 5.25, + "learning_rate": 9.981652780216281e-06, + "loss": 1.68402672, + "memory(GiB)": 58.14, + "step": 5985, + "train_speed(iter/s)": 0.451617 + }, + { + "acc": 0.65454082, + "epoch": 0.1519533231861999, + "grad_norm": 6.6875, + "learning_rate": 9.981562920336915e-06, + "loss": 1.67746849, + "memory(GiB)": 58.14, + "step": 5990, + "train_speed(iter/s)": 0.451818 + }, + { + "acc": 0.62003784, + "epoch": 0.15208016235413496, + "grad_norm": 5.5, + "learning_rate": 9.98147284134617e-06, + "loss": 1.77328529, + "memory(GiB)": 58.14, + "step": 5995, + "train_speed(iter/s)": 0.452019 + }, + { + "acc": 0.63403778, + "epoch": 0.15220700152207, + "grad_norm": 6.15625, + "learning_rate": 9.981382543248011e-06, + "loss": 1.74843636, + "memory(GiB)": 58.14, + "step": 6000, + "train_speed(iter/s)": 0.452218 + }, + { + "epoch": 0.15220700152207, + "eval_acc": 0.6291751142698033, + "eval_loss": 1.6928975582122803, + "eval_runtime": 113.8661, + "eval_samples_per_second": 55.943, + "eval_steps_per_second": 27.971, + "step": 6000 + }, + { + "acc": 0.63630781, + "epoch": 0.15233384069000508, + "grad_norm": 5.53125, + "learning_rate": 9.981292026046406e-06, + "loss": 1.70675201, + "memory(GiB)": 58.14, + "step": 6005, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.63288488, + "epoch": 0.15246067985794012, + "grad_norm": 5.375, + "learning_rate": 9.981201289745337e-06, + "loss": 1.69798717, + "memory(GiB)": 58.14, + "step": 6010, + "train_speed(iter/s)": 0.448619 + }, + { + "acc": 0.62116013, + "epoch": 0.1525875190258752, + "grad_norm": 5.46875, + "learning_rate": 9.981110334348796e-06, + "loss": 1.7549902, + "memory(GiB)": 58.14, + "step": 6015, + "train_speed(iter/s)": 0.44882 + }, + { + "acc": 0.64530439, + "epoch": 0.15271435819381024, + "grad_norm": 5.84375, + "learning_rate": 9.981019159860782e-06, + "loss": 1.67368908, + "memory(GiB)": 58.14, + "step": 6020, + "train_speed(iter/s)": 0.449022 + }, + { + "acc": 0.63427982, + "epoch": 0.1528411973617453, + "grad_norm": 5.34375, + "learning_rate": 9.98092776628531e-06, + "loss": 1.68455658, + "memory(GiB)": 58.14, + "step": 6025, + "train_speed(iter/s)": 0.449222 + }, + { + "acc": 0.64319792, + "epoch": 0.15296803652968036, + "grad_norm": 4.6875, + "learning_rate": 9.980836153626396e-06, + "loss": 1.65393372, + "memory(GiB)": 58.14, + "step": 6030, + "train_speed(iter/s)": 0.449422 + }, + { + "acc": 0.64806957, + "epoch": 0.15309487569761543, + "grad_norm": 5.4375, + "learning_rate": 9.980744321888068e-06, + "loss": 1.69325905, + "memory(GiB)": 58.14, + "step": 6035, + "train_speed(iter/s)": 0.44962 + }, + { + "acc": 0.62998781, + "epoch": 0.15322171486555047, + "grad_norm": 4.8125, + "learning_rate": 9.98065227107437e-06, + "loss": 1.80609856, + "memory(GiB)": 58.14, + "step": 6040, + "train_speed(iter/s)": 0.449818 + }, + { + "acc": 0.63349195, + "epoch": 0.15334855403348555, + "grad_norm": 4.90625, + "learning_rate": 9.980560001189346e-06, + "loss": 1.73680534, + "memory(GiB)": 58.14, + "step": 6045, + "train_speed(iter/s)": 0.450017 + }, + { + "acc": 0.62060304, + "epoch": 0.1534753932014206, + "grad_norm": 5.0, + "learning_rate": 9.980467512237058e-06, + "loss": 1.77974815, + "memory(GiB)": 58.14, + "step": 6050, + "train_speed(iter/s)": 0.450215 + }, + { + "acc": 0.62257667, + "epoch": 0.15360223236935566, + "grad_norm": 6.25, + "learning_rate": 9.98037480422157e-06, + "loss": 1.68618126, + "memory(GiB)": 71.21, + "step": 6055, + "train_speed(iter/s)": 0.450413 + }, + { + "acc": 0.64822035, + "epoch": 0.1537290715372907, + "grad_norm": 5.8125, + "learning_rate": 9.980281877146964e-06, + "loss": 1.69027042, + "memory(GiB)": 71.21, + "step": 6060, + "train_speed(iter/s)": 0.450613 + }, + { + "acc": 0.63800855, + "epoch": 0.15385591070522578, + "grad_norm": 4.8125, + "learning_rate": 9.980188731017327e-06, + "loss": 1.74775219, + "memory(GiB)": 71.21, + "step": 6065, + "train_speed(iter/s)": 0.450812 + }, + { + "acc": 0.64689889, + "epoch": 0.15398274987316082, + "grad_norm": 5.53125, + "learning_rate": 9.980095365836753e-06, + "loss": 1.64242783, + "memory(GiB)": 71.21, + "step": 6070, + "train_speed(iter/s)": 0.451009 + }, + { + "acc": 0.63862944, + "epoch": 0.1541095890410959, + "grad_norm": 5.90625, + "learning_rate": 9.980001781609353e-06, + "loss": 1.72136669, + "memory(GiB)": 71.21, + "step": 6075, + "train_speed(iter/s)": 0.451208 + }, + { + "acc": 0.62790833, + "epoch": 0.15423642820903094, + "grad_norm": 5.96875, + "learning_rate": 9.979907978339236e-06, + "loss": 1.722258, + "memory(GiB)": 71.21, + "step": 6080, + "train_speed(iter/s)": 0.451406 + }, + { + "acc": 0.6412046, + "epoch": 0.154363267376966, + "grad_norm": 5.96875, + "learning_rate": 9.979813956030535e-06, + "loss": 1.66379547, + "memory(GiB)": 71.21, + "step": 6085, + "train_speed(iter/s)": 0.451604 + }, + { + "acc": 0.63239679, + "epoch": 0.15449010654490106, + "grad_norm": 5.78125, + "learning_rate": 9.979719714687384e-06, + "loss": 1.70074711, + "memory(GiB)": 71.21, + "step": 6090, + "train_speed(iter/s)": 0.451802 + }, + { + "acc": 0.63261719, + "epoch": 0.15461694571283613, + "grad_norm": 4.9375, + "learning_rate": 9.979625254313924e-06, + "loss": 1.79346123, + "memory(GiB)": 71.21, + "step": 6095, + "train_speed(iter/s)": 0.452 + }, + { + "acc": 0.63210621, + "epoch": 0.15474378488077117, + "grad_norm": 6.65625, + "learning_rate": 9.979530574914316e-06, + "loss": 1.7295742, + "memory(GiB)": 71.21, + "step": 6100, + "train_speed(iter/s)": 0.452199 + }, + { + "epoch": 0.15474378488077117, + "eval_acc": 0.6294165837746697, + "eval_loss": 1.6914069652557373, + "eval_runtime": 113.5501, + "eval_samples_per_second": 56.099, + "eval_steps_per_second": 28.049, + "step": 6100 + }, + { + "acc": 0.63916278, + "epoch": 0.15487062404870625, + "grad_norm": 6.40625, + "learning_rate": 9.97943567649272e-06, + "loss": 1.68640079, + "memory(GiB)": 71.21, + "step": 6105, + "train_speed(iter/s)": 0.448473 + }, + { + "acc": 0.62674141, + "epoch": 0.1549974632166413, + "grad_norm": 5.125, + "learning_rate": 9.979340559053311e-06, + "loss": 1.65725746, + "memory(GiB)": 71.21, + "step": 6110, + "train_speed(iter/s)": 0.448668 + }, + { + "acc": 0.63233347, + "epoch": 0.15512430238457636, + "grad_norm": 4.34375, + "learning_rate": 9.979245222600273e-06, + "loss": 1.7305582, + "memory(GiB)": 71.21, + "step": 6115, + "train_speed(iter/s)": 0.448864 + }, + { + "acc": 0.62707701, + "epoch": 0.1552511415525114, + "grad_norm": 4.9375, + "learning_rate": 9.979149667137801e-06, + "loss": 1.77526169, + "memory(GiB)": 71.21, + "step": 6120, + "train_speed(iter/s)": 0.449061 + }, + { + "acc": 0.62613888, + "epoch": 0.15537798072044648, + "grad_norm": 5.46875, + "learning_rate": 9.979053892670094e-06, + "loss": 1.76586418, + "memory(GiB)": 71.21, + "step": 6125, + "train_speed(iter/s)": 0.449259 + }, + { + "acc": 0.6409863, + "epoch": 0.15550481988838152, + "grad_norm": 4.59375, + "learning_rate": 9.978957899201369e-06, + "loss": 1.75378647, + "memory(GiB)": 71.21, + "step": 6130, + "train_speed(iter/s)": 0.449455 + }, + { + "acc": 0.64654946, + "epoch": 0.1556316590563166, + "grad_norm": 4.65625, + "learning_rate": 9.978861686735845e-06, + "loss": 1.67361908, + "memory(GiB)": 71.21, + "step": 6135, + "train_speed(iter/s)": 0.449652 + }, + { + "acc": 0.6318131, + "epoch": 0.15575849822425164, + "grad_norm": 7.34375, + "learning_rate": 9.978765255277756e-06, + "loss": 1.72088928, + "memory(GiB)": 71.21, + "step": 6140, + "train_speed(iter/s)": 0.449848 + }, + { + "acc": 0.63990841, + "epoch": 0.1558853373921867, + "grad_norm": 4.90625, + "learning_rate": 9.97866860483134e-06, + "loss": 1.71816597, + "memory(GiB)": 71.21, + "step": 6145, + "train_speed(iter/s)": 0.450044 + }, + { + "acc": 0.65172644, + "epoch": 0.15601217656012176, + "grad_norm": 5.46875, + "learning_rate": 9.978571735400853e-06, + "loss": 1.69172668, + "memory(GiB)": 71.21, + "step": 6150, + "train_speed(iter/s)": 0.450238 + }, + { + "acc": 0.62753701, + "epoch": 0.15613901572805683, + "grad_norm": 6.0, + "learning_rate": 9.978474646990552e-06, + "loss": 1.68023739, + "memory(GiB)": 71.21, + "step": 6155, + "train_speed(iter/s)": 0.450433 + }, + { + "acc": 0.63415184, + "epoch": 0.15626585489599187, + "grad_norm": 5.65625, + "learning_rate": 9.97837733960471e-06, + "loss": 1.73192654, + "memory(GiB)": 71.21, + "step": 6160, + "train_speed(iter/s)": 0.450627 + }, + { + "acc": 0.65774403, + "epoch": 0.15639269406392695, + "grad_norm": 6.46875, + "learning_rate": 9.978279813247605e-06, + "loss": 1.67506714, + "memory(GiB)": 71.21, + "step": 6165, + "train_speed(iter/s)": 0.450823 + }, + { + "acc": 0.64377122, + "epoch": 0.156519533231862, + "grad_norm": 4.84375, + "learning_rate": 9.978182067923528e-06, + "loss": 1.66583214, + "memory(GiB)": 71.21, + "step": 6170, + "train_speed(iter/s)": 0.451018 + }, + { + "acc": 0.62760563, + "epoch": 0.15664637239979706, + "grad_norm": 4.46875, + "learning_rate": 9.978084103636778e-06, + "loss": 1.80464134, + "memory(GiB)": 71.21, + "step": 6175, + "train_speed(iter/s)": 0.451212 + }, + { + "acc": 0.65065422, + "epoch": 0.1567732115677321, + "grad_norm": 5.90625, + "learning_rate": 9.977985920391661e-06, + "loss": 1.6794405, + "memory(GiB)": 71.21, + "step": 6180, + "train_speed(iter/s)": 0.451407 + }, + { + "acc": 0.63304777, + "epoch": 0.15690005073566718, + "grad_norm": 5.375, + "learning_rate": 9.977887518192501e-06, + "loss": 1.77069626, + "memory(GiB)": 71.21, + "step": 6185, + "train_speed(iter/s)": 0.451603 + }, + { + "acc": 0.61158085, + "epoch": 0.15702688990360222, + "grad_norm": 4.40625, + "learning_rate": 9.977788897043622e-06, + "loss": 1.80420361, + "memory(GiB)": 71.21, + "step": 6190, + "train_speed(iter/s)": 0.451798 + }, + { + "acc": 0.64230924, + "epoch": 0.1571537290715373, + "grad_norm": 5.84375, + "learning_rate": 9.977690056949363e-06, + "loss": 1.68551178, + "memory(GiB)": 71.21, + "step": 6195, + "train_speed(iter/s)": 0.451992 + }, + { + "acc": 0.63077068, + "epoch": 0.15728056823947234, + "grad_norm": 6.59375, + "learning_rate": 9.977590997914072e-06, + "loss": 1.75044956, + "memory(GiB)": 71.21, + "step": 6200, + "train_speed(iter/s)": 0.452185 + }, + { + "epoch": 0.15728056823947234, + "eval_acc": 0.6295770064215013, + "eval_loss": 1.6895726919174194, + "eval_runtime": 113.4251, + "eval_samples_per_second": 56.16, + "eval_steps_per_second": 28.08, + "step": 6200 + }, + { + "acc": 0.63319101, + "epoch": 0.1574074074074074, + "grad_norm": 5.09375, + "learning_rate": 9.977491719942106e-06, + "loss": 1.69404125, + "memory(GiB)": 71.21, + "step": 6205, + "train_speed(iter/s)": 0.448521 + }, + { + "acc": 0.63302393, + "epoch": 0.15753424657534246, + "grad_norm": 5.625, + "learning_rate": 9.97739222303783e-06, + "loss": 1.74673462, + "memory(GiB)": 71.21, + "step": 6210, + "train_speed(iter/s)": 0.448714 + }, + { + "acc": 0.64599352, + "epoch": 0.15766108574327753, + "grad_norm": 5.46875, + "learning_rate": 9.977292507205623e-06, + "loss": 1.65102501, + "memory(GiB)": 71.21, + "step": 6215, + "train_speed(iter/s)": 0.448907 + }, + { + "acc": 0.61004953, + "epoch": 0.15778792491121257, + "grad_norm": 5.8125, + "learning_rate": 9.977192572449868e-06, + "loss": 1.81020546, + "memory(GiB)": 71.21, + "step": 6220, + "train_speed(iter/s)": 0.4491 + }, + { + "acc": 0.63879442, + "epoch": 0.15791476407914765, + "grad_norm": 6.1875, + "learning_rate": 9.977092418774962e-06, + "loss": 1.73230076, + "memory(GiB)": 71.21, + "step": 6225, + "train_speed(iter/s)": 0.449294 + }, + { + "acc": 0.61847792, + "epoch": 0.1580416032470827, + "grad_norm": 5.78125, + "learning_rate": 9.976992046185313e-06, + "loss": 1.79007607, + "memory(GiB)": 71.21, + "step": 6230, + "train_speed(iter/s)": 0.449486 + }, + { + "acc": 0.63165073, + "epoch": 0.15816844241501776, + "grad_norm": 4.875, + "learning_rate": 9.97689145468533e-06, + "loss": 1.74281464, + "memory(GiB)": 71.21, + "step": 6235, + "train_speed(iter/s)": 0.449679 + }, + { + "acc": 0.64975591, + "epoch": 0.1582952815829528, + "grad_norm": 5.5625, + "learning_rate": 9.976790644279442e-06, + "loss": 1.64184265, + "memory(GiB)": 71.21, + "step": 6240, + "train_speed(iter/s)": 0.449872 + }, + { + "acc": 0.6411006, + "epoch": 0.15842212075088788, + "grad_norm": 5.28125, + "learning_rate": 9.976689614972082e-06, + "loss": 1.64428825, + "memory(GiB)": 71.21, + "step": 6245, + "train_speed(iter/s)": 0.450063 + }, + { + "acc": 0.640096, + "epoch": 0.15854895991882292, + "grad_norm": 7.15625, + "learning_rate": 9.976588366767693e-06, + "loss": 1.73388939, + "memory(GiB)": 71.21, + "step": 6250, + "train_speed(iter/s)": 0.450257 + }, + { + "acc": 0.63372641, + "epoch": 0.158675799086758, + "grad_norm": 5.90625, + "learning_rate": 9.976486899670729e-06, + "loss": 1.67146568, + "memory(GiB)": 71.21, + "step": 6255, + "train_speed(iter/s)": 0.450449 + }, + { + "acc": 0.63916836, + "epoch": 0.15880263825469304, + "grad_norm": 5.46875, + "learning_rate": 9.976385213685652e-06, + "loss": 1.71383762, + "memory(GiB)": 71.21, + "step": 6260, + "train_speed(iter/s)": 0.450641 + }, + { + "acc": 0.63503208, + "epoch": 0.1589294774226281, + "grad_norm": 5.59375, + "learning_rate": 9.976283308816937e-06, + "loss": 1.77095871, + "memory(GiB)": 71.21, + "step": 6265, + "train_speed(iter/s)": 0.450836 + }, + { + "acc": 0.64217548, + "epoch": 0.15905631659056316, + "grad_norm": 6.03125, + "learning_rate": 9.976181185069063e-06, + "loss": 1.6796814, + "memory(GiB)": 71.21, + "step": 6270, + "train_speed(iter/s)": 0.45103 + }, + { + "acc": 0.65000076, + "epoch": 0.15918315575849823, + "grad_norm": 4.15625, + "learning_rate": 9.976078842446522e-06, + "loss": 1.67571526, + "memory(GiB)": 71.21, + "step": 6275, + "train_speed(iter/s)": 0.451218 + }, + { + "acc": 0.63133268, + "epoch": 0.15930999492643327, + "grad_norm": 4.8125, + "learning_rate": 9.97597628095382e-06, + "loss": 1.83774338, + "memory(GiB)": 71.21, + "step": 6280, + "train_speed(iter/s)": 0.451409 + }, + { + "acc": 0.62406836, + "epoch": 0.15943683409436835, + "grad_norm": 5.59375, + "learning_rate": 9.975873500595464e-06, + "loss": 1.8288662, + "memory(GiB)": 71.21, + "step": 6285, + "train_speed(iter/s)": 0.451601 + }, + { + "acc": 0.6443151, + "epoch": 0.1595636732623034, + "grad_norm": 5.875, + "learning_rate": 9.975770501375974e-06, + "loss": 1.73548775, + "memory(GiB)": 71.21, + "step": 6290, + "train_speed(iter/s)": 0.45179 + }, + { + "acc": 0.63668132, + "epoch": 0.15969051243023846, + "grad_norm": 5.28125, + "learning_rate": 9.975667283299884e-06, + "loss": 1.67912426, + "memory(GiB)": 71.21, + "step": 6295, + "train_speed(iter/s)": 0.451982 + }, + { + "acc": 0.64422007, + "epoch": 0.1598173515981735, + "grad_norm": 6.28125, + "learning_rate": 9.975563846371732e-06, + "loss": 1.71651382, + "memory(GiB)": 71.21, + "step": 6300, + "train_speed(iter/s)": 0.452174 + }, + { + "epoch": 0.1598173515981735, + "eval_acc": 0.629977645271271, + "eval_loss": 1.6877862215042114, + "eval_runtime": 112.9937, + "eval_samples_per_second": 56.375, + "eval_steps_per_second": 28.187, + "step": 6300 + }, + { + "acc": 0.63653145, + "epoch": 0.15994419076610858, + "grad_norm": 4.90625, + "learning_rate": 9.975460190596068e-06, + "loss": 1.71957417, + "memory(GiB)": 71.21, + "step": 6305, + "train_speed(iter/s)": 0.448582 + }, + { + "acc": 0.63068991, + "epoch": 0.16007102993404362, + "grad_norm": 6.1875, + "learning_rate": 9.975356315977451e-06, + "loss": 1.7084877, + "memory(GiB)": 71.21, + "step": 6310, + "train_speed(iter/s)": 0.448773 + }, + { + "acc": 0.63710327, + "epoch": 0.1601978691019787, + "grad_norm": 4.9375, + "learning_rate": 9.975252222520449e-06, + "loss": 1.70445976, + "memory(GiB)": 71.21, + "step": 6315, + "train_speed(iter/s)": 0.448961 + }, + { + "acc": 0.63732691, + "epoch": 0.16032470826991374, + "grad_norm": 4.6875, + "learning_rate": 9.97514791022964e-06, + "loss": 1.71315804, + "memory(GiB)": 71.21, + "step": 6320, + "train_speed(iter/s)": 0.449151 + }, + { + "acc": 0.6422514, + "epoch": 0.1604515474378488, + "grad_norm": 5.0625, + "learning_rate": 9.975043379109617e-06, + "loss": 1.71896667, + "memory(GiB)": 71.21, + "step": 6325, + "train_speed(iter/s)": 0.449342 + }, + { + "acc": 0.64499207, + "epoch": 0.16057838660578386, + "grad_norm": 4.8125, + "learning_rate": 9.974938629164973e-06, + "loss": 1.692416, + "memory(GiB)": 71.21, + "step": 6330, + "train_speed(iter/s)": 0.44953 + }, + { + "acc": 0.63671074, + "epoch": 0.16070522577371893, + "grad_norm": 4.4375, + "learning_rate": 9.974833660400315e-06, + "loss": 1.73145828, + "memory(GiB)": 71.21, + "step": 6335, + "train_speed(iter/s)": 0.44972 + }, + { + "acc": 0.62892737, + "epoch": 0.16083206494165397, + "grad_norm": 6.1875, + "learning_rate": 9.974728472820264e-06, + "loss": 1.71347389, + "memory(GiB)": 71.21, + "step": 6340, + "train_speed(iter/s)": 0.449909 + }, + { + "acc": 0.63510127, + "epoch": 0.16095890410958905, + "grad_norm": 4.3125, + "learning_rate": 9.97462306642944e-06, + "loss": 1.72145348, + "memory(GiB)": 71.21, + "step": 6345, + "train_speed(iter/s)": 0.450096 + }, + { + "acc": 0.65161772, + "epoch": 0.1610857432775241, + "grad_norm": 4.65625, + "learning_rate": 9.974517441232487e-06, + "loss": 1.61755943, + "memory(GiB)": 71.21, + "step": 6350, + "train_speed(iter/s)": 0.450286 + }, + { + "acc": 0.63160658, + "epoch": 0.16121258244545916, + "grad_norm": 5.28125, + "learning_rate": 9.974411597234046e-06, + "loss": 1.73004341, + "memory(GiB)": 71.21, + "step": 6355, + "train_speed(iter/s)": 0.450475 + }, + { + "acc": 0.63577652, + "epoch": 0.1613394216133942, + "grad_norm": 6.09375, + "learning_rate": 9.974305534438774e-06, + "loss": 1.66141491, + "memory(GiB)": 71.21, + "step": 6360, + "train_speed(iter/s)": 0.450664 + }, + { + "acc": 0.64566479, + "epoch": 0.16146626078132928, + "grad_norm": 4.75, + "learning_rate": 9.974199252851338e-06, + "loss": 1.70336876, + "memory(GiB)": 71.21, + "step": 6365, + "train_speed(iter/s)": 0.450854 + }, + { + "acc": 0.64617376, + "epoch": 0.16159309994926432, + "grad_norm": 5.34375, + "learning_rate": 9.974092752476408e-06, + "loss": 1.66555977, + "memory(GiB)": 71.21, + "step": 6370, + "train_speed(iter/s)": 0.451043 + }, + { + "acc": 0.65135059, + "epoch": 0.1617199391171994, + "grad_norm": 4.875, + "learning_rate": 9.973986033318673e-06, + "loss": 1.71009521, + "memory(GiB)": 71.21, + "step": 6375, + "train_speed(iter/s)": 0.451231 + }, + { + "acc": 0.65685863, + "epoch": 0.16184677828513444, + "grad_norm": 5.03125, + "learning_rate": 9.973879095382824e-06, + "loss": 1.62660656, + "memory(GiB)": 71.21, + "step": 6380, + "train_speed(iter/s)": 0.451419 + }, + { + "acc": 0.64332199, + "epoch": 0.1619736174530695, + "grad_norm": 6.1875, + "learning_rate": 9.973771938673564e-06, + "loss": 1.71407089, + "memory(GiB)": 71.21, + "step": 6385, + "train_speed(iter/s)": 0.451608 + }, + { + "acc": 0.63628197, + "epoch": 0.16210045662100456, + "grad_norm": 6.09375, + "learning_rate": 9.973664563195609e-06, + "loss": 1.76624718, + "memory(GiB)": 71.21, + "step": 6390, + "train_speed(iter/s)": 0.451796 + }, + { + "acc": 0.63456173, + "epoch": 0.16222729578893963, + "grad_norm": 5.96875, + "learning_rate": 9.973556968953682e-06, + "loss": 1.72168999, + "memory(GiB)": 71.21, + "step": 6395, + "train_speed(iter/s)": 0.451984 + }, + { + "acc": 0.63222179, + "epoch": 0.16235413495687467, + "grad_norm": 5.34375, + "learning_rate": 9.973449155952512e-06, + "loss": 1.70554523, + "memory(GiB)": 71.21, + "step": 6400, + "train_speed(iter/s)": 0.452173 + }, + { + "epoch": 0.16235413495687467, + "eval_acc": 0.6302111771972576, + "eval_loss": 1.6864054203033447, + "eval_runtime": 115.1051, + "eval_samples_per_second": 55.341, + "eval_steps_per_second": 27.67, + "step": 6400 + }, + { + "acc": 0.63519077, + "epoch": 0.16248097412480975, + "grad_norm": 5.15625, + "learning_rate": 9.973341124196847e-06, + "loss": 1.67466583, + "memory(GiB)": 71.21, + "step": 6405, + "train_speed(iter/s)": 0.448572 + }, + { + "acc": 0.65170107, + "epoch": 0.1626078132927448, + "grad_norm": 4.875, + "learning_rate": 9.973232873691431e-06, + "loss": 1.66719913, + "memory(GiB)": 71.21, + "step": 6410, + "train_speed(iter/s)": 0.448758 + }, + { + "acc": 0.63416324, + "epoch": 0.16273465246067986, + "grad_norm": 4.71875, + "learning_rate": 9.973124404441031e-06, + "loss": 1.72526855, + "memory(GiB)": 71.21, + "step": 6415, + "train_speed(iter/s)": 0.448944 + }, + { + "acc": 0.62481432, + "epoch": 0.1628614916286149, + "grad_norm": 5.375, + "learning_rate": 9.973015716450416e-06, + "loss": 1.82407475, + "memory(GiB)": 71.21, + "step": 6420, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.64744358, + "epoch": 0.16298833079654998, + "grad_norm": 5.5, + "learning_rate": 9.972906809724367e-06, + "loss": 1.66186638, + "memory(GiB)": 71.21, + "step": 6425, + "train_speed(iter/s)": 0.449317 + }, + { + "acc": 0.6531373, + "epoch": 0.16311516996448502, + "grad_norm": 5.46875, + "learning_rate": 9.972797684267674e-06, + "loss": 1.64693489, + "memory(GiB)": 71.21, + "step": 6430, + "train_speed(iter/s)": 0.449503 + }, + { + "acc": 0.6519381, + "epoch": 0.1632420091324201, + "grad_norm": 5.0, + "learning_rate": 9.972688340085137e-06, + "loss": 1.63905487, + "memory(GiB)": 71.21, + "step": 6435, + "train_speed(iter/s)": 0.44969 + }, + { + "acc": 0.64304528, + "epoch": 0.16336884830035514, + "grad_norm": 6.3125, + "learning_rate": 9.972578777181565e-06, + "loss": 1.72123661, + "memory(GiB)": 71.21, + "step": 6440, + "train_speed(iter/s)": 0.449877 + }, + { + "acc": 0.63808084, + "epoch": 0.1634956874682902, + "grad_norm": 5.6875, + "learning_rate": 9.972468995561778e-06, + "loss": 1.64513397, + "memory(GiB)": 71.21, + "step": 6445, + "train_speed(iter/s)": 0.450065 + }, + { + "acc": 0.64115577, + "epoch": 0.16362252663622526, + "grad_norm": 5.4375, + "learning_rate": 9.972358995230604e-06, + "loss": 1.60937214, + "memory(GiB)": 71.21, + "step": 6450, + "train_speed(iter/s)": 0.450251 + }, + { + "acc": 0.6481483, + "epoch": 0.16374936580416033, + "grad_norm": 6.28125, + "learning_rate": 9.97224877619288e-06, + "loss": 1.69629669, + "memory(GiB)": 71.21, + "step": 6455, + "train_speed(iter/s)": 0.450438 + }, + { + "acc": 0.63374949, + "epoch": 0.16387620497209537, + "grad_norm": 5.03125, + "learning_rate": 9.972138338453457e-06, + "loss": 1.74168835, + "memory(GiB)": 71.21, + "step": 6460, + "train_speed(iter/s)": 0.450625 + }, + { + "acc": 0.63819251, + "epoch": 0.16400304414003045, + "grad_norm": 4.75, + "learning_rate": 9.972027682017191e-06, + "loss": 1.70580139, + "memory(GiB)": 71.21, + "step": 6465, + "train_speed(iter/s)": 0.45081 + }, + { + "acc": 0.62667751, + "epoch": 0.1641298833079655, + "grad_norm": 4.4375, + "learning_rate": 9.971916806888948e-06, + "loss": 1.79486923, + "memory(GiB)": 71.21, + "step": 6470, + "train_speed(iter/s)": 0.450996 + }, + { + "acc": 0.63105898, + "epoch": 0.16425672247590056, + "grad_norm": 6.0, + "learning_rate": 9.971805713073606e-06, + "loss": 1.76685257, + "memory(GiB)": 71.21, + "step": 6475, + "train_speed(iter/s)": 0.451183 + }, + { + "acc": 0.63640289, + "epoch": 0.1643835616438356, + "grad_norm": 4.625, + "learning_rate": 9.971694400576053e-06, + "loss": 1.65031548, + "memory(GiB)": 71.21, + "step": 6480, + "train_speed(iter/s)": 0.451368 + }, + { + "acc": 0.64500246, + "epoch": 0.16451040081177068, + "grad_norm": 4.53125, + "learning_rate": 9.971582869401182e-06, + "loss": 1.69581451, + "memory(GiB)": 71.21, + "step": 6485, + "train_speed(iter/s)": 0.451552 + }, + { + "acc": 0.64254198, + "epoch": 0.16463723997970572, + "grad_norm": 4.78125, + "learning_rate": 9.9714711195539e-06, + "loss": 1.71236401, + "memory(GiB)": 71.21, + "step": 6490, + "train_speed(iter/s)": 0.451738 + }, + { + "acc": 0.63174195, + "epoch": 0.1647640791476408, + "grad_norm": 6.65625, + "learning_rate": 9.97135915103912e-06, + "loss": 1.71433392, + "memory(GiB)": 71.21, + "step": 6495, + "train_speed(iter/s)": 0.451925 + }, + { + "acc": 0.63656821, + "epoch": 0.16489091831557584, + "grad_norm": 5.40625, + "learning_rate": 9.971246963861772e-06, + "loss": 1.6948307, + "memory(GiB)": 71.21, + "step": 6500, + "train_speed(iter/s)": 0.452111 + }, + { + "epoch": 0.16489091831557584, + "eval_acc": 0.6304238207577714, + "eval_loss": 1.6844685077667236, + "eval_runtime": 115.3889, + "eval_samples_per_second": 55.205, + "eval_steps_per_second": 27.602, + "step": 6500 + }, + { + "acc": 0.63033233, + "epoch": 0.1650177574835109, + "grad_norm": 4.78125, + "learning_rate": 9.971134558026786e-06, + "loss": 1.69573898, + "memory(GiB)": 71.21, + "step": 6505, + "train_speed(iter/s)": 0.448555 + }, + { + "acc": 0.64669685, + "epoch": 0.16514459665144596, + "grad_norm": 5.375, + "learning_rate": 9.971021933539108e-06, + "loss": 1.66779327, + "memory(GiB)": 71.21, + "step": 6510, + "train_speed(iter/s)": 0.448739 + }, + { + "acc": 0.64687338, + "epoch": 0.16527143581938103, + "grad_norm": 5.8125, + "learning_rate": 9.97090909040369e-06, + "loss": 1.66354771, + "memory(GiB)": 71.21, + "step": 6515, + "train_speed(iter/s)": 0.448925 + }, + { + "acc": 0.63365822, + "epoch": 0.16539827498731607, + "grad_norm": 4.75, + "learning_rate": 9.970796028625499e-06, + "loss": 1.73496552, + "memory(GiB)": 71.21, + "step": 6520, + "train_speed(iter/s)": 0.449109 + }, + { + "acc": 0.63935966, + "epoch": 0.16552511415525115, + "grad_norm": 5.375, + "learning_rate": 9.970682748209505e-06, + "loss": 1.67517166, + "memory(GiB)": 71.21, + "step": 6525, + "train_speed(iter/s)": 0.449295 + }, + { + "acc": 0.64406037, + "epoch": 0.1656519533231862, + "grad_norm": 4.1875, + "learning_rate": 9.97056924916069e-06, + "loss": 1.67376709, + "memory(GiB)": 71.21, + "step": 6530, + "train_speed(iter/s)": 0.449479 + }, + { + "acc": 0.65336609, + "epoch": 0.16577879249112126, + "grad_norm": 5.0625, + "learning_rate": 9.970455531484049e-06, + "loss": 1.65930328, + "memory(GiB)": 71.21, + "step": 6535, + "train_speed(iter/s)": 0.44966 + }, + { + "acc": 0.6460598, + "epoch": 0.1659056316590563, + "grad_norm": 5.375, + "learning_rate": 9.97034159518458e-06, + "loss": 1.68185635, + "memory(GiB)": 71.21, + "step": 6540, + "train_speed(iter/s)": 0.449844 + }, + { + "acc": 0.64707355, + "epoch": 0.16603247082699138, + "grad_norm": 5.96875, + "learning_rate": 9.9702274402673e-06, + "loss": 1.6996788, + "memory(GiB)": 71.21, + "step": 6545, + "train_speed(iter/s)": 0.450029 + }, + { + "acc": 0.6494092, + "epoch": 0.16615930999492642, + "grad_norm": 5.1875, + "learning_rate": 9.970113066737223e-06, + "loss": 1.7148674, + "memory(GiB)": 71.21, + "step": 6550, + "train_speed(iter/s)": 0.450211 + }, + { + "acc": 0.6416522, + "epoch": 0.1662861491628615, + "grad_norm": 5.40625, + "learning_rate": 9.969998474599386e-06, + "loss": 1.68036804, + "memory(GiB)": 71.21, + "step": 6555, + "train_speed(iter/s)": 0.450394 + }, + { + "acc": 0.64480743, + "epoch": 0.16641298833079654, + "grad_norm": 6.28125, + "learning_rate": 9.969883663858826e-06, + "loss": 1.63234901, + "memory(GiB)": 71.21, + "step": 6560, + "train_speed(iter/s)": 0.450579 + }, + { + "acc": 0.65155292, + "epoch": 0.1665398274987316, + "grad_norm": 5.90625, + "learning_rate": 9.969768634520593e-06, + "loss": 1.68875046, + "memory(GiB)": 71.21, + "step": 6565, + "train_speed(iter/s)": 0.450763 + }, + { + "acc": 0.65460062, + "epoch": 0.16666666666666666, + "grad_norm": 4.5625, + "learning_rate": 9.969653386589749e-06, + "loss": 1.62505722, + "memory(GiB)": 71.21, + "step": 6570, + "train_speed(iter/s)": 0.450946 + }, + { + "acc": 0.65035658, + "epoch": 0.16679350583460173, + "grad_norm": 5.0625, + "learning_rate": 9.96953792007136e-06, + "loss": 1.70844345, + "memory(GiB)": 71.21, + "step": 6575, + "train_speed(iter/s)": 0.451129 + }, + { + "acc": 0.62463398, + "epoch": 0.16692034500253677, + "grad_norm": 4.625, + "learning_rate": 9.969422234970506e-06, + "loss": 1.72562618, + "memory(GiB)": 71.21, + "step": 6580, + "train_speed(iter/s)": 0.451311 + }, + { + "acc": 0.6577569, + "epoch": 0.16704718417047185, + "grad_norm": 4.84375, + "learning_rate": 9.969306331292273e-06, + "loss": 1.60892105, + "memory(GiB)": 71.21, + "step": 6585, + "train_speed(iter/s)": 0.451492 + }, + { + "acc": 0.62938375, + "epoch": 0.1671740233384069, + "grad_norm": 5.53125, + "learning_rate": 9.969190209041764e-06, + "loss": 1.75925503, + "memory(GiB)": 71.21, + "step": 6590, + "train_speed(iter/s)": 0.451675 + }, + { + "acc": 0.63579378, + "epoch": 0.16730086250634196, + "grad_norm": 5.9375, + "learning_rate": 9.969073868224082e-06, + "loss": 1.74828606, + "memory(GiB)": 71.21, + "step": 6595, + "train_speed(iter/s)": 0.451856 + }, + { + "acc": 0.62777534, + "epoch": 0.167427701674277, + "grad_norm": 4.90625, + "learning_rate": 9.968957308844346e-06, + "loss": 1.78514938, + "memory(GiB)": 71.21, + "step": 6600, + "train_speed(iter/s)": 0.45204 + }, + { + "epoch": 0.167427701674277, + "eval_acc": 0.6306757344453742, + "eval_loss": 1.6830366849899292, + "eval_runtime": 114.5252, + "eval_samples_per_second": 55.621, + "eval_steps_per_second": 27.81, + "step": 6600 + }, + { + "acc": 0.63884621, + "epoch": 0.16755454084221208, + "grad_norm": 6.15625, + "learning_rate": 9.968840530907684e-06, + "loss": 1.74698486, + "memory(GiB)": 71.21, + "step": 6605, + "train_speed(iter/s)": 0.448567 + }, + { + "acc": 0.62307458, + "epoch": 0.16768138001014712, + "grad_norm": 6.6875, + "learning_rate": 9.96872353441923e-06, + "loss": 1.74410114, + "memory(GiB)": 71.21, + "step": 6610, + "train_speed(iter/s)": 0.448749 + }, + { + "acc": 0.63508086, + "epoch": 0.1678082191780822, + "grad_norm": 6.25, + "learning_rate": 9.968606319384131e-06, + "loss": 1.75212784, + "memory(GiB)": 71.21, + "step": 6615, + "train_speed(iter/s)": 0.44893 + }, + { + "acc": 0.6378727, + "epoch": 0.16793505834601724, + "grad_norm": 6.84375, + "learning_rate": 9.968488885807544e-06, + "loss": 1.76319771, + "memory(GiB)": 71.21, + "step": 6620, + "train_speed(iter/s)": 0.449112 + }, + { + "acc": 0.64346189, + "epoch": 0.1680618975139523, + "grad_norm": 4.65625, + "learning_rate": 9.968371233694633e-06, + "loss": 1.7074379, + "memory(GiB)": 71.21, + "step": 6625, + "train_speed(iter/s)": 0.449294 + }, + { + "acc": 0.64173932, + "epoch": 0.16818873668188736, + "grad_norm": 5.46875, + "learning_rate": 9.968253363050573e-06, + "loss": 1.71890678, + "memory(GiB)": 71.21, + "step": 6630, + "train_speed(iter/s)": 0.449474 + }, + { + "acc": 0.64332647, + "epoch": 0.16831557584982243, + "grad_norm": 4.34375, + "learning_rate": 9.968135273880547e-06, + "loss": 1.65478706, + "memory(GiB)": 71.21, + "step": 6635, + "train_speed(iter/s)": 0.449655 + }, + { + "acc": 0.6396049, + "epoch": 0.16844241501775747, + "grad_norm": 5.59375, + "learning_rate": 9.968016966189753e-06, + "loss": 1.66964741, + "memory(GiB)": 71.21, + "step": 6640, + "train_speed(iter/s)": 0.449836 + }, + { + "acc": 0.64336157, + "epoch": 0.16856925418569255, + "grad_norm": 5.96875, + "learning_rate": 9.96789843998339e-06, + "loss": 1.63839188, + "memory(GiB)": 71.21, + "step": 6645, + "train_speed(iter/s)": 0.450017 + }, + { + "acc": 0.63814754, + "epoch": 0.1686960933536276, + "grad_norm": 6.25, + "learning_rate": 9.967779695266675e-06, + "loss": 1.70653496, + "memory(GiB)": 71.21, + "step": 6650, + "train_speed(iter/s)": 0.450199 + }, + { + "acc": 0.64745874, + "epoch": 0.16882293252156266, + "grad_norm": 6.21875, + "learning_rate": 9.967660732044828e-06, + "loss": 1.69663105, + "memory(GiB)": 71.21, + "step": 6655, + "train_speed(iter/s)": 0.450379 + }, + { + "acc": 0.64206862, + "epoch": 0.1689497716894977, + "grad_norm": 4.875, + "learning_rate": 9.967541550323085e-06, + "loss": 1.6930912, + "memory(GiB)": 71.21, + "step": 6660, + "train_speed(iter/s)": 0.45056 + }, + { + "acc": 0.64384375, + "epoch": 0.16907661085743278, + "grad_norm": 5.53125, + "learning_rate": 9.967422150106685e-06, + "loss": 1.70717201, + "memory(GiB)": 71.21, + "step": 6665, + "train_speed(iter/s)": 0.450742 + }, + { + "acc": 0.63365135, + "epoch": 0.16920345002536782, + "grad_norm": 4.875, + "learning_rate": 9.96730253140088e-06, + "loss": 1.6828104, + "memory(GiB)": 71.21, + "step": 6670, + "train_speed(iter/s)": 0.450923 + }, + { + "acc": 0.63285046, + "epoch": 0.1693302891933029, + "grad_norm": 5.5, + "learning_rate": 9.967182694210933e-06, + "loss": 1.74362907, + "memory(GiB)": 71.21, + "step": 6675, + "train_speed(iter/s)": 0.451104 + }, + { + "acc": 0.64584789, + "epoch": 0.16945712836123794, + "grad_norm": 4.625, + "learning_rate": 9.967062638542116e-06, + "loss": 1.6860466, + "memory(GiB)": 71.21, + "step": 6680, + "train_speed(iter/s)": 0.451285 + }, + { + "acc": 0.63993616, + "epoch": 0.169583967529173, + "grad_norm": 5.46875, + "learning_rate": 9.966942364399706e-06, + "loss": 1.70842514, + "memory(GiB)": 71.21, + "step": 6685, + "train_speed(iter/s)": 0.451466 + }, + { + "acc": 0.63192701, + "epoch": 0.16971080669710806, + "grad_norm": 5.21875, + "learning_rate": 9.966821871788995e-06, + "loss": 1.73490124, + "memory(GiB)": 71.21, + "step": 6690, + "train_speed(iter/s)": 0.451646 + }, + { + "acc": 0.64266176, + "epoch": 0.16983764586504313, + "grad_norm": 6.09375, + "learning_rate": 9.966701160715283e-06, + "loss": 1.71660919, + "memory(GiB)": 71.21, + "step": 6695, + "train_speed(iter/s)": 0.451827 + }, + { + "acc": 0.65317831, + "epoch": 0.16996448503297817, + "grad_norm": 5.40625, + "learning_rate": 9.96658023118388e-06, + "loss": 1.63388596, + "memory(GiB)": 71.21, + "step": 6700, + "train_speed(iter/s)": 0.452006 + }, + { + "epoch": 0.16996448503297817, + "eval_acc": 0.6308257129094694, + "eval_loss": 1.6815400123596191, + "eval_runtime": 113.4152, + "eval_samples_per_second": 56.165, + "eval_steps_per_second": 28.083, + "step": 6700 + }, + { + "acc": 0.64241967, + "epoch": 0.17009132420091325, + "grad_norm": 6.0, + "learning_rate": 9.966459083200102e-06, + "loss": 1.73983688, + "memory(GiB)": 71.21, + "step": 6705, + "train_speed(iter/s)": 0.448615 + }, + { + "acc": 0.62719831, + "epoch": 0.1702181633688483, + "grad_norm": 4.71875, + "learning_rate": 9.966337716769283e-06, + "loss": 1.7396513, + "memory(GiB)": 71.21, + "step": 6710, + "train_speed(iter/s)": 0.44879 + }, + { + "acc": 0.63033934, + "epoch": 0.17034500253678336, + "grad_norm": 5.75, + "learning_rate": 9.966216131896755e-06, + "loss": 1.75303535, + "memory(GiB)": 71.21, + "step": 6715, + "train_speed(iter/s)": 0.448965 + }, + { + "acc": 0.62729473, + "epoch": 0.1704718417047184, + "grad_norm": 5.71875, + "learning_rate": 9.966094328587871e-06, + "loss": 1.73549404, + "memory(GiB)": 71.21, + "step": 6720, + "train_speed(iter/s)": 0.449141 + }, + { + "acc": 0.63383093, + "epoch": 0.17059868087265348, + "grad_norm": 5.125, + "learning_rate": 9.965972306847986e-06, + "loss": 1.76836853, + "memory(GiB)": 71.21, + "step": 6725, + "train_speed(iter/s)": 0.449315 + }, + { + "acc": 0.65538917, + "epoch": 0.17072552004058852, + "grad_norm": 5.65625, + "learning_rate": 9.965850066682468e-06, + "loss": 1.59685116, + "memory(GiB)": 71.21, + "step": 6730, + "train_speed(iter/s)": 0.44949 + }, + { + "acc": 0.62905073, + "epoch": 0.1708523592085236, + "grad_norm": 4.96875, + "learning_rate": 9.965727608096692e-06, + "loss": 1.69803505, + "memory(GiB)": 71.21, + "step": 6735, + "train_speed(iter/s)": 0.449667 + }, + { + "acc": 0.64299374, + "epoch": 0.17097919837645864, + "grad_norm": 4.375, + "learning_rate": 9.965604931096045e-06, + "loss": 1.7507885, + "memory(GiB)": 71.21, + "step": 6740, + "train_speed(iter/s)": 0.449843 + }, + { + "acc": 0.63420615, + "epoch": 0.1711060375443937, + "grad_norm": 5.21875, + "learning_rate": 9.965482035685925e-06, + "loss": 1.73899078, + "memory(GiB)": 71.21, + "step": 6745, + "train_speed(iter/s)": 0.450017 + }, + { + "acc": 0.64873242, + "epoch": 0.17123287671232876, + "grad_norm": 5.1875, + "learning_rate": 9.965358921871735e-06, + "loss": 1.71025467, + "memory(GiB)": 71.21, + "step": 6750, + "train_speed(iter/s)": 0.450193 + }, + { + "acc": 0.65254641, + "epoch": 0.17135971588026383, + "grad_norm": 4.96875, + "learning_rate": 9.965235589658891e-06, + "loss": 1.64431171, + "memory(GiB)": 71.21, + "step": 6755, + "train_speed(iter/s)": 0.450369 + }, + { + "acc": 0.64476228, + "epoch": 0.17148655504819887, + "grad_norm": 3.921875, + "learning_rate": 9.965112039052817e-06, + "loss": 1.69672508, + "memory(GiB)": 71.21, + "step": 6760, + "train_speed(iter/s)": 0.450543 + }, + { + "acc": 0.63649068, + "epoch": 0.17161339421613395, + "grad_norm": 5.65625, + "learning_rate": 9.964988270058948e-06, + "loss": 1.75114594, + "memory(GiB)": 71.21, + "step": 6765, + "train_speed(iter/s)": 0.450718 + }, + { + "acc": 0.63136678, + "epoch": 0.171740233384069, + "grad_norm": 4.75, + "learning_rate": 9.96486428268273e-06, + "loss": 1.6766922, + "memory(GiB)": 71.21, + "step": 6770, + "train_speed(iter/s)": 0.450892 + }, + { + "acc": 0.63834257, + "epoch": 0.17186707255200406, + "grad_norm": 5.84375, + "learning_rate": 9.964740076929612e-06, + "loss": 1.68672562, + "memory(GiB)": 71.21, + "step": 6775, + "train_speed(iter/s)": 0.451067 + }, + { + "acc": 0.62989707, + "epoch": 0.1719939117199391, + "grad_norm": 5.8125, + "learning_rate": 9.964615652805059e-06, + "loss": 1.75441475, + "memory(GiB)": 71.21, + "step": 6780, + "train_speed(iter/s)": 0.451241 + }, + { + "acc": 0.64719915, + "epoch": 0.17212075088787418, + "grad_norm": 5.78125, + "learning_rate": 9.964491010314545e-06, + "loss": 1.66278801, + "memory(GiB)": 71.21, + "step": 6785, + "train_speed(iter/s)": 0.451415 + }, + { + "acc": 0.64117589, + "epoch": 0.17224759005580922, + "grad_norm": 5.0625, + "learning_rate": 9.964366149463552e-06, + "loss": 1.71033897, + "memory(GiB)": 71.21, + "step": 6790, + "train_speed(iter/s)": 0.451589 + }, + { + "acc": 0.65882111, + "epoch": 0.1723744292237443, + "grad_norm": 4.65625, + "learning_rate": 9.96424107025757e-06, + "loss": 1.64602737, + "memory(GiB)": 71.21, + "step": 6795, + "train_speed(iter/s)": 0.451763 + }, + { + "acc": 0.64780502, + "epoch": 0.17250126839167934, + "grad_norm": 4.40625, + "learning_rate": 9.964115772702104e-06, + "loss": 1.68545761, + "memory(GiB)": 71.21, + "step": 6800, + "train_speed(iter/s)": 0.451939 + }, + { + "epoch": 0.17250126839167934, + "eval_acc": 0.6312635330497808, + "eval_loss": 1.6801426410675049, + "eval_runtime": 114.1408, + "eval_samples_per_second": 55.808, + "eval_steps_per_second": 27.904, + "step": 6800 + }, + { + "acc": 0.63717914, + "epoch": 0.1726281075596144, + "grad_norm": 4.71875, + "learning_rate": 9.963990256802662e-06, + "loss": 1.74001732, + "memory(GiB)": 71.21, + "step": 6805, + "train_speed(iter/s)": 0.448578 + }, + { + "acc": 0.64864168, + "epoch": 0.17275494672754946, + "grad_norm": 6.375, + "learning_rate": 9.963864522564765e-06, + "loss": 1.66755104, + "memory(GiB)": 71.21, + "step": 6810, + "train_speed(iter/s)": 0.448752 + }, + { + "acc": 0.63915405, + "epoch": 0.17288178589548453, + "grad_norm": 4.9375, + "learning_rate": 9.963738569993945e-06, + "loss": 1.72978821, + "memory(GiB)": 71.21, + "step": 6815, + "train_speed(iter/s)": 0.448926 + }, + { + "acc": 0.64949522, + "epoch": 0.17300862506341957, + "grad_norm": 6.65625, + "learning_rate": 9.963612399095743e-06, + "loss": 1.70425949, + "memory(GiB)": 71.21, + "step": 6820, + "train_speed(iter/s)": 0.4491 + }, + { + "acc": 0.62626896, + "epoch": 0.17313546423135465, + "grad_norm": 5.5, + "learning_rate": 9.963486009875705e-06, + "loss": 1.77209167, + "memory(GiB)": 71.21, + "step": 6825, + "train_speed(iter/s)": 0.449274 + }, + { + "acc": 0.63922653, + "epoch": 0.1732623033992897, + "grad_norm": 5.5625, + "learning_rate": 9.963359402339393e-06, + "loss": 1.6831665, + "memory(GiB)": 71.21, + "step": 6830, + "train_speed(iter/s)": 0.449446 + }, + { + "acc": 0.63800831, + "epoch": 0.17338914256722476, + "grad_norm": 6.71875, + "learning_rate": 9.963232576492373e-06, + "loss": 1.75432358, + "memory(GiB)": 71.21, + "step": 6835, + "train_speed(iter/s)": 0.449619 + }, + { + "acc": 0.63258686, + "epoch": 0.1735159817351598, + "grad_norm": 4.9375, + "learning_rate": 9.963105532340226e-06, + "loss": 1.7624651, + "memory(GiB)": 71.21, + "step": 6840, + "train_speed(iter/s)": 0.449792 + }, + { + "acc": 0.62508526, + "epoch": 0.17364282090309488, + "grad_norm": 6.09375, + "learning_rate": 9.962978269888538e-06, + "loss": 1.73552055, + "memory(GiB)": 71.21, + "step": 6845, + "train_speed(iter/s)": 0.449965 + }, + { + "acc": 0.62747498, + "epoch": 0.17376966007102992, + "grad_norm": 4.625, + "learning_rate": 9.96285078914291e-06, + "loss": 1.69868374, + "memory(GiB)": 71.21, + "step": 6850, + "train_speed(iter/s)": 0.450137 + }, + { + "acc": 0.63322325, + "epoch": 0.173896499238965, + "grad_norm": 4.90625, + "learning_rate": 9.962723090108944e-06, + "loss": 1.70937462, + "memory(GiB)": 71.21, + "step": 6855, + "train_speed(iter/s)": 0.450311 + }, + { + "acc": 0.63472128, + "epoch": 0.17402333840690004, + "grad_norm": 5.15625, + "learning_rate": 9.962595172792261e-06, + "loss": 1.71486435, + "memory(GiB)": 71.21, + "step": 6860, + "train_speed(iter/s)": 0.450484 + }, + { + "acc": 0.64128575, + "epoch": 0.1741501775748351, + "grad_norm": 4.90625, + "learning_rate": 9.962467037198487e-06, + "loss": 1.64592934, + "memory(GiB)": 71.21, + "step": 6865, + "train_speed(iter/s)": 0.450656 + }, + { + "acc": 0.641115, + "epoch": 0.17427701674277016, + "grad_norm": 5.0, + "learning_rate": 9.962338683333254e-06, + "loss": 1.70540276, + "memory(GiB)": 71.21, + "step": 6870, + "train_speed(iter/s)": 0.450829 + }, + { + "acc": 0.63209133, + "epoch": 0.17440385591070523, + "grad_norm": 4.75, + "learning_rate": 9.962210111202212e-06, + "loss": 1.78955193, + "memory(GiB)": 71.21, + "step": 6875, + "train_speed(iter/s)": 0.451001 + }, + { + "acc": 0.63830833, + "epoch": 0.17453069507864027, + "grad_norm": 6.46875, + "learning_rate": 9.962081320811015e-06, + "loss": 1.71247025, + "memory(GiB)": 71.21, + "step": 6880, + "train_speed(iter/s)": 0.451173 + }, + { + "acc": 0.64899645, + "epoch": 0.17465753424657535, + "grad_norm": 5.8125, + "learning_rate": 9.961952312165327e-06, + "loss": 1.68242855, + "memory(GiB)": 71.21, + "step": 6885, + "train_speed(iter/s)": 0.451346 + }, + { + "acc": 0.64467893, + "epoch": 0.1747843734145104, + "grad_norm": 4.75, + "learning_rate": 9.961823085270823e-06, + "loss": 1.67668209, + "memory(GiB)": 71.21, + "step": 6890, + "train_speed(iter/s)": 0.451517 + }, + { + "acc": 0.6508585, + "epoch": 0.17491121258244546, + "grad_norm": 4.96875, + "learning_rate": 9.961693640133187e-06, + "loss": 1.60053349, + "memory(GiB)": 71.21, + "step": 6895, + "train_speed(iter/s)": 0.451689 + }, + { + "acc": 0.63960209, + "epoch": 0.1750380517503805, + "grad_norm": 5.15625, + "learning_rate": 9.961563976758112e-06, + "loss": 1.71107082, + "memory(GiB)": 71.21, + "step": 6900, + "train_speed(iter/s)": 0.45186 + }, + { + "epoch": 0.1750380517503805, + "eval_acc": 0.6314807720506985, + "eval_loss": 1.679044485092163, + "eval_runtime": 114.2953, + "eval_samples_per_second": 55.733, + "eval_steps_per_second": 27.866, + "step": 6900 + }, + { + "acc": 0.63297606, + "epoch": 0.17516489091831558, + "grad_norm": 5.65625, + "learning_rate": 9.961434095151301e-06, + "loss": 1.75422115, + "memory(GiB)": 71.21, + "step": 6905, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.63639727, + "epoch": 0.17529173008625062, + "grad_norm": 4.28125, + "learning_rate": 9.961303995318467e-06, + "loss": 1.63829842, + "memory(GiB)": 71.21, + "step": 6910, + "train_speed(iter/s)": 0.448716 + }, + { + "acc": 0.64464526, + "epoch": 0.1754185692541857, + "grad_norm": 5.375, + "learning_rate": 9.961173677265334e-06, + "loss": 1.66234818, + "memory(GiB)": 71.21, + "step": 6915, + "train_speed(iter/s)": 0.448887 + }, + { + "acc": 0.63808126, + "epoch": 0.17554540842212074, + "grad_norm": 4.625, + "learning_rate": 9.961043140997632e-06, + "loss": 1.68159714, + "memory(GiB)": 71.21, + "step": 6920, + "train_speed(iter/s)": 0.449056 + }, + { + "acc": 0.62776995, + "epoch": 0.1756722475900558, + "grad_norm": 5.875, + "learning_rate": 9.960912386521104e-06, + "loss": 1.72295723, + "memory(GiB)": 71.21, + "step": 6925, + "train_speed(iter/s)": 0.449228 + }, + { + "acc": 0.63374896, + "epoch": 0.17579908675799086, + "grad_norm": 5.28125, + "learning_rate": 9.9607814138415e-06, + "loss": 1.69503231, + "memory(GiB)": 71.21, + "step": 6930, + "train_speed(iter/s)": 0.449398 + }, + { + "acc": 0.64268875, + "epoch": 0.17592592592592593, + "grad_norm": 6.40625, + "learning_rate": 9.96065022296458e-06, + "loss": 1.71260147, + "memory(GiB)": 71.21, + "step": 6935, + "train_speed(iter/s)": 0.44957 + }, + { + "acc": 0.64920301, + "epoch": 0.17605276509386097, + "grad_norm": 5.9375, + "learning_rate": 9.960518813896117e-06, + "loss": 1.65298824, + "memory(GiB)": 71.21, + "step": 6940, + "train_speed(iter/s)": 0.449739 + }, + { + "acc": 0.63320346, + "epoch": 0.17617960426179605, + "grad_norm": 5.4375, + "learning_rate": 9.960387186641887e-06, + "loss": 1.69064255, + "memory(GiB)": 71.21, + "step": 6945, + "train_speed(iter/s)": 0.449911 + }, + { + "acc": 0.64141626, + "epoch": 0.1763064434297311, + "grad_norm": 5.40625, + "learning_rate": 9.960255341207686e-06, + "loss": 1.66379223, + "memory(GiB)": 71.21, + "step": 6950, + "train_speed(iter/s)": 0.450079 + }, + { + "acc": 0.65310822, + "epoch": 0.17643328259766616, + "grad_norm": 4.78125, + "learning_rate": 9.960123277599305e-06, + "loss": 1.67130661, + "memory(GiB)": 71.21, + "step": 6955, + "train_speed(iter/s)": 0.450249 + }, + { + "acc": 0.62346492, + "epoch": 0.1765601217656012, + "grad_norm": 6.03125, + "learning_rate": 9.959990995822559e-06, + "loss": 1.71594238, + "memory(GiB)": 71.21, + "step": 6960, + "train_speed(iter/s)": 0.45042 + }, + { + "acc": 0.64713631, + "epoch": 0.17668696093353628, + "grad_norm": 4.4375, + "learning_rate": 9.959858495883263e-06, + "loss": 1.66925926, + "memory(GiB)": 71.21, + "step": 6965, + "train_speed(iter/s)": 0.450591 + }, + { + "acc": 0.64613409, + "epoch": 0.17681380010147132, + "grad_norm": 4.40625, + "learning_rate": 9.959725777787249e-06, + "loss": 1.64243317, + "memory(GiB)": 71.21, + "step": 6970, + "train_speed(iter/s)": 0.450762 + }, + { + "acc": 0.64114819, + "epoch": 0.1769406392694064, + "grad_norm": 6.375, + "learning_rate": 9.959592841540349e-06, + "loss": 1.72644653, + "memory(GiB)": 71.21, + "step": 6975, + "train_speed(iter/s)": 0.450932 + }, + { + "acc": 0.6248683, + "epoch": 0.17706747843734144, + "grad_norm": 4.96875, + "learning_rate": 9.959459687148414e-06, + "loss": 1.71904373, + "memory(GiB)": 71.21, + "step": 6980, + "train_speed(iter/s)": 0.451102 + }, + { + "acc": 0.6215354, + "epoch": 0.1771943176052765, + "grad_norm": 5.53125, + "learning_rate": 9.959326314617299e-06, + "loss": 1.75260105, + "memory(GiB)": 71.21, + "step": 6985, + "train_speed(iter/s)": 0.45127 + }, + { + "acc": 0.63244171, + "epoch": 0.17732115677321156, + "grad_norm": 5.875, + "learning_rate": 9.95919272395287e-06, + "loss": 1.73920155, + "memory(GiB)": 71.21, + "step": 6990, + "train_speed(iter/s)": 0.451439 + }, + { + "acc": 0.6280045, + "epoch": 0.17744799594114663, + "grad_norm": 5.375, + "learning_rate": 9.959058915161006e-06, + "loss": 1.8079855, + "memory(GiB)": 71.21, + "step": 6995, + "train_speed(iter/s)": 0.451611 + }, + { + "acc": 0.62093716, + "epoch": 0.17757483510908167, + "grad_norm": 5.25, + "learning_rate": 9.95892488824759e-06, + "loss": 1.74558754, + "memory(GiB)": 71.21, + "step": 7000, + "train_speed(iter/s)": 0.451782 + }, + { + "epoch": 0.17757483510908167, + "eval_acc": 0.631599000199275, + "eval_loss": 1.6772499084472656, + "eval_runtime": 114.821, + "eval_samples_per_second": 55.478, + "eval_steps_per_second": 27.739, + "step": 7000 + }, + { + "acc": 0.65774221, + "epoch": 0.17770167427701675, + "grad_norm": 5.4375, + "learning_rate": 9.958790643218515e-06, + "loss": 1.57542114, + "memory(GiB)": 71.21, + "step": 7005, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.6377305, + "epoch": 0.1778285134449518, + "grad_norm": 4.3125, + "learning_rate": 9.95865618007969e-06, + "loss": 1.63670177, + "memory(GiB)": 71.21, + "step": 7010, + "train_speed(iter/s)": 0.44867 + }, + { + "acc": 0.64239469, + "epoch": 0.17795535261288686, + "grad_norm": 4.9375, + "learning_rate": 9.958521498837029e-06, + "loss": 1.66215572, + "memory(GiB)": 71.21, + "step": 7015, + "train_speed(iter/s)": 0.448841 + }, + { + "acc": 0.63991089, + "epoch": 0.1780821917808219, + "grad_norm": 5.5, + "learning_rate": 9.95838659949645e-06, + "loss": 1.73151035, + "memory(GiB)": 71.21, + "step": 7020, + "train_speed(iter/s)": 0.449012 + }, + { + "acc": 0.65910544, + "epoch": 0.17820903094875698, + "grad_norm": 5.5625, + "learning_rate": 9.958251482063894e-06, + "loss": 1.67527275, + "memory(GiB)": 71.21, + "step": 7025, + "train_speed(iter/s)": 0.449183 + }, + { + "acc": 0.63096104, + "epoch": 0.17833587011669202, + "grad_norm": 5.4375, + "learning_rate": 9.9581161465453e-06, + "loss": 1.74462528, + "memory(GiB)": 71.21, + "step": 7030, + "train_speed(iter/s)": 0.449354 + }, + { + "acc": 0.62816329, + "epoch": 0.1784627092846271, + "grad_norm": 5.46875, + "learning_rate": 9.957980592946621e-06, + "loss": 1.73820324, + "memory(GiB)": 71.21, + "step": 7035, + "train_speed(iter/s)": 0.449526 + }, + { + "acc": 0.65771322, + "epoch": 0.17858954845256214, + "grad_norm": 5.4375, + "learning_rate": 9.957844821273822e-06, + "loss": 1.61181011, + "memory(GiB)": 71.21, + "step": 7040, + "train_speed(iter/s)": 0.449697 + }, + { + "acc": 0.65727053, + "epoch": 0.1787163876204972, + "grad_norm": 5.0625, + "learning_rate": 9.95770883153287e-06, + "loss": 1.59334517, + "memory(GiB)": 71.21, + "step": 7045, + "train_speed(iter/s)": 0.449867 + }, + { + "acc": 0.64151306, + "epoch": 0.17884322678843226, + "grad_norm": 5.28125, + "learning_rate": 9.957572623729749e-06, + "loss": 1.75311317, + "memory(GiB)": 71.21, + "step": 7050, + "train_speed(iter/s)": 0.450038 + }, + { + "acc": 0.6288065, + "epoch": 0.17897006595636733, + "grad_norm": 5.90625, + "learning_rate": 9.957436197870451e-06, + "loss": 1.702075, + "memory(GiB)": 71.21, + "step": 7055, + "train_speed(iter/s)": 0.450207 + }, + { + "acc": 0.6372962, + "epoch": 0.17909690512430237, + "grad_norm": 6.09375, + "learning_rate": 9.957299553960975e-06, + "loss": 1.73042221, + "memory(GiB)": 71.21, + "step": 7060, + "train_speed(iter/s)": 0.450377 + }, + { + "acc": 0.64321432, + "epoch": 0.17922374429223745, + "grad_norm": 6.03125, + "learning_rate": 9.957162692007334e-06, + "loss": 1.64974308, + "memory(GiB)": 71.21, + "step": 7065, + "train_speed(iter/s)": 0.450547 + }, + { + "acc": 0.64608669, + "epoch": 0.1793505834601725, + "grad_norm": 5.625, + "learning_rate": 9.957025612015543e-06, + "loss": 1.68608284, + "memory(GiB)": 71.21, + "step": 7070, + "train_speed(iter/s)": 0.450716 + }, + { + "acc": 0.65064063, + "epoch": 0.17947742262810756, + "grad_norm": 4.8125, + "learning_rate": 9.956888313991636e-06, + "loss": 1.63163643, + "memory(GiB)": 71.21, + "step": 7075, + "train_speed(iter/s)": 0.450885 + }, + { + "acc": 0.63756208, + "epoch": 0.1796042617960426, + "grad_norm": 5.65625, + "learning_rate": 9.956750797941648e-06, + "loss": 1.6888073, + "memory(GiB)": 71.21, + "step": 7080, + "train_speed(iter/s)": 0.451056 + }, + { + "acc": 0.63542538, + "epoch": 0.17973110096397768, + "grad_norm": 4.71875, + "learning_rate": 9.95661306387163e-06, + "loss": 1.74927101, + "memory(GiB)": 71.21, + "step": 7085, + "train_speed(iter/s)": 0.451225 + }, + { + "acc": 0.64372125, + "epoch": 0.17985794013191272, + "grad_norm": 4.40625, + "learning_rate": 9.95647511178764e-06, + "loss": 1.60645237, + "memory(GiB)": 71.21, + "step": 7090, + "train_speed(iter/s)": 0.451395 + }, + { + "acc": 0.64498444, + "epoch": 0.1799847792998478, + "grad_norm": 6.15625, + "learning_rate": 9.956336941695747e-06, + "loss": 1.71201668, + "memory(GiB)": 71.21, + "step": 7095, + "train_speed(iter/s)": 0.451564 + }, + { + "acc": 0.64342675, + "epoch": 0.18011161846778284, + "grad_norm": 5.3125, + "learning_rate": 9.956198553602026e-06, + "loss": 1.64535561, + "memory(GiB)": 71.21, + "step": 7100, + "train_speed(iter/s)": 0.451733 + }, + { + "epoch": 0.18011161846778284, + "eval_acc": 0.6319010459640126, + "eval_loss": 1.675980567932129, + "eval_runtime": 114.7422, + "eval_samples_per_second": 55.516, + "eval_steps_per_second": 27.758, + "step": 7100 + }, + { + "acc": 0.60847073, + "epoch": 0.1802384576357179, + "grad_norm": 4.875, + "learning_rate": 9.956059947512563e-06, + "loss": 1.79434357, + "memory(GiB)": 71.21, + "step": 7105, + "train_speed(iter/s)": 0.448499 + }, + { + "acc": 0.63685389, + "epoch": 0.18036529680365296, + "grad_norm": 5.0625, + "learning_rate": 9.95592112343346e-06, + "loss": 1.66323643, + "memory(GiB)": 71.21, + "step": 7110, + "train_speed(iter/s)": 0.448666 + }, + { + "acc": 0.64174376, + "epoch": 0.18049213597158803, + "grad_norm": 5.5625, + "learning_rate": 9.955782081370818e-06, + "loss": 1.64723377, + "memory(GiB)": 71.21, + "step": 7115, + "train_speed(iter/s)": 0.448831 + }, + { + "acc": 0.65300169, + "epoch": 0.18061897513952307, + "grad_norm": 4.96875, + "learning_rate": 9.955642821330752e-06, + "loss": 1.67541733, + "memory(GiB)": 71.21, + "step": 7120, + "train_speed(iter/s)": 0.448998 + }, + { + "acc": 0.65841947, + "epoch": 0.18074581430745815, + "grad_norm": 4.96875, + "learning_rate": 9.95550334331939e-06, + "loss": 1.63094025, + "memory(GiB)": 71.21, + "step": 7125, + "train_speed(iter/s)": 0.449163 + }, + { + "acc": 0.6470685, + "epoch": 0.1808726534753932, + "grad_norm": 5.25, + "learning_rate": 9.955363647342868e-06, + "loss": 1.64027538, + "memory(GiB)": 71.21, + "step": 7130, + "train_speed(iter/s)": 0.44933 + }, + { + "acc": 0.62691331, + "epoch": 0.18099949264332826, + "grad_norm": 6.5, + "learning_rate": 9.955223733407327e-06, + "loss": 1.74702682, + "memory(GiB)": 71.21, + "step": 7135, + "train_speed(iter/s)": 0.449499 + }, + { + "acc": 0.64267902, + "epoch": 0.1811263318112633, + "grad_norm": 5.3125, + "learning_rate": 9.955083601518924e-06, + "loss": 1.7266552, + "memory(GiB)": 71.21, + "step": 7140, + "train_speed(iter/s)": 0.449668 + }, + { + "acc": 0.62211342, + "epoch": 0.18125317097919838, + "grad_norm": 5.3125, + "learning_rate": 9.95494325168382e-06, + "loss": 1.75653534, + "memory(GiB)": 71.21, + "step": 7145, + "train_speed(iter/s)": 0.449837 + }, + { + "acc": 0.6333622, + "epoch": 0.18138001014713342, + "grad_norm": 4.90625, + "learning_rate": 9.954802683908192e-06, + "loss": 1.74033585, + "memory(GiB)": 71.21, + "step": 7150, + "train_speed(iter/s)": 0.450006 + }, + { + "acc": 0.64277658, + "epoch": 0.1815068493150685, + "grad_norm": 4.84375, + "learning_rate": 9.954661898198216e-06, + "loss": 1.64676437, + "memory(GiB)": 71.21, + "step": 7155, + "train_speed(iter/s)": 0.450175 + }, + { + "acc": 0.63284101, + "epoch": 0.18163368848300354, + "grad_norm": 5.75, + "learning_rate": 9.954520894560092e-06, + "loss": 1.73367424, + "memory(GiB)": 71.21, + "step": 7160, + "train_speed(iter/s)": 0.450345 + }, + { + "acc": 0.63460722, + "epoch": 0.1817605276509386, + "grad_norm": 4.46875, + "learning_rate": 9.954379673000018e-06, + "loss": 1.72355309, + "memory(GiB)": 71.21, + "step": 7165, + "train_speed(iter/s)": 0.450513 + }, + { + "acc": 0.63592048, + "epoch": 0.18188736681887366, + "grad_norm": 5.0, + "learning_rate": 9.954238233524208e-06, + "loss": 1.72763519, + "memory(GiB)": 71.21, + "step": 7170, + "train_speed(iter/s)": 0.450681 + }, + { + "acc": 0.63759942, + "epoch": 0.18201420598680873, + "grad_norm": 5.1875, + "learning_rate": 9.954096576138879e-06, + "loss": 1.73296318, + "memory(GiB)": 71.21, + "step": 7175, + "train_speed(iter/s)": 0.450849 + }, + { + "acc": 0.63229589, + "epoch": 0.18214104515474377, + "grad_norm": 5.09375, + "learning_rate": 9.953954700850264e-06, + "loss": 1.73471756, + "memory(GiB)": 71.21, + "step": 7180, + "train_speed(iter/s)": 0.451018 + }, + { + "acc": 0.62964411, + "epoch": 0.18226788432267885, + "grad_norm": 4.40625, + "learning_rate": 9.953812607664607e-06, + "loss": 1.74249134, + "memory(GiB)": 71.21, + "step": 7185, + "train_speed(iter/s)": 0.451185 + }, + { + "acc": 0.6472333, + "epoch": 0.1823947234906139, + "grad_norm": 5.5, + "learning_rate": 9.95367029658815e-06, + "loss": 1.71487122, + "memory(GiB)": 71.21, + "step": 7190, + "train_speed(iter/s)": 0.451354 + }, + { + "acc": 0.63836093, + "epoch": 0.18252156265854896, + "grad_norm": 6.0625, + "learning_rate": 9.953527767627159e-06, + "loss": 1.68553467, + "memory(GiB)": 71.21, + "step": 7195, + "train_speed(iter/s)": 0.451521 + }, + { + "acc": 0.66171017, + "epoch": 0.182648401826484, + "grad_norm": 4.875, + "learning_rate": 9.9533850207879e-06, + "loss": 1.60840836, + "memory(GiB)": 71.21, + "step": 7200, + "train_speed(iter/s)": 0.451688 + }, + { + "epoch": 0.182648401826484, + "eval_acc": 0.6320497711261795, + "eval_loss": 1.6746814250946045, + "eval_runtime": 113.9395, + "eval_samples_per_second": 55.907, + "eval_steps_per_second": 27.953, + "step": 7200 + }, + { + "acc": 0.65047646, + "epoch": 0.18277524099441908, + "grad_norm": 5.5, + "learning_rate": 9.953242056076652e-06, + "loss": 1.70351143, + "memory(GiB)": 71.21, + "step": 7205, + "train_speed(iter/s)": 0.448524 + }, + { + "acc": 0.62818956, + "epoch": 0.18290208016235412, + "grad_norm": 5.8125, + "learning_rate": 9.953098873499705e-06, + "loss": 1.75705948, + "memory(GiB)": 71.21, + "step": 7210, + "train_speed(iter/s)": 0.448692 + }, + { + "acc": 0.64600315, + "epoch": 0.1830289193302892, + "grad_norm": 5.0, + "learning_rate": 9.952955473063356e-06, + "loss": 1.62562675, + "memory(GiB)": 71.21, + "step": 7215, + "train_speed(iter/s)": 0.448859 + }, + { + "acc": 0.63142004, + "epoch": 0.18315575849822424, + "grad_norm": 5.1875, + "learning_rate": 9.952811854773911e-06, + "loss": 1.73863335, + "memory(GiB)": 71.21, + "step": 7220, + "train_speed(iter/s)": 0.449026 + }, + { + "acc": 0.63053617, + "epoch": 0.1832825976661593, + "grad_norm": 6.09375, + "learning_rate": 9.952668018637687e-06, + "loss": 1.73417759, + "memory(GiB)": 71.21, + "step": 7225, + "train_speed(iter/s)": 0.449191 + }, + { + "acc": 0.6388804, + "epoch": 0.18340943683409436, + "grad_norm": 5.5, + "learning_rate": 9.952523964661014e-06, + "loss": 1.68569317, + "memory(GiB)": 71.21, + "step": 7230, + "train_speed(iter/s)": 0.449354 + }, + { + "acc": 0.63938942, + "epoch": 0.18353627600202943, + "grad_norm": 5.375, + "learning_rate": 9.952379692850222e-06, + "loss": 1.70200844, + "memory(GiB)": 71.21, + "step": 7235, + "train_speed(iter/s)": 0.449518 + }, + { + "acc": 0.64353642, + "epoch": 0.18366311516996447, + "grad_norm": 6.5, + "learning_rate": 9.952235203211663e-06, + "loss": 1.63670692, + "memory(GiB)": 71.21, + "step": 7240, + "train_speed(iter/s)": 0.449681 + }, + { + "acc": 0.65047665, + "epoch": 0.18378995433789955, + "grad_norm": 5.03125, + "learning_rate": 9.952090495751689e-06, + "loss": 1.67845173, + "memory(GiB)": 71.21, + "step": 7245, + "train_speed(iter/s)": 0.449843 + }, + { + "acc": 0.64037266, + "epoch": 0.1839167935058346, + "grad_norm": 5.53125, + "learning_rate": 9.951945570476666e-06, + "loss": 1.68576107, + "memory(GiB)": 71.21, + "step": 7250, + "train_speed(iter/s)": 0.450008 + }, + { + "acc": 0.62808704, + "epoch": 0.18404363267376966, + "grad_norm": 5.875, + "learning_rate": 9.951800427392968e-06, + "loss": 1.77198658, + "memory(GiB)": 71.21, + "step": 7255, + "train_speed(iter/s)": 0.450174 + }, + { + "acc": 0.63405037, + "epoch": 0.1841704718417047, + "grad_norm": 5.3125, + "learning_rate": 9.951655066506977e-06, + "loss": 1.75762978, + "memory(GiB)": 71.21, + "step": 7260, + "train_speed(iter/s)": 0.450339 + }, + { + "acc": 0.63699379, + "epoch": 0.18429731100963978, + "grad_norm": 5.25, + "learning_rate": 9.951509487825091e-06, + "loss": 1.68818207, + "memory(GiB)": 71.21, + "step": 7265, + "train_speed(iter/s)": 0.450503 + }, + { + "acc": 0.62087727, + "epoch": 0.18442415017757482, + "grad_norm": 6.21875, + "learning_rate": 9.95136369135371e-06, + "loss": 1.75531082, + "memory(GiB)": 71.21, + "step": 7270, + "train_speed(iter/s)": 0.450669 + }, + { + "acc": 0.64249229, + "epoch": 0.1845509893455099, + "grad_norm": 4.96875, + "learning_rate": 9.951217677099248e-06, + "loss": 1.70273247, + "memory(GiB)": 71.21, + "step": 7275, + "train_speed(iter/s)": 0.450834 + }, + { + "acc": 0.63505716, + "epoch": 0.18467782851344494, + "grad_norm": 5.125, + "learning_rate": 9.951071445068125e-06, + "loss": 1.67429008, + "memory(GiB)": 71.21, + "step": 7280, + "train_speed(iter/s)": 0.450997 + }, + { + "acc": 0.63318262, + "epoch": 0.18480466768138, + "grad_norm": 5.46875, + "learning_rate": 9.950924995266778e-06, + "loss": 1.73183746, + "memory(GiB)": 71.21, + "step": 7285, + "train_speed(iter/s)": 0.45116 + }, + { + "acc": 0.6422121, + "epoch": 0.18493150684931506, + "grad_norm": 5.1875, + "learning_rate": 9.950778327701643e-06, + "loss": 1.71187248, + "memory(GiB)": 71.21, + "step": 7290, + "train_speed(iter/s)": 0.451323 + }, + { + "acc": 0.65183239, + "epoch": 0.18505834601725013, + "grad_norm": 4.46875, + "learning_rate": 9.950631442379175e-06, + "loss": 1.68032799, + "memory(GiB)": 71.21, + "step": 7295, + "train_speed(iter/s)": 0.451486 + }, + { + "acc": 0.64511795, + "epoch": 0.18518518518518517, + "grad_norm": 5.90625, + "learning_rate": 9.950484339305832e-06, + "loss": 1.64890652, + "memory(GiB)": 71.21, + "step": 7300, + "train_speed(iter/s)": 0.45165 + }, + { + "epoch": 0.18518518518518517, + "eval_acc": 0.6323739585583185, + "eval_loss": 1.673172950744629, + "eval_runtime": 113.8426, + "eval_samples_per_second": 55.954, + "eval_steps_per_second": 27.977, + "step": 7300 + }, + { + "acc": 0.64879112, + "epoch": 0.18531202435312025, + "grad_norm": 5.4375, + "learning_rate": 9.950337018488086e-06, + "loss": 1.72085114, + "memory(GiB)": 71.21, + "step": 7305, + "train_speed(iter/s)": 0.448531 + }, + { + "acc": 0.63489003, + "epoch": 0.1854388635210553, + "grad_norm": 4.8125, + "learning_rate": 9.950189479932417e-06, + "loss": 1.71967411, + "memory(GiB)": 71.21, + "step": 7310, + "train_speed(iter/s)": 0.448692 + }, + { + "acc": 0.64130874, + "epoch": 0.18556570268899036, + "grad_norm": 5.0, + "learning_rate": 9.950041723645312e-06, + "loss": 1.69104652, + "memory(GiB)": 71.21, + "step": 7315, + "train_speed(iter/s)": 0.448853 + }, + { + "acc": 0.63875937, + "epoch": 0.1856925418569254, + "grad_norm": 4.75, + "learning_rate": 9.949893749633273e-06, + "loss": 1.7242691, + "memory(GiB)": 71.21, + "step": 7320, + "train_speed(iter/s)": 0.449015 + }, + { + "acc": 0.64443955, + "epoch": 0.18581938102486048, + "grad_norm": 5.1875, + "learning_rate": 9.949745557902806e-06, + "loss": 1.75857048, + "memory(GiB)": 71.21, + "step": 7325, + "train_speed(iter/s)": 0.449176 + }, + { + "acc": 0.64022522, + "epoch": 0.18594622019279552, + "grad_norm": 4.40625, + "learning_rate": 9.949597148460433e-06, + "loss": 1.70952644, + "memory(GiB)": 71.21, + "step": 7330, + "train_speed(iter/s)": 0.449338 + }, + { + "acc": 0.63151064, + "epoch": 0.1860730593607306, + "grad_norm": 5.46875, + "learning_rate": 9.949448521312676e-06, + "loss": 1.7911808, + "memory(GiB)": 71.21, + "step": 7335, + "train_speed(iter/s)": 0.4495 + }, + { + "acc": 0.63607311, + "epoch": 0.18619989852866564, + "grad_norm": 6.84375, + "learning_rate": 9.949299676466077e-06, + "loss": 1.74132233, + "memory(GiB)": 71.21, + "step": 7340, + "train_speed(iter/s)": 0.44966 + }, + { + "acc": 0.63019934, + "epoch": 0.1863267376966007, + "grad_norm": 5.25, + "learning_rate": 9.94915061392718e-06, + "loss": 1.69024429, + "memory(GiB)": 71.21, + "step": 7345, + "train_speed(iter/s)": 0.449819 + }, + { + "acc": 0.64417286, + "epoch": 0.18645357686453576, + "grad_norm": 5.3125, + "learning_rate": 9.949001333702543e-06, + "loss": 1.64219131, + "memory(GiB)": 71.21, + "step": 7350, + "train_speed(iter/s)": 0.449979 + }, + { + "acc": 0.62558489, + "epoch": 0.18658041603247083, + "grad_norm": 5.5, + "learning_rate": 9.948851835798732e-06, + "loss": 1.81280899, + "memory(GiB)": 71.21, + "step": 7355, + "train_speed(iter/s)": 0.450139 + }, + { + "acc": 0.62541652, + "epoch": 0.18670725520040587, + "grad_norm": 4.8125, + "learning_rate": 9.948702120222323e-06, + "loss": 1.73590984, + "memory(GiB)": 71.21, + "step": 7360, + "train_speed(iter/s)": 0.450298 + }, + { + "acc": 0.65073023, + "epoch": 0.18683409436834095, + "grad_norm": 5.4375, + "learning_rate": 9.9485521869799e-06, + "loss": 1.64357224, + "memory(GiB)": 71.21, + "step": 7365, + "train_speed(iter/s)": 0.450457 + }, + { + "acc": 0.65757856, + "epoch": 0.186960933536276, + "grad_norm": 5.1875, + "learning_rate": 9.948402036078057e-06, + "loss": 1.62586269, + "memory(GiB)": 71.21, + "step": 7370, + "train_speed(iter/s)": 0.450615 + }, + { + "acc": 0.62736578, + "epoch": 0.18708777270421106, + "grad_norm": 5.125, + "learning_rate": 9.948251667523401e-06, + "loss": 1.74679489, + "memory(GiB)": 71.21, + "step": 7375, + "train_speed(iter/s)": 0.450772 + }, + { + "acc": 0.62891831, + "epoch": 0.1872146118721461, + "grad_norm": 5.03125, + "learning_rate": 9.948101081322544e-06, + "loss": 1.67257652, + "memory(GiB)": 71.21, + "step": 7380, + "train_speed(iter/s)": 0.45093 + }, + { + "acc": 0.64120245, + "epoch": 0.18734145104008118, + "grad_norm": 4.59375, + "learning_rate": 9.947950277482109e-06, + "loss": 1.66956902, + "memory(GiB)": 71.21, + "step": 7385, + "train_speed(iter/s)": 0.451088 + }, + { + "acc": 0.64241848, + "epoch": 0.18746829020801623, + "grad_norm": 5.3125, + "learning_rate": 9.94779925600873e-06, + "loss": 1.71430054, + "memory(GiB)": 71.21, + "step": 7390, + "train_speed(iter/s)": 0.451246 + }, + { + "acc": 0.65908499, + "epoch": 0.1875951293759513, + "grad_norm": 5.09375, + "learning_rate": 9.947648016909048e-06, + "loss": 1.66054764, + "memory(GiB)": 71.21, + "step": 7395, + "train_speed(iter/s)": 0.451405 + }, + { + "acc": 0.6492085, + "epoch": 0.18772196854388634, + "grad_norm": 5.3125, + "learning_rate": 9.947496560189717e-06, + "loss": 1.69584789, + "memory(GiB)": 71.21, + "step": 7400, + "train_speed(iter/s)": 0.451563 + }, + { + "epoch": 0.18772196854388634, + "eval_acc": 0.6326066549496862, + "eval_loss": 1.6725375652313232, + "eval_runtime": 113.9438, + "eval_samples_per_second": 55.905, + "eval_steps_per_second": 27.952, + "step": 7400 + }, + { + "acc": 0.64540453, + "epoch": 0.18784880771182141, + "grad_norm": 4.71875, + "learning_rate": 9.9473448858574e-06, + "loss": 1.72471313, + "memory(GiB)": 71.21, + "step": 7405, + "train_speed(iter/s)": 0.448482 + }, + { + "acc": 0.61458941, + "epoch": 0.18797564687975646, + "grad_norm": 5.71875, + "learning_rate": 9.947192993918765e-06, + "loss": 1.79372635, + "memory(GiB)": 71.21, + "step": 7410, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.62938728, + "epoch": 0.18810248604769153, + "grad_norm": 5.53125, + "learning_rate": 9.947040884380496e-06, + "loss": 1.71874619, + "memory(GiB)": 71.21, + "step": 7415, + "train_speed(iter/s)": 0.448798 + }, + { + "acc": 0.65056539, + "epoch": 0.18822932521562658, + "grad_norm": 5.65625, + "learning_rate": 9.946888557249281e-06, + "loss": 1.64476013, + "memory(GiB)": 71.21, + "step": 7420, + "train_speed(iter/s)": 0.448957 + }, + { + "acc": 0.6448823, + "epoch": 0.18835616438356165, + "grad_norm": 5.53125, + "learning_rate": 9.946736012531821e-06, + "loss": 1.68225822, + "memory(GiB)": 71.21, + "step": 7425, + "train_speed(iter/s)": 0.449116 + }, + { + "acc": 0.64309673, + "epoch": 0.1884830035514967, + "grad_norm": 4.84375, + "learning_rate": 9.946583250234826e-06, + "loss": 1.68420677, + "memory(GiB)": 71.21, + "step": 7430, + "train_speed(iter/s)": 0.449274 + }, + { + "acc": 0.64436255, + "epoch": 0.18860984271943176, + "grad_norm": 4.78125, + "learning_rate": 9.946430270365015e-06, + "loss": 1.64312897, + "memory(GiB)": 71.21, + "step": 7435, + "train_speed(iter/s)": 0.449432 + }, + { + "acc": 0.64921837, + "epoch": 0.1887366818873668, + "grad_norm": 5.25, + "learning_rate": 9.946277072929115e-06, + "loss": 1.66890221, + "memory(GiB)": 71.21, + "step": 7440, + "train_speed(iter/s)": 0.44959 + }, + { + "acc": 0.63899708, + "epoch": 0.18886352105530188, + "grad_norm": 4.8125, + "learning_rate": 9.946123657933867e-06, + "loss": 1.70184975, + "memory(GiB)": 71.21, + "step": 7445, + "train_speed(iter/s)": 0.449745 + }, + { + "acc": 0.63540998, + "epoch": 0.18899036022323693, + "grad_norm": 5.34375, + "learning_rate": 9.945970025386018e-06, + "loss": 1.6520401, + "memory(GiB)": 71.21, + "step": 7450, + "train_speed(iter/s)": 0.449902 + }, + { + "acc": 0.6401722, + "epoch": 0.189117199391172, + "grad_norm": 6.09375, + "learning_rate": 9.945816175292326e-06, + "loss": 1.65837421, + "memory(GiB)": 71.21, + "step": 7455, + "train_speed(iter/s)": 0.45006 + }, + { + "acc": 0.64298801, + "epoch": 0.18924403855910704, + "grad_norm": 5.09375, + "learning_rate": 9.945662107659554e-06, + "loss": 1.71494408, + "memory(GiB)": 71.21, + "step": 7460, + "train_speed(iter/s)": 0.450216 + }, + { + "acc": 0.65179439, + "epoch": 0.18937087772704211, + "grad_norm": 6.125, + "learning_rate": 9.945507822494485e-06, + "loss": 1.65016613, + "memory(GiB)": 71.21, + "step": 7465, + "train_speed(iter/s)": 0.450372 + }, + { + "acc": 0.63841887, + "epoch": 0.18949771689497716, + "grad_norm": 5.9375, + "learning_rate": 9.9453533198039e-06, + "loss": 1.69674721, + "memory(GiB)": 71.21, + "step": 7470, + "train_speed(iter/s)": 0.45053 + }, + { + "acc": 0.63471417, + "epoch": 0.18962455606291223, + "grad_norm": 5.03125, + "learning_rate": 9.945198599594598e-06, + "loss": 1.68193779, + "memory(GiB)": 71.21, + "step": 7475, + "train_speed(iter/s)": 0.450687 + }, + { + "acc": 0.63170328, + "epoch": 0.18975139523084728, + "grad_norm": 4.71875, + "learning_rate": 9.945043661873381e-06, + "loss": 1.75789623, + "memory(GiB)": 71.21, + "step": 7480, + "train_speed(iter/s)": 0.450842 + }, + { + "acc": 0.65170374, + "epoch": 0.18987823439878235, + "grad_norm": 4.78125, + "learning_rate": 9.944888506647066e-06, + "loss": 1.66185989, + "memory(GiB)": 71.21, + "step": 7485, + "train_speed(iter/s)": 0.450999 + }, + { + "acc": 0.64946141, + "epoch": 0.1900050735667174, + "grad_norm": 5.6875, + "learning_rate": 9.944733133922479e-06, + "loss": 1.68779831, + "memory(GiB)": 71.21, + "step": 7490, + "train_speed(iter/s)": 0.451157 + }, + { + "acc": 0.64097347, + "epoch": 0.19013191273465246, + "grad_norm": 5.625, + "learning_rate": 9.944577543706451e-06, + "loss": 1.66502228, + "memory(GiB)": 71.21, + "step": 7495, + "train_speed(iter/s)": 0.451312 + }, + { + "acc": 0.6335701, + "epoch": 0.1902587519025875, + "grad_norm": 5.25, + "learning_rate": 9.944421736005825e-06, + "loss": 1.68425674, + "memory(GiB)": 71.21, + "step": 7500, + "train_speed(iter/s)": 0.451469 + }, + { + "epoch": 0.1902587519025875, + "eval_acc": 0.6326546981902738, + "eval_loss": 1.671066164970398, + "eval_runtime": 112.5248, + "eval_samples_per_second": 56.61, + "eval_steps_per_second": 28.305, + "step": 7500 + }, + { + "acc": 0.63305182, + "epoch": 0.19038559107052258, + "grad_norm": 5.0, + "learning_rate": 9.944265710827459e-06, + "loss": 1.78227558, + "memory(GiB)": 71.21, + "step": 7505, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.65126085, + "epoch": 0.19051243023845763, + "grad_norm": 4.96875, + "learning_rate": 9.944109468178208e-06, + "loss": 1.66644154, + "memory(GiB)": 71.21, + "step": 7510, + "train_speed(iter/s)": 0.448623 + }, + { + "acc": 0.64263673, + "epoch": 0.1906392694063927, + "grad_norm": 4.375, + "learning_rate": 9.943953008064953e-06, + "loss": 1.69451485, + "memory(GiB)": 71.21, + "step": 7515, + "train_speed(iter/s)": 0.448777 + }, + { + "acc": 0.63827748, + "epoch": 0.19076610857432774, + "grad_norm": 5.71875, + "learning_rate": 9.94379633049457e-06, + "loss": 1.70436745, + "memory(GiB)": 71.21, + "step": 7520, + "train_speed(iter/s)": 0.448932 + }, + { + "acc": 0.64319239, + "epoch": 0.19089294774226281, + "grad_norm": 4.71875, + "learning_rate": 9.943639435473952e-06, + "loss": 1.70651016, + "memory(GiB)": 71.21, + "step": 7525, + "train_speed(iter/s)": 0.449087 + }, + { + "acc": 0.65277729, + "epoch": 0.19101978691019786, + "grad_norm": 5.125, + "learning_rate": 9.94348232301e-06, + "loss": 1.67304573, + "memory(GiB)": 71.21, + "step": 7530, + "train_speed(iter/s)": 0.449083 + }, + { + "acc": 0.62471237, + "epoch": 0.19114662607813293, + "grad_norm": 5.59375, + "learning_rate": 9.943324993109624e-06, + "loss": 1.72396278, + "memory(GiB)": 71.21, + "step": 7535, + "train_speed(iter/s)": 0.449241 + }, + { + "acc": 0.64704294, + "epoch": 0.19127346524606798, + "grad_norm": 5.4375, + "learning_rate": 9.943167445779745e-06, + "loss": 1.65588303, + "memory(GiB)": 71.21, + "step": 7540, + "train_speed(iter/s)": 0.449397 + }, + { + "acc": 0.66720619, + "epoch": 0.19140030441400305, + "grad_norm": 5.5, + "learning_rate": 9.94300968102729e-06, + "loss": 1.62361031, + "memory(GiB)": 71.21, + "step": 7545, + "train_speed(iter/s)": 0.449552 + }, + { + "acc": 0.64876785, + "epoch": 0.1915271435819381, + "grad_norm": 5.03125, + "learning_rate": 9.942851698859204e-06, + "loss": 1.67092438, + "memory(GiB)": 71.21, + "step": 7550, + "train_speed(iter/s)": 0.449708 + }, + { + "acc": 0.65490861, + "epoch": 0.19165398274987316, + "grad_norm": 5.59375, + "learning_rate": 9.94269349928243e-06, + "loss": 1.62048988, + "memory(GiB)": 71.21, + "step": 7555, + "train_speed(iter/s)": 0.449863 + }, + { + "acc": 0.65470638, + "epoch": 0.1917808219178082, + "grad_norm": 5.25, + "learning_rate": 9.942535082303927e-06, + "loss": 1.62141399, + "memory(GiB)": 71.21, + "step": 7560, + "train_speed(iter/s)": 0.450019 + }, + { + "acc": 0.63746181, + "epoch": 0.19190766108574328, + "grad_norm": 5.28125, + "learning_rate": 9.942376447930666e-06, + "loss": 1.68490295, + "memory(GiB)": 71.21, + "step": 7565, + "train_speed(iter/s)": 0.450175 + }, + { + "acc": 0.65492516, + "epoch": 0.19203450025367833, + "grad_norm": 4.84375, + "learning_rate": 9.942217596169623e-06, + "loss": 1.60066757, + "memory(GiB)": 71.21, + "step": 7570, + "train_speed(iter/s)": 0.45033 + }, + { + "acc": 0.62345309, + "epoch": 0.1921613394216134, + "grad_norm": 4.625, + "learning_rate": 9.942058527027785e-06, + "loss": 1.74955063, + "memory(GiB)": 71.21, + "step": 7575, + "train_speed(iter/s)": 0.450487 + }, + { + "acc": 0.62559452, + "epoch": 0.19228817858954844, + "grad_norm": 5.125, + "learning_rate": 9.941899240512147e-06, + "loss": 1.7148037, + "memory(GiB)": 71.21, + "step": 7580, + "train_speed(iter/s)": 0.450643 + }, + { + "acc": 0.6384922, + "epoch": 0.19241501775748351, + "grad_norm": 5.34375, + "learning_rate": 9.941739736629716e-06, + "loss": 1.69868488, + "memory(GiB)": 71.21, + "step": 7585, + "train_speed(iter/s)": 0.450799 + }, + { + "acc": 0.64508724, + "epoch": 0.19254185692541856, + "grad_norm": 5.09375, + "learning_rate": 9.941580015387509e-06, + "loss": 1.6488596, + "memory(GiB)": 71.21, + "step": 7590, + "train_speed(iter/s)": 0.450955 + }, + { + "acc": 0.6188859, + "epoch": 0.19266869609335363, + "grad_norm": 7.375, + "learning_rate": 9.94142007679255e-06, + "loss": 1.7915451, + "memory(GiB)": 71.21, + "step": 7595, + "train_speed(iter/s)": 0.451111 + }, + { + "acc": 0.65783496, + "epoch": 0.19279553526128868, + "grad_norm": 6.40625, + "learning_rate": 9.941259920851874e-06, + "loss": 1.65056438, + "memory(GiB)": 71.21, + "step": 7600, + "train_speed(iter/s)": 0.451267 + }, + { + "epoch": 0.19279553526128868, + "eval_acc": 0.6329128783875184, + "eval_loss": 1.6699492931365967, + "eval_runtime": 112.9926, + "eval_samples_per_second": 56.375, + "eval_steps_per_second": 28.188, + "step": 7600 + }, + { + "acc": 0.63844094, + "epoch": 0.19292237442922375, + "grad_norm": 7.53125, + "learning_rate": 9.941099547572527e-06, + "loss": 1.71425858, + "memory(GiB)": 71.21, + "step": 7605, + "train_speed(iter/s)": 0.448297 + }, + { + "acc": 0.64833999, + "epoch": 0.1930492135971588, + "grad_norm": 4.90625, + "learning_rate": 9.94093895696156e-06, + "loss": 1.6759016, + "memory(GiB)": 71.21, + "step": 7610, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.63412328, + "epoch": 0.19317605276509386, + "grad_norm": 5.6875, + "learning_rate": 9.940778149026038e-06, + "loss": 1.71386738, + "memory(GiB)": 71.21, + "step": 7615, + "train_speed(iter/s)": 0.448604 + }, + { + "acc": 0.63581343, + "epoch": 0.1933028919330289, + "grad_norm": 5.125, + "learning_rate": 9.940617123773036e-06, + "loss": 1.69941521, + "memory(GiB)": 71.21, + "step": 7620, + "train_speed(iter/s)": 0.448758 + }, + { + "acc": 0.65294085, + "epoch": 0.19342973110096398, + "grad_norm": 5.53125, + "learning_rate": 9.940455881209632e-06, + "loss": 1.59593277, + "memory(GiB)": 71.21, + "step": 7625, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.64158449, + "epoch": 0.19355657026889903, + "grad_norm": 5.25, + "learning_rate": 9.940294421342922e-06, + "loss": 1.63792305, + "memory(GiB)": 71.21, + "step": 7630, + "train_speed(iter/s)": 0.449068 + }, + { + "acc": 0.6248044, + "epoch": 0.1936834094368341, + "grad_norm": 5.90625, + "learning_rate": 9.940132744180007e-06, + "loss": 1.74248085, + "memory(GiB)": 71.21, + "step": 7635, + "train_speed(iter/s)": 0.449223 + }, + { + "acc": 0.6530365, + "epoch": 0.19381024860476914, + "grad_norm": 5.09375, + "learning_rate": 9.939970849727995e-06, + "loss": 1.62175636, + "memory(GiB)": 71.21, + "step": 7640, + "train_speed(iter/s)": 0.449378 + }, + { + "acc": 0.62256699, + "epoch": 0.19393708777270421, + "grad_norm": 5.4375, + "learning_rate": 9.939808737994013e-06, + "loss": 1.7660099, + "memory(GiB)": 71.21, + "step": 7645, + "train_speed(iter/s)": 0.449532 + }, + { + "acc": 0.64405804, + "epoch": 0.19406392694063926, + "grad_norm": 5.03125, + "learning_rate": 9.939646408985186e-06, + "loss": 1.63705406, + "memory(GiB)": 71.21, + "step": 7650, + "train_speed(iter/s)": 0.449686 + }, + { + "acc": 0.65222592, + "epoch": 0.19419076610857433, + "grad_norm": 5.65625, + "learning_rate": 9.939483862708658e-06, + "loss": 1.62148476, + "memory(GiB)": 71.21, + "step": 7655, + "train_speed(iter/s)": 0.44984 + }, + { + "acc": 0.63589873, + "epoch": 0.19431760527650938, + "grad_norm": 5.59375, + "learning_rate": 9.939321099171575e-06, + "loss": 1.65142441, + "memory(GiB)": 71.21, + "step": 7660, + "train_speed(iter/s)": 0.449993 + }, + { + "acc": 0.63626051, + "epoch": 0.19444444444444445, + "grad_norm": 4.8125, + "learning_rate": 9.939158118381097e-06, + "loss": 1.71466312, + "memory(GiB)": 71.21, + "step": 7665, + "train_speed(iter/s)": 0.450146 + }, + { + "acc": 0.63825407, + "epoch": 0.1945712836123795, + "grad_norm": 5.71875, + "learning_rate": 9.938994920344395e-06, + "loss": 1.69848709, + "memory(GiB)": 71.21, + "step": 7670, + "train_speed(iter/s)": 0.4503 + }, + { + "acc": 0.64246707, + "epoch": 0.19469812278031456, + "grad_norm": 5.03125, + "learning_rate": 9.938831505068645e-06, + "loss": 1.65335846, + "memory(GiB)": 71.21, + "step": 7675, + "train_speed(iter/s)": 0.450456 + }, + { + "acc": 0.64241924, + "epoch": 0.1948249619482496, + "grad_norm": 5.25, + "learning_rate": 9.938667872561035e-06, + "loss": 1.62154961, + "memory(GiB)": 71.21, + "step": 7680, + "train_speed(iter/s)": 0.45061 + }, + { + "acc": 0.6567832, + "epoch": 0.19495180111618468, + "grad_norm": 5.25, + "learning_rate": 9.938504022828762e-06, + "loss": 1.63594303, + "memory(GiB)": 71.21, + "step": 7685, + "train_speed(iter/s)": 0.450764 + }, + { + "acc": 0.64295287, + "epoch": 0.19507864028411973, + "grad_norm": 4.125, + "learning_rate": 9.938339955879033e-06, + "loss": 1.65605888, + "memory(GiB)": 71.21, + "step": 7690, + "train_speed(iter/s)": 0.450918 + }, + { + "acc": 0.63837428, + "epoch": 0.1952054794520548, + "grad_norm": 4.9375, + "learning_rate": 9.938175671719064e-06, + "loss": 1.71003571, + "memory(GiB)": 71.21, + "step": 7695, + "train_speed(iter/s)": 0.451071 + }, + { + "acc": 0.62362108, + "epoch": 0.19533231861998984, + "grad_norm": 5.0, + "learning_rate": 9.938011170356083e-06, + "loss": 1.66456642, + "memory(GiB)": 71.21, + "step": 7700, + "train_speed(iter/s)": 0.451226 + }, + { + "epoch": 0.19533231861998984, + "eval_acc": 0.6330766431728257, + "eval_loss": 1.6685415506362915, + "eval_runtime": 113.1115, + "eval_samples_per_second": 56.316, + "eval_steps_per_second": 28.158, + "step": 7700 + }, + { + "acc": 0.63569598, + "epoch": 0.19545915778792491, + "grad_norm": 7.96875, + "learning_rate": 9.937846451797324e-06, + "loss": 1.7004097, + "memory(GiB)": 71.21, + "step": 7705, + "train_speed(iter/s)": 0.448291 + }, + { + "acc": 0.64071584, + "epoch": 0.19558599695585996, + "grad_norm": 6.5625, + "learning_rate": 9.93768151605003e-06, + "loss": 1.64734421, + "memory(GiB)": 71.21, + "step": 7710, + "train_speed(iter/s)": 0.448443 + }, + { + "acc": 0.6473577, + "epoch": 0.19571283612379503, + "grad_norm": 6.0625, + "learning_rate": 9.93751636312146e-06, + "loss": 1.63463402, + "memory(GiB)": 71.21, + "step": 7715, + "train_speed(iter/s)": 0.448596 + }, + { + "acc": 0.62583714, + "epoch": 0.19583967529173008, + "grad_norm": 5.78125, + "learning_rate": 9.937350993018875e-06, + "loss": 1.78231087, + "memory(GiB)": 71.21, + "step": 7720, + "train_speed(iter/s)": 0.448748 + }, + { + "acc": 0.63588905, + "epoch": 0.19596651445966515, + "grad_norm": 4.6875, + "learning_rate": 9.93718540574955e-06, + "loss": 1.67965469, + "memory(GiB)": 71.21, + "step": 7725, + "train_speed(iter/s)": 0.448901 + }, + { + "acc": 0.62698727, + "epoch": 0.1960933536276002, + "grad_norm": 4.8125, + "learning_rate": 9.937019601320768e-06, + "loss": 1.75032501, + "memory(GiB)": 71.21, + "step": 7730, + "train_speed(iter/s)": 0.449055 + }, + { + "acc": 0.63055353, + "epoch": 0.19622019279553526, + "grad_norm": 5.5, + "learning_rate": 9.936853579739823e-06, + "loss": 1.69595947, + "memory(GiB)": 71.21, + "step": 7735, + "train_speed(iter/s)": 0.449207 + }, + { + "acc": 0.62087431, + "epoch": 0.1963470319634703, + "grad_norm": 6.625, + "learning_rate": 9.936687341014015e-06, + "loss": 1.74901276, + "memory(GiB)": 71.21, + "step": 7740, + "train_speed(iter/s)": 0.449361 + }, + { + "acc": 0.65325327, + "epoch": 0.19647387113140538, + "grad_norm": 4.9375, + "learning_rate": 9.936520885150655e-06, + "loss": 1.61029797, + "memory(GiB)": 71.21, + "step": 7745, + "train_speed(iter/s)": 0.449513 + }, + { + "acc": 0.6420105, + "epoch": 0.19660071029934043, + "grad_norm": 5.03125, + "learning_rate": 9.936354212157068e-06, + "loss": 1.735359, + "memory(GiB)": 71.21, + "step": 7750, + "train_speed(iter/s)": 0.449667 + }, + { + "acc": 0.6397995, + "epoch": 0.1967275494672755, + "grad_norm": 4.3125, + "learning_rate": 9.936187322040584e-06, + "loss": 1.70557671, + "memory(GiB)": 71.21, + "step": 7755, + "train_speed(iter/s)": 0.449819 + }, + { + "acc": 0.62862849, + "epoch": 0.19685438863521054, + "grad_norm": 4.96875, + "learning_rate": 9.936020214808544e-06, + "loss": 1.73480721, + "memory(GiB)": 71.21, + "step": 7760, + "train_speed(iter/s)": 0.449971 + }, + { + "acc": 0.64502392, + "epoch": 0.19698122780314561, + "grad_norm": 5.1875, + "learning_rate": 9.935852890468297e-06, + "loss": 1.67905083, + "memory(GiB)": 71.21, + "step": 7765, + "train_speed(iter/s)": 0.450123 + }, + { + "acc": 0.6280983, + "epoch": 0.19710806697108066, + "grad_norm": 6.84375, + "learning_rate": 9.935685349027201e-06, + "loss": 1.70416718, + "memory(GiB)": 71.21, + "step": 7770, + "train_speed(iter/s)": 0.450276 + }, + { + "acc": 0.63208303, + "epoch": 0.19723490613901573, + "grad_norm": 4.9375, + "learning_rate": 9.935517590492627e-06, + "loss": 1.69104424, + "memory(GiB)": 71.21, + "step": 7775, + "train_speed(iter/s)": 0.450428 + }, + { + "acc": 0.6499599, + "epoch": 0.19736174530695078, + "grad_norm": 5.0625, + "learning_rate": 9.935349614871957e-06, + "loss": 1.63721924, + "memory(GiB)": 71.21, + "step": 7780, + "train_speed(iter/s)": 0.45058 + }, + { + "acc": 0.64528522, + "epoch": 0.19748858447488585, + "grad_norm": 5.15625, + "learning_rate": 9.935181422172574e-06, + "loss": 1.67749386, + "memory(GiB)": 71.21, + "step": 7785, + "train_speed(iter/s)": 0.450733 + }, + { + "acc": 0.62817917, + "epoch": 0.1976154236428209, + "grad_norm": 6.09375, + "learning_rate": 9.935013012401878e-06, + "loss": 1.79723186, + "memory(GiB)": 71.21, + "step": 7790, + "train_speed(iter/s)": 0.450885 + }, + { + "acc": 0.65724416, + "epoch": 0.19774226281075596, + "grad_norm": 5.125, + "learning_rate": 9.934844385567275e-06, + "loss": 1.61300316, + "memory(GiB)": 71.21, + "step": 7795, + "train_speed(iter/s)": 0.451039 + }, + { + "acc": 0.64110975, + "epoch": 0.197869101978691, + "grad_norm": 5.875, + "learning_rate": 9.934675541676186e-06, + "loss": 1.69013252, + "memory(GiB)": 71.21, + "step": 7800, + "train_speed(iter/s)": 0.451192 + }, + { + "epoch": 0.197869101978691, + "eval_acc": 0.6332262038696115, + "eval_loss": 1.6671642065048218, + "eval_runtime": 113.4528, + "eval_samples_per_second": 56.147, + "eval_steps_per_second": 28.073, + "step": 7800 + }, + { + "acc": 0.6523982, + "epoch": 0.19799594114662608, + "grad_norm": 6.46875, + "learning_rate": 9.934506480736034e-06, + "loss": 1.63810654, + "memory(GiB)": 71.21, + "step": 7805, + "train_speed(iter/s)": 0.448287 + }, + { + "acc": 0.65067701, + "epoch": 0.19812278031456113, + "grad_norm": 5.34375, + "learning_rate": 9.934337202754257e-06, + "loss": 1.67652397, + "memory(GiB)": 71.21, + "step": 7810, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.64005933, + "epoch": 0.1982496194824962, + "grad_norm": 6.0, + "learning_rate": 9.934167707738298e-06, + "loss": 1.73353348, + "memory(GiB)": 71.21, + "step": 7815, + "train_speed(iter/s)": 0.448593 + }, + { + "acc": 0.646842, + "epoch": 0.19837645865043124, + "grad_norm": 5.625, + "learning_rate": 9.933997995695615e-06, + "loss": 1.71121407, + "memory(GiB)": 71.21, + "step": 7820, + "train_speed(iter/s)": 0.448744 + }, + { + "acc": 0.6337667, + "epoch": 0.19850329781836631, + "grad_norm": 5.21875, + "learning_rate": 9.93382806663367e-06, + "loss": 1.7432972, + "memory(GiB)": 71.21, + "step": 7825, + "train_speed(iter/s)": 0.448896 + }, + { + "acc": 0.64032583, + "epoch": 0.19863013698630136, + "grad_norm": 5.15625, + "learning_rate": 9.933657920559939e-06, + "loss": 1.68478546, + "memory(GiB)": 71.21, + "step": 7830, + "train_speed(iter/s)": 0.449047 + }, + { + "acc": 0.64448905, + "epoch": 0.19875697615423643, + "grad_norm": 5.03125, + "learning_rate": 9.933487557481905e-06, + "loss": 1.69543915, + "memory(GiB)": 71.21, + "step": 7835, + "train_speed(iter/s)": 0.449198 + }, + { + "acc": 0.64522381, + "epoch": 0.19888381532217148, + "grad_norm": 5.0625, + "learning_rate": 9.933316977407063e-06, + "loss": 1.73755608, + "memory(GiB)": 71.21, + "step": 7840, + "train_speed(iter/s)": 0.449349 + }, + { + "acc": 0.65792408, + "epoch": 0.19901065449010655, + "grad_norm": 4.78125, + "learning_rate": 9.933146180342914e-06, + "loss": 1.56299667, + "memory(GiB)": 71.21, + "step": 7845, + "train_speed(iter/s)": 0.449501 + }, + { + "acc": 0.6573257, + "epoch": 0.1991374936580416, + "grad_norm": 4.65625, + "learning_rate": 9.932975166296972e-06, + "loss": 1.67728119, + "memory(GiB)": 71.21, + "step": 7850, + "train_speed(iter/s)": 0.449651 + }, + { + "acc": 0.64167662, + "epoch": 0.19926433282597666, + "grad_norm": 5.90625, + "learning_rate": 9.932803935276757e-06, + "loss": 1.63266239, + "memory(GiB)": 71.21, + "step": 7855, + "train_speed(iter/s)": 0.449801 + }, + { + "acc": 0.62765436, + "epoch": 0.1993911719939117, + "grad_norm": 6.21875, + "learning_rate": 9.932632487289802e-06, + "loss": 1.75699196, + "memory(GiB)": 71.21, + "step": 7860, + "train_speed(iter/s)": 0.449952 + }, + { + "acc": 0.63972826, + "epoch": 0.19951801116184678, + "grad_norm": 5.15625, + "learning_rate": 9.932460822343649e-06, + "loss": 1.67174435, + "memory(GiB)": 71.21, + "step": 7865, + "train_speed(iter/s)": 0.450103 + }, + { + "acc": 0.62979274, + "epoch": 0.19964485032978183, + "grad_norm": 5.78125, + "learning_rate": 9.932288940445845e-06, + "loss": 1.81322289, + "memory(GiB)": 71.21, + "step": 7870, + "train_speed(iter/s)": 0.450254 + }, + { + "acc": 0.63745799, + "epoch": 0.1997716894977169, + "grad_norm": 5.875, + "learning_rate": 9.932116841603954e-06, + "loss": 1.64593296, + "memory(GiB)": 71.21, + "step": 7875, + "train_speed(iter/s)": 0.450406 + }, + { + "acc": 0.64807987, + "epoch": 0.19989852866565194, + "grad_norm": 6.875, + "learning_rate": 9.931944525825542e-06, + "loss": 1.69942818, + "memory(GiB)": 71.21, + "step": 7880, + "train_speed(iter/s)": 0.450557 + }, + { + "acc": 0.63484945, + "epoch": 0.20002536783358701, + "grad_norm": 5.875, + "learning_rate": 9.931771993118191e-06, + "loss": 1.61795902, + "memory(GiB)": 71.21, + "step": 7885, + "train_speed(iter/s)": 0.450707 + }, + { + "acc": 0.64385376, + "epoch": 0.20015220700152206, + "grad_norm": 4.3125, + "learning_rate": 9.931599243489489e-06, + "loss": 1.72382717, + "memory(GiB)": 71.21, + "step": 7890, + "train_speed(iter/s)": 0.450857 + }, + { + "acc": 0.64237194, + "epoch": 0.20027904616945713, + "grad_norm": 5.5625, + "learning_rate": 9.931426276947037e-06, + "loss": 1.65209923, + "memory(GiB)": 71.21, + "step": 7895, + "train_speed(iter/s)": 0.451007 + }, + { + "acc": 0.63275084, + "epoch": 0.20040588533739218, + "grad_norm": 5.0, + "learning_rate": 9.931253093498437e-06, + "loss": 1.67295685, + "memory(GiB)": 71.21, + "step": 7900, + "train_speed(iter/s)": 0.451157 + }, + { + "epoch": 0.20040588533739218, + "eval_acc": 0.6334927394130453, + "eval_loss": 1.6666460037231445, + "eval_runtime": 113.2497, + "eval_samples_per_second": 56.247, + "eval_steps_per_second": 28.124, + "step": 7900 + }, + { + "acc": 0.63523302, + "epoch": 0.20053272450532725, + "grad_norm": 4.0625, + "learning_rate": 9.93107969315131e-06, + "loss": 1.69254799, + "memory(GiB)": 71.21, + "step": 7905, + "train_speed(iter/s)": 0.448295 + }, + { + "acc": 0.65175066, + "epoch": 0.2006595636732623, + "grad_norm": 4.65625, + "learning_rate": 9.930906075913281e-06, + "loss": 1.68473587, + "memory(GiB)": 71.21, + "step": 7910, + "train_speed(iter/s)": 0.448445 + }, + { + "acc": 0.65698614, + "epoch": 0.20078640284119736, + "grad_norm": 4.5625, + "learning_rate": 9.93073224179199e-06, + "loss": 1.61309128, + "memory(GiB)": 71.21, + "step": 7915, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.63625751, + "epoch": 0.2009132420091324, + "grad_norm": 6.03125, + "learning_rate": 9.93055819079508e-06, + "loss": 1.76167679, + "memory(GiB)": 71.21, + "step": 7920, + "train_speed(iter/s)": 0.448745 + }, + { + "acc": 0.64582615, + "epoch": 0.20104008117706748, + "grad_norm": 4.4375, + "learning_rate": 9.930383922930207e-06, + "loss": 1.73017941, + "memory(GiB)": 71.21, + "step": 7925, + "train_speed(iter/s)": 0.448742 + }, + { + "acc": 0.6340169, + "epoch": 0.20116692034500253, + "grad_norm": 5.84375, + "learning_rate": 9.930209438205038e-06, + "loss": 1.72401924, + "memory(GiB)": 71.21, + "step": 7930, + "train_speed(iter/s)": 0.448891 + }, + { + "acc": 0.6383214, + "epoch": 0.2012937595129376, + "grad_norm": 4.71875, + "learning_rate": 9.930034736627245e-06, + "loss": 1.69098244, + "memory(GiB)": 71.21, + "step": 7935, + "train_speed(iter/s)": 0.449041 + }, + { + "acc": 0.65146866, + "epoch": 0.20142059868087264, + "grad_norm": 5.125, + "learning_rate": 9.929859818204514e-06, + "loss": 1.66798973, + "memory(GiB)": 71.21, + "step": 7940, + "train_speed(iter/s)": 0.44919 + }, + { + "acc": 0.66242647, + "epoch": 0.20154743784880771, + "grad_norm": 4.8125, + "learning_rate": 9.929684682944538e-06, + "loss": 1.54506779, + "memory(GiB)": 71.21, + "step": 7945, + "train_speed(iter/s)": 0.44934 + }, + { + "acc": 0.63335748, + "epoch": 0.20167427701674276, + "grad_norm": 5.53125, + "learning_rate": 9.929509330855018e-06, + "loss": 1.71219521, + "memory(GiB)": 71.21, + "step": 7950, + "train_speed(iter/s)": 0.44949 + }, + { + "acc": 0.64373326, + "epoch": 0.20180111618467783, + "grad_norm": 4.90625, + "learning_rate": 9.929333761943672e-06, + "loss": 1.73696709, + "memory(GiB)": 71.21, + "step": 7955, + "train_speed(iter/s)": 0.449639 + }, + { + "acc": 0.62869987, + "epoch": 0.20192795535261288, + "grad_norm": 5.90625, + "learning_rate": 9.929157976218218e-06, + "loss": 1.75948296, + "memory(GiB)": 71.21, + "step": 7960, + "train_speed(iter/s)": 0.449788 + }, + { + "acc": 0.63892293, + "epoch": 0.20205479452054795, + "grad_norm": 6.0, + "learning_rate": 9.928981973686388e-06, + "loss": 1.73255825, + "memory(GiB)": 71.21, + "step": 7965, + "train_speed(iter/s)": 0.449937 + }, + { + "acc": 0.65110106, + "epoch": 0.202181633688483, + "grad_norm": 4.5, + "learning_rate": 9.928805754355926e-06, + "loss": 1.60163498, + "memory(GiB)": 71.21, + "step": 7970, + "train_speed(iter/s)": 0.450085 + }, + { + "acc": 0.63845358, + "epoch": 0.20230847285641806, + "grad_norm": 5.9375, + "learning_rate": 9.92862931823458e-06, + "loss": 1.66533623, + "memory(GiB)": 71.21, + "step": 7975, + "train_speed(iter/s)": 0.450236 + }, + { + "acc": 0.64788485, + "epoch": 0.2024353120243531, + "grad_norm": 5.34375, + "learning_rate": 9.928452665330113e-06, + "loss": 1.71982994, + "memory(GiB)": 71.21, + "step": 7980, + "train_speed(iter/s)": 0.450385 + }, + { + "acc": 0.63263912, + "epoch": 0.20256215119228818, + "grad_norm": 4.84375, + "learning_rate": 9.928275795650293e-06, + "loss": 1.72257652, + "memory(GiB)": 71.21, + "step": 7985, + "train_speed(iter/s)": 0.450535 + }, + { + "acc": 0.64622374, + "epoch": 0.20268899036022323, + "grad_norm": 5.0625, + "learning_rate": 9.928098709202901e-06, + "loss": 1.66977768, + "memory(GiB)": 71.21, + "step": 7990, + "train_speed(iter/s)": 0.450683 + }, + { + "acc": 0.6344285, + "epoch": 0.2028158295281583, + "grad_norm": 7.28125, + "learning_rate": 9.927921405995727e-06, + "loss": 1.73109589, + "memory(GiB)": 71.21, + "step": 7995, + "train_speed(iter/s)": 0.450832 + }, + { + "acc": 0.65770769, + "epoch": 0.20294266869609334, + "grad_norm": 5.15625, + "learning_rate": 9.927743886036566e-06, + "loss": 1.6832737, + "memory(GiB)": 71.21, + "step": 8000, + "train_speed(iter/s)": 0.450981 + }, + { + "epoch": 0.20294266869609334, + "eval_acc": 0.6335274140997302, + "eval_loss": 1.665919303894043, + "eval_runtime": 113.4819, + "eval_samples_per_second": 56.132, + "eval_steps_per_second": 28.066, + "step": 8000 + }, + { + "acc": 0.65675073, + "epoch": 0.20306950786402841, + "grad_norm": 5.0, + "learning_rate": 9.927566149333228e-06, + "loss": 1.60690155, + "memory(GiB)": 71.21, + "step": 8005, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.6301918, + "epoch": 0.20319634703196346, + "grad_norm": 5.65625, + "learning_rate": 9.92738819589353e-06, + "loss": 1.67938118, + "memory(GiB)": 71.21, + "step": 8010, + "train_speed(iter/s)": 0.448293 + }, + { + "acc": 0.63772144, + "epoch": 0.20332318619989853, + "grad_norm": 5.9375, + "learning_rate": 9.927210025725301e-06, + "loss": 1.66887665, + "memory(GiB)": 71.21, + "step": 8015, + "train_speed(iter/s)": 0.448441 + }, + { + "acc": 0.63797626, + "epoch": 0.20345002536783358, + "grad_norm": 5.4375, + "learning_rate": 9.927031638836377e-06, + "loss": 1.68546829, + "memory(GiB)": 71.21, + "step": 8020, + "train_speed(iter/s)": 0.448589 + }, + { + "acc": 0.64352541, + "epoch": 0.20357686453576865, + "grad_norm": 6.71875, + "learning_rate": 9.926853035234603e-06, + "loss": 1.69362869, + "memory(GiB)": 71.21, + "step": 8025, + "train_speed(iter/s)": 0.448737 + }, + { + "acc": 0.64745083, + "epoch": 0.2037037037037037, + "grad_norm": 5.71875, + "learning_rate": 9.926674214927836e-06, + "loss": 1.74950905, + "memory(GiB)": 71.21, + "step": 8030, + "train_speed(iter/s)": 0.448883 + }, + { + "acc": 0.63772974, + "epoch": 0.20383054287163876, + "grad_norm": 4.5625, + "learning_rate": 9.926495177923941e-06, + "loss": 1.66350918, + "memory(GiB)": 71.21, + "step": 8035, + "train_speed(iter/s)": 0.449032 + }, + { + "acc": 0.64252672, + "epoch": 0.2039573820395738, + "grad_norm": 5.625, + "learning_rate": 9.926315924230794e-06, + "loss": 1.70670547, + "memory(GiB)": 71.21, + "step": 8040, + "train_speed(iter/s)": 0.449181 + }, + { + "acc": 0.63815174, + "epoch": 0.20408422120750888, + "grad_norm": 4.96875, + "learning_rate": 9.926136453856277e-06, + "loss": 1.67611732, + "memory(GiB)": 71.21, + "step": 8045, + "train_speed(iter/s)": 0.449325 + }, + { + "acc": 0.64188876, + "epoch": 0.20421106037544393, + "grad_norm": 5.09375, + "learning_rate": 9.925956766808286e-06, + "loss": 1.68995171, + "memory(GiB)": 71.21, + "step": 8050, + "train_speed(iter/s)": 0.449473 + }, + { + "acc": 0.62811117, + "epoch": 0.204337899543379, + "grad_norm": 4.6875, + "learning_rate": 9.925776863094723e-06, + "loss": 1.70620766, + "memory(GiB)": 71.21, + "step": 8055, + "train_speed(iter/s)": 0.449621 + }, + { + "acc": 0.63177853, + "epoch": 0.20446473871131404, + "grad_norm": 4.90625, + "learning_rate": 9.9255967427235e-06, + "loss": 1.72925949, + "memory(GiB)": 71.21, + "step": 8060, + "train_speed(iter/s)": 0.449769 + }, + { + "acc": 0.64552116, + "epoch": 0.20459157787924911, + "grad_norm": 5.1875, + "learning_rate": 9.925416405702544e-06, + "loss": 1.67127571, + "memory(GiB)": 71.21, + "step": 8065, + "train_speed(iter/s)": 0.449916 + }, + { + "acc": 0.64164724, + "epoch": 0.20471841704718416, + "grad_norm": 5.125, + "learning_rate": 9.925235852039783e-06, + "loss": 1.70809174, + "memory(GiB)": 71.21, + "step": 8070, + "train_speed(iter/s)": 0.450063 + }, + { + "acc": 0.63770175, + "epoch": 0.20484525621511923, + "grad_norm": 5.0, + "learning_rate": 9.92505508174316e-06, + "loss": 1.74251461, + "memory(GiB)": 71.21, + "step": 8075, + "train_speed(iter/s)": 0.450211 + }, + { + "acc": 0.6505959, + "epoch": 0.20497209538305428, + "grad_norm": 7.0625, + "learning_rate": 9.924874094820625e-06, + "loss": 1.65471878, + "memory(GiB)": 71.21, + "step": 8080, + "train_speed(iter/s)": 0.450358 + }, + { + "acc": 0.64262419, + "epoch": 0.20509893455098935, + "grad_norm": 6.46875, + "learning_rate": 9.924692891280139e-06, + "loss": 1.6950058, + "memory(GiB)": 71.21, + "step": 8085, + "train_speed(iter/s)": 0.450506 + }, + { + "acc": 0.6454524, + "epoch": 0.2052257737189244, + "grad_norm": 5.28125, + "learning_rate": 9.924511471129673e-06, + "loss": 1.65451469, + "memory(GiB)": 71.21, + "step": 8090, + "train_speed(iter/s)": 0.450653 + }, + { + "acc": 0.63959908, + "epoch": 0.20535261288685946, + "grad_norm": 5.15625, + "learning_rate": 9.924329834377206e-06, + "loss": 1.65727654, + "memory(GiB)": 71.21, + "step": 8095, + "train_speed(iter/s)": 0.450799 + }, + { + "acc": 0.64031057, + "epoch": 0.2054794520547945, + "grad_norm": 5.875, + "learning_rate": 9.924147981030728e-06, + "loss": 1.68533363, + "memory(GiB)": 71.21, + "step": 8100, + "train_speed(iter/s)": 0.450946 + }, + { + "epoch": 0.2054794520547945, + "eval_acc": 0.6337396398929346, + "eval_loss": 1.6647824048995972, + "eval_runtime": 112.79, + "eval_samples_per_second": 56.477, + "eval_steps_per_second": 28.238, + "step": 8100 + }, + { + "acc": 0.65359106, + "epoch": 0.20560629122272958, + "grad_norm": 5.03125, + "learning_rate": 9.923965911098235e-06, + "loss": 1.65648003, + "memory(GiB)": 71.21, + "step": 8105, + "train_speed(iter/s)": 0.448166 + }, + { + "acc": 0.62566323, + "epoch": 0.20573313039066463, + "grad_norm": 7.0625, + "learning_rate": 9.92378362458774e-06, + "loss": 1.77997799, + "memory(GiB)": 71.21, + "step": 8110, + "train_speed(iter/s)": 0.448312 + }, + { + "acc": 0.64280653, + "epoch": 0.2058599695585997, + "grad_norm": 5.0625, + "learning_rate": 9.923601121507256e-06, + "loss": 1.69552689, + "memory(GiB)": 71.21, + "step": 8115, + "train_speed(iter/s)": 0.448456 + }, + { + "acc": 0.6256156, + "epoch": 0.20598680872653474, + "grad_norm": 4.875, + "learning_rate": 9.923418401864812e-06, + "loss": 1.71542816, + "memory(GiB)": 71.21, + "step": 8120, + "train_speed(iter/s)": 0.448601 + }, + { + "acc": 0.63545871, + "epoch": 0.20611364789446981, + "grad_norm": 5.5625, + "learning_rate": 9.923235465668447e-06, + "loss": 1.65908432, + "memory(GiB)": 71.21, + "step": 8125, + "train_speed(iter/s)": 0.448746 + }, + { + "acc": 0.64828691, + "epoch": 0.20624048706240486, + "grad_norm": 5.46875, + "learning_rate": 9.923052312926204e-06, + "loss": 1.66910572, + "memory(GiB)": 71.21, + "step": 8130, + "train_speed(iter/s)": 0.448891 + }, + { + "acc": 0.6419198, + "epoch": 0.20636732623033993, + "grad_norm": 4.46875, + "learning_rate": 9.922868943646142e-06, + "loss": 1.6480854, + "memory(GiB)": 71.21, + "step": 8135, + "train_speed(iter/s)": 0.449035 + }, + { + "acc": 0.63529701, + "epoch": 0.20649416539827498, + "grad_norm": 5.75, + "learning_rate": 9.922685357836324e-06, + "loss": 1.69019833, + "memory(GiB)": 71.21, + "step": 8140, + "train_speed(iter/s)": 0.449179 + }, + { + "acc": 0.63884268, + "epoch": 0.20662100456621005, + "grad_norm": 5.40625, + "learning_rate": 9.922501555504827e-06, + "loss": 1.69838448, + "memory(GiB)": 71.21, + "step": 8145, + "train_speed(iter/s)": 0.449324 + }, + { + "acc": 0.64210405, + "epoch": 0.2067478437341451, + "grad_norm": 4.71875, + "learning_rate": 9.922317536659733e-06, + "loss": 1.63487358, + "memory(GiB)": 71.21, + "step": 8150, + "train_speed(iter/s)": 0.449468 + }, + { + "acc": 0.64256277, + "epoch": 0.20687468290208016, + "grad_norm": 4.90625, + "learning_rate": 9.922133301309136e-06, + "loss": 1.69277077, + "memory(GiB)": 71.21, + "step": 8155, + "train_speed(iter/s)": 0.449613 + }, + { + "acc": 0.63550062, + "epoch": 0.2070015220700152, + "grad_norm": 6.21875, + "learning_rate": 9.921948849461142e-06, + "loss": 1.70646725, + "memory(GiB)": 71.21, + "step": 8160, + "train_speed(iter/s)": 0.449758 + }, + { + "acc": 0.65031271, + "epoch": 0.20712836123795028, + "grad_norm": 5.25, + "learning_rate": 9.921764181123864e-06, + "loss": 1.5951025, + "memory(GiB)": 71.21, + "step": 8165, + "train_speed(iter/s)": 0.449902 + }, + { + "acc": 0.65544124, + "epoch": 0.20725520040588533, + "grad_norm": 6.65625, + "learning_rate": 9.921579296305421e-06, + "loss": 1.62906513, + "memory(GiB)": 71.21, + "step": 8170, + "train_speed(iter/s)": 0.450047 + }, + { + "acc": 0.63458395, + "epoch": 0.2073820395738204, + "grad_norm": 5.5625, + "learning_rate": 9.921394195013949e-06, + "loss": 1.63672066, + "memory(GiB)": 71.21, + "step": 8175, + "train_speed(iter/s)": 0.450192 + }, + { + "acc": 0.64468884, + "epoch": 0.20750887874175544, + "grad_norm": 4.8125, + "learning_rate": 9.921208877257586e-06, + "loss": 1.67782326, + "memory(GiB)": 71.21, + "step": 8180, + "train_speed(iter/s)": 0.450335 + }, + { + "acc": 0.63539209, + "epoch": 0.20763571790969051, + "grad_norm": 6.25, + "learning_rate": 9.921023343044486e-06, + "loss": 1.66944523, + "memory(GiB)": 71.21, + "step": 8185, + "train_speed(iter/s)": 0.450479 + }, + { + "acc": 0.64285927, + "epoch": 0.20776255707762556, + "grad_norm": 7.21875, + "learning_rate": 9.92083759238281e-06, + "loss": 1.62081184, + "memory(GiB)": 71.21, + "step": 8190, + "train_speed(iter/s)": 0.450622 + }, + { + "acc": 0.64389386, + "epoch": 0.20788939624556063, + "grad_norm": 5.0625, + "learning_rate": 9.920651625280725e-06, + "loss": 1.7454895, + "memory(GiB)": 71.21, + "step": 8195, + "train_speed(iter/s)": 0.450767 + }, + { + "acc": 0.63974924, + "epoch": 0.20801623541349568, + "grad_norm": 5.8125, + "learning_rate": 9.920465441746412e-06, + "loss": 1.71090851, + "memory(GiB)": 71.21, + "step": 8200, + "train_speed(iter/s)": 0.450913 + }, + { + "epoch": 0.20801623541349568, + "eval_acc": 0.6339146843955973, + "eval_loss": 1.6637836694717407, + "eval_runtime": 113.5595, + "eval_samples_per_second": 56.094, + "eval_steps_per_second": 28.047, + "step": 8200 + }, + { + "acc": 0.63911419, + "epoch": 0.20814307458143075, + "grad_norm": 4.71875, + "learning_rate": 9.920279041788062e-06, + "loss": 1.68976269, + "memory(GiB)": 71.21, + "step": 8205, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.6528338, + "epoch": 0.2082699137493658, + "grad_norm": 5.375, + "learning_rate": 9.920092425413871e-06, + "loss": 1.68342724, + "memory(GiB)": 71.21, + "step": 8210, + "train_speed(iter/s)": 0.44829 + }, + { + "acc": 0.64426146, + "epoch": 0.20839675291730086, + "grad_norm": 5.375, + "learning_rate": 9.919905592632048e-06, + "loss": 1.62852592, + "memory(GiB)": 71.21, + "step": 8215, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.6371129, + "epoch": 0.2085235920852359, + "grad_norm": 6.15625, + "learning_rate": 9.919718543450813e-06, + "loss": 1.64365387, + "memory(GiB)": 71.21, + "step": 8220, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.64223194, + "epoch": 0.20865043125317098, + "grad_norm": 5.59375, + "learning_rate": 9.919531277878391e-06, + "loss": 1.68323517, + "memory(GiB)": 71.21, + "step": 8225, + "train_speed(iter/s)": 0.44872 + }, + { + "acc": 0.64952016, + "epoch": 0.20877727042110603, + "grad_norm": 7.75, + "learning_rate": 9.91934379592302e-06, + "loss": 1.55598221, + "memory(GiB)": 71.21, + "step": 8230, + "train_speed(iter/s)": 0.448864 + }, + { + "acc": 0.6534863, + "epoch": 0.2089041095890411, + "grad_norm": 4.59375, + "learning_rate": 9.919156097592944e-06, + "loss": 1.59364061, + "memory(GiB)": 71.21, + "step": 8235, + "train_speed(iter/s)": 0.449007 + }, + { + "acc": 0.65045528, + "epoch": 0.20903094875697614, + "grad_norm": 5.40625, + "learning_rate": 9.91896818289642e-06, + "loss": 1.63226776, + "memory(GiB)": 71.21, + "step": 8240, + "train_speed(iter/s)": 0.44915 + }, + { + "acc": 0.65358953, + "epoch": 0.20915778792491121, + "grad_norm": 5.0625, + "learning_rate": 9.918780051841716e-06, + "loss": 1.67622566, + "memory(GiB)": 71.21, + "step": 8245, + "train_speed(iter/s)": 0.449292 + }, + { + "acc": 0.64208689, + "epoch": 0.20928462709284626, + "grad_norm": 4.9375, + "learning_rate": 9.918591704437103e-06, + "loss": 1.67470646, + "memory(GiB)": 71.21, + "step": 8250, + "train_speed(iter/s)": 0.449436 + }, + { + "acc": 0.63879557, + "epoch": 0.20941146626078133, + "grad_norm": 5.34375, + "learning_rate": 9.918403140690866e-06, + "loss": 1.66502895, + "memory(GiB)": 71.21, + "step": 8255, + "train_speed(iter/s)": 0.44958 + }, + { + "acc": 0.63374705, + "epoch": 0.20953830542871638, + "grad_norm": 4.59375, + "learning_rate": 9.918214360611302e-06, + "loss": 1.68402672, + "memory(GiB)": 71.21, + "step": 8260, + "train_speed(iter/s)": 0.449721 + }, + { + "acc": 0.63472929, + "epoch": 0.20966514459665145, + "grad_norm": 6.875, + "learning_rate": 9.918025364206712e-06, + "loss": 1.69764442, + "memory(GiB)": 71.21, + "step": 8265, + "train_speed(iter/s)": 0.449862 + }, + { + "acc": 0.64071884, + "epoch": 0.2097919837645865, + "grad_norm": 4.84375, + "learning_rate": 9.917836151485407e-06, + "loss": 1.67343712, + "memory(GiB)": 71.21, + "step": 8270, + "train_speed(iter/s)": 0.450004 + }, + { + "acc": 0.65173993, + "epoch": 0.20991882293252156, + "grad_norm": 4.71875, + "learning_rate": 9.917646722455713e-06, + "loss": 1.69439087, + "memory(GiB)": 71.21, + "step": 8275, + "train_speed(iter/s)": 0.450145 + }, + { + "acc": 0.6344492, + "epoch": 0.2100456621004566, + "grad_norm": 5.96875, + "learning_rate": 9.91745707712596e-06, + "loss": 1.66014977, + "memory(GiB)": 71.21, + "step": 8280, + "train_speed(iter/s)": 0.450286 + }, + { + "acc": 0.64204292, + "epoch": 0.21017250126839168, + "grad_norm": 4.71875, + "learning_rate": 9.91726721550449e-06, + "loss": 1.67773266, + "memory(GiB)": 71.21, + "step": 8285, + "train_speed(iter/s)": 0.450427 + }, + { + "acc": 0.63515577, + "epoch": 0.21029934043632673, + "grad_norm": 4.34375, + "learning_rate": 9.917077137599653e-06, + "loss": 1.64633846, + "memory(GiB)": 71.21, + "step": 8290, + "train_speed(iter/s)": 0.450568 + }, + { + "acc": 0.63695116, + "epoch": 0.2104261796042618, + "grad_norm": 4.78125, + "learning_rate": 9.916886843419811e-06, + "loss": 1.71501465, + "memory(GiB)": 71.21, + "step": 8295, + "train_speed(iter/s)": 0.450709 + }, + { + "acc": 0.65506315, + "epoch": 0.21055301877219684, + "grad_norm": 5.65625, + "learning_rate": 9.916696332973334e-06, + "loss": 1.70812073, + "memory(GiB)": 71.21, + "step": 8300, + "train_speed(iter/s)": 0.45085 + }, + { + "epoch": 0.21055301877219684, + "eval_acc": 0.6341152127041367, + "eval_loss": 1.6625590324401855, + "eval_runtime": 112.9259, + "eval_samples_per_second": 56.409, + "eval_steps_per_second": 28.204, + "step": 8300 + }, + { + "acc": 0.64055576, + "epoch": 0.21067985794013191, + "grad_norm": 5.28125, + "learning_rate": 9.9165056062686e-06, + "loss": 1.63725185, + "memory(GiB)": 71.21, + "step": 8305, + "train_speed(iter/s)": 0.448135 + }, + { + "acc": 0.63156252, + "epoch": 0.21080669710806696, + "grad_norm": 6.0, + "learning_rate": 9.916314663314e-06, + "loss": 1.71484127, + "memory(GiB)": 71.21, + "step": 8310, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.65101252, + "epoch": 0.21093353627600203, + "grad_norm": 4.6875, + "learning_rate": 9.91612350411793e-06, + "loss": 1.69785118, + "memory(GiB)": 71.21, + "step": 8315, + "train_speed(iter/s)": 0.448421 + }, + { + "acc": 0.65721731, + "epoch": 0.21106037544393708, + "grad_norm": 5.875, + "learning_rate": 9.9159321286888e-06, + "loss": 1.62851791, + "memory(GiB)": 71.21, + "step": 8320, + "train_speed(iter/s)": 0.448564 + }, + { + "acc": 0.63737774, + "epoch": 0.21118721461187215, + "grad_norm": 4.875, + "learning_rate": 9.915740537035026e-06, + "loss": 1.73724747, + "memory(GiB)": 71.21, + "step": 8325, + "train_speed(iter/s)": 0.448707 + }, + { + "acc": 0.63812456, + "epoch": 0.2113140537798072, + "grad_norm": 5.8125, + "learning_rate": 9.915548729165036e-06, + "loss": 1.73263206, + "memory(GiB)": 71.21, + "step": 8330, + "train_speed(iter/s)": 0.448851 + }, + { + "acc": 0.66573229, + "epoch": 0.21144089294774226, + "grad_norm": 5.4375, + "learning_rate": 9.915356705087269e-06, + "loss": 1.60339928, + "memory(GiB)": 71.21, + "step": 8335, + "train_speed(iter/s)": 0.448995 + }, + { + "acc": 0.62934456, + "epoch": 0.2115677321156773, + "grad_norm": 5.53125, + "learning_rate": 9.915164464810166e-06, + "loss": 1.71792564, + "memory(GiB)": 71.21, + "step": 8340, + "train_speed(iter/s)": 0.449138 + }, + { + "acc": 0.64654231, + "epoch": 0.21169457128361238, + "grad_norm": 5.625, + "learning_rate": 9.914972008342186e-06, + "loss": 1.66225758, + "memory(GiB)": 71.21, + "step": 8345, + "train_speed(iter/s)": 0.449281 + }, + { + "acc": 0.62663965, + "epoch": 0.21182141045154743, + "grad_norm": 5.25, + "learning_rate": 9.914779335691793e-06, + "loss": 1.75027809, + "memory(GiB)": 71.21, + "step": 8350, + "train_speed(iter/s)": 0.449424 + }, + { + "acc": 0.62927217, + "epoch": 0.2119482496194825, + "grad_norm": 4.8125, + "learning_rate": 9.914586446867463e-06, + "loss": 1.72834949, + "memory(GiB)": 71.21, + "step": 8355, + "train_speed(iter/s)": 0.449567 + }, + { + "acc": 0.64732742, + "epoch": 0.21207508878741754, + "grad_norm": 5.3125, + "learning_rate": 9.914393341877678e-06, + "loss": 1.68878078, + "memory(GiB)": 71.21, + "step": 8360, + "train_speed(iter/s)": 0.44971 + }, + { + "acc": 0.6347538, + "epoch": 0.21220192795535261, + "grad_norm": 5.53125, + "learning_rate": 9.914200020730932e-06, + "loss": 1.70048676, + "memory(GiB)": 71.21, + "step": 8365, + "train_speed(iter/s)": 0.449853 + }, + { + "acc": 0.64807692, + "epoch": 0.21232876712328766, + "grad_norm": 5.03125, + "learning_rate": 9.914006483435732e-06, + "loss": 1.65754147, + "memory(GiB)": 71.21, + "step": 8370, + "train_speed(iter/s)": 0.449997 + }, + { + "acc": 0.62769446, + "epoch": 0.21245560629122273, + "grad_norm": 5.1875, + "learning_rate": 9.913812730000585e-06, + "loss": 1.72713394, + "memory(GiB)": 71.21, + "step": 8375, + "train_speed(iter/s)": 0.450139 + }, + { + "acc": 0.6506166, + "epoch": 0.21258244545915778, + "grad_norm": 6.34375, + "learning_rate": 9.913618760434015e-06, + "loss": 1.67680492, + "memory(GiB)": 71.21, + "step": 8380, + "train_speed(iter/s)": 0.450284 + }, + { + "acc": 0.64880624, + "epoch": 0.21270928462709285, + "grad_norm": 5.875, + "learning_rate": 9.913424574744555e-06, + "loss": 1.6197855, + "memory(GiB)": 71.21, + "step": 8385, + "train_speed(iter/s)": 0.450427 + }, + { + "acc": 0.65454459, + "epoch": 0.2128361237950279, + "grad_norm": 5.96875, + "learning_rate": 9.913230172940744e-06, + "loss": 1.70528831, + "memory(GiB)": 71.21, + "step": 8390, + "train_speed(iter/s)": 0.45057 + }, + { + "acc": 0.6289444, + "epoch": 0.21296296296296297, + "grad_norm": 5.09375, + "learning_rate": 9.913035555031136e-06, + "loss": 1.69058723, + "memory(GiB)": 71.21, + "step": 8395, + "train_speed(iter/s)": 0.450714 + }, + { + "acc": 0.64075446, + "epoch": 0.213089802130898, + "grad_norm": 5.34375, + "learning_rate": 9.912840721024288e-06, + "loss": 1.61358147, + "memory(GiB)": 71.21, + "step": 8400, + "train_speed(iter/s)": 0.450858 + }, + { + "epoch": 0.213089802130898, + "eval_acc": 0.6343157410126763, + "eval_loss": 1.6613247394561768, + "eval_runtime": 113.829, + "eval_samples_per_second": 55.961, + "eval_steps_per_second": 27.981, + "step": 8400 + }, + { + "acc": 0.66029735, + "epoch": 0.21321664129883308, + "grad_norm": 5.09375, + "learning_rate": 9.91264567092877e-06, + "loss": 1.56634007, + "memory(GiB)": 71.21, + "step": 8405, + "train_speed(iter/s)": 0.448156 + }, + { + "acc": 0.63390112, + "epoch": 0.21334348046676813, + "grad_norm": 6.4375, + "learning_rate": 9.912450404753164e-06, + "loss": 1.75464439, + "memory(GiB)": 71.21, + "step": 8410, + "train_speed(iter/s)": 0.448301 + }, + { + "acc": 0.65367498, + "epoch": 0.2134703196347032, + "grad_norm": 4.65625, + "learning_rate": 9.912254922506057e-06, + "loss": 1.61806946, + "memory(GiB)": 71.21, + "step": 8415, + "train_speed(iter/s)": 0.448445 + }, + { + "acc": 0.64051189, + "epoch": 0.21359715880263824, + "grad_norm": 5.375, + "learning_rate": 9.912059224196044e-06, + "loss": 1.70011864, + "memory(GiB)": 71.21, + "step": 8420, + "train_speed(iter/s)": 0.448588 + }, + { + "acc": 0.6440187, + "epoch": 0.21372399797057332, + "grad_norm": 4.5625, + "learning_rate": 9.911863309831738e-06, + "loss": 1.62829304, + "memory(GiB)": 71.21, + "step": 8425, + "train_speed(iter/s)": 0.44873 + }, + { + "acc": 0.62482624, + "epoch": 0.21385083713850836, + "grad_norm": 5.0, + "learning_rate": 9.911667179421753e-06, + "loss": 1.76024208, + "memory(GiB)": 71.21, + "step": 8430, + "train_speed(iter/s)": 0.448874 + }, + { + "acc": 0.64880376, + "epoch": 0.21397767630644343, + "grad_norm": 4.59375, + "learning_rate": 9.911470832974717e-06, + "loss": 1.62155228, + "memory(GiB)": 71.21, + "step": 8435, + "train_speed(iter/s)": 0.449016 + }, + { + "acc": 0.63791022, + "epoch": 0.21410451547437848, + "grad_norm": 5.0625, + "learning_rate": 9.911274270499265e-06, + "loss": 1.63118515, + "memory(GiB)": 71.21, + "step": 8440, + "train_speed(iter/s)": 0.449158 + }, + { + "acc": 0.64686451, + "epoch": 0.21423135464231355, + "grad_norm": 5.84375, + "learning_rate": 9.911077492004044e-06, + "loss": 1.66607552, + "memory(GiB)": 71.21, + "step": 8445, + "train_speed(iter/s)": 0.4493 + }, + { + "acc": 0.63298426, + "epoch": 0.2143581938102486, + "grad_norm": 5.625, + "learning_rate": 9.910880497497707e-06, + "loss": 1.67723961, + "memory(GiB)": 71.21, + "step": 8450, + "train_speed(iter/s)": 0.449442 + }, + { + "acc": 0.64578667, + "epoch": 0.21448503297818367, + "grad_norm": 5.0625, + "learning_rate": 9.910683286988922e-06, + "loss": 1.66332283, + "memory(GiB)": 71.21, + "step": 8455, + "train_speed(iter/s)": 0.449584 + }, + { + "acc": 0.63888454, + "epoch": 0.2146118721461187, + "grad_norm": 5.34375, + "learning_rate": 9.910485860486361e-06, + "loss": 1.67283554, + "memory(GiB)": 71.21, + "step": 8460, + "train_speed(iter/s)": 0.449727 + }, + { + "acc": 0.63577676, + "epoch": 0.21473871131405378, + "grad_norm": 5.4375, + "learning_rate": 9.910288217998707e-06, + "loss": 1.73443718, + "memory(GiB)": 71.21, + "step": 8465, + "train_speed(iter/s)": 0.44987 + }, + { + "acc": 0.64198065, + "epoch": 0.21486555048198883, + "grad_norm": 5.21875, + "learning_rate": 9.910090359534654e-06, + "loss": 1.71854591, + "memory(GiB)": 71.21, + "step": 8470, + "train_speed(iter/s)": 0.450011 + }, + { + "acc": 0.64926138, + "epoch": 0.2149923896499239, + "grad_norm": 5.65625, + "learning_rate": 9.909892285102907e-06, + "loss": 1.68130035, + "memory(GiB)": 71.21, + "step": 8475, + "train_speed(iter/s)": 0.450155 + }, + { + "acc": 0.65523891, + "epoch": 0.21511922881785894, + "grad_norm": 5.71875, + "learning_rate": 9.909693994712174e-06, + "loss": 1.64424038, + "memory(GiB)": 71.21, + "step": 8480, + "train_speed(iter/s)": 0.450296 + }, + { + "acc": 0.66069756, + "epoch": 0.21524606798579402, + "grad_norm": 5.28125, + "learning_rate": 9.909495488371181e-06, + "loss": 1.57171192, + "memory(GiB)": 71.21, + "step": 8485, + "train_speed(iter/s)": 0.450437 + }, + { + "acc": 0.6437088, + "epoch": 0.21537290715372906, + "grad_norm": 5.25, + "learning_rate": 9.909296766088657e-06, + "loss": 1.66058884, + "memory(GiB)": 71.21, + "step": 8490, + "train_speed(iter/s)": 0.450579 + }, + { + "acc": 0.65637722, + "epoch": 0.21549974632166413, + "grad_norm": 6.0625, + "learning_rate": 9.909097827873341e-06, + "loss": 1.64300213, + "memory(GiB)": 71.21, + "step": 8495, + "train_speed(iter/s)": 0.45072 + }, + { + "acc": 0.64597721, + "epoch": 0.21562658548959918, + "grad_norm": 5.78125, + "learning_rate": 9.908898673733986e-06, + "loss": 1.69559383, + "memory(GiB)": 71.21, + "step": 8500, + "train_speed(iter/s)": 0.450862 + }, + { + "epoch": 0.21562658548959918, + "eval_acc": 0.6345003941634565, + "eval_loss": 1.6606484651565552, + "eval_runtime": 113.5763, + "eval_samples_per_second": 56.086, + "eval_steps_per_second": 28.043, + "step": 8500 + }, + { + "acc": 0.65200815, + "epoch": 0.21575342465753425, + "grad_norm": 5.53125, + "learning_rate": 9.90869930367935e-06, + "loss": 1.65521774, + "memory(GiB)": 71.21, + "step": 8505, + "train_speed(iter/s)": 0.448197 + }, + { + "acc": 0.6272964, + "epoch": 0.2158802638254693, + "grad_norm": 6.25, + "learning_rate": 9.908499717718203e-06, + "loss": 1.75980797, + "memory(GiB)": 71.21, + "step": 8510, + "train_speed(iter/s)": 0.448338 + }, + { + "acc": 0.65032406, + "epoch": 0.21600710299340437, + "grad_norm": 5.46875, + "learning_rate": 9.908299915859325e-06, + "loss": 1.68360138, + "memory(GiB)": 71.21, + "step": 8515, + "train_speed(iter/s)": 0.448478 + }, + { + "acc": 0.65043573, + "epoch": 0.2161339421613394, + "grad_norm": 4.84375, + "learning_rate": 9.908099898111502e-06, + "loss": 1.65408058, + "memory(GiB)": 71.21, + "step": 8520, + "train_speed(iter/s)": 0.44862 + }, + { + "acc": 0.64116797, + "epoch": 0.21626078132927448, + "grad_norm": 5.59375, + "learning_rate": 9.907899664483533e-06, + "loss": 1.64792061, + "memory(GiB)": 71.21, + "step": 8525, + "train_speed(iter/s)": 0.44876 + }, + { + "acc": 0.6256278, + "epoch": 0.21638762049720953, + "grad_norm": 4.78125, + "learning_rate": 9.907699214984223e-06, + "loss": 1.72127037, + "memory(GiB)": 71.21, + "step": 8530, + "train_speed(iter/s)": 0.4489 + }, + { + "acc": 0.6327477, + "epoch": 0.2165144596651446, + "grad_norm": 5.25, + "learning_rate": 9.90749854962239e-06, + "loss": 1.67684021, + "memory(GiB)": 71.21, + "step": 8535, + "train_speed(iter/s)": 0.449039 + }, + { + "acc": 0.6455873, + "epoch": 0.21664129883307964, + "grad_norm": 5.65625, + "learning_rate": 9.907297668406863e-06, + "loss": 1.66346493, + "memory(GiB)": 71.21, + "step": 8540, + "train_speed(iter/s)": 0.44918 + }, + { + "acc": 0.63578806, + "epoch": 0.21676813800101472, + "grad_norm": 4.9375, + "learning_rate": 9.907096571346474e-06, + "loss": 1.67655296, + "memory(GiB)": 71.21, + "step": 8545, + "train_speed(iter/s)": 0.449321 + }, + { + "acc": 0.63263969, + "epoch": 0.21689497716894976, + "grad_norm": 4.65625, + "learning_rate": 9.906895258450067e-06, + "loss": 1.62459526, + "memory(GiB)": 71.21, + "step": 8550, + "train_speed(iter/s)": 0.449462 + }, + { + "acc": 0.65021734, + "epoch": 0.21702181633688483, + "grad_norm": 4.59375, + "learning_rate": 9.9066937297265e-06, + "loss": 1.64175014, + "memory(GiB)": 71.21, + "step": 8555, + "train_speed(iter/s)": 0.449603 + }, + { + "acc": 0.63873334, + "epoch": 0.21714865550481988, + "grad_norm": 5.15625, + "learning_rate": 9.906491985184637e-06, + "loss": 1.62002945, + "memory(GiB)": 71.21, + "step": 8560, + "train_speed(iter/s)": 0.449744 + }, + { + "acc": 0.6204812, + "epoch": 0.21727549467275495, + "grad_norm": 5.96875, + "learning_rate": 9.906290024833349e-06, + "loss": 1.71428185, + "memory(GiB)": 71.21, + "step": 8565, + "train_speed(iter/s)": 0.449884 + }, + { + "acc": 0.64224968, + "epoch": 0.21740233384069, + "grad_norm": 5.59375, + "learning_rate": 9.906087848681523e-06, + "loss": 1.6708149, + "memory(GiB)": 71.21, + "step": 8570, + "train_speed(iter/s)": 0.450026 + }, + { + "acc": 0.64272375, + "epoch": 0.21752917300862507, + "grad_norm": 5.46875, + "learning_rate": 9.905885456738046e-06, + "loss": 1.69182625, + "memory(GiB)": 71.21, + "step": 8575, + "train_speed(iter/s)": 0.450166 + }, + { + "acc": 0.64687839, + "epoch": 0.2176560121765601, + "grad_norm": 5.40625, + "learning_rate": 9.905682849011826e-06, + "loss": 1.68394508, + "memory(GiB)": 71.21, + "step": 8580, + "train_speed(iter/s)": 0.450307 + }, + { + "acc": 0.63993306, + "epoch": 0.21778285134449518, + "grad_norm": 5.9375, + "learning_rate": 9.905480025511772e-06, + "loss": 1.72913322, + "memory(GiB)": 71.21, + "step": 8585, + "train_speed(iter/s)": 0.450447 + }, + { + "acc": 0.63887682, + "epoch": 0.21790969051243023, + "grad_norm": 4.90625, + "learning_rate": 9.905276986246804e-06, + "loss": 1.68368549, + "memory(GiB)": 71.21, + "step": 8590, + "train_speed(iter/s)": 0.450587 + }, + { + "acc": 0.64382825, + "epoch": 0.2180365296803653, + "grad_norm": 4.71875, + "learning_rate": 9.905073731225854e-06, + "loss": 1.62338028, + "memory(GiB)": 71.21, + "step": 8595, + "train_speed(iter/s)": 0.450727 + }, + { + "acc": 0.62571907, + "epoch": 0.21816336884830034, + "grad_norm": 6.0, + "learning_rate": 9.904870260457861e-06, + "loss": 1.67393875, + "memory(GiB)": 71.21, + "step": 8600, + "train_speed(iter/s)": 0.450866 + }, + { + "epoch": 0.21816336884830034, + "eval_acc": 0.6345434241963306, + "eval_loss": 1.6599725484848022, + "eval_runtime": 112.8545, + "eval_samples_per_second": 56.444, + "eval_steps_per_second": 28.222, + "step": 8600 + }, + { + "acc": 0.62467356, + "epoch": 0.21829020801623542, + "grad_norm": 6.09375, + "learning_rate": 9.904666573951777e-06, + "loss": 1.73779869, + "memory(GiB)": 71.21, + "step": 8605, + "train_speed(iter/s)": 0.448248 + }, + { + "acc": 0.63310089, + "epoch": 0.21841704718417046, + "grad_norm": 5.59375, + "learning_rate": 9.904462671716559e-06, + "loss": 1.71177025, + "memory(GiB)": 71.21, + "step": 8610, + "train_speed(iter/s)": 0.448386 + }, + { + "acc": 0.62540512, + "epoch": 0.21854388635210553, + "grad_norm": 4.875, + "learning_rate": 9.904258553761175e-06, + "loss": 1.76584167, + "memory(GiB)": 71.21, + "step": 8615, + "train_speed(iter/s)": 0.448525 + }, + { + "acc": 0.64137201, + "epoch": 0.21867072552004058, + "grad_norm": 4.8125, + "learning_rate": 9.904054220094603e-06, + "loss": 1.70458794, + "memory(GiB)": 71.21, + "step": 8620, + "train_speed(iter/s)": 0.448662 + }, + { + "acc": 0.63665075, + "epoch": 0.21879756468797565, + "grad_norm": 5.375, + "learning_rate": 9.903849670725833e-06, + "loss": 1.69028702, + "memory(GiB)": 71.21, + "step": 8625, + "train_speed(iter/s)": 0.448801 + }, + { + "acc": 0.63703203, + "epoch": 0.2189244038559107, + "grad_norm": 5.46875, + "learning_rate": 9.903644905663861e-06, + "loss": 1.65207024, + "memory(GiB)": 71.21, + "step": 8630, + "train_speed(iter/s)": 0.44894 + }, + { + "acc": 0.64025235, + "epoch": 0.21905124302384577, + "grad_norm": 4.84375, + "learning_rate": 9.90343992491769e-06, + "loss": 1.62149334, + "memory(GiB)": 71.21, + "step": 8635, + "train_speed(iter/s)": 0.449077 + }, + { + "acc": 0.64674225, + "epoch": 0.2191780821917808, + "grad_norm": 5.03125, + "learning_rate": 9.903234728496341e-06, + "loss": 1.6584301, + "memory(GiB)": 71.21, + "step": 8640, + "train_speed(iter/s)": 0.449216 + }, + { + "acc": 0.63305817, + "epoch": 0.21930492135971588, + "grad_norm": 5.8125, + "learning_rate": 9.903029316408838e-06, + "loss": 1.74811535, + "memory(GiB)": 71.21, + "step": 8645, + "train_speed(iter/s)": 0.449355 + }, + { + "acc": 0.65412693, + "epoch": 0.21943176052765093, + "grad_norm": 4.96875, + "learning_rate": 9.902823688664214e-06, + "loss": 1.63295174, + "memory(GiB)": 71.21, + "step": 8650, + "train_speed(iter/s)": 0.449494 + }, + { + "acc": 0.64470105, + "epoch": 0.219558599695586, + "grad_norm": 5.03125, + "learning_rate": 9.902617845271514e-06, + "loss": 1.62208443, + "memory(GiB)": 71.21, + "step": 8655, + "train_speed(iter/s)": 0.449631 + }, + { + "acc": 0.65957475, + "epoch": 0.21968543886352104, + "grad_norm": 4.875, + "learning_rate": 9.902411786239794e-06, + "loss": 1.61606712, + "memory(GiB)": 71.21, + "step": 8660, + "train_speed(iter/s)": 0.44977 + }, + { + "acc": 0.63803854, + "epoch": 0.21981227803145612, + "grad_norm": 4.75, + "learning_rate": 9.902205511578114e-06, + "loss": 1.66391106, + "memory(GiB)": 71.21, + "step": 8665, + "train_speed(iter/s)": 0.449909 + }, + { + "acc": 0.63988657, + "epoch": 0.21993911719939116, + "grad_norm": 6.1875, + "learning_rate": 9.90199902129555e-06, + "loss": 1.68073959, + "memory(GiB)": 71.21, + "step": 8670, + "train_speed(iter/s)": 0.450048 + }, + { + "acc": 0.62764416, + "epoch": 0.22006595636732623, + "grad_norm": 5.40625, + "learning_rate": 9.901792315401184e-06, + "loss": 1.72427216, + "memory(GiB)": 71.21, + "step": 8675, + "train_speed(iter/s)": 0.450187 + }, + { + "acc": 0.64597545, + "epoch": 0.22019279553526128, + "grad_norm": 4.78125, + "learning_rate": 9.901585393904104e-06, + "loss": 1.60278053, + "memory(GiB)": 71.21, + "step": 8680, + "train_speed(iter/s)": 0.450324 + }, + { + "acc": 0.63844781, + "epoch": 0.22031963470319635, + "grad_norm": 4.78125, + "learning_rate": 9.901378256813418e-06, + "loss": 1.74130955, + "memory(GiB)": 71.21, + "step": 8685, + "train_speed(iter/s)": 0.450462 + }, + { + "acc": 0.64623966, + "epoch": 0.2204464738711314, + "grad_norm": 5.5, + "learning_rate": 9.901170904138232e-06, + "loss": 1.64620361, + "memory(GiB)": 71.21, + "step": 8690, + "train_speed(iter/s)": 0.450599 + }, + { + "acc": 0.62689481, + "epoch": 0.22057331303906647, + "grad_norm": 4.78125, + "learning_rate": 9.900963335887667e-06, + "loss": 1.77168198, + "memory(GiB)": 71.21, + "step": 8695, + "train_speed(iter/s)": 0.450736 + }, + { + "acc": 0.64292135, + "epoch": 0.2207001522070015, + "grad_norm": 5.53125, + "learning_rate": 9.900755552070852e-06, + "loss": 1.69609146, + "memory(GiB)": 71.21, + "step": 8700, + "train_speed(iter/s)": 0.450873 + }, + { + "epoch": 0.2207001522070015, + "eval_acc": 0.6346411817467436, + "eval_loss": 1.65890371799469, + "eval_runtime": 113.7787, + "eval_samples_per_second": 55.986, + "eval_steps_per_second": 27.993, + "step": 8700 + }, + { + "acc": 0.64919968, + "epoch": 0.22082699137493658, + "grad_norm": 5.34375, + "learning_rate": 9.900547552696931e-06, + "loss": 1.66205635, + "memory(GiB)": 71.21, + "step": 8705, + "train_speed(iter/s)": 0.448262 + }, + { + "acc": 0.65390587, + "epoch": 0.22095383054287163, + "grad_norm": 4.6875, + "learning_rate": 9.900339337775046e-06, + "loss": 1.71874924, + "memory(GiB)": 71.21, + "step": 8710, + "train_speed(iter/s)": 0.448399 + }, + { + "acc": 0.64157581, + "epoch": 0.2210806697108067, + "grad_norm": 5.84375, + "learning_rate": 9.90013090731436e-06, + "loss": 1.71507683, + "memory(GiB)": 71.21, + "step": 8715, + "train_speed(iter/s)": 0.448537 + }, + { + "acc": 0.62691565, + "epoch": 0.22120750887874174, + "grad_norm": 5.46875, + "learning_rate": 9.89992226132404e-06, + "loss": 1.73959789, + "memory(GiB)": 71.21, + "step": 8720, + "train_speed(iter/s)": 0.448673 + }, + { + "acc": 0.64337192, + "epoch": 0.22133434804667682, + "grad_norm": 4.5625, + "learning_rate": 9.899713399813261e-06, + "loss": 1.63869781, + "memory(GiB)": 71.21, + "step": 8725, + "train_speed(iter/s)": 0.44881 + }, + { + "acc": 0.64540224, + "epoch": 0.22146118721461186, + "grad_norm": 5.375, + "learning_rate": 9.899504322791212e-06, + "loss": 1.6667572, + "memory(GiB)": 71.21, + "step": 8730, + "train_speed(iter/s)": 0.448947 + }, + { + "acc": 0.64097061, + "epoch": 0.22158802638254693, + "grad_norm": 4.96875, + "learning_rate": 9.899295030267086e-06, + "loss": 1.72067375, + "memory(GiB)": 71.21, + "step": 8735, + "train_speed(iter/s)": 0.449083 + }, + { + "acc": 0.63403368, + "epoch": 0.22171486555048198, + "grad_norm": 4.875, + "learning_rate": 9.899085522250094e-06, + "loss": 1.6453619, + "memory(GiB)": 71.21, + "step": 8740, + "train_speed(iter/s)": 0.44922 + }, + { + "acc": 0.65026455, + "epoch": 0.22184170471841705, + "grad_norm": 5.03125, + "learning_rate": 9.898875798749446e-06, + "loss": 1.65235004, + "memory(GiB)": 71.21, + "step": 8745, + "train_speed(iter/s)": 0.449357 + }, + { + "acc": 0.64297285, + "epoch": 0.2219685438863521, + "grad_norm": 5.46875, + "learning_rate": 9.898665859774367e-06, + "loss": 1.70155563, + "memory(GiB)": 71.21, + "step": 8750, + "train_speed(iter/s)": 0.449493 + }, + { + "acc": 0.64260025, + "epoch": 0.22209538305428717, + "grad_norm": 6.0625, + "learning_rate": 9.898455705334095e-06, + "loss": 1.70446529, + "memory(GiB)": 71.21, + "step": 8755, + "train_speed(iter/s)": 0.44963 + }, + { + "acc": 0.63793778, + "epoch": 0.2222222222222222, + "grad_norm": 4.90625, + "learning_rate": 9.89824533543787e-06, + "loss": 1.69004879, + "memory(GiB)": 71.21, + "step": 8760, + "train_speed(iter/s)": 0.449764 + }, + { + "acc": 0.62443075, + "epoch": 0.22234906139015728, + "grad_norm": 5.03125, + "learning_rate": 9.898034750094946e-06, + "loss": 1.70333672, + "memory(GiB)": 71.21, + "step": 8765, + "train_speed(iter/s)": 0.4499 + }, + { + "acc": 0.64417849, + "epoch": 0.22247590055809233, + "grad_norm": 6.0, + "learning_rate": 9.897823949314586e-06, + "loss": 1.73804665, + "memory(GiB)": 71.21, + "step": 8770, + "train_speed(iter/s)": 0.450038 + }, + { + "acc": 0.6413599, + "epoch": 0.2226027397260274, + "grad_norm": 5.78125, + "learning_rate": 9.897612933106061e-06, + "loss": 1.73241959, + "memory(GiB)": 71.21, + "step": 8775, + "train_speed(iter/s)": 0.450174 + }, + { + "acc": 0.6458806, + "epoch": 0.22272957889396244, + "grad_norm": 5.0, + "learning_rate": 9.897401701478654e-06, + "loss": 1.66043892, + "memory(GiB)": 71.21, + "step": 8780, + "train_speed(iter/s)": 0.450311 + }, + { + "acc": 0.63821559, + "epoch": 0.22285641806189752, + "grad_norm": 6.09375, + "learning_rate": 9.897190254441653e-06, + "loss": 1.71675835, + "memory(GiB)": 71.21, + "step": 8785, + "train_speed(iter/s)": 0.450446 + }, + { + "acc": 0.6251699, + "epoch": 0.22298325722983256, + "grad_norm": 4.96875, + "learning_rate": 9.896978592004363e-06, + "loss": 1.69117851, + "memory(GiB)": 71.21, + "step": 8790, + "train_speed(iter/s)": 0.450583 + }, + { + "acc": 0.63237815, + "epoch": 0.22311009639776763, + "grad_norm": 5.84375, + "learning_rate": 9.896766714176089e-06, + "loss": 1.71420631, + "memory(GiB)": 71.21, + "step": 8795, + "train_speed(iter/s)": 0.45072 + }, + { + "acc": 0.6399663, + "epoch": 0.22323693556570268, + "grad_norm": 6.0, + "learning_rate": 9.896554620966152e-06, + "loss": 1.71319427, + "memory(GiB)": 71.21, + "step": 8800, + "train_speed(iter/s)": 0.450858 + }, + { + "epoch": 0.22323693556570268, + "eval_acc": 0.634847976564925, + "eval_loss": 1.6580780744552612, + "eval_runtime": 113.4003, + "eval_samples_per_second": 56.173, + "eval_steps_per_second": 28.086, + "step": 8800 + }, + { + "acc": 0.65060911, + "epoch": 0.22336377473363775, + "grad_norm": 5.96875, + "learning_rate": 9.896342312383883e-06, + "loss": 1.69036503, + "memory(GiB)": 71.21, + "step": 8805, + "train_speed(iter/s)": 0.448286 + }, + { + "acc": 0.64468422, + "epoch": 0.2234906139015728, + "grad_norm": 6.40625, + "learning_rate": 9.896129788438617e-06, + "loss": 1.67757721, + "memory(GiB)": 71.21, + "step": 8810, + "train_speed(iter/s)": 0.44842 + }, + { + "acc": 0.65232911, + "epoch": 0.22361745306950787, + "grad_norm": 4.8125, + "learning_rate": 9.895917049139704e-06, + "loss": 1.66300545, + "memory(GiB)": 71.21, + "step": 8815, + "train_speed(iter/s)": 0.448553 + }, + { + "acc": 0.63518257, + "epoch": 0.2237442922374429, + "grad_norm": 6.59375, + "learning_rate": 9.895704094496502e-06, + "loss": 1.72634087, + "memory(GiB)": 71.21, + "step": 8820, + "train_speed(iter/s)": 0.448688 + }, + { + "acc": 0.66409698, + "epoch": 0.22387113140537798, + "grad_norm": 4.65625, + "learning_rate": 9.895490924518372e-06, + "loss": 1.61879234, + "memory(GiB)": 71.21, + "step": 8825, + "train_speed(iter/s)": 0.448823 + }, + { + "acc": 0.6458086, + "epoch": 0.22399797057331303, + "grad_norm": 4.75, + "learning_rate": 9.895277539214698e-06, + "loss": 1.57342072, + "memory(GiB)": 71.21, + "step": 8830, + "train_speed(iter/s)": 0.448956 + }, + { + "acc": 0.63458805, + "epoch": 0.2241248097412481, + "grad_norm": 5.125, + "learning_rate": 9.895063938594859e-06, + "loss": 1.71501255, + "memory(GiB)": 71.21, + "step": 8835, + "train_speed(iter/s)": 0.449091 + }, + { + "acc": 0.63261671, + "epoch": 0.22425164890918314, + "grad_norm": 4.25, + "learning_rate": 9.894850122668256e-06, + "loss": 1.71544609, + "memory(GiB)": 71.21, + "step": 8840, + "train_speed(iter/s)": 0.449225 + }, + { + "acc": 0.64980526, + "epoch": 0.22437848807711822, + "grad_norm": 4.4375, + "learning_rate": 9.89463609144429e-06, + "loss": 1.63574715, + "memory(GiB)": 71.21, + "step": 8845, + "train_speed(iter/s)": 0.449359 + }, + { + "acc": 0.64912205, + "epoch": 0.22450532724505326, + "grad_norm": 5.5, + "learning_rate": 9.894421844932375e-06, + "loss": 1.67726002, + "memory(GiB)": 71.21, + "step": 8850, + "train_speed(iter/s)": 0.449493 + }, + { + "acc": 0.66804757, + "epoch": 0.22463216641298833, + "grad_norm": 5.5, + "learning_rate": 9.894207383141937e-06, + "loss": 1.60480156, + "memory(GiB)": 71.21, + "step": 8855, + "train_speed(iter/s)": 0.449628 + }, + { + "acc": 0.64424849, + "epoch": 0.22475900558092338, + "grad_norm": 5.96875, + "learning_rate": 9.893992706082405e-06, + "loss": 1.62976532, + "memory(GiB)": 71.21, + "step": 8860, + "train_speed(iter/s)": 0.449763 + }, + { + "acc": 0.63582263, + "epoch": 0.22488584474885845, + "grad_norm": 4.9375, + "learning_rate": 9.893777813763223e-06, + "loss": 1.69842148, + "memory(GiB)": 71.21, + "step": 8865, + "train_speed(iter/s)": 0.449898 + }, + { + "acc": 0.63942838, + "epoch": 0.2250126839167935, + "grad_norm": 7.21875, + "learning_rate": 9.893562706193847e-06, + "loss": 1.74765472, + "memory(GiB)": 71.21, + "step": 8870, + "train_speed(iter/s)": 0.450033 + }, + { + "acc": 0.64789085, + "epoch": 0.22513952308472857, + "grad_norm": 4.8125, + "learning_rate": 9.893347383383732e-06, + "loss": 1.670755, + "memory(GiB)": 71.21, + "step": 8875, + "train_speed(iter/s)": 0.450168 + }, + { + "acc": 0.66058598, + "epoch": 0.2252663622526636, + "grad_norm": 5.34375, + "learning_rate": 9.893131845342352e-06, + "loss": 1.66837845, + "memory(GiB)": 71.21, + "step": 8880, + "train_speed(iter/s)": 0.450304 + }, + { + "acc": 0.64740686, + "epoch": 0.22539320142059868, + "grad_norm": 5.25, + "learning_rate": 9.892916092079188e-06, + "loss": 1.69717522, + "memory(GiB)": 71.21, + "step": 8885, + "train_speed(iter/s)": 0.450438 + }, + { + "acc": 0.64078178, + "epoch": 0.22552004058853373, + "grad_norm": 4.78125, + "learning_rate": 9.89270012360373e-06, + "loss": 1.71598701, + "memory(GiB)": 71.21, + "step": 8890, + "train_speed(iter/s)": 0.450572 + }, + { + "acc": 0.65165362, + "epoch": 0.2256468797564688, + "grad_norm": 5.25, + "learning_rate": 9.892483939925476e-06, + "loss": 1.6560318, + "memory(GiB)": 71.21, + "step": 8895, + "train_speed(iter/s)": 0.450706 + }, + { + "acc": 0.63041105, + "epoch": 0.22577371892440384, + "grad_norm": 6.125, + "learning_rate": 9.892267541053933e-06, + "loss": 1.69148598, + "memory(GiB)": 71.21, + "step": 8900, + "train_speed(iter/s)": 0.45084 + }, + { + "epoch": 0.22577371892440384, + "eval_acc": 0.6348859933900857, + "eval_loss": 1.657177209854126, + "eval_runtime": 112.9822, + "eval_samples_per_second": 56.381, + "eval_steps_per_second": 28.19, + "step": 8900 + }, + { + "acc": 0.64178405, + "epoch": 0.22590055809233892, + "grad_norm": 5.75, + "learning_rate": 9.892050926998624e-06, + "loss": 1.64071465, + "memory(GiB)": 71.21, + "step": 8905, + "train_speed(iter/s)": 0.448308 + }, + { + "acc": 0.63683686, + "epoch": 0.22602739726027396, + "grad_norm": 4.90625, + "learning_rate": 9.891834097769071e-06, + "loss": 1.73163795, + "memory(GiB)": 71.21, + "step": 8910, + "train_speed(iter/s)": 0.448442 + }, + { + "acc": 0.62672396, + "epoch": 0.22615423642820903, + "grad_norm": 4.6875, + "learning_rate": 9.891617053374816e-06, + "loss": 1.75903835, + "memory(GiB)": 71.21, + "step": 8915, + "train_speed(iter/s)": 0.448574 + }, + { + "acc": 0.6349349, + "epoch": 0.22628107559614408, + "grad_norm": 5.375, + "learning_rate": 9.891399793825403e-06, + "loss": 1.68106956, + "memory(GiB)": 71.21, + "step": 8920, + "train_speed(iter/s)": 0.448709 + }, + { + "acc": 0.64478931, + "epoch": 0.22640791476407915, + "grad_norm": 5.09375, + "learning_rate": 9.891182319130387e-06, + "loss": 1.74965858, + "memory(GiB)": 71.21, + "step": 8925, + "train_speed(iter/s)": 0.448844 + }, + { + "acc": 0.64513016, + "epoch": 0.2265347539320142, + "grad_norm": 5.40625, + "learning_rate": 9.890964629299336e-06, + "loss": 1.72498131, + "memory(GiB)": 71.21, + "step": 8930, + "train_speed(iter/s)": 0.448978 + }, + { + "acc": 0.63588705, + "epoch": 0.22666159309994927, + "grad_norm": 4.90625, + "learning_rate": 9.890746724341825e-06, + "loss": 1.72935143, + "memory(GiB)": 71.21, + "step": 8935, + "train_speed(iter/s)": 0.44911 + }, + { + "acc": 0.64235659, + "epoch": 0.2267884322678843, + "grad_norm": 5.40625, + "learning_rate": 9.890528604267436e-06, + "loss": 1.61256599, + "memory(GiB)": 71.21, + "step": 8940, + "train_speed(iter/s)": 0.449242 + }, + { + "acc": 0.65340319, + "epoch": 0.22691527143581938, + "grad_norm": 5.15625, + "learning_rate": 9.890310269085765e-06, + "loss": 1.64577961, + "memory(GiB)": 71.21, + "step": 8945, + "train_speed(iter/s)": 0.449376 + }, + { + "acc": 0.62669792, + "epoch": 0.22704211060375443, + "grad_norm": 5.25, + "learning_rate": 9.890091718806414e-06, + "loss": 1.73854828, + "memory(GiB)": 71.21, + "step": 8950, + "train_speed(iter/s)": 0.449508 + }, + { + "acc": 0.6441967, + "epoch": 0.2271689497716895, + "grad_norm": 6.0, + "learning_rate": 9.889872953438996e-06, + "loss": 1.69498367, + "memory(GiB)": 71.21, + "step": 8955, + "train_speed(iter/s)": 0.44964 + }, + { + "acc": 0.63234682, + "epoch": 0.22729578893962454, + "grad_norm": 6.375, + "learning_rate": 9.889653972993136e-06, + "loss": 1.69548874, + "memory(GiB)": 71.21, + "step": 8960, + "train_speed(iter/s)": 0.449773 + }, + { + "acc": 0.64013309, + "epoch": 0.22742262810755962, + "grad_norm": 5.125, + "learning_rate": 9.889434777478464e-06, + "loss": 1.68811188, + "memory(GiB)": 71.21, + "step": 8965, + "train_speed(iter/s)": 0.449907 + }, + { + "acc": 0.64788342, + "epoch": 0.22754946727549466, + "grad_norm": 5.125, + "learning_rate": 9.88921536690462e-06, + "loss": 1.67843094, + "memory(GiB)": 71.21, + "step": 8970, + "train_speed(iter/s)": 0.450039 + }, + { + "acc": 0.62794642, + "epoch": 0.22767630644342973, + "grad_norm": 4.84375, + "learning_rate": 9.888995741281252e-06, + "loss": 1.76861305, + "memory(GiB)": 71.21, + "step": 8975, + "train_speed(iter/s)": 0.450172 + }, + { + "acc": 0.64210501, + "epoch": 0.22780314561136478, + "grad_norm": 5.5, + "learning_rate": 9.888775900618028e-06, + "loss": 1.63038273, + "memory(GiB)": 71.21, + "step": 8980, + "train_speed(iter/s)": 0.450305 + }, + { + "acc": 0.6400279, + "epoch": 0.22792998477929985, + "grad_norm": 5.75, + "learning_rate": 9.88855584492461e-06, + "loss": 1.6897522, + "memory(GiB)": 71.21, + "step": 8985, + "train_speed(iter/s)": 0.450437 + }, + { + "acc": 0.63977609, + "epoch": 0.2280568239472349, + "grad_norm": 5.5625, + "learning_rate": 9.888335574210681e-06, + "loss": 1.69920502, + "memory(GiB)": 71.21, + "step": 8990, + "train_speed(iter/s)": 0.450568 + }, + { + "acc": 0.6471415, + "epoch": 0.22818366311516997, + "grad_norm": 5.53125, + "learning_rate": 9.888115088485931e-06, + "loss": 1.57227955, + "memory(GiB)": 71.21, + "step": 8995, + "train_speed(iter/s)": 0.450699 + }, + { + "acc": 0.64659061, + "epoch": 0.228310502283105, + "grad_norm": 5.03125, + "learning_rate": 9.887894387760053e-06, + "loss": 1.67709599, + "memory(GiB)": 71.21, + "step": 9000, + "train_speed(iter/s)": 0.450831 + }, + { + "epoch": 0.228310502283105, + "eval_acc": 0.6350518470119402, + "eval_loss": 1.656386137008667, + "eval_runtime": 114.2974, + "eval_samples_per_second": 55.732, + "eval_steps_per_second": 27.866, + "step": 9000 + }, + { + "acc": 0.64150748, + "epoch": 0.22843734145104008, + "grad_norm": 6.25, + "learning_rate": 9.887673472042757e-06, + "loss": 1.6895256, + "memory(GiB)": 71.21, + "step": 9005, + "train_speed(iter/s)": 0.448296 + }, + { + "acc": 0.64918518, + "epoch": 0.22856418061897513, + "grad_norm": 5.46875, + "learning_rate": 9.88745234134376e-06, + "loss": 1.68153152, + "memory(GiB)": 71.21, + "step": 9010, + "train_speed(iter/s)": 0.448428 + }, + { + "acc": 0.63698549, + "epoch": 0.2286910197869102, + "grad_norm": 4.96875, + "learning_rate": 9.887230995672789e-06, + "loss": 1.68130035, + "memory(GiB)": 71.21, + "step": 9015, + "train_speed(iter/s)": 0.448558 + }, + { + "acc": 0.64020567, + "epoch": 0.22881785895484524, + "grad_norm": 5.90625, + "learning_rate": 9.887009435039578e-06, + "loss": 1.61104927, + "memory(GiB)": 71.21, + "step": 9020, + "train_speed(iter/s)": 0.448689 + }, + { + "acc": 0.63508744, + "epoch": 0.22894469812278032, + "grad_norm": 5.78125, + "learning_rate": 9.886787659453873e-06, + "loss": 1.66029129, + "memory(GiB)": 71.21, + "step": 9025, + "train_speed(iter/s)": 0.44882 + }, + { + "acc": 0.63999081, + "epoch": 0.22907153729071536, + "grad_norm": 5.15625, + "learning_rate": 9.886565668925429e-06, + "loss": 1.6706646, + "memory(GiB)": 71.21, + "step": 9030, + "train_speed(iter/s)": 0.448949 + }, + { + "acc": 0.64018488, + "epoch": 0.22919837645865043, + "grad_norm": 6.375, + "learning_rate": 9.88634346346401e-06, + "loss": 1.70037613, + "memory(GiB)": 71.21, + "step": 9035, + "train_speed(iter/s)": 0.44908 + }, + { + "acc": 0.63816729, + "epoch": 0.22932521562658548, + "grad_norm": 5.03125, + "learning_rate": 9.88612104307939e-06, + "loss": 1.64146347, + "memory(GiB)": 71.21, + "step": 9040, + "train_speed(iter/s)": 0.449211 + }, + { + "acc": 0.6534162, + "epoch": 0.22945205479452055, + "grad_norm": 4.90625, + "learning_rate": 9.885898407781352e-06, + "loss": 1.62353325, + "memory(GiB)": 71.21, + "step": 9045, + "train_speed(iter/s)": 0.449342 + }, + { + "acc": 0.64851255, + "epoch": 0.2295788939624556, + "grad_norm": 5.4375, + "learning_rate": 9.885675557579686e-06, + "loss": 1.75118752, + "memory(GiB)": 71.21, + "step": 9050, + "train_speed(iter/s)": 0.449473 + }, + { + "acc": 0.64811354, + "epoch": 0.22970573313039067, + "grad_norm": 5.6875, + "learning_rate": 9.885452492484198e-06, + "loss": 1.57105618, + "memory(GiB)": 71.21, + "step": 9055, + "train_speed(iter/s)": 0.449604 + }, + { + "acc": 0.64184046, + "epoch": 0.2298325722983257, + "grad_norm": 5.5625, + "learning_rate": 9.885229212504697e-06, + "loss": 1.67570992, + "memory(GiB)": 71.21, + "step": 9060, + "train_speed(iter/s)": 0.449735 + }, + { + "acc": 0.63943548, + "epoch": 0.22995941146626078, + "grad_norm": 4.59375, + "learning_rate": 9.885005717651002e-06, + "loss": 1.63892593, + "memory(GiB)": 71.21, + "step": 9065, + "train_speed(iter/s)": 0.449864 + }, + { + "acc": 0.64715137, + "epoch": 0.23008625063419583, + "grad_norm": 5.90625, + "learning_rate": 9.88478200793295e-06, + "loss": 1.6347887, + "memory(GiB)": 71.21, + "step": 9070, + "train_speed(iter/s)": 0.449994 + }, + { + "acc": 0.64410233, + "epoch": 0.2302130898021309, + "grad_norm": 5.25, + "learning_rate": 9.884558083360372e-06, + "loss": 1.63467331, + "memory(GiB)": 71.21, + "step": 9075, + "train_speed(iter/s)": 0.450125 + }, + { + "acc": 0.65830746, + "epoch": 0.23033992897006594, + "grad_norm": 4.84375, + "learning_rate": 9.884333943943123e-06, + "loss": 1.56917982, + "memory(GiB)": 71.21, + "step": 9080, + "train_speed(iter/s)": 0.450254 + }, + { + "acc": 0.64271107, + "epoch": 0.23046676813800102, + "grad_norm": 4.96875, + "learning_rate": 9.884109589691062e-06, + "loss": 1.69438782, + "memory(GiB)": 71.21, + "step": 9085, + "train_speed(iter/s)": 0.450386 + }, + { + "acc": 0.6374619, + "epoch": 0.23059360730593606, + "grad_norm": 5.21875, + "learning_rate": 9.883885020614052e-06, + "loss": 1.71255112, + "memory(GiB)": 71.21, + "step": 9090, + "train_speed(iter/s)": 0.450518 + }, + { + "acc": 0.66035328, + "epoch": 0.23072044647387113, + "grad_norm": 4.71875, + "learning_rate": 9.883660236721977e-06, + "loss": 1.57349663, + "memory(GiB)": 71.21, + "step": 9095, + "train_speed(iter/s)": 0.450648 + }, + { + "acc": 0.63642597, + "epoch": 0.23084728564180618, + "grad_norm": 5.0625, + "learning_rate": 9.883435238024718e-06, + "loss": 1.67853546, + "memory(GiB)": 71.21, + "step": 9100, + "train_speed(iter/s)": 0.450778 + }, + { + "epoch": 0.23084728564180618, + "eval_acc": 0.6351416669834735, + "eval_loss": 1.6557594537734985, + "eval_runtime": 113.1007, + "eval_samples_per_second": 56.321, + "eval_steps_per_second": 28.161, + "step": 9100 + }, + { + "acc": 0.65952053, + "epoch": 0.23097412480974125, + "grad_norm": 4.90625, + "learning_rate": 9.883210024532176e-06, + "loss": 1.61474571, + "memory(GiB)": 71.21, + "step": 9105, + "train_speed(iter/s)": 0.448299 + }, + { + "acc": 0.65104585, + "epoch": 0.2311009639776763, + "grad_norm": 5.6875, + "learning_rate": 9.882984596254255e-06, + "loss": 1.67403049, + "memory(GiB)": 71.21, + "step": 9110, + "train_speed(iter/s)": 0.44843 + }, + { + "acc": 0.64161625, + "epoch": 0.23122780314561137, + "grad_norm": 5.40625, + "learning_rate": 9.88275895320087e-06, + "loss": 1.70848083, + "memory(GiB)": 71.21, + "step": 9115, + "train_speed(iter/s)": 0.448561 + }, + { + "acc": 0.6435461, + "epoch": 0.2313546423135464, + "grad_norm": 5.34375, + "learning_rate": 9.882533095381947e-06, + "loss": 1.64648685, + "memory(GiB)": 71.21, + "step": 9120, + "train_speed(iter/s)": 0.448692 + }, + { + "acc": 0.62880745, + "epoch": 0.23148148148148148, + "grad_norm": 7.0, + "learning_rate": 9.882307022807419e-06, + "loss": 1.66570415, + "memory(GiB)": 71.21, + "step": 9125, + "train_speed(iter/s)": 0.448822 + }, + { + "acc": 0.63962955, + "epoch": 0.23160832064941653, + "grad_norm": 4.71875, + "learning_rate": 9.88208073548723e-06, + "loss": 1.73283863, + "memory(GiB)": 71.21, + "step": 9130, + "train_speed(iter/s)": 0.44895 + }, + { + "acc": 0.64338145, + "epoch": 0.2317351598173516, + "grad_norm": 4.625, + "learning_rate": 9.881854233431333e-06, + "loss": 1.68782501, + "memory(GiB)": 71.21, + "step": 9135, + "train_speed(iter/s)": 0.449078 + }, + { + "acc": 0.64351912, + "epoch": 0.23186199898528664, + "grad_norm": 4.8125, + "learning_rate": 9.881627516649692e-06, + "loss": 1.66155205, + "memory(GiB)": 71.21, + "step": 9140, + "train_speed(iter/s)": 0.449207 + }, + { + "acc": 0.62770853, + "epoch": 0.23198883815322172, + "grad_norm": 4.78125, + "learning_rate": 9.881400585152278e-06, + "loss": 1.75321198, + "memory(GiB)": 71.21, + "step": 9145, + "train_speed(iter/s)": 0.449337 + }, + { + "acc": 0.65417624, + "epoch": 0.23211567732115676, + "grad_norm": 4.5625, + "learning_rate": 9.881173438949072e-06, + "loss": 1.67187366, + "memory(GiB)": 71.21, + "step": 9150, + "train_speed(iter/s)": 0.449464 + }, + { + "acc": 0.65229292, + "epoch": 0.23224251648909183, + "grad_norm": 4.75, + "learning_rate": 9.880946078050064e-06, + "loss": 1.63622475, + "memory(GiB)": 71.21, + "step": 9155, + "train_speed(iter/s)": 0.449593 + }, + { + "acc": 0.62765646, + "epoch": 0.23236935565702688, + "grad_norm": 5.09375, + "learning_rate": 9.880718502465258e-06, + "loss": 1.74452744, + "memory(GiB)": 71.21, + "step": 9160, + "train_speed(iter/s)": 0.449722 + }, + { + "acc": 0.62743311, + "epoch": 0.23249619482496195, + "grad_norm": 4.8125, + "learning_rate": 9.88049071220466e-06, + "loss": 1.72903748, + "memory(GiB)": 71.21, + "step": 9165, + "train_speed(iter/s)": 0.449851 + }, + { + "acc": 0.64528837, + "epoch": 0.232623033992897, + "grad_norm": 5.0625, + "learning_rate": 9.88026270727829e-06, + "loss": 1.65912437, + "memory(GiB)": 71.21, + "step": 9170, + "train_speed(iter/s)": 0.449979 + }, + { + "acc": 0.6376646, + "epoch": 0.23274987316083207, + "grad_norm": 4.9375, + "learning_rate": 9.880034487696179e-06, + "loss": 1.65191498, + "memory(GiB)": 71.21, + "step": 9175, + "train_speed(iter/s)": 0.450107 + }, + { + "acc": 0.63751621, + "epoch": 0.2328767123287671, + "grad_norm": 7.34375, + "learning_rate": 9.879806053468361e-06, + "loss": 1.6735714, + "memory(GiB)": 71.21, + "step": 9180, + "train_speed(iter/s)": 0.450236 + }, + { + "acc": 0.63770809, + "epoch": 0.23300355149670218, + "grad_norm": 4.53125, + "learning_rate": 9.879577404604889e-06, + "loss": 1.66351414, + "memory(GiB)": 71.21, + "step": 9185, + "train_speed(iter/s)": 0.450365 + }, + { + "acc": 0.6309947, + "epoch": 0.23313039066463723, + "grad_norm": 5.09375, + "learning_rate": 9.879348541115816e-06, + "loss": 1.67180634, + "memory(GiB)": 71.21, + "step": 9190, + "train_speed(iter/s)": 0.450492 + }, + { + "acc": 0.63793411, + "epoch": 0.2332572298325723, + "grad_norm": 4.375, + "learning_rate": 9.879119463011208e-06, + "loss": 1.6578392, + "memory(GiB)": 71.21, + "step": 9195, + "train_speed(iter/s)": 0.450621 + }, + { + "acc": 0.65422168, + "epoch": 0.23338406900050734, + "grad_norm": 5.125, + "learning_rate": 9.878890170301143e-06, + "loss": 1.6297781, + "memory(GiB)": 71.21, + "step": 9200, + "train_speed(iter/s)": 0.45075 + }, + { + "epoch": 0.23338406900050734, + "eval_acc": 0.6352903921456403, + "eval_loss": 1.654702067375183, + "eval_runtime": 113.2272, + "eval_samples_per_second": 56.259, + "eval_steps_per_second": 28.129, + "step": 9200 + }, + { + "acc": 0.65784197, + "epoch": 0.23351090816844242, + "grad_norm": 4.4375, + "learning_rate": 9.878660662995706e-06, + "loss": 1.60854683, + "memory(GiB)": 71.21, + "step": 9205, + "train_speed(iter/s)": 0.448293 + }, + { + "acc": 0.64525433, + "epoch": 0.23363774733637746, + "grad_norm": 6.4375, + "learning_rate": 9.878430941104991e-06, + "loss": 1.69421883, + "memory(GiB)": 71.21, + "step": 9210, + "train_speed(iter/s)": 0.44842 + }, + { + "acc": 0.641152, + "epoch": 0.23376458650431253, + "grad_norm": 5.0625, + "learning_rate": 9.878201004639104e-06, + "loss": 1.72842541, + "memory(GiB)": 71.21, + "step": 9215, + "train_speed(iter/s)": 0.448548 + }, + { + "acc": 0.6653368, + "epoch": 0.23389142567224758, + "grad_norm": 5.34375, + "learning_rate": 9.877970853608156e-06, + "loss": 1.61517982, + "memory(GiB)": 71.21, + "step": 9220, + "train_speed(iter/s)": 0.448675 + }, + { + "acc": 0.64513693, + "epoch": 0.23401826484018265, + "grad_norm": 5.78125, + "learning_rate": 9.87774048802227e-06, + "loss": 1.67071304, + "memory(GiB)": 71.21, + "step": 9225, + "train_speed(iter/s)": 0.448802 + }, + { + "acc": 0.63552632, + "epoch": 0.2341451040081177, + "grad_norm": 5.15625, + "learning_rate": 9.877509907891583e-06, + "loss": 1.6041235, + "memory(GiB)": 71.21, + "step": 9230, + "train_speed(iter/s)": 0.448929 + }, + { + "acc": 0.65549068, + "epoch": 0.23427194317605277, + "grad_norm": 5.125, + "learning_rate": 9.877279113226232e-06, + "loss": 1.5990778, + "memory(GiB)": 71.21, + "step": 9235, + "train_speed(iter/s)": 0.449056 + }, + { + "acc": 0.64702358, + "epoch": 0.2343987823439878, + "grad_norm": 5.3125, + "learning_rate": 9.87704810403637e-06, + "loss": 1.62338562, + "memory(GiB)": 71.21, + "step": 9240, + "train_speed(iter/s)": 0.449182 + }, + { + "acc": 0.63972073, + "epoch": 0.23452562151192288, + "grad_norm": 5.125, + "learning_rate": 9.876816880332157e-06, + "loss": 1.69598198, + "memory(GiB)": 71.21, + "step": 9245, + "train_speed(iter/s)": 0.449309 + }, + { + "acc": 0.64490156, + "epoch": 0.23465246067985793, + "grad_norm": 4.15625, + "learning_rate": 9.876585442123765e-06, + "loss": 1.67653465, + "memory(GiB)": 71.21, + "step": 9250, + "train_speed(iter/s)": 0.449435 + }, + { + "acc": 0.64251409, + "epoch": 0.234779299847793, + "grad_norm": 5.0625, + "learning_rate": 9.876353789421373e-06, + "loss": 1.73206291, + "memory(GiB)": 71.21, + "step": 9255, + "train_speed(iter/s)": 0.449562 + }, + { + "acc": 0.63710775, + "epoch": 0.23490613901572804, + "grad_norm": 6.875, + "learning_rate": 9.876121922235171e-06, + "loss": 1.75852489, + "memory(GiB)": 71.21, + "step": 9260, + "train_speed(iter/s)": 0.449688 + }, + { + "acc": 0.64012041, + "epoch": 0.23503297818366312, + "grad_norm": 5.03125, + "learning_rate": 9.875889840575356e-06, + "loss": 1.73697338, + "memory(GiB)": 71.21, + "step": 9265, + "train_speed(iter/s)": 0.449816 + }, + { + "acc": 0.63503313, + "epoch": 0.23515981735159816, + "grad_norm": 4.59375, + "learning_rate": 9.875657544452135e-06, + "loss": 1.67308979, + "memory(GiB)": 71.21, + "step": 9270, + "train_speed(iter/s)": 0.449942 + }, + { + "acc": 0.62886133, + "epoch": 0.23528665651953323, + "grad_norm": 5.4375, + "learning_rate": 9.875425033875728e-06, + "loss": 1.70906258, + "memory(GiB)": 71.21, + "step": 9275, + "train_speed(iter/s)": 0.450069 + }, + { + "acc": 0.65527387, + "epoch": 0.23541349568746828, + "grad_norm": 5.3125, + "learning_rate": 9.875192308856363e-06, + "loss": 1.65108147, + "memory(GiB)": 71.21, + "step": 9280, + "train_speed(iter/s)": 0.450195 + }, + { + "acc": 0.65551763, + "epoch": 0.23554033485540335, + "grad_norm": 5.84375, + "learning_rate": 9.87495936940427e-06, + "loss": 1.60943241, + "memory(GiB)": 71.21, + "step": 9285, + "train_speed(iter/s)": 0.450322 + }, + { + "acc": 0.62830639, + "epoch": 0.2356671740233384, + "grad_norm": 5.375, + "learning_rate": 9.874726215529702e-06, + "loss": 1.7238945, + "memory(GiB)": 71.21, + "step": 9290, + "train_speed(iter/s)": 0.45045 + }, + { + "acc": 0.64427629, + "epoch": 0.23579401319127347, + "grad_norm": 6.03125, + "learning_rate": 9.87449284724291e-06, + "loss": 1.62295799, + "memory(GiB)": 71.21, + "step": 9295, + "train_speed(iter/s)": 0.450576 + }, + { + "acc": 0.6435245, + "epoch": 0.2359208523592085, + "grad_norm": 5.25, + "learning_rate": 9.874259264554159e-06, + "loss": 1.70718803, + "memory(GiB)": 71.21, + "step": 9300, + "train_speed(iter/s)": 0.450702 + }, + { + "epoch": 0.2359208523592085, + "eval_acc": 0.6354466371193774, + "eval_loss": 1.6541646718978882, + "eval_runtime": 113.6514, + "eval_samples_per_second": 56.049, + "eval_steps_per_second": 28.024, + "step": 9300 + }, + { + "acc": 0.65446739, + "epoch": 0.23604769152714358, + "grad_norm": 4.78125, + "learning_rate": 9.874025467473722e-06, + "loss": 1.58965683, + "memory(GiB)": 71.21, + "step": 9305, + "train_speed(iter/s)": 0.448263 + }, + { + "acc": 0.62812486, + "epoch": 0.23617453069507863, + "grad_norm": 5.40625, + "learning_rate": 9.873791456011887e-06, + "loss": 1.70619049, + "memory(GiB)": 71.21, + "step": 9310, + "train_speed(iter/s)": 0.448389 + }, + { + "acc": 0.64351168, + "epoch": 0.2363013698630137, + "grad_norm": 6.0625, + "learning_rate": 9.873557230178942e-06, + "loss": 1.66674194, + "memory(GiB)": 71.21, + "step": 9315, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.64316278, + "epoch": 0.23642820903094874, + "grad_norm": 5.09375, + "learning_rate": 9.873322789985191e-06, + "loss": 1.64037819, + "memory(GiB)": 71.21, + "step": 9320, + "train_speed(iter/s)": 0.448639 + }, + { + "acc": 0.63499861, + "epoch": 0.23655504819888382, + "grad_norm": 5.5625, + "learning_rate": 9.873088135440949e-06, + "loss": 1.72765312, + "memory(GiB)": 71.21, + "step": 9325, + "train_speed(iter/s)": 0.448764 + }, + { + "acc": 0.61725483, + "epoch": 0.23668188736681886, + "grad_norm": 5.09375, + "learning_rate": 9.87285326655653e-06, + "loss": 1.74182358, + "memory(GiB)": 71.21, + "step": 9330, + "train_speed(iter/s)": 0.448889 + }, + { + "acc": 0.64406857, + "epoch": 0.23680872653475393, + "grad_norm": 5.0625, + "learning_rate": 9.87261818334227e-06, + "loss": 1.64370708, + "memory(GiB)": 71.21, + "step": 9335, + "train_speed(iter/s)": 0.449015 + }, + { + "acc": 0.62091289, + "epoch": 0.23693556570268898, + "grad_norm": 5.375, + "learning_rate": 9.87238288580851e-06, + "loss": 1.75851574, + "memory(GiB)": 71.21, + "step": 9340, + "train_speed(iter/s)": 0.44914 + }, + { + "acc": 0.64080849, + "epoch": 0.23706240487062405, + "grad_norm": 5.03125, + "learning_rate": 9.872147373965594e-06, + "loss": 1.68316193, + "memory(GiB)": 71.21, + "step": 9345, + "train_speed(iter/s)": 0.449266 + }, + { + "acc": 0.63035998, + "epoch": 0.2371892440385591, + "grad_norm": 5.40625, + "learning_rate": 9.871911647823884e-06, + "loss": 1.78609562, + "memory(GiB)": 71.21, + "step": 9350, + "train_speed(iter/s)": 0.449392 + }, + { + "acc": 0.64652686, + "epoch": 0.23731608320649417, + "grad_norm": 4.34375, + "learning_rate": 9.871675707393749e-06, + "loss": 1.65325546, + "memory(GiB)": 71.21, + "step": 9355, + "train_speed(iter/s)": 0.449518 + }, + { + "acc": 0.64698601, + "epoch": 0.2374429223744292, + "grad_norm": 5.40625, + "learning_rate": 9.871439552685566e-06, + "loss": 1.68890839, + "memory(GiB)": 71.21, + "step": 9360, + "train_speed(iter/s)": 0.449644 + }, + { + "acc": 0.64148483, + "epoch": 0.23756976154236428, + "grad_norm": 4.8125, + "learning_rate": 9.871203183709723e-06, + "loss": 1.68862762, + "memory(GiB)": 71.21, + "step": 9365, + "train_speed(iter/s)": 0.44977 + }, + { + "acc": 0.6490633, + "epoch": 0.23769660071029933, + "grad_norm": 5.96875, + "learning_rate": 9.870966600476614e-06, + "loss": 1.64669571, + "memory(GiB)": 71.21, + "step": 9370, + "train_speed(iter/s)": 0.449897 + }, + { + "acc": 0.65917749, + "epoch": 0.2378234398782344, + "grad_norm": 5.71875, + "learning_rate": 9.870729802996647e-06, + "loss": 1.68023338, + "memory(GiB)": 71.21, + "step": 9375, + "train_speed(iter/s)": 0.450022 + }, + { + "acc": 0.63755732, + "epoch": 0.23795027904616944, + "grad_norm": 6.03125, + "learning_rate": 9.870492791280239e-06, + "loss": 1.6498703, + "memory(GiB)": 71.21, + "step": 9380, + "train_speed(iter/s)": 0.450149 + }, + { + "acc": 0.63591413, + "epoch": 0.23807711821410452, + "grad_norm": 8.25, + "learning_rate": 9.87025556533781e-06, + "loss": 1.66962547, + "memory(GiB)": 71.21, + "step": 9385, + "train_speed(iter/s)": 0.450274 + }, + { + "acc": 0.67321892, + "epoch": 0.23820395738203956, + "grad_norm": 5.625, + "learning_rate": 9.870018125179799e-06, + "loss": 1.56504402, + "memory(GiB)": 71.21, + "step": 9390, + "train_speed(iter/s)": 0.4504 + }, + { + "acc": 0.65984888, + "epoch": 0.23833079654997463, + "grad_norm": 5.21875, + "learning_rate": 9.869780470816647e-06, + "loss": 1.55633698, + "memory(GiB)": 71.21, + "step": 9395, + "train_speed(iter/s)": 0.450526 + }, + { + "acc": 0.65697632, + "epoch": 0.23845763571790968, + "grad_norm": 4.21875, + "learning_rate": 9.869542602258809e-06, + "loss": 1.60326118, + "memory(GiB)": 71.21, + "step": 9400, + "train_speed(iter/s)": 0.45065 + }, + { + "epoch": 0.23845763571790968, + "eval_acc": 0.635622517156659, + "eval_loss": 1.6531695127487183, + "eval_runtime": 113.5957, + "eval_samples_per_second": 56.076, + "eval_steps_per_second": 28.038, + "step": 9400 + }, + { + "acc": 0.63990774, + "epoch": 0.23858447488584475, + "grad_norm": 6.28125, + "learning_rate": 9.869304519516745e-06, + "loss": 1.75399323, + "memory(GiB)": 71.21, + "step": 9405, + "train_speed(iter/s)": 0.448239 + }, + { + "acc": 0.64788933, + "epoch": 0.2387113140537798, + "grad_norm": 4.84375, + "learning_rate": 9.869066222600928e-06, + "loss": 1.70030308, + "memory(GiB)": 71.21, + "step": 9410, + "train_speed(iter/s)": 0.448365 + }, + { + "acc": 0.63833828, + "epoch": 0.23883815322171487, + "grad_norm": 4.59375, + "learning_rate": 9.86882771152184e-06, + "loss": 1.68622322, + "memory(GiB)": 71.21, + "step": 9415, + "train_speed(iter/s)": 0.448491 + }, + { + "acc": 0.63343077, + "epoch": 0.2389649923896499, + "grad_norm": 5.1875, + "learning_rate": 9.868588986289973e-06, + "loss": 1.78013668, + "memory(GiB)": 71.21, + "step": 9420, + "train_speed(iter/s)": 0.448618 + }, + { + "acc": 0.62793851, + "epoch": 0.23909183155758498, + "grad_norm": 5.28125, + "learning_rate": 9.868350046915825e-06, + "loss": 1.74199944, + "memory(GiB)": 71.21, + "step": 9425, + "train_speed(iter/s)": 0.448745 + }, + { + "acc": 0.63901939, + "epoch": 0.23921867072552003, + "grad_norm": 5.875, + "learning_rate": 9.868110893409906e-06, + "loss": 1.74208241, + "memory(GiB)": 71.21, + "step": 9430, + "train_speed(iter/s)": 0.448871 + }, + { + "acc": 0.66115327, + "epoch": 0.2393455098934551, + "grad_norm": 5.875, + "learning_rate": 9.867871525782735e-06, + "loss": 1.6152317, + "memory(GiB)": 71.21, + "step": 9435, + "train_speed(iter/s)": 0.448998 + }, + { + "acc": 0.64950633, + "epoch": 0.23947234906139014, + "grad_norm": 5.09375, + "learning_rate": 9.86763194404484e-06, + "loss": 1.65739822, + "memory(GiB)": 71.21, + "step": 9440, + "train_speed(iter/s)": 0.449126 + }, + { + "acc": 0.62799082, + "epoch": 0.23959918822932522, + "grad_norm": 5.28125, + "learning_rate": 9.867392148206762e-06, + "loss": 1.73607903, + "memory(GiB)": 71.21, + "step": 9445, + "train_speed(iter/s)": 0.449253 + }, + { + "acc": 0.64420195, + "epoch": 0.23972602739726026, + "grad_norm": 6.15625, + "learning_rate": 9.867152138279043e-06, + "loss": 1.67406158, + "memory(GiB)": 71.21, + "step": 9450, + "train_speed(iter/s)": 0.44938 + }, + { + "acc": 0.64051237, + "epoch": 0.23985286656519533, + "grad_norm": 4.65625, + "learning_rate": 9.866911914272246e-06, + "loss": 1.66110554, + "memory(GiB)": 71.21, + "step": 9455, + "train_speed(iter/s)": 0.449506 + }, + { + "acc": 0.64643135, + "epoch": 0.23997970573313038, + "grad_norm": 5.125, + "learning_rate": 9.866671476196931e-06, + "loss": 1.6352808, + "memory(GiB)": 71.21, + "step": 9460, + "train_speed(iter/s)": 0.449633 + }, + { + "acc": 0.65710506, + "epoch": 0.24010654490106545, + "grad_norm": 4.9375, + "learning_rate": 9.866430824063678e-06, + "loss": 1.61370506, + "memory(GiB)": 71.21, + "step": 9465, + "train_speed(iter/s)": 0.44976 + }, + { + "acc": 0.65134134, + "epoch": 0.2402333840690005, + "grad_norm": 5.28125, + "learning_rate": 9.86618995788307e-06, + "loss": 1.61107235, + "memory(GiB)": 84.18, + "step": 9470, + "train_speed(iter/s)": 0.449886 + }, + { + "acc": 0.62410879, + "epoch": 0.24036022323693557, + "grad_norm": 6.0, + "learning_rate": 9.865948877665702e-06, + "loss": 1.69452362, + "memory(GiB)": 84.18, + "step": 9475, + "train_speed(iter/s)": 0.450013 + }, + { + "acc": 0.64340334, + "epoch": 0.2404870624048706, + "grad_norm": 5.9375, + "learning_rate": 9.865707583422178e-06, + "loss": 1.71322784, + "memory(GiB)": 84.18, + "step": 9480, + "train_speed(iter/s)": 0.45014 + }, + { + "acc": 0.63803396, + "epoch": 0.24061390157280568, + "grad_norm": 5.34375, + "learning_rate": 9.865466075163108e-06, + "loss": 1.69009895, + "memory(GiB)": 84.18, + "step": 9485, + "train_speed(iter/s)": 0.450266 + }, + { + "acc": 0.65077553, + "epoch": 0.24074074074074073, + "grad_norm": 4.78125, + "learning_rate": 9.86522435289912e-06, + "loss": 1.6221138, + "memory(GiB)": 84.18, + "step": 9490, + "train_speed(iter/s)": 0.450393 + }, + { + "acc": 0.63570275, + "epoch": 0.2408675799086758, + "grad_norm": 5.34375, + "learning_rate": 9.864982416640843e-06, + "loss": 1.73583221, + "memory(GiB)": 84.18, + "step": 9495, + "train_speed(iter/s)": 0.450519 + }, + { + "acc": 0.64468632, + "epoch": 0.24099441907661084, + "grad_norm": 5.125, + "learning_rate": 9.864740266398918e-06, + "loss": 1.67825508, + "memory(GiB)": 84.18, + "step": 9500, + "train_speed(iter/s)": 0.450647 + }, + { + "epoch": 0.24099441907661084, + "eval_acc": 0.6357670646457312, + "eval_loss": 1.6526737213134766, + "eval_runtime": 112.8007, + "eval_samples_per_second": 56.471, + "eval_steps_per_second": 28.236, + "step": 9500 + }, + { + "acc": 0.65441008, + "epoch": 0.24112125824454592, + "grad_norm": 5.28125, + "learning_rate": 9.864497902183996e-06, + "loss": 1.62061348, + "memory(GiB)": 84.18, + "step": 9505, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.64899931, + "epoch": 0.24124809741248096, + "grad_norm": 5.5625, + "learning_rate": 9.864255324006738e-06, + "loss": 1.66408501, + "memory(GiB)": 84.18, + "step": 9510, + "train_speed(iter/s)": 0.448405 + }, + { + "acc": 0.62924976, + "epoch": 0.24137493658041603, + "grad_norm": 5.5625, + "learning_rate": 9.864012531877814e-06, + "loss": 1.72995281, + "memory(GiB)": 84.18, + "step": 9515, + "train_speed(iter/s)": 0.448531 + }, + { + "acc": 0.63176484, + "epoch": 0.24150177574835108, + "grad_norm": 5.1875, + "learning_rate": 9.863769525807903e-06, + "loss": 1.73512478, + "memory(GiB)": 84.18, + "step": 9520, + "train_speed(iter/s)": 0.448658 + }, + { + "acc": 0.66021013, + "epoch": 0.24162861491628615, + "grad_norm": 5.5, + "learning_rate": 9.863526305807694e-06, + "loss": 1.60844841, + "memory(GiB)": 84.18, + "step": 9525, + "train_speed(iter/s)": 0.448785 + }, + { + "acc": 0.64407992, + "epoch": 0.2417554540842212, + "grad_norm": 7.46875, + "learning_rate": 9.863282871887882e-06, + "loss": 1.75947914, + "memory(GiB)": 84.18, + "step": 9530, + "train_speed(iter/s)": 0.448911 + }, + { + "acc": 0.64780474, + "epoch": 0.24188229325215627, + "grad_norm": 4.9375, + "learning_rate": 9.863039224059177e-06, + "loss": 1.66074753, + "memory(GiB)": 84.18, + "step": 9535, + "train_speed(iter/s)": 0.449037 + }, + { + "acc": 0.6472455, + "epoch": 0.2420091324200913, + "grad_norm": 4.75, + "learning_rate": 9.862795362332293e-06, + "loss": 1.62877121, + "memory(GiB)": 84.18, + "step": 9540, + "train_speed(iter/s)": 0.449163 + }, + { + "acc": 0.66378317, + "epoch": 0.24213597158802638, + "grad_norm": 5.0625, + "learning_rate": 9.862551286717961e-06, + "loss": 1.61465034, + "memory(GiB)": 84.18, + "step": 9545, + "train_speed(iter/s)": 0.449288 + }, + { + "acc": 0.65136733, + "epoch": 0.24226281075596143, + "grad_norm": 4.625, + "learning_rate": 9.862306997226914e-06, + "loss": 1.61032352, + "memory(GiB)": 84.18, + "step": 9550, + "train_speed(iter/s)": 0.449413 + }, + { + "acc": 0.64301291, + "epoch": 0.2423896499238965, + "grad_norm": 6.28125, + "learning_rate": 9.862062493869895e-06, + "loss": 1.71672344, + "memory(GiB)": 84.18, + "step": 9555, + "train_speed(iter/s)": 0.449539 + }, + { + "acc": 0.64165573, + "epoch": 0.24251648909183154, + "grad_norm": 5.46875, + "learning_rate": 9.861817776657661e-06, + "loss": 1.61028442, + "memory(GiB)": 84.18, + "step": 9560, + "train_speed(iter/s)": 0.449665 + }, + { + "acc": 0.62407374, + "epoch": 0.24264332825976662, + "grad_norm": 6.75, + "learning_rate": 9.861572845600973e-06, + "loss": 1.70165749, + "memory(GiB)": 84.18, + "step": 9565, + "train_speed(iter/s)": 0.449792 + }, + { + "acc": 0.64178395, + "epoch": 0.24277016742770166, + "grad_norm": 4.65625, + "learning_rate": 9.861327700710608e-06, + "loss": 1.67437096, + "memory(GiB)": 84.18, + "step": 9570, + "train_speed(iter/s)": 0.449917 + }, + { + "acc": 0.63719091, + "epoch": 0.24289700659563673, + "grad_norm": 5.34375, + "learning_rate": 9.861082341997345e-06, + "loss": 1.71820889, + "memory(GiB)": 84.18, + "step": 9575, + "train_speed(iter/s)": 0.450043 + }, + { + "acc": 0.64579763, + "epoch": 0.24302384576357178, + "grad_norm": 4.34375, + "learning_rate": 9.860836769471977e-06, + "loss": 1.658815, + "memory(GiB)": 84.18, + "step": 9580, + "train_speed(iter/s)": 0.450168 + }, + { + "acc": 0.63311477, + "epoch": 0.24315068493150685, + "grad_norm": 5.1875, + "learning_rate": 9.860590983145307e-06, + "loss": 1.73535118, + "memory(GiB)": 84.18, + "step": 9585, + "train_speed(iter/s)": 0.450293 + }, + { + "acc": 0.64140406, + "epoch": 0.2432775240994419, + "grad_norm": 5.6875, + "learning_rate": 9.860344983028146e-06, + "loss": 1.69727478, + "memory(GiB)": 84.18, + "step": 9590, + "train_speed(iter/s)": 0.450419 + }, + { + "acc": 0.63631096, + "epoch": 0.24340436326737697, + "grad_norm": 4.9375, + "learning_rate": 9.86009876913131e-06, + "loss": 1.71041374, + "memory(GiB)": 84.18, + "step": 9595, + "train_speed(iter/s)": 0.450545 + }, + { + "acc": 0.63834791, + "epoch": 0.243531202435312, + "grad_norm": 3.984375, + "learning_rate": 9.859852341465633e-06, + "loss": 1.74432907, + "memory(GiB)": 84.18, + "step": 9600, + "train_speed(iter/s)": 0.45067 + }, + { + "epoch": 0.243531202435312, + "eval_acc": 0.6357954728227744, + "eval_loss": 1.6525287628173828, + "eval_runtime": 113.7275, + "eval_samples_per_second": 56.011, + "eval_steps_per_second": 28.006, + "step": 9600 + }, + { + "acc": 0.63986053, + "epoch": 0.24365804160324708, + "grad_norm": 5.1875, + "learning_rate": 9.859605700041951e-06, + "loss": 1.71194496, + "memory(GiB)": 84.18, + "step": 9605, + "train_speed(iter/s)": 0.448307 + }, + { + "acc": 0.65924358, + "epoch": 0.24378488077118213, + "grad_norm": 5.65625, + "learning_rate": 9.859358844871113e-06, + "loss": 1.61128387, + "memory(GiB)": 84.18, + "step": 9610, + "train_speed(iter/s)": 0.448432 + }, + { + "acc": 0.64468589, + "epoch": 0.2439117199391172, + "grad_norm": 5.1875, + "learning_rate": 9.859111775963981e-06, + "loss": 1.68093033, + "memory(GiB)": 84.18, + "step": 9615, + "train_speed(iter/s)": 0.448556 + }, + { + "acc": 0.65342407, + "epoch": 0.24403855910705224, + "grad_norm": 5.1875, + "learning_rate": 9.858864493331417e-06, + "loss": 1.6401659, + "memory(GiB)": 84.18, + "step": 9620, + "train_speed(iter/s)": 0.44868 + }, + { + "acc": 0.63621445, + "epoch": 0.24416539827498732, + "grad_norm": 5.75, + "learning_rate": 9.858616996984297e-06, + "loss": 1.72697945, + "memory(GiB)": 84.18, + "step": 9625, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.63256578, + "epoch": 0.24429223744292236, + "grad_norm": 5.5, + "learning_rate": 9.858369286933513e-06, + "loss": 1.71205883, + "memory(GiB)": 84.18, + "step": 9630, + "train_speed(iter/s)": 0.448931 + }, + { + "acc": 0.64839354, + "epoch": 0.24441907661085743, + "grad_norm": 5.09375, + "learning_rate": 9.858121363189954e-06, + "loss": 1.66816864, + "memory(GiB)": 84.18, + "step": 9635, + "train_speed(iter/s)": 0.449055 + }, + { + "acc": 0.63687744, + "epoch": 0.24454591577879248, + "grad_norm": 5.34375, + "learning_rate": 9.85787322576453e-06, + "loss": 1.7540205, + "memory(GiB)": 84.18, + "step": 9640, + "train_speed(iter/s)": 0.44918 + }, + { + "acc": 0.64409537, + "epoch": 0.24467275494672755, + "grad_norm": 4.15625, + "learning_rate": 9.85762487466815e-06, + "loss": 1.64781952, + "memory(GiB)": 84.18, + "step": 9645, + "train_speed(iter/s)": 0.449303 + }, + { + "acc": 0.6551383, + "epoch": 0.2447995941146626, + "grad_norm": 5.1875, + "learning_rate": 9.857376309911741e-06, + "loss": 1.53300838, + "memory(GiB)": 84.18, + "step": 9650, + "train_speed(iter/s)": 0.449427 + }, + { + "acc": 0.63431463, + "epoch": 0.24492643328259767, + "grad_norm": 4.59375, + "learning_rate": 9.857127531506237e-06, + "loss": 1.7407032, + "memory(GiB)": 84.18, + "step": 9655, + "train_speed(iter/s)": 0.449551 + }, + { + "acc": 0.64181747, + "epoch": 0.2450532724505327, + "grad_norm": 5.625, + "learning_rate": 9.856878539462577e-06, + "loss": 1.67271271, + "memory(GiB)": 84.18, + "step": 9660, + "train_speed(iter/s)": 0.449673 + }, + { + "acc": 0.66206522, + "epoch": 0.24518011161846778, + "grad_norm": 5.1875, + "learning_rate": 9.856629333791716e-06, + "loss": 1.59543552, + "memory(GiB)": 84.18, + "step": 9665, + "train_speed(iter/s)": 0.449797 + }, + { + "acc": 0.64142909, + "epoch": 0.24530695078640283, + "grad_norm": 4.9375, + "learning_rate": 9.856379914504612e-06, + "loss": 1.71039314, + "memory(GiB)": 84.18, + "step": 9670, + "train_speed(iter/s)": 0.449921 + }, + { + "acc": 0.64408822, + "epoch": 0.2454337899543379, + "grad_norm": 4.59375, + "learning_rate": 9.856130281612237e-06, + "loss": 1.69994144, + "memory(GiB)": 84.18, + "step": 9675, + "train_speed(iter/s)": 0.450045 + }, + { + "acc": 0.63171906, + "epoch": 0.24556062912227294, + "grad_norm": 5.09375, + "learning_rate": 9.855880435125572e-06, + "loss": 1.63647842, + "memory(GiB)": 84.18, + "step": 9680, + "train_speed(iter/s)": 0.450169 + }, + { + "acc": 0.63819122, + "epoch": 0.24568746829020802, + "grad_norm": 5.53125, + "learning_rate": 9.855630375055604e-06, + "loss": 1.64949474, + "memory(GiB)": 84.18, + "step": 9685, + "train_speed(iter/s)": 0.450292 + }, + { + "acc": 0.64347181, + "epoch": 0.24581430745814306, + "grad_norm": 5.09375, + "learning_rate": 9.855380101413336e-06, + "loss": 1.69322319, + "memory(GiB)": 84.18, + "step": 9690, + "train_speed(iter/s)": 0.450415 + }, + { + "acc": 0.63900776, + "epoch": 0.24594114662607813, + "grad_norm": 5.28125, + "learning_rate": 9.855129614209771e-06, + "loss": 1.6777544, + "memory(GiB)": 84.18, + "step": 9695, + "train_speed(iter/s)": 0.450539 + }, + { + "acc": 0.64113932, + "epoch": 0.24606798579401318, + "grad_norm": 5.71875, + "learning_rate": 9.85487891345593e-06, + "loss": 1.72149887, + "memory(GiB)": 84.18, + "step": 9700, + "train_speed(iter/s)": 0.450663 + }, + { + "epoch": 0.24606798579401318, + "eval_acc": 0.6359888990870531, + "eval_loss": 1.6516163349151611, + "eval_runtime": 113.3401, + "eval_samples_per_second": 56.203, + "eval_steps_per_second": 28.101, + "step": 9700 + }, + { + "acc": 0.63458576, + "epoch": 0.24619482496194825, + "grad_norm": 6.28125, + "learning_rate": 9.85462799916284e-06, + "loss": 1.74244194, + "memory(GiB)": 84.18, + "step": 9705, + "train_speed(iter/s)": 0.448333 + }, + { + "acc": 0.63018379, + "epoch": 0.2463216641298833, + "grad_norm": 5.40625, + "learning_rate": 9.854376871341535e-06, + "loss": 1.70108566, + "memory(GiB)": 84.18, + "step": 9710, + "train_speed(iter/s)": 0.448457 + }, + { + "acc": 0.63088713, + "epoch": 0.24644850329781837, + "grad_norm": 6.5, + "learning_rate": 9.854125530003063e-06, + "loss": 1.73135223, + "memory(GiB)": 84.18, + "step": 9715, + "train_speed(iter/s)": 0.448581 + }, + { + "acc": 0.64723148, + "epoch": 0.2465753424657534, + "grad_norm": 6.0625, + "learning_rate": 9.853873975158476e-06, + "loss": 1.64123268, + "memory(GiB)": 84.18, + "step": 9720, + "train_speed(iter/s)": 0.448704 + }, + { + "acc": 0.63749256, + "epoch": 0.24670218163368848, + "grad_norm": 5.4375, + "learning_rate": 9.853622206818842e-06, + "loss": 1.75529137, + "memory(GiB)": 84.18, + "step": 9725, + "train_speed(iter/s)": 0.448828 + }, + { + "acc": 0.64337592, + "epoch": 0.24682902080162353, + "grad_norm": 4.6875, + "learning_rate": 9.853370224995233e-06, + "loss": 1.64793549, + "memory(GiB)": 84.18, + "step": 9730, + "train_speed(iter/s)": 0.44895 + }, + { + "acc": 0.65726776, + "epoch": 0.2469558599695586, + "grad_norm": 5.90625, + "learning_rate": 9.853118029698733e-06, + "loss": 1.60163612, + "memory(GiB)": 84.18, + "step": 9735, + "train_speed(iter/s)": 0.449072 + }, + { + "acc": 0.6583992, + "epoch": 0.24708269913749364, + "grad_norm": 5.84375, + "learning_rate": 9.852865620940436e-06, + "loss": 1.69895668, + "memory(GiB)": 84.18, + "step": 9740, + "train_speed(iter/s)": 0.449195 + }, + { + "acc": 0.64079418, + "epoch": 0.24720953830542872, + "grad_norm": 4.71875, + "learning_rate": 9.85261299873144e-06, + "loss": 1.66521473, + "memory(GiB)": 84.18, + "step": 9745, + "train_speed(iter/s)": 0.449318 + }, + { + "acc": 0.6563859, + "epoch": 0.24733637747336376, + "grad_norm": 5.0625, + "learning_rate": 9.85236016308286e-06, + "loss": 1.64144573, + "memory(GiB)": 84.18, + "step": 9750, + "train_speed(iter/s)": 0.449441 + }, + { + "acc": 0.6437757, + "epoch": 0.24746321664129883, + "grad_norm": 4.9375, + "learning_rate": 9.852107114005816e-06, + "loss": 1.68185616, + "memory(GiB)": 84.18, + "step": 9755, + "train_speed(iter/s)": 0.449565 + }, + { + "acc": 0.66282849, + "epoch": 0.24759005580923388, + "grad_norm": 5.03125, + "learning_rate": 9.851853851511437e-06, + "loss": 1.63104763, + "memory(GiB)": 84.18, + "step": 9760, + "train_speed(iter/s)": 0.449689 + }, + { + "acc": 0.6452096, + "epoch": 0.24771689497716895, + "grad_norm": 5.90625, + "learning_rate": 9.851600375610864e-06, + "loss": 1.64493637, + "memory(GiB)": 84.18, + "step": 9765, + "train_speed(iter/s)": 0.449813 + }, + { + "acc": 0.64191589, + "epoch": 0.247843734145104, + "grad_norm": 5.375, + "learning_rate": 9.851346686315246e-06, + "loss": 1.66515808, + "memory(GiB)": 84.18, + "step": 9770, + "train_speed(iter/s)": 0.449936 + }, + { + "acc": 0.63775434, + "epoch": 0.24797057331303907, + "grad_norm": 4.28125, + "learning_rate": 9.851092783635742e-06, + "loss": 1.73454018, + "memory(GiB)": 84.18, + "step": 9775, + "train_speed(iter/s)": 0.450056 + }, + { + "acc": 0.64558911, + "epoch": 0.2480974124809741, + "grad_norm": 6.0, + "learning_rate": 9.850838667583518e-06, + "loss": 1.67103024, + "memory(GiB)": 84.18, + "step": 9780, + "train_speed(iter/s)": 0.450178 + }, + { + "acc": 0.65088477, + "epoch": 0.24822425164890918, + "grad_norm": 5.5625, + "learning_rate": 9.850584338169752e-06, + "loss": 1.6459053, + "memory(GiB)": 84.18, + "step": 9785, + "train_speed(iter/s)": 0.450301 + }, + { + "acc": 0.63854332, + "epoch": 0.24835109081684423, + "grad_norm": 5.0, + "learning_rate": 9.85032979540563e-06, + "loss": 1.69356747, + "memory(GiB)": 84.18, + "step": 9790, + "train_speed(iter/s)": 0.450422 + }, + { + "acc": 0.63480186, + "epoch": 0.2484779299847793, + "grad_norm": 4.90625, + "learning_rate": 9.85007503930235e-06, + "loss": 1.71151695, + "memory(GiB)": 84.18, + "step": 9795, + "train_speed(iter/s)": 0.450543 + }, + { + "acc": 0.65237126, + "epoch": 0.24860476915271434, + "grad_norm": 5.125, + "learning_rate": 9.849820069871114e-06, + "loss": 1.67856102, + "memory(GiB)": 84.18, + "step": 9800, + "train_speed(iter/s)": 0.450664 + }, + { + "epoch": 0.24860476915271434, + "eval_acc": 0.635965921885033, + "eval_loss": 1.6508677005767822, + "eval_runtime": 113.9102, + "eval_samples_per_second": 55.921, + "eval_steps_per_second": 27.961, + "step": 9800 + }, + { + "acc": 0.63831329, + "epoch": 0.24873160832064942, + "grad_norm": 6.3125, + "learning_rate": 9.849564887123138e-06, + "loss": 1.69200115, + "memory(GiB)": 84.18, + "step": 9805, + "train_speed(iter/s)": 0.448344 + }, + { + "acc": 0.63247623, + "epoch": 0.24885844748858446, + "grad_norm": 5.0, + "learning_rate": 9.849309491069647e-06, + "loss": 1.68875885, + "memory(GiB)": 84.18, + "step": 9810, + "train_speed(iter/s)": 0.448465 + }, + { + "acc": 0.64661217, + "epoch": 0.24898528665651953, + "grad_norm": 6.375, + "learning_rate": 9.849053881721876e-06, + "loss": 1.63333359, + "memory(GiB)": 84.18, + "step": 9815, + "train_speed(iter/s)": 0.448585 + }, + { + "acc": 0.64804544, + "epoch": 0.24911212582445458, + "grad_norm": 4.84375, + "learning_rate": 9.848798059091064e-06, + "loss": 1.71267929, + "memory(GiB)": 84.18, + "step": 9820, + "train_speed(iter/s)": 0.448706 + }, + { + "acc": 0.64320364, + "epoch": 0.24923896499238965, + "grad_norm": 5.375, + "learning_rate": 9.848542023188466e-06, + "loss": 1.717276, + "memory(GiB)": 84.18, + "step": 9825, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.66500702, + "epoch": 0.2493658041603247, + "grad_norm": 5.53125, + "learning_rate": 9.848285774025342e-06, + "loss": 1.55275669, + "memory(GiB)": 84.18, + "step": 9830, + "train_speed(iter/s)": 0.448949 + }, + { + "acc": 0.64382815, + "epoch": 0.24949264332825977, + "grad_norm": 5.4375, + "learning_rate": 9.848029311612963e-06, + "loss": 1.67573204, + "memory(GiB)": 84.18, + "step": 9835, + "train_speed(iter/s)": 0.449069 + }, + { + "acc": 0.64469352, + "epoch": 0.2496194824961948, + "grad_norm": 5.96875, + "learning_rate": 9.84777263596261e-06, + "loss": 1.68478775, + "memory(GiB)": 84.18, + "step": 9840, + "train_speed(iter/s)": 0.449191 + }, + { + "acc": 0.63684015, + "epoch": 0.24974632166412988, + "grad_norm": 4.90625, + "learning_rate": 9.847515747085573e-06, + "loss": 1.70679092, + "memory(GiB)": 84.18, + "step": 9845, + "train_speed(iter/s)": 0.449312 + }, + { + "acc": 0.64473124, + "epoch": 0.24987316083206493, + "grad_norm": 4.8125, + "learning_rate": 9.847258644993151e-06, + "loss": 1.68105564, + "memory(GiB)": 84.18, + "step": 9850, + "train_speed(iter/s)": 0.449434 + }, + { + "acc": 0.62908134, + "epoch": 0.25, + "grad_norm": 4.75, + "learning_rate": 9.847001329696653e-06, + "loss": 1.71537418, + "memory(GiB)": 84.18, + "step": 9855, + "train_speed(iter/s)": 0.449555 + }, + { + "acc": 0.64225731, + "epoch": 0.25012683916793504, + "grad_norm": 4.875, + "learning_rate": 9.846743801207395e-06, + "loss": 1.66853409, + "memory(GiB)": 84.18, + "step": 9860, + "train_speed(iter/s)": 0.449677 + }, + { + "acc": 0.62420464, + "epoch": 0.25025367833587014, + "grad_norm": 5.375, + "learning_rate": 9.846486059536706e-06, + "loss": 1.72606525, + "memory(GiB)": 97.17, + "step": 9865, + "train_speed(iter/s)": 0.449797 + }, + { + "acc": 0.64164133, + "epoch": 0.2503805175038052, + "grad_norm": 4.84375, + "learning_rate": 9.846228104695922e-06, + "loss": 1.67052307, + "memory(GiB)": 97.17, + "step": 9870, + "train_speed(iter/s)": 0.449918 + }, + { + "acc": 0.63010015, + "epoch": 0.25050735667174023, + "grad_norm": 5.46875, + "learning_rate": 9.84596993669639e-06, + "loss": 1.77863979, + "memory(GiB)": 97.17, + "step": 9875, + "train_speed(iter/s)": 0.450038 + }, + { + "acc": 0.64390774, + "epoch": 0.2506341958396753, + "grad_norm": 7.625, + "learning_rate": 9.845711555549464e-06, + "loss": 1.68367977, + "memory(GiB)": 97.17, + "step": 9880, + "train_speed(iter/s)": 0.450158 + }, + { + "acc": 0.64109015, + "epoch": 0.2507610350076104, + "grad_norm": 4.96875, + "learning_rate": 9.845452961266509e-06, + "loss": 1.64917259, + "memory(GiB)": 97.17, + "step": 9885, + "train_speed(iter/s)": 0.450277 + }, + { + "acc": 0.65007424, + "epoch": 0.2508878741755454, + "grad_norm": 5.4375, + "learning_rate": 9.845194153858899e-06, + "loss": 1.6526825, + "memory(GiB)": 97.17, + "step": 9890, + "train_speed(iter/s)": 0.450397 + }, + { + "acc": 0.64389496, + "epoch": 0.25101471334348047, + "grad_norm": 4.78125, + "learning_rate": 9.844935133338018e-06, + "loss": 1.65119019, + "memory(GiB)": 97.17, + "step": 9895, + "train_speed(iter/s)": 0.450516 + }, + { + "acc": 0.62651777, + "epoch": 0.2511415525114155, + "grad_norm": 4.84375, + "learning_rate": 9.84467589971526e-06, + "loss": 1.78537216, + "memory(GiB)": 97.17, + "step": 9900, + "train_speed(iter/s)": 0.450635 + }, + { + "epoch": 0.2511415525114155, + "eval_acc": 0.6361025317952255, + "eval_loss": 1.650057077407837, + "eval_runtime": 112.2369, + "eval_samples_per_second": 56.755, + "eval_steps_per_second": 28.377, + "step": 9900 + }, + { + "acc": 0.64691887, + "epoch": 0.2512683916793506, + "grad_norm": 5.75, + "learning_rate": 9.844416453002027e-06, + "loss": 1.69708214, + "memory(GiB)": 97.17, + "step": 9905, + "train_speed(iter/s)": 0.448373 + }, + { + "acc": 0.64094362, + "epoch": 0.25139523084728566, + "grad_norm": 5.125, + "learning_rate": 9.844156793209725e-06, + "loss": 1.65398254, + "memory(GiB)": 97.17, + "step": 9910, + "train_speed(iter/s)": 0.448493 + }, + { + "acc": 0.64647436, + "epoch": 0.2515220700152207, + "grad_norm": 5.21875, + "learning_rate": 9.843896920349783e-06, + "loss": 1.64679718, + "memory(GiB)": 97.17, + "step": 9915, + "train_speed(iter/s)": 0.448611 + }, + { + "acc": 0.64076943, + "epoch": 0.25164890918315574, + "grad_norm": 4.5, + "learning_rate": 9.843636834433627e-06, + "loss": 1.69570084, + "memory(GiB)": 97.17, + "step": 9920, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.63992963, + "epoch": 0.25177574835109084, + "grad_norm": 5.5, + "learning_rate": 9.843376535472698e-06, + "loss": 1.67204361, + "memory(GiB)": 97.17, + "step": 9925, + "train_speed(iter/s)": 0.44885 + }, + { + "acc": 0.65698419, + "epoch": 0.2519025875190259, + "grad_norm": 4.28125, + "learning_rate": 9.843116023478445e-06, + "loss": 1.59887962, + "memory(GiB)": 97.17, + "step": 9930, + "train_speed(iter/s)": 0.448969 + }, + { + "acc": 0.6405468, + "epoch": 0.25202942668696093, + "grad_norm": 4.84375, + "learning_rate": 9.842855298462327e-06, + "loss": 1.69015961, + "memory(GiB)": 97.17, + "step": 9935, + "train_speed(iter/s)": 0.449087 + }, + { + "acc": 0.63615909, + "epoch": 0.252156265854896, + "grad_norm": 6.21875, + "learning_rate": 9.84259436043581e-06, + "loss": 1.71440792, + "memory(GiB)": 97.17, + "step": 9940, + "train_speed(iter/s)": 0.449207 + }, + { + "acc": 0.64654627, + "epoch": 0.2522831050228311, + "grad_norm": 5.65625, + "learning_rate": 9.842333209410372e-06, + "loss": 1.63373108, + "memory(GiB)": 97.17, + "step": 9945, + "train_speed(iter/s)": 0.449325 + }, + { + "acc": 0.6533143, + "epoch": 0.2524099441907661, + "grad_norm": 5.03125, + "learning_rate": 9.842071845397502e-06, + "loss": 1.56941128, + "memory(GiB)": 97.17, + "step": 9950, + "train_speed(iter/s)": 0.449444 + }, + { + "acc": 0.63823442, + "epoch": 0.25253678335870117, + "grad_norm": 5.0, + "learning_rate": 9.841810268408692e-06, + "loss": 1.71749573, + "memory(GiB)": 97.17, + "step": 9955, + "train_speed(iter/s)": 0.449563 + }, + { + "acc": 0.64907484, + "epoch": 0.2526636225266362, + "grad_norm": 4.65625, + "learning_rate": 9.841548478455451e-06, + "loss": 1.72018814, + "memory(GiB)": 97.17, + "step": 9960, + "train_speed(iter/s)": 0.449681 + }, + { + "acc": 0.64448767, + "epoch": 0.2527904616945713, + "grad_norm": 5.375, + "learning_rate": 9.841286475549291e-06, + "loss": 1.70823631, + "memory(GiB)": 97.17, + "step": 9965, + "train_speed(iter/s)": 0.449799 + }, + { + "acc": 0.64359264, + "epoch": 0.25291730086250636, + "grad_norm": 5.65625, + "learning_rate": 9.841024259701737e-06, + "loss": 1.65259724, + "memory(GiB)": 97.17, + "step": 9970, + "train_speed(iter/s)": 0.449917 + }, + { + "acc": 0.64129829, + "epoch": 0.2530441400304414, + "grad_norm": 4.875, + "learning_rate": 9.840761830924323e-06, + "loss": 1.67311554, + "memory(GiB)": 97.17, + "step": 9975, + "train_speed(iter/s)": 0.450037 + }, + { + "acc": 0.64141607, + "epoch": 0.25317097919837644, + "grad_norm": 7.0, + "learning_rate": 9.84049918922859e-06, + "loss": 1.69971275, + "memory(GiB)": 97.17, + "step": 9980, + "train_speed(iter/s)": 0.450155 + }, + { + "acc": 0.63405867, + "epoch": 0.25329781836631154, + "grad_norm": 5.3125, + "learning_rate": 9.840236334626091e-06, + "loss": 1.68541508, + "memory(GiB)": 97.17, + "step": 9985, + "train_speed(iter/s)": 0.450274 + }, + { + "acc": 0.6443634, + "epoch": 0.2534246575342466, + "grad_norm": 4.21875, + "learning_rate": 9.83997326712839e-06, + "loss": 1.64873981, + "memory(GiB)": 97.17, + "step": 9990, + "train_speed(iter/s)": 0.450393 + }, + { + "acc": 0.65700703, + "epoch": 0.25355149670218163, + "grad_norm": 4.4375, + "learning_rate": 9.839709986747054e-06, + "loss": 1.59298191, + "memory(GiB)": 97.17, + "step": 9995, + "train_speed(iter/s)": 0.450512 + }, + { + "acc": 0.65080976, + "epoch": 0.2536783358701167, + "grad_norm": 4.875, + "learning_rate": 9.839446493493667e-06, + "loss": 1.57753963, + "memory(GiB)": 97.17, + "step": 10000, + "train_speed(iter/s)": 0.450629 + }, + { + "epoch": 0.2536783358701167, + "eval_acc": 0.6362470792842978, + "eval_loss": 1.649816632270813, + "eval_runtime": 112.9138, + "eval_samples_per_second": 56.415, + "eval_steps_per_second": 28.207, + "step": 10000 + }, + { + "acc": 0.63965268, + "epoch": 0.2538051750380518, + "grad_norm": 4.75, + "learning_rate": 9.839182787379815e-06, + "loss": 1.61130695, + "memory(GiB)": 97.17, + "step": 10005, + "train_speed(iter/s)": 0.448375 + }, + { + "acc": 0.64882364, + "epoch": 0.2539320142059868, + "grad_norm": 5.3125, + "learning_rate": 9.8389188684171e-06, + "loss": 1.6103384, + "memory(GiB)": 97.17, + "step": 10010, + "train_speed(iter/s)": 0.448492 + }, + { + "acc": 0.65004439, + "epoch": 0.25405885337392187, + "grad_norm": 5.15625, + "learning_rate": 9.838654736617128e-06, + "loss": 1.57968521, + "memory(GiB)": 97.17, + "step": 10015, + "train_speed(iter/s)": 0.44861 + }, + { + "acc": 0.64287462, + "epoch": 0.2541856925418569, + "grad_norm": 5.09375, + "learning_rate": 9.838390391991517e-06, + "loss": 1.66347008, + "memory(GiB)": 97.17, + "step": 10020, + "train_speed(iter/s)": 0.448729 + }, + { + "acc": 0.63177919, + "epoch": 0.254312531709792, + "grad_norm": 5.5625, + "learning_rate": 9.838125834551895e-06, + "loss": 1.69404659, + "memory(GiB)": 97.17, + "step": 10025, + "train_speed(iter/s)": 0.448846 + }, + { + "acc": 0.64552422, + "epoch": 0.25443937087772706, + "grad_norm": 5.03125, + "learning_rate": 9.837861064309899e-06, + "loss": 1.69221878, + "memory(GiB)": 97.17, + "step": 10030, + "train_speed(iter/s)": 0.448964 + }, + { + "acc": 0.64279327, + "epoch": 0.2545662100456621, + "grad_norm": 5.125, + "learning_rate": 9.837596081277173e-06, + "loss": 1.63055134, + "memory(GiB)": 97.17, + "step": 10035, + "train_speed(iter/s)": 0.449081 + }, + { + "acc": 0.64182806, + "epoch": 0.25469304921359714, + "grad_norm": 5.5, + "learning_rate": 9.837330885465373e-06, + "loss": 1.66612244, + "memory(GiB)": 97.17, + "step": 10040, + "train_speed(iter/s)": 0.449198 + }, + { + "acc": 0.6442668, + "epoch": 0.25481988838153224, + "grad_norm": 4.875, + "learning_rate": 9.837065476886163e-06, + "loss": 1.70503654, + "memory(GiB)": 97.17, + "step": 10045, + "train_speed(iter/s)": 0.449315 + }, + { + "acc": 0.64067979, + "epoch": 0.2549467275494673, + "grad_norm": 5.5, + "learning_rate": 9.83679985555122e-06, + "loss": 1.64878254, + "memory(GiB)": 97.17, + "step": 10050, + "train_speed(iter/s)": 0.449431 + }, + { + "acc": 0.63483868, + "epoch": 0.25507356671740233, + "grad_norm": 5.5, + "learning_rate": 9.836534021472222e-06, + "loss": 1.68517342, + "memory(GiB)": 97.17, + "step": 10055, + "train_speed(iter/s)": 0.449548 + }, + { + "acc": 0.64606409, + "epoch": 0.2552004058853374, + "grad_norm": 5.3125, + "learning_rate": 9.836267974660866e-06, + "loss": 1.66882305, + "memory(GiB)": 97.17, + "step": 10060, + "train_speed(iter/s)": 0.449663 + }, + { + "acc": 0.64150124, + "epoch": 0.2553272450532725, + "grad_norm": 5.125, + "learning_rate": 9.836001715128851e-06, + "loss": 1.7048481, + "memory(GiB)": 97.17, + "step": 10065, + "train_speed(iter/s)": 0.449778 + }, + { + "acc": 0.6441721, + "epoch": 0.2554540842212075, + "grad_norm": 5.65625, + "learning_rate": 9.835735242887889e-06, + "loss": 1.62405491, + "memory(GiB)": 97.17, + "step": 10070, + "train_speed(iter/s)": 0.449895 + }, + { + "acc": 0.64245281, + "epoch": 0.25558092338914257, + "grad_norm": 4.84375, + "learning_rate": 9.835468557949701e-06, + "loss": 1.63910599, + "memory(GiB)": 97.17, + "step": 10075, + "train_speed(iter/s)": 0.450011 + }, + { + "acc": 0.64228315, + "epoch": 0.2557077625570776, + "grad_norm": 4.8125, + "learning_rate": 9.83520166032602e-06, + "loss": 1.66814384, + "memory(GiB)": 97.17, + "step": 10080, + "train_speed(iter/s)": 0.450127 + }, + { + "acc": 0.63693943, + "epoch": 0.2558346017250127, + "grad_norm": 5.71875, + "learning_rate": 9.834934550028579e-06, + "loss": 1.65157509, + "memory(GiB)": 97.17, + "step": 10085, + "train_speed(iter/s)": 0.450243 + }, + { + "acc": 0.63770728, + "epoch": 0.25596144089294776, + "grad_norm": 5.53125, + "learning_rate": 9.83466722706913e-06, + "loss": 1.69734192, + "memory(GiB)": 97.17, + "step": 10090, + "train_speed(iter/s)": 0.450359 + }, + { + "acc": 0.64027219, + "epoch": 0.2560882800608828, + "grad_norm": 5.125, + "learning_rate": 9.834399691459433e-06, + "loss": 1.68638973, + "memory(GiB)": 97.17, + "step": 10095, + "train_speed(iter/s)": 0.450474 + }, + { + "acc": 0.63692484, + "epoch": 0.25621511922881784, + "grad_norm": 5.25, + "learning_rate": 9.83413194321125e-06, + "loss": 1.64527855, + "memory(GiB)": 97.17, + "step": 10100, + "train_speed(iter/s)": 0.450589 + }, + { + "epoch": 0.25621511922881784, + "eval_acc": 0.6363252017711662, + "eval_loss": 1.64874267578125, + "eval_runtime": 113.6142, + "eval_samples_per_second": 56.067, + "eval_steps_per_second": 28.033, + "step": 10100 + }, + { + "acc": 0.65384083, + "epoch": 0.25634195839675294, + "grad_norm": 5.625, + "learning_rate": 9.833863982336365e-06, + "loss": 1.60704098, + "memory(GiB)": 97.17, + "step": 10105, + "train_speed(iter/s)": 0.448344 + }, + { + "acc": 0.65403714, + "epoch": 0.256468797564688, + "grad_norm": 4.625, + "learning_rate": 9.83359580884656e-06, + "loss": 1.6452343, + "memory(GiB)": 97.17, + "step": 10110, + "train_speed(iter/s)": 0.44846 + }, + { + "acc": 0.65271111, + "epoch": 0.25659563673262303, + "grad_norm": 5.1875, + "learning_rate": 9.83332742275363e-06, + "loss": 1.66658325, + "memory(GiB)": 97.17, + "step": 10115, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.62488079, + "epoch": 0.2567224759005581, + "grad_norm": 6.03125, + "learning_rate": 9.833058824069382e-06, + "loss": 1.70323849, + "memory(GiB)": 97.17, + "step": 10120, + "train_speed(iter/s)": 0.448693 + }, + { + "acc": 0.65034437, + "epoch": 0.2568493150684932, + "grad_norm": 6.0625, + "learning_rate": 9.832790012805626e-06, + "loss": 1.65741673, + "memory(GiB)": 97.17, + "step": 10125, + "train_speed(iter/s)": 0.44881 + }, + { + "acc": 0.6508049, + "epoch": 0.2569761542364282, + "grad_norm": 5.8125, + "learning_rate": 9.832520988974191e-06, + "loss": 1.60273743, + "memory(GiB)": 97.17, + "step": 10130, + "train_speed(iter/s)": 0.448928 + }, + { + "acc": 0.65117588, + "epoch": 0.25710299340436327, + "grad_norm": 5.03125, + "learning_rate": 9.832251752586907e-06, + "loss": 1.58609066, + "memory(GiB)": 97.17, + "step": 10135, + "train_speed(iter/s)": 0.449044 + }, + { + "acc": 0.64600563, + "epoch": 0.2572298325722983, + "grad_norm": 4.625, + "learning_rate": 9.831982303655617e-06, + "loss": 1.62340183, + "memory(GiB)": 97.17, + "step": 10140, + "train_speed(iter/s)": 0.449161 + }, + { + "acc": 0.64842343, + "epoch": 0.2573566717402334, + "grad_norm": 5.3125, + "learning_rate": 9.83171264219217e-06, + "loss": 1.63590279, + "memory(GiB)": 97.17, + "step": 10145, + "train_speed(iter/s)": 0.449276 + }, + { + "acc": 0.67493429, + "epoch": 0.25748351090816846, + "grad_norm": 5.25, + "learning_rate": 9.831442768208429e-06, + "loss": 1.47832823, + "memory(GiB)": 97.17, + "step": 10150, + "train_speed(iter/s)": 0.449392 + }, + { + "acc": 0.66122627, + "epoch": 0.2576103500761035, + "grad_norm": 4.65625, + "learning_rate": 9.831172681716265e-06, + "loss": 1.63144588, + "memory(GiB)": 97.17, + "step": 10155, + "train_speed(iter/s)": 0.449507 + }, + { + "acc": 0.65866194, + "epoch": 0.25773718924403854, + "grad_norm": 4.96875, + "learning_rate": 9.830902382727556e-06, + "loss": 1.63804474, + "memory(GiB)": 97.17, + "step": 10160, + "train_speed(iter/s)": 0.449623 + }, + { + "acc": 0.65340271, + "epoch": 0.25786402841197364, + "grad_norm": 4.84375, + "learning_rate": 9.830631871254193e-06, + "loss": 1.60548134, + "memory(GiB)": 97.17, + "step": 10165, + "train_speed(iter/s)": 0.449739 + }, + { + "acc": 0.64797201, + "epoch": 0.2579908675799087, + "grad_norm": 5.09375, + "learning_rate": 9.830361147308074e-06, + "loss": 1.64140205, + "memory(GiB)": 97.17, + "step": 10170, + "train_speed(iter/s)": 0.449855 + }, + { + "acc": 0.63142986, + "epoch": 0.25811770674784373, + "grad_norm": 4.53125, + "learning_rate": 9.830090210901104e-06, + "loss": 1.66541443, + "memory(GiB)": 97.17, + "step": 10175, + "train_speed(iter/s)": 0.449971 + }, + { + "acc": 0.63583708, + "epoch": 0.2582445459157788, + "grad_norm": 5.90625, + "learning_rate": 9.829819062045203e-06, + "loss": 1.68908386, + "memory(GiB)": 97.17, + "step": 10180, + "train_speed(iter/s)": 0.450086 + }, + { + "acc": 0.66478491, + "epoch": 0.2583713850837139, + "grad_norm": 5.15625, + "learning_rate": 9.829547700752295e-06, + "loss": 1.5695116, + "memory(GiB)": 97.17, + "step": 10185, + "train_speed(iter/s)": 0.450203 + }, + { + "acc": 0.62844906, + "epoch": 0.2584982242516489, + "grad_norm": 5.1875, + "learning_rate": 9.829276127034315e-06, + "loss": 1.73666248, + "memory(GiB)": 97.17, + "step": 10190, + "train_speed(iter/s)": 0.450319 + }, + { + "acc": 0.62936883, + "epoch": 0.25862506341958397, + "grad_norm": 5.625, + "learning_rate": 9.829004340903214e-06, + "loss": 1.70294666, + "memory(GiB)": 97.17, + "step": 10195, + "train_speed(iter/s)": 0.450435 + }, + { + "acc": 0.63833456, + "epoch": 0.258751902587519, + "grad_norm": 4.375, + "learning_rate": 9.82873234237094e-06, + "loss": 1.70547714, + "memory(GiB)": 97.17, + "step": 10200, + "train_speed(iter/s)": 0.450552 + }, + { + "epoch": 0.258751902587519, + "eval_acc": 0.6364868777199263, + "eval_loss": 1.648376226425171, + "eval_runtime": 114.1916, + "eval_samples_per_second": 55.783, + "eval_steps_per_second": 27.892, + "step": 10200 + }, + { + "acc": 0.63217478, + "epoch": 0.2588787417554541, + "grad_norm": 6.28125, + "learning_rate": 9.828460131449457e-06, + "loss": 1.71564789, + "memory(GiB)": 97.17, + "step": 10205, + "train_speed(iter/s)": 0.448319 + }, + { + "acc": 0.63032894, + "epoch": 0.25900558092338916, + "grad_norm": 5.28125, + "learning_rate": 9.828187708150743e-06, + "loss": 1.70845108, + "memory(GiB)": 97.17, + "step": 10210, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.6530694, + "epoch": 0.2591324200913242, + "grad_norm": 5.21875, + "learning_rate": 9.827915072486776e-06, + "loss": 1.62753773, + "memory(GiB)": 97.17, + "step": 10215, + "train_speed(iter/s)": 0.448547 + }, + { + "acc": 0.65005016, + "epoch": 0.25925925925925924, + "grad_norm": 5.75, + "learning_rate": 9.827642224469547e-06, + "loss": 1.6148571, + "memory(GiB)": 111.15, + "step": 10220, + "train_speed(iter/s)": 0.448663 + }, + { + "acc": 0.64397888, + "epoch": 0.25938609842719434, + "grad_norm": 6.25, + "learning_rate": 9.827369164111062e-06, + "loss": 1.71176014, + "memory(GiB)": 111.15, + "step": 10225, + "train_speed(iter/s)": 0.448779 + }, + { + "acc": 0.63397889, + "epoch": 0.2595129375951294, + "grad_norm": 6.03125, + "learning_rate": 9.827095891423328e-06, + "loss": 1.77851276, + "memory(GiB)": 111.15, + "step": 10230, + "train_speed(iter/s)": 0.448895 + }, + { + "acc": 0.65936146, + "epoch": 0.25963977676306443, + "grad_norm": 6.03125, + "learning_rate": 9.826822406418366e-06, + "loss": 1.55915585, + "memory(GiB)": 111.15, + "step": 10235, + "train_speed(iter/s)": 0.449012 + }, + { + "acc": 0.64927692, + "epoch": 0.2597666159309995, + "grad_norm": 4.75, + "learning_rate": 9.826548709108202e-06, + "loss": 1.56409826, + "memory(GiB)": 111.15, + "step": 10240, + "train_speed(iter/s)": 0.449128 + }, + { + "acc": 0.63607216, + "epoch": 0.2598934550989346, + "grad_norm": 5.625, + "learning_rate": 9.826274799504878e-06, + "loss": 1.63129997, + "memory(GiB)": 111.15, + "step": 10245, + "train_speed(iter/s)": 0.449244 + }, + { + "acc": 0.64446115, + "epoch": 0.2600202942668696, + "grad_norm": 4.71875, + "learning_rate": 9.82600067762044e-06, + "loss": 1.65774345, + "memory(GiB)": 111.15, + "step": 10250, + "train_speed(iter/s)": 0.449361 + }, + { + "acc": 0.65417261, + "epoch": 0.26014713343480467, + "grad_norm": 5.46875, + "learning_rate": 9.825726343466947e-06, + "loss": 1.63800316, + "memory(GiB)": 111.15, + "step": 10255, + "train_speed(iter/s)": 0.449477 + }, + { + "acc": 0.64329796, + "epoch": 0.2602739726027397, + "grad_norm": 5.4375, + "learning_rate": 9.825451797056462e-06, + "loss": 1.68297005, + "memory(GiB)": 111.15, + "step": 10260, + "train_speed(iter/s)": 0.449591 + }, + { + "acc": 0.63969059, + "epoch": 0.2604008117706748, + "grad_norm": 6.09375, + "learning_rate": 9.825177038401064e-06, + "loss": 1.64901123, + "memory(GiB)": 111.15, + "step": 10265, + "train_speed(iter/s)": 0.449706 + }, + { + "acc": 0.63883801, + "epoch": 0.26052765093860986, + "grad_norm": 4.90625, + "learning_rate": 9.824902067512838e-06, + "loss": 1.70557919, + "memory(GiB)": 111.15, + "step": 10270, + "train_speed(iter/s)": 0.449823 + }, + { + "acc": 0.65232162, + "epoch": 0.2606544901065449, + "grad_norm": 5.0, + "learning_rate": 9.824626884403877e-06, + "loss": 1.64757271, + "memory(GiB)": 111.15, + "step": 10275, + "train_speed(iter/s)": 0.449939 + }, + { + "acc": 0.64612293, + "epoch": 0.26078132927447994, + "grad_norm": 4.78125, + "learning_rate": 9.824351489086283e-06, + "loss": 1.63203411, + "memory(GiB)": 111.15, + "step": 10280, + "train_speed(iter/s)": 0.450054 + }, + { + "acc": 0.65411334, + "epoch": 0.26090816844241504, + "grad_norm": 6.15625, + "learning_rate": 9.824075881572176e-06, + "loss": 1.63782921, + "memory(GiB)": 111.15, + "step": 10285, + "train_speed(iter/s)": 0.450171 + }, + { + "acc": 0.65749311, + "epoch": 0.2610350076103501, + "grad_norm": 5.59375, + "learning_rate": 9.823800061873669e-06, + "loss": 1.65367374, + "memory(GiB)": 111.15, + "step": 10290, + "train_speed(iter/s)": 0.450287 + }, + { + "acc": 0.64824762, + "epoch": 0.26116184677828513, + "grad_norm": 6.65625, + "learning_rate": 9.8235240300029e-06, + "loss": 1.70516548, + "memory(GiB)": 111.15, + "step": 10295, + "train_speed(iter/s)": 0.450403 + }, + { + "acc": 0.6479887, + "epoch": 0.2612886859462202, + "grad_norm": 5.0, + "learning_rate": 9.82324778597201e-06, + "loss": 1.62311974, + "memory(GiB)": 111.15, + "step": 10300, + "train_speed(iter/s)": 0.450519 + }, + { + "epoch": 0.2612886859462202, + "eval_acc": 0.6365065127834708, + "eval_loss": 1.6477584838867188, + "eval_runtime": 112.8502, + "eval_samples_per_second": 56.447, + "eval_steps_per_second": 28.223, + "step": 10300 + }, + { + "acc": 0.64832945, + "epoch": 0.2614155251141553, + "grad_norm": 5.1875, + "learning_rate": 9.822971329793147e-06, + "loss": 1.65533886, + "memory(GiB)": 111.15, + "step": 10305, + "train_speed(iter/s)": 0.448334 + }, + { + "acc": 0.65736933, + "epoch": 0.2615423642820903, + "grad_norm": 4.96875, + "learning_rate": 9.822694661478471e-06, + "loss": 1.60440922, + "memory(GiB)": 111.15, + "step": 10310, + "train_speed(iter/s)": 0.448448 + }, + { + "acc": 0.63942556, + "epoch": 0.26166920345002537, + "grad_norm": 5.6875, + "learning_rate": 9.822417781040154e-06, + "loss": 1.66899605, + "memory(GiB)": 111.15, + "step": 10315, + "train_speed(iter/s)": 0.448562 + }, + { + "acc": 0.63693323, + "epoch": 0.2617960426179604, + "grad_norm": 4.6875, + "learning_rate": 9.822140688490372e-06, + "loss": 1.66239643, + "memory(GiB)": 111.15, + "step": 10320, + "train_speed(iter/s)": 0.448676 + }, + { + "acc": 0.6457778, + "epoch": 0.2619228817858955, + "grad_norm": 5.40625, + "learning_rate": 9.821863383841312e-06, + "loss": 1.6458786, + "memory(GiB)": 111.15, + "step": 10325, + "train_speed(iter/s)": 0.44879 + }, + { + "acc": 0.66147504, + "epoch": 0.26204972095383056, + "grad_norm": 5.3125, + "learning_rate": 9.821585867105173e-06, + "loss": 1.67228203, + "memory(GiB)": 111.15, + "step": 10330, + "train_speed(iter/s)": 0.448904 + }, + { + "acc": 0.63241463, + "epoch": 0.2621765601217656, + "grad_norm": 4.96875, + "learning_rate": 9.821308138294162e-06, + "loss": 1.72343426, + "memory(GiB)": 111.15, + "step": 10335, + "train_speed(iter/s)": 0.449018 + }, + { + "acc": 0.63823109, + "epoch": 0.26230339928970065, + "grad_norm": 5.59375, + "learning_rate": 9.821030197420492e-06, + "loss": 1.68619862, + "memory(GiB)": 111.15, + "step": 10340, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.64724827, + "epoch": 0.26243023845763574, + "grad_norm": 5.34375, + "learning_rate": 9.820752044496389e-06, + "loss": 1.64986401, + "memory(GiB)": 111.15, + "step": 10345, + "train_speed(iter/s)": 0.449245 + }, + { + "acc": 0.6319047, + "epoch": 0.2625570776255708, + "grad_norm": 5.0625, + "learning_rate": 9.82047367953409e-06, + "loss": 1.65586586, + "memory(GiB)": 111.15, + "step": 10350, + "train_speed(iter/s)": 0.449357 + }, + { + "acc": 0.65941129, + "epoch": 0.26268391679350583, + "grad_norm": 6.0, + "learning_rate": 9.820195102545835e-06, + "loss": 1.61650639, + "memory(GiB)": 111.15, + "step": 10355, + "train_speed(iter/s)": 0.449472 + }, + { + "acc": 0.62752819, + "epoch": 0.2628107559614409, + "grad_norm": 4.6875, + "learning_rate": 9.81991631354388e-06, + "loss": 1.64996834, + "memory(GiB)": 111.15, + "step": 10360, + "train_speed(iter/s)": 0.449588 + }, + { + "acc": 0.63833857, + "epoch": 0.262937595129376, + "grad_norm": 5.59375, + "learning_rate": 9.819637312540485e-06, + "loss": 1.7260973, + "memory(GiB)": 111.15, + "step": 10365, + "train_speed(iter/s)": 0.449701 + }, + { + "acc": 0.64258757, + "epoch": 0.263064434297311, + "grad_norm": 4.6875, + "learning_rate": 9.819358099547923e-06, + "loss": 1.60989456, + "memory(GiB)": 111.15, + "step": 10370, + "train_speed(iter/s)": 0.449815 + }, + { + "acc": 0.62938113, + "epoch": 0.26319127346524607, + "grad_norm": 5.25, + "learning_rate": 9.819078674578474e-06, + "loss": 1.75894127, + "memory(GiB)": 111.15, + "step": 10375, + "train_speed(iter/s)": 0.44993 + }, + { + "acc": 0.62849383, + "epoch": 0.2633181126331811, + "grad_norm": 5.25, + "learning_rate": 9.818799037644432e-06, + "loss": 1.72301483, + "memory(GiB)": 111.15, + "step": 10380, + "train_speed(iter/s)": 0.450045 + }, + { + "acc": 0.64598904, + "epoch": 0.2634449518011162, + "grad_norm": 4.5, + "learning_rate": 9.818519188758092e-06, + "loss": 1.67254715, + "memory(GiB)": 111.15, + "step": 10385, + "train_speed(iter/s)": 0.450159 + }, + { + "acc": 0.64492431, + "epoch": 0.26357179096905126, + "grad_norm": 5.25, + "learning_rate": 9.818239127931765e-06, + "loss": 1.65526237, + "memory(GiB)": 111.15, + "step": 10390, + "train_speed(iter/s)": 0.450272 + }, + { + "acc": 0.63072276, + "epoch": 0.2636986301369863, + "grad_norm": 5.09375, + "learning_rate": 9.81795885517777e-06, + "loss": 1.74420052, + "memory(GiB)": 111.15, + "step": 10395, + "train_speed(iter/s)": 0.450386 + }, + { + "acc": 0.64704728, + "epoch": 0.26382546930492135, + "grad_norm": 4.53125, + "learning_rate": 9.817678370508434e-06, + "loss": 1.69675446, + "memory(GiB)": 111.15, + "step": 10400, + "train_speed(iter/s)": 0.450499 + }, + { + "epoch": 0.26382546930492135, + "eval_acc": 0.6365796220626259, + "eval_loss": 1.6467974185943604, + "eval_runtime": 113.8276, + "eval_samples_per_second": 55.962, + "eval_steps_per_second": 27.981, + "step": 10400 + }, + { + "acc": 0.63342175, + "epoch": 0.26395230847285645, + "grad_norm": 5.59375, + "learning_rate": 9.817397673936093e-06, + "loss": 1.73191185, + "memory(GiB)": 111.15, + "step": 10405, + "train_speed(iter/s)": 0.448316 + }, + { + "acc": 0.63455944, + "epoch": 0.2640791476407915, + "grad_norm": 5.625, + "learning_rate": 9.817116765473095e-06, + "loss": 1.71335678, + "memory(GiB)": 111.15, + "step": 10410, + "train_speed(iter/s)": 0.448429 + }, + { + "acc": 0.6439085, + "epoch": 0.26420598680872653, + "grad_norm": 5.15625, + "learning_rate": 9.816835645131795e-06, + "loss": 1.6211647, + "memory(GiB)": 111.15, + "step": 10415, + "train_speed(iter/s)": 0.448543 + }, + { + "acc": 0.64781103, + "epoch": 0.2643328259766616, + "grad_norm": 5.5, + "learning_rate": 9.816554312924555e-06, + "loss": 1.62224064, + "memory(GiB)": 111.15, + "step": 10420, + "train_speed(iter/s)": 0.448656 + }, + { + "acc": 0.66037989, + "epoch": 0.2644596651445967, + "grad_norm": 5.25, + "learning_rate": 9.816272768863756e-06, + "loss": 1.61004105, + "memory(GiB)": 111.15, + "step": 10425, + "train_speed(iter/s)": 0.44877 + }, + { + "acc": 0.65064759, + "epoch": 0.2645865043125317, + "grad_norm": 5.28125, + "learning_rate": 9.815991012961773e-06, + "loss": 1.63398972, + "memory(GiB)": 111.15, + "step": 10430, + "train_speed(iter/s)": 0.448883 + }, + { + "acc": 0.64851537, + "epoch": 0.26471334348046677, + "grad_norm": 4.9375, + "learning_rate": 9.815709045231008e-06, + "loss": 1.71657066, + "memory(GiB)": 111.15, + "step": 10435, + "train_speed(iter/s)": 0.448995 + }, + { + "acc": 0.63785, + "epoch": 0.2648401826484018, + "grad_norm": 4.625, + "learning_rate": 9.815426865683858e-06, + "loss": 1.76409569, + "memory(GiB)": 111.15, + "step": 10440, + "train_speed(iter/s)": 0.449109 + }, + { + "acc": 0.6514431, + "epoch": 0.2649670218163369, + "grad_norm": 5.125, + "learning_rate": 9.815144474332732e-06, + "loss": 1.61515751, + "memory(GiB)": 111.15, + "step": 10445, + "train_speed(iter/s)": 0.449223 + }, + { + "acc": 0.65421677, + "epoch": 0.26509386098427196, + "grad_norm": 4.90625, + "learning_rate": 9.814861871190056e-06, + "loss": 1.618647, + "memory(GiB)": 111.15, + "step": 10450, + "train_speed(iter/s)": 0.449336 + }, + { + "acc": 0.64979601, + "epoch": 0.265220700152207, + "grad_norm": 5.59375, + "learning_rate": 9.814579056268256e-06, + "loss": 1.58884907, + "memory(GiB)": 111.15, + "step": 10455, + "train_speed(iter/s)": 0.44945 + }, + { + "acc": 0.65148792, + "epoch": 0.26534753932014205, + "grad_norm": 5.1875, + "learning_rate": 9.814296029579776e-06, + "loss": 1.62890167, + "memory(GiB)": 111.15, + "step": 10460, + "train_speed(iter/s)": 0.449564 + }, + { + "acc": 0.62713518, + "epoch": 0.26547437848807715, + "grad_norm": 5.21875, + "learning_rate": 9.814012791137063e-06, + "loss": 1.76987038, + "memory(GiB)": 111.15, + "step": 10465, + "train_speed(iter/s)": 0.449678 + }, + { + "acc": 0.63424377, + "epoch": 0.2656012176560122, + "grad_norm": 5.40625, + "learning_rate": 9.81372934095257e-06, + "loss": 1.6876936, + "memory(GiB)": 111.15, + "step": 10470, + "train_speed(iter/s)": 0.44979 + }, + { + "acc": 0.63029704, + "epoch": 0.26572805682394723, + "grad_norm": 5.5, + "learning_rate": 9.813445679038773e-06, + "loss": 1.73932724, + "memory(GiB)": 111.15, + "step": 10475, + "train_speed(iter/s)": 0.449904 + }, + { + "acc": 0.63514605, + "epoch": 0.2658548959918823, + "grad_norm": 6.1875, + "learning_rate": 9.813161805408145e-06, + "loss": 1.67673512, + "memory(GiB)": 111.15, + "step": 10480, + "train_speed(iter/s)": 0.450018 + }, + { + "acc": 0.65030417, + "epoch": 0.2659817351598174, + "grad_norm": 5.28125, + "learning_rate": 9.812877720073169e-06, + "loss": 1.67042656, + "memory(GiB)": 111.15, + "step": 10485, + "train_speed(iter/s)": 0.450131 + }, + { + "acc": 0.64088945, + "epoch": 0.2661085743277524, + "grad_norm": 6.0625, + "learning_rate": 9.812593423046344e-06, + "loss": 1.71995163, + "memory(GiB)": 111.15, + "step": 10490, + "train_speed(iter/s)": 0.450244 + }, + { + "acc": 0.63417873, + "epoch": 0.26623541349568747, + "grad_norm": 4.65625, + "learning_rate": 9.812308914340174e-06, + "loss": 1.7628397, + "memory(GiB)": 111.15, + "step": 10495, + "train_speed(iter/s)": 0.450357 + }, + { + "acc": 0.63861637, + "epoch": 0.2663622526636225, + "grad_norm": 5.90625, + "learning_rate": 9.812024193967171e-06, + "loss": 1.7220993, + "memory(GiB)": 111.15, + "step": 10500, + "train_speed(iter/s)": 0.45047 + }, + { + "epoch": 0.2663622526636225, + "eval_acc": 0.6365754443895313, + "eval_loss": 1.6466397047042847, + "eval_runtime": 113.4611, + "eval_samples_per_second": 56.143, + "eval_steps_per_second": 28.071, + "step": 10500 + }, + { + "acc": 0.64456587, + "epoch": 0.2664890918315576, + "grad_norm": 5.4375, + "learning_rate": 9.811739261939861e-06, + "loss": 1.69936275, + "memory(GiB)": 111.15, + "step": 10505, + "train_speed(iter/s)": 0.448315 + }, + { + "acc": 0.64456005, + "epoch": 0.26661593099949266, + "grad_norm": 4.40625, + "learning_rate": 9.811454118270775e-06, + "loss": 1.68555069, + "memory(GiB)": 111.15, + "step": 10510, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.63295393, + "epoch": 0.2667427701674277, + "grad_norm": 5.0625, + "learning_rate": 9.811168762972457e-06, + "loss": 1.68106022, + "memory(GiB)": 111.15, + "step": 10515, + "train_speed(iter/s)": 0.448539 + }, + { + "acc": 0.64319992, + "epoch": 0.26686960933536275, + "grad_norm": 5.34375, + "learning_rate": 9.810883196057454e-06, + "loss": 1.68998985, + "memory(GiB)": 111.15, + "step": 10520, + "train_speed(iter/s)": 0.448651 + }, + { + "acc": 0.64584904, + "epoch": 0.26699644850329785, + "grad_norm": 5.28125, + "learning_rate": 9.81059741753833e-06, + "loss": 1.62842865, + "memory(GiB)": 111.15, + "step": 10525, + "train_speed(iter/s)": 0.448762 + }, + { + "acc": 0.64315882, + "epoch": 0.2671232876712329, + "grad_norm": 5.375, + "learning_rate": 9.810311427427653e-06, + "loss": 1.66778412, + "memory(GiB)": 111.15, + "step": 10530, + "train_speed(iter/s)": 0.448874 + }, + { + "acc": 0.65436134, + "epoch": 0.26725012683916793, + "grad_norm": 4.8125, + "learning_rate": 9.810025225738005e-06, + "loss": 1.66106319, + "memory(GiB)": 111.15, + "step": 10535, + "train_speed(iter/s)": 0.448987 + }, + { + "acc": 0.61416044, + "epoch": 0.267376966007103, + "grad_norm": 6.75, + "learning_rate": 9.809738812481971e-06, + "loss": 1.79050331, + "memory(GiB)": 111.15, + "step": 10540, + "train_speed(iter/s)": 0.449099 + }, + { + "acc": 0.64566631, + "epoch": 0.2675038051750381, + "grad_norm": 5.25, + "learning_rate": 9.809452187672149e-06, + "loss": 1.64396667, + "memory(GiB)": 111.15, + "step": 10545, + "train_speed(iter/s)": 0.449211 + }, + { + "acc": 0.63508387, + "epoch": 0.2676306443429731, + "grad_norm": 4.78125, + "learning_rate": 9.809165351321149e-06, + "loss": 1.68749962, + "memory(GiB)": 111.15, + "step": 10550, + "train_speed(iter/s)": 0.449324 + }, + { + "acc": 0.66122913, + "epoch": 0.26775748351090817, + "grad_norm": 5.375, + "learning_rate": 9.808878303441585e-06, + "loss": 1.65110321, + "memory(GiB)": 111.15, + "step": 10555, + "train_speed(iter/s)": 0.449436 + }, + { + "acc": 0.64469032, + "epoch": 0.2678843226788432, + "grad_norm": 6.84375, + "learning_rate": 9.808591044046083e-06, + "loss": 1.69892731, + "memory(GiB)": 111.15, + "step": 10560, + "train_speed(iter/s)": 0.449548 + }, + { + "acc": 0.64150982, + "epoch": 0.2680111618467783, + "grad_norm": 5.65625, + "learning_rate": 9.808303573147277e-06, + "loss": 1.70587845, + "memory(GiB)": 111.15, + "step": 10565, + "train_speed(iter/s)": 0.44966 + }, + { + "acc": 0.64790754, + "epoch": 0.26813800101471336, + "grad_norm": 6.25, + "learning_rate": 9.808015890757812e-06, + "loss": 1.6574131, + "memory(GiB)": 111.15, + "step": 10570, + "train_speed(iter/s)": 0.449773 + }, + { + "acc": 0.65174494, + "epoch": 0.2682648401826484, + "grad_norm": 4.9375, + "learning_rate": 9.807727996890343e-06, + "loss": 1.61414452, + "memory(GiB)": 111.15, + "step": 10575, + "train_speed(iter/s)": 0.449887 + }, + { + "acc": 0.64311385, + "epoch": 0.26839167935058345, + "grad_norm": 5.53125, + "learning_rate": 9.807439891557533e-06, + "loss": 1.66213684, + "memory(GiB)": 111.15, + "step": 10580, + "train_speed(iter/s)": 0.449999 + }, + { + "acc": 0.63658705, + "epoch": 0.26851851851851855, + "grad_norm": 5.78125, + "learning_rate": 9.80715157477205e-06, + "loss": 1.72441101, + "memory(GiB)": 111.15, + "step": 10585, + "train_speed(iter/s)": 0.450112 + }, + { + "acc": 0.63954258, + "epoch": 0.2686453576864536, + "grad_norm": 5.03125, + "learning_rate": 9.806863046546581e-06, + "loss": 1.71004639, + "memory(GiB)": 111.15, + "step": 10590, + "train_speed(iter/s)": 0.450223 + }, + { + "acc": 0.64068861, + "epoch": 0.26877219685438863, + "grad_norm": 4.34375, + "learning_rate": 9.806574306893814e-06, + "loss": 1.69977074, + "memory(GiB)": 111.15, + "step": 10595, + "train_speed(iter/s)": 0.450336 + }, + { + "acc": 0.65321507, + "epoch": 0.2688990360223237, + "grad_norm": 5.65625, + "learning_rate": 9.806285355826447e-06, + "loss": 1.60649109, + "memory(GiB)": 111.15, + "step": 10600, + "train_speed(iter/s)": 0.450449 + }, + { + "epoch": 0.2688990360223237, + "eval_acc": 0.6369589547796132, + "eval_loss": 1.6462167501449585, + "eval_runtime": 113.7921, + "eval_samples_per_second": 55.979, + "eval_steps_per_second": 27.99, + "step": 10600 + }, + { + "acc": 0.66060572, + "epoch": 0.2690258751902588, + "grad_norm": 5.71875, + "learning_rate": 9.805996193357194e-06, + "loss": 1.60208435, + "memory(GiB)": 111.15, + "step": 10605, + "train_speed(iter/s)": 0.448307 + }, + { + "acc": 0.62219057, + "epoch": 0.2691527143581938, + "grad_norm": 5.15625, + "learning_rate": 9.80570681949877e-06, + "loss": 1.6789711, + "memory(GiB)": 111.15, + "step": 10610, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.63864217, + "epoch": 0.26927955352612887, + "grad_norm": 4.78125, + "learning_rate": 9.805417234263905e-06, + "loss": 1.68765221, + "memory(GiB)": 111.15, + "step": 10615, + "train_speed(iter/s)": 0.448531 + }, + { + "acc": 0.63425088, + "epoch": 0.2694063926940639, + "grad_norm": 4.78125, + "learning_rate": 9.805127437665333e-06, + "loss": 1.74543762, + "memory(GiB)": 111.15, + "step": 10620, + "train_speed(iter/s)": 0.448643 + }, + { + "acc": 0.6350512, + "epoch": 0.269533231861999, + "grad_norm": 4.65625, + "learning_rate": 9.804837429715805e-06, + "loss": 1.73776474, + "memory(GiB)": 111.15, + "step": 10625, + "train_speed(iter/s)": 0.448754 + }, + { + "acc": 0.65661583, + "epoch": 0.26966007102993406, + "grad_norm": 4.71875, + "learning_rate": 9.804547210428074e-06, + "loss": 1.64035664, + "memory(GiB)": 111.15, + "step": 10630, + "train_speed(iter/s)": 0.448867 + }, + { + "acc": 0.6331749, + "epoch": 0.2697869101978691, + "grad_norm": 5.5, + "learning_rate": 9.804256779814906e-06, + "loss": 1.70343304, + "memory(GiB)": 111.15, + "step": 10635, + "train_speed(iter/s)": 0.448979 + }, + { + "acc": 0.64547281, + "epoch": 0.26991374936580415, + "grad_norm": 5.03125, + "learning_rate": 9.803966137889076e-06, + "loss": 1.58225842, + "memory(GiB)": 111.15, + "step": 10640, + "train_speed(iter/s)": 0.449091 + }, + { + "acc": 0.62487526, + "epoch": 0.27004058853373925, + "grad_norm": 5.5, + "learning_rate": 9.803675284663368e-06, + "loss": 1.75015697, + "memory(GiB)": 111.15, + "step": 10645, + "train_speed(iter/s)": 0.449204 + }, + { + "acc": 0.64729156, + "epoch": 0.2701674277016743, + "grad_norm": 5.75, + "learning_rate": 9.803384220150571e-06, + "loss": 1.7088501, + "memory(GiB)": 111.15, + "step": 10650, + "train_speed(iter/s)": 0.449316 + }, + { + "acc": 0.63795233, + "epoch": 0.27029426686960933, + "grad_norm": 6.0625, + "learning_rate": 9.803092944363493e-06, + "loss": 1.72100296, + "memory(GiB)": 111.15, + "step": 10655, + "train_speed(iter/s)": 0.449429 + }, + { + "acc": 0.66238565, + "epoch": 0.2704211060375444, + "grad_norm": 5.84375, + "learning_rate": 9.802801457314943e-06, + "loss": 1.64964027, + "memory(GiB)": 111.15, + "step": 10660, + "train_speed(iter/s)": 0.449542 + }, + { + "acc": 0.65660119, + "epoch": 0.2705479452054795, + "grad_norm": 7.0625, + "learning_rate": 9.802509759017741e-06, + "loss": 1.62792206, + "memory(GiB)": 111.15, + "step": 10665, + "train_speed(iter/s)": 0.449654 + }, + { + "acc": 0.63391347, + "epoch": 0.2706747843734145, + "grad_norm": 4.09375, + "learning_rate": 9.802217849484719e-06, + "loss": 1.72473183, + "memory(GiB)": 111.15, + "step": 10670, + "train_speed(iter/s)": 0.449766 + }, + { + "acc": 0.64480619, + "epoch": 0.27080162354134957, + "grad_norm": 5.65625, + "learning_rate": 9.801925728728715e-06, + "loss": 1.66173363, + "memory(GiB)": 111.15, + "step": 10675, + "train_speed(iter/s)": 0.449878 + }, + { + "acc": 0.64621096, + "epoch": 0.2709284627092846, + "grad_norm": 5.0, + "learning_rate": 9.801633396762577e-06, + "loss": 1.69071236, + "memory(GiB)": 111.15, + "step": 10680, + "train_speed(iter/s)": 0.449989 + }, + { + "acc": 0.64803853, + "epoch": 0.2710553018772197, + "grad_norm": 5.53125, + "learning_rate": 9.801340853599167e-06, + "loss": 1.71612778, + "memory(GiB)": 111.15, + "step": 10685, + "train_speed(iter/s)": 0.4501 + }, + { + "acc": 0.66075945, + "epoch": 0.27118214104515476, + "grad_norm": 6.96875, + "learning_rate": 9.801048099251348e-06, + "loss": 1.61173706, + "memory(GiB)": 111.15, + "step": 10690, + "train_speed(iter/s)": 0.450212 + }, + { + "acc": 0.63571215, + "epoch": 0.2713089802130898, + "grad_norm": 5.03125, + "learning_rate": 9.800755133731999e-06, + "loss": 1.6736496, + "memory(GiB)": 111.15, + "step": 10695, + "train_speed(iter/s)": 0.450324 + }, + { + "acc": 0.65049639, + "epoch": 0.27143581938102485, + "grad_norm": 5.1875, + "learning_rate": 9.800461957054006e-06, + "loss": 1.67568741, + "memory(GiB)": 111.15, + "step": 10700, + "train_speed(iter/s)": 0.450435 + }, + { + "epoch": 0.27143581938102485, + "eval_acc": 0.6369209379544525, + "eval_loss": 1.6455292701721191, + "eval_runtime": 113.4786, + "eval_samples_per_second": 56.134, + "eval_steps_per_second": 28.067, + "step": 10700 + }, + { + "acc": 0.64458227, + "epoch": 0.27156265854895995, + "grad_norm": 5.8125, + "learning_rate": 9.800168569230261e-06, + "loss": 1.62625351, + "memory(GiB)": 111.15, + "step": 10705, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.65337734, + "epoch": 0.271689497716895, + "grad_norm": 4.96875, + "learning_rate": 9.799874970273674e-06, + "loss": 1.62052536, + "memory(GiB)": 111.15, + "step": 10710, + "train_speed(iter/s)": 0.448431 + }, + { + "acc": 0.64829273, + "epoch": 0.27181633688483003, + "grad_norm": 4.78125, + "learning_rate": 9.799581160197156e-06, + "loss": 1.6696476, + "memory(GiB)": 111.15, + "step": 10715, + "train_speed(iter/s)": 0.448542 + }, + { + "acc": 0.65377245, + "epoch": 0.2719431760527651, + "grad_norm": 4.1875, + "learning_rate": 9.799287139013628e-06, + "loss": 1.61526585, + "memory(GiB)": 111.15, + "step": 10720, + "train_speed(iter/s)": 0.448652 + }, + { + "acc": 0.65760822, + "epoch": 0.2720700152207002, + "grad_norm": 5.25, + "learning_rate": 9.798992906736028e-06, + "loss": 1.57026005, + "memory(GiB)": 111.15, + "step": 10725, + "train_speed(iter/s)": 0.448763 + }, + { + "acc": 0.64212141, + "epoch": 0.2721968543886352, + "grad_norm": 5.09375, + "learning_rate": 9.79869846337729e-06, + "loss": 1.68452644, + "memory(GiB)": 111.15, + "step": 10730, + "train_speed(iter/s)": 0.448874 + }, + { + "acc": 0.63777227, + "epoch": 0.27232369355657027, + "grad_norm": 6.09375, + "learning_rate": 9.79840380895037e-06, + "loss": 1.68718719, + "memory(GiB)": 111.15, + "step": 10735, + "train_speed(iter/s)": 0.448983 + }, + { + "acc": 0.65433426, + "epoch": 0.2724505327245053, + "grad_norm": 4.90625, + "learning_rate": 9.798108943468228e-06, + "loss": 1.5699688, + "memory(GiB)": 111.15, + "step": 10740, + "train_speed(iter/s)": 0.449094 + }, + { + "acc": 0.63904018, + "epoch": 0.2725773718924404, + "grad_norm": 4.8125, + "learning_rate": 9.797813866943832e-06, + "loss": 1.62258835, + "memory(GiB)": 111.15, + "step": 10745, + "train_speed(iter/s)": 0.449205 + }, + { + "acc": 0.65388136, + "epoch": 0.27270421106037546, + "grad_norm": 5.53125, + "learning_rate": 9.797518579390162e-06, + "loss": 1.54560442, + "memory(GiB)": 111.15, + "step": 10750, + "train_speed(iter/s)": 0.449316 + }, + { + "acc": 0.65189171, + "epoch": 0.2728310502283105, + "grad_norm": 4.375, + "learning_rate": 9.797223080820204e-06, + "loss": 1.62971153, + "memory(GiB)": 111.15, + "step": 10755, + "train_speed(iter/s)": 0.449426 + }, + { + "acc": 0.63916759, + "epoch": 0.27295788939624555, + "grad_norm": 4.625, + "learning_rate": 9.796927371246958e-06, + "loss": 1.66935959, + "memory(GiB)": 111.15, + "step": 10760, + "train_speed(iter/s)": 0.449537 + }, + { + "acc": 0.6458045, + "epoch": 0.27308472856418065, + "grad_norm": 4.96875, + "learning_rate": 9.796631450683431e-06, + "loss": 1.64870815, + "memory(GiB)": 111.15, + "step": 10765, + "train_speed(iter/s)": 0.449647 + }, + { + "acc": 0.63077965, + "epoch": 0.2732115677321157, + "grad_norm": 5.1875, + "learning_rate": 9.796335319142637e-06, + "loss": 1.65166626, + "memory(GiB)": 111.15, + "step": 10770, + "train_speed(iter/s)": 0.449758 + }, + { + "acc": 0.64339128, + "epoch": 0.27333840690005073, + "grad_norm": 5.6875, + "learning_rate": 9.796038976637599e-06, + "loss": 1.66504822, + "memory(GiB)": 111.15, + "step": 10775, + "train_speed(iter/s)": 0.449869 + }, + { + "acc": 0.6482748, + "epoch": 0.2734652460679858, + "grad_norm": 5.21875, + "learning_rate": 9.795742423181355e-06, + "loss": 1.65122929, + "memory(GiB)": 111.15, + "step": 10780, + "train_speed(iter/s)": 0.44998 + }, + { + "acc": 0.63234205, + "epoch": 0.2735920852359209, + "grad_norm": 4.84375, + "learning_rate": 9.795445658786948e-06, + "loss": 1.76725044, + "memory(GiB)": 111.15, + "step": 10785, + "train_speed(iter/s)": 0.450091 + }, + { + "acc": 0.64653225, + "epoch": 0.2737189244038559, + "grad_norm": 5.0, + "learning_rate": 9.795148683467431e-06, + "loss": 1.66851044, + "memory(GiB)": 111.15, + "step": 10790, + "train_speed(iter/s)": 0.450201 + }, + { + "acc": 0.6354784, + "epoch": 0.27384576357179097, + "grad_norm": 5.28125, + "learning_rate": 9.794851497235866e-06, + "loss": 1.66212234, + "memory(GiB)": 111.15, + "step": 10795, + "train_speed(iter/s)": 0.450313 + }, + { + "acc": 0.6351965, + "epoch": 0.273972602739726, + "grad_norm": 5.65625, + "learning_rate": 9.794554100105325e-06, + "loss": 1.62322807, + "memory(GiB)": 111.15, + "step": 10800, + "train_speed(iter/s)": 0.450424 + }, + { + "epoch": 0.273972602739726, + "eval_acc": 0.637052952424241, + "eval_loss": 1.64460027217865, + "eval_runtime": 112.9875, + "eval_samples_per_second": 56.378, + "eval_steps_per_second": 28.189, + "step": 10800 + }, + { + "acc": 0.63959303, + "epoch": 0.2740994419076611, + "grad_norm": 4.6875, + "learning_rate": 9.794256492088888e-06, + "loss": 1.64184456, + "memory(GiB)": 111.15, + "step": 10805, + "train_speed(iter/s)": 0.448338 + }, + { + "acc": 0.64204593, + "epoch": 0.27422628107559616, + "grad_norm": 5.90625, + "learning_rate": 9.793958673199647e-06, + "loss": 1.69917564, + "memory(GiB)": 111.15, + "step": 10810, + "train_speed(iter/s)": 0.448448 + }, + { + "acc": 0.65081358, + "epoch": 0.2743531202435312, + "grad_norm": 4.46875, + "learning_rate": 9.793660643450697e-06, + "loss": 1.63717918, + "memory(GiB)": 111.15, + "step": 10815, + "train_speed(iter/s)": 0.448559 + }, + { + "acc": 0.6347259, + "epoch": 0.27447995941146625, + "grad_norm": 4.9375, + "learning_rate": 9.793362402855152e-06, + "loss": 1.66847305, + "memory(GiB)": 111.15, + "step": 10820, + "train_speed(iter/s)": 0.448669 + }, + { + "acc": 0.63835955, + "epoch": 0.27460679857940135, + "grad_norm": 5.5, + "learning_rate": 9.79306395142613e-06, + "loss": 1.63952446, + "memory(GiB)": 111.15, + "step": 10825, + "train_speed(iter/s)": 0.44878 + }, + { + "acc": 0.65294881, + "epoch": 0.2747336377473364, + "grad_norm": 4.3125, + "learning_rate": 9.792765289176751e-06, + "loss": 1.62337971, + "memory(GiB)": 111.15, + "step": 10830, + "train_speed(iter/s)": 0.44889 + }, + { + "acc": 0.63839631, + "epoch": 0.27486047691527143, + "grad_norm": 5.09375, + "learning_rate": 9.79246641612016e-06, + "loss": 1.66194191, + "memory(GiB)": 111.15, + "step": 10835, + "train_speed(iter/s)": 0.449001 + }, + { + "acc": 0.64459667, + "epoch": 0.2749873160832065, + "grad_norm": 5.0625, + "learning_rate": 9.792167332269498e-06, + "loss": 1.65412235, + "memory(GiB)": 111.15, + "step": 10840, + "train_speed(iter/s)": 0.449112 + }, + { + "acc": 0.65780134, + "epoch": 0.2751141552511416, + "grad_norm": 4.625, + "learning_rate": 9.791868037637922e-06, + "loss": 1.62622795, + "memory(GiB)": 111.15, + "step": 10845, + "train_speed(iter/s)": 0.449223 + }, + { + "acc": 0.64601293, + "epoch": 0.2752409944190766, + "grad_norm": 5.625, + "learning_rate": 9.791568532238594e-06, + "loss": 1.6990694, + "memory(GiB)": 111.15, + "step": 10850, + "train_speed(iter/s)": 0.449334 + }, + { + "acc": 0.64925914, + "epoch": 0.27536783358701167, + "grad_norm": 5.5, + "learning_rate": 9.79126881608469e-06, + "loss": 1.68230667, + "memory(GiB)": 111.15, + "step": 10855, + "train_speed(iter/s)": 0.449444 + }, + { + "acc": 0.64796381, + "epoch": 0.2754946727549467, + "grad_norm": 5.6875, + "learning_rate": 9.790968889189392e-06, + "loss": 1.66537838, + "memory(GiB)": 111.15, + "step": 10860, + "train_speed(iter/s)": 0.449555 + }, + { + "acc": 0.65023007, + "epoch": 0.2756215119228818, + "grad_norm": 5.78125, + "learning_rate": 9.790668751565893e-06, + "loss": 1.65628929, + "memory(GiB)": 111.15, + "step": 10865, + "train_speed(iter/s)": 0.449666 + }, + { + "acc": 0.65258899, + "epoch": 0.27574835109081686, + "grad_norm": 4.78125, + "learning_rate": 9.790368403227391e-06, + "loss": 1.70908203, + "memory(GiB)": 111.15, + "step": 10870, + "train_speed(iter/s)": 0.449776 + }, + { + "acc": 0.64060316, + "epoch": 0.2758751902587519, + "grad_norm": 4.5, + "learning_rate": 9.7900678441871e-06, + "loss": 1.62029591, + "memory(GiB)": 111.15, + "step": 10875, + "train_speed(iter/s)": 0.449886 + }, + { + "acc": 0.63888922, + "epoch": 0.27600202942668695, + "grad_norm": 5.5, + "learning_rate": 9.78976707445824e-06, + "loss": 1.63015079, + "memory(GiB)": 111.15, + "step": 10880, + "train_speed(iter/s)": 0.449996 + }, + { + "acc": 0.64855566, + "epoch": 0.27612886859462205, + "grad_norm": 5.6875, + "learning_rate": 9.78946609405404e-06, + "loss": 1.69419975, + "memory(GiB)": 111.15, + "step": 10885, + "train_speed(iter/s)": 0.450106 + }, + { + "acc": 0.65410242, + "epoch": 0.2762557077625571, + "grad_norm": 6.96875, + "learning_rate": 9.789164902987738e-06, + "loss": 1.63207436, + "memory(GiB)": 111.15, + "step": 10890, + "train_speed(iter/s)": 0.450216 + }, + { + "acc": 0.66555386, + "epoch": 0.27638254693049213, + "grad_norm": 4.96875, + "learning_rate": 9.78886350127258e-06, + "loss": 1.59287901, + "memory(GiB)": 111.15, + "step": 10895, + "train_speed(iter/s)": 0.450326 + }, + { + "acc": 0.63766856, + "epoch": 0.2765093860984272, + "grad_norm": 5.125, + "learning_rate": 9.788561888921825e-06, + "loss": 1.71051807, + "memory(GiB)": 111.15, + "step": 10900, + "train_speed(iter/s)": 0.450436 + }, + { + "epoch": 0.2765093860984272, + "eval_acc": 0.6372250725557375, + "eval_loss": 1.6446298360824585, + "eval_runtime": 112.6253, + "eval_samples_per_second": 56.559, + "eval_steps_per_second": 28.28, + "step": 10900 + }, + { + "acc": 0.65730972, + "epoch": 0.2766362252663623, + "grad_norm": 5.9375, + "learning_rate": 9.788260065948738e-06, + "loss": 1.67935715, + "memory(GiB)": 111.15, + "step": 10905, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.62518539, + "epoch": 0.2767630644342973, + "grad_norm": 5.75, + "learning_rate": 9.787958032366596e-06, + "loss": 1.72752686, + "memory(GiB)": 111.15, + "step": 10910, + "train_speed(iter/s)": 0.448486 + }, + { + "acc": 0.65110226, + "epoch": 0.27688990360223237, + "grad_norm": 4.84375, + "learning_rate": 9.787655788188684e-06, + "loss": 1.59783077, + "memory(GiB)": 111.15, + "step": 10915, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.65954733, + "epoch": 0.2770167427701674, + "grad_norm": 5.375, + "learning_rate": 9.787353333428293e-06, + "loss": 1.5819643, + "memory(GiB)": 111.15, + "step": 10920, + "train_speed(iter/s)": 0.448703 + }, + { + "acc": 0.63955789, + "epoch": 0.2771435819381025, + "grad_norm": 5.6875, + "learning_rate": 9.78705066809873e-06, + "loss": 1.67073822, + "memory(GiB)": 111.15, + "step": 10925, + "train_speed(iter/s)": 0.448813 + }, + { + "acc": 0.63286953, + "epoch": 0.27727042110603756, + "grad_norm": 5.3125, + "learning_rate": 9.786747792213304e-06, + "loss": 1.72865868, + "memory(GiB)": 111.15, + "step": 10930, + "train_speed(iter/s)": 0.448921 + }, + { + "acc": 0.65135775, + "epoch": 0.2773972602739726, + "grad_norm": 5.84375, + "learning_rate": 9.78644470578534e-06, + "loss": 1.64770889, + "memory(GiB)": 111.15, + "step": 10935, + "train_speed(iter/s)": 0.449031 + }, + { + "acc": 0.6353941, + "epoch": 0.27752409944190765, + "grad_norm": 4.78125, + "learning_rate": 9.78614140882817e-06, + "loss": 1.71120033, + "memory(GiB)": 111.15, + "step": 10940, + "train_speed(iter/s)": 0.44914 + }, + { + "acc": 0.66260204, + "epoch": 0.27765093860984275, + "grad_norm": 5.09375, + "learning_rate": 9.78583790135513e-06, + "loss": 1.6121273, + "memory(GiB)": 111.15, + "step": 10945, + "train_speed(iter/s)": 0.449249 + }, + { + "acc": 0.65313053, + "epoch": 0.2777777777777778, + "grad_norm": 4.96875, + "learning_rate": 9.785534183379571e-06, + "loss": 1.64291039, + "memory(GiB)": 111.15, + "step": 10950, + "train_speed(iter/s)": 0.449357 + }, + { + "acc": 0.63889456, + "epoch": 0.27790461694571283, + "grad_norm": 5.09375, + "learning_rate": 9.785230254914855e-06, + "loss": 1.69854431, + "memory(GiB)": 111.15, + "step": 10955, + "train_speed(iter/s)": 0.449466 + }, + { + "acc": 0.63251867, + "epoch": 0.2780314561136479, + "grad_norm": 4.78125, + "learning_rate": 9.784926115974346e-06, + "loss": 1.70529423, + "memory(GiB)": 111.15, + "step": 10960, + "train_speed(iter/s)": 0.449575 + }, + { + "acc": 0.65269289, + "epoch": 0.278158295281583, + "grad_norm": 4.75, + "learning_rate": 9.784621766571424e-06, + "loss": 1.64479561, + "memory(GiB)": 111.15, + "step": 10965, + "train_speed(iter/s)": 0.449684 + }, + { + "acc": 0.64205313, + "epoch": 0.278285134449518, + "grad_norm": 5.1875, + "learning_rate": 9.784317206719475e-06, + "loss": 1.7293457, + "memory(GiB)": 111.15, + "step": 10970, + "train_speed(iter/s)": 0.449793 + }, + { + "acc": 0.63773346, + "epoch": 0.27841197361745307, + "grad_norm": 4.3125, + "learning_rate": 9.784012436431896e-06, + "loss": 1.66916771, + "memory(GiB)": 111.15, + "step": 10975, + "train_speed(iter/s)": 0.449902 + }, + { + "acc": 0.65136795, + "epoch": 0.2785388127853881, + "grad_norm": 4.84375, + "learning_rate": 9.78370745572209e-06, + "loss": 1.68317585, + "memory(GiB)": 111.15, + "step": 10980, + "train_speed(iter/s)": 0.450011 + }, + { + "acc": 0.63882689, + "epoch": 0.2786656519533232, + "grad_norm": 5.875, + "learning_rate": 9.783402264603471e-06, + "loss": 1.60392971, + "memory(GiB)": 111.15, + "step": 10985, + "train_speed(iter/s)": 0.450119 + }, + { + "acc": 0.63746691, + "epoch": 0.27879249112125826, + "grad_norm": 4.53125, + "learning_rate": 9.783096863089465e-06, + "loss": 1.64590454, + "memory(GiB)": 111.15, + "step": 10990, + "train_speed(iter/s)": 0.450227 + }, + { + "acc": 0.65722866, + "epoch": 0.2789193302891933, + "grad_norm": 5.5625, + "learning_rate": 9.782791251193505e-06, + "loss": 1.52460938, + "memory(GiB)": 111.15, + "step": 10995, + "train_speed(iter/s)": 0.450337 + }, + { + "acc": 0.64458003, + "epoch": 0.27904616945712835, + "grad_norm": 4.3125, + "learning_rate": 9.782485428929032e-06, + "loss": 1.64803886, + "memory(GiB)": 111.15, + "step": 11000, + "train_speed(iter/s)": 0.450445 + }, + { + "epoch": 0.27904616945712835, + "eval_acc": 0.6371958288440754, + "eval_loss": 1.6440985202789307, + "eval_runtime": 113.9778, + "eval_samples_per_second": 55.888, + "eval_steps_per_second": 27.944, + "step": 11000 + }, + { + "acc": 0.65625963, + "epoch": 0.27917300862506345, + "grad_norm": 5.125, + "learning_rate": 9.782179396309496e-06, + "loss": 1.68092861, + "memory(GiB)": 111.15, + "step": 11005, + "train_speed(iter/s)": 0.448378 + }, + { + "acc": 0.63286467, + "epoch": 0.2792998477929985, + "grad_norm": 4.6875, + "learning_rate": 9.78187315334836e-06, + "loss": 1.66688595, + "memory(GiB)": 111.15, + "step": 11010, + "train_speed(iter/s)": 0.448486 + }, + { + "acc": 0.637713, + "epoch": 0.27942668696093353, + "grad_norm": 5.125, + "learning_rate": 9.781566700059094e-06, + "loss": 1.65753059, + "memory(GiB)": 111.15, + "step": 11015, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.63673034, + "epoch": 0.2795535261288686, + "grad_norm": 6.5625, + "learning_rate": 9.781260036455176e-06, + "loss": 1.73183899, + "memory(GiB)": 111.15, + "step": 11020, + "train_speed(iter/s)": 0.448701 + }, + { + "acc": 0.66248817, + "epoch": 0.2796803652968037, + "grad_norm": 4.5, + "learning_rate": 9.780953162550093e-06, + "loss": 1.61990871, + "memory(GiB)": 111.15, + "step": 11025, + "train_speed(iter/s)": 0.448808 + }, + { + "acc": 0.63879566, + "epoch": 0.2798072044647387, + "grad_norm": 5.53125, + "learning_rate": 9.780646078357346e-06, + "loss": 1.67689247, + "memory(GiB)": 111.15, + "step": 11030, + "train_speed(iter/s)": 0.448916 + }, + { + "acc": 0.64713459, + "epoch": 0.27993404363267377, + "grad_norm": 5.03125, + "learning_rate": 9.78033878389044e-06, + "loss": 1.72274246, + "memory(GiB)": 111.15, + "step": 11035, + "train_speed(iter/s)": 0.449024 + }, + { + "acc": 0.65327854, + "epoch": 0.2800608828006088, + "grad_norm": 6.8125, + "learning_rate": 9.780031279162892e-06, + "loss": 1.6879549, + "memory(GiB)": 111.15, + "step": 11040, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.65426345, + "epoch": 0.2801877219685439, + "grad_norm": 7.5, + "learning_rate": 9.779723564188228e-06, + "loss": 1.71193504, + "memory(GiB)": 111.15, + "step": 11045, + "train_speed(iter/s)": 0.44924 + }, + { + "acc": 0.63709173, + "epoch": 0.28031456113647896, + "grad_norm": 5.8125, + "learning_rate": 9.77941563897998e-06, + "loss": 1.68700409, + "memory(GiB)": 111.15, + "step": 11050, + "train_speed(iter/s)": 0.449348 + }, + { + "acc": 0.64473991, + "epoch": 0.280441400304414, + "grad_norm": 5.0, + "learning_rate": 9.779107503551695e-06, + "loss": 1.66767731, + "memory(GiB)": 111.15, + "step": 11055, + "train_speed(iter/s)": 0.449456 + }, + { + "acc": 0.6568327, + "epoch": 0.28056823947234905, + "grad_norm": 5.40625, + "learning_rate": 9.778799157916926e-06, + "loss": 1.6142765, + "memory(GiB)": 111.15, + "step": 11060, + "train_speed(iter/s)": 0.449562 + }, + { + "acc": 0.6405899, + "epoch": 0.28069507864028415, + "grad_norm": 4.40625, + "learning_rate": 9.77849060208923e-06, + "loss": 1.68636894, + "memory(GiB)": 111.15, + "step": 11065, + "train_speed(iter/s)": 0.449669 + }, + { + "acc": 0.63269148, + "epoch": 0.2808219178082192, + "grad_norm": 4.6875, + "learning_rate": 9.778181836082185e-06, + "loss": 1.69798508, + "memory(GiB)": 111.15, + "step": 11070, + "train_speed(iter/s)": 0.449776 + }, + { + "acc": 0.6535573, + "epoch": 0.28094875697615423, + "grad_norm": 6.28125, + "learning_rate": 9.777872859909373e-06, + "loss": 1.64425087, + "memory(GiB)": 111.15, + "step": 11075, + "train_speed(iter/s)": 0.449884 + }, + { + "acc": 0.66931486, + "epoch": 0.2810755961440893, + "grad_norm": 5.3125, + "learning_rate": 9.777563673584376e-06, + "loss": 1.57153587, + "memory(GiB)": 111.15, + "step": 11080, + "train_speed(iter/s)": 0.449991 + }, + { + "acc": 0.64883709, + "epoch": 0.2812024353120244, + "grad_norm": 5.9375, + "learning_rate": 9.777254277120801e-06, + "loss": 1.64784203, + "memory(GiB)": 111.15, + "step": 11085, + "train_speed(iter/s)": 0.450098 + }, + { + "acc": 0.66168542, + "epoch": 0.2813292744799594, + "grad_norm": 5.0625, + "learning_rate": 9.776944670532253e-06, + "loss": 1.64538383, + "memory(GiB)": 111.15, + "step": 11090, + "train_speed(iter/s)": 0.450205 + }, + { + "acc": 0.64872818, + "epoch": 0.28145611364789447, + "grad_norm": 5.75, + "learning_rate": 9.776634853832352e-06, + "loss": 1.62329617, + "memory(GiB)": 111.15, + "step": 11095, + "train_speed(iter/s)": 0.450313 + }, + { + "acc": 0.65059605, + "epoch": 0.2815829528158295, + "grad_norm": 5.3125, + "learning_rate": 9.776324827034724e-06, + "loss": 1.66148243, + "memory(GiB)": 111.15, + "step": 11100, + "train_speed(iter/s)": 0.45042 + }, + { + "epoch": 0.2815829528158295, + "eval_acc": 0.6373219945715316, + "eval_loss": 1.6432462930679321, + "eval_runtime": 112.8929, + "eval_samples_per_second": 56.425, + "eval_steps_per_second": 28.213, + "step": 11100 + }, + { + "acc": 0.64392147, + "epoch": 0.2817097919837646, + "grad_norm": 4.71875, + "learning_rate": 9.776014590153005e-06, + "loss": 1.66989422, + "memory(GiB)": 111.15, + "step": 11105, + "train_speed(iter/s)": 0.448391 + }, + { + "acc": 0.65874119, + "epoch": 0.28183663115169966, + "grad_norm": 5.3125, + "learning_rate": 9.77570414320084e-06, + "loss": 1.60502071, + "memory(GiB)": 111.15, + "step": 11110, + "train_speed(iter/s)": 0.448497 + }, + { + "acc": 0.65444789, + "epoch": 0.2819634703196347, + "grad_norm": 5.5, + "learning_rate": 9.775393486191884e-06, + "loss": 1.6413517, + "memory(GiB)": 111.15, + "step": 11115, + "train_speed(iter/s)": 0.448603 + }, + { + "acc": 0.63810673, + "epoch": 0.28209030948756975, + "grad_norm": 5.875, + "learning_rate": 9.775082619139805e-06, + "loss": 1.67183723, + "memory(GiB)": 111.15, + "step": 11120, + "train_speed(iter/s)": 0.44871 + }, + { + "acc": 0.66055446, + "epoch": 0.28221714865550485, + "grad_norm": 5.75, + "learning_rate": 9.77477154205827e-06, + "loss": 1.63652782, + "memory(GiB)": 111.15, + "step": 11125, + "train_speed(iter/s)": 0.448817 + }, + { + "acc": 0.63723497, + "epoch": 0.2823439878234399, + "grad_norm": 5.21875, + "learning_rate": 9.774460254960968e-06, + "loss": 1.64890442, + "memory(GiB)": 111.15, + "step": 11130, + "train_speed(iter/s)": 0.448925 + }, + { + "acc": 0.65797944, + "epoch": 0.28247082699137493, + "grad_norm": 5.59375, + "learning_rate": 9.774148757861584e-06, + "loss": 1.59983253, + "memory(GiB)": 111.15, + "step": 11135, + "train_speed(iter/s)": 0.449031 + }, + { + "acc": 0.64013643, + "epoch": 0.28259766615931, + "grad_norm": 5.75, + "learning_rate": 9.773837050773824e-06, + "loss": 1.72286663, + "memory(GiB)": 111.15, + "step": 11140, + "train_speed(iter/s)": 0.449138 + }, + { + "acc": 0.62790108, + "epoch": 0.2827245053272451, + "grad_norm": 5.34375, + "learning_rate": 9.773525133711399e-06, + "loss": 1.71524029, + "memory(GiB)": 111.15, + "step": 11145, + "train_speed(iter/s)": 0.449245 + }, + { + "acc": 0.64691563, + "epoch": 0.2828513444951801, + "grad_norm": 4.625, + "learning_rate": 9.773213006688024e-06, + "loss": 1.68762417, + "memory(GiB)": 111.15, + "step": 11150, + "train_speed(iter/s)": 0.449351 + }, + { + "acc": 0.62633705, + "epoch": 0.28297818366311517, + "grad_norm": 5.0, + "learning_rate": 9.77290066971743e-06, + "loss": 1.73390636, + "memory(GiB)": 111.15, + "step": 11155, + "train_speed(iter/s)": 0.449458 + }, + { + "acc": 0.63712168, + "epoch": 0.2831050228310502, + "grad_norm": 4.625, + "learning_rate": 9.772588122813358e-06, + "loss": 1.679105, + "memory(GiB)": 111.15, + "step": 11160, + "train_speed(iter/s)": 0.449564 + }, + { + "acc": 0.65788794, + "epoch": 0.2832318619989853, + "grad_norm": 5.40625, + "learning_rate": 9.772275365989548e-06, + "loss": 1.65855522, + "memory(GiB)": 111.15, + "step": 11165, + "train_speed(iter/s)": 0.449671 + }, + { + "acc": 0.65132484, + "epoch": 0.28335870116692036, + "grad_norm": 5.375, + "learning_rate": 9.771962399259764e-06, + "loss": 1.64876175, + "memory(GiB)": 111.15, + "step": 11170, + "train_speed(iter/s)": 0.449777 + }, + { + "acc": 0.65645885, + "epoch": 0.2834855403348554, + "grad_norm": 5.125, + "learning_rate": 9.771649222637767e-06, + "loss": 1.55534706, + "memory(GiB)": 111.15, + "step": 11175, + "train_speed(iter/s)": 0.449883 + }, + { + "acc": 0.61657734, + "epoch": 0.28361237950279045, + "grad_norm": 6.09375, + "learning_rate": 9.771335836137332e-06, + "loss": 1.73665085, + "memory(GiB)": 111.15, + "step": 11180, + "train_speed(iter/s)": 0.44999 + }, + { + "acc": 0.65149679, + "epoch": 0.28373921867072555, + "grad_norm": 5.125, + "learning_rate": 9.771022239772248e-06, + "loss": 1.59150524, + "memory(GiB)": 111.15, + "step": 11185, + "train_speed(iter/s)": 0.450095 + }, + { + "acc": 0.64141812, + "epoch": 0.2838660578386606, + "grad_norm": 5.53125, + "learning_rate": 9.770708433556302e-06, + "loss": 1.72291679, + "memory(GiB)": 111.15, + "step": 11190, + "train_speed(iter/s)": 0.450201 + }, + { + "acc": 0.64601059, + "epoch": 0.28399289700659563, + "grad_norm": 5.25, + "learning_rate": 9.7703944175033e-06, + "loss": 1.66540737, + "memory(GiB)": 111.15, + "step": 11195, + "train_speed(iter/s)": 0.450307 + }, + { + "acc": 0.65243759, + "epoch": 0.2841197361745307, + "grad_norm": 5.09375, + "learning_rate": 9.770080191627054e-06, + "loss": 1.61659222, + "memory(GiB)": 111.15, + "step": 11200, + "train_speed(iter/s)": 0.450413 + }, + { + "epoch": 0.2841197361745307, + "eval_acc": 0.6374435648585837, + "eval_loss": 1.6428890228271484, + "eval_runtime": 114.2028, + "eval_samples_per_second": 55.778, + "eval_steps_per_second": 27.889, + "step": 11200 + }, + { + "acc": 0.65002193, + "epoch": 0.2842465753424658, + "grad_norm": 4.40625, + "learning_rate": 9.769765755941383e-06, + "loss": 1.63769264, + "memory(GiB)": 111.15, + "step": 11205, + "train_speed(iter/s)": 0.448379 + }, + { + "acc": 0.64336171, + "epoch": 0.2843734145104008, + "grad_norm": 5.53125, + "learning_rate": 9.76945111046012e-06, + "loss": 1.66362648, + "memory(GiB)": 111.15, + "step": 11210, + "train_speed(iter/s)": 0.448485 + }, + { + "acc": 0.63385491, + "epoch": 0.28450025367833587, + "grad_norm": 5.5, + "learning_rate": 9.769136255197103e-06, + "loss": 1.73203564, + "memory(GiB)": 111.15, + "step": 11215, + "train_speed(iter/s)": 0.448591 + }, + { + "acc": 0.64736862, + "epoch": 0.2846270928462709, + "grad_norm": 5.75, + "learning_rate": 9.768821190166179e-06, + "loss": 1.65053291, + "memory(GiB)": 111.15, + "step": 11220, + "train_speed(iter/s)": 0.448698 + }, + { + "acc": 0.64030943, + "epoch": 0.284753932014206, + "grad_norm": 4.90625, + "learning_rate": 9.76850591538121e-06, + "loss": 1.60505371, + "memory(GiB)": 111.15, + "step": 11225, + "train_speed(iter/s)": 0.448804 + }, + { + "acc": 0.6424767, + "epoch": 0.28488077118214106, + "grad_norm": 5.4375, + "learning_rate": 9.76819043085606e-06, + "loss": 1.61811295, + "memory(GiB)": 111.15, + "step": 11230, + "train_speed(iter/s)": 0.448911 + }, + { + "acc": 0.65504217, + "epoch": 0.2850076103500761, + "grad_norm": 4.4375, + "learning_rate": 9.767874736604605e-06, + "loss": 1.61249866, + "memory(GiB)": 111.15, + "step": 11235, + "train_speed(iter/s)": 0.449017 + }, + { + "acc": 0.65184245, + "epoch": 0.28513444951801115, + "grad_norm": 5.40625, + "learning_rate": 9.767558832640734e-06, + "loss": 1.5913414, + "memory(GiB)": 111.15, + "step": 11240, + "train_speed(iter/s)": 0.449123 + }, + { + "acc": 0.65228891, + "epoch": 0.28526128868594625, + "grad_norm": 5.3125, + "learning_rate": 9.76724271897834e-06, + "loss": 1.55513945, + "memory(GiB)": 111.15, + "step": 11245, + "train_speed(iter/s)": 0.449228 + }, + { + "acc": 0.63418932, + "epoch": 0.2853881278538813, + "grad_norm": 5.3125, + "learning_rate": 9.766926395631326e-06, + "loss": 1.73253365, + "memory(GiB)": 111.15, + "step": 11250, + "train_speed(iter/s)": 0.449334 + }, + { + "acc": 0.6566186, + "epoch": 0.28551496702181633, + "grad_norm": 6.125, + "learning_rate": 9.766609862613607e-06, + "loss": 1.62139435, + "memory(GiB)": 111.15, + "step": 11255, + "train_speed(iter/s)": 0.44944 + }, + { + "acc": 0.65278406, + "epoch": 0.2856418061897514, + "grad_norm": 5.65625, + "learning_rate": 9.766293119939104e-06, + "loss": 1.68457394, + "memory(GiB)": 111.15, + "step": 11260, + "train_speed(iter/s)": 0.449547 + }, + { + "acc": 0.65203304, + "epoch": 0.2857686453576865, + "grad_norm": 5.0, + "learning_rate": 9.76597616762175e-06, + "loss": 1.64673347, + "memory(GiB)": 111.15, + "step": 11265, + "train_speed(iter/s)": 0.449652 + }, + { + "acc": 0.63427362, + "epoch": 0.2858954845256215, + "grad_norm": 4.28125, + "learning_rate": 9.765659005675488e-06, + "loss": 1.63265877, + "memory(GiB)": 111.15, + "step": 11270, + "train_speed(iter/s)": 0.449757 + }, + { + "acc": 0.6284483, + "epoch": 0.28602232369355657, + "grad_norm": 4.75, + "learning_rate": 9.765341634114263e-06, + "loss": 1.71627026, + "memory(GiB)": 111.15, + "step": 11275, + "train_speed(iter/s)": 0.449863 + }, + { + "acc": 0.64157157, + "epoch": 0.2861491628614916, + "grad_norm": 4.84375, + "learning_rate": 9.765024052952037e-06, + "loss": 1.65192795, + "memory(GiB)": 111.15, + "step": 11280, + "train_speed(iter/s)": 0.449969 + }, + { + "acc": 0.63879046, + "epoch": 0.2862760020294267, + "grad_norm": 4.8125, + "learning_rate": 9.76470626220278e-06, + "loss": 1.70751534, + "memory(GiB)": 111.15, + "step": 11285, + "train_speed(iter/s)": 0.450075 + }, + { + "acc": 0.63971567, + "epoch": 0.28640284119736176, + "grad_norm": 5.3125, + "learning_rate": 9.76438826188047e-06, + "loss": 1.66216564, + "memory(GiB)": 111.15, + "step": 11290, + "train_speed(iter/s)": 0.45018 + }, + { + "acc": 0.6330039, + "epoch": 0.2865296803652968, + "grad_norm": 6.15625, + "learning_rate": 9.76407005199909e-06, + "loss": 1.69981003, + "memory(GiB)": 111.15, + "step": 11295, + "train_speed(iter/s)": 0.450285 + }, + { + "acc": 0.6402606, + "epoch": 0.28665651953323185, + "grad_norm": 4.1875, + "learning_rate": 9.76375163257264e-06, + "loss": 1.63105011, + "memory(GiB)": 111.15, + "step": 11300, + "train_speed(iter/s)": 0.45039 + }, + { + "epoch": 0.28665651953323185, + "eval_acc": 0.6375617930071601, + "eval_loss": 1.6423982381820679, + "eval_runtime": 112.6458, + "eval_samples_per_second": 56.549, + "eval_steps_per_second": 28.274, + "step": 11300 + }, + { + "acc": 0.64987841, + "epoch": 0.28678335870116695, + "grad_norm": 5.46875, + "learning_rate": 9.763433003615124e-06, + "loss": 1.64602165, + "memory(GiB)": 111.15, + "step": 11305, + "train_speed(iter/s)": 0.448402 + }, + { + "acc": 0.64360776, + "epoch": 0.286910197869102, + "grad_norm": 5.53125, + "learning_rate": 9.763114165140559e-06, + "loss": 1.74809456, + "memory(GiB)": 111.15, + "step": 11310, + "train_speed(iter/s)": 0.448507 + }, + { + "acc": 0.62160263, + "epoch": 0.28703703703703703, + "grad_norm": 5.0625, + "learning_rate": 9.762795117162967e-06, + "loss": 1.71172638, + "memory(GiB)": 111.15, + "step": 11315, + "train_speed(iter/s)": 0.448613 + }, + { + "acc": 0.66071787, + "epoch": 0.2871638762049721, + "grad_norm": 5.34375, + "learning_rate": 9.76247585969638e-06, + "loss": 1.61466675, + "memory(GiB)": 111.15, + "step": 11320, + "train_speed(iter/s)": 0.448716 + }, + { + "acc": 0.65368338, + "epoch": 0.2872907153729072, + "grad_norm": 5.0, + "learning_rate": 9.762156392754842e-06, + "loss": 1.58079624, + "memory(GiB)": 111.15, + "step": 11325, + "train_speed(iter/s)": 0.448821 + }, + { + "acc": 0.62557964, + "epoch": 0.2874175545408422, + "grad_norm": 5.28125, + "learning_rate": 9.761836716352405e-06, + "loss": 1.69626083, + "memory(GiB)": 111.15, + "step": 11330, + "train_speed(iter/s)": 0.448925 + }, + { + "acc": 0.63066788, + "epoch": 0.28754439370877727, + "grad_norm": 5.0, + "learning_rate": 9.761516830503128e-06, + "loss": 1.73517246, + "memory(GiB)": 111.15, + "step": 11335, + "train_speed(iter/s)": 0.44903 + }, + { + "acc": 0.65028868, + "epoch": 0.2876712328767123, + "grad_norm": 5.1875, + "learning_rate": 9.761196735221083e-06, + "loss": 1.68195801, + "memory(GiB)": 111.15, + "step": 11340, + "train_speed(iter/s)": 0.449135 + }, + { + "acc": 0.64325228, + "epoch": 0.2877980720446474, + "grad_norm": 5.40625, + "learning_rate": 9.76087643052035e-06, + "loss": 1.65014992, + "memory(GiB)": 111.15, + "step": 11345, + "train_speed(iter/s)": 0.449239 + }, + { + "acc": 0.64079103, + "epoch": 0.28792491121258246, + "grad_norm": 4.875, + "learning_rate": 9.760555916415015e-06, + "loss": 1.68327618, + "memory(GiB)": 111.15, + "step": 11350, + "train_speed(iter/s)": 0.449343 + }, + { + "acc": 0.65408354, + "epoch": 0.2880517503805175, + "grad_norm": 4.90625, + "learning_rate": 9.760235192919175e-06, + "loss": 1.63960381, + "memory(GiB)": 111.15, + "step": 11355, + "train_speed(iter/s)": 0.449448 + }, + { + "acc": 0.64269376, + "epoch": 0.28817858954845255, + "grad_norm": 5.71875, + "learning_rate": 9.75991426004694e-06, + "loss": 1.67405891, + "memory(GiB)": 111.15, + "step": 11360, + "train_speed(iter/s)": 0.449552 + }, + { + "acc": 0.64675093, + "epoch": 0.28830542871638765, + "grad_norm": 4.9375, + "learning_rate": 9.759593117812423e-06, + "loss": 1.65332355, + "memory(GiB)": 111.15, + "step": 11365, + "train_speed(iter/s)": 0.449657 + }, + { + "acc": 0.66907692, + "epoch": 0.2884322678843227, + "grad_norm": 5.34375, + "learning_rate": 9.75927176622975e-06, + "loss": 1.61816444, + "memory(GiB)": 111.15, + "step": 11370, + "train_speed(iter/s)": 0.449762 + }, + { + "acc": 0.64397087, + "epoch": 0.28855910705225774, + "grad_norm": 4.96875, + "learning_rate": 9.758950205313057e-06, + "loss": 1.69099274, + "memory(GiB)": 111.15, + "step": 11375, + "train_speed(iter/s)": 0.449866 + }, + { + "acc": 0.6582695, + "epoch": 0.2886859462201928, + "grad_norm": 4.34375, + "learning_rate": 9.758628435076488e-06, + "loss": 1.61282654, + "memory(GiB)": 111.15, + "step": 11380, + "train_speed(iter/s)": 0.44997 + }, + { + "acc": 0.63627133, + "epoch": 0.2888127853881279, + "grad_norm": 5.03125, + "learning_rate": 9.758306455534193e-06, + "loss": 1.67416039, + "memory(GiB)": 111.15, + "step": 11385, + "train_speed(iter/s)": 0.450074 + }, + { + "acc": 0.63031754, + "epoch": 0.2889396245560629, + "grad_norm": 5.375, + "learning_rate": 9.757984266700336e-06, + "loss": 1.69458675, + "memory(GiB)": 111.15, + "step": 11390, + "train_speed(iter/s)": 0.450178 + }, + { + "acc": 0.64530091, + "epoch": 0.28906646372399797, + "grad_norm": 5.8125, + "learning_rate": 9.75766186858909e-06, + "loss": 1.6250721, + "memory(GiB)": 111.15, + "step": 11395, + "train_speed(iter/s)": 0.450283 + }, + { + "acc": 0.64924459, + "epoch": 0.289193302891933, + "grad_norm": 4.96875, + "learning_rate": 9.757339261214631e-06, + "loss": 1.70327377, + "memory(GiB)": 111.15, + "step": 11400, + "train_speed(iter/s)": 0.450387 + }, + { + "epoch": 0.289193302891933, + "eval_acc": 0.6376223692670314, + "eval_loss": 1.6421945095062256, + "eval_runtime": 112.5113, + "eval_samples_per_second": 56.617, + "eval_steps_per_second": 28.308, + "step": 11400 + }, + { + "acc": 0.63925867, + "epoch": 0.2893201420598681, + "grad_norm": 4.78125, + "learning_rate": 9.757016444591152e-06, + "loss": 1.66531487, + "memory(GiB)": 111.15, + "step": 11405, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.6477879, + "epoch": 0.28944698122780316, + "grad_norm": 4.4375, + "learning_rate": 9.756693418732852e-06, + "loss": 1.65402031, + "memory(GiB)": 111.15, + "step": 11410, + "train_speed(iter/s)": 0.448523 + }, + { + "acc": 0.64859414, + "epoch": 0.2895738203957382, + "grad_norm": 5.375, + "learning_rate": 9.756370183653938e-06, + "loss": 1.70133362, + "memory(GiB)": 111.15, + "step": 11415, + "train_speed(iter/s)": 0.448627 + }, + { + "acc": 0.64122958, + "epoch": 0.28970065956367325, + "grad_norm": 5.875, + "learning_rate": 9.756046739368628e-06, + "loss": 1.65705261, + "memory(GiB)": 111.15, + "step": 11420, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.63132238, + "epoch": 0.28982749873160835, + "grad_norm": 5.21875, + "learning_rate": 9.755723085891147e-06, + "loss": 1.71997509, + "memory(GiB)": 111.15, + "step": 11425, + "train_speed(iter/s)": 0.448836 + }, + { + "acc": 0.65546174, + "epoch": 0.2899543378995434, + "grad_norm": 5.71875, + "learning_rate": 9.755399223235734e-06, + "loss": 1.59276905, + "memory(GiB)": 111.15, + "step": 11430, + "train_speed(iter/s)": 0.44894 + }, + { + "acc": 0.6415092, + "epoch": 0.29008117706747844, + "grad_norm": 6.28125, + "learning_rate": 9.75507515141663e-06, + "loss": 1.67783546, + "memory(GiB)": 111.15, + "step": 11435, + "train_speed(iter/s)": 0.449044 + }, + { + "acc": 0.63919439, + "epoch": 0.2902080162354135, + "grad_norm": 5.78125, + "learning_rate": 9.75475087044809e-06, + "loss": 1.66107254, + "memory(GiB)": 111.15, + "step": 11440, + "train_speed(iter/s)": 0.449147 + }, + { + "acc": 0.64341927, + "epoch": 0.2903348554033486, + "grad_norm": 5.28125, + "learning_rate": 9.754426380344382e-06, + "loss": 1.68818626, + "memory(GiB)": 111.15, + "step": 11445, + "train_speed(iter/s)": 0.449251 + }, + { + "acc": 0.63659678, + "epoch": 0.2904616945712836, + "grad_norm": 5.09375, + "learning_rate": 9.754101681119772e-06, + "loss": 1.71427994, + "memory(GiB)": 111.15, + "step": 11450, + "train_speed(iter/s)": 0.449355 + }, + { + "acc": 0.62910376, + "epoch": 0.29058853373921867, + "grad_norm": 6.3125, + "learning_rate": 9.753776772788545e-06, + "loss": 1.68716812, + "memory(GiB)": 111.15, + "step": 11455, + "train_speed(iter/s)": 0.44946 + }, + { + "acc": 0.64524622, + "epoch": 0.2907153729071537, + "grad_norm": 4.75, + "learning_rate": 9.753451655364992e-06, + "loss": 1.65006809, + "memory(GiB)": 111.15, + "step": 11460, + "train_speed(iter/s)": 0.449564 + }, + { + "acc": 0.63962708, + "epoch": 0.2908422120750888, + "grad_norm": 5.28125, + "learning_rate": 9.75312632886341e-06, + "loss": 1.73162804, + "memory(GiB)": 111.15, + "step": 11465, + "train_speed(iter/s)": 0.449668 + }, + { + "acc": 0.67126164, + "epoch": 0.29096905124302386, + "grad_norm": 4.6875, + "learning_rate": 9.752800793298113e-06, + "loss": 1.60011826, + "memory(GiB)": 111.15, + "step": 11470, + "train_speed(iter/s)": 0.449772 + }, + { + "acc": 0.64818296, + "epoch": 0.2910958904109589, + "grad_norm": 5.375, + "learning_rate": 9.752475048683419e-06, + "loss": 1.62799568, + "memory(GiB)": 111.15, + "step": 11475, + "train_speed(iter/s)": 0.449875 + }, + { + "acc": 0.64708767, + "epoch": 0.29122272957889395, + "grad_norm": 4.75, + "learning_rate": 9.752149095033651e-06, + "loss": 1.63794365, + "memory(GiB)": 111.15, + "step": 11480, + "train_speed(iter/s)": 0.44998 + }, + { + "acc": 0.64788084, + "epoch": 0.29134956874682905, + "grad_norm": 6.0625, + "learning_rate": 9.75182293236315e-06, + "loss": 1.62897167, + "memory(GiB)": 111.15, + "step": 11485, + "train_speed(iter/s)": 0.450084 + }, + { + "acc": 0.63815374, + "epoch": 0.2914764079147641, + "grad_norm": 6.21875, + "learning_rate": 9.751496560686262e-06, + "loss": 1.59192648, + "memory(GiB)": 111.15, + "step": 11490, + "train_speed(iter/s)": 0.450188 + }, + { + "acc": 0.64463158, + "epoch": 0.29160324708269914, + "grad_norm": 5.1875, + "learning_rate": 9.751169980017341e-06, + "loss": 1.61655426, + "memory(GiB)": 111.15, + "step": 11495, + "train_speed(iter/s)": 0.450291 + }, + { + "acc": 0.64837766, + "epoch": 0.2917300862506342, + "grad_norm": 4.59375, + "learning_rate": 9.750843190370752e-06, + "loss": 1.62248306, + "memory(GiB)": 111.15, + "step": 11500, + "train_speed(iter/s)": 0.450393 + }, + { + "epoch": 0.2917300862506342, + "eval_acc": 0.6376060763419626, + "eval_loss": 1.6415537595748901, + "eval_runtime": 113.8922, + "eval_samples_per_second": 55.93, + "eval_steps_per_second": 27.965, + "step": 11500 + }, + { + "acc": 0.66487856, + "epoch": 0.2918569254185693, + "grad_norm": 4.0, + "learning_rate": 9.750516191760868e-06, + "loss": 1.61725578, + "memory(GiB)": 111.15, + "step": 11505, + "train_speed(iter/s)": 0.448418 + }, + { + "acc": 0.64256449, + "epoch": 0.2919837645865043, + "grad_norm": 4.0625, + "learning_rate": 9.750188984202073e-06, + "loss": 1.62876587, + "memory(GiB)": 111.15, + "step": 11510, + "train_speed(iter/s)": 0.448522 + }, + { + "acc": 0.63722296, + "epoch": 0.29211060375443937, + "grad_norm": 5.75, + "learning_rate": 9.749861567708759e-06, + "loss": 1.74470482, + "memory(GiB)": 111.15, + "step": 11515, + "train_speed(iter/s)": 0.448625 + }, + { + "acc": 0.62911801, + "epoch": 0.2922374429223744, + "grad_norm": 5.46875, + "learning_rate": 9.749533942295323e-06, + "loss": 1.72435951, + "memory(GiB)": 111.15, + "step": 11520, + "train_speed(iter/s)": 0.448729 + }, + { + "acc": 0.64836416, + "epoch": 0.2923642820903095, + "grad_norm": 6.4375, + "learning_rate": 9.749206107976183e-06, + "loss": 1.63874817, + "memory(GiB)": 111.15, + "step": 11525, + "train_speed(iter/s)": 0.448834 + }, + { + "acc": 0.63467135, + "epoch": 0.29249112125824456, + "grad_norm": 6.0, + "learning_rate": 9.748878064765753e-06, + "loss": 1.67717838, + "memory(GiB)": 111.15, + "step": 11530, + "train_speed(iter/s)": 0.448937 + }, + { + "acc": 0.64459701, + "epoch": 0.2926179604261796, + "grad_norm": 5.90625, + "learning_rate": 9.748549812678466e-06, + "loss": 1.69776421, + "memory(GiB)": 111.15, + "step": 11535, + "train_speed(iter/s)": 0.44904 + }, + { + "acc": 0.64658208, + "epoch": 0.29274479959411465, + "grad_norm": 4.78125, + "learning_rate": 9.748221351728754e-06, + "loss": 1.69286003, + "memory(GiB)": 111.15, + "step": 11540, + "train_speed(iter/s)": 0.449143 + }, + { + "acc": 0.65305119, + "epoch": 0.29287163876204975, + "grad_norm": 5.59375, + "learning_rate": 9.747892681931067e-06, + "loss": 1.62567425, + "memory(GiB)": 111.15, + "step": 11545, + "train_speed(iter/s)": 0.449246 + }, + { + "acc": 0.64375591, + "epoch": 0.2929984779299848, + "grad_norm": 5.375, + "learning_rate": 9.747563803299865e-06, + "loss": 1.65536385, + "memory(GiB)": 111.15, + "step": 11550, + "train_speed(iter/s)": 0.449349 + }, + { + "acc": 0.64389753, + "epoch": 0.29312531709791984, + "grad_norm": 5.0625, + "learning_rate": 9.74723471584961e-06, + "loss": 1.63563118, + "memory(GiB)": 111.15, + "step": 11555, + "train_speed(iter/s)": 0.449452 + }, + { + "acc": 0.64412231, + "epoch": 0.2932521562658549, + "grad_norm": 5.0625, + "learning_rate": 9.746905419594777e-06, + "loss": 1.61791515, + "memory(GiB)": 111.15, + "step": 11560, + "train_speed(iter/s)": 0.449556 + }, + { + "acc": 0.64138718, + "epoch": 0.29337899543379, + "grad_norm": 5.53125, + "learning_rate": 9.746575914549851e-06, + "loss": 1.66518135, + "memory(GiB)": 111.15, + "step": 11565, + "train_speed(iter/s)": 0.449658 + }, + { + "acc": 0.62727528, + "epoch": 0.293505834601725, + "grad_norm": 5.0625, + "learning_rate": 9.746246200729323e-06, + "loss": 1.78143463, + "memory(GiB)": 111.15, + "step": 11570, + "train_speed(iter/s)": 0.44976 + }, + { + "acc": 0.65039501, + "epoch": 0.29363267376966007, + "grad_norm": 4.375, + "learning_rate": 9.745916278147696e-06, + "loss": 1.60077534, + "memory(GiB)": 111.15, + "step": 11575, + "train_speed(iter/s)": 0.449863 + }, + { + "acc": 0.64748378, + "epoch": 0.2937595129375951, + "grad_norm": 7.0, + "learning_rate": 9.745586146819484e-06, + "loss": 1.67576866, + "memory(GiB)": 111.15, + "step": 11580, + "train_speed(iter/s)": 0.449965 + }, + { + "acc": 0.64018588, + "epoch": 0.2938863521055302, + "grad_norm": 4.71875, + "learning_rate": 9.745255806759205e-06, + "loss": 1.74319305, + "memory(GiB)": 111.15, + "step": 11585, + "train_speed(iter/s)": 0.450068 + }, + { + "acc": 0.64772768, + "epoch": 0.29401319127346526, + "grad_norm": 5.4375, + "learning_rate": 9.74492525798139e-06, + "loss": 1.69699211, + "memory(GiB)": 111.15, + "step": 11590, + "train_speed(iter/s)": 0.45017 + }, + { + "acc": 0.63093505, + "epoch": 0.2941400304414003, + "grad_norm": 5.0625, + "learning_rate": 9.744594500500578e-06, + "loss": 1.69472408, + "memory(GiB)": 111.15, + "step": 11595, + "train_speed(iter/s)": 0.450272 + }, + { + "acc": 0.63869567, + "epoch": 0.29426686960933535, + "grad_norm": 5.125, + "learning_rate": 9.744263534331315e-06, + "loss": 1.68456745, + "memory(GiB)": 111.15, + "step": 11600, + "train_speed(iter/s)": 0.450376 + }, + { + "epoch": 0.29426686960933535, + "eval_acc": 0.6376929719423298, + "eval_loss": 1.6413490772247314, + "eval_runtime": 112.4334, + "eval_samples_per_second": 56.656, + "eval_steps_per_second": 28.328, + "step": 11600 + }, + { + "acc": 0.65008388, + "epoch": 0.29439370877727045, + "grad_norm": 5.46875, + "learning_rate": 9.743932359488161e-06, + "loss": 1.65295372, + "memory(GiB)": 111.15, + "step": 11605, + "train_speed(iter/s)": 0.448442 + }, + { + "acc": 0.64418478, + "epoch": 0.2945205479452055, + "grad_norm": 4.5625, + "learning_rate": 9.743600975985681e-06, + "loss": 1.67906227, + "memory(GiB)": 111.15, + "step": 11610, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.64813824, + "epoch": 0.29464738711314054, + "grad_norm": 5.0625, + "learning_rate": 9.743269383838452e-06, + "loss": 1.68654861, + "memory(GiB)": 111.15, + "step": 11615, + "train_speed(iter/s)": 0.448648 + }, + { + "acc": 0.64886427, + "epoch": 0.2947742262810756, + "grad_norm": 5.96875, + "learning_rate": 9.74293758306106e-06, + "loss": 1.65968666, + "memory(GiB)": 111.15, + "step": 11620, + "train_speed(iter/s)": 0.448751 + }, + { + "acc": 0.62220774, + "epoch": 0.2949010654490107, + "grad_norm": 6.3125, + "learning_rate": 9.742605573668096e-06, + "loss": 1.68922386, + "memory(GiB)": 111.15, + "step": 11625, + "train_speed(iter/s)": 0.448854 + }, + { + "acc": 0.64330578, + "epoch": 0.2950279046169457, + "grad_norm": 6.09375, + "learning_rate": 9.742273355674164e-06, + "loss": 1.70709038, + "memory(GiB)": 111.15, + "step": 11630, + "train_speed(iter/s)": 0.448957 + }, + { + "acc": 0.6558012, + "epoch": 0.29515474378488077, + "grad_norm": 4.90625, + "learning_rate": 9.741940929093879e-06, + "loss": 1.66321907, + "memory(GiB)": 111.15, + "step": 11635, + "train_speed(iter/s)": 0.449059 + }, + { + "acc": 0.65294905, + "epoch": 0.2952815829528158, + "grad_norm": 5.03125, + "learning_rate": 9.741608293941858e-06, + "loss": 1.61639748, + "memory(GiB)": 111.15, + "step": 11640, + "train_speed(iter/s)": 0.449161 + }, + { + "acc": 0.6511591, + "epoch": 0.2954084221207509, + "grad_norm": 4.90625, + "learning_rate": 9.741275450232736e-06, + "loss": 1.66115055, + "memory(GiB)": 111.15, + "step": 11645, + "train_speed(iter/s)": 0.449265 + }, + { + "acc": 0.65216026, + "epoch": 0.29553526128868596, + "grad_norm": 5.53125, + "learning_rate": 9.740942397981151e-06, + "loss": 1.57517376, + "memory(GiB)": 111.15, + "step": 11650, + "train_speed(iter/s)": 0.449367 + }, + { + "acc": 0.64622383, + "epoch": 0.295662100456621, + "grad_norm": 5.09375, + "learning_rate": 9.740609137201752e-06, + "loss": 1.64572754, + "memory(GiB)": 111.15, + "step": 11655, + "train_speed(iter/s)": 0.44947 + }, + { + "acc": 0.64864187, + "epoch": 0.29578893962455605, + "grad_norm": 5.15625, + "learning_rate": 9.7402756679092e-06, + "loss": 1.67011795, + "memory(GiB)": 111.15, + "step": 11660, + "train_speed(iter/s)": 0.449572 + }, + { + "acc": 0.63881364, + "epoch": 0.29591577879249115, + "grad_norm": 5.5, + "learning_rate": 9.739941990118157e-06, + "loss": 1.63815231, + "memory(GiB)": 111.15, + "step": 11665, + "train_speed(iter/s)": 0.449674 + }, + { + "acc": 0.65184441, + "epoch": 0.2960426179604262, + "grad_norm": 4.5, + "learning_rate": 9.739608103843306e-06, + "loss": 1.64906082, + "memory(GiB)": 111.15, + "step": 11670, + "train_speed(iter/s)": 0.449776 + }, + { + "acc": 0.64520464, + "epoch": 0.29616945712836124, + "grad_norm": 4.8125, + "learning_rate": 9.739274009099328e-06, + "loss": 1.66979141, + "memory(GiB)": 111.15, + "step": 11675, + "train_speed(iter/s)": 0.449878 + }, + { + "acc": 0.65894494, + "epoch": 0.2962962962962963, + "grad_norm": 5.125, + "learning_rate": 9.738939705900922e-06, + "loss": 1.63615952, + "memory(GiB)": 111.15, + "step": 11680, + "train_speed(iter/s)": 0.44998 + }, + { + "acc": 0.65550718, + "epoch": 0.2964231354642314, + "grad_norm": 4.46875, + "learning_rate": 9.738605194262787e-06, + "loss": 1.62016964, + "memory(GiB)": 111.15, + "step": 11685, + "train_speed(iter/s)": 0.450082 + }, + { + "acc": 0.64233932, + "epoch": 0.2965499746321664, + "grad_norm": 4.71875, + "learning_rate": 9.738270474199641e-06, + "loss": 1.68456612, + "memory(GiB)": 111.15, + "step": 11690, + "train_speed(iter/s)": 0.450183 + }, + { + "acc": 0.63042569, + "epoch": 0.29667681380010147, + "grad_norm": 6.15625, + "learning_rate": 9.737935545726205e-06, + "loss": 1.70305901, + "memory(GiB)": 111.15, + "step": 11695, + "train_speed(iter/s)": 0.450285 + }, + { + "acc": 0.64669423, + "epoch": 0.2968036529680365, + "grad_norm": 5.1875, + "learning_rate": 9.737600408857208e-06, + "loss": 1.64072266, + "memory(GiB)": 111.15, + "step": 11700, + "train_speed(iter/s)": 0.450388 + }, + { + "epoch": 0.2968036529680365, + "eval_acc": 0.6376883765019257, + "eval_loss": 1.640526533126831, + "eval_runtime": 113.5937, + "eval_samples_per_second": 56.077, + "eval_steps_per_second": 28.039, + "step": 11700 + }, + { + "acc": 0.63824759, + "epoch": 0.2969304921359716, + "grad_norm": 5.65625, + "learning_rate": 9.737265063607395e-06, + "loss": 1.67866592, + "memory(GiB)": 111.15, + "step": 11705, + "train_speed(iter/s)": 0.448452 + }, + { + "acc": 0.64700489, + "epoch": 0.29705733130390666, + "grad_norm": 5.53125, + "learning_rate": 9.736929509991515e-06, + "loss": 1.64573975, + "memory(GiB)": 111.15, + "step": 11710, + "train_speed(iter/s)": 0.448552 + }, + { + "acc": 0.65277472, + "epoch": 0.2971841704718417, + "grad_norm": 5.75, + "learning_rate": 9.736593748024325e-06, + "loss": 1.67160969, + "memory(GiB)": 111.15, + "step": 11715, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.65333595, + "epoch": 0.29731100963977675, + "grad_norm": 4.65625, + "learning_rate": 9.736257777720595e-06, + "loss": 1.60299435, + "memory(GiB)": 111.15, + "step": 11720, + "train_speed(iter/s)": 0.448755 + }, + { + "acc": 0.6571207, + "epoch": 0.29743784880771185, + "grad_norm": 6.09375, + "learning_rate": 9.735921599095101e-06, + "loss": 1.63371067, + "memory(GiB)": 111.15, + "step": 11725, + "train_speed(iter/s)": 0.448856 + }, + { + "acc": 0.65891581, + "epoch": 0.2975646879756469, + "grad_norm": 4.84375, + "learning_rate": 9.735585212162633e-06, + "loss": 1.62234383, + "memory(GiB)": 111.15, + "step": 11730, + "train_speed(iter/s)": 0.448955 + }, + { + "acc": 0.63484511, + "epoch": 0.29769152714358194, + "grad_norm": 6.125, + "learning_rate": 9.735248616937983e-06, + "loss": 1.68009872, + "memory(GiB)": 111.15, + "step": 11735, + "train_speed(iter/s)": 0.449055 + }, + { + "acc": 0.6496913, + "epoch": 0.297818366311517, + "grad_norm": 4.625, + "learning_rate": 9.734911813435957e-06, + "loss": 1.63371468, + "memory(GiB)": 111.15, + "step": 11740, + "train_speed(iter/s)": 0.449155 + }, + { + "acc": 0.65094013, + "epoch": 0.2979452054794521, + "grad_norm": 4.6875, + "learning_rate": 9.73457480167137e-06, + "loss": 1.75572929, + "memory(GiB)": 111.15, + "step": 11745, + "train_speed(iter/s)": 0.449256 + }, + { + "acc": 0.65438247, + "epoch": 0.2980720446473871, + "grad_norm": 5.53125, + "learning_rate": 9.734237581659045e-06, + "loss": 1.66425533, + "memory(GiB)": 111.15, + "step": 11750, + "train_speed(iter/s)": 0.449356 + }, + { + "acc": 0.64325409, + "epoch": 0.29819888381532217, + "grad_norm": 5.25, + "learning_rate": 9.733900153413813e-06, + "loss": 1.6219595, + "memory(GiB)": 111.15, + "step": 11755, + "train_speed(iter/s)": 0.449457 + }, + { + "acc": 0.64971428, + "epoch": 0.2983257229832572, + "grad_norm": 5.53125, + "learning_rate": 9.733562516950519e-06, + "loss": 1.62884407, + "memory(GiB)": 111.15, + "step": 11760, + "train_speed(iter/s)": 0.449557 + }, + { + "acc": 0.63364954, + "epoch": 0.2984525621511923, + "grad_norm": 5.8125, + "learning_rate": 9.73322467228401e-06, + "loss": 1.65487595, + "memory(GiB)": 111.15, + "step": 11765, + "train_speed(iter/s)": 0.449658 + }, + { + "acc": 0.64463124, + "epoch": 0.29857940131912736, + "grad_norm": 6.65625, + "learning_rate": 9.73288661942915e-06, + "loss": 1.63990517, + "memory(GiB)": 111.15, + "step": 11770, + "train_speed(iter/s)": 0.449758 + }, + { + "acc": 0.65357952, + "epoch": 0.2987062404870624, + "grad_norm": 5.90625, + "learning_rate": 9.732548358400802e-06, + "loss": 1.62151356, + "memory(GiB)": 111.15, + "step": 11775, + "train_speed(iter/s)": 0.449858 + }, + { + "acc": 0.65091891, + "epoch": 0.29883307965499745, + "grad_norm": 4.71875, + "learning_rate": 9.73220988921385e-06, + "loss": 1.60483646, + "memory(GiB)": 111.15, + "step": 11780, + "train_speed(iter/s)": 0.449958 + }, + { + "acc": 0.64542446, + "epoch": 0.29895991882293255, + "grad_norm": 4.84375, + "learning_rate": 9.73187121188318e-06, + "loss": 1.6259985, + "memory(GiB)": 111.15, + "step": 11785, + "train_speed(iter/s)": 0.450059 + }, + { + "acc": 0.65885277, + "epoch": 0.2990867579908676, + "grad_norm": 6.34375, + "learning_rate": 9.731532326423686e-06, + "loss": 1.62264538, + "memory(GiB)": 111.15, + "step": 11790, + "train_speed(iter/s)": 0.450159 + }, + { + "acc": 0.65812302, + "epoch": 0.29921359715880264, + "grad_norm": 4.65625, + "learning_rate": 9.731193232850277e-06, + "loss": 1.60783558, + "memory(GiB)": 111.15, + "step": 11795, + "train_speed(iter/s)": 0.450259 + }, + { + "acc": 0.63924456, + "epoch": 0.2993404363267377, + "grad_norm": 4.59375, + "learning_rate": 9.730853931177866e-06, + "loss": 1.68779984, + "memory(GiB)": 111.15, + "step": 11800, + "train_speed(iter/s)": 0.450358 + }, + { + "epoch": 0.2993404363267377, + "eval_acc": 0.6377614857810807, + "eval_loss": 1.6401467323303223, + "eval_runtime": 113.6777, + "eval_samples_per_second": 56.036, + "eval_steps_per_second": 28.018, + "step": 11800 + }, + { + "acc": 0.6511086, + "epoch": 0.2994672754946728, + "grad_norm": 5.96875, + "learning_rate": 9.730514421421378e-06, + "loss": 1.62002869, + "memory(GiB)": 111.15, + "step": 11805, + "train_speed(iter/s)": 0.448435 + }, + { + "acc": 0.63997355, + "epoch": 0.2995941146626078, + "grad_norm": 4.40625, + "learning_rate": 9.730174703595745e-06, + "loss": 1.66207867, + "memory(GiB)": 111.15, + "step": 11810, + "train_speed(iter/s)": 0.448532 + }, + { + "acc": 0.64688449, + "epoch": 0.29972095383054287, + "grad_norm": 4.875, + "learning_rate": 9.72983477771591e-06, + "loss": 1.60341206, + "memory(GiB)": 111.15, + "step": 11815, + "train_speed(iter/s)": 0.448629 + }, + { + "acc": 0.63591094, + "epoch": 0.2998477929984779, + "grad_norm": 4.75, + "learning_rate": 9.729494643796823e-06, + "loss": 1.73419876, + "memory(GiB)": 111.15, + "step": 11820, + "train_speed(iter/s)": 0.448728 + }, + { + "acc": 0.6449708, + "epoch": 0.299974632166413, + "grad_norm": 4.90625, + "learning_rate": 9.729154301853448e-06, + "loss": 1.64082012, + "memory(GiB)": 111.15, + "step": 11825, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.64127045, + "epoch": 0.30010147133434806, + "grad_norm": 7.375, + "learning_rate": 9.72881375190075e-06, + "loss": 1.71435051, + "memory(GiB)": 111.15, + "step": 11830, + "train_speed(iter/s)": 0.448925 + }, + { + "acc": 0.65814962, + "epoch": 0.3002283105022831, + "grad_norm": 4.5, + "learning_rate": 9.728472993953712e-06, + "loss": 1.62002068, + "memory(GiB)": 111.15, + "step": 11835, + "train_speed(iter/s)": 0.449024 + }, + { + "acc": 0.64119005, + "epoch": 0.30035514967021815, + "grad_norm": 4.5625, + "learning_rate": 9.728132028027323e-06, + "loss": 1.59561958, + "memory(GiB)": 111.15, + "step": 11840, + "train_speed(iter/s)": 0.449123 + }, + { + "acc": 0.66369143, + "epoch": 0.30048198883815325, + "grad_norm": 4.65625, + "learning_rate": 9.727790854136573e-06, + "loss": 1.67817039, + "memory(GiB)": 111.15, + "step": 11845, + "train_speed(iter/s)": 0.449222 + }, + { + "acc": 0.65159359, + "epoch": 0.3006088280060883, + "grad_norm": 4.5, + "learning_rate": 9.727449472296476e-06, + "loss": 1.58684721, + "memory(GiB)": 111.15, + "step": 11850, + "train_speed(iter/s)": 0.449322 + }, + { + "acc": 0.64275713, + "epoch": 0.30073566717402334, + "grad_norm": 5.96875, + "learning_rate": 9.727107882522045e-06, + "loss": 1.70808487, + "memory(GiB)": 111.15, + "step": 11855, + "train_speed(iter/s)": 0.449421 + }, + { + "acc": 0.65096769, + "epoch": 0.3008625063419584, + "grad_norm": 5.5, + "learning_rate": 9.726766084828303e-06, + "loss": 1.62070122, + "memory(GiB)": 111.15, + "step": 11860, + "train_speed(iter/s)": 0.449519 + }, + { + "acc": 0.63954062, + "epoch": 0.3009893455098935, + "grad_norm": 5.15625, + "learning_rate": 9.726424079230286e-06, + "loss": 1.64515018, + "memory(GiB)": 111.15, + "step": 11865, + "train_speed(iter/s)": 0.449616 + }, + { + "acc": 0.64372387, + "epoch": 0.3011161846778285, + "grad_norm": 5.09375, + "learning_rate": 9.726081865743036e-06, + "loss": 1.65789051, + "memory(GiB)": 111.15, + "step": 11870, + "train_speed(iter/s)": 0.449716 + }, + { + "acc": 0.66590734, + "epoch": 0.30124302384576357, + "grad_norm": 4.84375, + "learning_rate": 9.725739444381603e-06, + "loss": 1.64140377, + "memory(GiB)": 111.15, + "step": 11875, + "train_speed(iter/s)": 0.449815 + }, + { + "acc": 0.64857073, + "epoch": 0.3013698630136986, + "grad_norm": 5.0, + "learning_rate": 9.725396815161053e-06, + "loss": 1.67433014, + "memory(GiB)": 111.15, + "step": 11880, + "train_speed(iter/s)": 0.449914 + }, + { + "acc": 0.6443572, + "epoch": 0.3014967021816337, + "grad_norm": 5.28125, + "learning_rate": 9.725053978096453e-06, + "loss": 1.71241341, + "memory(GiB)": 111.15, + "step": 11885, + "train_speed(iter/s)": 0.450012 + }, + { + "acc": 0.63792281, + "epoch": 0.30162354134956876, + "grad_norm": 5.03125, + "learning_rate": 9.724710933202884e-06, + "loss": 1.7287796, + "memory(GiB)": 111.15, + "step": 11890, + "train_speed(iter/s)": 0.450111 + }, + { + "acc": 0.64801517, + "epoch": 0.3017503805175038, + "grad_norm": 5.09375, + "learning_rate": 9.724367680495432e-06, + "loss": 1.66857796, + "memory(GiB)": 111.15, + "step": 11895, + "train_speed(iter/s)": 0.450209 + }, + { + "acc": 0.65664587, + "epoch": 0.30187721968543885, + "grad_norm": 5.125, + "learning_rate": 9.724024219989198e-06, + "loss": 1.58443737, + "memory(GiB)": 111.15, + "step": 11900, + "train_speed(iter/s)": 0.450308 + }, + { + "epoch": 0.30187721968543885, + "eval_acc": 0.6377021628231377, + "eval_loss": 1.6403369903564453, + "eval_runtime": 113.5819, + "eval_samples_per_second": 56.083, + "eval_steps_per_second": 28.041, + "step": 11900 + }, + { + "acc": 0.63888397, + "epoch": 0.30200405885337395, + "grad_norm": 5.75, + "learning_rate": 9.723680551699286e-06, + "loss": 1.64306831, + "memory(GiB)": 111.15, + "step": 11905, + "train_speed(iter/s)": 0.448403 + }, + { + "acc": 0.63390312, + "epoch": 0.302130898021309, + "grad_norm": 5.75, + "learning_rate": 9.723336675640815e-06, + "loss": 1.73192673, + "memory(GiB)": 111.15, + "step": 11910, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.64023352, + "epoch": 0.30225773718924404, + "grad_norm": 4.9375, + "learning_rate": 9.722992591828908e-06, + "loss": 1.64327412, + "memory(GiB)": 111.15, + "step": 11915, + "train_speed(iter/s)": 0.4486 + }, + { + "acc": 0.65822115, + "epoch": 0.3023845763571791, + "grad_norm": 5.5, + "learning_rate": 9.722648300278701e-06, + "loss": 1.57156754, + "memory(GiB)": 111.15, + "step": 11920, + "train_speed(iter/s)": 0.448698 + }, + { + "acc": 0.66270533, + "epoch": 0.3025114155251142, + "grad_norm": 5.0, + "learning_rate": 9.722303801005338e-06, + "loss": 1.65449562, + "memory(GiB)": 111.15, + "step": 11925, + "train_speed(iter/s)": 0.448796 + }, + { + "acc": 0.64662967, + "epoch": 0.3026382546930492, + "grad_norm": 6.0625, + "learning_rate": 9.721959094023968e-06, + "loss": 1.70118828, + "memory(GiB)": 111.15, + "step": 11930, + "train_speed(iter/s)": 0.448895 + }, + { + "acc": 0.6626874, + "epoch": 0.30276509386098427, + "grad_norm": 5.9375, + "learning_rate": 9.721614179349754e-06, + "loss": 1.57054834, + "memory(GiB)": 111.15, + "step": 11935, + "train_speed(iter/s)": 0.448993 + }, + { + "acc": 0.65682116, + "epoch": 0.3028919330289193, + "grad_norm": 5.65625, + "learning_rate": 9.72126905699787e-06, + "loss": 1.67868423, + "memory(GiB)": 111.15, + "step": 11940, + "train_speed(iter/s)": 0.449092 + }, + { + "acc": 0.63211412, + "epoch": 0.3030187721968544, + "grad_norm": 4.625, + "learning_rate": 9.720923726983493e-06, + "loss": 1.63951874, + "memory(GiB)": 111.15, + "step": 11945, + "train_speed(iter/s)": 0.44919 + }, + { + "acc": 0.64202328, + "epoch": 0.30314561136478946, + "grad_norm": 4.71875, + "learning_rate": 9.720578189321814e-06, + "loss": 1.64672623, + "memory(GiB)": 111.15, + "step": 11950, + "train_speed(iter/s)": 0.449288 + }, + { + "acc": 0.65513144, + "epoch": 0.3032724505327245, + "grad_norm": 6.21875, + "learning_rate": 9.72023244402803e-06, + "loss": 1.63461227, + "memory(GiB)": 111.15, + "step": 11955, + "train_speed(iter/s)": 0.449386 + }, + { + "acc": 0.64464426, + "epoch": 0.30339928970065955, + "grad_norm": 5.1875, + "learning_rate": 9.719886491117348e-06, + "loss": 1.65310287, + "memory(GiB)": 111.15, + "step": 11960, + "train_speed(iter/s)": 0.449485 + }, + { + "acc": 0.64655538, + "epoch": 0.30352612886859465, + "grad_norm": 5.625, + "learning_rate": 9.719540330604986e-06, + "loss": 1.58601408, + "memory(GiB)": 111.15, + "step": 11965, + "train_speed(iter/s)": 0.449583 + }, + { + "acc": 0.65397029, + "epoch": 0.3036529680365297, + "grad_norm": 5.59375, + "learning_rate": 9.71919396250617e-06, + "loss": 1.66997547, + "memory(GiB)": 111.15, + "step": 11970, + "train_speed(iter/s)": 0.449681 + }, + { + "acc": 0.62513132, + "epoch": 0.30377980720446474, + "grad_norm": 5.09375, + "learning_rate": 9.718847386836131e-06, + "loss": 1.72465439, + "memory(GiB)": 111.15, + "step": 11975, + "train_speed(iter/s)": 0.449779 + }, + { + "acc": 0.64613333, + "epoch": 0.3039066463723998, + "grad_norm": 5.15625, + "learning_rate": 9.718500603610119e-06, + "loss": 1.68342171, + "memory(GiB)": 111.15, + "step": 11980, + "train_speed(iter/s)": 0.449878 + }, + { + "acc": 0.63610115, + "epoch": 0.3040334855403349, + "grad_norm": 5.75, + "learning_rate": 9.718153612843382e-06, + "loss": 1.68873291, + "memory(GiB)": 111.15, + "step": 11985, + "train_speed(iter/s)": 0.449976 + }, + { + "acc": 0.63843594, + "epoch": 0.3041603247082699, + "grad_norm": 4.59375, + "learning_rate": 9.717806414551186e-06, + "loss": 1.70086899, + "memory(GiB)": 111.15, + "step": 11990, + "train_speed(iter/s)": 0.450074 + }, + { + "acc": 0.65368214, + "epoch": 0.30428716387620497, + "grad_norm": 6.6875, + "learning_rate": 9.717459008748798e-06, + "loss": 1.64928856, + "memory(GiB)": 111.15, + "step": 11995, + "train_speed(iter/s)": 0.450172 + }, + { + "acc": 0.64351168, + "epoch": 0.30441400304414, + "grad_norm": 5.875, + "learning_rate": 9.717111395451501e-06, + "loss": 1.63215752, + "memory(GiB)": 111.15, + "step": 12000, + "train_speed(iter/s)": 0.450272 + }, + { + "epoch": 0.30441400304414, + "eval_acc": 0.6379365302837434, + "eval_loss": 1.6398106813430786, + "eval_runtime": 113.7597, + "eval_samples_per_second": 55.995, + "eval_steps_per_second": 27.998, + "step": 12000 + }, + { + "acc": 0.63126626, + "epoch": 0.3045408422120751, + "grad_norm": 4.71875, + "learning_rate": 9.716763574674586e-06, + "loss": 1.73478889, + "memory(GiB)": 111.15, + "step": 12005, + "train_speed(iter/s)": 0.448382 + }, + { + "acc": 0.65735931, + "epoch": 0.30466768138001016, + "grad_norm": 5.0625, + "learning_rate": 9.71641554643335e-06, + "loss": 1.57218084, + "memory(GiB)": 111.15, + "step": 12010, + "train_speed(iter/s)": 0.448478 + }, + { + "acc": 0.6517272, + "epoch": 0.3047945205479452, + "grad_norm": 5.28125, + "learning_rate": 9.7160673107431e-06, + "loss": 1.59773941, + "memory(GiB)": 111.15, + "step": 12015, + "train_speed(iter/s)": 0.448576 + }, + { + "acc": 0.62148085, + "epoch": 0.30492135971588025, + "grad_norm": 6.09375, + "learning_rate": 9.715718867619155e-06, + "loss": 1.74924793, + "memory(GiB)": 111.15, + "step": 12020, + "train_speed(iter/s)": 0.448674 + }, + { + "acc": 0.6463058, + "epoch": 0.30504819888381535, + "grad_norm": 4.71875, + "learning_rate": 9.715370217076838e-06, + "loss": 1.63156624, + "memory(GiB)": 111.15, + "step": 12025, + "train_speed(iter/s)": 0.448772 + }, + { + "acc": 0.63073778, + "epoch": 0.3051750380517504, + "grad_norm": 5.375, + "learning_rate": 9.715021359131489e-06, + "loss": 1.69368858, + "memory(GiB)": 111.15, + "step": 12030, + "train_speed(iter/s)": 0.448869 + }, + { + "acc": 0.6340353, + "epoch": 0.30530187721968544, + "grad_norm": 5.8125, + "learning_rate": 9.714672293798449e-06, + "loss": 1.71521568, + "memory(GiB)": 111.15, + "step": 12035, + "train_speed(iter/s)": 0.448968 + }, + { + "acc": 0.6300971, + "epoch": 0.3054287163876205, + "grad_norm": 4.5625, + "learning_rate": 9.71432302109307e-06, + "loss": 1.70242882, + "memory(GiB)": 111.15, + "step": 12040, + "train_speed(iter/s)": 0.449065 + }, + { + "acc": 0.62911949, + "epoch": 0.3055555555555556, + "grad_norm": 6.0625, + "learning_rate": 9.713973541030716e-06, + "loss": 1.66565475, + "memory(GiB)": 111.15, + "step": 12045, + "train_speed(iter/s)": 0.449162 + }, + { + "acc": 0.64725952, + "epoch": 0.3056823947234906, + "grad_norm": 5.28125, + "learning_rate": 9.713623853626763e-06, + "loss": 1.60451374, + "memory(GiB)": 111.15, + "step": 12050, + "train_speed(iter/s)": 0.44926 + }, + { + "acc": 0.65198832, + "epoch": 0.30580923389142567, + "grad_norm": 5.0625, + "learning_rate": 9.713273958896586e-06, + "loss": 1.66147537, + "memory(GiB)": 111.15, + "step": 12055, + "train_speed(iter/s)": 0.449357 + }, + { + "acc": 0.6500145, + "epoch": 0.3059360730593607, + "grad_norm": 4.53125, + "learning_rate": 9.712923856855578e-06, + "loss": 1.59981766, + "memory(GiB)": 111.15, + "step": 12060, + "train_speed(iter/s)": 0.449455 + }, + { + "acc": 0.63221989, + "epoch": 0.3060629122272958, + "grad_norm": 5.65625, + "learning_rate": 9.712573547519134e-06, + "loss": 1.71357021, + "memory(GiB)": 111.15, + "step": 12065, + "train_speed(iter/s)": 0.449553 + }, + { + "acc": 0.63088374, + "epoch": 0.30618975139523086, + "grad_norm": 6.3125, + "learning_rate": 9.712223030902668e-06, + "loss": 1.74134827, + "memory(GiB)": 111.15, + "step": 12070, + "train_speed(iter/s)": 0.44965 + }, + { + "acc": 0.65225515, + "epoch": 0.3063165905631659, + "grad_norm": 5.78125, + "learning_rate": 9.711872307021594e-06, + "loss": 1.65614586, + "memory(GiB)": 111.15, + "step": 12075, + "train_speed(iter/s)": 0.449747 + }, + { + "acc": 0.64735193, + "epoch": 0.30644342973110095, + "grad_norm": 5.625, + "learning_rate": 9.711521375891339e-06, + "loss": 1.67787609, + "memory(GiB)": 111.15, + "step": 12080, + "train_speed(iter/s)": 0.449844 + }, + { + "acc": 0.64399862, + "epoch": 0.30657026889903605, + "grad_norm": 5.34375, + "learning_rate": 9.71117023752734e-06, + "loss": 1.67920876, + "memory(GiB)": 111.15, + "step": 12085, + "train_speed(iter/s)": 0.449941 + }, + { + "acc": 0.6430006, + "epoch": 0.3066971080669711, + "grad_norm": 5.4375, + "learning_rate": 9.710818891945037e-06, + "loss": 1.61480503, + "memory(GiB)": 111.15, + "step": 12090, + "train_speed(iter/s)": 0.450037 + }, + { + "acc": 0.65409584, + "epoch": 0.30682394723490614, + "grad_norm": 5.28125, + "learning_rate": 9.71046733915989e-06, + "loss": 1.63893089, + "memory(GiB)": 111.15, + "step": 12095, + "train_speed(iter/s)": 0.450134 + }, + { + "acc": 0.64327707, + "epoch": 0.3069507864028412, + "grad_norm": 4.65625, + "learning_rate": 9.710115579187356e-06, + "loss": 1.64909363, + "memory(GiB)": 111.15, + "step": 12100, + "train_speed(iter/s)": 0.450232 + }, + { + "epoch": 0.3069507864028412, + "eval_acc": 0.6378630032372788, + "eval_loss": 1.6393685340881348, + "eval_runtime": 111.9765, + "eval_samples_per_second": 56.887, + "eval_steps_per_second": 28.443, + "step": 12100 + }, + { + "acc": 0.63206072, + "epoch": 0.3070776255707763, + "grad_norm": 5.84375, + "learning_rate": 9.709763612042911e-06, + "loss": 1.7169445, + "memory(GiB)": 111.15, + "step": 12105, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.64454517, + "epoch": 0.3072044647387113, + "grad_norm": 5.6875, + "learning_rate": 9.709411437742035e-06, + "loss": 1.58953419, + "memory(GiB)": 111.15, + "step": 12110, + "train_speed(iter/s)": 0.448483 + }, + { + "acc": 0.65365906, + "epoch": 0.30733130390664637, + "grad_norm": 6.25, + "learning_rate": 9.709059056300218e-06, + "loss": 1.61971188, + "memory(GiB)": 111.15, + "step": 12115, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.64947519, + "epoch": 0.3074581430745814, + "grad_norm": 5.09375, + "learning_rate": 9.708706467732958e-06, + "loss": 1.6222126, + "memory(GiB)": 111.15, + "step": 12120, + "train_speed(iter/s)": 0.448673 + }, + { + "acc": 0.64905167, + "epoch": 0.3075849822425165, + "grad_norm": 5.25, + "learning_rate": 9.708353672055766e-06, + "loss": 1.67694397, + "memory(GiB)": 111.15, + "step": 12125, + "train_speed(iter/s)": 0.44877 + }, + { + "acc": 0.64748888, + "epoch": 0.30771182141045156, + "grad_norm": 5.78125, + "learning_rate": 9.708000669284158e-06, + "loss": 1.69502411, + "memory(GiB)": 111.15, + "step": 12130, + "train_speed(iter/s)": 0.448867 + }, + { + "acc": 0.64687757, + "epoch": 0.3078386605783866, + "grad_norm": 5.84375, + "learning_rate": 9.707647459433661e-06, + "loss": 1.67606163, + "memory(GiB)": 111.15, + "step": 12135, + "train_speed(iter/s)": 0.448964 + }, + { + "acc": 0.64536748, + "epoch": 0.30796549974632165, + "grad_norm": 5.28125, + "learning_rate": 9.707294042519808e-06, + "loss": 1.62308407, + "memory(GiB)": 111.15, + "step": 12140, + "train_speed(iter/s)": 0.449061 + }, + { + "acc": 0.65040302, + "epoch": 0.30809233891425675, + "grad_norm": 5.15625, + "learning_rate": 9.70694041855815e-06, + "loss": 1.72639656, + "memory(GiB)": 111.15, + "step": 12145, + "train_speed(iter/s)": 0.449158 + }, + { + "acc": 0.64523211, + "epoch": 0.3082191780821918, + "grad_norm": 4.59375, + "learning_rate": 9.706586587564236e-06, + "loss": 1.7079071, + "memory(GiB)": 111.15, + "step": 12150, + "train_speed(iter/s)": 0.449253 + }, + { + "acc": 0.64028225, + "epoch": 0.30834601725012684, + "grad_norm": 5.5625, + "learning_rate": 9.70623254955363e-06, + "loss": 1.66565628, + "memory(GiB)": 111.15, + "step": 12155, + "train_speed(iter/s)": 0.449349 + }, + { + "acc": 0.65672913, + "epoch": 0.3084728564180619, + "grad_norm": 7.46875, + "learning_rate": 9.705878304541905e-06, + "loss": 1.58745451, + "memory(GiB)": 111.15, + "step": 12160, + "train_speed(iter/s)": 0.449446 + }, + { + "acc": 0.64548979, + "epoch": 0.308599695585997, + "grad_norm": 5.15625, + "learning_rate": 9.705523852544643e-06, + "loss": 1.65273056, + "memory(GiB)": 111.15, + "step": 12165, + "train_speed(iter/s)": 0.449543 + }, + { + "acc": 0.63662605, + "epoch": 0.308726534753932, + "grad_norm": 6.25, + "learning_rate": 9.705169193577434e-06, + "loss": 1.70854816, + "memory(GiB)": 111.15, + "step": 12170, + "train_speed(iter/s)": 0.449639 + }, + { + "acc": 0.64997644, + "epoch": 0.30885337392186707, + "grad_norm": 5.78125, + "learning_rate": 9.704814327655874e-06, + "loss": 1.64368496, + "memory(GiB)": 111.15, + "step": 12175, + "train_speed(iter/s)": 0.449736 + }, + { + "acc": 0.65820026, + "epoch": 0.3089802130898021, + "grad_norm": 5.3125, + "learning_rate": 9.704459254795575e-06, + "loss": 1.63453636, + "memory(GiB)": 111.15, + "step": 12180, + "train_speed(iter/s)": 0.449831 + }, + { + "acc": 0.6321578, + "epoch": 0.3091070522577372, + "grad_norm": 4.59375, + "learning_rate": 9.704103975012155e-06, + "loss": 1.71723938, + "memory(GiB)": 111.15, + "step": 12185, + "train_speed(iter/s)": 0.449928 + }, + { + "acc": 0.65098791, + "epoch": 0.30923389142567226, + "grad_norm": 5.0625, + "learning_rate": 9.703748488321238e-06, + "loss": 1.59529457, + "memory(GiB)": 111.15, + "step": 12190, + "train_speed(iter/s)": 0.450023 + }, + { + "acc": 0.67134185, + "epoch": 0.3093607305936073, + "grad_norm": 5.375, + "learning_rate": 9.703392794738464e-06, + "loss": 1.59396152, + "memory(GiB)": 111.15, + "step": 12195, + "train_speed(iter/s)": 0.450117 + }, + { + "acc": 0.64275575, + "epoch": 0.30948756976154235, + "grad_norm": 5.59375, + "learning_rate": 9.703036894279476e-06, + "loss": 1.69781876, + "memory(GiB)": 111.15, + "step": 12200, + "train_speed(iter/s)": 0.450212 + }, + { + "epoch": 0.30948756976154235, + "eval_acc": 0.6380956996286467, + "eval_loss": 1.6388194561004639, + "eval_runtime": 114.6282, + "eval_samples_per_second": 55.571, + "eval_steps_per_second": 27.785, + "step": 12200 + }, + { + "acc": 0.66670141, + "epoch": 0.30961440892947745, + "grad_norm": 6.125, + "learning_rate": 9.702680786959925e-06, + "loss": 1.57581778, + "memory(GiB)": 111.15, + "step": 12205, + "train_speed(iter/s)": 0.448337 + }, + { + "acc": 0.65414839, + "epoch": 0.3097412480974125, + "grad_norm": 5.0625, + "learning_rate": 9.702324472795481e-06, + "loss": 1.65087891, + "memory(GiB)": 111.15, + "step": 12210, + "train_speed(iter/s)": 0.448432 + }, + { + "acc": 0.63975921, + "epoch": 0.30986808726534754, + "grad_norm": 5.0, + "learning_rate": 9.70196795180181e-06, + "loss": 1.65003872, + "memory(GiB)": 111.15, + "step": 12215, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.64131489, + "epoch": 0.3099949264332826, + "grad_norm": 5.6875, + "learning_rate": 9.701611223994596e-06, + "loss": 1.62780056, + "memory(GiB)": 111.15, + "step": 12220, + "train_speed(iter/s)": 0.448623 + }, + { + "acc": 0.64259577, + "epoch": 0.3101217656012177, + "grad_norm": 4.65625, + "learning_rate": 9.701254289389529e-06, + "loss": 1.63561344, + "memory(GiB)": 111.15, + "step": 12225, + "train_speed(iter/s)": 0.44872 + }, + { + "acc": 0.64992857, + "epoch": 0.3102486047691527, + "grad_norm": 5.375, + "learning_rate": 9.700897148002308e-06, + "loss": 1.67261581, + "memory(GiB)": 111.15, + "step": 12230, + "train_speed(iter/s)": 0.448815 + }, + { + "acc": 0.64694147, + "epoch": 0.31037544393708777, + "grad_norm": 5.25, + "learning_rate": 9.700539799848645e-06, + "loss": 1.68389206, + "memory(GiB)": 111.15, + "step": 12235, + "train_speed(iter/s)": 0.448911 + }, + { + "acc": 0.64467955, + "epoch": 0.3105022831050228, + "grad_norm": 4.8125, + "learning_rate": 9.700182244944252e-06, + "loss": 1.63858376, + "memory(GiB)": 111.15, + "step": 12240, + "train_speed(iter/s)": 0.449006 + }, + { + "acc": 0.63759542, + "epoch": 0.3106291222729579, + "grad_norm": 5.21875, + "learning_rate": 9.69982448330486e-06, + "loss": 1.68621025, + "memory(GiB)": 111.15, + "step": 12245, + "train_speed(iter/s)": 0.449101 + }, + { + "acc": 0.63918104, + "epoch": 0.31075596144089296, + "grad_norm": 5.78125, + "learning_rate": 9.699466514946206e-06, + "loss": 1.78194199, + "memory(GiB)": 111.15, + "step": 12250, + "train_speed(iter/s)": 0.449197 + }, + { + "acc": 0.66187096, + "epoch": 0.310882800608828, + "grad_norm": 6.6875, + "learning_rate": 9.699108339884032e-06, + "loss": 1.5962779, + "memory(GiB)": 111.15, + "step": 12255, + "train_speed(iter/s)": 0.449292 + }, + { + "acc": 0.62815957, + "epoch": 0.31100963977676305, + "grad_norm": 5.25, + "learning_rate": 9.698749958134093e-06, + "loss": 1.709412, + "memory(GiB)": 111.15, + "step": 12260, + "train_speed(iter/s)": 0.449388 + }, + { + "acc": 0.63694048, + "epoch": 0.31113647894469815, + "grad_norm": 5.78125, + "learning_rate": 9.698391369712152e-06, + "loss": 1.74542732, + "memory(GiB)": 111.15, + "step": 12265, + "train_speed(iter/s)": 0.449485 + }, + { + "acc": 0.6515121, + "epoch": 0.3112633181126332, + "grad_norm": 5.3125, + "learning_rate": 9.698032574633982e-06, + "loss": 1.62864571, + "memory(GiB)": 111.15, + "step": 12270, + "train_speed(iter/s)": 0.449581 + }, + { + "acc": 0.64555387, + "epoch": 0.31139015728056824, + "grad_norm": 4.9375, + "learning_rate": 9.697673572915364e-06, + "loss": 1.65919189, + "memory(GiB)": 111.15, + "step": 12275, + "train_speed(iter/s)": 0.449677 + }, + { + "acc": 0.64611149, + "epoch": 0.3115169964485033, + "grad_norm": 5.28125, + "learning_rate": 9.697314364572087e-06, + "loss": 1.67409039, + "memory(GiB)": 111.15, + "step": 12280, + "train_speed(iter/s)": 0.449773 + }, + { + "acc": 0.65357752, + "epoch": 0.3116438356164384, + "grad_norm": 5.25, + "learning_rate": 9.696954949619955e-06, + "loss": 1.68172207, + "memory(GiB)": 111.15, + "step": 12285, + "train_speed(iter/s)": 0.449869 + }, + { + "acc": 0.65429983, + "epoch": 0.3117706747843734, + "grad_norm": 4.6875, + "learning_rate": 9.696595328074774e-06, + "loss": 1.63068848, + "memory(GiB)": 111.15, + "step": 12290, + "train_speed(iter/s)": 0.449965 + }, + { + "acc": 0.65027237, + "epoch": 0.31189751395230847, + "grad_norm": 5.21875, + "learning_rate": 9.69623549995236e-06, + "loss": 1.66726665, + "memory(GiB)": 111.15, + "step": 12295, + "train_speed(iter/s)": 0.450062 + }, + { + "acc": 0.63774481, + "epoch": 0.3120243531202435, + "grad_norm": 5.625, + "learning_rate": 9.695875465268543e-06, + "loss": 1.68268318, + "memory(GiB)": 111.15, + "step": 12300, + "train_speed(iter/s)": 0.450158 + }, + { + "epoch": 0.3120243531202435, + "eval_acc": 0.6380873442824575, + "eval_loss": 1.6378207206726074, + "eval_runtime": 112.1795, + "eval_samples_per_second": 56.784, + "eval_steps_per_second": 28.392, + "step": 12300 + }, + { + "acc": 0.6519865, + "epoch": 0.3121511922881786, + "grad_norm": 6.28125, + "learning_rate": 9.695515224039156e-06, + "loss": 1.67220707, + "memory(GiB)": 111.15, + "step": 12305, + "train_speed(iter/s)": 0.44834 + }, + { + "acc": 0.64857898, + "epoch": 0.31227803145611366, + "grad_norm": 5.53125, + "learning_rate": 9.695154776280047e-06, + "loss": 1.65074673, + "memory(GiB)": 111.15, + "step": 12310, + "train_speed(iter/s)": 0.448436 + }, + { + "acc": 0.63974056, + "epoch": 0.3124048706240487, + "grad_norm": 5.375, + "learning_rate": 9.694794122007067e-06, + "loss": 1.70130463, + "memory(GiB)": 111.15, + "step": 12315, + "train_speed(iter/s)": 0.448531 + }, + { + "acc": 0.64638128, + "epoch": 0.31253170979198375, + "grad_norm": 4.9375, + "learning_rate": 9.694433261236083e-06, + "loss": 1.65410366, + "memory(GiB)": 111.15, + "step": 12320, + "train_speed(iter/s)": 0.448626 + }, + { + "acc": 0.63894725, + "epoch": 0.31265854895991885, + "grad_norm": 5.40625, + "learning_rate": 9.694072193982962e-06, + "loss": 1.73535461, + "memory(GiB)": 111.15, + "step": 12325, + "train_speed(iter/s)": 0.448721 + }, + { + "acc": 0.64942036, + "epoch": 0.3127853881278539, + "grad_norm": 5.46875, + "learning_rate": 9.69371092026359e-06, + "loss": 1.65363121, + "memory(GiB)": 111.15, + "step": 12330, + "train_speed(iter/s)": 0.448816 + }, + { + "acc": 0.67198801, + "epoch": 0.31291222729578894, + "grad_norm": 5.15625, + "learning_rate": 9.693349440093855e-06, + "loss": 1.59727898, + "memory(GiB)": 111.15, + "step": 12335, + "train_speed(iter/s)": 0.448913 + }, + { + "acc": 0.66462412, + "epoch": 0.313039066463724, + "grad_norm": 4.65625, + "learning_rate": 9.69298775348966e-06, + "loss": 1.54071484, + "memory(GiB)": 111.15, + "step": 12340, + "train_speed(iter/s)": 0.449009 + }, + { + "acc": 0.64421482, + "epoch": 0.3131659056316591, + "grad_norm": 5.71875, + "learning_rate": 9.69262586046691e-06, + "loss": 1.6330759, + "memory(GiB)": 111.15, + "step": 12345, + "train_speed(iter/s)": 0.449105 + }, + { + "acc": 0.65639176, + "epoch": 0.3132927447995941, + "grad_norm": 5.625, + "learning_rate": 9.692263761041521e-06, + "loss": 1.66559887, + "memory(GiB)": 111.15, + "step": 12350, + "train_speed(iter/s)": 0.449202 + }, + { + "acc": 0.64486132, + "epoch": 0.31341958396752917, + "grad_norm": 4.4375, + "learning_rate": 9.691901455229425e-06, + "loss": 1.67320061, + "memory(GiB)": 111.15, + "step": 12355, + "train_speed(iter/s)": 0.449299 + }, + { + "acc": 0.64498558, + "epoch": 0.3135464231354642, + "grad_norm": 5.125, + "learning_rate": 9.691538943046552e-06, + "loss": 1.66187878, + "memory(GiB)": 111.15, + "step": 12360, + "train_speed(iter/s)": 0.449396 + }, + { + "acc": 0.64981194, + "epoch": 0.3136732623033993, + "grad_norm": 4.625, + "learning_rate": 9.691176224508853e-06, + "loss": 1.68947449, + "memory(GiB)": 111.15, + "step": 12365, + "train_speed(iter/s)": 0.449492 + }, + { + "acc": 0.6405592, + "epoch": 0.31380010147133436, + "grad_norm": 5.4375, + "learning_rate": 9.690813299632278e-06, + "loss": 1.64771156, + "memory(GiB)": 111.15, + "step": 12370, + "train_speed(iter/s)": 0.449589 + }, + { + "acc": 0.65196357, + "epoch": 0.3139269406392694, + "grad_norm": 5.25, + "learning_rate": 9.690450168432793e-06, + "loss": 1.68153496, + "memory(GiB)": 111.15, + "step": 12375, + "train_speed(iter/s)": 0.449685 + }, + { + "acc": 0.63152599, + "epoch": 0.31405377980720445, + "grad_norm": 4.84375, + "learning_rate": 9.690086830926366e-06, + "loss": 1.67459259, + "memory(GiB)": 111.15, + "step": 12380, + "train_speed(iter/s)": 0.449782 + }, + { + "acc": 0.62347646, + "epoch": 0.31418061897513955, + "grad_norm": 4.875, + "learning_rate": 9.689723287128981e-06, + "loss": 1.69290962, + "memory(GiB)": 111.15, + "step": 12385, + "train_speed(iter/s)": 0.449879 + }, + { + "acc": 0.64841156, + "epoch": 0.3143074581430746, + "grad_norm": 5.4375, + "learning_rate": 9.689359537056628e-06, + "loss": 1.65416145, + "memory(GiB)": 111.15, + "step": 12390, + "train_speed(iter/s)": 0.449976 + }, + { + "acc": 0.64357009, + "epoch": 0.31443429731100964, + "grad_norm": 5.84375, + "learning_rate": 9.688995580725304e-06, + "loss": 1.63249969, + "memory(GiB)": 111.15, + "step": 12395, + "train_speed(iter/s)": 0.450073 + }, + { + "acc": 0.64715986, + "epoch": 0.3145611364789447, + "grad_norm": 5.5625, + "learning_rate": 9.688631418151022e-06, + "loss": 1.61093616, + "memory(GiB)": 111.15, + "step": 12400, + "train_speed(iter/s)": 0.450169 + }, + { + "epoch": 0.3145611364789447, + "eval_acc": 0.6382030658271772, + "eval_loss": 1.6377592086791992, + "eval_runtime": 113.5274, + "eval_samples_per_second": 56.11, + "eval_steps_per_second": 28.055, + "step": 12400 + }, + { + "acc": 0.65611143, + "epoch": 0.3146879756468798, + "grad_norm": 5.5625, + "learning_rate": 9.688267049349796e-06, + "loss": 1.61102505, + "memory(GiB)": 111.15, + "step": 12405, + "train_speed(iter/s)": 0.448344 + }, + { + "acc": 0.63575783, + "epoch": 0.3148148148148148, + "grad_norm": 5.0, + "learning_rate": 9.687902474337654e-06, + "loss": 1.72477722, + "memory(GiB)": 111.15, + "step": 12410, + "train_speed(iter/s)": 0.448439 + }, + { + "acc": 0.6424655, + "epoch": 0.31494165398274987, + "grad_norm": 4.9375, + "learning_rate": 9.687537693130631e-06, + "loss": 1.61407776, + "memory(GiB)": 111.15, + "step": 12415, + "train_speed(iter/s)": 0.448535 + }, + { + "acc": 0.63960419, + "epoch": 0.3150684931506849, + "grad_norm": 5.15625, + "learning_rate": 9.687172705744773e-06, + "loss": 1.67211838, + "memory(GiB)": 111.15, + "step": 12420, + "train_speed(iter/s)": 0.44863 + }, + { + "acc": 0.63361168, + "epoch": 0.31519533231862, + "grad_norm": 5.4375, + "learning_rate": 9.686807512196132e-06, + "loss": 1.73157749, + "memory(GiB)": 111.15, + "step": 12425, + "train_speed(iter/s)": 0.448726 + }, + { + "acc": 0.64837651, + "epoch": 0.31532217148655506, + "grad_norm": 5.71875, + "learning_rate": 9.68644211250077e-06, + "loss": 1.61457081, + "memory(GiB)": 111.15, + "step": 12430, + "train_speed(iter/s)": 0.448823 + }, + { + "acc": 0.62967625, + "epoch": 0.3154490106544901, + "grad_norm": 5.4375, + "learning_rate": 9.686076506674761e-06, + "loss": 1.68003998, + "memory(GiB)": 111.15, + "step": 12435, + "train_speed(iter/s)": 0.448918 + }, + { + "acc": 0.63506322, + "epoch": 0.31557584982242515, + "grad_norm": 4.8125, + "learning_rate": 9.685710694734187e-06, + "loss": 1.69748344, + "memory(GiB)": 111.15, + "step": 12440, + "train_speed(iter/s)": 0.449014 + }, + { + "acc": 0.64497838, + "epoch": 0.31570268899036025, + "grad_norm": 5.53125, + "learning_rate": 9.685344676695135e-06, + "loss": 1.64053726, + "memory(GiB)": 111.15, + "step": 12445, + "train_speed(iter/s)": 0.44911 + }, + { + "acc": 0.64180107, + "epoch": 0.3158295281582953, + "grad_norm": 5.5625, + "learning_rate": 9.684978452573706e-06, + "loss": 1.70237026, + "memory(GiB)": 111.15, + "step": 12450, + "train_speed(iter/s)": 0.449205 + }, + { + "acc": 0.65929174, + "epoch": 0.31595636732623034, + "grad_norm": 4.65625, + "learning_rate": 9.684612022386008e-06, + "loss": 1.57421932, + "memory(GiB)": 111.15, + "step": 12455, + "train_speed(iter/s)": 0.449299 + }, + { + "acc": 0.63754921, + "epoch": 0.3160832064941654, + "grad_norm": 4.71875, + "learning_rate": 9.68424538614816e-06, + "loss": 1.67094002, + "memory(GiB)": 111.15, + "step": 12460, + "train_speed(iter/s)": 0.449395 + }, + { + "acc": 0.6456542, + "epoch": 0.3162100456621005, + "grad_norm": 5.0625, + "learning_rate": 9.683878543876281e-06, + "loss": 1.67341232, + "memory(GiB)": 111.15, + "step": 12465, + "train_speed(iter/s)": 0.44949 + }, + { + "acc": 0.65152187, + "epoch": 0.3163368848300355, + "grad_norm": 5.46875, + "learning_rate": 9.683511495586516e-06, + "loss": 1.62917557, + "memory(GiB)": 111.15, + "step": 12470, + "train_speed(iter/s)": 0.449586 + }, + { + "acc": 0.64967985, + "epoch": 0.31646372399797057, + "grad_norm": 5.25, + "learning_rate": 9.683144241295003e-06, + "loss": 1.60640774, + "memory(GiB)": 111.15, + "step": 12475, + "train_speed(iter/s)": 0.44968 + }, + { + "acc": 0.65012331, + "epoch": 0.3165905631659056, + "grad_norm": 6.21875, + "learning_rate": 9.682776781017899e-06, + "loss": 1.62433281, + "memory(GiB)": 111.15, + "step": 12480, + "train_speed(iter/s)": 0.449775 + }, + { + "acc": 0.63326378, + "epoch": 0.3167174023338407, + "grad_norm": 5.625, + "learning_rate": 9.682409114771364e-06, + "loss": 1.69007587, + "memory(GiB)": 111.15, + "step": 12485, + "train_speed(iter/s)": 0.449869 + }, + { + "acc": 0.63097239, + "epoch": 0.31684424150177576, + "grad_norm": 4.78125, + "learning_rate": 9.682041242571571e-06, + "loss": 1.67244854, + "memory(GiB)": 111.15, + "step": 12490, + "train_speed(iter/s)": 0.449964 + }, + { + "acc": 0.64923487, + "epoch": 0.3169710806697108, + "grad_norm": 5.90625, + "learning_rate": 9.681673164434701e-06, + "loss": 1.67347546, + "memory(GiB)": 111.15, + "step": 12495, + "train_speed(iter/s)": 0.450059 + }, + { + "acc": 0.65236993, + "epoch": 0.31709791983764585, + "grad_norm": 4.46875, + "learning_rate": 9.681304880376942e-06, + "loss": 1.59691343, + "memory(GiB)": 111.15, + "step": 12500, + "train_speed(iter/s)": 0.450153 + }, + { + "epoch": 0.31709791983764585, + "eval_acc": 0.6383233828123009, + "eval_loss": 1.6375657320022583, + "eval_runtime": 113.5589, + "eval_samples_per_second": 56.094, + "eval_steps_per_second": 28.047, + "step": 12500 + }, + { + "acc": 0.64003592, + "epoch": 0.31722475900558095, + "grad_norm": 5.3125, + "learning_rate": 9.680936390414495e-06, + "loss": 1.7397686, + "memory(GiB)": 111.15, + "step": 12505, + "train_speed(iter/s)": 0.448341 + }, + { + "acc": 0.64458485, + "epoch": 0.317351598173516, + "grad_norm": 6.09375, + "learning_rate": 9.680567694563566e-06, + "loss": 1.68002281, + "memory(GiB)": 111.15, + "step": 12510, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.64521823, + "epoch": 0.31747843734145104, + "grad_norm": 5.5625, + "learning_rate": 9.680198792840371e-06, + "loss": 1.66868973, + "memory(GiB)": 111.15, + "step": 12515, + "train_speed(iter/s)": 0.448526 + }, + { + "acc": 0.63747115, + "epoch": 0.3176052765093861, + "grad_norm": 5.4375, + "learning_rate": 9.67982968526114e-06, + "loss": 1.64109364, + "memory(GiB)": 111.15, + "step": 12520, + "train_speed(iter/s)": 0.448618 + }, + { + "acc": 0.65895777, + "epoch": 0.3177321156773212, + "grad_norm": 5.15625, + "learning_rate": 9.679460371842104e-06, + "loss": 1.61014729, + "memory(GiB)": 111.15, + "step": 12525, + "train_speed(iter/s)": 0.448712 + }, + { + "acc": 0.63992023, + "epoch": 0.3178589548452562, + "grad_norm": 5.34375, + "learning_rate": 9.679090852599508e-06, + "loss": 1.65574684, + "memory(GiB)": 111.15, + "step": 12530, + "train_speed(iter/s)": 0.448805 + }, + { + "acc": 0.63760061, + "epoch": 0.31798579401319127, + "grad_norm": 5.3125, + "learning_rate": 9.678721127549608e-06, + "loss": 1.69728069, + "memory(GiB)": 111.15, + "step": 12535, + "train_speed(iter/s)": 0.448899 + }, + { + "acc": 0.62461891, + "epoch": 0.3181126331811263, + "grad_norm": 6.46875, + "learning_rate": 9.678351196708662e-06, + "loss": 1.77512836, + "memory(GiB)": 111.15, + "step": 12540, + "train_speed(iter/s)": 0.448993 + }, + { + "acc": 0.64404902, + "epoch": 0.3182394723490614, + "grad_norm": 5.78125, + "learning_rate": 9.677981060092943e-06, + "loss": 1.71545658, + "memory(GiB)": 111.15, + "step": 12545, + "train_speed(iter/s)": 0.449087 + }, + { + "acc": 0.62593794, + "epoch": 0.31836631151699646, + "grad_norm": 4.34375, + "learning_rate": 9.677610717718732e-06, + "loss": 1.72600632, + "memory(GiB)": 111.15, + "step": 12550, + "train_speed(iter/s)": 0.44918 + }, + { + "acc": 0.62508011, + "epoch": 0.3184931506849315, + "grad_norm": 6.25, + "learning_rate": 9.677240169602317e-06, + "loss": 1.73107262, + "memory(GiB)": 111.15, + "step": 12555, + "train_speed(iter/s)": 0.449274 + }, + { + "acc": 0.64505749, + "epoch": 0.31861998985286655, + "grad_norm": 5.84375, + "learning_rate": 9.676869415759999e-06, + "loss": 1.66729317, + "memory(GiB)": 111.15, + "step": 12560, + "train_speed(iter/s)": 0.449366 + }, + { + "acc": 0.62894316, + "epoch": 0.31874682902080165, + "grad_norm": 4.84375, + "learning_rate": 9.67649845620808e-06, + "loss": 1.75995369, + "memory(GiB)": 111.15, + "step": 12565, + "train_speed(iter/s)": 0.449459 + }, + { + "acc": 0.64986734, + "epoch": 0.3188736681887367, + "grad_norm": 4.625, + "learning_rate": 9.676127290962883e-06, + "loss": 1.69117355, + "memory(GiB)": 111.15, + "step": 12570, + "train_speed(iter/s)": 0.449552 + }, + { + "acc": 0.64299374, + "epoch": 0.31900050735667174, + "grad_norm": 5.40625, + "learning_rate": 9.675755920040728e-06, + "loss": 1.6517168, + "memory(GiB)": 111.15, + "step": 12575, + "train_speed(iter/s)": 0.449646 + }, + { + "acc": 0.65798578, + "epoch": 0.3191273465246068, + "grad_norm": 5.0, + "learning_rate": 9.675384343457954e-06, + "loss": 1.64493504, + "memory(GiB)": 111.15, + "step": 12580, + "train_speed(iter/s)": 0.449739 + }, + { + "acc": 0.64022484, + "epoch": 0.3192541856925419, + "grad_norm": 5.03125, + "learning_rate": 9.675012561230901e-06, + "loss": 1.68173428, + "memory(GiB)": 111.15, + "step": 12585, + "train_speed(iter/s)": 0.449832 + }, + { + "acc": 0.65287528, + "epoch": 0.3193810248604769, + "grad_norm": 5.3125, + "learning_rate": 9.674640573375924e-06, + "loss": 1.6714222, + "memory(GiB)": 111.15, + "step": 12590, + "train_speed(iter/s)": 0.449925 + }, + { + "acc": 0.6428266, + "epoch": 0.31950786402841197, + "grad_norm": 6.21875, + "learning_rate": 9.674268379909383e-06, + "loss": 1.72566566, + "memory(GiB)": 111.15, + "step": 12595, + "train_speed(iter/s)": 0.450018 + }, + { + "acc": 0.64935522, + "epoch": 0.319634703196347, + "grad_norm": 6.625, + "learning_rate": 9.67389598084765e-06, + "loss": 1.66134434, + "memory(GiB)": 111.15, + "step": 12600, + "train_speed(iter/s)": 0.450111 + }, + { + "epoch": 0.319634703196347, + "eval_acc": 0.6382615532505013, + "eval_loss": 1.6370997428894043, + "eval_runtime": 113.4572, + "eval_samples_per_second": 56.145, + "eval_steps_per_second": 28.072, + "step": 12600 + }, + { + "acc": 0.6462431, + "epoch": 0.3197615423642821, + "grad_norm": 6.65625, + "learning_rate": 9.673523376207103e-06, + "loss": 1.63447762, + "memory(GiB)": 111.15, + "step": 12605, + "train_speed(iter/s)": 0.448315 + }, + { + "acc": 0.6332521, + "epoch": 0.31988838153221716, + "grad_norm": 6.8125, + "learning_rate": 9.673150566004135e-06, + "loss": 1.66529655, + "memory(GiB)": 111.15, + "step": 12610, + "train_speed(iter/s)": 0.448408 + }, + { + "acc": 0.63859196, + "epoch": 0.3200152207001522, + "grad_norm": 4.875, + "learning_rate": 9.672777550255137e-06, + "loss": 1.74135742, + "memory(GiB)": 111.15, + "step": 12615, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.65100927, + "epoch": 0.32014205986808725, + "grad_norm": 5.1875, + "learning_rate": 9.672404328976523e-06, + "loss": 1.6732151, + "memory(GiB)": 111.15, + "step": 12620, + "train_speed(iter/s)": 0.448593 + }, + { + "acc": 0.64866238, + "epoch": 0.32026889903602235, + "grad_norm": 5.28125, + "learning_rate": 9.672030902184706e-06, + "loss": 1.6350605, + "memory(GiB)": 111.15, + "step": 12625, + "train_speed(iter/s)": 0.448685 + }, + { + "acc": 0.6432375, + "epoch": 0.3203957382039574, + "grad_norm": 4.875, + "learning_rate": 9.671657269896108e-06, + "loss": 1.73939285, + "memory(GiB)": 111.15, + "step": 12630, + "train_speed(iter/s)": 0.448778 + }, + { + "acc": 0.64651279, + "epoch": 0.32052257737189244, + "grad_norm": 4.625, + "learning_rate": 9.671283432127169e-06, + "loss": 1.6644062, + "memory(GiB)": 111.15, + "step": 12635, + "train_speed(iter/s)": 0.448871 + }, + { + "acc": 0.65070481, + "epoch": 0.3206494165398275, + "grad_norm": 5.15625, + "learning_rate": 9.670909388894328e-06, + "loss": 1.62857304, + "memory(GiB)": 111.15, + "step": 12640, + "train_speed(iter/s)": 0.448964 + }, + { + "acc": 0.64438763, + "epoch": 0.3207762557077626, + "grad_norm": 4.53125, + "learning_rate": 9.670535140214037e-06, + "loss": 1.68114624, + "memory(GiB)": 111.15, + "step": 12645, + "train_speed(iter/s)": 0.449058 + }, + { + "acc": 0.64042883, + "epoch": 0.3209030948756976, + "grad_norm": 5.40625, + "learning_rate": 9.670160686102759e-06, + "loss": 1.67542953, + "memory(GiB)": 111.15, + "step": 12650, + "train_speed(iter/s)": 0.44915 + }, + { + "acc": 0.63703537, + "epoch": 0.32102993404363267, + "grad_norm": 5.25, + "learning_rate": 9.669786026576962e-06, + "loss": 1.6852375, + "memory(GiB)": 111.15, + "step": 12655, + "train_speed(iter/s)": 0.449242 + }, + { + "acc": 0.65243816, + "epoch": 0.3211567732115677, + "grad_norm": 5.03125, + "learning_rate": 9.669411161653127e-06, + "loss": 1.65030746, + "memory(GiB)": 111.15, + "step": 12660, + "train_speed(iter/s)": 0.449335 + }, + { + "acc": 0.63336363, + "epoch": 0.3212836123795028, + "grad_norm": 5.0625, + "learning_rate": 9.669036091347742e-06, + "loss": 1.6585041, + "memory(GiB)": 111.15, + "step": 12665, + "train_speed(iter/s)": 0.449428 + }, + { + "acc": 0.64557114, + "epoch": 0.32141045154743786, + "grad_norm": 6.09375, + "learning_rate": 9.668660815677304e-06, + "loss": 1.61463661, + "memory(GiB)": 111.15, + "step": 12670, + "train_speed(iter/s)": 0.44952 + }, + { + "acc": 0.66961851, + "epoch": 0.3215372907153729, + "grad_norm": 6.21875, + "learning_rate": 9.668285334658319e-06, + "loss": 1.60739269, + "memory(GiB)": 111.15, + "step": 12675, + "train_speed(iter/s)": 0.449613 + }, + { + "acc": 0.63705478, + "epoch": 0.32166412988330795, + "grad_norm": 4.875, + "learning_rate": 9.667909648307302e-06, + "loss": 1.69984627, + "memory(GiB)": 111.15, + "step": 12680, + "train_speed(iter/s)": 0.449707 + }, + { + "acc": 0.66240554, + "epoch": 0.32179096905124305, + "grad_norm": 6.96875, + "learning_rate": 9.66753375664078e-06, + "loss": 1.5973053, + "memory(GiB)": 111.15, + "step": 12685, + "train_speed(iter/s)": 0.449801 + }, + { + "acc": 0.64298182, + "epoch": 0.3219178082191781, + "grad_norm": 5.75, + "learning_rate": 9.667157659675284e-06, + "loss": 1.64596786, + "memory(GiB)": 111.15, + "step": 12690, + "train_speed(iter/s)": 0.449894 + }, + { + "acc": 0.64712687, + "epoch": 0.32204464738711314, + "grad_norm": 5.46875, + "learning_rate": 9.666781357427355e-06, + "loss": 1.68261433, + "memory(GiB)": 111.15, + "step": 12695, + "train_speed(iter/s)": 0.449987 + }, + { + "acc": 0.63881989, + "epoch": 0.3221714865550482, + "grad_norm": 5.21875, + "learning_rate": 9.666404849913546e-06, + "loss": 1.67943859, + "memory(GiB)": 111.15, + "step": 12700, + "train_speed(iter/s)": 0.450081 + }, + { + "epoch": 0.3221714865550482, + "eval_acc": 0.6383309026238712, + "eval_loss": 1.636739730834961, + "eval_runtime": 113.1193, + "eval_samples_per_second": 56.312, + "eval_steps_per_second": 28.156, + "step": 12700 + }, + { + "acc": 0.65242548, + "epoch": 0.3222983257229833, + "grad_norm": 6.1875, + "learning_rate": 9.66602813715042e-06, + "loss": 1.62917042, + "memory(GiB)": 111.15, + "step": 12705, + "train_speed(iter/s)": 0.448305 + }, + { + "acc": 0.65058274, + "epoch": 0.3224251648909183, + "grad_norm": 5.34375, + "learning_rate": 9.665651219154543e-06, + "loss": 1.67290039, + "memory(GiB)": 111.15, + "step": 12710, + "train_speed(iter/s)": 0.448398 + }, + { + "acc": 0.6516675, + "epoch": 0.32255200405885337, + "grad_norm": 4.78125, + "learning_rate": 9.665274095942495e-06, + "loss": 1.6879673, + "memory(GiB)": 111.15, + "step": 12715, + "train_speed(iter/s)": 0.448491 + }, + { + "acc": 0.64853749, + "epoch": 0.3226788432267884, + "grad_norm": 5.125, + "learning_rate": 9.664896767530862e-06, + "loss": 1.72316685, + "memory(GiB)": 111.15, + "step": 12720, + "train_speed(iter/s)": 0.448584 + }, + { + "acc": 0.63993325, + "epoch": 0.3228056823947235, + "grad_norm": 4.59375, + "learning_rate": 9.664519233936242e-06, + "loss": 1.66685562, + "memory(GiB)": 111.15, + "step": 12725, + "train_speed(iter/s)": 0.448677 + }, + { + "acc": 0.64207273, + "epoch": 0.32293252156265856, + "grad_norm": 4.90625, + "learning_rate": 9.664141495175242e-06, + "loss": 1.61380424, + "memory(GiB)": 111.15, + "step": 12730, + "train_speed(iter/s)": 0.448768 + }, + { + "acc": 0.64839392, + "epoch": 0.3230593607305936, + "grad_norm": 4.875, + "learning_rate": 9.663763551264476e-06, + "loss": 1.68673096, + "memory(GiB)": 111.15, + "step": 12735, + "train_speed(iter/s)": 0.44886 + }, + { + "acc": 0.64906898, + "epoch": 0.32318619989852865, + "grad_norm": 5.25, + "learning_rate": 9.663385402220565e-06, + "loss": 1.61658974, + "memory(GiB)": 111.15, + "step": 12740, + "train_speed(iter/s)": 0.448952 + }, + { + "acc": 0.64483638, + "epoch": 0.32331303906646375, + "grad_norm": 5.90625, + "learning_rate": 9.663007048060144e-06, + "loss": 1.73548241, + "memory(GiB)": 111.15, + "step": 12745, + "train_speed(iter/s)": 0.449045 + }, + { + "acc": 0.6556385, + "epoch": 0.3234398782343988, + "grad_norm": 5.03125, + "learning_rate": 9.662628488799854e-06, + "loss": 1.58471451, + "memory(GiB)": 111.15, + "step": 12750, + "train_speed(iter/s)": 0.449137 + }, + { + "acc": 0.63845682, + "epoch": 0.32356671740233384, + "grad_norm": 5.90625, + "learning_rate": 9.662249724456346e-06, + "loss": 1.62975445, + "memory(GiB)": 111.15, + "step": 12755, + "train_speed(iter/s)": 0.449229 + }, + { + "acc": 0.65083241, + "epoch": 0.3236935565702689, + "grad_norm": 5.9375, + "learning_rate": 9.661870755046278e-06, + "loss": 1.65941849, + "memory(GiB)": 111.15, + "step": 12760, + "train_speed(iter/s)": 0.449323 + }, + { + "acc": 0.64471331, + "epoch": 0.323820395738204, + "grad_norm": 4.65625, + "learning_rate": 9.661491580586322e-06, + "loss": 1.62121983, + "memory(GiB)": 111.15, + "step": 12765, + "train_speed(iter/s)": 0.449415 + }, + { + "acc": 0.63827276, + "epoch": 0.323947234906139, + "grad_norm": 6.15625, + "learning_rate": 9.661112201093156e-06, + "loss": 1.65420418, + "memory(GiB)": 111.15, + "step": 12770, + "train_speed(iter/s)": 0.449507 + }, + { + "acc": 0.64150987, + "epoch": 0.32407407407407407, + "grad_norm": 5.46875, + "learning_rate": 9.660732616583463e-06, + "loss": 1.68162708, + "memory(GiB)": 111.15, + "step": 12775, + "train_speed(iter/s)": 0.449599 + }, + { + "acc": 0.6585186, + "epoch": 0.3242009132420091, + "grad_norm": 5.40625, + "learning_rate": 9.660352827073941e-06, + "loss": 1.62037201, + "memory(GiB)": 111.15, + "step": 12780, + "train_speed(iter/s)": 0.449691 + }, + { + "acc": 0.62741652, + "epoch": 0.3243277524099442, + "grad_norm": 5.09375, + "learning_rate": 9.659972832581295e-06, + "loss": 1.68121338, + "memory(GiB)": 111.15, + "step": 12785, + "train_speed(iter/s)": 0.449784 + }, + { + "acc": 0.64315205, + "epoch": 0.32445459157787926, + "grad_norm": 5.1875, + "learning_rate": 9.659592633122238e-06, + "loss": 1.64644413, + "memory(GiB)": 111.15, + "step": 12790, + "train_speed(iter/s)": 0.449876 + }, + { + "acc": 0.63680639, + "epoch": 0.3245814307458143, + "grad_norm": 6.03125, + "learning_rate": 9.659212228713495e-06, + "loss": 1.68993225, + "memory(GiB)": 111.15, + "step": 12795, + "train_speed(iter/s)": 0.449969 + }, + { + "acc": 0.65505972, + "epoch": 0.32470826991374935, + "grad_norm": 5.6875, + "learning_rate": 9.658831619371793e-06, + "loss": 1.64265099, + "memory(GiB)": 111.15, + "step": 12800, + "train_speed(iter/s)": 0.450061 + }, + { + "epoch": 0.32470826991374935, + "eval_acc": 0.6385264177246972, + "eval_loss": 1.6363639831542969, + "eval_runtime": 113.5116, + "eval_samples_per_second": 56.118, + "eval_steps_per_second": 28.059, + "step": 12800 + }, + { + "acc": 0.64229908, + "epoch": 0.32483510908168445, + "grad_norm": 5.9375, + "learning_rate": 9.658450805113879e-06, + "loss": 1.71740494, + "memory(GiB)": 111.15, + "step": 12805, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.65582676, + "epoch": 0.3249619482496195, + "grad_norm": 4.5625, + "learning_rate": 9.6580697859565e-06, + "loss": 1.59375496, + "memory(GiB)": 111.15, + "step": 12810, + "train_speed(iter/s)": 0.448386 + }, + { + "acc": 0.6705183, + "epoch": 0.32508878741755454, + "grad_norm": 4.78125, + "learning_rate": 9.657688561916414e-06, + "loss": 1.51463938, + "memory(GiB)": 111.15, + "step": 12815, + "train_speed(iter/s)": 0.448476 + }, + { + "acc": 0.65902934, + "epoch": 0.3252156265854896, + "grad_norm": 6.0, + "learning_rate": 9.65730713301039e-06, + "loss": 1.59709816, + "memory(GiB)": 111.15, + "step": 12820, + "train_speed(iter/s)": 0.448567 + }, + { + "acc": 0.63833179, + "epoch": 0.3253424657534247, + "grad_norm": 5.5, + "learning_rate": 9.656925499255206e-06, + "loss": 1.65652084, + "memory(GiB)": 111.15, + "step": 12825, + "train_speed(iter/s)": 0.448659 + }, + { + "acc": 0.6510891, + "epoch": 0.3254693049213597, + "grad_norm": 4.84375, + "learning_rate": 9.656543660667646e-06, + "loss": 1.62483921, + "memory(GiB)": 111.15, + "step": 12830, + "train_speed(iter/s)": 0.448751 + }, + { + "acc": 0.66133995, + "epoch": 0.32559614408929477, + "grad_norm": 5.3125, + "learning_rate": 9.656161617264507e-06, + "loss": 1.56089907, + "memory(GiB)": 111.15, + "step": 12835, + "train_speed(iter/s)": 0.448843 + }, + { + "acc": 0.63487082, + "epoch": 0.3257229832572298, + "grad_norm": 6.25, + "learning_rate": 9.65577936906259e-06, + "loss": 1.66533298, + "memory(GiB)": 111.15, + "step": 12840, + "train_speed(iter/s)": 0.448933 + }, + { + "acc": 0.63235302, + "epoch": 0.3258498224251649, + "grad_norm": 6.4375, + "learning_rate": 9.65539691607871e-06, + "loss": 1.67911472, + "memory(GiB)": 111.15, + "step": 12845, + "train_speed(iter/s)": 0.449025 + }, + { + "acc": 0.64030523, + "epoch": 0.32597666159309996, + "grad_norm": 4.71875, + "learning_rate": 9.65501425832969e-06, + "loss": 1.69944878, + "memory(GiB)": 111.15, + "step": 12850, + "train_speed(iter/s)": 0.449116 + }, + { + "acc": 0.63727961, + "epoch": 0.326103500761035, + "grad_norm": 4.96875, + "learning_rate": 9.65463139583236e-06, + "loss": 1.69278393, + "memory(GiB)": 111.15, + "step": 12855, + "train_speed(iter/s)": 0.449207 + }, + { + "acc": 0.64086809, + "epoch": 0.32623033992897005, + "grad_norm": 5.4375, + "learning_rate": 9.65424832860356e-06, + "loss": 1.68437691, + "memory(GiB)": 111.15, + "step": 12860, + "train_speed(iter/s)": 0.449299 + }, + { + "acc": 0.64703484, + "epoch": 0.32635717909690515, + "grad_norm": 5.1875, + "learning_rate": 9.653865056660136e-06, + "loss": 1.6538414, + "memory(GiB)": 111.15, + "step": 12865, + "train_speed(iter/s)": 0.44939 + }, + { + "acc": 0.64520769, + "epoch": 0.3264840182648402, + "grad_norm": 5.34375, + "learning_rate": 9.653481580018951e-06, + "loss": 1.69008942, + "memory(GiB)": 111.15, + "step": 12870, + "train_speed(iter/s)": 0.449481 + }, + { + "acc": 0.64410734, + "epoch": 0.32661085743277524, + "grad_norm": 6.03125, + "learning_rate": 9.653097898696869e-06, + "loss": 1.7099102, + "memory(GiB)": 111.15, + "step": 12875, + "train_speed(iter/s)": 0.449572 + }, + { + "acc": 0.63412018, + "epoch": 0.3267376966007103, + "grad_norm": 6.125, + "learning_rate": 9.652714012710766e-06, + "loss": 1.70931396, + "memory(GiB)": 111.15, + "step": 12880, + "train_speed(iter/s)": 0.449663 + }, + { + "acc": 0.64472704, + "epoch": 0.3268645357686454, + "grad_norm": 6.09375, + "learning_rate": 9.65232992207753e-06, + "loss": 1.72259197, + "memory(GiB)": 111.15, + "step": 12885, + "train_speed(iter/s)": 0.449755 + }, + { + "acc": 0.63794131, + "epoch": 0.3269913749365804, + "grad_norm": 4.59375, + "learning_rate": 9.651945626814052e-06, + "loss": 1.70574207, + "memory(GiB)": 111.15, + "step": 12890, + "train_speed(iter/s)": 0.449848 + }, + { + "acc": 0.649508, + "epoch": 0.32711821410451547, + "grad_norm": 5.53125, + "learning_rate": 9.651561126937236e-06, + "loss": 1.63064919, + "memory(GiB)": 111.15, + "step": 12895, + "train_speed(iter/s)": 0.449938 + }, + { + "acc": 0.63328896, + "epoch": 0.3272450532724505, + "grad_norm": 5.875, + "learning_rate": 9.651176422463994e-06, + "loss": 1.74399433, + "memory(GiB)": 111.15, + "step": 12900, + "train_speed(iter/s)": 0.45003 + }, + { + "epoch": 0.3272450532724505, + "eval_acc": 0.6384775389494907, + "eval_loss": 1.6361290216445923, + "eval_runtime": 113.0341, + "eval_samples_per_second": 56.355, + "eval_steps_per_second": 28.177, + "step": 12900 + }, + { + "acc": 0.64975996, + "epoch": 0.3273718924403856, + "grad_norm": 4.5625, + "learning_rate": 9.650791513411246e-06, + "loss": 1.61017075, + "memory(GiB)": 111.15, + "step": 12905, + "train_speed(iter/s)": 0.448284 + }, + { + "acc": 0.62648792, + "epoch": 0.32749873160832066, + "grad_norm": 4.4375, + "learning_rate": 9.650406399795924e-06, + "loss": 1.70466881, + "memory(GiB)": 111.15, + "step": 12910, + "train_speed(iter/s)": 0.448374 + }, + { + "acc": 0.64123712, + "epoch": 0.3276255707762557, + "grad_norm": 6.375, + "learning_rate": 9.650021081634965e-06, + "loss": 1.71838818, + "memory(GiB)": 111.15, + "step": 12915, + "train_speed(iter/s)": 0.448466 + }, + { + "acc": 0.64075165, + "epoch": 0.32775240994419075, + "grad_norm": 5.375, + "learning_rate": 9.649635558945318e-06, + "loss": 1.64437752, + "memory(GiB)": 111.15, + "step": 12920, + "train_speed(iter/s)": 0.448557 + }, + { + "acc": 0.63724203, + "epoch": 0.32787924911212585, + "grad_norm": 4.5625, + "learning_rate": 9.649249831743941e-06, + "loss": 1.69252949, + "memory(GiB)": 111.15, + "step": 12925, + "train_speed(iter/s)": 0.448647 + }, + { + "acc": 0.63944511, + "epoch": 0.3280060882800609, + "grad_norm": 5.21875, + "learning_rate": 9.6488639000478e-06, + "loss": 1.70934372, + "memory(GiB)": 111.15, + "step": 12930, + "train_speed(iter/s)": 0.448739 + }, + { + "acc": 0.6546886, + "epoch": 0.32813292744799594, + "grad_norm": 5.125, + "learning_rate": 9.648477763873868e-06, + "loss": 1.64684982, + "memory(GiB)": 111.15, + "step": 12935, + "train_speed(iter/s)": 0.448831 + }, + { + "acc": 0.64837551, + "epoch": 0.328259766615931, + "grad_norm": 5.5625, + "learning_rate": 9.648091423239128e-06, + "loss": 1.61827412, + "memory(GiB)": 111.15, + "step": 12940, + "train_speed(iter/s)": 0.448922 + }, + { + "acc": 0.65476255, + "epoch": 0.3283866057838661, + "grad_norm": 5.375, + "learning_rate": 9.647704878160576e-06, + "loss": 1.5966938, + "memory(GiB)": 111.15, + "step": 12945, + "train_speed(iter/s)": 0.449013 + }, + { + "acc": 0.628016, + "epoch": 0.3285134449518011, + "grad_norm": 5.65625, + "learning_rate": 9.647318128655213e-06, + "loss": 1.65439758, + "memory(GiB)": 111.15, + "step": 12950, + "train_speed(iter/s)": 0.449105 + }, + { + "acc": 0.66120496, + "epoch": 0.32864028411973617, + "grad_norm": 4.625, + "learning_rate": 9.64693117474005e-06, + "loss": 1.59857464, + "memory(GiB)": 111.15, + "step": 12955, + "train_speed(iter/s)": 0.449194 + }, + { + "acc": 0.64784303, + "epoch": 0.3287671232876712, + "grad_norm": 5.1875, + "learning_rate": 9.646544016432109e-06, + "loss": 1.65207672, + "memory(GiB)": 111.15, + "step": 12960, + "train_speed(iter/s)": 0.449285 + }, + { + "acc": 0.65713844, + "epoch": 0.3288939624556063, + "grad_norm": 5.78125, + "learning_rate": 9.646156653748415e-06, + "loss": 1.63149567, + "memory(GiB)": 111.15, + "step": 12965, + "train_speed(iter/s)": 0.449376 + }, + { + "acc": 0.65116572, + "epoch": 0.32902080162354136, + "grad_norm": 5.5, + "learning_rate": 9.645769086706008e-06, + "loss": 1.6997673, + "memory(GiB)": 111.15, + "step": 12970, + "train_speed(iter/s)": 0.449466 + }, + { + "acc": 0.64783249, + "epoch": 0.3291476407914764, + "grad_norm": 4.34375, + "learning_rate": 9.645381315321934e-06, + "loss": 1.64681797, + "memory(GiB)": 111.15, + "step": 12975, + "train_speed(iter/s)": 0.449557 + }, + { + "acc": 0.64741354, + "epoch": 0.32927447995941145, + "grad_norm": 4.46875, + "learning_rate": 9.64499333961325e-06, + "loss": 1.68768044, + "memory(GiB)": 111.15, + "step": 12980, + "train_speed(iter/s)": 0.449649 + }, + { + "acc": 0.66060028, + "epoch": 0.32940131912734655, + "grad_norm": 4.875, + "learning_rate": 9.64460515959702e-06, + "loss": 1.56873646, + "memory(GiB)": 111.15, + "step": 12985, + "train_speed(iter/s)": 0.44974 + }, + { + "acc": 0.63774815, + "epoch": 0.3295281582952816, + "grad_norm": 5.84375, + "learning_rate": 9.64421677529032e-06, + "loss": 1.66719456, + "memory(GiB)": 111.15, + "step": 12990, + "train_speed(iter/s)": 0.449832 + }, + { + "acc": 0.66481404, + "epoch": 0.32965499746321664, + "grad_norm": 4.5625, + "learning_rate": 9.64382818671023e-06, + "loss": 1.57818432, + "memory(GiB)": 111.15, + "step": 12995, + "train_speed(iter/s)": 0.449923 + }, + { + "acc": 0.65654202, + "epoch": 0.3297818366311517, + "grad_norm": 5.71875, + "learning_rate": 9.643439393873844e-06, + "loss": 1.58309574, + "memory(GiB)": 111.15, + "step": 13000, + "train_speed(iter/s)": 0.450013 + }, + { + "epoch": 0.3297818366311517, + "eval_acc": 0.638631277319371, + "eval_loss": 1.635608196258545, + "eval_runtime": 113.3342, + "eval_samples_per_second": 56.205, + "eval_steps_per_second": 28.103, + "step": 13000 + }, + { + "acc": 0.64650068, + "epoch": 0.3299086757990868, + "grad_norm": 4.78125, + "learning_rate": 9.643050396798262e-06, + "loss": 1.67794151, + "memory(GiB)": 111.15, + "step": 13005, + "train_speed(iter/s)": 0.448277 + }, + { + "acc": 0.65416927, + "epoch": 0.3300355149670218, + "grad_norm": 5.03125, + "learning_rate": 9.642661195500593e-06, + "loss": 1.60028248, + "memory(GiB)": 111.15, + "step": 13010, + "train_speed(iter/s)": 0.448369 + }, + { + "acc": 0.63268671, + "epoch": 0.33016235413495687, + "grad_norm": 5.75, + "learning_rate": 9.642271789997956e-06, + "loss": 1.66915569, + "memory(GiB)": 111.15, + "step": 13015, + "train_speed(iter/s)": 0.448459 + }, + { + "acc": 0.65095663, + "epoch": 0.3302891933028919, + "grad_norm": 4.625, + "learning_rate": 9.64188218030748e-06, + "loss": 1.582341, + "memory(GiB)": 111.15, + "step": 13020, + "train_speed(iter/s)": 0.44855 + }, + { + "acc": 0.65387135, + "epoch": 0.330416032470827, + "grad_norm": 4.78125, + "learning_rate": 9.641492366446301e-06, + "loss": 1.61686783, + "memory(GiB)": 111.15, + "step": 13025, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.64295979, + "epoch": 0.33054287163876206, + "grad_norm": 4.625, + "learning_rate": 9.641102348431565e-06, + "loss": 1.64997482, + "memory(GiB)": 111.15, + "step": 13030, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.63614974, + "epoch": 0.3306697108066971, + "grad_norm": 4.96875, + "learning_rate": 9.640712126280429e-06, + "loss": 1.67680168, + "memory(GiB)": 111.15, + "step": 13035, + "train_speed(iter/s)": 0.448822 + }, + { + "acc": 0.64576697, + "epoch": 0.33079654997463215, + "grad_norm": 5.0, + "learning_rate": 9.640321700010053e-06, + "loss": 1.62277756, + "memory(GiB)": 111.15, + "step": 13040, + "train_speed(iter/s)": 0.448913 + }, + { + "acc": 0.64266219, + "epoch": 0.33092338914256725, + "grad_norm": 5.15625, + "learning_rate": 9.63993106963761e-06, + "loss": 1.66203728, + "memory(GiB)": 111.15, + "step": 13045, + "train_speed(iter/s)": 0.449005 + }, + { + "acc": 0.63997221, + "epoch": 0.3310502283105023, + "grad_norm": 4.65625, + "learning_rate": 9.639540235180283e-06, + "loss": 1.67431717, + "memory(GiB)": 111.15, + "step": 13050, + "train_speed(iter/s)": 0.449095 + }, + { + "acc": 0.65796938, + "epoch": 0.33117706747843734, + "grad_norm": 5.96875, + "learning_rate": 9.639149196655263e-06, + "loss": 1.59994869, + "memory(GiB)": 111.15, + "step": 13055, + "train_speed(iter/s)": 0.449186 + }, + { + "acc": 0.64478292, + "epoch": 0.3313039066463724, + "grad_norm": 4.96875, + "learning_rate": 9.638757954079749e-06, + "loss": 1.60334072, + "memory(GiB)": 111.15, + "step": 13060, + "train_speed(iter/s)": 0.449277 + }, + { + "acc": 0.65211725, + "epoch": 0.3314307458143075, + "grad_norm": 6.0, + "learning_rate": 9.638366507470948e-06, + "loss": 1.61802597, + "memory(GiB)": 111.15, + "step": 13065, + "train_speed(iter/s)": 0.449368 + }, + { + "acc": 0.65717516, + "epoch": 0.3315575849822425, + "grad_norm": 5.875, + "learning_rate": 9.637974856846082e-06, + "loss": 1.65632782, + "memory(GiB)": 111.15, + "step": 13070, + "train_speed(iter/s)": 0.449458 + }, + { + "acc": 0.65569501, + "epoch": 0.33168442415017757, + "grad_norm": 5.40625, + "learning_rate": 9.637583002222373e-06, + "loss": 1.56628113, + "memory(GiB)": 111.15, + "step": 13075, + "train_speed(iter/s)": 0.449549 + }, + { + "acc": 0.65104508, + "epoch": 0.3318112633181126, + "grad_norm": 4.9375, + "learning_rate": 9.637190943617059e-06, + "loss": 1.66362152, + "memory(GiB)": 111.15, + "step": 13080, + "train_speed(iter/s)": 0.44964 + }, + { + "acc": 0.6433291, + "epoch": 0.3319381024860477, + "grad_norm": 5.5625, + "learning_rate": 9.636798681047383e-06, + "loss": 1.61164646, + "memory(GiB)": 111.15, + "step": 13085, + "train_speed(iter/s)": 0.44973 + }, + { + "acc": 0.66432476, + "epoch": 0.33206494165398276, + "grad_norm": 5.875, + "learning_rate": 9.6364062145306e-06, + "loss": 1.63647938, + "memory(GiB)": 111.15, + "step": 13090, + "train_speed(iter/s)": 0.449821 + }, + { + "acc": 0.65174584, + "epoch": 0.3321917808219178, + "grad_norm": 4.5, + "learning_rate": 9.636013544083971e-06, + "loss": 1.59489069, + "memory(GiB)": 111.15, + "step": 13095, + "train_speed(iter/s)": 0.449911 + }, + { + "acc": 0.63728123, + "epoch": 0.33231861998985285, + "grad_norm": 5.09375, + "learning_rate": 9.635620669724768e-06, + "loss": 1.68059349, + "memory(GiB)": 111.15, + "step": 13100, + "train_speed(iter/s)": 0.450002 + }, + { + "epoch": 0.33231861998985285, + "eval_acc": 0.6385765498018321, + "eval_loss": 1.635573387145996, + "eval_runtime": 113.5173, + "eval_samples_per_second": 56.115, + "eval_steps_per_second": 28.057, + "step": 13100 + }, + { + "acc": 0.65193415, + "epoch": 0.33244545915778795, + "grad_norm": 5.25, + "learning_rate": 9.635227591470272e-06, + "loss": 1.68287163, + "memory(GiB)": 111.15, + "step": 13105, + "train_speed(iter/s)": 0.448275 + }, + { + "acc": 0.65815716, + "epoch": 0.332572298325723, + "grad_norm": 5.21875, + "learning_rate": 9.63483430933777e-06, + "loss": 1.61004143, + "memory(GiB)": 111.15, + "step": 13110, + "train_speed(iter/s)": 0.448364 + }, + { + "acc": 0.65156751, + "epoch": 0.33269913749365804, + "grad_norm": 5.28125, + "learning_rate": 9.634440823344565e-06, + "loss": 1.67079315, + "memory(GiB)": 111.15, + "step": 13115, + "train_speed(iter/s)": 0.448453 + }, + { + "acc": 0.66256595, + "epoch": 0.3328259766615931, + "grad_norm": 5.46875, + "learning_rate": 9.634047133507959e-06, + "loss": 1.57034683, + "memory(GiB)": 111.15, + "step": 13120, + "train_speed(iter/s)": 0.448544 + }, + { + "acc": 0.66299314, + "epoch": 0.3329528158295282, + "grad_norm": 4.84375, + "learning_rate": 9.63365323984527e-06, + "loss": 1.58265743, + "memory(GiB)": 111.15, + "step": 13125, + "train_speed(iter/s)": 0.448633 + }, + { + "acc": 0.65500031, + "epoch": 0.3330796549974632, + "grad_norm": 5.125, + "learning_rate": 9.633259142373825e-06, + "loss": 1.57403793, + "memory(GiB)": 111.15, + "step": 13130, + "train_speed(iter/s)": 0.448722 + }, + { + "acc": 0.63279281, + "epoch": 0.33320649416539827, + "grad_norm": 4.90625, + "learning_rate": 9.632864841110957e-06, + "loss": 1.68904915, + "memory(GiB)": 111.15, + "step": 13135, + "train_speed(iter/s)": 0.44881 + }, + { + "acc": 0.63893814, + "epoch": 0.3333333333333333, + "grad_norm": 6.1875, + "learning_rate": 9.632470336074009e-06, + "loss": 1.70369148, + "memory(GiB)": 111.15, + "step": 13140, + "train_speed(iter/s)": 0.4489 + }, + { + "acc": 0.65731258, + "epoch": 0.3334601725012684, + "grad_norm": 5.75, + "learning_rate": 9.632075627280333e-06, + "loss": 1.5899931, + "memory(GiB)": 111.15, + "step": 13145, + "train_speed(iter/s)": 0.44899 + }, + { + "acc": 0.64074097, + "epoch": 0.33358701166920346, + "grad_norm": 5.0625, + "learning_rate": 9.631680714747292e-06, + "loss": 1.64995422, + "memory(GiB)": 111.15, + "step": 13150, + "train_speed(iter/s)": 0.44908 + }, + { + "acc": 0.65549002, + "epoch": 0.3337138508371385, + "grad_norm": 7.59375, + "learning_rate": 9.63128559849225e-06, + "loss": 1.58083076, + "memory(GiB)": 111.15, + "step": 13155, + "train_speed(iter/s)": 0.44917 + }, + { + "acc": 0.64021378, + "epoch": 0.33384069000507355, + "grad_norm": 6.21875, + "learning_rate": 9.630890278532594e-06, + "loss": 1.60837212, + "memory(GiB)": 111.15, + "step": 13160, + "train_speed(iter/s)": 0.44926 + }, + { + "acc": 0.62284355, + "epoch": 0.33396752917300865, + "grad_norm": 4.875, + "learning_rate": 9.630494754885706e-06, + "loss": 1.7300766, + "memory(GiB)": 111.15, + "step": 13165, + "train_speed(iter/s)": 0.44935 + }, + { + "acc": 0.64853106, + "epoch": 0.3340943683409437, + "grad_norm": 5.15625, + "learning_rate": 9.630099027568986e-06, + "loss": 1.62550392, + "memory(GiB)": 111.15, + "step": 13170, + "train_speed(iter/s)": 0.449439 + }, + { + "acc": 0.6525857, + "epoch": 0.33422120750887874, + "grad_norm": 5.375, + "learning_rate": 9.629703096599839e-06, + "loss": 1.68082466, + "memory(GiB)": 111.15, + "step": 13175, + "train_speed(iter/s)": 0.44953 + }, + { + "acc": 0.65101528, + "epoch": 0.3343480466768138, + "grad_norm": 4.71875, + "learning_rate": 9.629306961995678e-06, + "loss": 1.64511185, + "memory(GiB)": 111.15, + "step": 13180, + "train_speed(iter/s)": 0.449618 + }, + { + "acc": 0.63455973, + "epoch": 0.3344748858447489, + "grad_norm": 4.96875, + "learning_rate": 9.62891062377393e-06, + "loss": 1.69095974, + "memory(GiB)": 111.15, + "step": 13185, + "train_speed(iter/s)": 0.449708 + }, + { + "acc": 0.65014539, + "epoch": 0.3346017250126839, + "grad_norm": 5.375, + "learning_rate": 9.628514081952026e-06, + "loss": 1.59857016, + "memory(GiB)": 111.15, + "step": 13190, + "train_speed(iter/s)": 0.449798 + }, + { + "acc": 0.62542477, + "epoch": 0.33472856418061897, + "grad_norm": 5.8125, + "learning_rate": 9.628117336547408e-06, + "loss": 1.74039078, + "memory(GiB)": 111.15, + "step": 13195, + "train_speed(iter/s)": 0.449888 + }, + { + "acc": 0.63493738, + "epoch": 0.334855403348554, + "grad_norm": 4.375, + "learning_rate": 9.627720387577525e-06, + "loss": 1.64815445, + "memory(GiB)": 111.15, + "step": 13200, + "train_speed(iter/s)": 0.449977 + }, + { + "epoch": 0.334855403348554, + "eval_acc": 0.6386550900560101, + "eval_loss": 1.6349986791610718, + "eval_runtime": 113.5552, + "eval_samples_per_second": 56.096, + "eval_steps_per_second": 28.048, + "step": 13200 + }, + { + "acc": 0.64301682, + "epoch": 0.3349822425164891, + "grad_norm": 6.15625, + "learning_rate": 9.62732323505984e-06, + "loss": 1.7193306, + "memory(GiB)": 111.15, + "step": 13205, + "train_speed(iter/s)": 0.448263 + }, + { + "acc": 0.65336485, + "epoch": 0.33510908168442416, + "grad_norm": 4.71875, + "learning_rate": 9.62692587901182e-06, + "loss": 1.5975771, + "memory(GiB)": 111.15, + "step": 13210, + "train_speed(iter/s)": 0.448351 + }, + { + "acc": 0.64600048, + "epoch": 0.3352359208523592, + "grad_norm": 5.09375, + "learning_rate": 9.62652831945094e-06, + "loss": 1.68663177, + "memory(GiB)": 111.15, + "step": 13215, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.65562434, + "epoch": 0.33536276002029425, + "grad_norm": 5.5625, + "learning_rate": 9.626130556394689e-06, + "loss": 1.67380371, + "memory(GiB)": 111.15, + "step": 13220, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.65275626, + "epoch": 0.33548959918822935, + "grad_norm": 4.90625, + "learning_rate": 9.625732589860562e-06, + "loss": 1.67726994, + "memory(GiB)": 111.15, + "step": 13225, + "train_speed(iter/s)": 0.448618 + }, + { + "acc": 0.63073645, + "epoch": 0.3356164383561644, + "grad_norm": 5.78125, + "learning_rate": 9.625334419866064e-06, + "loss": 1.73389359, + "memory(GiB)": 111.15, + "step": 13230, + "train_speed(iter/s)": 0.448708 + }, + { + "acc": 0.64667993, + "epoch": 0.33574327752409944, + "grad_norm": 5.28125, + "learning_rate": 9.624936046428708e-06, + "loss": 1.62015228, + "memory(GiB)": 111.15, + "step": 13235, + "train_speed(iter/s)": 0.448797 + }, + { + "acc": 0.65902152, + "epoch": 0.3358701166920345, + "grad_norm": 5.0625, + "learning_rate": 9.624537469566015e-06, + "loss": 1.62713299, + "memory(GiB)": 111.15, + "step": 13240, + "train_speed(iter/s)": 0.448886 + }, + { + "acc": 0.6513711, + "epoch": 0.3359969558599696, + "grad_norm": 5.875, + "learning_rate": 9.624138689295516e-06, + "loss": 1.59884453, + "memory(GiB)": 111.15, + "step": 13245, + "train_speed(iter/s)": 0.448975 + }, + { + "acc": 0.64085379, + "epoch": 0.3361237950279046, + "grad_norm": 5.21875, + "learning_rate": 9.623739705634753e-06, + "loss": 1.62877579, + "memory(GiB)": 111.15, + "step": 13250, + "train_speed(iter/s)": 0.449064 + }, + { + "acc": 0.65831618, + "epoch": 0.33625063419583967, + "grad_norm": 5.6875, + "learning_rate": 9.623340518601274e-06, + "loss": 1.62732716, + "memory(GiB)": 111.15, + "step": 13255, + "train_speed(iter/s)": 0.449152 + }, + { + "acc": 0.64107018, + "epoch": 0.3363774733637747, + "grad_norm": 7.15625, + "learning_rate": 9.622941128212639e-06, + "loss": 1.74835548, + "memory(GiB)": 111.15, + "step": 13260, + "train_speed(iter/s)": 0.449241 + }, + { + "acc": 0.64420753, + "epoch": 0.3365043125317098, + "grad_norm": 5.6875, + "learning_rate": 9.622541534486411e-06, + "loss": 1.6488121, + "memory(GiB)": 111.15, + "step": 13265, + "train_speed(iter/s)": 0.449332 + }, + { + "acc": 0.64763708, + "epoch": 0.33663115169964486, + "grad_norm": 6.25, + "learning_rate": 9.62214173744017e-06, + "loss": 1.66007347, + "memory(GiB)": 111.15, + "step": 13270, + "train_speed(iter/s)": 0.44942 + }, + { + "acc": 0.64130392, + "epoch": 0.3367579908675799, + "grad_norm": 5.21875, + "learning_rate": 9.6217417370915e-06, + "loss": 1.65013523, + "memory(GiB)": 111.15, + "step": 13275, + "train_speed(iter/s)": 0.449508 + }, + { + "acc": 0.65157938, + "epoch": 0.33688483003551495, + "grad_norm": 5.34375, + "learning_rate": 9.62134153345799e-06, + "loss": 1.62035675, + "memory(GiB)": 111.15, + "step": 13280, + "train_speed(iter/s)": 0.449597 + }, + { + "acc": 0.6455739, + "epoch": 0.33701166920345005, + "grad_norm": 5.5625, + "learning_rate": 9.620941126557248e-06, + "loss": 1.60122757, + "memory(GiB)": 111.15, + "step": 13285, + "train_speed(iter/s)": 0.449686 + }, + { + "acc": 0.66329541, + "epoch": 0.3371385083713851, + "grad_norm": 4.5625, + "learning_rate": 9.620540516406885e-06, + "loss": 1.52728844, + "memory(GiB)": 111.15, + "step": 13290, + "train_speed(iter/s)": 0.449774 + }, + { + "acc": 0.63905411, + "epoch": 0.33726534753932014, + "grad_norm": 4.9375, + "learning_rate": 9.620139703024522e-06, + "loss": 1.65136719, + "memory(GiB)": 111.15, + "step": 13295, + "train_speed(iter/s)": 0.449864 + }, + { + "acc": 0.63869944, + "epoch": 0.3373921867072552, + "grad_norm": 5.90625, + "learning_rate": 9.619738686427785e-06, + "loss": 1.73343887, + "memory(GiB)": 111.15, + "step": 13300, + "train_speed(iter/s)": 0.449953 + }, + { + "epoch": 0.3373921867072552, + "eval_acc": 0.6388393254394807, + "eval_loss": 1.6349447965621948, + "eval_runtime": 113.8407, + "eval_samples_per_second": 55.955, + "eval_steps_per_second": 27.978, + "step": 13300 + }, + { + "acc": 0.64281597, + "epoch": 0.3375190258751903, + "grad_norm": 4.875, + "learning_rate": 9.619337466634317e-06, + "loss": 1.63470974, + "memory(GiB)": 111.15, + "step": 13305, + "train_speed(iter/s)": 0.448249 + }, + { + "acc": 0.66472673, + "epoch": 0.3376458650431253, + "grad_norm": 5.71875, + "learning_rate": 9.618936043661762e-06, + "loss": 1.58528605, + "memory(GiB)": 111.15, + "step": 13310, + "train_speed(iter/s)": 0.448338 + }, + { + "acc": 0.63050585, + "epoch": 0.33777270421106037, + "grad_norm": 5.84375, + "learning_rate": 9.618534417527779e-06, + "loss": 1.66478615, + "memory(GiB)": 111.15, + "step": 13315, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.64970269, + "epoch": 0.3378995433789954, + "grad_norm": 5.875, + "learning_rate": 9.61813258825003e-06, + "loss": 1.58948174, + "memory(GiB)": 111.15, + "step": 13320, + "train_speed(iter/s)": 0.448516 + }, + { + "acc": 0.63918457, + "epoch": 0.3380263825469305, + "grad_norm": 5.40625, + "learning_rate": 9.617730555846191e-06, + "loss": 1.65543938, + "memory(GiB)": 111.15, + "step": 13325, + "train_speed(iter/s)": 0.448605 + }, + { + "acc": 0.64256334, + "epoch": 0.33815322171486556, + "grad_norm": 5.34375, + "learning_rate": 9.617328320333947e-06, + "loss": 1.63879738, + "memory(GiB)": 111.15, + "step": 13330, + "train_speed(iter/s)": 0.448694 + }, + { + "acc": 0.66426125, + "epoch": 0.3382800608828006, + "grad_norm": 5.3125, + "learning_rate": 9.616925881730989e-06, + "loss": 1.54026623, + "memory(GiB)": 111.15, + "step": 13335, + "train_speed(iter/s)": 0.448783 + }, + { + "acc": 0.65528955, + "epoch": 0.33840690005073565, + "grad_norm": 5.125, + "learning_rate": 9.616523240055017e-06, + "loss": 1.6255867, + "memory(GiB)": 111.15, + "step": 13340, + "train_speed(iter/s)": 0.448873 + }, + { + "acc": 0.63299441, + "epoch": 0.33853373921867075, + "grad_norm": 5.15625, + "learning_rate": 9.616120395323743e-06, + "loss": 1.71563339, + "memory(GiB)": 111.15, + "step": 13345, + "train_speed(iter/s)": 0.448961 + }, + { + "acc": 0.64787416, + "epoch": 0.3386605783866058, + "grad_norm": 4.90625, + "learning_rate": 9.615717347554882e-06, + "loss": 1.62935066, + "memory(GiB)": 111.15, + "step": 13350, + "train_speed(iter/s)": 0.44905 + }, + { + "acc": 0.63082991, + "epoch": 0.33878741755454084, + "grad_norm": 4.6875, + "learning_rate": 9.615314096766166e-06, + "loss": 1.65929337, + "memory(GiB)": 111.15, + "step": 13355, + "train_speed(iter/s)": 0.449139 + }, + { + "acc": 0.64211025, + "epoch": 0.3389142567224759, + "grad_norm": 6.34375, + "learning_rate": 9.61491064297533e-06, + "loss": 1.71746864, + "memory(GiB)": 111.15, + "step": 13360, + "train_speed(iter/s)": 0.449227 + }, + { + "acc": 0.64315557, + "epoch": 0.339041095890411, + "grad_norm": 5.375, + "learning_rate": 9.614506986200119e-06, + "loss": 1.70846329, + "memory(GiB)": 111.15, + "step": 13365, + "train_speed(iter/s)": 0.449317 + }, + { + "acc": 0.64313512, + "epoch": 0.339167935058346, + "grad_norm": 6.75, + "learning_rate": 9.61410312645829e-06, + "loss": 1.6734972, + "memory(GiB)": 111.15, + "step": 13370, + "train_speed(iter/s)": 0.449405 + }, + { + "acc": 0.64325175, + "epoch": 0.33929477422628107, + "grad_norm": 6.0625, + "learning_rate": 9.613699063767603e-06, + "loss": 1.70340042, + "memory(GiB)": 111.15, + "step": 13375, + "train_speed(iter/s)": 0.449494 + }, + { + "acc": 0.64963217, + "epoch": 0.3394216133942161, + "grad_norm": 5.5625, + "learning_rate": 9.613294798145833e-06, + "loss": 1.68396988, + "memory(GiB)": 111.15, + "step": 13380, + "train_speed(iter/s)": 0.449583 + }, + { + "acc": 0.65194283, + "epoch": 0.3395484525621512, + "grad_norm": 5.59375, + "learning_rate": 9.612890329610762e-06, + "loss": 1.6042963, + "memory(GiB)": 111.15, + "step": 13385, + "train_speed(iter/s)": 0.449672 + }, + { + "acc": 0.65384626, + "epoch": 0.33967529173008626, + "grad_norm": 5.5625, + "learning_rate": 9.612485658180178e-06, + "loss": 1.63723431, + "memory(GiB)": 111.15, + "step": 13390, + "train_speed(iter/s)": 0.44976 + }, + { + "acc": 0.6560977, + "epoch": 0.3398021308980213, + "grad_norm": 5.28125, + "learning_rate": 9.612080783871882e-06, + "loss": 1.62229919, + "memory(GiB)": 111.15, + "step": 13395, + "train_speed(iter/s)": 0.449849 + }, + { + "acc": 0.64801359, + "epoch": 0.33992897006595635, + "grad_norm": 4.875, + "learning_rate": 9.611675706703682e-06, + "loss": 1.62219238, + "memory(GiB)": 111.15, + "step": 13400, + "train_speed(iter/s)": 0.449937 + }, + { + "epoch": 0.33992897006595635, + "eval_acc": 0.6388823554723548, + "eval_loss": 1.6347674131393433, + "eval_runtime": 113.4718, + "eval_samples_per_second": 56.137, + "eval_steps_per_second": 28.069, + "step": 13400 + }, + { + "acc": 0.63912477, + "epoch": 0.34005580923389145, + "grad_norm": 6.34375, + "learning_rate": 9.611270426693395e-06, + "loss": 1.69518585, + "memory(GiB)": 111.15, + "step": 13405, + "train_speed(iter/s)": 0.44825 + }, + { + "acc": 0.61878805, + "epoch": 0.3401826484018265, + "grad_norm": 6.625, + "learning_rate": 9.610864943858847e-06, + "loss": 1.68853245, + "memory(GiB)": 111.15, + "step": 13410, + "train_speed(iter/s)": 0.448338 + }, + { + "acc": 0.65231886, + "epoch": 0.34030948756976154, + "grad_norm": 5.875, + "learning_rate": 9.61045925821787e-06, + "loss": 1.63459167, + "memory(GiB)": 111.15, + "step": 13415, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.63874903, + "epoch": 0.3404363267376966, + "grad_norm": 4.84375, + "learning_rate": 9.610053369788314e-06, + "loss": 1.67132282, + "memory(GiB)": 111.15, + "step": 13420, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.64549313, + "epoch": 0.3405631659056317, + "grad_norm": 5.09375, + "learning_rate": 9.609647278588027e-06, + "loss": 1.74758263, + "memory(GiB)": 111.15, + "step": 13425, + "train_speed(iter/s)": 0.448601 + }, + { + "acc": 0.64561815, + "epoch": 0.3406900050735667, + "grad_norm": 4.25, + "learning_rate": 9.609240984634871e-06, + "loss": 1.62978745, + "memory(GiB)": 111.15, + "step": 13430, + "train_speed(iter/s)": 0.448689 + }, + { + "acc": 0.65488977, + "epoch": 0.34081684424150177, + "grad_norm": 4.8125, + "learning_rate": 9.608834487946719e-06, + "loss": 1.65245972, + "memory(GiB)": 111.15, + "step": 13435, + "train_speed(iter/s)": 0.448776 + }, + { + "acc": 0.6309679, + "epoch": 0.3409436834094368, + "grad_norm": 5.46875, + "learning_rate": 9.60842778854145e-06, + "loss": 1.67359505, + "memory(GiB)": 111.15, + "step": 13440, + "train_speed(iter/s)": 0.448865 + }, + { + "acc": 0.65254688, + "epoch": 0.3410705225773719, + "grad_norm": 4.9375, + "learning_rate": 9.60802088643695e-06, + "loss": 1.62262802, + "memory(GiB)": 111.15, + "step": 13445, + "train_speed(iter/s)": 0.448953 + }, + { + "acc": 0.63615122, + "epoch": 0.34119736174530696, + "grad_norm": 4.625, + "learning_rate": 9.60761378165112e-06, + "loss": 1.64737816, + "memory(GiB)": 111.15, + "step": 13450, + "train_speed(iter/s)": 0.44904 + }, + { + "acc": 0.66723995, + "epoch": 0.341324200913242, + "grad_norm": 5.34375, + "learning_rate": 9.607206474201863e-06, + "loss": 1.61577473, + "memory(GiB)": 111.15, + "step": 13455, + "train_speed(iter/s)": 0.449127 + }, + { + "acc": 0.64097991, + "epoch": 0.34145104008117705, + "grad_norm": 5.28125, + "learning_rate": 9.606798964107096e-06, + "loss": 1.72571659, + "memory(GiB)": 111.15, + "step": 13460, + "train_speed(iter/s)": 0.449215 + }, + { + "acc": 0.66033754, + "epoch": 0.34157787924911215, + "grad_norm": 5.65625, + "learning_rate": 9.60639125138474e-06, + "loss": 1.54907274, + "memory(GiB)": 111.15, + "step": 13465, + "train_speed(iter/s)": 0.449304 + }, + { + "acc": 0.64955292, + "epoch": 0.3417047184170472, + "grad_norm": 5.40625, + "learning_rate": 9.605983336052735e-06, + "loss": 1.64906578, + "memory(GiB)": 111.15, + "step": 13470, + "train_speed(iter/s)": 0.449392 + }, + { + "acc": 0.63080535, + "epoch": 0.34183155758498224, + "grad_norm": 5.25, + "learning_rate": 9.605575218129017e-06, + "loss": 1.66043396, + "memory(GiB)": 111.15, + "step": 13475, + "train_speed(iter/s)": 0.44948 + }, + { + "acc": 0.63029566, + "epoch": 0.3419583967529173, + "grad_norm": 5.25, + "learning_rate": 9.605166897631539e-06, + "loss": 1.72559814, + "memory(GiB)": 111.15, + "step": 13480, + "train_speed(iter/s)": 0.449567 + }, + { + "acc": 0.6558557, + "epoch": 0.3420852359208524, + "grad_norm": 5.5625, + "learning_rate": 9.604758374578259e-06, + "loss": 1.65277557, + "memory(GiB)": 111.15, + "step": 13485, + "train_speed(iter/s)": 0.449655 + }, + { + "acc": 0.64086399, + "epoch": 0.3422120750887874, + "grad_norm": 4.96875, + "learning_rate": 9.604349648987148e-06, + "loss": 1.68258934, + "memory(GiB)": 111.15, + "step": 13490, + "train_speed(iter/s)": 0.449742 + }, + { + "acc": 0.64346614, + "epoch": 0.34233891425672247, + "grad_norm": 5.96875, + "learning_rate": 9.603940720876181e-06, + "loss": 1.67253494, + "memory(GiB)": 111.15, + "step": 13495, + "train_speed(iter/s)": 0.44983 + }, + { + "acc": 0.6423337, + "epoch": 0.3424657534246575, + "grad_norm": 5.15625, + "learning_rate": 9.603531590263348e-06, + "loss": 1.6550499, + "memory(GiB)": 111.15, + "step": 13500, + "train_speed(iter/s)": 0.449918 + }, + { + "epoch": 0.3424657534246575, + "eval_acc": 0.6389015727685899, + "eval_loss": 1.634238600730896, + "eval_runtime": 113.527, + "eval_samples_per_second": 56.11, + "eval_steps_per_second": 28.055, + "step": 13500 + }, + { + "acc": 0.64177632, + "epoch": 0.3425925925925926, + "grad_norm": 4.59375, + "learning_rate": 9.603122257166641e-06, + "loss": 1.66742134, + "memory(GiB)": 111.15, + "step": 13505, + "train_speed(iter/s)": 0.448243 + }, + { + "acc": 0.65725021, + "epoch": 0.34271943176052766, + "grad_norm": 5.15625, + "learning_rate": 9.602712721604066e-06, + "loss": 1.61912212, + "memory(GiB)": 111.15, + "step": 13510, + "train_speed(iter/s)": 0.448331 + }, + { + "acc": 0.63438482, + "epoch": 0.3428462709284627, + "grad_norm": 6.0625, + "learning_rate": 9.602302983593637e-06, + "loss": 1.7111002, + "memory(GiB)": 111.15, + "step": 13515, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.64421062, + "epoch": 0.34297311009639775, + "grad_norm": 5.84375, + "learning_rate": 9.601893043153372e-06, + "loss": 1.65377235, + "memory(GiB)": 111.15, + "step": 13520, + "train_speed(iter/s)": 0.448505 + }, + { + "acc": 0.64046326, + "epoch": 0.34309994926433285, + "grad_norm": 5.125, + "learning_rate": 9.601482900301308e-06, + "loss": 1.62992496, + "memory(GiB)": 111.15, + "step": 13525, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.65164299, + "epoch": 0.3432267884322679, + "grad_norm": 5.40625, + "learning_rate": 9.60107255505548e-06, + "loss": 1.66345749, + "memory(GiB)": 111.15, + "step": 13530, + "train_speed(iter/s)": 0.44868 + }, + { + "acc": 0.63073874, + "epoch": 0.34335362760020294, + "grad_norm": 4.8125, + "learning_rate": 9.60066200743394e-06, + "loss": 1.77850189, + "memory(GiB)": 111.15, + "step": 13535, + "train_speed(iter/s)": 0.448769 + }, + { + "acc": 0.64757633, + "epoch": 0.343480466768138, + "grad_norm": 6.15625, + "learning_rate": 9.600251257454744e-06, + "loss": 1.63806038, + "memory(GiB)": 111.15, + "step": 13540, + "train_speed(iter/s)": 0.448858 + }, + { + "acc": 0.63056636, + "epoch": 0.3436073059360731, + "grad_norm": 5.0, + "learning_rate": 9.599840305135959e-06, + "loss": 1.67879333, + "memory(GiB)": 111.15, + "step": 13545, + "train_speed(iter/s)": 0.448945 + }, + { + "acc": 0.64656172, + "epoch": 0.3437341451040081, + "grad_norm": 5.15625, + "learning_rate": 9.59942915049566e-06, + "loss": 1.6023735, + "memory(GiB)": 111.15, + "step": 13550, + "train_speed(iter/s)": 0.449033 + }, + { + "acc": 0.63602276, + "epoch": 0.34386098427194317, + "grad_norm": 5.0, + "learning_rate": 9.599017793551933e-06, + "loss": 1.65858822, + "memory(GiB)": 111.15, + "step": 13555, + "train_speed(iter/s)": 0.449121 + }, + { + "acc": 0.63266535, + "epoch": 0.3439878234398782, + "grad_norm": 5.46875, + "learning_rate": 9.598606234322869e-06, + "loss": 1.75436592, + "memory(GiB)": 111.15, + "step": 13560, + "train_speed(iter/s)": 0.44921 + }, + { + "acc": 0.65314236, + "epoch": 0.3441146626078133, + "grad_norm": 5.0625, + "learning_rate": 9.598194472826574e-06, + "loss": 1.64844475, + "memory(GiB)": 111.15, + "step": 13565, + "train_speed(iter/s)": 0.449297 + }, + { + "acc": 0.64052863, + "epoch": 0.34424150177574836, + "grad_norm": 6.03125, + "learning_rate": 9.597782509081154e-06, + "loss": 1.6744709, + "memory(GiB)": 111.15, + "step": 13570, + "train_speed(iter/s)": 0.449385 + }, + { + "acc": 0.64789963, + "epoch": 0.3443683409436834, + "grad_norm": 4.5625, + "learning_rate": 9.597370343104733e-06, + "loss": 1.60511379, + "memory(GiB)": 111.15, + "step": 13575, + "train_speed(iter/s)": 0.449473 + }, + { + "acc": 0.66312146, + "epoch": 0.34449518011161845, + "grad_norm": 5.28125, + "learning_rate": 9.596957974915438e-06, + "loss": 1.5928648, + "memory(GiB)": 111.15, + "step": 13580, + "train_speed(iter/s)": 0.44956 + }, + { + "acc": 0.64366326, + "epoch": 0.34462201927955355, + "grad_norm": 4.28125, + "learning_rate": 9.596545404531408e-06, + "loss": 1.66371078, + "memory(GiB)": 111.15, + "step": 13585, + "train_speed(iter/s)": 0.449648 + }, + { + "acc": 0.63196435, + "epoch": 0.3447488584474886, + "grad_norm": 4.6875, + "learning_rate": 9.596132631970788e-06, + "loss": 1.64398956, + "memory(GiB)": 111.15, + "step": 13590, + "train_speed(iter/s)": 0.449737 + }, + { + "acc": 0.65817604, + "epoch": 0.34487569761542364, + "grad_norm": 5.375, + "learning_rate": 9.595719657251735e-06, + "loss": 1.60683346, + "memory(GiB)": 111.15, + "step": 13595, + "train_speed(iter/s)": 0.449825 + }, + { + "acc": 0.6601769, + "epoch": 0.3450025367833587, + "grad_norm": 5.34375, + "learning_rate": 9.595306480392413e-06, + "loss": 1.62262859, + "memory(GiB)": 111.15, + "step": 13600, + "train_speed(iter/s)": 0.449913 + }, + { + "epoch": 0.3450025367833587, + "eval_acc": 0.6388364010683145, + "eval_loss": 1.6343255043029785, + "eval_runtime": 113.2881, + "eval_samples_per_second": 56.228, + "eval_steps_per_second": 28.114, + "step": 13600 + }, + { + "acc": 0.64074392, + "epoch": 0.3451293759512938, + "grad_norm": 6.15625, + "learning_rate": 9.594893101410995e-06, + "loss": 1.69255333, + "memory(GiB)": 111.15, + "step": 13605, + "train_speed(iter/s)": 0.448253 + }, + { + "acc": 0.63475628, + "epoch": 0.3452562151192288, + "grad_norm": 5.0, + "learning_rate": 9.594479520325665e-06, + "loss": 1.6626339, + "memory(GiB)": 111.15, + "step": 13610, + "train_speed(iter/s)": 0.448339 + }, + { + "acc": 0.63485117, + "epoch": 0.34538305428716387, + "grad_norm": 5.75, + "learning_rate": 9.594065737154611e-06, + "loss": 1.63332291, + "memory(GiB)": 111.15, + "step": 13615, + "train_speed(iter/s)": 0.448426 + }, + { + "acc": 0.64619331, + "epoch": 0.3455098934550989, + "grad_norm": 5.5625, + "learning_rate": 9.593651751916037e-06, + "loss": 1.62646847, + "memory(GiB)": 111.15, + "step": 13620, + "train_speed(iter/s)": 0.448512 + }, + { + "acc": 0.66380672, + "epoch": 0.345636732623034, + "grad_norm": 6.34375, + "learning_rate": 9.593237564628149e-06, + "loss": 1.54840393, + "memory(GiB)": 111.15, + "step": 13625, + "train_speed(iter/s)": 0.448599 + }, + { + "acc": 0.66192174, + "epoch": 0.34576357179096906, + "grad_norm": 4.34375, + "learning_rate": 9.592823175309164e-06, + "loss": 1.5609189, + "memory(GiB)": 111.15, + "step": 13630, + "train_speed(iter/s)": 0.448684 + }, + { + "acc": 0.65221529, + "epoch": 0.3458904109589041, + "grad_norm": 4.34375, + "learning_rate": 9.592408583977311e-06, + "loss": 1.61033363, + "memory(GiB)": 111.15, + "step": 13635, + "train_speed(iter/s)": 0.44877 + }, + { + "acc": 0.63230114, + "epoch": 0.34601725012683915, + "grad_norm": 5.21875, + "learning_rate": 9.591993790650826e-06, + "loss": 1.62142143, + "memory(GiB)": 111.15, + "step": 13640, + "train_speed(iter/s)": 0.448858 + }, + { + "acc": 0.65384998, + "epoch": 0.34614408929477425, + "grad_norm": 5.75, + "learning_rate": 9.591578795347952e-06, + "loss": 1.66564922, + "memory(GiB)": 111.15, + "step": 13645, + "train_speed(iter/s)": 0.448946 + }, + { + "acc": 0.64846134, + "epoch": 0.3462709284627093, + "grad_norm": 6.03125, + "learning_rate": 9.591163598086943e-06, + "loss": 1.6676302, + "memory(GiB)": 111.15, + "step": 13650, + "train_speed(iter/s)": 0.449034 + }, + { + "acc": 0.64900293, + "epoch": 0.34639776763064434, + "grad_norm": 4.84375, + "learning_rate": 9.59074819888606e-06, + "loss": 1.65944977, + "memory(GiB)": 111.15, + "step": 13655, + "train_speed(iter/s)": 0.449122 + }, + { + "acc": 0.64978294, + "epoch": 0.3465246067985794, + "grad_norm": 5.1875, + "learning_rate": 9.590332597763575e-06, + "loss": 1.60361557, + "memory(GiB)": 111.15, + "step": 13660, + "train_speed(iter/s)": 0.449209 + }, + { + "acc": 0.64699898, + "epoch": 0.3466514459665145, + "grad_norm": 5.3125, + "learning_rate": 9.589916794737768e-06, + "loss": 1.69255486, + "memory(GiB)": 111.15, + "step": 13665, + "train_speed(iter/s)": 0.449297 + }, + { + "acc": 0.64601269, + "epoch": 0.3467782851344495, + "grad_norm": 5.40625, + "learning_rate": 9.589500789826927e-06, + "loss": 1.7080019, + "memory(GiB)": 111.15, + "step": 13670, + "train_speed(iter/s)": 0.449384 + }, + { + "acc": 0.63440256, + "epoch": 0.34690512430238457, + "grad_norm": 5.65625, + "learning_rate": 9.589084583049353e-06, + "loss": 1.7775383, + "memory(GiB)": 111.15, + "step": 13675, + "train_speed(iter/s)": 0.449471 + }, + { + "acc": 0.64539881, + "epoch": 0.3470319634703196, + "grad_norm": 6.15625, + "learning_rate": 9.588668174423348e-06, + "loss": 1.65382347, + "memory(GiB)": 111.15, + "step": 13680, + "train_speed(iter/s)": 0.449559 + }, + { + "acc": 0.62981782, + "epoch": 0.3471588026382547, + "grad_norm": 5.09375, + "learning_rate": 9.588251563967232e-06, + "loss": 1.70327568, + "memory(GiB)": 111.15, + "step": 13685, + "train_speed(iter/s)": 0.449645 + }, + { + "acc": 0.63958158, + "epoch": 0.34728564180618976, + "grad_norm": 4.78125, + "learning_rate": 9.587834751699326e-06, + "loss": 1.7377182, + "memory(GiB)": 111.15, + "step": 13690, + "train_speed(iter/s)": 0.449732 + }, + { + "acc": 0.64236112, + "epoch": 0.3474124809741248, + "grad_norm": 5.8125, + "learning_rate": 9.587417737637963e-06, + "loss": 1.69043465, + "memory(GiB)": 111.15, + "step": 13695, + "train_speed(iter/s)": 0.449819 + }, + { + "acc": 0.65243244, + "epoch": 0.34753932014205985, + "grad_norm": 5.4375, + "learning_rate": 9.587000521801488e-06, + "loss": 1.60939236, + "memory(GiB)": 111.15, + "step": 13700, + "train_speed(iter/s)": 0.449905 + }, + { + "epoch": 0.34753932014205985, + "eval_acc": 0.6389550469842005, + "eval_loss": 1.6337294578552246, + "eval_runtime": 114.2841, + "eval_samples_per_second": 55.738, + "eval_steps_per_second": 27.869, + "step": 13700 + }, + { + "acc": 0.65224667, + "epoch": 0.34766615930999495, + "grad_norm": 5.6875, + "learning_rate": 9.58658310420825e-06, + "loss": 1.64363079, + "memory(GiB)": 111.15, + "step": 13705, + "train_speed(iter/s)": 0.448243 + }, + { + "acc": 0.64518862, + "epoch": 0.34779299847793, + "grad_norm": 5.65625, + "learning_rate": 9.58616548487661e-06, + "loss": 1.60551453, + "memory(GiB)": 111.15, + "step": 13710, + "train_speed(iter/s)": 0.448331 + }, + { + "acc": 0.64422979, + "epoch": 0.34791983764586504, + "grad_norm": 5.90625, + "learning_rate": 9.585747663824936e-06, + "loss": 1.72791748, + "memory(GiB)": 111.15, + "step": 13715, + "train_speed(iter/s)": 0.448418 + }, + { + "acc": 0.65144629, + "epoch": 0.3480466768138001, + "grad_norm": 5.5, + "learning_rate": 9.585329641071606e-06, + "loss": 1.62911205, + "memory(GiB)": 111.15, + "step": 13720, + "train_speed(iter/s)": 0.448506 + }, + { + "acc": 0.64711027, + "epoch": 0.3481735159817352, + "grad_norm": 6.15625, + "learning_rate": 9.584911416635007e-06, + "loss": 1.64262257, + "memory(GiB)": 111.15, + "step": 13725, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.64410057, + "epoch": 0.3483003551496702, + "grad_norm": 4.90625, + "learning_rate": 9.584492990533533e-06, + "loss": 1.65934143, + "memory(GiB)": 111.15, + "step": 13730, + "train_speed(iter/s)": 0.448681 + }, + { + "acc": 0.65518126, + "epoch": 0.34842719431760527, + "grad_norm": 5.53125, + "learning_rate": 9.58407436278559e-06, + "loss": 1.56461582, + "memory(GiB)": 111.15, + "step": 13735, + "train_speed(iter/s)": 0.448768 + }, + { + "acc": 0.63987513, + "epoch": 0.3485540334855403, + "grad_norm": 4.84375, + "learning_rate": 9.583655533409588e-06, + "loss": 1.65193958, + "memory(GiB)": 111.15, + "step": 13740, + "train_speed(iter/s)": 0.448856 + }, + { + "acc": 0.64441519, + "epoch": 0.3486808726534754, + "grad_norm": 4.40625, + "learning_rate": 9.583236502423952e-06, + "loss": 1.67047768, + "memory(GiB)": 111.15, + "step": 13745, + "train_speed(iter/s)": 0.448943 + }, + { + "acc": 0.64345765, + "epoch": 0.34880771182141046, + "grad_norm": 5.28125, + "learning_rate": 9.582817269847112e-06, + "loss": 1.71998463, + "memory(GiB)": 111.15, + "step": 13750, + "train_speed(iter/s)": 0.449032 + }, + { + "acc": 0.65165386, + "epoch": 0.3489345509893455, + "grad_norm": 4.8125, + "learning_rate": 9.582397835697509e-06, + "loss": 1.59828377, + "memory(GiB)": 111.15, + "step": 13755, + "train_speed(iter/s)": 0.449119 + }, + { + "acc": 0.65344687, + "epoch": 0.34906139015728055, + "grad_norm": 4.90625, + "learning_rate": 9.581978199993587e-06, + "loss": 1.61674061, + "memory(GiB)": 111.15, + "step": 13760, + "train_speed(iter/s)": 0.449206 + }, + { + "acc": 0.64435925, + "epoch": 0.34918822932521565, + "grad_norm": 4.65625, + "learning_rate": 9.58155836275381e-06, + "loss": 1.63897324, + "memory(GiB)": 111.15, + "step": 13765, + "train_speed(iter/s)": 0.449293 + }, + { + "acc": 0.64087219, + "epoch": 0.3493150684931507, + "grad_norm": 4.65625, + "learning_rate": 9.581138323996639e-06, + "loss": 1.70425644, + "memory(GiB)": 111.15, + "step": 13770, + "train_speed(iter/s)": 0.449379 + }, + { + "acc": 0.65538464, + "epoch": 0.34944190766108574, + "grad_norm": 5.34375, + "learning_rate": 9.580718083740553e-06, + "loss": 1.69704266, + "memory(GiB)": 111.15, + "step": 13775, + "train_speed(iter/s)": 0.449466 + }, + { + "acc": 0.67970333, + "epoch": 0.3495687468290208, + "grad_norm": 5.1875, + "learning_rate": 9.580297642004032e-06, + "loss": 1.5640316, + "memory(GiB)": 111.15, + "step": 13780, + "train_speed(iter/s)": 0.449552 + }, + { + "acc": 0.64756861, + "epoch": 0.3496955859969559, + "grad_norm": 4.75, + "learning_rate": 9.579876998805573e-06, + "loss": 1.70120335, + "memory(GiB)": 111.15, + "step": 13785, + "train_speed(iter/s)": 0.449639 + }, + { + "acc": 0.64993563, + "epoch": 0.3498224251648909, + "grad_norm": 5.46875, + "learning_rate": 9.579456154163676e-06, + "loss": 1.66384811, + "memory(GiB)": 111.15, + "step": 13790, + "train_speed(iter/s)": 0.449726 + }, + { + "acc": 0.63879771, + "epoch": 0.34994926433282597, + "grad_norm": 5.15625, + "learning_rate": 9.57903510809685e-06, + "loss": 1.63221931, + "memory(GiB)": 111.15, + "step": 13795, + "train_speed(iter/s)": 0.449813 + }, + { + "acc": 0.64708543, + "epoch": 0.350076103500761, + "grad_norm": 6.0, + "learning_rate": 9.578613860623617e-06, + "loss": 1.6336792, + "memory(GiB)": 111.15, + "step": 13800, + "train_speed(iter/s)": 0.449899 + }, + { + "epoch": 0.350076103500761, + "eval_acc": 0.6388978128628048, + "eval_loss": 1.633205771446228, + "eval_runtime": 113.431, + "eval_samples_per_second": 56.158, + "eval_steps_per_second": 28.079, + "step": 13800 + }, + { + "acc": 0.65197029, + "epoch": 0.3502029426686961, + "grad_norm": 5.5625, + "learning_rate": 9.578192411762503e-06, + "loss": 1.55447464, + "memory(GiB)": 111.15, + "step": 13805, + "train_speed(iter/s)": 0.448262 + }, + { + "acc": 0.64415607, + "epoch": 0.35032978183663116, + "grad_norm": 4.375, + "learning_rate": 9.577770761532049e-06, + "loss": 1.68373661, + "memory(GiB)": 111.15, + "step": 13810, + "train_speed(iter/s)": 0.448348 + }, + { + "acc": 0.64257393, + "epoch": 0.3504566210045662, + "grad_norm": 5.84375, + "learning_rate": 9.577348909950797e-06, + "loss": 1.66234207, + "memory(GiB)": 111.15, + "step": 13815, + "train_speed(iter/s)": 0.448434 + }, + { + "acc": 0.6567091, + "epoch": 0.35058346017250125, + "grad_norm": 5.25, + "learning_rate": 9.576926857037303e-06, + "loss": 1.50343027, + "memory(GiB)": 111.15, + "step": 13820, + "train_speed(iter/s)": 0.448521 + }, + { + "acc": 0.66023369, + "epoch": 0.35071029934043635, + "grad_norm": 5.53125, + "learning_rate": 9.576504602810133e-06, + "loss": 1.64527283, + "memory(GiB)": 111.15, + "step": 13825, + "train_speed(iter/s)": 0.448608 + }, + { + "acc": 0.64658866, + "epoch": 0.3508371385083714, + "grad_norm": 5.125, + "learning_rate": 9.576082147287858e-06, + "loss": 1.62077484, + "memory(GiB)": 111.15, + "step": 13830, + "train_speed(iter/s)": 0.448694 + }, + { + "acc": 0.64606075, + "epoch": 0.35096397767630644, + "grad_norm": 5.53125, + "learning_rate": 9.575659490489058e-06, + "loss": 1.59905949, + "memory(GiB)": 111.15, + "step": 13835, + "train_speed(iter/s)": 0.448781 + }, + { + "acc": 0.64534016, + "epoch": 0.3510908168442415, + "grad_norm": 5.15625, + "learning_rate": 9.575236632432325e-06, + "loss": 1.6225029, + "memory(GiB)": 111.15, + "step": 13840, + "train_speed(iter/s)": 0.448869 + }, + { + "acc": 0.65817809, + "epoch": 0.3512176560121766, + "grad_norm": 5.75, + "learning_rate": 9.574813573136259e-06, + "loss": 1.62662258, + "memory(GiB)": 111.15, + "step": 13845, + "train_speed(iter/s)": 0.448956 + }, + { + "acc": 0.65324006, + "epoch": 0.3513444951801116, + "grad_norm": 4.625, + "learning_rate": 9.574390312619466e-06, + "loss": 1.65058708, + "memory(GiB)": 111.15, + "step": 13850, + "train_speed(iter/s)": 0.449043 + }, + { + "acc": 0.65746584, + "epoch": 0.35147133434804667, + "grad_norm": 5.03125, + "learning_rate": 9.573966850900565e-06, + "loss": 1.61844025, + "memory(GiB)": 111.15, + "step": 13855, + "train_speed(iter/s)": 0.449129 + }, + { + "acc": 0.64182949, + "epoch": 0.3515981735159817, + "grad_norm": 5.59375, + "learning_rate": 9.57354318799818e-06, + "loss": 1.61763935, + "memory(GiB)": 111.15, + "step": 13860, + "train_speed(iter/s)": 0.449216 + }, + { + "acc": 0.64992461, + "epoch": 0.3517250126839168, + "grad_norm": 5.1875, + "learning_rate": 9.573119323930946e-06, + "loss": 1.64605808, + "memory(GiB)": 111.15, + "step": 13865, + "train_speed(iter/s)": 0.449303 + }, + { + "acc": 0.64900827, + "epoch": 0.35185185185185186, + "grad_norm": 4.75, + "learning_rate": 9.572695258717507e-06, + "loss": 1.63551769, + "memory(GiB)": 111.15, + "step": 13870, + "train_speed(iter/s)": 0.44939 + }, + { + "acc": 0.6451992, + "epoch": 0.3519786910197869, + "grad_norm": 4.96875, + "learning_rate": 9.572270992376513e-06, + "loss": 1.66867371, + "memory(GiB)": 111.15, + "step": 13875, + "train_speed(iter/s)": 0.449476 + }, + { + "acc": 0.65296707, + "epoch": 0.35210553018772195, + "grad_norm": 4.65625, + "learning_rate": 9.571846524926629e-06, + "loss": 1.64710693, + "memory(GiB)": 111.15, + "step": 13880, + "train_speed(iter/s)": 0.449564 + }, + { + "acc": 0.65420208, + "epoch": 0.35223236935565705, + "grad_norm": 4.90625, + "learning_rate": 9.571421856386522e-06, + "loss": 1.69653473, + "memory(GiB)": 111.15, + "step": 13885, + "train_speed(iter/s)": 0.44965 + }, + { + "acc": 0.64517431, + "epoch": 0.3523592085235921, + "grad_norm": 5.0625, + "learning_rate": 9.570996986774872e-06, + "loss": 1.65888557, + "memory(GiB)": 111.15, + "step": 13890, + "train_speed(iter/s)": 0.449736 + }, + { + "acc": 0.64347372, + "epoch": 0.35248604769152714, + "grad_norm": 6.34375, + "learning_rate": 9.570571916110366e-06, + "loss": 1.67334843, + "memory(GiB)": 111.15, + "step": 13895, + "train_speed(iter/s)": 0.449822 + }, + { + "acc": 0.64887338, + "epoch": 0.3526128868594622, + "grad_norm": 4.875, + "learning_rate": 9.570146644411705e-06, + "loss": 1.66215553, + "memory(GiB)": 111.15, + "step": 13900, + "train_speed(iter/s)": 0.449909 + }, + { + "epoch": 0.3526128868594622, + "eval_acc": 0.639013116640215, + "eval_loss": 1.6327142715454102, + "eval_runtime": 112.9348, + "eval_samples_per_second": 56.404, + "eval_steps_per_second": 28.202, + "step": 13900 + }, + { + "acc": 0.65493631, + "epoch": 0.3527397260273973, + "grad_norm": 5.40625, + "learning_rate": 9.569721171697587e-06, + "loss": 1.65564365, + "memory(GiB)": 111.15, + "step": 13905, + "train_speed(iter/s)": 0.448292 + }, + { + "acc": 0.64626961, + "epoch": 0.3528665651953323, + "grad_norm": 4.71875, + "learning_rate": 9.569295497986727e-06, + "loss": 1.67892075, + "memory(GiB)": 111.15, + "step": 13910, + "train_speed(iter/s)": 0.448378 + }, + { + "acc": 0.65952797, + "epoch": 0.35299340436326737, + "grad_norm": 5.34375, + "learning_rate": 9.568869623297855e-06, + "loss": 1.63244972, + "memory(GiB)": 111.15, + "step": 13915, + "train_speed(iter/s)": 0.448464 + }, + { + "acc": 0.6325058, + "epoch": 0.3531202435312024, + "grad_norm": 4.90625, + "learning_rate": 9.568443547649697e-06, + "loss": 1.73353481, + "memory(GiB)": 111.15, + "step": 13920, + "train_speed(iter/s)": 0.448551 + }, + { + "acc": 0.65298038, + "epoch": 0.3532470826991375, + "grad_norm": 5.84375, + "learning_rate": 9.568017271060994e-06, + "loss": 1.66758442, + "memory(GiB)": 111.15, + "step": 13925, + "train_speed(iter/s)": 0.448637 + }, + { + "acc": 0.66020384, + "epoch": 0.35337392186707256, + "grad_norm": 6.9375, + "learning_rate": 9.567590793550498e-06, + "loss": 1.59989319, + "memory(GiB)": 111.15, + "step": 13930, + "train_speed(iter/s)": 0.448723 + }, + { + "acc": 0.64962511, + "epoch": 0.3535007610350076, + "grad_norm": 5.3125, + "learning_rate": 9.567164115136965e-06, + "loss": 1.6726553, + "memory(GiB)": 111.15, + "step": 13935, + "train_speed(iter/s)": 0.448809 + }, + { + "acc": 0.65230341, + "epoch": 0.35362760020294265, + "grad_norm": 4.96875, + "learning_rate": 9.566737235839166e-06, + "loss": 1.64690189, + "memory(GiB)": 111.15, + "step": 13940, + "train_speed(iter/s)": 0.448896 + }, + { + "acc": 0.63917646, + "epoch": 0.35375443937087775, + "grad_norm": 5.5, + "learning_rate": 9.566310155675871e-06, + "loss": 1.65717449, + "memory(GiB)": 111.15, + "step": 13945, + "train_speed(iter/s)": 0.448983 + }, + { + "acc": 0.63990064, + "epoch": 0.3538812785388128, + "grad_norm": 4.0625, + "learning_rate": 9.56588287466587e-06, + "loss": 1.58915138, + "memory(GiB)": 111.15, + "step": 13950, + "train_speed(iter/s)": 0.44907 + }, + { + "acc": 0.66091614, + "epoch": 0.35400811770674784, + "grad_norm": 5.46875, + "learning_rate": 9.565455392827954e-06, + "loss": 1.59982376, + "memory(GiB)": 111.15, + "step": 13955, + "train_speed(iter/s)": 0.449156 + }, + { + "acc": 0.64128933, + "epoch": 0.3541349568746829, + "grad_norm": 4.96875, + "learning_rate": 9.565027710180927e-06, + "loss": 1.6735693, + "memory(GiB)": 111.15, + "step": 13960, + "train_speed(iter/s)": 0.449242 + }, + { + "acc": 0.66160359, + "epoch": 0.354261796042618, + "grad_norm": 5.59375, + "learning_rate": 9.5645998267436e-06, + "loss": 1.54278316, + "memory(GiB)": 111.15, + "step": 13965, + "train_speed(iter/s)": 0.449329 + }, + { + "acc": 0.63823557, + "epoch": 0.354388635210553, + "grad_norm": 4.625, + "learning_rate": 9.564171742534794e-06, + "loss": 1.66355858, + "memory(GiB)": 111.15, + "step": 13970, + "train_speed(iter/s)": 0.449415 + }, + { + "acc": 0.6439827, + "epoch": 0.35451547437848807, + "grad_norm": 5.25, + "learning_rate": 9.563743457573336e-06, + "loss": 1.67273636, + "memory(GiB)": 111.15, + "step": 13975, + "train_speed(iter/s)": 0.4495 + }, + { + "acc": 0.63638186, + "epoch": 0.3546423135464231, + "grad_norm": 4.875, + "learning_rate": 9.563314971878065e-06, + "loss": 1.68128262, + "memory(GiB)": 111.15, + "step": 13980, + "train_speed(iter/s)": 0.449585 + }, + { + "acc": 0.64579544, + "epoch": 0.3547691527143582, + "grad_norm": 5.28125, + "learning_rate": 9.562886285467828e-06, + "loss": 1.65092735, + "memory(GiB)": 111.15, + "step": 13985, + "train_speed(iter/s)": 0.449672 + }, + { + "acc": 0.63931093, + "epoch": 0.35489599188229326, + "grad_norm": 5.28125, + "learning_rate": 9.56245739836148e-06, + "loss": 1.67836208, + "memory(GiB)": 111.15, + "step": 13990, + "train_speed(iter/s)": 0.449757 + }, + { + "acc": 0.6450654, + "epoch": 0.3550228310502283, + "grad_norm": 4.875, + "learning_rate": 9.562028310577887e-06, + "loss": 1.6372509, + "memory(GiB)": 111.15, + "step": 13995, + "train_speed(iter/s)": 0.449843 + }, + { + "acc": 0.63708959, + "epoch": 0.35514967021816335, + "grad_norm": 5.15625, + "learning_rate": 9.56159902213592e-06, + "loss": 1.67382431, + "memory(GiB)": 111.15, + "step": 14000, + "train_speed(iter/s)": 0.449929 + }, + { + "epoch": 0.35514967021816335, + "eval_acc": 0.6390778705731809, + "eval_loss": 1.632528305053711, + "eval_runtime": 113.9799, + "eval_samples_per_second": 55.887, + "eval_steps_per_second": 27.944, + "step": 14000 + }, + { + "acc": 0.64792118, + "epoch": 0.35527650938609845, + "grad_norm": 5.1875, + "learning_rate": 9.561169533054462e-06, + "loss": 1.63665886, + "memory(GiB)": 111.15, + "step": 14005, + "train_speed(iter/s)": 0.448306 + }, + { + "acc": 0.65470276, + "epoch": 0.3554033485540335, + "grad_norm": 6.65625, + "learning_rate": 9.560739843352404e-06, + "loss": 1.65799561, + "memory(GiB)": 111.15, + "step": 14010, + "train_speed(iter/s)": 0.448391 + }, + { + "acc": 0.64416289, + "epoch": 0.35553018772196854, + "grad_norm": 5.28125, + "learning_rate": 9.560309953048645e-06, + "loss": 1.68351078, + "memory(GiB)": 111.15, + "step": 14015, + "train_speed(iter/s)": 0.448476 + }, + { + "acc": 0.63143368, + "epoch": 0.3556570268899036, + "grad_norm": 5.09375, + "learning_rate": 9.559879862162095e-06, + "loss": 1.69894142, + "memory(GiB)": 111.15, + "step": 14020, + "train_speed(iter/s)": 0.448561 + }, + { + "acc": 0.6448947, + "epoch": 0.3557838660578387, + "grad_norm": 5.4375, + "learning_rate": 9.55944957071167e-06, + "loss": 1.70590954, + "memory(GiB)": 111.15, + "step": 14025, + "train_speed(iter/s)": 0.448646 + }, + { + "acc": 0.64087458, + "epoch": 0.3559107052257737, + "grad_norm": 5.75, + "learning_rate": 9.559019078716295e-06, + "loss": 1.65334091, + "memory(GiB)": 111.15, + "step": 14030, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.64791336, + "epoch": 0.35603754439370877, + "grad_norm": 5.96875, + "learning_rate": 9.558588386194907e-06, + "loss": 1.70946712, + "memory(GiB)": 111.15, + "step": 14035, + "train_speed(iter/s)": 0.448817 + }, + { + "acc": 0.63646221, + "epoch": 0.3561643835616438, + "grad_norm": 6.1875, + "learning_rate": 9.55815749316645e-06, + "loss": 1.62360001, + "memory(GiB)": 111.15, + "step": 14040, + "train_speed(iter/s)": 0.448901 + }, + { + "acc": 0.63083696, + "epoch": 0.3562912227295789, + "grad_norm": 5.28125, + "learning_rate": 9.557726399649875e-06, + "loss": 1.69806862, + "memory(GiB)": 111.15, + "step": 14045, + "train_speed(iter/s)": 0.448987 + }, + { + "acc": 0.63959489, + "epoch": 0.35641806189751396, + "grad_norm": 5.65625, + "learning_rate": 9.557295105664144e-06, + "loss": 1.66109428, + "memory(GiB)": 111.15, + "step": 14050, + "train_speed(iter/s)": 0.449072 + }, + { + "acc": 0.63467441, + "epoch": 0.356544901065449, + "grad_norm": 5.625, + "learning_rate": 9.556863611228228e-06, + "loss": 1.72277565, + "memory(GiB)": 111.15, + "step": 14055, + "train_speed(iter/s)": 0.449158 + }, + { + "acc": 0.65098476, + "epoch": 0.35667174023338405, + "grad_norm": 5.53125, + "learning_rate": 9.556431916361105e-06, + "loss": 1.66806259, + "memory(GiB)": 111.15, + "step": 14060, + "train_speed(iter/s)": 0.449243 + }, + { + "acc": 0.63266835, + "epoch": 0.35679857940131915, + "grad_norm": 5.3125, + "learning_rate": 9.556000021081764e-06, + "loss": 1.65586433, + "memory(GiB)": 111.15, + "step": 14065, + "train_speed(iter/s)": 0.449328 + }, + { + "acc": 0.65270166, + "epoch": 0.3569254185692542, + "grad_norm": 6.15625, + "learning_rate": 9.5555679254092e-06, + "loss": 1.58572121, + "memory(GiB)": 111.15, + "step": 14070, + "train_speed(iter/s)": 0.449413 + }, + { + "acc": 0.63574028, + "epoch": 0.35705225773718924, + "grad_norm": 5.6875, + "learning_rate": 9.55513562936242e-06, + "loss": 1.73494415, + "memory(GiB)": 111.15, + "step": 14075, + "train_speed(iter/s)": 0.449498 + }, + { + "acc": 0.63240619, + "epoch": 0.3571790969051243, + "grad_norm": 5.65625, + "learning_rate": 9.554703132960437e-06, + "loss": 1.74862785, + "memory(GiB)": 111.15, + "step": 14080, + "train_speed(iter/s)": 0.449584 + }, + { + "acc": 0.65498648, + "epoch": 0.3573059360730594, + "grad_norm": 4.34375, + "learning_rate": 9.554270436222277e-06, + "loss": 1.57103758, + "memory(GiB)": 111.15, + "step": 14085, + "train_speed(iter/s)": 0.449669 + }, + { + "acc": 0.64250689, + "epoch": 0.3574327752409944, + "grad_norm": 5.25, + "learning_rate": 9.553837539166969e-06, + "loss": 1.63600712, + "memory(GiB)": 111.15, + "step": 14090, + "train_speed(iter/s)": 0.449754 + }, + { + "acc": 0.64493866, + "epoch": 0.35755961440892947, + "grad_norm": 5.34375, + "learning_rate": 9.553404441813554e-06, + "loss": 1.7062685, + "memory(GiB)": 111.15, + "step": 14095, + "train_speed(iter/s)": 0.44984 + }, + { + "acc": 0.65040212, + "epoch": 0.3576864535768645, + "grad_norm": 7.21875, + "learning_rate": 9.552971144181083e-06, + "loss": 1.62010555, + "memory(GiB)": 111.15, + "step": 14100, + "train_speed(iter/s)": 0.449924 + }, + { + "epoch": 0.3576864535768645, + "eval_acc": 0.6392600171201044, + "eval_loss": 1.6321802139282227, + "eval_runtime": 113.2553, + "eval_samples_per_second": 56.245, + "eval_steps_per_second": 28.122, + "step": 14100 + }, + { + "acc": 0.64428568, + "epoch": 0.3578132927447996, + "grad_norm": 4.9375, + "learning_rate": 9.552537646288612e-06, + "loss": 1.69233932, + "memory(GiB)": 111.15, + "step": 14105, + "train_speed(iter/s)": 0.448325 + }, + { + "acc": 0.6589201, + "epoch": 0.35794013191273466, + "grad_norm": 5.34375, + "learning_rate": 9.552103948155211e-06, + "loss": 1.61974697, + "memory(GiB)": 111.15, + "step": 14110, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.65990062, + "epoch": 0.3580669710806697, + "grad_norm": 6.125, + "learning_rate": 9.551670049799954e-06, + "loss": 1.63631229, + "memory(GiB)": 111.15, + "step": 14115, + "train_speed(iter/s)": 0.448495 + }, + { + "acc": 0.6388341, + "epoch": 0.35819381024860475, + "grad_norm": 5.28125, + "learning_rate": 9.551235951241927e-06, + "loss": 1.6283268, + "memory(GiB)": 111.15, + "step": 14120, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.63620405, + "epoch": 0.35832064941653985, + "grad_norm": 5.4375, + "learning_rate": 9.550801652500223e-06, + "loss": 1.72039032, + "memory(GiB)": 111.15, + "step": 14125, + "train_speed(iter/s)": 0.448664 + }, + { + "acc": 0.64916501, + "epoch": 0.3584474885844749, + "grad_norm": 5.65625, + "learning_rate": 9.550367153593944e-06, + "loss": 1.63811283, + "memory(GiB)": 111.15, + "step": 14130, + "train_speed(iter/s)": 0.448749 + }, + { + "acc": 0.63713417, + "epoch": 0.35857432775240994, + "grad_norm": 4.625, + "learning_rate": 9.549932454542202e-06, + "loss": 1.66669979, + "memory(GiB)": 111.15, + "step": 14135, + "train_speed(iter/s)": 0.448833 + }, + { + "acc": 0.65722542, + "epoch": 0.358701166920345, + "grad_norm": 7.59375, + "learning_rate": 9.549497555364115e-06, + "loss": 1.62094383, + "memory(GiB)": 111.15, + "step": 14140, + "train_speed(iter/s)": 0.448919 + }, + { + "acc": 0.64861345, + "epoch": 0.3588280060882801, + "grad_norm": 6.15625, + "learning_rate": 9.549062456078816e-06, + "loss": 1.61638603, + "memory(GiB)": 111.15, + "step": 14145, + "train_speed(iter/s)": 0.449003 + }, + { + "acc": 0.64791088, + "epoch": 0.3589548452562151, + "grad_norm": 5.1875, + "learning_rate": 9.54862715670544e-06, + "loss": 1.63125668, + "memory(GiB)": 111.15, + "step": 14150, + "train_speed(iter/s)": 0.449088 + }, + { + "acc": 0.64804392, + "epoch": 0.35908168442415017, + "grad_norm": 4.75, + "learning_rate": 9.548191657263132e-06, + "loss": 1.64996834, + "memory(GiB)": 111.15, + "step": 14155, + "train_speed(iter/s)": 0.449173 + }, + { + "acc": 0.6534133, + "epoch": 0.3592085235920852, + "grad_norm": 5.0625, + "learning_rate": 9.547755957771049e-06, + "loss": 1.63778038, + "memory(GiB)": 111.15, + "step": 14160, + "train_speed(iter/s)": 0.449258 + }, + { + "acc": 0.65406156, + "epoch": 0.3593353627600203, + "grad_norm": 4.8125, + "learning_rate": 9.547320058248356e-06, + "loss": 1.64388351, + "memory(GiB)": 111.15, + "step": 14165, + "train_speed(iter/s)": 0.449344 + }, + { + "acc": 0.65060816, + "epoch": 0.35946220192795536, + "grad_norm": 6.03125, + "learning_rate": 9.546883958714223e-06, + "loss": 1.6840313, + "memory(GiB)": 111.15, + "step": 14170, + "train_speed(iter/s)": 0.449429 + }, + { + "acc": 0.61958141, + "epoch": 0.3595890410958904, + "grad_norm": 5.28125, + "learning_rate": 9.546447659187834e-06, + "loss": 1.7229599, + "memory(GiB)": 111.15, + "step": 14175, + "train_speed(iter/s)": 0.449514 + }, + { + "acc": 0.64923544, + "epoch": 0.35971588026382545, + "grad_norm": 4.875, + "learning_rate": 9.546011159688377e-06, + "loss": 1.61892262, + "memory(GiB)": 111.15, + "step": 14180, + "train_speed(iter/s)": 0.449598 + }, + { + "acc": 0.65543013, + "epoch": 0.35984271943176055, + "grad_norm": 5.1875, + "learning_rate": 9.545574460235055e-06, + "loss": 1.64788437, + "memory(GiB)": 111.15, + "step": 14185, + "train_speed(iter/s)": 0.449683 + }, + { + "acc": 0.64655871, + "epoch": 0.3599695585996956, + "grad_norm": 5.21875, + "learning_rate": 9.545137560847071e-06, + "loss": 1.58711977, + "memory(GiB)": 111.15, + "step": 14190, + "train_speed(iter/s)": 0.449767 + }, + { + "acc": 0.63546391, + "epoch": 0.36009639776763064, + "grad_norm": 5.3125, + "learning_rate": 9.544700461543647e-06, + "loss": 1.6802248, + "memory(GiB)": 111.15, + "step": 14195, + "train_speed(iter/s)": 0.449852 + }, + { + "acc": 0.66254797, + "epoch": 0.3602232369355657, + "grad_norm": 4.90625, + "learning_rate": 9.544263162344005e-06, + "loss": 1.61615257, + "memory(GiB)": 111.15, + "step": 14200, + "train_speed(iter/s)": 0.449937 + }, + { + "epoch": 0.3602232369355657, + "eval_acc": 0.639140953436909, + "eval_loss": 1.6318191289901733, + "eval_runtime": 113.0153, + "eval_samples_per_second": 56.364, + "eval_steps_per_second": 28.182, + "step": 14200 + }, + { + "acc": 0.63858023, + "epoch": 0.3603500761035008, + "grad_norm": 5.34375, + "learning_rate": 9.54382566326738e-06, + "loss": 1.68927975, + "memory(GiB)": 111.15, + "step": 14205, + "train_speed(iter/s)": 0.448352 + }, + { + "acc": 0.63646684, + "epoch": 0.3604769152714358, + "grad_norm": 4.96875, + "learning_rate": 9.543387964333018e-06, + "loss": 1.65857124, + "memory(GiB)": 111.15, + "step": 14210, + "train_speed(iter/s)": 0.448436 + }, + { + "acc": 0.6283659, + "epoch": 0.36060375443937087, + "grad_norm": 5.65625, + "learning_rate": 9.542950065560165e-06, + "loss": 1.65683708, + "memory(GiB)": 111.15, + "step": 14215, + "train_speed(iter/s)": 0.448521 + }, + { + "acc": 0.62544022, + "epoch": 0.3607305936073059, + "grad_norm": 4.90625, + "learning_rate": 9.542511966968087e-06, + "loss": 1.7676384, + "memory(GiB)": 111.15, + "step": 14220, + "train_speed(iter/s)": 0.448606 + }, + { + "acc": 0.64924555, + "epoch": 0.360857432775241, + "grad_norm": 5.5, + "learning_rate": 9.542073668576052e-06, + "loss": 1.58760176, + "memory(GiB)": 111.15, + "step": 14225, + "train_speed(iter/s)": 0.448689 + }, + { + "acc": 0.6560461, + "epoch": 0.36098427194317606, + "grad_norm": 4.5625, + "learning_rate": 9.541635170403338e-06, + "loss": 1.61116943, + "memory(GiB)": 111.15, + "step": 14230, + "train_speed(iter/s)": 0.448774 + }, + { + "acc": 0.64099092, + "epoch": 0.3611111111111111, + "grad_norm": 5.0625, + "learning_rate": 9.541196472469234e-06, + "loss": 1.64024811, + "memory(GiB)": 111.15, + "step": 14235, + "train_speed(iter/s)": 0.448859 + }, + { + "acc": 0.64503651, + "epoch": 0.36123795027904615, + "grad_norm": 6.3125, + "learning_rate": 9.540757574793032e-06, + "loss": 1.65698242, + "memory(GiB)": 111.15, + "step": 14240, + "train_speed(iter/s)": 0.448943 + }, + { + "acc": 0.63416624, + "epoch": 0.36136478944698125, + "grad_norm": 5.21875, + "learning_rate": 9.540318477394039e-06, + "loss": 1.66830101, + "memory(GiB)": 111.15, + "step": 14245, + "train_speed(iter/s)": 0.449028 + }, + { + "acc": 0.66051869, + "epoch": 0.3614916286149163, + "grad_norm": 4.625, + "learning_rate": 9.539879180291568e-06, + "loss": 1.60914898, + "memory(GiB)": 111.15, + "step": 14250, + "train_speed(iter/s)": 0.449112 + }, + { + "acc": 0.64642611, + "epoch": 0.36161846778285134, + "grad_norm": 4.78125, + "learning_rate": 9.539439683504943e-06, + "loss": 1.64850788, + "memory(GiB)": 111.15, + "step": 14255, + "train_speed(iter/s)": 0.449197 + }, + { + "acc": 0.63902354, + "epoch": 0.3617453069507864, + "grad_norm": 5.3125, + "learning_rate": 9.538999987053492e-06, + "loss": 1.73722286, + "memory(GiB)": 111.15, + "step": 14260, + "train_speed(iter/s)": 0.449282 + }, + { + "acc": 0.63937635, + "epoch": 0.3618721461187215, + "grad_norm": 4.9375, + "learning_rate": 9.538560090956557e-06, + "loss": 1.67937088, + "memory(GiB)": 111.15, + "step": 14265, + "train_speed(iter/s)": 0.449366 + }, + { + "acc": 0.65090427, + "epoch": 0.3619989852866565, + "grad_norm": 5.34375, + "learning_rate": 9.538119995233485e-06, + "loss": 1.61618996, + "memory(GiB)": 111.15, + "step": 14270, + "train_speed(iter/s)": 0.44945 + }, + { + "acc": 0.65808105, + "epoch": 0.36212582445459157, + "grad_norm": 4.53125, + "learning_rate": 9.537679699903637e-06, + "loss": 1.65665131, + "memory(GiB)": 111.15, + "step": 14275, + "train_speed(iter/s)": 0.449534 + }, + { + "acc": 0.66959643, + "epoch": 0.3622526636225266, + "grad_norm": 6.34375, + "learning_rate": 9.537239204986375e-06, + "loss": 1.50684471, + "memory(GiB)": 111.15, + "step": 14280, + "train_speed(iter/s)": 0.449619 + }, + { + "acc": 0.6284399, + "epoch": 0.3623795027904617, + "grad_norm": 4.875, + "learning_rate": 9.536798510501075e-06, + "loss": 1.66848011, + "memory(GiB)": 111.15, + "step": 14285, + "train_speed(iter/s)": 0.449704 + }, + { + "acc": 0.64889402, + "epoch": 0.36250634195839676, + "grad_norm": 6.625, + "learning_rate": 9.536357616467123e-06, + "loss": 1.63340073, + "memory(GiB)": 111.15, + "step": 14290, + "train_speed(iter/s)": 0.449789 + }, + { + "acc": 0.65227442, + "epoch": 0.3626331811263318, + "grad_norm": 4.78125, + "learning_rate": 9.535916522903908e-06, + "loss": 1.62053204, + "memory(GiB)": 111.15, + "step": 14295, + "train_speed(iter/s)": 0.449872 + }, + { + "acc": 0.63406906, + "epoch": 0.36276002029426685, + "grad_norm": 6.09375, + "learning_rate": 9.535475229830832e-06, + "loss": 1.71453323, + "memory(GiB)": 111.15, + "step": 14300, + "train_speed(iter/s)": 0.449956 + }, + { + "epoch": 0.36276002029426685, + "eval_acc": 0.6393302020280932, + "eval_loss": 1.6315418481826782, + "eval_runtime": 113.6861, + "eval_samples_per_second": 56.031, + "eval_steps_per_second": 28.016, + "step": 14300 + }, + { + "acc": 0.64229002, + "epoch": 0.36288685946220195, + "grad_norm": 5.0, + "learning_rate": 9.535033737267308e-06, + "loss": 1.60360203, + "memory(GiB)": 111.15, + "step": 14305, + "train_speed(iter/s)": 0.448373 + }, + { + "acc": 0.6455997, + "epoch": 0.363013698630137, + "grad_norm": 5.8125, + "learning_rate": 9.534592045232752e-06, + "loss": 1.58670921, + "memory(GiB)": 111.15, + "step": 14310, + "train_speed(iter/s)": 0.448457 + }, + { + "acc": 0.63988686, + "epoch": 0.36314053779807204, + "grad_norm": 5.46875, + "learning_rate": 9.534150153746591e-06, + "loss": 1.72967033, + "memory(GiB)": 111.15, + "step": 14315, + "train_speed(iter/s)": 0.448541 + }, + { + "acc": 0.64051681, + "epoch": 0.3632673769660071, + "grad_norm": 5.65625, + "learning_rate": 9.533708062828264e-06, + "loss": 1.66630898, + "memory(GiB)": 111.15, + "step": 14320, + "train_speed(iter/s)": 0.448626 + }, + { + "acc": 0.64569559, + "epoch": 0.3633942161339422, + "grad_norm": 4.96875, + "learning_rate": 9.533265772497216e-06, + "loss": 1.70899239, + "memory(GiB)": 111.15, + "step": 14325, + "train_speed(iter/s)": 0.44871 + }, + { + "acc": 0.65536766, + "epoch": 0.3635210553018772, + "grad_norm": 5.5625, + "learning_rate": 9.532823282772899e-06, + "loss": 1.61132374, + "memory(GiB)": 111.15, + "step": 14330, + "train_speed(iter/s)": 0.448794 + }, + { + "acc": 0.64040947, + "epoch": 0.36364789446981227, + "grad_norm": 5.90625, + "learning_rate": 9.532380593674775e-06, + "loss": 1.65101051, + "memory(GiB)": 111.15, + "step": 14335, + "train_speed(iter/s)": 0.448879 + }, + { + "acc": 0.64864426, + "epoch": 0.3637747336377473, + "grad_norm": 6.46875, + "learning_rate": 9.531937705222319e-06, + "loss": 1.66865959, + "memory(GiB)": 111.15, + "step": 14340, + "train_speed(iter/s)": 0.448963 + }, + { + "acc": 0.6335063, + "epoch": 0.3639015728056824, + "grad_norm": 5.1875, + "learning_rate": 9.531494617435006e-06, + "loss": 1.65864449, + "memory(GiB)": 111.15, + "step": 14345, + "train_speed(iter/s)": 0.449047 + }, + { + "acc": 0.65094366, + "epoch": 0.36402841197361746, + "grad_norm": 6.71875, + "learning_rate": 9.531051330332331e-06, + "loss": 1.6091671, + "memory(GiB)": 111.15, + "step": 14350, + "train_speed(iter/s)": 0.449132 + }, + { + "acc": 0.6294487, + "epoch": 0.3641552511415525, + "grad_norm": 5.15625, + "learning_rate": 9.530607843933788e-06, + "loss": 1.68662071, + "memory(GiB)": 111.15, + "step": 14355, + "train_speed(iter/s)": 0.449217 + }, + { + "acc": 0.64478893, + "epoch": 0.36428209030948755, + "grad_norm": 5.28125, + "learning_rate": 9.530164158258883e-06, + "loss": 1.61809845, + "memory(GiB)": 111.15, + "step": 14360, + "train_speed(iter/s)": 0.449301 + }, + { + "acc": 0.65816355, + "epoch": 0.36440892947742265, + "grad_norm": 5.15625, + "learning_rate": 9.529720273327135e-06, + "loss": 1.60836754, + "memory(GiB)": 111.15, + "step": 14365, + "train_speed(iter/s)": 0.449385 + }, + { + "acc": 0.6479497, + "epoch": 0.3645357686453577, + "grad_norm": 5.03125, + "learning_rate": 9.529276189158063e-06, + "loss": 1.68301163, + "memory(GiB)": 111.15, + "step": 14370, + "train_speed(iter/s)": 0.44947 + }, + { + "acc": 0.65491405, + "epoch": 0.36466260781329274, + "grad_norm": 4.875, + "learning_rate": 9.528831905771205e-06, + "loss": 1.64202576, + "memory(GiB)": 111.15, + "step": 14375, + "train_speed(iter/s)": 0.449554 + }, + { + "acc": 0.64020581, + "epoch": 0.3647894469812278, + "grad_norm": 5.46875, + "learning_rate": 9.528387423186098e-06, + "loss": 1.66654663, + "memory(GiB)": 111.15, + "step": 14380, + "train_speed(iter/s)": 0.449639 + }, + { + "acc": 0.64381261, + "epoch": 0.3649162861491629, + "grad_norm": 4.75, + "learning_rate": 9.527942741422297e-06, + "loss": 1.65318909, + "memory(GiB)": 111.15, + "step": 14385, + "train_speed(iter/s)": 0.449723 + }, + { + "acc": 0.63821669, + "epoch": 0.3650431253170979, + "grad_norm": 5.5625, + "learning_rate": 9.527497860499355e-06, + "loss": 1.72965393, + "memory(GiB)": 111.15, + "step": 14390, + "train_speed(iter/s)": 0.449807 + }, + { + "acc": 0.64864984, + "epoch": 0.36516996448503297, + "grad_norm": 5.71875, + "learning_rate": 9.527052780436845e-06, + "loss": 1.55539169, + "memory(GiB)": 111.15, + "step": 14395, + "train_speed(iter/s)": 0.449891 + }, + { + "acc": 0.64167666, + "epoch": 0.365296803652968, + "grad_norm": 5.09375, + "learning_rate": 9.52660750125434e-06, + "loss": 1.63269711, + "memory(GiB)": 111.15, + "step": 14400, + "train_speed(iter/s)": 0.449975 + }, + { + "epoch": 0.365296803652968, + "eval_acc": 0.6392704613028408, + "eval_loss": 1.6314353942871094, + "eval_runtime": 112.9253, + "eval_samples_per_second": 56.409, + "eval_steps_per_second": 28.204, + "step": 14400 + }, + { + "acc": 0.6418047, + "epoch": 0.3654236428209031, + "grad_norm": 4.71875, + "learning_rate": 9.52616202297143e-06, + "loss": 1.64433746, + "memory(GiB)": 111.15, + "step": 14405, + "train_speed(iter/s)": 0.448413 + }, + { + "acc": 0.64049568, + "epoch": 0.36555048198883816, + "grad_norm": 4.40625, + "learning_rate": 9.525716345607706e-06, + "loss": 1.62900429, + "memory(GiB)": 111.15, + "step": 14410, + "train_speed(iter/s)": 0.448497 + }, + { + "acc": 0.62558689, + "epoch": 0.3656773211567732, + "grad_norm": 4.75, + "learning_rate": 9.52527046918277e-06, + "loss": 1.72462845, + "memory(GiB)": 111.15, + "step": 14415, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.62971029, + "epoch": 0.36580416032470825, + "grad_norm": 4.6875, + "learning_rate": 9.524824393716235e-06, + "loss": 1.72703705, + "memory(GiB)": 111.15, + "step": 14420, + "train_speed(iter/s)": 0.448663 + }, + { + "acc": 0.63854952, + "epoch": 0.36593099949264335, + "grad_norm": 5.65625, + "learning_rate": 9.524378119227722e-06, + "loss": 1.65682793, + "memory(GiB)": 111.15, + "step": 14425, + "train_speed(iter/s)": 0.448747 + }, + { + "acc": 0.6552175, + "epoch": 0.3660578386605784, + "grad_norm": 5.5, + "learning_rate": 9.523931645736858e-06, + "loss": 1.68306732, + "memory(GiB)": 111.15, + "step": 14430, + "train_speed(iter/s)": 0.448831 + }, + { + "acc": 0.64867997, + "epoch": 0.36618467782851344, + "grad_norm": 5.21875, + "learning_rate": 9.523484973263283e-06, + "loss": 1.68383541, + "memory(GiB)": 111.15, + "step": 14435, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.65961008, + "epoch": 0.3663115169964485, + "grad_norm": 6.0625, + "learning_rate": 9.523038101826644e-06, + "loss": 1.60269737, + "memory(GiB)": 111.15, + "step": 14440, + "train_speed(iter/s)": 0.448998 + }, + { + "acc": 0.6243906, + "epoch": 0.3664383561643836, + "grad_norm": 5.5, + "learning_rate": 9.522591031446596e-06, + "loss": 1.73108921, + "memory(GiB)": 111.15, + "step": 14445, + "train_speed(iter/s)": 0.449082 + }, + { + "acc": 0.64948978, + "epoch": 0.3665651953323186, + "grad_norm": 5.75, + "learning_rate": 9.522143762142801e-06, + "loss": 1.6035965, + "memory(GiB)": 111.15, + "step": 14450, + "train_speed(iter/s)": 0.449165 + }, + { + "acc": 0.66550035, + "epoch": 0.36669203450025367, + "grad_norm": 5.78125, + "learning_rate": 9.521696293934934e-06, + "loss": 1.62463913, + "memory(GiB)": 111.15, + "step": 14455, + "train_speed(iter/s)": 0.449249 + }, + { + "acc": 0.64511032, + "epoch": 0.3668188736681887, + "grad_norm": 4.8125, + "learning_rate": 9.521248626842676e-06, + "loss": 1.68508854, + "memory(GiB)": 111.15, + "step": 14460, + "train_speed(iter/s)": 0.449333 + }, + { + "acc": 0.63800745, + "epoch": 0.3669457128361238, + "grad_norm": 4.84375, + "learning_rate": 9.520800760885716e-06, + "loss": 1.70908775, + "memory(GiB)": 111.15, + "step": 14465, + "train_speed(iter/s)": 0.449416 + }, + { + "acc": 0.64224854, + "epoch": 0.36707255200405886, + "grad_norm": 5.28125, + "learning_rate": 9.520352696083756e-06, + "loss": 1.586619, + "memory(GiB)": 111.15, + "step": 14470, + "train_speed(iter/s)": 0.4495 + }, + { + "acc": 0.65365391, + "epoch": 0.3671993911719939, + "grad_norm": 5.71875, + "learning_rate": 9.519904432456504e-06, + "loss": 1.56092949, + "memory(GiB)": 111.15, + "step": 14475, + "train_speed(iter/s)": 0.449584 + }, + { + "acc": 0.66726441, + "epoch": 0.36732623033992895, + "grad_norm": 5.5, + "learning_rate": 9.519455970023672e-06, + "loss": 1.54890289, + "memory(GiB)": 111.15, + "step": 14480, + "train_speed(iter/s)": 0.449668 + }, + { + "acc": 0.64416842, + "epoch": 0.36745306950786405, + "grad_norm": 5.59375, + "learning_rate": 9.519007308804991e-06, + "loss": 1.63817139, + "memory(GiB)": 111.15, + "step": 14485, + "train_speed(iter/s)": 0.449752 + }, + { + "acc": 0.64854178, + "epoch": 0.3675799086757991, + "grad_norm": 4.5, + "learning_rate": 9.518558448820193e-06, + "loss": 1.59827461, + "memory(GiB)": 111.15, + "step": 14490, + "train_speed(iter/s)": 0.449836 + }, + { + "acc": 0.64438305, + "epoch": 0.36770674784373414, + "grad_norm": 6.03125, + "learning_rate": 9.518109390089017e-06, + "loss": 1.66951408, + "memory(GiB)": 111.15, + "step": 14495, + "train_speed(iter/s)": 0.44992 + }, + { + "acc": 0.6367979, + "epoch": 0.3678335870116692, + "grad_norm": 6.84375, + "learning_rate": 9.517660132631222e-06, + "loss": 1.69788647, + "memory(GiB)": 111.15, + "step": 14500, + "train_speed(iter/s)": 0.450004 + }, + { + "epoch": 0.3678335870116692, + "eval_acc": 0.6392583460508665, + "eval_loss": 1.630948543548584, + "eval_runtime": 112.2222, + "eval_samples_per_second": 56.762, + "eval_steps_per_second": 28.381, + "step": 14500 + }, + { + "acc": 0.6574049, + "epoch": 0.3679604261796043, + "grad_norm": 5.28125, + "learning_rate": 9.517210676466561e-06, + "loss": 1.65641937, + "memory(GiB)": 111.15, + "step": 14505, + "train_speed(iter/s)": 0.448463 + }, + { + "acc": 0.650071, + "epoch": 0.3680872653475393, + "grad_norm": 4.25, + "learning_rate": 9.516761021614809e-06, + "loss": 1.60505524, + "memory(GiB)": 111.15, + "step": 14510, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.64554081, + "epoch": 0.36821410451547437, + "grad_norm": 4.6875, + "learning_rate": 9.51631116809574e-06, + "loss": 1.63457813, + "memory(GiB)": 111.15, + "step": 14515, + "train_speed(iter/s)": 0.448628 + }, + { + "acc": 0.64037876, + "epoch": 0.3683409436834094, + "grad_norm": 5.25, + "learning_rate": 9.515861115929144e-06, + "loss": 1.70555191, + "memory(GiB)": 111.15, + "step": 14520, + "train_speed(iter/s)": 0.448711 + }, + { + "acc": 0.64018316, + "epoch": 0.3684677828513445, + "grad_norm": 7.34375, + "learning_rate": 9.515410865134812e-06, + "loss": 1.63516159, + "memory(GiB)": 111.15, + "step": 14525, + "train_speed(iter/s)": 0.448794 + }, + { + "acc": 0.65630131, + "epoch": 0.36859462201927956, + "grad_norm": 5.46875, + "learning_rate": 9.514960415732551e-06, + "loss": 1.58691101, + "memory(GiB)": 111.15, + "step": 14530, + "train_speed(iter/s)": 0.448876 + }, + { + "acc": 0.64040222, + "epoch": 0.3687214611872146, + "grad_norm": 6.0, + "learning_rate": 9.514509767742172e-06, + "loss": 1.69796944, + "memory(GiB)": 111.15, + "step": 14535, + "train_speed(iter/s)": 0.448959 + }, + { + "acc": 0.64557772, + "epoch": 0.36884830035514965, + "grad_norm": 6.75, + "learning_rate": 9.5140589211835e-06, + "loss": 1.64887981, + "memory(GiB)": 111.15, + "step": 14540, + "train_speed(iter/s)": 0.449042 + }, + { + "acc": 0.65082846, + "epoch": 0.36897513952308475, + "grad_norm": 5.5, + "learning_rate": 9.513607876076363e-06, + "loss": 1.69332352, + "memory(GiB)": 111.15, + "step": 14545, + "train_speed(iter/s)": 0.449125 + }, + { + "acc": 0.65439215, + "epoch": 0.3691019786910198, + "grad_norm": 6.59375, + "learning_rate": 9.513156632440598e-06, + "loss": 1.60593147, + "memory(GiB)": 111.15, + "step": 14550, + "train_speed(iter/s)": 0.449208 + }, + { + "acc": 0.64734979, + "epoch": 0.36922881785895484, + "grad_norm": 5.03125, + "learning_rate": 9.512705190296055e-06, + "loss": 1.66978092, + "memory(GiB)": 111.15, + "step": 14555, + "train_speed(iter/s)": 0.44929 + }, + { + "acc": 0.65431437, + "epoch": 0.3693556570268899, + "grad_norm": 4.96875, + "learning_rate": 9.512253549662588e-06, + "loss": 1.59671278, + "memory(GiB)": 111.15, + "step": 14560, + "train_speed(iter/s)": 0.449372 + }, + { + "acc": 0.6595489, + "epoch": 0.369482496194825, + "grad_norm": 5.53125, + "learning_rate": 9.511801710560066e-06, + "loss": 1.70670662, + "memory(GiB)": 111.15, + "step": 14565, + "train_speed(iter/s)": 0.449454 + }, + { + "acc": 0.65455818, + "epoch": 0.36960933536276, + "grad_norm": 5.25, + "learning_rate": 9.511349673008364e-06, + "loss": 1.60160637, + "memory(GiB)": 111.15, + "step": 14570, + "train_speed(iter/s)": 0.449536 + }, + { + "acc": 0.64083128, + "epoch": 0.36973617453069507, + "grad_norm": 5.53125, + "learning_rate": 9.510897437027358e-06, + "loss": 1.66876411, + "memory(GiB)": 111.15, + "step": 14575, + "train_speed(iter/s)": 0.449618 + }, + { + "acc": 0.6513443, + "epoch": 0.3698630136986301, + "grad_norm": 5.625, + "learning_rate": 9.510445002636943e-06, + "loss": 1.65052776, + "memory(GiB)": 111.15, + "step": 14580, + "train_speed(iter/s)": 0.449701 + }, + { + "acc": 0.66851597, + "epoch": 0.3699898528665652, + "grad_norm": 5.59375, + "learning_rate": 9.50999236985702e-06, + "loss": 1.633741, + "memory(GiB)": 111.15, + "step": 14585, + "train_speed(iter/s)": 0.449783 + }, + { + "acc": 0.65297441, + "epoch": 0.37011669203450026, + "grad_norm": 5.25, + "learning_rate": 9.509539538707497e-06, + "loss": 1.71499596, + "memory(GiB)": 111.15, + "step": 14590, + "train_speed(iter/s)": 0.449865 + }, + { + "acc": 0.64281874, + "epoch": 0.3702435312024353, + "grad_norm": 6.40625, + "learning_rate": 9.50908650920829e-06, + "loss": 1.66126995, + "memory(GiB)": 111.15, + "step": 14595, + "train_speed(iter/s)": 0.449947 + }, + { + "acc": 0.64553657, + "epoch": 0.37037037037037035, + "grad_norm": 4.71875, + "learning_rate": 9.50863328137933e-06, + "loss": 1.62382393, + "memory(GiB)": 111.15, + "step": 14600, + "train_speed(iter/s)": 0.450028 + }, + { + "epoch": 0.37037037037037035, + "eval_acc": 0.6393928671245118, + "eval_loss": 1.6308417320251465, + "eval_runtime": 113.7183, + "eval_samples_per_second": 56.016, + "eval_steps_per_second": 28.008, + "step": 14600 + }, + { + "acc": 0.6516736, + "epoch": 0.37049720953830545, + "grad_norm": 5.125, + "learning_rate": 9.508179855240545e-06, + "loss": 1.63735027, + "memory(GiB)": 111.15, + "step": 14605, + "train_speed(iter/s)": 0.448476 + }, + { + "acc": 0.65449963, + "epoch": 0.3706240487062405, + "grad_norm": 5.21875, + "learning_rate": 9.507726230811884e-06, + "loss": 1.70705204, + "memory(GiB)": 111.15, + "step": 14610, + "train_speed(iter/s)": 0.448558 + }, + { + "acc": 0.64333277, + "epoch": 0.37075088787417554, + "grad_norm": 5.5625, + "learning_rate": 9.507272408113298e-06, + "loss": 1.62672024, + "memory(GiB)": 111.15, + "step": 14615, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.63390422, + "epoch": 0.3708777270421106, + "grad_norm": 4.5, + "learning_rate": 9.506818387164748e-06, + "loss": 1.68107338, + "memory(GiB)": 111.15, + "step": 14620, + "train_speed(iter/s)": 0.448722 + }, + { + "acc": 0.65500851, + "epoch": 0.3710045662100457, + "grad_norm": 5.0, + "learning_rate": 9.506364167986204e-06, + "loss": 1.65059471, + "memory(GiB)": 111.15, + "step": 14625, + "train_speed(iter/s)": 0.448803 + }, + { + "acc": 0.64964819, + "epoch": 0.3711314053779807, + "grad_norm": 6.0625, + "learning_rate": 9.505909750597644e-06, + "loss": 1.66220169, + "memory(GiB)": 111.15, + "step": 14630, + "train_speed(iter/s)": 0.448885 + }, + { + "acc": 0.6452774, + "epoch": 0.37125824454591577, + "grad_norm": 5.15625, + "learning_rate": 9.505455135019055e-06, + "loss": 1.6050169, + "memory(GiB)": 111.15, + "step": 14635, + "train_speed(iter/s)": 0.448967 + }, + { + "acc": 0.64545937, + "epoch": 0.3713850837138508, + "grad_norm": 5.9375, + "learning_rate": 9.505000321270435e-06, + "loss": 1.62233124, + "memory(GiB)": 111.15, + "step": 14640, + "train_speed(iter/s)": 0.449049 + }, + { + "acc": 0.65504727, + "epoch": 0.3715119228817859, + "grad_norm": 6.0625, + "learning_rate": 9.504545309371786e-06, + "loss": 1.63773575, + "memory(GiB)": 111.15, + "step": 14645, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.62760553, + "epoch": 0.37163876204972096, + "grad_norm": 4.84375, + "learning_rate": 9.504090099343125e-06, + "loss": 1.6988781, + "memory(GiB)": 111.15, + "step": 14650, + "train_speed(iter/s)": 0.449214 + }, + { + "acc": 0.67126055, + "epoch": 0.371765601217656, + "grad_norm": 5.84375, + "learning_rate": 9.50363469120447e-06, + "loss": 1.55456266, + "memory(GiB)": 111.15, + "step": 14655, + "train_speed(iter/s)": 0.449296 + }, + { + "acc": 0.6206193, + "epoch": 0.37189244038559105, + "grad_norm": 5.3125, + "learning_rate": 9.503179084975855e-06, + "loss": 1.72508469, + "memory(GiB)": 111.15, + "step": 14660, + "train_speed(iter/s)": 0.449378 + }, + { + "acc": 0.63771548, + "epoch": 0.37201927955352615, + "grad_norm": 5.125, + "learning_rate": 9.502723280677319e-06, + "loss": 1.67332745, + "memory(GiB)": 111.15, + "step": 14665, + "train_speed(iter/s)": 0.449459 + }, + { + "acc": 0.65977945, + "epoch": 0.3721461187214612, + "grad_norm": 5.65625, + "learning_rate": 9.50226727832891e-06, + "loss": 1.6745182, + "memory(GiB)": 111.15, + "step": 14670, + "train_speed(iter/s)": 0.449541 + }, + { + "acc": 0.63518057, + "epoch": 0.37227295788939624, + "grad_norm": 5.9375, + "learning_rate": 9.501811077950685e-06, + "loss": 1.72175121, + "memory(GiB)": 111.15, + "step": 14675, + "train_speed(iter/s)": 0.449624 + }, + { + "acc": 0.64569473, + "epoch": 0.3723997970573313, + "grad_norm": 5.75, + "learning_rate": 9.501354679562708e-06, + "loss": 1.63712082, + "memory(GiB)": 111.15, + "step": 14680, + "train_speed(iter/s)": 0.449706 + }, + { + "acc": 0.65331955, + "epoch": 0.3725266362252664, + "grad_norm": 5.59375, + "learning_rate": 9.500898083185058e-06, + "loss": 1.65383987, + "memory(GiB)": 111.15, + "step": 14685, + "train_speed(iter/s)": 0.449788 + }, + { + "acc": 0.64626818, + "epoch": 0.3726534753932014, + "grad_norm": 5.03125, + "learning_rate": 9.500441288837812e-06, + "loss": 1.60931568, + "memory(GiB)": 111.15, + "step": 14690, + "train_speed(iter/s)": 0.44987 + }, + { + "acc": 0.64005337, + "epoch": 0.37278031456113647, + "grad_norm": 4.9375, + "learning_rate": 9.499984296541066e-06, + "loss": 1.6972126, + "memory(GiB)": 111.15, + "step": 14695, + "train_speed(iter/s)": 0.449951 + }, + { + "acc": 0.62346907, + "epoch": 0.3729071537290715, + "grad_norm": 5.625, + "learning_rate": 9.49952710631492e-06, + "loss": 1.77166595, + "memory(GiB)": 111.15, + "step": 14700, + "train_speed(iter/s)": 0.450033 + }, + { + "epoch": 0.3729071537290715, + "eval_acc": 0.6393882716841077, + "eval_loss": 1.6302801370620728, + "eval_runtime": 113.7874, + "eval_samples_per_second": 55.982, + "eval_steps_per_second": 27.991, + "step": 14700 + }, + { + "acc": 0.64275198, + "epoch": 0.3730339928970066, + "grad_norm": 5.15625, + "learning_rate": 9.499069718179484e-06, + "loss": 1.613451, + "memory(GiB)": 111.15, + "step": 14705, + "train_speed(iter/s)": 0.44849 + }, + { + "acc": 0.65590305, + "epoch": 0.37316083206494166, + "grad_norm": 5.53125, + "learning_rate": 9.498612132154874e-06, + "loss": 1.58161278, + "memory(GiB)": 111.15, + "step": 14710, + "train_speed(iter/s)": 0.448572 + }, + { + "acc": 0.64549217, + "epoch": 0.3732876712328767, + "grad_norm": 4.03125, + "learning_rate": 9.498154348261217e-06, + "loss": 1.62372856, + "memory(GiB)": 111.15, + "step": 14715, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.65355778, + "epoch": 0.37341451040081175, + "grad_norm": 5.46875, + "learning_rate": 9.497696366518649e-06, + "loss": 1.65267181, + "memory(GiB)": 111.15, + "step": 14720, + "train_speed(iter/s)": 0.448735 + }, + { + "acc": 0.64496713, + "epoch": 0.37354134956874685, + "grad_norm": 4.65625, + "learning_rate": 9.497238186947315e-06, + "loss": 1.66919975, + "memory(GiB)": 111.15, + "step": 14725, + "train_speed(iter/s)": 0.448816 + }, + { + "acc": 0.63596821, + "epoch": 0.3736681887366819, + "grad_norm": 5.15625, + "learning_rate": 9.496779809567367e-06, + "loss": 1.68401794, + "memory(GiB)": 111.15, + "step": 14730, + "train_speed(iter/s)": 0.448898 + }, + { + "acc": 0.64546413, + "epoch": 0.37379502790461694, + "grad_norm": 4.5, + "learning_rate": 9.496321234398967e-06, + "loss": 1.60097618, + "memory(GiB)": 111.15, + "step": 14735, + "train_speed(iter/s)": 0.44898 + }, + { + "acc": 0.64233036, + "epoch": 0.373921867072552, + "grad_norm": 4.3125, + "learning_rate": 9.495862461462282e-06, + "loss": 1.64110184, + "memory(GiB)": 111.15, + "step": 14740, + "train_speed(iter/s)": 0.449061 + }, + { + "acc": 0.64893937, + "epoch": 0.3740487062404871, + "grad_norm": 4.75, + "learning_rate": 9.495403490777495e-06, + "loss": 1.66180439, + "memory(GiB)": 111.15, + "step": 14745, + "train_speed(iter/s)": 0.449142 + }, + { + "acc": 0.66583433, + "epoch": 0.3741755454084221, + "grad_norm": 5.25, + "learning_rate": 9.49494432236479e-06, + "loss": 1.58811522, + "memory(GiB)": 111.15, + "step": 14750, + "train_speed(iter/s)": 0.449224 + }, + { + "acc": 0.64432797, + "epoch": 0.37430238457635717, + "grad_norm": 5.125, + "learning_rate": 9.494484956244368e-06, + "loss": 1.65499039, + "memory(GiB)": 111.15, + "step": 14755, + "train_speed(iter/s)": 0.449306 + }, + { + "acc": 0.64863954, + "epoch": 0.3744292237442922, + "grad_norm": 6.5625, + "learning_rate": 9.49402539243643e-06, + "loss": 1.69186001, + "memory(GiB)": 111.15, + "step": 14760, + "train_speed(iter/s)": 0.449388 + }, + { + "acc": 0.64270353, + "epoch": 0.3745560629122273, + "grad_norm": 5.0, + "learning_rate": 9.49356563096119e-06, + "loss": 1.65822678, + "memory(GiB)": 111.15, + "step": 14765, + "train_speed(iter/s)": 0.449469 + }, + { + "acc": 0.6442338, + "epoch": 0.37468290208016236, + "grad_norm": 4.4375, + "learning_rate": 9.49310567183887e-06, + "loss": 1.61382446, + "memory(GiB)": 111.15, + "step": 14770, + "train_speed(iter/s)": 0.449551 + }, + { + "acc": 0.64676294, + "epoch": 0.3748097412480974, + "grad_norm": 5.84375, + "learning_rate": 9.492645515089706e-06, + "loss": 1.63478985, + "memory(GiB)": 111.15, + "step": 14775, + "train_speed(iter/s)": 0.449634 + }, + { + "acc": 0.64007363, + "epoch": 0.37493658041603245, + "grad_norm": 5.90625, + "learning_rate": 9.492185160733934e-06, + "loss": 1.64263306, + "memory(GiB)": 111.15, + "step": 14780, + "train_speed(iter/s)": 0.449715 + }, + { + "acc": 0.64391584, + "epoch": 0.37506341958396755, + "grad_norm": 6.125, + "learning_rate": 9.491724608791798e-06, + "loss": 1.67460728, + "memory(GiB)": 111.15, + "step": 14785, + "train_speed(iter/s)": 0.449797 + }, + { + "acc": 0.63908739, + "epoch": 0.3751902587519026, + "grad_norm": 5.09375, + "learning_rate": 9.491263859283563e-06, + "loss": 1.63415985, + "memory(GiB)": 111.15, + "step": 14790, + "train_speed(iter/s)": 0.449878 + }, + { + "acc": 0.65625801, + "epoch": 0.37531709791983764, + "grad_norm": 4.25, + "learning_rate": 9.490802912229491e-06, + "loss": 1.64570007, + "memory(GiB)": 111.15, + "step": 14795, + "train_speed(iter/s)": 0.44996 + }, + { + "acc": 0.63277636, + "epoch": 0.3754439370877727, + "grad_norm": 4.5, + "learning_rate": 9.490341767649858e-06, + "loss": 1.73162193, + "memory(GiB)": 111.15, + "step": 14800, + "train_speed(iter/s)": 0.450041 + }, + { + "epoch": 0.3754439370877727, + "eval_acc": 0.6394375682266237, + "eval_loss": 1.6301159858703613, + "eval_runtime": 113.3927, + "eval_samples_per_second": 56.176, + "eval_steps_per_second": 28.088, + "step": 14800 + }, + { + "acc": 0.65420423, + "epoch": 0.3755707762557078, + "grad_norm": 5.71875, + "learning_rate": 9.489880425564944e-06, + "loss": 1.59041624, + "memory(GiB)": 111.15, + "step": 14805, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.63566408, + "epoch": 0.37569761542364283, + "grad_norm": 4.46875, + "learning_rate": 9.489418885995043e-06, + "loss": 1.60790405, + "memory(GiB)": 111.15, + "step": 14810, + "train_speed(iter/s)": 0.448596 + }, + { + "acc": 0.62919316, + "epoch": 0.3758244545915779, + "grad_norm": 5.1875, + "learning_rate": 9.488957148960457e-06, + "loss": 1.72803001, + "memory(GiB)": 111.15, + "step": 14815, + "train_speed(iter/s)": 0.448677 + }, + { + "acc": 0.63234234, + "epoch": 0.3759512937595129, + "grad_norm": 5.0, + "learning_rate": 9.488495214481494e-06, + "loss": 1.64002342, + "memory(GiB)": 111.15, + "step": 14820, + "train_speed(iter/s)": 0.448758 + }, + { + "acc": 0.65110502, + "epoch": 0.376078132927448, + "grad_norm": 5.0, + "learning_rate": 9.48803308257847e-06, + "loss": 1.66407471, + "memory(GiB)": 111.15, + "step": 14825, + "train_speed(iter/s)": 0.448839 + }, + { + "acc": 0.6602035, + "epoch": 0.37620497209538306, + "grad_norm": 5.3125, + "learning_rate": 9.487570753271716e-06, + "loss": 1.62914314, + "memory(GiB)": 111.15, + "step": 14830, + "train_speed(iter/s)": 0.44892 + }, + { + "acc": 0.63964972, + "epoch": 0.3763318112633181, + "grad_norm": 5.9375, + "learning_rate": 9.487108226581564e-06, + "loss": 1.67306557, + "memory(GiB)": 111.15, + "step": 14835, + "train_speed(iter/s)": 0.449 + }, + { + "acc": 0.65277367, + "epoch": 0.37645865043125315, + "grad_norm": 4.625, + "learning_rate": 9.486645502528355e-06, + "loss": 1.63380127, + "memory(GiB)": 111.15, + "step": 14840, + "train_speed(iter/s)": 0.449079 + }, + { + "acc": 0.6532012, + "epoch": 0.37658548959918825, + "grad_norm": 4.625, + "learning_rate": 9.486182581132449e-06, + "loss": 1.58466768, + "memory(GiB)": 111.15, + "step": 14845, + "train_speed(iter/s)": 0.449159 + }, + { + "acc": 0.64996367, + "epoch": 0.3767123287671233, + "grad_norm": 5.65625, + "learning_rate": 9.485719462414202e-06, + "loss": 1.58853722, + "memory(GiB)": 111.15, + "step": 14850, + "train_speed(iter/s)": 0.449238 + }, + { + "acc": 0.66455765, + "epoch": 0.37683916793505834, + "grad_norm": 5.53125, + "learning_rate": 9.485256146393987e-06, + "loss": 1.65410004, + "memory(GiB)": 111.15, + "step": 14855, + "train_speed(iter/s)": 0.449319 + }, + { + "acc": 0.65352192, + "epoch": 0.3769660071029934, + "grad_norm": 5.8125, + "learning_rate": 9.484792633092182e-06, + "loss": 1.63899994, + "memory(GiB)": 111.15, + "step": 14860, + "train_speed(iter/s)": 0.4494 + }, + { + "acc": 0.64974709, + "epoch": 0.3770928462709285, + "grad_norm": 5.21875, + "learning_rate": 9.484328922529172e-06, + "loss": 1.5913332, + "memory(GiB)": 111.15, + "step": 14865, + "train_speed(iter/s)": 0.44948 + }, + { + "acc": 0.64389739, + "epoch": 0.37721968543886353, + "grad_norm": 5.0, + "learning_rate": 9.483865014725356e-06, + "loss": 1.59500332, + "memory(GiB)": 111.15, + "step": 14870, + "train_speed(iter/s)": 0.44956 + }, + { + "acc": 0.64750729, + "epoch": 0.3773465246067986, + "grad_norm": 5.875, + "learning_rate": 9.483400909701139e-06, + "loss": 1.67172165, + "memory(GiB)": 111.15, + "step": 14875, + "train_speed(iter/s)": 0.44964 + }, + { + "acc": 0.64214354, + "epoch": 0.3774733637747336, + "grad_norm": 4.9375, + "learning_rate": 9.482936607476931e-06, + "loss": 1.65352135, + "memory(GiB)": 111.15, + "step": 14880, + "train_speed(iter/s)": 0.44972 + }, + { + "acc": 0.64087601, + "epoch": 0.3776002029426687, + "grad_norm": 5.84375, + "learning_rate": 9.482472108073157e-06, + "loss": 1.70822086, + "memory(GiB)": 111.15, + "step": 14885, + "train_speed(iter/s)": 0.449801 + }, + { + "acc": 0.63795581, + "epoch": 0.37772704211060376, + "grad_norm": 5.5, + "learning_rate": 9.482007411510245e-06, + "loss": 1.63081284, + "memory(GiB)": 111.15, + "step": 14890, + "train_speed(iter/s)": 0.44988 + }, + { + "acc": 0.64351187, + "epoch": 0.3778538812785388, + "grad_norm": 5.21875, + "learning_rate": 9.48154251780864e-06, + "loss": 1.59781742, + "memory(GiB)": 111.15, + "step": 14895, + "train_speed(iter/s)": 0.449961 + }, + { + "acc": 0.65495691, + "epoch": 0.37798072044647385, + "grad_norm": 5.375, + "learning_rate": 9.481077426988782e-06, + "loss": 1.66050072, + "memory(GiB)": 111.15, + "step": 14900, + "train_speed(iter/s)": 0.45004 + }, + { + "epoch": 0.37798072044647385, + "eval_acc": 0.6395954842695987, + "eval_loss": 1.629775047302246, + "eval_runtime": 114.0557, + "eval_samples_per_second": 55.85, + "eval_steps_per_second": 27.925, + "step": 14900 + }, + { + "acc": 0.63240757, + "epoch": 0.37810755961440895, + "grad_norm": 5.71875, + "learning_rate": 9.480612139071134e-06, + "loss": 1.66841755, + "memory(GiB)": 111.15, + "step": 14905, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.63805861, + "epoch": 0.378234398782344, + "grad_norm": 5.75, + "learning_rate": 9.48014665407616e-06, + "loss": 1.67057076, + "memory(GiB)": 111.15, + "step": 14910, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.66378422, + "epoch": 0.37836123795027904, + "grad_norm": 5.375, + "learning_rate": 9.479680972024334e-06, + "loss": 1.60718403, + "memory(GiB)": 111.15, + "step": 14915, + "train_speed(iter/s)": 0.448671 + }, + { + "acc": 0.66384716, + "epoch": 0.3784880771182141, + "grad_norm": 6.15625, + "learning_rate": 9.47921509293614e-06, + "loss": 1.57608624, + "memory(GiB)": 111.15, + "step": 14920, + "train_speed(iter/s)": 0.448751 + }, + { + "acc": 0.65116968, + "epoch": 0.3786149162861492, + "grad_norm": 5.3125, + "learning_rate": 9.478749016832066e-06, + "loss": 1.68787537, + "memory(GiB)": 111.15, + "step": 14925, + "train_speed(iter/s)": 0.448829 + }, + { + "acc": 0.65183778, + "epoch": 0.37874175545408423, + "grad_norm": 4.5625, + "learning_rate": 9.478282743732613e-06, + "loss": 1.60239487, + "memory(GiB)": 111.15, + "step": 14930, + "train_speed(iter/s)": 0.448907 + }, + { + "acc": 0.64813204, + "epoch": 0.3788685946220193, + "grad_norm": 5.53125, + "learning_rate": 9.477816273658293e-06, + "loss": 1.64867096, + "memory(GiB)": 111.15, + "step": 14935, + "train_speed(iter/s)": 0.448987 + }, + { + "acc": 0.64645214, + "epoch": 0.3789954337899543, + "grad_norm": 5.21875, + "learning_rate": 9.47734960662962e-06, + "loss": 1.7227047, + "memory(GiB)": 111.15, + "step": 14940, + "train_speed(iter/s)": 0.449065 + }, + { + "acc": 0.64930363, + "epoch": 0.3791222729578894, + "grad_norm": 4.5625, + "learning_rate": 9.476882742667122e-06, + "loss": 1.61137047, + "memory(GiB)": 111.15, + "step": 14945, + "train_speed(iter/s)": 0.449145 + }, + { + "acc": 0.64239426, + "epoch": 0.37924911212582446, + "grad_norm": 5.1875, + "learning_rate": 9.476415681791333e-06, + "loss": 1.6158699, + "memory(GiB)": 111.15, + "step": 14950, + "train_speed(iter/s)": 0.449224 + }, + { + "acc": 0.64700909, + "epoch": 0.3793759512937595, + "grad_norm": 5.3125, + "learning_rate": 9.475948424022798e-06, + "loss": 1.63355522, + "memory(GiB)": 111.15, + "step": 14955, + "train_speed(iter/s)": 0.449304 + }, + { + "acc": 0.66545525, + "epoch": 0.37950279046169455, + "grad_norm": 5.0625, + "learning_rate": 9.475480969382065e-06, + "loss": 1.59146023, + "memory(GiB)": 111.15, + "step": 14960, + "train_speed(iter/s)": 0.449382 + }, + { + "acc": 0.64744444, + "epoch": 0.37962962962962965, + "grad_norm": 4.65625, + "learning_rate": 9.475013317889699e-06, + "loss": 1.62458935, + "memory(GiB)": 111.15, + "step": 14965, + "train_speed(iter/s)": 0.449462 + }, + { + "acc": 0.64174986, + "epoch": 0.3797564687975647, + "grad_norm": 5.46875, + "learning_rate": 9.474545469566267e-06, + "loss": 1.6196907, + "memory(GiB)": 111.15, + "step": 14970, + "train_speed(iter/s)": 0.449542 + }, + { + "acc": 0.63776031, + "epoch": 0.37988330796549974, + "grad_norm": 4.9375, + "learning_rate": 9.474077424432348e-06, + "loss": 1.6416954, + "memory(GiB)": 111.15, + "step": 14975, + "train_speed(iter/s)": 0.449621 + }, + { + "acc": 0.64091244, + "epoch": 0.3800101471334348, + "grad_norm": 5.71875, + "learning_rate": 9.47360918250853e-06, + "loss": 1.6896759, + "memory(GiB)": 111.15, + "step": 14980, + "train_speed(iter/s)": 0.449625 + }, + { + "acc": 0.64674959, + "epoch": 0.3801369863013699, + "grad_norm": 6.375, + "learning_rate": 9.473140743815405e-06, + "loss": 1.72840748, + "memory(GiB)": 111.15, + "step": 14985, + "train_speed(iter/s)": 0.449705 + }, + { + "acc": 0.66726432, + "epoch": 0.38026382546930493, + "grad_norm": 4.875, + "learning_rate": 9.47267210837358e-06, + "loss": 1.59296646, + "memory(GiB)": 111.15, + "step": 14990, + "train_speed(iter/s)": 0.449784 + }, + { + "acc": 0.65054026, + "epoch": 0.38039066463724, + "grad_norm": 5.875, + "learning_rate": 9.472203276203667e-06, + "loss": 1.65335712, + "memory(GiB)": 111.15, + "step": 14995, + "train_speed(iter/s)": 0.449863 + }, + { + "acc": 0.64996519, + "epoch": 0.380517503805175, + "grad_norm": 4.875, + "learning_rate": 9.471734247326284e-06, + "loss": 1.67458668, + "memory(GiB)": 111.15, + "step": 15000, + "train_speed(iter/s)": 0.449942 + }, + { + "epoch": 0.380517503805175, + "eval_acc": 0.6394764205864033, + "eval_loss": 1.629930019378662, + "eval_runtime": 114.205, + "eval_samples_per_second": 55.777, + "eval_steps_per_second": 27.888, + "step": 15000 + }, + { + "acc": 0.65467911, + "epoch": 0.3806443429731101, + "grad_norm": 5.75, + "learning_rate": 9.471265021762067e-06, + "loss": 1.65644875, + "memory(GiB)": 111.15, + "step": 15005, + "train_speed(iter/s)": 0.448423 + }, + { + "acc": 0.6757957, + "epoch": 0.38077118214104516, + "grad_norm": 5.625, + "learning_rate": 9.47079559953165e-06, + "loss": 1.51892376, + "memory(GiB)": 111.15, + "step": 15010, + "train_speed(iter/s)": 0.448502 + }, + { + "acc": 0.62558093, + "epoch": 0.3808980213089802, + "grad_norm": 5.8125, + "learning_rate": 9.470325980655683e-06, + "loss": 1.69560833, + "memory(GiB)": 111.15, + "step": 15015, + "train_speed(iter/s)": 0.448581 + }, + { + "acc": 0.6407486, + "epoch": 0.38102486047691525, + "grad_norm": 7.78125, + "learning_rate": 9.46985616515482e-06, + "loss": 1.69873848, + "memory(GiB)": 111.15, + "step": 15020, + "train_speed(iter/s)": 0.44866 + }, + { + "acc": 0.64284592, + "epoch": 0.38115169964485035, + "grad_norm": 5.78125, + "learning_rate": 9.469386153049727e-06, + "loss": 1.64029083, + "memory(GiB)": 111.15, + "step": 15025, + "train_speed(iter/s)": 0.448739 + }, + { + "acc": 0.64997482, + "epoch": 0.3812785388127854, + "grad_norm": 5.6875, + "learning_rate": 9.468915944361076e-06, + "loss": 1.57378559, + "memory(GiB)": 111.15, + "step": 15030, + "train_speed(iter/s)": 0.448818 + }, + { + "acc": 0.65559368, + "epoch": 0.38140537798072044, + "grad_norm": 4.375, + "learning_rate": 9.468445539109551e-06, + "loss": 1.65336628, + "memory(GiB)": 111.15, + "step": 15035, + "train_speed(iter/s)": 0.448897 + }, + { + "acc": 0.6344059, + "epoch": 0.3815322171486555, + "grad_norm": 5.5, + "learning_rate": 9.46797493731584e-06, + "loss": 1.72369843, + "memory(GiB)": 111.15, + "step": 15040, + "train_speed(iter/s)": 0.448976 + }, + { + "acc": 0.65586414, + "epoch": 0.3816590563165906, + "grad_norm": 4.6875, + "learning_rate": 9.467504139000642e-06, + "loss": 1.63554211, + "memory(GiB)": 111.15, + "step": 15045, + "train_speed(iter/s)": 0.449055 + }, + { + "acc": 0.64697371, + "epoch": 0.38178589548452563, + "grad_norm": 4.625, + "learning_rate": 9.467033144184667e-06, + "loss": 1.68755455, + "memory(GiB)": 111.15, + "step": 15050, + "train_speed(iter/s)": 0.449134 + }, + { + "acc": 0.65605583, + "epoch": 0.3819127346524607, + "grad_norm": 5.75, + "learning_rate": 9.466561952888632e-06, + "loss": 1.64963093, + "memory(GiB)": 111.15, + "step": 15055, + "train_speed(iter/s)": 0.449212 + }, + { + "acc": 0.64498358, + "epoch": 0.3820395738203957, + "grad_norm": 5.3125, + "learning_rate": 9.466090565133259e-06, + "loss": 1.64929657, + "memory(GiB)": 111.15, + "step": 15060, + "train_speed(iter/s)": 0.449291 + }, + { + "acc": 0.65483985, + "epoch": 0.3821664129883308, + "grad_norm": 4.59375, + "learning_rate": 9.465618980939284e-06, + "loss": 1.63532448, + "memory(GiB)": 111.15, + "step": 15065, + "train_speed(iter/s)": 0.44937 + }, + { + "acc": 0.65672026, + "epoch": 0.38229325215626586, + "grad_norm": 5.125, + "learning_rate": 9.465147200327446e-06, + "loss": 1.67566223, + "memory(GiB)": 111.15, + "step": 15070, + "train_speed(iter/s)": 0.449448 + }, + { + "acc": 0.64983163, + "epoch": 0.3824200913242009, + "grad_norm": 5.96875, + "learning_rate": 9.464675223318503e-06, + "loss": 1.59508696, + "memory(GiB)": 111.15, + "step": 15075, + "train_speed(iter/s)": 0.449527 + }, + { + "acc": 0.66478987, + "epoch": 0.38254693049213595, + "grad_norm": 5.84375, + "learning_rate": 9.464203049933207e-06, + "loss": 1.60530891, + "memory(GiB)": 111.15, + "step": 15080, + "train_speed(iter/s)": 0.449605 + }, + { + "acc": 0.67865434, + "epoch": 0.38267376966007105, + "grad_norm": 5.4375, + "learning_rate": 9.463730680192332e-06, + "loss": 1.57984028, + "memory(GiB)": 111.15, + "step": 15085, + "train_speed(iter/s)": 0.449684 + }, + { + "acc": 0.64460173, + "epoch": 0.3828006088280061, + "grad_norm": 5.1875, + "learning_rate": 9.46325811411665e-06, + "loss": 1.70267296, + "memory(GiB)": 111.15, + "step": 15090, + "train_speed(iter/s)": 0.449763 + }, + { + "acc": 0.65097761, + "epoch": 0.38292744799594114, + "grad_norm": 6.25, + "learning_rate": 9.462785351726951e-06, + "loss": 1.63867168, + "memory(GiB)": 111.15, + "step": 15095, + "train_speed(iter/s)": 0.449841 + }, + { + "acc": 0.65100784, + "epoch": 0.3830542871638762, + "grad_norm": 5.125, + "learning_rate": 9.462312393044027e-06, + "loss": 1.63593254, + "memory(GiB)": 111.15, + "step": 15100, + "train_speed(iter/s)": 0.449919 + }, + { + "epoch": 0.3830542871638762, + "eval_acc": 0.639693659587321, + "eval_loss": 1.6291992664337158, + "eval_runtime": 113.0649, + "eval_samples_per_second": 56.339, + "eval_steps_per_second": 28.17, + "step": 15100 + }, + { + "acc": 0.66500111, + "epoch": 0.3831811263318113, + "grad_norm": 5.25, + "learning_rate": 9.46183923808868e-06, + "loss": 1.61016483, + "memory(GiB)": 111.15, + "step": 15105, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.65316582, + "epoch": 0.38330796549974633, + "grad_norm": 5.15625, + "learning_rate": 9.461365886881724e-06, + "loss": 1.66185608, + "memory(GiB)": 111.15, + "step": 15110, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.65410786, + "epoch": 0.3834348046676814, + "grad_norm": 5.09375, + "learning_rate": 9.460892339443977e-06, + "loss": 1.69379997, + "memory(GiB)": 111.15, + "step": 15115, + "train_speed(iter/s)": 0.448506 + }, + { + "acc": 0.64868379, + "epoch": 0.3835616438356164, + "grad_norm": 6.34375, + "learning_rate": 9.460418595796268e-06, + "loss": 1.60279846, + "memory(GiB)": 111.15, + "step": 15120, + "train_speed(iter/s)": 0.448585 + }, + { + "acc": 0.6314846, + "epoch": 0.3836884830035515, + "grad_norm": 4.59375, + "learning_rate": 9.459944655959437e-06, + "loss": 1.68295822, + "memory(GiB)": 111.15, + "step": 15125, + "train_speed(iter/s)": 0.448663 + }, + { + "acc": 0.65969458, + "epoch": 0.38381532217148656, + "grad_norm": 5.1875, + "learning_rate": 9.459470519954325e-06, + "loss": 1.57525578, + "memory(GiB)": 111.15, + "step": 15130, + "train_speed(iter/s)": 0.448742 + }, + { + "acc": 0.6498271, + "epoch": 0.3839421613394216, + "grad_norm": 5.9375, + "learning_rate": 9.458996187801791e-06, + "loss": 1.62640457, + "memory(GiB)": 111.15, + "step": 15135, + "train_speed(iter/s)": 0.448821 + }, + { + "acc": 0.63278399, + "epoch": 0.38406900050735665, + "grad_norm": 5.625, + "learning_rate": 9.458521659522697e-06, + "loss": 1.70377178, + "memory(GiB)": 111.15, + "step": 15140, + "train_speed(iter/s)": 0.448899 + }, + { + "acc": 0.63956375, + "epoch": 0.38419583967529175, + "grad_norm": 5.21875, + "learning_rate": 9.458046935137913e-06, + "loss": 1.71881905, + "memory(GiB)": 111.15, + "step": 15145, + "train_speed(iter/s)": 0.448978 + }, + { + "acc": 0.65624151, + "epoch": 0.3843226788432268, + "grad_norm": 5.25, + "learning_rate": 9.457572014668323e-06, + "loss": 1.64339256, + "memory(GiB)": 111.15, + "step": 15150, + "train_speed(iter/s)": 0.449057 + }, + { + "acc": 0.6456933, + "epoch": 0.38444951801116184, + "grad_norm": 4.8125, + "learning_rate": 9.457096898134813e-06, + "loss": 1.62951183, + "memory(GiB)": 111.15, + "step": 15155, + "train_speed(iter/s)": 0.449136 + }, + { + "acc": 0.65966244, + "epoch": 0.3845763571790969, + "grad_norm": 6.90625, + "learning_rate": 9.45662158555828e-06, + "loss": 1.65218697, + "memory(GiB)": 111.15, + "step": 15160, + "train_speed(iter/s)": 0.449215 + }, + { + "acc": 0.6442337, + "epoch": 0.384703196347032, + "grad_norm": 7.4375, + "learning_rate": 9.456146076959636e-06, + "loss": 1.65164452, + "memory(GiB)": 111.15, + "step": 15165, + "train_speed(iter/s)": 0.449294 + }, + { + "acc": 0.65274773, + "epoch": 0.38483003551496703, + "grad_norm": 5.09375, + "learning_rate": 9.455670372359791e-06, + "loss": 1.68195763, + "memory(GiB)": 111.15, + "step": 15170, + "train_speed(iter/s)": 0.449373 + }, + { + "acc": 0.63407898, + "epoch": 0.3849568746829021, + "grad_norm": 4.90625, + "learning_rate": 9.45519447177967e-06, + "loss": 1.67016945, + "memory(GiB)": 111.15, + "step": 15175, + "train_speed(iter/s)": 0.449451 + }, + { + "acc": 0.6460537, + "epoch": 0.3850837138508371, + "grad_norm": 5.375, + "learning_rate": 9.454718375240204e-06, + "loss": 1.70183525, + "memory(GiB)": 111.15, + "step": 15180, + "train_speed(iter/s)": 0.44953 + }, + { + "acc": 0.65817518, + "epoch": 0.3852105530187722, + "grad_norm": 6.53125, + "learning_rate": 9.454242082762336e-06, + "loss": 1.60628033, + "memory(GiB)": 111.15, + "step": 15185, + "train_speed(iter/s)": 0.449609 + }, + { + "acc": 0.65647864, + "epoch": 0.38533739218670726, + "grad_norm": 6.90625, + "learning_rate": 9.453765594367014e-06, + "loss": 1.65025444, + "memory(GiB)": 111.15, + "step": 15190, + "train_speed(iter/s)": 0.449687 + }, + { + "acc": 0.65487814, + "epoch": 0.3854642313546423, + "grad_norm": 5.0625, + "learning_rate": 9.453288910075196e-06, + "loss": 1.62743378, + "memory(GiB)": 111.15, + "step": 15195, + "train_speed(iter/s)": 0.449766 + }, + { + "acc": 0.64754405, + "epoch": 0.38559107052257735, + "grad_norm": 6.375, + "learning_rate": 9.452812029907849e-06, + "loss": 1.65987091, + "memory(GiB)": 111.15, + "step": 15200, + "train_speed(iter/s)": 0.449845 + }, + { + "epoch": 0.38559107052257735, + "eval_acc": 0.6397020149335102, + "eval_loss": 1.6289994716644287, + "eval_runtime": 112.6041, + "eval_samples_per_second": 56.57, + "eval_steps_per_second": 28.285, + "step": 15200 + }, + { + "acc": 0.63997431, + "epoch": 0.38571790969051245, + "grad_norm": 5.0625, + "learning_rate": 9.452334953885951e-06, + "loss": 1.65040684, + "memory(GiB)": 111.15, + "step": 15205, + "train_speed(iter/s)": 0.448369 + }, + { + "acc": 0.64217138, + "epoch": 0.3858447488584475, + "grad_norm": 5.46875, + "learning_rate": 9.451857682030481e-06, + "loss": 1.69071159, + "memory(GiB)": 111.15, + "step": 15210, + "train_speed(iter/s)": 0.448448 + }, + { + "acc": 0.66231194, + "epoch": 0.38597158802638254, + "grad_norm": 5.5625, + "learning_rate": 9.451380214362436e-06, + "loss": 1.62620201, + "memory(GiB)": 111.15, + "step": 15215, + "train_speed(iter/s)": 0.448526 + }, + { + "acc": 0.65221553, + "epoch": 0.3860984271943176, + "grad_norm": 4.90625, + "learning_rate": 9.450902550902814e-06, + "loss": 1.62973309, + "memory(GiB)": 111.15, + "step": 15220, + "train_speed(iter/s)": 0.448604 + }, + { + "acc": 0.63812065, + "epoch": 0.3862252663622527, + "grad_norm": 5.28125, + "learning_rate": 9.450424691672626e-06, + "loss": 1.70745049, + "memory(GiB)": 111.15, + "step": 15225, + "train_speed(iter/s)": 0.448682 + }, + { + "acc": 0.6533679, + "epoch": 0.38635210553018773, + "grad_norm": 5.53125, + "learning_rate": 9.449946636692891e-06, + "loss": 1.61814213, + "memory(GiB)": 111.15, + "step": 15230, + "train_speed(iter/s)": 0.448759 + }, + { + "acc": 0.64956169, + "epoch": 0.3864789446981228, + "grad_norm": 5.78125, + "learning_rate": 9.449468385984634e-06, + "loss": 1.6199892, + "memory(GiB)": 111.15, + "step": 15235, + "train_speed(iter/s)": 0.448838 + }, + { + "acc": 0.65661983, + "epoch": 0.3866057838660578, + "grad_norm": 6.03125, + "learning_rate": 9.448989939568892e-06, + "loss": 1.65606689, + "memory(GiB)": 111.15, + "step": 15240, + "train_speed(iter/s)": 0.448916 + }, + { + "acc": 0.62977877, + "epoch": 0.3867326230339929, + "grad_norm": 4.65625, + "learning_rate": 9.448511297466708e-06, + "loss": 1.70530624, + "memory(GiB)": 111.15, + "step": 15245, + "train_speed(iter/s)": 0.448994 + }, + { + "acc": 0.64273548, + "epoch": 0.38685946220192796, + "grad_norm": 5.125, + "learning_rate": 9.448032459699139e-06, + "loss": 1.64327698, + "memory(GiB)": 111.15, + "step": 15250, + "train_speed(iter/s)": 0.449072 + }, + { + "acc": 0.63000369, + "epoch": 0.386986301369863, + "grad_norm": 5.1875, + "learning_rate": 9.447553426287244e-06, + "loss": 1.73388176, + "memory(GiB)": 111.15, + "step": 15255, + "train_speed(iter/s)": 0.44915 + }, + { + "acc": 0.64731398, + "epoch": 0.38711314053779805, + "grad_norm": 4.75, + "learning_rate": 9.44707419725209e-06, + "loss": 1.63816948, + "memory(GiB)": 111.15, + "step": 15260, + "train_speed(iter/s)": 0.449229 + }, + { + "acc": 0.63824806, + "epoch": 0.38723997970573315, + "grad_norm": 5.375, + "learning_rate": 9.446594772614759e-06, + "loss": 1.69819431, + "memory(GiB)": 111.15, + "step": 15265, + "train_speed(iter/s)": 0.449306 + }, + { + "acc": 0.64304609, + "epoch": 0.3873668188736682, + "grad_norm": 5.1875, + "learning_rate": 9.446115152396335e-06, + "loss": 1.61759911, + "memory(GiB)": 111.15, + "step": 15270, + "train_speed(iter/s)": 0.449384 + }, + { + "acc": 0.66238666, + "epoch": 0.38749365804160324, + "grad_norm": 6.0625, + "learning_rate": 9.445635336617919e-06, + "loss": 1.56813908, + "memory(GiB)": 111.15, + "step": 15275, + "train_speed(iter/s)": 0.449462 + }, + { + "acc": 0.64201412, + "epoch": 0.3876204972095383, + "grad_norm": 4.09375, + "learning_rate": 9.445155325300612e-06, + "loss": 1.68102913, + "memory(GiB)": 111.15, + "step": 15280, + "train_speed(iter/s)": 0.44954 + }, + { + "acc": 0.65047398, + "epoch": 0.3877473363774734, + "grad_norm": 5.5, + "learning_rate": 9.444675118465528e-06, + "loss": 1.65923939, + "memory(GiB)": 111.15, + "step": 15285, + "train_speed(iter/s)": 0.449617 + }, + { + "acc": 0.64145823, + "epoch": 0.38787417554540843, + "grad_norm": 4.84375, + "learning_rate": 9.444194716133785e-06, + "loss": 1.7031292, + "memory(GiB)": 111.15, + "step": 15290, + "train_speed(iter/s)": 0.449695 + }, + { + "acc": 0.65741835, + "epoch": 0.3880010147133435, + "grad_norm": 4.9375, + "learning_rate": 9.44371411832652e-06, + "loss": 1.57439222, + "memory(GiB)": 111.15, + "step": 15295, + "train_speed(iter/s)": 0.449773 + }, + { + "acc": 0.63437672, + "epoch": 0.3881278538812785, + "grad_norm": 5.25, + "learning_rate": 9.443233325064867e-06, + "loss": 1.70402832, + "memory(GiB)": 111.15, + "step": 15300, + "train_speed(iter/s)": 0.449851 + }, + { + "epoch": 0.3881278538812785, + "eval_acc": 0.6397149657201034, + "eval_loss": 1.6287158727645874, + "eval_runtime": 112.032, + "eval_samples_per_second": 56.859, + "eval_steps_per_second": 28.429, + "step": 15300 + }, + { + "acc": 0.64509773, + "epoch": 0.3882546930492136, + "grad_norm": 4.75, + "learning_rate": 9.442752336369976e-06, + "loss": 1.61056633, + "memory(GiB)": 111.15, + "step": 15305, + "train_speed(iter/s)": 0.448391 + }, + { + "acc": 0.66699753, + "epoch": 0.38838153221714866, + "grad_norm": 5.09375, + "learning_rate": 9.442271152263e-06, + "loss": 1.55269375, + "memory(GiB)": 111.15, + "step": 15310, + "train_speed(iter/s)": 0.448467 + }, + { + "acc": 0.64906425, + "epoch": 0.3885083713850837, + "grad_norm": 4.6875, + "learning_rate": 9.441789772765107e-06, + "loss": 1.65491867, + "memory(GiB)": 111.15, + "step": 15315, + "train_speed(iter/s)": 0.448544 + }, + { + "acc": 0.64233923, + "epoch": 0.38863521055301875, + "grad_norm": 5.03125, + "learning_rate": 9.441308197897467e-06, + "loss": 1.66000843, + "memory(GiB)": 111.15, + "step": 15320, + "train_speed(iter/s)": 0.448621 + }, + { + "acc": 0.66120858, + "epoch": 0.38876204972095385, + "grad_norm": 5.6875, + "learning_rate": 9.440826427681264e-06, + "loss": 1.70942841, + "memory(GiB)": 111.15, + "step": 15325, + "train_speed(iter/s)": 0.448698 + }, + { + "acc": 0.64997606, + "epoch": 0.3888888888888889, + "grad_norm": 6.21875, + "learning_rate": 9.44034446213769e-06, + "loss": 1.65782032, + "memory(GiB)": 111.15, + "step": 15330, + "train_speed(iter/s)": 0.448775 + }, + { + "acc": 0.64528971, + "epoch": 0.38901572805682394, + "grad_norm": 5.34375, + "learning_rate": 9.439862301287939e-06, + "loss": 1.63670292, + "memory(GiB)": 111.15, + "step": 15335, + "train_speed(iter/s)": 0.448851 + }, + { + "acc": 0.64188776, + "epoch": 0.389142567224759, + "grad_norm": 5.21875, + "learning_rate": 9.439379945153223e-06, + "loss": 1.71291142, + "memory(GiB)": 111.15, + "step": 15340, + "train_speed(iter/s)": 0.448929 + }, + { + "acc": 0.66085248, + "epoch": 0.3892694063926941, + "grad_norm": 5.53125, + "learning_rate": 9.438897393754755e-06, + "loss": 1.58937817, + "memory(GiB)": 111.15, + "step": 15345, + "train_speed(iter/s)": 0.449007 + }, + { + "acc": 0.65259256, + "epoch": 0.38939624556062913, + "grad_norm": 4.9375, + "learning_rate": 9.438414647113762e-06, + "loss": 1.65180779, + "memory(GiB)": 111.15, + "step": 15350, + "train_speed(iter/s)": 0.449083 + }, + { + "acc": 0.64670777, + "epoch": 0.3895230847285642, + "grad_norm": 5.03125, + "learning_rate": 9.437931705251478e-06, + "loss": 1.65523129, + "memory(GiB)": 111.15, + "step": 15355, + "train_speed(iter/s)": 0.44916 + }, + { + "acc": 0.64699469, + "epoch": 0.3896499238964992, + "grad_norm": 4.65625, + "learning_rate": 9.437448568189142e-06, + "loss": 1.60090027, + "memory(GiB)": 111.15, + "step": 15360, + "train_speed(iter/s)": 0.449237 + }, + { + "acc": 0.65053358, + "epoch": 0.3897767630644343, + "grad_norm": 4.90625, + "learning_rate": 9.436965235948008e-06, + "loss": 1.58004894, + "memory(GiB)": 111.15, + "step": 15365, + "train_speed(iter/s)": 0.449314 + }, + { + "acc": 0.65488625, + "epoch": 0.38990360223236936, + "grad_norm": 4.84375, + "learning_rate": 9.436481708549332e-06, + "loss": 1.6598877, + "memory(GiB)": 111.15, + "step": 15370, + "train_speed(iter/s)": 0.449391 + }, + { + "acc": 0.64302225, + "epoch": 0.3900304414003044, + "grad_norm": 5.15625, + "learning_rate": 9.435997986014382e-06, + "loss": 1.70952606, + "memory(GiB)": 111.15, + "step": 15375, + "train_speed(iter/s)": 0.449468 + }, + { + "acc": 0.63212361, + "epoch": 0.39015728056823945, + "grad_norm": 5.375, + "learning_rate": 9.435514068364437e-06, + "loss": 1.76018486, + "memory(GiB)": 111.15, + "step": 15380, + "train_speed(iter/s)": 0.449545 + }, + { + "acc": 0.65054703, + "epoch": 0.39028411973617455, + "grad_norm": 5.25, + "learning_rate": 9.43502995562078e-06, + "loss": 1.64097748, + "memory(GiB)": 111.15, + "step": 15385, + "train_speed(iter/s)": 0.44962 + }, + { + "acc": 0.64780569, + "epoch": 0.3904109589041096, + "grad_norm": 6.625, + "learning_rate": 9.434545647804703e-06, + "loss": 1.67113419, + "memory(GiB)": 111.15, + "step": 15390, + "train_speed(iter/s)": 0.449697 + }, + { + "acc": 0.65801506, + "epoch": 0.39053779807204464, + "grad_norm": 5.6875, + "learning_rate": 9.434061144937512e-06, + "loss": 1.61923561, + "memory(GiB)": 111.15, + "step": 15395, + "train_speed(iter/s)": 0.449774 + }, + { + "acc": 0.65816479, + "epoch": 0.3906646372399797, + "grad_norm": 5.625, + "learning_rate": 9.433576447040513e-06, + "loss": 1.59084454, + "memory(GiB)": 111.15, + "step": 15400, + "train_speed(iter/s)": 0.44985 + }, + { + "epoch": 0.3906646372399797, + "eval_acc": 0.6397124591162466, + "eval_loss": 1.62845778465271, + "eval_runtime": 113.8332, + "eval_samples_per_second": 55.959, + "eval_steps_per_second": 27.98, + "step": 15400 + }, + { + "acc": 0.65243464, + "epoch": 0.3907914764079148, + "grad_norm": 4.75, + "learning_rate": 9.433091554135029e-06, + "loss": 1.67814903, + "memory(GiB)": 111.15, + "step": 15405, + "train_speed(iter/s)": 0.448377 + }, + { + "acc": 0.64383297, + "epoch": 0.39091831557584983, + "grad_norm": 5.625, + "learning_rate": 9.432606466242384e-06, + "loss": 1.68169594, + "memory(GiB)": 111.15, + "step": 15410, + "train_speed(iter/s)": 0.448453 + }, + { + "acc": 0.64857321, + "epoch": 0.3910451547437849, + "grad_norm": 4.71875, + "learning_rate": 9.43212118338392e-06, + "loss": 1.62442493, + "memory(GiB)": 111.15, + "step": 15415, + "train_speed(iter/s)": 0.44853 + }, + { + "acc": 0.64737511, + "epoch": 0.3911719939117199, + "grad_norm": 4.875, + "learning_rate": 9.431635705580975e-06, + "loss": 1.6328764, + "memory(GiB)": 111.15, + "step": 15420, + "train_speed(iter/s)": 0.448606 + }, + { + "acc": 0.64507885, + "epoch": 0.391298833079655, + "grad_norm": 5.875, + "learning_rate": 9.431150032854907e-06, + "loss": 1.64898911, + "memory(GiB)": 111.15, + "step": 15425, + "train_speed(iter/s)": 0.448683 + }, + { + "acc": 0.64797716, + "epoch": 0.39142567224759006, + "grad_norm": 5.71875, + "learning_rate": 9.430664165227077e-06, + "loss": 1.58477478, + "memory(GiB)": 111.15, + "step": 15430, + "train_speed(iter/s)": 0.448759 + }, + { + "acc": 0.65512757, + "epoch": 0.3915525114155251, + "grad_norm": 4.71875, + "learning_rate": 9.430178102718857e-06, + "loss": 1.65286331, + "memory(GiB)": 111.15, + "step": 15435, + "train_speed(iter/s)": 0.448836 + }, + { + "acc": 0.65792456, + "epoch": 0.39167935058346015, + "grad_norm": 5.1875, + "learning_rate": 9.429691845351623e-06, + "loss": 1.60286751, + "memory(GiB)": 111.15, + "step": 15440, + "train_speed(iter/s)": 0.448912 + }, + { + "acc": 0.64617133, + "epoch": 0.39180618975139525, + "grad_norm": 5.125, + "learning_rate": 9.429205393146763e-06, + "loss": 1.59066696, + "memory(GiB)": 111.15, + "step": 15445, + "train_speed(iter/s)": 0.448989 + }, + { + "acc": 0.63782969, + "epoch": 0.3919330289193303, + "grad_norm": 5.03125, + "learning_rate": 9.428718746125678e-06, + "loss": 1.69889832, + "memory(GiB)": 111.15, + "step": 15450, + "train_speed(iter/s)": 0.449066 + }, + { + "acc": 0.65231819, + "epoch": 0.39205986808726534, + "grad_norm": 5.59375, + "learning_rate": 9.428231904309768e-06, + "loss": 1.67045956, + "memory(GiB)": 111.15, + "step": 15455, + "train_speed(iter/s)": 0.449142 + }, + { + "acc": 0.63913121, + "epoch": 0.3921867072552004, + "grad_norm": 5.34375, + "learning_rate": 9.427744867720448e-06, + "loss": 1.70090332, + "memory(GiB)": 111.15, + "step": 15460, + "train_speed(iter/s)": 0.449219 + }, + { + "acc": 0.6478672, + "epoch": 0.3923135464231355, + "grad_norm": 5.59375, + "learning_rate": 9.42725763637914e-06, + "loss": 1.65929279, + "memory(GiB)": 111.15, + "step": 15465, + "train_speed(iter/s)": 0.449295 + }, + { + "acc": 0.65675707, + "epoch": 0.39244038559107053, + "grad_norm": 5.34375, + "learning_rate": 9.426770210307277e-06, + "loss": 1.6144619, + "memory(GiB)": 111.15, + "step": 15470, + "train_speed(iter/s)": 0.449372 + }, + { + "acc": 0.63946505, + "epoch": 0.3925672247590056, + "grad_norm": 5.3125, + "learning_rate": 9.426282589526294e-06, + "loss": 1.67030449, + "memory(GiB)": 111.15, + "step": 15475, + "train_speed(iter/s)": 0.449449 + }, + { + "acc": 0.65694008, + "epoch": 0.3926940639269406, + "grad_norm": 4.5625, + "learning_rate": 9.425794774057641e-06, + "loss": 1.58313389, + "memory(GiB)": 111.15, + "step": 15480, + "train_speed(iter/s)": 0.449525 + }, + { + "acc": 0.65447054, + "epoch": 0.3928209030948757, + "grad_norm": 4.875, + "learning_rate": 9.425306763922775e-06, + "loss": 1.61988182, + "memory(GiB)": 111.15, + "step": 15485, + "train_speed(iter/s)": 0.449601 + }, + { + "acc": 0.6263772, + "epoch": 0.39294774226281076, + "grad_norm": 5.84375, + "learning_rate": 9.42481855914316e-06, + "loss": 1.70543098, + "memory(GiB)": 111.15, + "step": 15490, + "train_speed(iter/s)": 0.449677 + }, + { + "acc": 0.63312292, + "epoch": 0.3930745814307458, + "grad_norm": 5.6875, + "learning_rate": 9.424330159740269e-06, + "loss": 1.650597, + "memory(GiB)": 111.15, + "step": 15495, + "train_speed(iter/s)": 0.449754 + }, + { + "acc": 0.6612793, + "epoch": 0.39320142059868085, + "grad_norm": 4.84375, + "learning_rate": 9.423841565735582e-06, + "loss": 1.56348343, + "memory(GiB)": 111.15, + "step": 15500, + "train_speed(iter/s)": 0.449828 + }, + { + "epoch": 0.39320142059868085, + "eval_acc": 0.639915494028643, + "eval_loss": 1.6281678676605225, + "eval_runtime": 113.8609, + "eval_samples_per_second": 55.945, + "eval_steps_per_second": 27.973, + "step": 15500 + }, + { + "acc": 0.64893351, + "epoch": 0.39332825976661595, + "grad_norm": 5.25, + "learning_rate": 9.423352777150597e-06, + "loss": 1.70626698, + "memory(GiB)": 111.15, + "step": 15505, + "train_speed(iter/s)": 0.448365 + }, + { + "acc": 0.65988979, + "epoch": 0.393455098934551, + "grad_norm": 6.0625, + "learning_rate": 9.422863794006804e-06, + "loss": 1.63095131, + "memory(GiB)": 111.15, + "step": 15510, + "train_speed(iter/s)": 0.448442 + }, + { + "acc": 0.63694048, + "epoch": 0.39358193810248604, + "grad_norm": 4.71875, + "learning_rate": 9.422374616325716e-06, + "loss": 1.67222519, + "memory(GiB)": 111.15, + "step": 15515, + "train_speed(iter/s)": 0.448518 + }, + { + "acc": 0.64123092, + "epoch": 0.3937087772704211, + "grad_norm": 5.71875, + "learning_rate": 9.421885244128847e-06, + "loss": 1.65485954, + "memory(GiB)": 111.15, + "step": 15520, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.64949789, + "epoch": 0.3938356164383562, + "grad_norm": 5.53125, + "learning_rate": 9.421395677437724e-06, + "loss": 1.62031002, + "memory(GiB)": 111.15, + "step": 15525, + "train_speed(iter/s)": 0.448671 + }, + { + "acc": 0.66366763, + "epoch": 0.39396245560629123, + "grad_norm": 6.125, + "learning_rate": 9.42090591627388e-06, + "loss": 1.61398182, + "memory(GiB)": 111.15, + "step": 15530, + "train_speed(iter/s)": 0.448746 + }, + { + "acc": 0.64333267, + "epoch": 0.3940892947742263, + "grad_norm": 6.375, + "learning_rate": 9.420415960658853e-06, + "loss": 1.62337646, + "memory(GiB)": 111.15, + "step": 15535, + "train_speed(iter/s)": 0.448823 + }, + { + "acc": 0.64654479, + "epoch": 0.3942161339421613, + "grad_norm": 5.28125, + "learning_rate": 9.419925810614196e-06, + "loss": 1.66029205, + "memory(GiB)": 111.15, + "step": 15540, + "train_speed(iter/s)": 0.448899 + }, + { + "acc": 0.63427815, + "epoch": 0.3943429731100964, + "grad_norm": 7.78125, + "learning_rate": 9.419435466161471e-06, + "loss": 1.66250477, + "memory(GiB)": 111.15, + "step": 15545, + "train_speed(iter/s)": 0.448976 + }, + { + "acc": 0.63966503, + "epoch": 0.39446981227803146, + "grad_norm": 5.375, + "learning_rate": 9.418944927322242e-06, + "loss": 1.7138319, + "memory(GiB)": 111.15, + "step": 15550, + "train_speed(iter/s)": 0.449052 + }, + { + "acc": 0.65564003, + "epoch": 0.3945966514459665, + "grad_norm": 5.59375, + "learning_rate": 9.418454194118085e-06, + "loss": 1.61816368, + "memory(GiB)": 111.15, + "step": 15555, + "train_speed(iter/s)": 0.449128 + }, + { + "acc": 0.66004262, + "epoch": 0.39472349061390155, + "grad_norm": 5.375, + "learning_rate": 9.417963266570587e-06, + "loss": 1.67379761, + "memory(GiB)": 111.15, + "step": 15560, + "train_speed(iter/s)": 0.449203 + }, + { + "acc": 0.64733167, + "epoch": 0.39485032978183665, + "grad_norm": 4.25, + "learning_rate": 9.417472144701338e-06, + "loss": 1.62819099, + "memory(GiB)": 111.15, + "step": 15565, + "train_speed(iter/s)": 0.44928 + }, + { + "acc": 0.66106672, + "epoch": 0.3949771689497717, + "grad_norm": 5.21875, + "learning_rate": 9.416980828531944e-06, + "loss": 1.56777678, + "memory(GiB)": 111.15, + "step": 15570, + "train_speed(iter/s)": 0.449356 + }, + { + "acc": 0.63666096, + "epoch": 0.39510400811770674, + "grad_norm": 5.3125, + "learning_rate": 9.41648931808401e-06, + "loss": 1.69103699, + "memory(GiB)": 111.15, + "step": 15575, + "train_speed(iter/s)": 0.449433 + }, + { + "acc": 0.64592295, + "epoch": 0.3952308472856418, + "grad_norm": 5.46875, + "learning_rate": 9.415997613379159e-06, + "loss": 1.62411423, + "memory(GiB)": 111.15, + "step": 15580, + "train_speed(iter/s)": 0.449509 + }, + { + "acc": 0.64935884, + "epoch": 0.3953576864535769, + "grad_norm": 5.4375, + "learning_rate": 9.415505714439016e-06, + "loss": 1.63873692, + "memory(GiB)": 111.15, + "step": 15585, + "train_speed(iter/s)": 0.449585 + }, + { + "acc": 0.63681211, + "epoch": 0.39548452562151193, + "grad_norm": 4.78125, + "learning_rate": 9.415013621285219e-06, + "loss": 1.66509705, + "memory(GiB)": 111.15, + "step": 15590, + "train_speed(iter/s)": 0.44966 + }, + { + "acc": 0.65156126, + "epoch": 0.395611364789447, + "grad_norm": 5.8125, + "learning_rate": 9.41452133393941e-06, + "loss": 1.56872368, + "memory(GiB)": 111.15, + "step": 15595, + "train_speed(iter/s)": 0.449737 + }, + { + "acc": 0.66403999, + "epoch": 0.395738203957382, + "grad_norm": 4.65625, + "learning_rate": 9.414028852423245e-06, + "loss": 1.61125298, + "memory(GiB)": 111.15, + "step": 15600, + "train_speed(iter/s)": 0.449813 + }, + { + "epoch": 0.395738203957382, + "eval_acc": 0.6397838973261639, + "eval_loss": 1.6280382871627808, + "eval_runtime": 114.3398, + "eval_samples_per_second": 55.711, + "eval_steps_per_second": 27.856, + "step": 15600 + }, + { + "acc": 0.65908194, + "epoch": 0.3958650431253171, + "grad_norm": 7.03125, + "learning_rate": 9.413536176758384e-06, + "loss": 1.60531044, + "memory(GiB)": 111.15, + "step": 15605, + "train_speed(iter/s)": 0.448352 + }, + { + "acc": 0.63965044, + "epoch": 0.39599188229325216, + "grad_norm": 5.5625, + "learning_rate": 9.413043306966496e-06, + "loss": 1.66303253, + "memory(GiB)": 111.15, + "step": 15610, + "train_speed(iter/s)": 0.448428 + }, + { + "acc": 0.66416321, + "epoch": 0.3961187214611872, + "grad_norm": 6.34375, + "learning_rate": 9.41255024306926e-06, + "loss": 1.58335915, + "memory(GiB)": 111.15, + "step": 15615, + "train_speed(iter/s)": 0.448504 + }, + { + "acc": 0.66371713, + "epoch": 0.39624556062912225, + "grad_norm": 5.78125, + "learning_rate": 9.412056985088364e-06, + "loss": 1.69100723, + "memory(GiB)": 111.15, + "step": 15620, + "train_speed(iter/s)": 0.448579 + }, + { + "acc": 0.64772377, + "epoch": 0.39637239979705735, + "grad_norm": 6.03125, + "learning_rate": 9.411563533045505e-06, + "loss": 1.63617744, + "memory(GiB)": 111.15, + "step": 15625, + "train_speed(iter/s)": 0.448655 + }, + { + "acc": 0.63429203, + "epoch": 0.3964992389649924, + "grad_norm": 6.1875, + "learning_rate": 9.411069886962383e-06, + "loss": 1.69695435, + "memory(GiB)": 111.15, + "step": 15630, + "train_speed(iter/s)": 0.44873 + }, + { + "acc": 0.63529005, + "epoch": 0.39662607813292744, + "grad_norm": 6.875, + "learning_rate": 9.410576046860716e-06, + "loss": 1.68165302, + "memory(GiB)": 111.15, + "step": 15635, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.65903158, + "epoch": 0.3967529173008625, + "grad_norm": 4.96875, + "learning_rate": 9.41008201276222e-06, + "loss": 1.61599846, + "memory(GiB)": 111.15, + "step": 15640, + "train_speed(iter/s)": 0.448881 + }, + { + "acc": 0.65936365, + "epoch": 0.3968797564687976, + "grad_norm": 4.96875, + "learning_rate": 9.409587784688629e-06, + "loss": 1.65689201, + "memory(GiB)": 111.15, + "step": 15645, + "train_speed(iter/s)": 0.448956 + }, + { + "acc": 0.6212923, + "epoch": 0.39700659563673263, + "grad_norm": 5.3125, + "learning_rate": 9.409093362661678e-06, + "loss": 1.73159027, + "memory(GiB)": 111.15, + "step": 15650, + "train_speed(iter/s)": 0.449032 + }, + { + "acc": 0.65028448, + "epoch": 0.3971334348046677, + "grad_norm": 5.875, + "learning_rate": 9.408598746703119e-06, + "loss": 1.62613831, + "memory(GiB)": 111.15, + "step": 15655, + "train_speed(iter/s)": 0.449107 + }, + { + "acc": 0.64887342, + "epoch": 0.3972602739726027, + "grad_norm": 5.875, + "learning_rate": 9.408103936834703e-06, + "loss": 1.68367844, + "memory(GiB)": 111.15, + "step": 15660, + "train_speed(iter/s)": 0.449183 + }, + { + "acc": 0.64645014, + "epoch": 0.3973871131405378, + "grad_norm": 5.25, + "learning_rate": 9.407608933078194e-06, + "loss": 1.67176781, + "memory(GiB)": 111.15, + "step": 15665, + "train_speed(iter/s)": 0.449258 + }, + { + "acc": 0.64907446, + "epoch": 0.39751395230847286, + "grad_norm": 4.6875, + "learning_rate": 9.407113735455366e-06, + "loss": 1.62852554, + "memory(GiB)": 111.15, + "step": 15670, + "train_speed(iter/s)": 0.449334 + }, + { + "acc": 0.65387764, + "epoch": 0.3976407914764079, + "grad_norm": 6.0625, + "learning_rate": 9.406618343988e-06, + "loss": 1.6108139, + "memory(GiB)": 111.15, + "step": 15675, + "train_speed(iter/s)": 0.44941 + }, + { + "acc": 0.64860325, + "epoch": 0.39776763064434295, + "grad_norm": 5.03125, + "learning_rate": 9.406122758697885e-06, + "loss": 1.65722046, + "memory(GiB)": 111.15, + "step": 15680, + "train_speed(iter/s)": 0.449485 + }, + { + "acc": 0.64742594, + "epoch": 0.39789446981227805, + "grad_norm": 4.71875, + "learning_rate": 9.405626979606819e-06, + "loss": 1.68592949, + "memory(GiB)": 111.15, + "step": 15685, + "train_speed(iter/s)": 0.44956 + }, + { + "acc": 0.64032421, + "epoch": 0.3980213089802131, + "grad_norm": 4.75, + "learning_rate": 9.405131006736608e-06, + "loss": 1.64602661, + "memory(GiB)": 111.15, + "step": 15690, + "train_speed(iter/s)": 0.449636 + }, + { + "acc": 0.65909767, + "epoch": 0.39814814814814814, + "grad_norm": 5.5, + "learning_rate": 9.404634840109069e-06, + "loss": 1.65061569, + "memory(GiB)": 111.15, + "step": 15695, + "train_speed(iter/s)": 0.449711 + }, + { + "acc": 0.64360504, + "epoch": 0.3982749873160832, + "grad_norm": 5.125, + "learning_rate": 9.404138479746022e-06, + "loss": 1.64114723, + "memory(GiB)": 111.15, + "step": 15700, + "train_speed(iter/s)": 0.449786 + }, + { + "epoch": 0.3982749873160832, + "eval_acc": 0.6398198253147772, + "eval_loss": 1.627565860748291, + "eval_runtime": 113.6024, + "eval_samples_per_second": 56.073, + "eval_steps_per_second": 28.036, + "step": 15700 + }, + { + "acc": 0.65060625, + "epoch": 0.3984018264840183, + "grad_norm": 5.28125, + "learning_rate": 9.403641925669304e-06, + "loss": 1.64542904, + "memory(GiB)": 111.15, + "step": 15705, + "train_speed(iter/s)": 0.448345 + }, + { + "acc": 0.6432219, + "epoch": 0.39852866565195333, + "grad_norm": 4.15625, + "learning_rate": 9.403145177900752e-06, + "loss": 1.67417469, + "memory(GiB)": 111.15, + "step": 15710, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.64616485, + "epoch": 0.3986555048198884, + "grad_norm": 5.75, + "learning_rate": 9.402648236462217e-06, + "loss": 1.64328499, + "memory(GiB)": 111.15, + "step": 15715, + "train_speed(iter/s)": 0.448494 + }, + { + "acc": 0.65105047, + "epoch": 0.3987823439878234, + "grad_norm": 4.46875, + "learning_rate": 9.402151101375557e-06, + "loss": 1.64415855, + "memory(GiB)": 111.15, + "step": 15720, + "train_speed(iter/s)": 0.448568 + }, + { + "acc": 0.63326621, + "epoch": 0.3989091831557585, + "grad_norm": 4.90625, + "learning_rate": 9.401653772662638e-06, + "loss": 1.66724129, + "memory(GiB)": 111.15, + "step": 15725, + "train_speed(iter/s)": 0.448643 + }, + { + "acc": 0.64757051, + "epoch": 0.39903602232369356, + "grad_norm": 4.78125, + "learning_rate": 9.401156250345331e-06, + "loss": 1.66217575, + "memory(GiB)": 111.15, + "step": 15730, + "train_speed(iter/s)": 0.448717 + }, + { + "acc": 0.65520287, + "epoch": 0.3991628614916286, + "grad_norm": 5.25, + "learning_rate": 9.400658534445524e-06, + "loss": 1.61325703, + "memory(GiB)": 111.15, + "step": 15735, + "train_speed(iter/s)": 0.448791 + }, + { + "acc": 0.63208036, + "epoch": 0.39928970065956365, + "grad_norm": 5.59375, + "learning_rate": 9.40016062498511e-06, + "loss": 1.68947563, + "memory(GiB)": 111.15, + "step": 15740, + "train_speed(iter/s)": 0.448866 + }, + { + "acc": 0.63864808, + "epoch": 0.39941653982749875, + "grad_norm": 5.65625, + "learning_rate": 9.399662521985982e-06, + "loss": 1.66203346, + "memory(GiB)": 111.15, + "step": 15745, + "train_speed(iter/s)": 0.44894 + }, + { + "acc": 0.65313511, + "epoch": 0.3995433789954338, + "grad_norm": 6.34375, + "learning_rate": 9.399164225470055e-06, + "loss": 1.69805832, + "memory(GiB)": 111.15, + "step": 15750, + "train_speed(iter/s)": 0.449014 + }, + { + "acc": 0.65095124, + "epoch": 0.39967021816336884, + "grad_norm": 5.03125, + "learning_rate": 9.398665735459245e-06, + "loss": 1.67277527, + "memory(GiB)": 111.15, + "step": 15755, + "train_speed(iter/s)": 0.44909 + }, + { + "acc": 0.62923899, + "epoch": 0.3997970573313039, + "grad_norm": 4.15625, + "learning_rate": 9.398167051975475e-06, + "loss": 1.69110222, + "memory(GiB)": 111.15, + "step": 15760, + "train_speed(iter/s)": 0.449164 + }, + { + "acc": 0.64613557, + "epoch": 0.399923896499239, + "grad_norm": 6.15625, + "learning_rate": 9.397668175040684e-06, + "loss": 1.63880711, + "memory(GiB)": 111.15, + "step": 15765, + "train_speed(iter/s)": 0.449239 + }, + { + "acc": 0.63017702, + "epoch": 0.40005073566717403, + "grad_norm": 5.125, + "learning_rate": 9.397169104676813e-06, + "loss": 1.68737602, + "memory(GiB)": 111.15, + "step": 15770, + "train_speed(iter/s)": 0.449313 + }, + { + "acc": 0.63325081, + "epoch": 0.4001775748351091, + "grad_norm": 4.6875, + "learning_rate": 9.39666984090581e-06, + "loss": 1.68690796, + "memory(GiB)": 111.15, + "step": 15775, + "train_speed(iter/s)": 0.449388 + }, + { + "acc": 0.65418549, + "epoch": 0.4003044140030441, + "grad_norm": 5.1875, + "learning_rate": 9.396170383749642e-06, + "loss": 1.62001362, + "memory(GiB)": 111.15, + "step": 15780, + "train_speed(iter/s)": 0.449462 + }, + { + "acc": 0.6529161, + "epoch": 0.4004312531709792, + "grad_norm": 4.59375, + "learning_rate": 9.39567073323027e-06, + "loss": 1.61634197, + "memory(GiB)": 111.15, + "step": 15785, + "train_speed(iter/s)": 0.449537 + }, + { + "acc": 0.63297606, + "epoch": 0.40055809233891426, + "grad_norm": 5.375, + "learning_rate": 9.395170889369674e-06, + "loss": 1.70829277, + "memory(GiB)": 111.15, + "step": 15790, + "train_speed(iter/s)": 0.449612 + }, + { + "acc": 0.65918269, + "epoch": 0.4006849315068493, + "grad_norm": 5.0625, + "learning_rate": 9.39467085218984e-06, + "loss": 1.65314789, + "memory(GiB)": 111.15, + "step": 15795, + "train_speed(iter/s)": 0.449686 + }, + { + "acc": 0.63931856, + "epoch": 0.40081177067478435, + "grad_norm": 4.8125, + "learning_rate": 9.39417062171276e-06, + "loss": 1.66523399, + "memory(GiB)": 111.15, + "step": 15800, + "train_speed(iter/s)": 0.449761 + }, + { + "epoch": 0.40081177067478435, + "eval_acc": 0.6398904279900756, + "eval_loss": 1.6277319192886353, + "eval_runtime": 113.7494, + "eval_samples_per_second": 56.0, + "eval_steps_per_second": 28.0, + "step": 15800 + }, + { + "acc": 0.64944453, + "epoch": 0.40093860984271945, + "grad_norm": 5.625, + "learning_rate": 9.393670197960439e-06, + "loss": 1.65363083, + "memory(GiB)": 111.15, + "step": 15805, + "train_speed(iter/s)": 0.448326 + }, + { + "acc": 0.64821367, + "epoch": 0.4010654490106545, + "grad_norm": 4.53125, + "learning_rate": 9.393169580954884e-06, + "loss": 1.62791157, + "memory(GiB)": 111.15, + "step": 15810, + "train_speed(iter/s)": 0.4484 + }, + { + "acc": 0.64016895, + "epoch": 0.40119228817858954, + "grad_norm": 4.46875, + "learning_rate": 9.392668770718118e-06, + "loss": 1.71036072, + "memory(GiB)": 111.15, + "step": 15815, + "train_speed(iter/s)": 0.448474 + }, + { + "acc": 0.65244312, + "epoch": 0.4013191273465246, + "grad_norm": 5.125, + "learning_rate": 9.392167767272169e-06, + "loss": 1.67083931, + "memory(GiB)": 111.15, + "step": 15820, + "train_speed(iter/s)": 0.448548 + }, + { + "acc": 0.64463968, + "epoch": 0.4014459665144597, + "grad_norm": 4.75, + "learning_rate": 9.39166657063907e-06, + "loss": 1.67240314, + "memory(GiB)": 111.15, + "step": 15825, + "train_speed(iter/s)": 0.448622 + }, + { + "acc": 0.63704596, + "epoch": 0.40157280568239473, + "grad_norm": 4.71875, + "learning_rate": 9.391165180840869e-06, + "loss": 1.65192661, + "memory(GiB)": 111.15, + "step": 15830, + "train_speed(iter/s)": 0.448696 + }, + { + "acc": 0.64868956, + "epoch": 0.4016996448503298, + "grad_norm": 6.21875, + "learning_rate": 9.390663597899619e-06, + "loss": 1.66151657, + "memory(GiB)": 111.15, + "step": 15835, + "train_speed(iter/s)": 0.448771 + }, + { + "acc": 0.65608826, + "epoch": 0.4018264840182648, + "grad_norm": 5.4375, + "learning_rate": 9.39016182183738e-06, + "loss": 1.58175411, + "memory(GiB)": 111.15, + "step": 15840, + "train_speed(iter/s)": 0.448846 + }, + { + "acc": 0.66116343, + "epoch": 0.4019533231861999, + "grad_norm": 5.40625, + "learning_rate": 9.389659852676223e-06, + "loss": 1.61717033, + "memory(GiB)": 111.15, + "step": 15845, + "train_speed(iter/s)": 0.448921 + }, + { + "acc": 0.66856432, + "epoch": 0.40208016235413496, + "grad_norm": 6.21875, + "learning_rate": 9.389157690438228e-06, + "loss": 1.52035027, + "memory(GiB)": 111.15, + "step": 15850, + "train_speed(iter/s)": 0.448996 + }, + { + "acc": 0.64993954, + "epoch": 0.40220700152207, + "grad_norm": 4.9375, + "learning_rate": 9.38865533514548e-06, + "loss": 1.60281887, + "memory(GiB)": 111.15, + "step": 15855, + "train_speed(iter/s)": 0.44907 + }, + { + "acc": 0.66053629, + "epoch": 0.40233384069000505, + "grad_norm": 4.71875, + "learning_rate": 9.388152786820078e-06, + "loss": 1.54785967, + "memory(GiB)": 111.15, + "step": 15860, + "train_speed(iter/s)": 0.449145 + }, + { + "acc": 0.63252573, + "epoch": 0.40246067985794015, + "grad_norm": 5.21875, + "learning_rate": 9.387650045484124e-06, + "loss": 1.77306099, + "memory(GiB)": 111.15, + "step": 15865, + "train_speed(iter/s)": 0.449219 + }, + { + "acc": 0.64841714, + "epoch": 0.4025875190258752, + "grad_norm": 5.46875, + "learning_rate": 9.387147111159734e-06, + "loss": 1.65072746, + "memory(GiB)": 111.15, + "step": 15870, + "train_speed(iter/s)": 0.449294 + }, + { + "acc": 0.64255686, + "epoch": 0.40271435819381024, + "grad_norm": 7.03125, + "learning_rate": 9.386643983869025e-06, + "loss": 1.6453064, + "memory(GiB)": 111.15, + "step": 15875, + "train_speed(iter/s)": 0.449368 + }, + { + "acc": 0.65005684, + "epoch": 0.4028411973617453, + "grad_norm": 6.0625, + "learning_rate": 9.386140663634127e-06, + "loss": 1.72945061, + "memory(GiB)": 111.15, + "step": 15880, + "train_speed(iter/s)": 0.449442 + }, + { + "acc": 0.6478961, + "epoch": 0.4029680365296804, + "grad_norm": 5.3125, + "learning_rate": 9.385637150477182e-06, + "loss": 1.68942375, + "memory(GiB)": 111.15, + "step": 15885, + "train_speed(iter/s)": 0.449517 + }, + { + "acc": 0.63437786, + "epoch": 0.40309487569761543, + "grad_norm": 6.03125, + "learning_rate": 9.385133444420333e-06, + "loss": 1.71772919, + "memory(GiB)": 111.15, + "step": 15890, + "train_speed(iter/s)": 0.449592 + }, + { + "acc": 0.66308455, + "epoch": 0.4032217148655505, + "grad_norm": 6.46875, + "learning_rate": 9.384629545485738e-06, + "loss": 1.56709146, + "memory(GiB)": 111.15, + "step": 15895, + "train_speed(iter/s)": 0.449665 + }, + { + "acc": 0.63499851, + "epoch": 0.4033485540334855, + "grad_norm": 5.03125, + "learning_rate": 9.38412545369556e-06, + "loss": 1.65431633, + "memory(GiB)": 111.15, + "step": 15900, + "train_speed(iter/s)": 0.449739 + }, + { + "epoch": 0.4033485540334855, + "eval_acc": 0.6398536644668432, + "eval_loss": 1.6273599863052368, + "eval_runtime": 112.5508, + "eval_samples_per_second": 56.597, + "eval_steps_per_second": 28.298, + "step": 15900 + }, + { + "acc": 0.63903146, + "epoch": 0.4034753932014206, + "grad_norm": 6.8125, + "learning_rate": 9.383621169071971e-06, + "loss": 1.61832771, + "memory(GiB)": 111.15, + "step": 15905, + "train_speed(iter/s)": 0.44833 + }, + { + "acc": 0.66349039, + "epoch": 0.40360223236935566, + "grad_norm": 5.0, + "learning_rate": 9.38311669163715e-06, + "loss": 1.64513531, + "memory(GiB)": 111.15, + "step": 15910, + "train_speed(iter/s)": 0.448403 + }, + { + "acc": 0.65134974, + "epoch": 0.4037290715372907, + "grad_norm": 6.59375, + "learning_rate": 9.38261202141329e-06, + "loss": 1.62714615, + "memory(GiB)": 111.15, + "step": 15915, + "train_speed(iter/s)": 0.448477 + }, + { + "acc": 0.64628363, + "epoch": 0.40385591070522575, + "grad_norm": 6.4375, + "learning_rate": 9.382107158422585e-06, + "loss": 1.64891701, + "memory(GiB)": 111.15, + "step": 15920, + "train_speed(iter/s)": 0.44855 + }, + { + "acc": 0.64765997, + "epoch": 0.40398274987316085, + "grad_norm": 5.53125, + "learning_rate": 9.381602102687241e-06, + "loss": 1.61613941, + "memory(GiB)": 111.15, + "step": 15925, + "train_speed(iter/s)": 0.448625 + }, + { + "acc": 0.65472775, + "epoch": 0.4041095890410959, + "grad_norm": 5.21875, + "learning_rate": 9.381096854229476e-06, + "loss": 1.63174019, + "memory(GiB)": 111.15, + "step": 15930, + "train_speed(iter/s)": 0.448699 + }, + { + "acc": 0.65799866, + "epoch": 0.40423642820903094, + "grad_norm": 4.90625, + "learning_rate": 9.38059141307151e-06, + "loss": 1.63090477, + "memory(GiB)": 111.15, + "step": 15935, + "train_speed(iter/s)": 0.448773 + }, + { + "acc": 0.64760499, + "epoch": 0.404363267376966, + "grad_norm": 4.96875, + "learning_rate": 9.380085779235577e-06, + "loss": 1.62491283, + "memory(GiB)": 111.15, + "step": 15940, + "train_speed(iter/s)": 0.448848 + }, + { + "acc": 0.64343209, + "epoch": 0.4044901065449011, + "grad_norm": 4.53125, + "learning_rate": 9.379579952743916e-06, + "loss": 1.64609375, + "memory(GiB)": 111.15, + "step": 15945, + "train_speed(iter/s)": 0.448924 + }, + { + "acc": 0.6467103, + "epoch": 0.40461694571283613, + "grad_norm": 5.6875, + "learning_rate": 9.379073933618774e-06, + "loss": 1.63088894, + "memory(GiB)": 111.15, + "step": 15950, + "train_speed(iter/s)": 0.448999 + }, + { + "acc": 0.63932209, + "epoch": 0.4047437848807712, + "grad_norm": 4.90625, + "learning_rate": 9.37856772188241e-06, + "loss": 1.67060413, + "memory(GiB)": 111.15, + "step": 15955, + "train_speed(iter/s)": 0.449075 + }, + { + "acc": 0.6417901, + "epoch": 0.4048706240487062, + "grad_norm": 5.09375, + "learning_rate": 9.378061317557088e-06, + "loss": 1.61358833, + "memory(GiB)": 111.15, + "step": 15960, + "train_speed(iter/s)": 0.44915 + }, + { + "acc": 0.66491165, + "epoch": 0.4049974632166413, + "grad_norm": 5.6875, + "learning_rate": 9.377554720665083e-06, + "loss": 1.62786751, + "memory(GiB)": 111.15, + "step": 15965, + "train_speed(iter/s)": 0.449226 + }, + { + "acc": 0.64388933, + "epoch": 0.40512430238457636, + "grad_norm": 4.875, + "learning_rate": 9.377047931228677e-06, + "loss": 1.70449886, + "memory(GiB)": 111.15, + "step": 15970, + "train_speed(iter/s)": 0.449301 + }, + { + "acc": 0.65411944, + "epoch": 0.4052511415525114, + "grad_norm": 6.3125, + "learning_rate": 9.376540949270161e-06, + "loss": 1.60068722, + "memory(GiB)": 111.15, + "step": 15975, + "train_speed(iter/s)": 0.449378 + }, + { + "acc": 0.65048256, + "epoch": 0.40537798072044645, + "grad_norm": 5.03125, + "learning_rate": 9.376033774811833e-06, + "loss": 1.60749359, + "memory(GiB)": 111.15, + "step": 15980, + "train_speed(iter/s)": 0.449453 + }, + { + "acc": 0.65982952, + "epoch": 0.40550481988838155, + "grad_norm": 5.6875, + "learning_rate": 9.375526407876003e-06, + "loss": 1.6095562, + "memory(GiB)": 111.15, + "step": 15985, + "train_speed(iter/s)": 0.449529 + }, + { + "acc": 0.63645515, + "epoch": 0.4056316590563166, + "grad_norm": 5.40625, + "learning_rate": 9.375018848484987e-06, + "loss": 1.7021246, + "memory(GiB)": 111.15, + "step": 15990, + "train_speed(iter/s)": 0.449605 + }, + { + "acc": 0.65723057, + "epoch": 0.40575849822425164, + "grad_norm": 4.65625, + "learning_rate": 9.374511096661108e-06, + "loss": 1.63122025, + "memory(GiB)": 111.15, + "step": 15995, + "train_speed(iter/s)": 0.44968 + }, + { + "acc": 0.64214497, + "epoch": 0.4058853373921867, + "grad_norm": 4.65625, + "learning_rate": 9.374003152426701e-06, + "loss": 1.66204491, + "memory(GiB)": 111.15, + "step": 16000, + "train_speed(iter/s)": 0.449755 + }, + { + "epoch": 0.4058853373921867, + "eval_acc": 0.6399221783055943, + "eval_loss": 1.6269961595535278, + "eval_runtime": 114.1445, + "eval_samples_per_second": 55.806, + "eval_steps_per_second": 27.903, + "step": 16000 + }, + { + "acc": 0.65143061, + "epoch": 0.4060121765601218, + "grad_norm": 5.15625, + "learning_rate": 9.373495015804106e-06, + "loss": 1.63724384, + "memory(GiB)": 111.15, + "step": 16005, + "train_speed(iter/s)": 0.448335 + }, + { + "acc": 0.64188366, + "epoch": 0.40613901572805683, + "grad_norm": 5.46875, + "learning_rate": 9.372986686815674e-06, + "loss": 1.70729752, + "memory(GiB)": 111.15, + "step": 16010, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.65008759, + "epoch": 0.4062658548959919, + "grad_norm": 5.75, + "learning_rate": 9.372478165483763e-06, + "loss": 1.60834465, + "memory(GiB)": 111.15, + "step": 16015, + "train_speed(iter/s)": 0.448486 + }, + { + "acc": 0.65186887, + "epoch": 0.4063926940639269, + "grad_norm": 5.25, + "learning_rate": 9.371969451830743e-06, + "loss": 1.68279152, + "memory(GiB)": 111.15, + "step": 16020, + "train_speed(iter/s)": 0.44856 + }, + { + "acc": 0.65010719, + "epoch": 0.406519533231862, + "grad_norm": 4.71875, + "learning_rate": 9.371460545878986e-06, + "loss": 1.62294769, + "memory(GiB)": 111.15, + "step": 16025, + "train_speed(iter/s)": 0.448635 + }, + { + "acc": 0.64782495, + "epoch": 0.40664637239979706, + "grad_norm": 5.5, + "learning_rate": 9.370951447650875e-06, + "loss": 1.61834831, + "memory(GiB)": 111.15, + "step": 16030, + "train_speed(iter/s)": 0.44871 + }, + { + "acc": 0.62972898, + "epoch": 0.4067732115677321, + "grad_norm": 5.125, + "learning_rate": 9.370442157168806e-06, + "loss": 1.70677662, + "memory(GiB)": 111.15, + "step": 16035, + "train_speed(iter/s)": 0.448785 + }, + { + "acc": 0.63901262, + "epoch": 0.40690005073566715, + "grad_norm": 4.78125, + "learning_rate": 9.369932674455177e-06, + "loss": 1.68983765, + "memory(GiB)": 111.15, + "step": 16040, + "train_speed(iter/s)": 0.448861 + }, + { + "acc": 0.6572814, + "epoch": 0.40702688990360225, + "grad_norm": 4.9375, + "learning_rate": 9.3694229995324e-06, + "loss": 1.57476215, + "memory(GiB)": 111.15, + "step": 16045, + "train_speed(iter/s)": 0.448935 + }, + { + "acc": 0.65458002, + "epoch": 0.4071537290715373, + "grad_norm": 5.65625, + "learning_rate": 9.368913132422891e-06, + "loss": 1.62649117, + "memory(GiB)": 111.15, + "step": 16050, + "train_speed(iter/s)": 0.449011 + }, + { + "acc": 0.64681015, + "epoch": 0.40728056823947234, + "grad_norm": 5.6875, + "learning_rate": 9.368403073149079e-06, + "loss": 1.66858234, + "memory(GiB)": 111.15, + "step": 16055, + "train_speed(iter/s)": 0.449087 + }, + { + "acc": 0.63391953, + "epoch": 0.4074074074074074, + "grad_norm": 6.34375, + "learning_rate": 9.367892821733393e-06, + "loss": 1.63374023, + "memory(GiB)": 111.15, + "step": 16060, + "train_speed(iter/s)": 0.449162 + }, + { + "acc": 0.64759874, + "epoch": 0.4075342465753425, + "grad_norm": 5.28125, + "learning_rate": 9.367382378198282e-06, + "loss": 1.63470249, + "memory(GiB)": 111.15, + "step": 16065, + "train_speed(iter/s)": 0.449237 + }, + { + "acc": 0.6456079, + "epoch": 0.40766108574327753, + "grad_norm": 6.5625, + "learning_rate": 9.366871742566193e-06, + "loss": 1.62195797, + "memory(GiB)": 111.15, + "step": 16070, + "train_speed(iter/s)": 0.449312 + }, + { + "acc": 0.63188477, + "epoch": 0.4077879249112126, + "grad_norm": 5.375, + "learning_rate": 9.36636091485959e-06, + "loss": 1.67230949, + "memory(GiB)": 111.15, + "step": 16075, + "train_speed(iter/s)": 0.449387 + }, + { + "acc": 0.65013676, + "epoch": 0.4079147640791476, + "grad_norm": 5.28125, + "learning_rate": 9.365849895100939e-06, + "loss": 1.59211006, + "memory(GiB)": 111.15, + "step": 16080, + "train_speed(iter/s)": 0.449461 + }, + { + "acc": 0.65820122, + "epoch": 0.4080416032470827, + "grad_norm": 7.0, + "learning_rate": 9.365338683312714e-06, + "loss": 1.59948635, + "memory(GiB)": 111.15, + "step": 16085, + "train_speed(iter/s)": 0.449536 + }, + { + "acc": 0.64477949, + "epoch": 0.40816844241501776, + "grad_norm": 4.90625, + "learning_rate": 9.364827279517408e-06, + "loss": 1.64739456, + "memory(GiB)": 111.15, + "step": 16090, + "train_speed(iter/s)": 0.44961 + }, + { + "acc": 0.63604441, + "epoch": 0.4082952815829528, + "grad_norm": 4.78125, + "learning_rate": 9.36431568373751e-06, + "loss": 1.68808937, + "memory(GiB)": 111.15, + "step": 16095, + "train_speed(iter/s)": 0.449684 + }, + { + "acc": 0.6508811, + "epoch": 0.40842212075088785, + "grad_norm": 5.1875, + "learning_rate": 9.363803895995522e-06, + "loss": 1.57515163, + "memory(GiB)": 111.15, + "step": 16100, + "train_speed(iter/s)": 0.449759 + }, + { + "epoch": 0.40842212075088785, + "eval_acc": 0.639984007867394, + "eval_loss": 1.6268733739852905, + "eval_runtime": 113.922, + "eval_samples_per_second": 55.915, + "eval_steps_per_second": 27.958, + "step": 16100 + }, + { + "acc": 0.64717054, + "epoch": 0.40854895991882295, + "grad_norm": 5.4375, + "learning_rate": 9.363291916313955e-06, + "loss": 1.61785889, + "memory(GiB)": 111.15, + "step": 16105, + "train_speed(iter/s)": 0.448351 + }, + { + "acc": 0.64519033, + "epoch": 0.408675799086758, + "grad_norm": 4.6875, + "learning_rate": 9.362779744715332e-06, + "loss": 1.65577869, + "memory(GiB)": 111.15, + "step": 16110, + "train_speed(iter/s)": 0.448426 + }, + { + "acc": 0.64788399, + "epoch": 0.40880263825469304, + "grad_norm": 5.65625, + "learning_rate": 9.362267381222174e-06, + "loss": 1.68617802, + "memory(GiB)": 111.15, + "step": 16115, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.65298834, + "epoch": 0.4089294774226281, + "grad_norm": 5.03125, + "learning_rate": 9.361754825857022e-06, + "loss": 1.63152237, + "memory(GiB)": 111.15, + "step": 16120, + "train_speed(iter/s)": 0.448576 + }, + { + "acc": 0.64395914, + "epoch": 0.4090563165905632, + "grad_norm": 5.40625, + "learning_rate": 9.36124207864242e-06, + "loss": 1.66268959, + "memory(GiB)": 111.15, + "step": 16125, + "train_speed(iter/s)": 0.448652 + }, + { + "acc": 0.63845472, + "epoch": 0.40918315575849823, + "grad_norm": 5.34375, + "learning_rate": 9.360729139600917e-06, + "loss": 1.65722351, + "memory(GiB)": 111.15, + "step": 16130, + "train_speed(iter/s)": 0.448728 + }, + { + "acc": 0.64443421, + "epoch": 0.4093099949264333, + "grad_norm": 5.125, + "learning_rate": 9.36021600875508e-06, + "loss": 1.6579586, + "memory(GiB)": 111.15, + "step": 16135, + "train_speed(iter/s)": 0.448804 + }, + { + "acc": 0.63106542, + "epoch": 0.4094368340943683, + "grad_norm": 5.0, + "learning_rate": 9.359702686127474e-06, + "loss": 1.71448345, + "memory(GiB)": 111.15, + "step": 16140, + "train_speed(iter/s)": 0.448879 + }, + { + "acc": 0.65172763, + "epoch": 0.4095636732623034, + "grad_norm": 5.125, + "learning_rate": 9.359189171740679e-06, + "loss": 1.65969353, + "memory(GiB)": 111.15, + "step": 16145, + "train_speed(iter/s)": 0.448954 + }, + { + "acc": 0.63784628, + "epoch": 0.40969051243023846, + "grad_norm": 5.09375, + "learning_rate": 9.358675465617283e-06, + "loss": 1.68275242, + "memory(GiB)": 111.15, + "step": 16150, + "train_speed(iter/s)": 0.44903 + }, + { + "acc": 0.6447082, + "epoch": 0.4098173515981735, + "grad_norm": 5.46875, + "learning_rate": 9.35816156777988e-06, + "loss": 1.66857929, + "memory(GiB)": 111.15, + "step": 16155, + "train_speed(iter/s)": 0.449105 + }, + { + "acc": 0.63655686, + "epoch": 0.40994419076610855, + "grad_norm": 6.46875, + "learning_rate": 9.357647478251072e-06, + "loss": 1.65050316, + "memory(GiB)": 111.15, + "step": 16160, + "train_speed(iter/s)": 0.44918 + }, + { + "acc": 0.64066706, + "epoch": 0.41007102993404365, + "grad_norm": 5.78125, + "learning_rate": 9.357133197053475e-06, + "loss": 1.70095825, + "memory(GiB)": 111.15, + "step": 16165, + "train_speed(iter/s)": 0.449255 + }, + { + "acc": 0.6547822, + "epoch": 0.4101978691019787, + "grad_norm": 6.1875, + "learning_rate": 9.356618724209704e-06, + "loss": 1.62589912, + "memory(GiB)": 111.15, + "step": 16170, + "train_speed(iter/s)": 0.44933 + }, + { + "acc": 0.65179415, + "epoch": 0.41032470826991374, + "grad_norm": 5.15625, + "learning_rate": 9.356104059742392e-06, + "loss": 1.60424728, + "memory(GiB)": 111.15, + "step": 16175, + "train_speed(iter/s)": 0.449405 + }, + { + "acc": 0.64641323, + "epoch": 0.4104515474378488, + "grad_norm": 4.875, + "learning_rate": 9.355589203674175e-06, + "loss": 1.63600464, + "memory(GiB)": 111.15, + "step": 16180, + "train_speed(iter/s)": 0.44948 + }, + { + "acc": 0.65505495, + "epoch": 0.4105783866057839, + "grad_norm": 6.03125, + "learning_rate": 9.355074156027699e-06, + "loss": 1.61012135, + "memory(GiB)": 111.15, + "step": 16185, + "train_speed(iter/s)": 0.449555 + }, + { + "acc": 0.64186916, + "epoch": 0.41070522577371893, + "grad_norm": 5.03125, + "learning_rate": 9.354558916825616e-06, + "loss": 1.69574356, + "memory(GiB)": 111.15, + "step": 16190, + "train_speed(iter/s)": 0.44963 + }, + { + "acc": 0.65749578, + "epoch": 0.410832064941654, + "grad_norm": 4.78125, + "learning_rate": 9.354043486090592e-06, + "loss": 1.61712971, + "memory(GiB)": 111.15, + "step": 16195, + "train_speed(iter/s)": 0.449705 + }, + { + "acc": 0.6378521, + "epoch": 0.410958904109589, + "grad_norm": 5.125, + "learning_rate": 9.353527863845296e-06, + "loss": 1.67580051, + "memory(GiB)": 111.15, + "step": 16200, + "train_speed(iter/s)": 0.44978 + }, + { + "epoch": 0.410958904109589, + "eval_acc": 0.6400266201329586, + "eval_loss": 1.6268789768218994, + "eval_runtime": 115.7535, + "eval_samples_per_second": 55.031, + "eval_steps_per_second": 27.515, + "step": 16200 + }, + { + "acc": 0.6622241, + "epoch": 0.4110857432775241, + "grad_norm": 4.78125, + "learning_rate": 9.353012050112405e-06, + "loss": 1.60279598, + "memory(GiB)": 111.15, + "step": 16205, + "train_speed(iter/s)": 0.448356 + }, + { + "acc": 0.6468565, + "epoch": 0.41121258244545916, + "grad_norm": 5.125, + "learning_rate": 9.352496044914611e-06, + "loss": 1.63705902, + "memory(GiB)": 111.15, + "step": 16210, + "train_speed(iter/s)": 0.44843 + }, + { + "acc": 0.65564222, + "epoch": 0.4113394216133942, + "grad_norm": 6.46875, + "learning_rate": 9.351979848274608e-06, + "loss": 1.61295776, + "memory(GiB)": 111.15, + "step": 16215, + "train_speed(iter/s)": 0.448503 + }, + { + "acc": 0.66352382, + "epoch": 0.41146626078132925, + "grad_norm": 4.84375, + "learning_rate": 9.351463460215102e-06, + "loss": 1.55455265, + "memory(GiB)": 111.15, + "step": 16220, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.65502863, + "epoch": 0.41159309994926435, + "grad_norm": 6.25, + "learning_rate": 9.350946880758804e-06, + "loss": 1.62152157, + "memory(GiB)": 111.15, + "step": 16225, + "train_speed(iter/s)": 0.448652 + }, + { + "acc": 0.65121527, + "epoch": 0.4117199391171994, + "grad_norm": 5.21875, + "learning_rate": 9.350430109928437e-06, + "loss": 1.63010521, + "memory(GiB)": 111.15, + "step": 16230, + "train_speed(iter/s)": 0.448725 + }, + { + "acc": 0.65481725, + "epoch": 0.41184677828513444, + "grad_norm": 5.96875, + "learning_rate": 9.349913147746731e-06, + "loss": 1.70053864, + "memory(GiB)": 111.15, + "step": 16235, + "train_speed(iter/s)": 0.448799 + }, + { + "acc": 0.65579934, + "epoch": 0.4119736174530695, + "grad_norm": 4.8125, + "learning_rate": 9.349395994236423e-06, + "loss": 1.5883482, + "memory(GiB)": 111.15, + "step": 16240, + "train_speed(iter/s)": 0.448873 + }, + { + "acc": 0.64541454, + "epoch": 0.4121004566210046, + "grad_norm": 5.3125, + "learning_rate": 9.348878649420262e-06, + "loss": 1.69941978, + "memory(GiB)": 111.15, + "step": 16245, + "train_speed(iter/s)": 0.448946 + }, + { + "acc": 0.63718967, + "epoch": 0.41222729578893963, + "grad_norm": 5.90625, + "learning_rate": 9.348361113321e-06, + "loss": 1.66328201, + "memory(GiB)": 111.15, + "step": 16250, + "train_speed(iter/s)": 0.44902 + }, + { + "acc": 0.6497962, + "epoch": 0.4123541349568747, + "grad_norm": 5.28125, + "learning_rate": 9.347843385961403e-06, + "loss": 1.65080185, + "memory(GiB)": 111.15, + "step": 16255, + "train_speed(iter/s)": 0.449093 + }, + { + "acc": 0.62718048, + "epoch": 0.4124809741248097, + "grad_norm": 5.1875, + "learning_rate": 9.347325467364242e-06, + "loss": 1.65719357, + "memory(GiB)": 111.15, + "step": 16260, + "train_speed(iter/s)": 0.449167 + }, + { + "acc": 0.65719485, + "epoch": 0.4126078132927448, + "grad_norm": 4.96875, + "learning_rate": 9.346807357552296e-06, + "loss": 1.63608742, + "memory(GiB)": 111.15, + "step": 16265, + "train_speed(iter/s)": 0.44924 + }, + { + "acc": 0.64354587, + "epoch": 0.41273465246067986, + "grad_norm": 4.78125, + "learning_rate": 9.346289056548357e-06, + "loss": 1.64914703, + "memory(GiB)": 111.15, + "step": 16270, + "train_speed(iter/s)": 0.449314 + }, + { + "acc": 0.66295147, + "epoch": 0.4128614916286149, + "grad_norm": 4.65625, + "learning_rate": 9.345770564375221e-06, + "loss": 1.58379822, + "memory(GiB)": 111.15, + "step": 16275, + "train_speed(iter/s)": 0.449387 + }, + { + "acc": 0.6384738, + "epoch": 0.41298833079654995, + "grad_norm": 4.6875, + "learning_rate": 9.345251881055692e-06, + "loss": 1.6455368, + "memory(GiB)": 111.15, + "step": 16280, + "train_speed(iter/s)": 0.44946 + }, + { + "acc": 0.64921432, + "epoch": 0.41311516996448505, + "grad_norm": 5.96875, + "learning_rate": 9.344733006612585e-06, + "loss": 1.68562775, + "memory(GiB)": 111.15, + "step": 16285, + "train_speed(iter/s)": 0.449534 + }, + { + "acc": 0.65300412, + "epoch": 0.4132420091324201, + "grad_norm": 5.40625, + "learning_rate": 9.344213941068724e-06, + "loss": 1.64478416, + "memory(GiB)": 111.15, + "step": 16290, + "train_speed(iter/s)": 0.449607 + }, + { + "acc": 0.64070544, + "epoch": 0.41336884830035514, + "grad_norm": 5.71875, + "learning_rate": 9.343694684446937e-06, + "loss": 1.6733345, + "memory(GiB)": 111.15, + "step": 16295, + "train_speed(iter/s)": 0.44968 + }, + { + "acc": 0.64769931, + "epoch": 0.4134956874682902, + "grad_norm": 5.03125, + "learning_rate": 9.343175236770065e-06, + "loss": 1.58220949, + "memory(GiB)": 111.15, + "step": 16300, + "train_speed(iter/s)": 0.449754 + }, + { + "epoch": 0.4134956874682902, + "eval_acc": 0.6400491795676693, + "eval_loss": 1.6263422966003418, + "eval_runtime": 115.5216, + "eval_samples_per_second": 55.141, + "eval_steps_per_second": 27.571, + "step": 16300 + }, + { + "acc": 0.65734739, + "epoch": 0.4136225266362253, + "grad_norm": 4.3125, + "learning_rate": 9.342655598060955e-06, + "loss": 1.60211163, + "memory(GiB)": 111.15, + "step": 16305, + "train_speed(iter/s)": 0.448343 + }, + { + "acc": 0.65160551, + "epoch": 0.41374936580416033, + "grad_norm": 4.5, + "learning_rate": 9.342135768342464e-06, + "loss": 1.59915905, + "memory(GiB)": 111.15, + "step": 16310, + "train_speed(iter/s)": 0.448416 + }, + { + "acc": 0.65592852, + "epoch": 0.4138762049720954, + "grad_norm": 5.1875, + "learning_rate": 9.341615747637454e-06, + "loss": 1.61348782, + "memory(GiB)": 111.15, + "step": 16315, + "train_speed(iter/s)": 0.44849 + }, + { + "acc": 0.63649092, + "epoch": 0.4140030441400304, + "grad_norm": 4.875, + "learning_rate": 9.3410955359688e-06, + "loss": 1.69064159, + "memory(GiB)": 111.15, + "step": 16320, + "train_speed(iter/s)": 0.448564 + }, + { + "acc": 0.65566339, + "epoch": 0.4141298833079655, + "grad_norm": 5.78125, + "learning_rate": 9.340575133359385e-06, + "loss": 1.64006691, + "memory(GiB)": 111.15, + "step": 16325, + "train_speed(iter/s)": 0.448637 + }, + { + "acc": 0.6503931, + "epoch": 0.41425672247590056, + "grad_norm": 5.25, + "learning_rate": 9.340054539832095e-06, + "loss": 1.59376688, + "memory(GiB)": 111.15, + "step": 16330, + "train_speed(iter/s)": 0.448711 + }, + { + "acc": 0.63787389, + "epoch": 0.4143835616438356, + "grad_norm": 6.34375, + "learning_rate": 9.339533755409828e-06, + "loss": 1.71170235, + "memory(GiB)": 111.15, + "step": 16335, + "train_speed(iter/s)": 0.448785 + }, + { + "acc": 0.64137039, + "epoch": 0.41451040081177065, + "grad_norm": 5.03125, + "learning_rate": 9.339012780115492e-06, + "loss": 1.68638439, + "memory(GiB)": 111.15, + "step": 16340, + "train_speed(iter/s)": 0.448859 + }, + { + "acc": 0.64735355, + "epoch": 0.41463723997970575, + "grad_norm": 4.65625, + "learning_rate": 9.338491613972002e-06, + "loss": 1.61712818, + "memory(GiB)": 111.15, + "step": 16345, + "train_speed(iter/s)": 0.448932 + }, + { + "acc": 0.66794863, + "epoch": 0.4147640791476408, + "grad_norm": 5.5625, + "learning_rate": 9.337970257002282e-06, + "loss": 1.64576454, + "memory(GiB)": 111.15, + "step": 16350, + "train_speed(iter/s)": 0.449005 + }, + { + "acc": 0.64591651, + "epoch": 0.41489091831557584, + "grad_norm": 6.34375, + "learning_rate": 9.337448709229261e-06, + "loss": 1.72351665, + "memory(GiB)": 111.15, + "step": 16355, + "train_speed(iter/s)": 0.449078 + }, + { + "acc": 0.66100492, + "epoch": 0.4150177574835109, + "grad_norm": 5.5625, + "learning_rate": 9.336926970675883e-06, + "loss": 1.63590851, + "memory(GiB)": 111.15, + "step": 16360, + "train_speed(iter/s)": 0.449152 + }, + { + "acc": 0.62602663, + "epoch": 0.415144596651446, + "grad_norm": 5.125, + "learning_rate": 9.33640504136509e-06, + "loss": 1.72774925, + "memory(GiB)": 111.15, + "step": 16365, + "train_speed(iter/s)": 0.449225 + }, + { + "acc": 0.64850082, + "epoch": 0.41527143581938103, + "grad_norm": 5.3125, + "learning_rate": 9.335882921319845e-06, + "loss": 1.62144756, + "memory(GiB)": 111.15, + "step": 16370, + "train_speed(iter/s)": 0.449299 + }, + { + "acc": 0.66400299, + "epoch": 0.4153982749873161, + "grad_norm": 4.875, + "learning_rate": 9.335360610563111e-06, + "loss": 1.61179256, + "memory(GiB)": 111.15, + "step": 16375, + "train_speed(iter/s)": 0.449373 + }, + { + "acc": 0.6542871, + "epoch": 0.4155251141552511, + "grad_norm": 5.25, + "learning_rate": 9.33483810911786e-06, + "loss": 1.63856068, + "memory(GiB)": 111.15, + "step": 16380, + "train_speed(iter/s)": 0.449446 + }, + { + "acc": 0.63096371, + "epoch": 0.4156519533231862, + "grad_norm": 5.5625, + "learning_rate": 9.334315417007079e-06, + "loss": 1.71098061, + "memory(GiB)": 111.15, + "step": 16385, + "train_speed(iter/s)": 0.449519 + }, + { + "acc": 0.63856692, + "epoch": 0.41577879249112126, + "grad_norm": 4.6875, + "learning_rate": 9.333792534253751e-06, + "loss": 1.66098347, + "memory(GiB)": 111.15, + "step": 16390, + "train_speed(iter/s)": 0.449592 + }, + { + "acc": 0.63820782, + "epoch": 0.4159056316590563, + "grad_norm": 4.96875, + "learning_rate": 9.333269460880879e-06, + "loss": 1.63703308, + "memory(GiB)": 111.15, + "step": 16395, + "train_speed(iter/s)": 0.449665 + }, + { + "acc": 0.65623446, + "epoch": 0.41603247082699135, + "grad_norm": 5.84375, + "learning_rate": 9.33274619691147e-06, + "loss": 1.56448746, + "memory(GiB)": 111.15, + "step": 16400, + "train_speed(iter/s)": 0.449738 + }, + { + "epoch": 0.41603247082699135, + "eval_acc": 0.6400441663599559, + "eval_loss": 1.626044750213623, + "eval_runtime": 113.216, + "eval_samples_per_second": 56.264, + "eval_steps_per_second": 28.132, + "step": 16400 + }, + { + "acc": 0.64643679, + "epoch": 0.41615930999492645, + "grad_norm": 5.71875, + "learning_rate": 9.332222742368537e-06, + "loss": 1.62306137, + "memory(GiB)": 111.15, + "step": 16405, + "train_speed(iter/s)": 0.448363 + }, + { + "acc": 0.63408146, + "epoch": 0.4162861491628615, + "grad_norm": 4.75, + "learning_rate": 9.331699097275108e-06, + "loss": 1.67109051, + "memory(GiB)": 111.15, + "step": 16410, + "train_speed(iter/s)": 0.448436 + }, + { + "acc": 0.65849137, + "epoch": 0.41641298833079654, + "grad_norm": 8.0, + "learning_rate": 9.331175261654213e-06, + "loss": 1.65627403, + "memory(GiB)": 111.15, + "step": 16415, + "train_speed(iter/s)": 0.448508 + }, + { + "acc": 0.64748726, + "epoch": 0.4165398274987316, + "grad_norm": 5.1875, + "learning_rate": 9.330651235528891e-06, + "loss": 1.69113121, + "memory(GiB)": 111.15, + "step": 16420, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.6414084, + "epoch": 0.4166666666666667, + "grad_norm": 5.09375, + "learning_rate": 9.330127018922195e-06, + "loss": 1.71781502, + "memory(GiB)": 111.15, + "step": 16425, + "train_speed(iter/s)": 0.448652 + }, + { + "acc": 0.66547098, + "epoch": 0.41679350583460173, + "grad_norm": 5.0, + "learning_rate": 9.329602611857179e-06, + "loss": 1.57332153, + "memory(GiB)": 111.15, + "step": 16430, + "train_speed(iter/s)": 0.448724 + }, + { + "acc": 0.64482942, + "epoch": 0.4169203450025368, + "grad_norm": 5.34375, + "learning_rate": 9.329078014356909e-06, + "loss": 1.61002789, + "memory(GiB)": 111.15, + "step": 16435, + "train_speed(iter/s)": 0.448797 + }, + { + "acc": 0.66598668, + "epoch": 0.4170471841704718, + "grad_norm": 5.65625, + "learning_rate": 9.32855322644446e-06, + "loss": 1.56462545, + "memory(GiB)": 111.15, + "step": 16440, + "train_speed(iter/s)": 0.448869 + }, + { + "acc": 0.65097532, + "epoch": 0.4171740233384069, + "grad_norm": 5.46875, + "learning_rate": 9.328028248142916e-06, + "loss": 1.58615246, + "memory(GiB)": 111.15, + "step": 16445, + "train_speed(iter/s)": 0.44894 + }, + { + "acc": 0.63982849, + "epoch": 0.41730086250634196, + "grad_norm": 4.4375, + "learning_rate": 9.327503079475365e-06, + "loss": 1.69326782, + "memory(GiB)": 111.15, + "step": 16450, + "train_speed(iter/s)": 0.449013 + }, + { + "acc": 0.63811693, + "epoch": 0.417427701674277, + "grad_norm": 5.78125, + "learning_rate": 9.326977720464908e-06, + "loss": 1.65198174, + "memory(GiB)": 111.15, + "step": 16455, + "train_speed(iter/s)": 0.449085 + }, + { + "acc": 0.65315723, + "epoch": 0.41755454084221205, + "grad_norm": 5.96875, + "learning_rate": 9.326452171134652e-06, + "loss": 1.63365364, + "memory(GiB)": 111.15, + "step": 16460, + "train_speed(iter/s)": 0.449157 + }, + { + "acc": 0.63213615, + "epoch": 0.41768138001014715, + "grad_norm": 5.34375, + "learning_rate": 9.325926431507714e-06, + "loss": 1.67402992, + "memory(GiB)": 111.15, + "step": 16465, + "train_speed(iter/s)": 0.449228 + }, + { + "acc": 0.64885883, + "epoch": 0.4178082191780822, + "grad_norm": 5.53125, + "learning_rate": 9.325400501607218e-06, + "loss": 1.62513809, + "memory(GiB)": 111.15, + "step": 16470, + "train_speed(iter/s)": 0.449301 + }, + { + "acc": 0.63371105, + "epoch": 0.41793505834601724, + "grad_norm": 5.59375, + "learning_rate": 9.324874381456295e-06, + "loss": 1.69248905, + "memory(GiB)": 111.15, + "step": 16475, + "train_speed(iter/s)": 0.449373 + }, + { + "acc": 0.65466413, + "epoch": 0.4180618975139523, + "grad_norm": 5.90625, + "learning_rate": 9.324348071078088e-06, + "loss": 1.61338005, + "memory(GiB)": 111.15, + "step": 16480, + "train_speed(iter/s)": 0.449444 + }, + { + "acc": 0.63324971, + "epoch": 0.4181887366818874, + "grad_norm": 5.0, + "learning_rate": 9.323821570495748e-06, + "loss": 1.68169384, + "memory(GiB)": 111.15, + "step": 16485, + "train_speed(iter/s)": 0.449515 + }, + { + "acc": 0.63939524, + "epoch": 0.41831557584982243, + "grad_norm": 5.40625, + "learning_rate": 9.32329487973243e-06, + "loss": 1.65264244, + "memory(GiB)": 111.15, + "step": 16490, + "train_speed(iter/s)": 0.449587 + }, + { + "acc": 0.65641127, + "epoch": 0.4184424150177575, + "grad_norm": 5.3125, + "learning_rate": 9.3227679988113e-06, + "loss": 1.61379433, + "memory(GiB)": 111.15, + "step": 16495, + "train_speed(iter/s)": 0.449658 + }, + { + "acc": 0.65871277, + "epoch": 0.4185692541856925, + "grad_norm": 6.0, + "learning_rate": 9.322240927755534e-06, + "loss": 1.60743122, + "memory(GiB)": 111.15, + "step": 16500, + "train_speed(iter/s)": 0.44973 + }, + { + "epoch": 0.4185692541856925, + "eval_acc": 0.6401247954506811, + "eval_loss": 1.6261812448501587, + "eval_runtime": 115.5391, + "eval_samples_per_second": 55.133, + "eval_steps_per_second": 27.566, + "step": 16500 + }, + { + "acc": 0.64120617, + "epoch": 0.4186960933536276, + "grad_norm": 6.125, + "learning_rate": 9.321713666588314e-06, + "loss": 1.64479294, + "memory(GiB)": 111.15, + "step": 16505, + "train_speed(iter/s)": 0.448335 + }, + { + "acc": 0.6487402, + "epoch": 0.41882293252156266, + "grad_norm": 4.96875, + "learning_rate": 9.321186215332833e-06, + "loss": 1.69038124, + "memory(GiB)": 111.15, + "step": 16510, + "train_speed(iter/s)": 0.448406 + }, + { + "acc": 0.65008621, + "epoch": 0.4189497716894977, + "grad_norm": 4.3125, + "learning_rate": 9.320658574012289e-06, + "loss": 1.67295723, + "memory(GiB)": 111.15, + "step": 16515, + "train_speed(iter/s)": 0.448478 + }, + { + "acc": 0.65017114, + "epoch": 0.41907661085743275, + "grad_norm": 5.25, + "learning_rate": 9.32013074264989e-06, + "loss": 1.63766842, + "memory(GiB)": 111.15, + "step": 16520, + "train_speed(iter/s)": 0.448549 + }, + { + "acc": 0.65635872, + "epoch": 0.41920345002536785, + "grad_norm": 5.9375, + "learning_rate": 9.319602721268853e-06, + "loss": 1.62925167, + "memory(GiB)": 111.15, + "step": 16525, + "train_speed(iter/s)": 0.44862 + }, + { + "acc": 0.6745924, + "epoch": 0.4193302891933029, + "grad_norm": 4.84375, + "learning_rate": 9.319074509892403e-06, + "loss": 1.56067848, + "memory(GiB)": 111.15, + "step": 16530, + "train_speed(iter/s)": 0.448691 + }, + { + "acc": 0.6559587, + "epoch": 0.41945712836123794, + "grad_norm": 4.65625, + "learning_rate": 9.318546108543774e-06, + "loss": 1.56671219, + "memory(GiB)": 111.15, + "step": 16535, + "train_speed(iter/s)": 0.448762 + }, + { + "acc": 0.62391901, + "epoch": 0.419583967529173, + "grad_norm": 4.875, + "learning_rate": 9.318017517246205e-06, + "loss": 1.70603447, + "memory(GiB)": 111.15, + "step": 16540, + "train_speed(iter/s)": 0.448834 + }, + { + "acc": 0.62648287, + "epoch": 0.4197108066971081, + "grad_norm": 5.25, + "learning_rate": 9.317488736022948e-06, + "loss": 1.71432819, + "memory(GiB)": 111.15, + "step": 16545, + "train_speed(iter/s)": 0.448904 + }, + { + "acc": 0.64563622, + "epoch": 0.41983764586504313, + "grad_norm": 6.0625, + "learning_rate": 9.316959764897259e-06, + "loss": 1.66305122, + "memory(GiB)": 111.15, + "step": 16550, + "train_speed(iter/s)": 0.448976 + }, + { + "acc": 0.63409739, + "epoch": 0.4199644850329782, + "grad_norm": 5.6875, + "learning_rate": 9.316430603892406e-06, + "loss": 1.63482513, + "memory(GiB)": 111.15, + "step": 16555, + "train_speed(iter/s)": 0.449047 + }, + { + "acc": 0.63901138, + "epoch": 0.4200913242009132, + "grad_norm": 4.8125, + "learning_rate": 9.315901253031663e-06, + "loss": 1.65722389, + "memory(GiB)": 111.15, + "step": 16560, + "train_speed(iter/s)": 0.449118 + }, + { + "acc": 0.63381476, + "epoch": 0.4202181633688483, + "grad_norm": 5.78125, + "learning_rate": 9.315371712338315e-06, + "loss": 1.78488026, + "memory(GiB)": 111.15, + "step": 16565, + "train_speed(iter/s)": 0.449189 + }, + { + "acc": 0.6543273, + "epoch": 0.42034500253678336, + "grad_norm": 5.15625, + "learning_rate": 9.314841981835652e-06, + "loss": 1.61914387, + "memory(GiB)": 111.15, + "step": 16570, + "train_speed(iter/s)": 0.449259 + }, + { + "acc": 0.63609886, + "epoch": 0.4204718417047184, + "grad_norm": 5.59375, + "learning_rate": 9.314312061546974e-06, + "loss": 1.70677261, + "memory(GiB)": 111.15, + "step": 16575, + "train_speed(iter/s)": 0.44933 + }, + { + "acc": 0.64558649, + "epoch": 0.42059868087265345, + "grad_norm": 4.34375, + "learning_rate": 9.313781951495588e-06, + "loss": 1.67297478, + "memory(GiB)": 111.15, + "step": 16580, + "train_speed(iter/s)": 0.449401 + }, + { + "acc": 0.64759083, + "epoch": 0.42072552004058855, + "grad_norm": 5.15625, + "learning_rate": 9.313251651704816e-06, + "loss": 1.71654301, + "memory(GiB)": 111.15, + "step": 16585, + "train_speed(iter/s)": 0.449472 + }, + { + "acc": 0.65371609, + "epoch": 0.4208523592085236, + "grad_norm": 4.71875, + "learning_rate": 9.312721162197975e-06, + "loss": 1.63245049, + "memory(GiB)": 111.15, + "step": 16590, + "train_speed(iter/s)": 0.449543 + }, + { + "acc": 0.65628662, + "epoch": 0.42097919837645864, + "grad_norm": 5.0625, + "learning_rate": 9.312190482998405e-06, + "loss": 1.60666885, + "memory(GiB)": 111.15, + "step": 16595, + "train_speed(iter/s)": 0.449615 + }, + { + "acc": 0.64406486, + "epoch": 0.4211060375443937, + "grad_norm": 5.375, + "learning_rate": 9.311659614129443e-06, + "loss": 1.68860016, + "memory(GiB)": 111.15, + "step": 16600, + "train_speed(iter/s)": 0.449685 + }, + { + "epoch": 0.4211060375443937, + "eval_acc": 0.640297333349487, + "eval_loss": 1.6258081197738647, + "eval_runtime": 115.017, + "eval_samples_per_second": 55.383, + "eval_steps_per_second": 27.692, + "step": 16600 + }, + { + "acc": 0.64297071, + "epoch": 0.4212328767123288, + "grad_norm": 5.625, + "learning_rate": 9.311128555614443e-06, + "loss": 1.69722977, + "memory(GiB)": 111.15, + "step": 16605, + "train_speed(iter/s)": 0.448305 + }, + { + "acc": 0.64769888, + "epoch": 0.42135971588026383, + "grad_norm": 4.625, + "learning_rate": 9.31059730747676e-06, + "loss": 1.5489953, + "memory(GiB)": 111.15, + "step": 16610, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.65323043, + "epoch": 0.4214865550481989, + "grad_norm": 5.28125, + "learning_rate": 9.310065869739763e-06, + "loss": 1.63912277, + "memory(GiB)": 111.15, + "step": 16615, + "train_speed(iter/s)": 0.448446 + }, + { + "acc": 0.64588718, + "epoch": 0.4216133942161339, + "grad_norm": 4.8125, + "learning_rate": 9.309534242426826e-06, + "loss": 1.6188818, + "memory(GiB)": 111.15, + "step": 16620, + "train_speed(iter/s)": 0.448517 + }, + { + "acc": 0.65608301, + "epoch": 0.421740233384069, + "grad_norm": 4.6875, + "learning_rate": 9.30900242556133e-06, + "loss": 1.61432457, + "memory(GiB)": 111.15, + "step": 16625, + "train_speed(iter/s)": 0.448588 + }, + { + "acc": 0.65423059, + "epoch": 0.42186707255200406, + "grad_norm": 5.375, + "learning_rate": 9.308470419166672e-06, + "loss": 1.56800117, + "memory(GiB)": 111.15, + "step": 16630, + "train_speed(iter/s)": 0.448658 + }, + { + "acc": 0.63786149, + "epoch": 0.4219939117199391, + "grad_norm": 5.125, + "learning_rate": 9.307938223266247e-06, + "loss": 1.6416153, + "memory(GiB)": 111.15, + "step": 16635, + "train_speed(iter/s)": 0.448729 + }, + { + "acc": 0.63211703, + "epoch": 0.42212075088787415, + "grad_norm": 4.4375, + "learning_rate": 9.307405837883467e-06, + "loss": 1.70988541, + "memory(GiB)": 111.15, + "step": 16640, + "train_speed(iter/s)": 0.4488 + }, + { + "acc": 0.6436923, + "epoch": 0.42224759005580925, + "grad_norm": 3.890625, + "learning_rate": 9.306873263041745e-06, + "loss": 1.62871246, + "memory(GiB)": 111.15, + "step": 16645, + "train_speed(iter/s)": 0.448871 + }, + { + "acc": 0.64910102, + "epoch": 0.4223744292237443, + "grad_norm": 5.5625, + "learning_rate": 9.30634049876451e-06, + "loss": 1.64532318, + "memory(GiB)": 111.15, + "step": 16650, + "train_speed(iter/s)": 0.448942 + }, + { + "acc": 0.64265566, + "epoch": 0.42250126839167934, + "grad_norm": 5.75, + "learning_rate": 9.305807545075194e-06, + "loss": 1.61112518, + "memory(GiB)": 111.15, + "step": 16655, + "train_speed(iter/s)": 0.449013 + }, + { + "acc": 0.65261889, + "epoch": 0.4226281075596144, + "grad_norm": 4.5, + "learning_rate": 9.305274401997237e-06, + "loss": 1.66084518, + "memory(GiB)": 111.15, + "step": 16660, + "train_speed(iter/s)": 0.449084 + }, + { + "acc": 0.64536171, + "epoch": 0.4227549467275495, + "grad_norm": 4.90625, + "learning_rate": 9.304741069554088e-06, + "loss": 1.61740913, + "memory(GiB)": 111.15, + "step": 16665, + "train_speed(iter/s)": 0.449154 + }, + { + "acc": 0.65130258, + "epoch": 0.42288178589548453, + "grad_norm": 5.0, + "learning_rate": 9.304207547769211e-06, + "loss": 1.60065498, + "memory(GiB)": 111.15, + "step": 16670, + "train_speed(iter/s)": 0.449226 + }, + { + "acc": 0.66398258, + "epoch": 0.4230086250634196, + "grad_norm": 6.5625, + "learning_rate": 9.303673836666067e-06, + "loss": 1.62464504, + "memory(GiB)": 111.15, + "step": 16675, + "train_speed(iter/s)": 0.449297 + }, + { + "acc": 0.63941708, + "epoch": 0.4231354642313546, + "grad_norm": 4.78125, + "learning_rate": 9.303139936268133e-06, + "loss": 1.67574768, + "memory(GiB)": 111.15, + "step": 16680, + "train_speed(iter/s)": 0.449369 + }, + { + "acc": 0.63693571, + "epoch": 0.4232623033992897, + "grad_norm": 4.875, + "learning_rate": 9.302605846598894e-06, + "loss": 1.68863716, + "memory(GiB)": 111.15, + "step": 16685, + "train_speed(iter/s)": 0.44944 + }, + { + "acc": 0.6443119, + "epoch": 0.42338914256722476, + "grad_norm": 6.53125, + "learning_rate": 9.30207156768184e-06, + "loss": 1.65081215, + "memory(GiB)": 111.15, + "step": 16690, + "train_speed(iter/s)": 0.449511 + }, + { + "acc": 0.65241866, + "epoch": 0.4235159817351598, + "grad_norm": 5.5, + "learning_rate": 9.30153709954047e-06, + "loss": 1.58175926, + "memory(GiB)": 111.15, + "step": 16695, + "train_speed(iter/s)": 0.449582 + }, + { + "acc": 0.65260615, + "epoch": 0.42364282090309485, + "grad_norm": 4.96875, + "learning_rate": 9.301002442198294e-06, + "loss": 1.58351049, + "memory(GiB)": 111.15, + "step": 16700, + "train_speed(iter/s)": 0.449654 + }, + { + "epoch": 0.42364282090309485, + "eval_acc": 0.6402451124358048, + "eval_loss": 1.6261844635009766, + "eval_runtime": 113.5013, + "eval_samples_per_second": 56.123, + "eval_steps_per_second": 28.061, + "step": 16700 + }, + { + "acc": 0.64703741, + "epoch": 0.42376966007102995, + "grad_norm": 5.40625, + "learning_rate": 9.300467595678829e-06, + "loss": 1.65279541, + "memory(GiB)": 111.15, + "step": 16705, + "train_speed(iter/s)": 0.4483 + }, + { + "acc": 0.64552221, + "epoch": 0.423896499238965, + "grad_norm": 6.78125, + "learning_rate": 9.299932560005596e-06, + "loss": 1.6721447, + "memory(GiB)": 111.15, + "step": 16710, + "train_speed(iter/s)": 0.44837 + }, + { + "acc": 0.65697675, + "epoch": 0.42402333840690004, + "grad_norm": 5.5, + "learning_rate": 9.299397335202133e-06, + "loss": 1.66039581, + "memory(GiB)": 111.15, + "step": 16715, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.6539917, + "epoch": 0.4241501775748351, + "grad_norm": 5.1875, + "learning_rate": 9.29886192129198e-06, + "loss": 1.61069336, + "memory(GiB)": 111.15, + "step": 16720, + "train_speed(iter/s)": 0.44851 + }, + { + "acc": 0.63343267, + "epoch": 0.4242770167427702, + "grad_norm": 5.46875, + "learning_rate": 9.298326318298688e-06, + "loss": 1.66645775, + "memory(GiB)": 111.15, + "step": 16725, + "train_speed(iter/s)": 0.448579 + }, + { + "acc": 0.64545574, + "epoch": 0.42440385591070523, + "grad_norm": 5.28125, + "learning_rate": 9.29779052624581e-06, + "loss": 1.59622297, + "memory(GiB)": 111.15, + "step": 16730, + "train_speed(iter/s)": 0.448649 + }, + { + "acc": 0.65057602, + "epoch": 0.4245306950786403, + "grad_norm": 5.96875, + "learning_rate": 9.29725454515692e-06, + "loss": 1.65726471, + "memory(GiB)": 111.15, + "step": 16735, + "train_speed(iter/s)": 0.44872 + }, + { + "acc": 0.65226707, + "epoch": 0.4246575342465753, + "grad_norm": 5.34375, + "learning_rate": 9.296718375055587e-06, + "loss": 1.57008324, + "memory(GiB)": 111.15, + "step": 16740, + "train_speed(iter/s)": 0.44879 + }, + { + "acc": 0.64606524, + "epoch": 0.4247843734145104, + "grad_norm": 5.21875, + "learning_rate": 9.296182015965399e-06, + "loss": 1.67107315, + "memory(GiB)": 111.15, + "step": 16745, + "train_speed(iter/s)": 0.448861 + }, + { + "acc": 0.65154357, + "epoch": 0.42491121258244546, + "grad_norm": 7.25, + "learning_rate": 9.295645467909942e-06, + "loss": 1.64996605, + "memory(GiB)": 111.15, + "step": 16750, + "train_speed(iter/s)": 0.448931 + }, + { + "acc": 0.6425993, + "epoch": 0.4250380517503805, + "grad_norm": 5.09375, + "learning_rate": 9.29510873091282e-06, + "loss": 1.59227448, + "memory(GiB)": 111.15, + "step": 16755, + "train_speed(iter/s)": 0.449001 + }, + { + "acc": 0.64310169, + "epoch": 0.42516489091831555, + "grad_norm": 5.875, + "learning_rate": 9.29457180499764e-06, + "loss": 1.66960564, + "memory(GiB)": 111.15, + "step": 16760, + "train_speed(iter/s)": 0.449072 + }, + { + "acc": 0.64456792, + "epoch": 0.42529173008625065, + "grad_norm": 5.5625, + "learning_rate": 9.294034690188016e-06, + "loss": 1.66299686, + "memory(GiB)": 111.15, + "step": 16765, + "train_speed(iter/s)": 0.449142 + }, + { + "acc": 0.64090328, + "epoch": 0.4254185692541857, + "grad_norm": 4.65625, + "learning_rate": 9.293497386507577e-06, + "loss": 1.68580379, + "memory(GiB)": 111.15, + "step": 16770, + "train_speed(iter/s)": 0.449212 + }, + { + "acc": 0.6614501, + "epoch": 0.42554540842212074, + "grad_norm": 6.09375, + "learning_rate": 9.292959893979953e-06, + "loss": 1.61997566, + "memory(GiB)": 111.15, + "step": 16775, + "train_speed(iter/s)": 0.449283 + }, + { + "acc": 0.64647217, + "epoch": 0.4256722475900558, + "grad_norm": 4.96875, + "learning_rate": 9.292422212628786e-06, + "loss": 1.67102852, + "memory(GiB)": 111.15, + "step": 16780, + "train_speed(iter/s)": 0.449353 + }, + { + "acc": 0.65543585, + "epoch": 0.4257990867579909, + "grad_norm": 5.21875, + "learning_rate": 9.291884342477728e-06, + "loss": 1.63762817, + "memory(GiB)": 111.15, + "step": 16785, + "train_speed(iter/s)": 0.449423 + }, + { + "acc": 0.64925184, + "epoch": 0.42592592592592593, + "grad_norm": 4.78125, + "learning_rate": 9.291346283550433e-06, + "loss": 1.62338123, + "memory(GiB)": 111.15, + "step": 16790, + "train_speed(iter/s)": 0.449494 + }, + { + "acc": 0.65038376, + "epoch": 0.426052765093861, + "grad_norm": 5.625, + "learning_rate": 9.290808035870569e-06, + "loss": 1.68389359, + "memory(GiB)": 111.15, + "step": 16795, + "train_speed(iter/s)": 0.449563 + }, + { + "acc": 0.63843489, + "epoch": 0.426179604261796, + "grad_norm": 5.6875, + "learning_rate": 9.29026959946181e-06, + "loss": 1.68883514, + "memory(GiB)": 111.15, + "step": 16800, + "train_speed(iter/s)": 0.449634 + }, + { + "epoch": 0.426179604261796, + "eval_acc": 0.6402421880646386, + "eval_loss": 1.6256159543991089, + "eval_runtime": 112.8047, + "eval_samples_per_second": 56.469, + "eval_steps_per_second": 28.235, + "step": 16800 + }, + { + "acc": 0.65532074, + "epoch": 0.4263064434297311, + "grad_norm": 4.625, + "learning_rate": 9.289730974347841e-06, + "loss": 1.61831551, + "memory(GiB)": 111.15, + "step": 16805, + "train_speed(iter/s)": 0.448296 + }, + { + "acc": 0.67026796, + "epoch": 0.42643328259766616, + "grad_norm": 5.40625, + "learning_rate": 9.28919216055235e-06, + "loss": 1.59910727, + "memory(GiB)": 111.15, + "step": 16810, + "train_speed(iter/s)": 0.448366 + }, + { + "acc": 0.66124544, + "epoch": 0.4265601217656012, + "grad_norm": 6.21875, + "learning_rate": 9.288653158099038e-06, + "loss": 1.58079872, + "memory(GiB)": 111.15, + "step": 16815, + "train_speed(iter/s)": 0.448436 + }, + { + "acc": 0.64744959, + "epoch": 0.42668696093353625, + "grad_norm": 5.125, + "learning_rate": 9.288113967011612e-06, + "loss": 1.689254, + "memory(GiB)": 111.15, + "step": 16820, + "train_speed(iter/s)": 0.448506 + }, + { + "acc": 0.6557457, + "epoch": 0.42681380010147135, + "grad_norm": 5.21875, + "learning_rate": 9.28757458731379e-06, + "loss": 1.60272255, + "memory(GiB)": 111.15, + "step": 16825, + "train_speed(iter/s)": 0.448576 + }, + { + "acc": 0.64284897, + "epoch": 0.4269406392694064, + "grad_norm": 5.40625, + "learning_rate": 9.287035019029295e-06, + "loss": 1.6989666, + "memory(GiB)": 111.15, + "step": 16830, + "train_speed(iter/s)": 0.448646 + }, + { + "acc": 0.64833803, + "epoch": 0.42706747843734144, + "grad_norm": 5.375, + "learning_rate": 9.286495262181859e-06, + "loss": 1.5594039, + "memory(GiB)": 111.15, + "step": 16835, + "train_speed(iter/s)": 0.448715 + }, + { + "acc": 0.64770517, + "epoch": 0.4271943176052765, + "grad_norm": 7.15625, + "learning_rate": 9.285955316795224e-06, + "loss": 1.69755535, + "memory(GiB)": 111.15, + "step": 16840, + "train_speed(iter/s)": 0.448785 + }, + { + "acc": 0.65035524, + "epoch": 0.4273211567732116, + "grad_norm": 6.8125, + "learning_rate": 9.285415182893138e-06, + "loss": 1.67875977, + "memory(GiB)": 111.15, + "step": 16845, + "train_speed(iter/s)": 0.448855 + }, + { + "acc": 0.64524832, + "epoch": 0.42744799594114663, + "grad_norm": 6.25, + "learning_rate": 9.28487486049936e-06, + "loss": 1.63049011, + "memory(GiB)": 111.15, + "step": 16850, + "train_speed(iter/s)": 0.448924 + }, + { + "acc": 0.64672074, + "epoch": 0.4275748351090817, + "grad_norm": 6.0625, + "learning_rate": 9.284334349637655e-06, + "loss": 1.71820488, + "memory(GiB)": 111.15, + "step": 16855, + "train_speed(iter/s)": 0.448994 + }, + { + "acc": 0.65825691, + "epoch": 0.4277016742770167, + "grad_norm": 5.3125, + "learning_rate": 9.283793650331798e-06, + "loss": 1.57684355, + "memory(GiB)": 111.15, + "step": 16860, + "train_speed(iter/s)": 0.449063 + }, + { + "acc": 0.66200809, + "epoch": 0.4278285134449518, + "grad_norm": 5.34375, + "learning_rate": 9.283252762605568e-06, + "loss": 1.60479908, + "memory(GiB)": 111.15, + "step": 16865, + "train_speed(iter/s)": 0.449133 + }, + { + "acc": 0.64627447, + "epoch": 0.42795535261288686, + "grad_norm": 5.0625, + "learning_rate": 9.28271168648276e-06, + "loss": 1.65301056, + "memory(GiB)": 111.15, + "step": 16870, + "train_speed(iter/s)": 0.449203 + }, + { + "acc": 0.6405756, + "epoch": 0.4280821917808219, + "grad_norm": 6.03125, + "learning_rate": 9.282170421987171e-06, + "loss": 1.68900375, + "memory(GiB)": 111.15, + "step": 16875, + "train_speed(iter/s)": 0.449273 + }, + { + "acc": 0.64207592, + "epoch": 0.42820903094875695, + "grad_norm": 5.90625, + "learning_rate": 9.281628969142609e-06, + "loss": 1.73415585, + "memory(GiB)": 111.15, + "step": 16880, + "train_speed(iter/s)": 0.449343 + }, + { + "acc": 0.63002329, + "epoch": 0.42833587011669205, + "grad_norm": 5.53125, + "learning_rate": 9.281087327972886e-06, + "loss": 1.73738594, + "memory(GiB)": 111.15, + "step": 16885, + "train_speed(iter/s)": 0.449413 + }, + { + "acc": 0.651647, + "epoch": 0.4284627092846271, + "grad_norm": 5.59375, + "learning_rate": 9.280545498501832e-06, + "loss": 1.56453705, + "memory(GiB)": 111.15, + "step": 16890, + "train_speed(iter/s)": 0.449483 + }, + { + "acc": 0.63814845, + "epoch": 0.42858954845256214, + "grad_norm": 5.125, + "learning_rate": 9.280003480753274e-06, + "loss": 1.72573929, + "memory(GiB)": 111.15, + "step": 16895, + "train_speed(iter/s)": 0.449553 + }, + { + "acc": 0.62625546, + "epoch": 0.4287163876204972, + "grad_norm": 5.875, + "learning_rate": 9.279461274751054e-06, + "loss": 1.71955948, + "memory(GiB)": 111.15, + "step": 16900, + "train_speed(iter/s)": 0.449623 + }, + { + "epoch": 0.4287163876204972, + "eval_acc": 0.6403086130668424, + "eval_loss": 1.6253241300582886, + "eval_runtime": 112.5476, + "eval_samples_per_second": 56.598, + "eval_steps_per_second": 28.299, + "step": 16900 + }, + { + "acc": 0.63471165, + "epoch": 0.4288432267884323, + "grad_norm": 5.0625, + "learning_rate": 9.27891888051902e-06, + "loss": 1.65053787, + "memory(GiB)": 111.15, + "step": 16905, + "train_speed(iter/s)": 0.448297 + }, + { + "acc": 0.65053587, + "epoch": 0.42897006595636733, + "grad_norm": 5.375, + "learning_rate": 9.278376298081032e-06, + "loss": 1.66243649, + "memory(GiB)": 111.15, + "step": 16910, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.64819679, + "epoch": 0.4290969051243024, + "grad_norm": 5.46875, + "learning_rate": 9.277833527460952e-06, + "loss": 1.65649776, + "memory(GiB)": 111.15, + "step": 16915, + "train_speed(iter/s)": 0.448437 + }, + { + "acc": 0.65822973, + "epoch": 0.4292237442922374, + "grad_norm": 5.25, + "learning_rate": 9.277290568682653e-06, + "loss": 1.67826843, + "memory(GiB)": 111.15, + "step": 16920, + "train_speed(iter/s)": 0.448507 + }, + { + "acc": 0.62949476, + "epoch": 0.4293505834601725, + "grad_norm": 5.375, + "learning_rate": 9.27674742177002e-06, + "loss": 1.71830711, + "memory(GiB)": 111.15, + "step": 16925, + "train_speed(iter/s)": 0.448576 + }, + { + "acc": 0.63309774, + "epoch": 0.42947742262810756, + "grad_norm": 5.09375, + "learning_rate": 9.27620408674694e-06, + "loss": 1.61059761, + "memory(GiB)": 111.15, + "step": 16930, + "train_speed(iter/s)": 0.448645 + }, + { + "acc": 0.63333473, + "epoch": 0.4296042617960426, + "grad_norm": 4.8125, + "learning_rate": 9.275660563637313e-06, + "loss": 1.63589363, + "memory(GiB)": 111.15, + "step": 16935, + "train_speed(iter/s)": 0.448715 + }, + { + "acc": 0.65663433, + "epoch": 0.42973110096397765, + "grad_norm": 5.40625, + "learning_rate": 9.275116852465043e-06, + "loss": 1.64043427, + "memory(GiB)": 111.15, + "step": 16940, + "train_speed(iter/s)": 0.448785 + }, + { + "acc": 0.64849672, + "epoch": 0.42985794013191275, + "grad_norm": 5.84375, + "learning_rate": 9.274572953254048e-06, + "loss": 1.60335484, + "memory(GiB)": 111.15, + "step": 16945, + "train_speed(iter/s)": 0.448856 + }, + { + "acc": 0.66199055, + "epoch": 0.4299847792998478, + "grad_norm": 6.1875, + "learning_rate": 9.27402886602825e-06, + "loss": 1.52564068, + "memory(GiB)": 111.15, + "step": 16950, + "train_speed(iter/s)": 0.448925 + }, + { + "acc": 0.63999491, + "epoch": 0.43011161846778284, + "grad_norm": 5.78125, + "learning_rate": 9.27348459081158e-06, + "loss": 1.68128662, + "memory(GiB)": 111.15, + "step": 16955, + "train_speed(iter/s)": 0.448996 + }, + { + "acc": 0.6496212, + "epoch": 0.4302384576357179, + "grad_norm": 4.90625, + "learning_rate": 9.272940127627979e-06, + "loss": 1.6156641, + "memory(GiB)": 111.15, + "step": 16960, + "train_speed(iter/s)": 0.449066 + }, + { + "acc": 0.64575748, + "epoch": 0.430365296803653, + "grad_norm": 6.40625, + "learning_rate": 9.272395476501392e-06, + "loss": 1.64408264, + "memory(GiB)": 111.15, + "step": 16965, + "train_speed(iter/s)": 0.449137 + }, + { + "acc": 0.63206034, + "epoch": 0.43049213597158803, + "grad_norm": 4.46875, + "learning_rate": 9.27185063745578e-06, + "loss": 1.71578598, + "memory(GiB)": 111.15, + "step": 16970, + "train_speed(iter/s)": 0.449207 + }, + { + "acc": 0.65571642, + "epoch": 0.4306189751395231, + "grad_norm": 5.0, + "learning_rate": 9.271305610515103e-06, + "loss": 1.59589539, + "memory(GiB)": 111.15, + "step": 16975, + "train_speed(iter/s)": 0.449277 + }, + { + "acc": 0.64838943, + "epoch": 0.4307458143074581, + "grad_norm": 4.625, + "learning_rate": 9.270760395703334e-06, + "loss": 1.65031166, + "memory(GiB)": 111.15, + "step": 16980, + "train_speed(iter/s)": 0.449347 + }, + { + "acc": 0.64349413, + "epoch": 0.4308726534753932, + "grad_norm": 5.15625, + "learning_rate": 9.270214993044456e-06, + "loss": 1.68148384, + "memory(GiB)": 111.15, + "step": 16985, + "train_speed(iter/s)": 0.449417 + }, + { + "acc": 0.64259553, + "epoch": 0.43099949264332826, + "grad_norm": 5.21875, + "learning_rate": 9.269669402562458e-06, + "loss": 1.64350758, + "memory(GiB)": 111.15, + "step": 16990, + "train_speed(iter/s)": 0.449487 + }, + { + "acc": 0.65871038, + "epoch": 0.4311263318112633, + "grad_norm": 5.5, + "learning_rate": 9.269123624281336e-06, + "loss": 1.62274895, + "memory(GiB)": 111.15, + "step": 16995, + "train_speed(iter/s)": 0.449557 + }, + { + "acc": 0.66687322, + "epoch": 0.43125317097919835, + "grad_norm": 4.75, + "learning_rate": 9.268577658225097e-06, + "loss": 1.50307837, + "memory(GiB)": 111.15, + "step": 17000, + "train_speed(iter/s)": 0.449627 + }, + { + "epoch": 0.43125317097919835, + "eval_acc": 0.6403203105515072, + "eval_loss": 1.625187635421753, + "eval_runtime": 113.4524, + "eval_samples_per_second": 56.147, + "eval_steps_per_second": 28.073, + "step": 17000 + }, + { + "acc": 0.65020485, + "epoch": 0.43138001014713345, + "grad_norm": 5.8125, + "learning_rate": 9.268031504417756e-06, + "loss": 1.62717075, + "memory(GiB)": 111.15, + "step": 17005, + "train_speed(iter/s)": 0.448298 + }, + { + "acc": 0.64776955, + "epoch": 0.4315068493150685, + "grad_norm": 5.90625, + "learning_rate": 9.267485162883334e-06, + "loss": 1.66278648, + "memory(GiB)": 111.15, + "step": 17010, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.63303723, + "epoch": 0.43163368848300354, + "grad_norm": 5.9375, + "learning_rate": 9.266938633645861e-06, + "loss": 1.73811512, + "memory(GiB)": 111.15, + "step": 17015, + "train_speed(iter/s)": 0.448437 + }, + { + "acc": 0.66061316, + "epoch": 0.4317605276509386, + "grad_norm": 6.15625, + "learning_rate": 9.266391916729376e-06, + "loss": 1.57575855, + "memory(GiB)": 111.15, + "step": 17020, + "train_speed(iter/s)": 0.448507 + }, + { + "acc": 0.63892803, + "epoch": 0.4318873668188737, + "grad_norm": 5.09375, + "learning_rate": 9.265845012157926e-06, + "loss": 1.66742783, + "memory(GiB)": 111.15, + "step": 17025, + "train_speed(iter/s)": 0.448576 + }, + { + "acc": 0.64368315, + "epoch": 0.43201420598680873, + "grad_norm": 6.09375, + "learning_rate": 9.265297919955566e-06, + "loss": 1.64607544, + "memory(GiB)": 111.15, + "step": 17030, + "train_speed(iter/s)": 0.448645 + }, + { + "acc": 0.64491901, + "epoch": 0.4321410451547438, + "grad_norm": 5.375, + "learning_rate": 9.264750640146363e-06, + "loss": 1.66307335, + "memory(GiB)": 111.15, + "step": 17035, + "train_speed(iter/s)": 0.448714 + }, + { + "acc": 0.64818497, + "epoch": 0.4322678843226788, + "grad_norm": 4.1875, + "learning_rate": 9.264203172754384e-06, + "loss": 1.64637089, + "memory(GiB)": 111.15, + "step": 17040, + "train_speed(iter/s)": 0.448783 + }, + { + "acc": 0.65443091, + "epoch": 0.4323947234906139, + "grad_norm": 5.34375, + "learning_rate": 9.263655517803713e-06, + "loss": 1.68622589, + "memory(GiB)": 111.15, + "step": 17045, + "train_speed(iter/s)": 0.448852 + }, + { + "acc": 0.64291744, + "epoch": 0.43252156265854896, + "grad_norm": 4.5625, + "learning_rate": 9.263107675318434e-06, + "loss": 1.70068359, + "memory(GiB)": 111.15, + "step": 17050, + "train_speed(iter/s)": 0.448922 + }, + { + "acc": 0.65749989, + "epoch": 0.432648401826484, + "grad_norm": 4.9375, + "learning_rate": 9.262559645322648e-06, + "loss": 1.62571373, + "memory(GiB)": 111.15, + "step": 17055, + "train_speed(iter/s)": 0.448991 + }, + { + "acc": 0.65325146, + "epoch": 0.43277524099441905, + "grad_norm": 5.53125, + "learning_rate": 9.262011427840459e-06, + "loss": 1.62780533, + "memory(GiB)": 111.15, + "step": 17060, + "train_speed(iter/s)": 0.44906 + }, + { + "acc": 0.65508704, + "epoch": 0.43290208016235415, + "grad_norm": 5.0, + "learning_rate": 9.261463022895976e-06, + "loss": 1.53218727, + "memory(GiB)": 111.15, + "step": 17065, + "train_speed(iter/s)": 0.449128 + }, + { + "acc": 0.65401487, + "epoch": 0.4330289193302892, + "grad_norm": 5.78125, + "learning_rate": 9.260914430513325e-06, + "loss": 1.61691055, + "memory(GiB)": 111.15, + "step": 17070, + "train_speed(iter/s)": 0.449197 + }, + { + "acc": 0.63701439, + "epoch": 0.43315575849822424, + "grad_norm": 5.21875, + "learning_rate": 9.260365650716632e-06, + "loss": 1.70378799, + "memory(GiB)": 111.15, + "step": 17075, + "train_speed(iter/s)": 0.449266 + }, + { + "acc": 0.64877682, + "epoch": 0.4332825976661593, + "grad_norm": 6.1875, + "learning_rate": 9.259816683530038e-06, + "loss": 1.67790031, + "memory(GiB)": 111.15, + "step": 17080, + "train_speed(iter/s)": 0.449335 + }, + { + "acc": 0.65751724, + "epoch": 0.4334094368340944, + "grad_norm": 5.25, + "learning_rate": 9.259267528977687e-06, + "loss": 1.57302475, + "memory(GiB)": 111.15, + "step": 17085, + "train_speed(iter/s)": 0.449405 + }, + { + "acc": 0.65287991, + "epoch": 0.43353627600202943, + "grad_norm": 5.09375, + "learning_rate": 9.258718187083734e-06, + "loss": 1.63934536, + "memory(GiB)": 111.15, + "step": 17090, + "train_speed(iter/s)": 0.449473 + }, + { + "acc": 0.65327225, + "epoch": 0.4336631151699645, + "grad_norm": 5.59375, + "learning_rate": 9.258168657872341e-06, + "loss": 1.62802601, + "memory(GiB)": 111.15, + "step": 17095, + "train_speed(iter/s)": 0.449543 + }, + { + "acc": 0.65321531, + "epoch": 0.4337899543378995, + "grad_norm": 5.21875, + "learning_rate": 9.25761894136768e-06, + "loss": 1.66509285, + "memory(GiB)": 111.15, + "step": 17100, + "train_speed(iter/s)": 0.449612 + }, + { + "epoch": 0.4337899543378995, + "eval_acc": 0.6403015110225816, + "eval_loss": 1.6251171827316284, + "eval_runtime": 113.1828, + "eval_samples_per_second": 56.281, + "eval_steps_per_second": 28.14, + "step": 17100 + }, + { + "acc": 0.64019852, + "epoch": 0.4339167935058346, + "grad_norm": 4.53125, + "learning_rate": 9.25706903759393e-06, + "loss": 1.6697237, + "memory(GiB)": 111.15, + "step": 17105, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.65722713, + "epoch": 0.43404363267376966, + "grad_norm": 5.8125, + "learning_rate": 9.256518946575274e-06, + "loss": 1.59425831, + "memory(GiB)": 111.15, + "step": 17110, + "train_speed(iter/s)": 0.448363 + }, + { + "acc": 0.65019989, + "epoch": 0.4341704718417047, + "grad_norm": 6.8125, + "learning_rate": 9.255968668335912e-06, + "loss": 1.6627388, + "memory(GiB)": 111.15, + "step": 17115, + "train_speed(iter/s)": 0.448432 + }, + { + "acc": 0.64703226, + "epoch": 0.43429731100963975, + "grad_norm": 6.09375, + "learning_rate": 9.255418202900048e-06, + "loss": 1.6077858, + "memory(GiB)": 111.15, + "step": 17120, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.63817945, + "epoch": 0.43442415017757485, + "grad_norm": 4.96875, + "learning_rate": 9.254867550291891e-06, + "loss": 1.68035316, + "memory(GiB)": 111.15, + "step": 17125, + "train_speed(iter/s)": 0.448569 + }, + { + "acc": 0.65268922, + "epoch": 0.4345509893455099, + "grad_norm": 5.34375, + "learning_rate": 9.254316710535662e-06, + "loss": 1.5683012, + "memory(GiB)": 111.15, + "step": 17130, + "train_speed(iter/s)": 0.448638 + }, + { + "acc": 0.65181632, + "epoch": 0.43467782851344494, + "grad_norm": 5.53125, + "learning_rate": 9.253765683655591e-06, + "loss": 1.63271027, + "memory(GiB)": 111.15, + "step": 17135, + "train_speed(iter/s)": 0.448707 + }, + { + "acc": 0.67040572, + "epoch": 0.43480466768138, + "grad_norm": 6.28125, + "learning_rate": 9.253214469675913e-06, + "loss": 1.57940731, + "memory(GiB)": 111.15, + "step": 17140, + "train_speed(iter/s)": 0.448776 + }, + { + "acc": 0.65769253, + "epoch": 0.4349315068493151, + "grad_norm": 5.53125, + "learning_rate": 9.252663068620874e-06, + "loss": 1.62087173, + "memory(GiB)": 111.15, + "step": 17145, + "train_speed(iter/s)": 0.448845 + }, + { + "acc": 0.63956175, + "epoch": 0.43505834601725013, + "grad_norm": 4.46875, + "learning_rate": 9.252111480514726e-06, + "loss": 1.67999439, + "memory(GiB)": 111.15, + "step": 17150, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.6475544, + "epoch": 0.4351851851851852, + "grad_norm": 5.71875, + "learning_rate": 9.251559705381731e-06, + "loss": 1.66605186, + "memory(GiB)": 111.15, + "step": 17155, + "train_speed(iter/s)": 0.448982 + }, + { + "acc": 0.62009058, + "epoch": 0.4353120243531202, + "grad_norm": 6.21875, + "learning_rate": 9.251007743246159e-06, + "loss": 1.73617134, + "memory(GiB)": 111.15, + "step": 17160, + "train_speed(iter/s)": 0.449052 + }, + { + "acc": 0.66004038, + "epoch": 0.4354388635210553, + "grad_norm": 6.625, + "learning_rate": 9.250455594132286e-06, + "loss": 1.62910156, + "memory(GiB)": 111.15, + "step": 17165, + "train_speed(iter/s)": 0.449121 + }, + { + "acc": 0.6510251, + "epoch": 0.43556570268899036, + "grad_norm": 5.0625, + "learning_rate": 9.249903258064399e-06, + "loss": 1.66936951, + "memory(GiB)": 111.15, + "step": 17170, + "train_speed(iter/s)": 0.44919 + }, + { + "acc": 0.65864077, + "epoch": 0.4356925418569254, + "grad_norm": 5.53125, + "learning_rate": 9.249350735066792e-06, + "loss": 1.58294449, + "memory(GiB)": 111.15, + "step": 17175, + "train_speed(iter/s)": 0.44926 + }, + { + "acc": 0.63140764, + "epoch": 0.43581938102486045, + "grad_norm": 5.09375, + "learning_rate": 9.248798025163768e-06, + "loss": 1.65163994, + "memory(GiB)": 111.15, + "step": 17180, + "train_speed(iter/s)": 0.449328 + }, + { + "acc": 0.65269561, + "epoch": 0.43594622019279555, + "grad_norm": 6.5, + "learning_rate": 9.248245128379638e-06, + "loss": 1.58572521, + "memory(GiB)": 111.15, + "step": 17185, + "train_speed(iter/s)": 0.449397 + }, + { + "acc": 0.64936581, + "epoch": 0.4360730593607306, + "grad_norm": 5.15625, + "learning_rate": 9.24769204473872e-06, + "loss": 1.61195831, + "memory(GiB)": 111.15, + "step": 17190, + "train_speed(iter/s)": 0.449467 + }, + { + "acc": 0.6508296, + "epoch": 0.43619989852866564, + "grad_norm": 5.09375, + "learning_rate": 9.24713877426534e-06, + "loss": 1.5736907, + "memory(GiB)": 111.15, + "step": 17195, + "train_speed(iter/s)": 0.449536 + }, + { + "acc": 0.65803494, + "epoch": 0.4363267376966007, + "grad_norm": 5.03125, + "learning_rate": 9.246585316983837e-06, + "loss": 1.63135242, + "memory(GiB)": 111.15, + "step": 17200, + "train_speed(iter/s)": 0.449605 + }, + { + "epoch": 0.4363267376966007, + "eval_acc": 0.6402597342916359, + "eval_loss": 1.6247406005859375, + "eval_runtime": 113.2372, + "eval_samples_per_second": 56.254, + "eval_steps_per_second": 28.127, + "step": 17200 + }, + { + "acc": 0.63855062, + "epoch": 0.4364535768645358, + "grad_norm": 4.71875, + "learning_rate": 9.24603167291855e-06, + "loss": 1.72843895, + "memory(GiB)": 111.15, + "step": 17205, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.64426422, + "epoch": 0.43658041603247083, + "grad_norm": 5.46875, + "learning_rate": 9.245477842093832e-06, + "loss": 1.62153549, + "memory(GiB)": 111.15, + "step": 17210, + "train_speed(iter/s)": 0.448364 + }, + { + "acc": 0.64279222, + "epoch": 0.4367072552004059, + "grad_norm": 5.5, + "learning_rate": 9.244923824534046e-06, + "loss": 1.63029099, + "memory(GiB)": 111.15, + "step": 17215, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.63957987, + "epoch": 0.4368340943683409, + "grad_norm": 5.28125, + "learning_rate": 9.244369620263558e-06, + "loss": 1.68073273, + "memory(GiB)": 111.15, + "step": 17220, + "train_speed(iter/s)": 0.448503 + }, + { + "acc": 0.64821095, + "epoch": 0.436960933536276, + "grad_norm": 5.5, + "learning_rate": 9.243815229306746e-06, + "loss": 1.60483761, + "memory(GiB)": 111.15, + "step": 17225, + "train_speed(iter/s)": 0.448571 + }, + { + "acc": 0.64310207, + "epoch": 0.43708777270421106, + "grad_norm": 6.125, + "learning_rate": 9.243260651687989e-06, + "loss": 1.69257698, + "memory(GiB)": 111.15, + "step": 17230, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.6435822, + "epoch": 0.4372146118721461, + "grad_norm": 5.1875, + "learning_rate": 9.242705887431685e-06, + "loss": 1.67983665, + "memory(GiB)": 111.15, + "step": 17235, + "train_speed(iter/s)": 0.44871 + }, + { + "acc": 0.64623799, + "epoch": 0.43734145104008115, + "grad_norm": 6.59375, + "learning_rate": 9.242150936562235e-06, + "loss": 1.71585026, + "memory(GiB)": 111.15, + "step": 17240, + "train_speed(iter/s)": 0.448779 + }, + { + "acc": 0.65051394, + "epoch": 0.43746829020801625, + "grad_norm": 4.875, + "learning_rate": 9.241595799104046e-06, + "loss": 1.58839073, + "memory(GiB)": 111.15, + "step": 17245, + "train_speed(iter/s)": 0.448848 + }, + { + "acc": 0.6440589, + "epoch": 0.4375951293759513, + "grad_norm": 5.15625, + "learning_rate": 9.241040475081537e-06, + "loss": 1.67115936, + "memory(GiB)": 111.15, + "step": 17250, + "train_speed(iter/s)": 0.448918 + }, + { + "acc": 0.63667784, + "epoch": 0.43772196854388634, + "grad_norm": 4.90625, + "learning_rate": 9.240484964519131e-06, + "loss": 1.66938286, + "memory(GiB)": 111.15, + "step": 17255, + "train_speed(iter/s)": 0.448987 + }, + { + "acc": 0.65027533, + "epoch": 0.4378488077118214, + "grad_norm": 6.46875, + "learning_rate": 9.239929267441267e-06, + "loss": 1.60519676, + "memory(GiB)": 111.15, + "step": 17260, + "train_speed(iter/s)": 0.449057 + }, + { + "acc": 0.65911837, + "epoch": 0.4379756468797565, + "grad_norm": 5.71875, + "learning_rate": 9.239373383872382e-06, + "loss": 1.6991806, + "memory(GiB)": 111.15, + "step": 17265, + "train_speed(iter/s)": 0.449126 + }, + { + "acc": 0.63225455, + "epoch": 0.43810248604769153, + "grad_norm": 4.875, + "learning_rate": 9.238817313836927e-06, + "loss": 1.6116333, + "memory(GiB)": 111.15, + "step": 17270, + "train_speed(iter/s)": 0.449196 + }, + { + "acc": 0.65202637, + "epoch": 0.4382293252156266, + "grad_norm": 4.71875, + "learning_rate": 9.238261057359365e-06, + "loss": 1.60404243, + "memory(GiB)": 111.15, + "step": 17275, + "train_speed(iter/s)": 0.449265 + }, + { + "acc": 0.64760871, + "epoch": 0.4383561643835616, + "grad_norm": 5.03125, + "learning_rate": 9.237704614464157e-06, + "loss": 1.65551243, + "memory(GiB)": 111.15, + "step": 17280, + "train_speed(iter/s)": 0.449334 + }, + { + "acc": 0.64110723, + "epoch": 0.4384830035514967, + "grad_norm": 4.6875, + "learning_rate": 9.237147985175781e-06, + "loss": 1.67709274, + "memory(GiB)": 111.15, + "step": 17285, + "train_speed(iter/s)": 0.449403 + }, + { + "acc": 0.65886908, + "epoch": 0.43860984271943176, + "grad_norm": 4.84375, + "learning_rate": 9.236591169518717e-06, + "loss": 1.62398186, + "memory(GiB)": 111.15, + "step": 17290, + "train_speed(iter/s)": 0.449473 + }, + { + "acc": 0.65168304, + "epoch": 0.4387366818873668, + "grad_norm": 5.625, + "learning_rate": 9.236034167517461e-06, + "loss": 1.65241547, + "memory(GiB)": 111.15, + "step": 17295, + "train_speed(iter/s)": 0.449542 + }, + { + "acc": 0.65537882, + "epoch": 0.43886352105530185, + "grad_norm": 6.5, + "learning_rate": 9.235476979196507e-06, + "loss": 1.61735764, + "memory(GiB)": 111.15, + "step": 17300, + "train_speed(iter/s)": 0.449611 + }, + { + "epoch": 0.43886352105530185, + "eval_acc": 0.6403086130668424, + "eval_loss": 1.624882459640503, + "eval_runtime": 113.5301, + "eval_samples_per_second": 56.108, + "eval_steps_per_second": 28.054, + "step": 17300 + }, + { + "acc": 0.65920978, + "epoch": 0.43899036022323695, + "grad_norm": 5.5, + "learning_rate": 9.234919604580368e-06, + "loss": 1.63529129, + "memory(GiB)": 111.15, + "step": 17305, + "train_speed(iter/s)": 0.448305 + }, + { + "acc": 0.66436696, + "epoch": 0.439117199391172, + "grad_norm": 5.3125, + "learning_rate": 9.234362043693556e-06, + "loss": 1.55953693, + "memory(GiB)": 111.15, + "step": 17310, + "train_speed(iter/s)": 0.448374 + }, + { + "acc": 0.63698516, + "epoch": 0.43924403855910704, + "grad_norm": 4.90625, + "learning_rate": 9.233804296560596e-06, + "loss": 1.64951096, + "memory(GiB)": 111.15, + "step": 17315, + "train_speed(iter/s)": 0.448441 + }, + { + "acc": 0.62880697, + "epoch": 0.4393708777270421, + "grad_norm": 5.1875, + "learning_rate": 9.233246363206021e-06, + "loss": 1.72565022, + "memory(GiB)": 111.15, + "step": 17320, + "train_speed(iter/s)": 0.44851 + }, + { + "acc": 0.64421535, + "epoch": 0.4394977168949772, + "grad_norm": 5.15625, + "learning_rate": 9.232688243654371e-06, + "loss": 1.70402832, + "memory(GiB)": 111.15, + "step": 17325, + "train_speed(iter/s)": 0.448579 + }, + { + "acc": 0.63451166, + "epoch": 0.43962455606291223, + "grad_norm": 5.34375, + "learning_rate": 9.232129937930194e-06, + "loss": 1.71982594, + "memory(GiB)": 111.15, + "step": 17330, + "train_speed(iter/s)": 0.448647 + }, + { + "acc": 0.65267391, + "epoch": 0.4397513952308473, + "grad_norm": 4.875, + "learning_rate": 9.231571446058047e-06, + "loss": 1.69477425, + "memory(GiB)": 111.15, + "step": 17335, + "train_speed(iter/s)": 0.448716 + }, + { + "acc": 0.64612331, + "epoch": 0.4398782343987823, + "grad_norm": 4.4375, + "learning_rate": 9.231012768062497e-06, + "loss": 1.63679028, + "memory(GiB)": 111.15, + "step": 17340, + "train_speed(iter/s)": 0.448784 + }, + { + "acc": 0.64650884, + "epoch": 0.4400050735667174, + "grad_norm": 4.9375, + "learning_rate": 9.230453903968112e-06, + "loss": 1.66090164, + "memory(GiB)": 111.15, + "step": 17345, + "train_speed(iter/s)": 0.448853 + }, + { + "acc": 0.65463758, + "epoch": 0.44013191273465246, + "grad_norm": 5.625, + "learning_rate": 9.22989485379948e-06, + "loss": 1.58351822, + "memory(GiB)": 111.15, + "step": 17350, + "train_speed(iter/s)": 0.448921 + }, + { + "acc": 0.65135288, + "epoch": 0.4402587519025875, + "grad_norm": 4.875, + "learning_rate": 9.229335617581187e-06, + "loss": 1.54890652, + "memory(GiB)": 111.15, + "step": 17355, + "train_speed(iter/s)": 0.44899 + }, + { + "acc": 0.65185184, + "epoch": 0.44038559107052255, + "grad_norm": 4.9375, + "learning_rate": 9.22877619533783e-06, + "loss": 1.5950943, + "memory(GiB)": 111.15, + "step": 17360, + "train_speed(iter/s)": 0.449058 + }, + { + "acc": 0.6463582, + "epoch": 0.44051243023845765, + "grad_norm": 4.5625, + "learning_rate": 9.228216587094014e-06, + "loss": 1.61898041, + "memory(GiB)": 111.15, + "step": 17365, + "train_speed(iter/s)": 0.449127 + }, + { + "acc": 0.66250005, + "epoch": 0.4406392694063927, + "grad_norm": 6.15625, + "learning_rate": 9.227656792874358e-06, + "loss": 1.57358389, + "memory(GiB)": 111.15, + "step": 17370, + "train_speed(iter/s)": 0.449196 + }, + { + "acc": 0.64613991, + "epoch": 0.44076610857432774, + "grad_norm": 5.78125, + "learning_rate": 9.227096812703479e-06, + "loss": 1.65252304, + "memory(GiB)": 111.15, + "step": 17375, + "train_speed(iter/s)": 0.449264 + }, + { + "acc": 0.6440485, + "epoch": 0.4408929477422628, + "grad_norm": 6.15625, + "learning_rate": 9.22653664660601e-06, + "loss": 1.64063301, + "memory(GiB)": 111.15, + "step": 17380, + "train_speed(iter/s)": 0.449333 + }, + { + "acc": 0.65983057, + "epoch": 0.4410197869101979, + "grad_norm": 4.59375, + "learning_rate": 9.225976294606589e-06, + "loss": 1.64178658, + "memory(GiB)": 111.15, + "step": 17385, + "train_speed(iter/s)": 0.449401 + }, + { + "acc": 0.64442163, + "epoch": 0.44114662607813293, + "grad_norm": 4.96875, + "learning_rate": 9.225415756729863e-06, + "loss": 1.61692619, + "memory(GiB)": 111.15, + "step": 17390, + "train_speed(iter/s)": 0.44947 + }, + { + "acc": 0.63393207, + "epoch": 0.441273465246068, + "grad_norm": 5.625, + "learning_rate": 9.224855033000489e-06, + "loss": 1.66131134, + "memory(GiB)": 111.15, + "step": 17395, + "train_speed(iter/s)": 0.449539 + }, + { + "acc": 0.66388664, + "epoch": 0.441400304414003, + "grad_norm": 4.90625, + "learning_rate": 9.224294123443125e-06, + "loss": 1.58479557, + "memory(GiB)": 111.15, + "step": 17400, + "train_speed(iter/s)": 0.449606 + }, + { + "epoch": 0.441400304414003, + "eval_acc": 0.6403959264345189, + "eval_loss": 1.62440824508667, + "eval_runtime": 113.8729, + "eval_samples_per_second": 55.94, + "eval_steps_per_second": 27.97, + "step": 17400 + }, + { + "acc": 0.65850697, + "epoch": 0.4415271435819381, + "grad_norm": 4.75, + "learning_rate": 9.223733028082447e-06, + "loss": 1.59983654, + "memory(GiB)": 111.15, + "step": 17405, + "train_speed(iter/s)": 0.448303 + }, + { + "acc": 0.64719419, + "epoch": 0.44165398274987316, + "grad_norm": 5.5, + "learning_rate": 9.223171746943132e-06, + "loss": 1.63351135, + "memory(GiB)": 111.15, + "step": 17410, + "train_speed(iter/s)": 0.448372 + }, + { + "acc": 0.6586267, + "epoch": 0.4417808219178082, + "grad_norm": 4.9375, + "learning_rate": 9.222610280049868e-06, + "loss": 1.66458321, + "memory(GiB)": 111.15, + "step": 17415, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.64104862, + "epoch": 0.44190766108574325, + "grad_norm": 5.65625, + "learning_rate": 9.222048627427352e-06, + "loss": 1.70456543, + "memory(GiB)": 111.15, + "step": 17420, + "train_speed(iter/s)": 0.448508 + }, + { + "acc": 0.6553834, + "epoch": 0.44203450025367835, + "grad_norm": 5.4375, + "learning_rate": 9.221486789100288e-06, + "loss": 1.66402206, + "memory(GiB)": 111.15, + "step": 17425, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.63825712, + "epoch": 0.4421613394216134, + "grad_norm": 5.90625, + "learning_rate": 9.220924765093386e-06, + "loss": 1.658078, + "memory(GiB)": 111.15, + "step": 17430, + "train_speed(iter/s)": 0.448644 + }, + { + "acc": 0.63809042, + "epoch": 0.44228817858954844, + "grad_norm": 4.625, + "learning_rate": 9.220362555431369e-06, + "loss": 1.66647434, + "memory(GiB)": 111.15, + "step": 17435, + "train_speed(iter/s)": 0.448713 + }, + { + "acc": 0.63973818, + "epoch": 0.4424150177574835, + "grad_norm": 4.8125, + "learning_rate": 9.219800160138964e-06, + "loss": 1.60111866, + "memory(GiB)": 111.15, + "step": 17440, + "train_speed(iter/s)": 0.448781 + }, + { + "acc": 0.64286842, + "epoch": 0.4425418569254186, + "grad_norm": 5.96875, + "learning_rate": 9.219237579240907e-06, + "loss": 1.63364887, + "memory(GiB)": 111.15, + "step": 17445, + "train_speed(iter/s)": 0.448849 + }, + { + "acc": 0.65482368, + "epoch": 0.44266869609335363, + "grad_norm": 5.34375, + "learning_rate": 9.218674812761946e-06, + "loss": 1.62116699, + "memory(GiB)": 111.15, + "step": 17450, + "train_speed(iter/s)": 0.448917 + }, + { + "acc": 0.64408908, + "epoch": 0.4427955352612887, + "grad_norm": 4.5, + "learning_rate": 9.21811186072683e-06, + "loss": 1.61038818, + "memory(GiB)": 111.15, + "step": 17455, + "train_speed(iter/s)": 0.448985 + }, + { + "acc": 0.65263801, + "epoch": 0.4429223744292237, + "grad_norm": 5.75, + "learning_rate": 9.21754872316032e-06, + "loss": 1.62329941, + "memory(GiB)": 111.15, + "step": 17460, + "train_speed(iter/s)": 0.449054 + }, + { + "acc": 0.67120581, + "epoch": 0.4430492135971588, + "grad_norm": 6.375, + "learning_rate": 9.21698540008719e-06, + "loss": 1.57654953, + "memory(GiB)": 111.15, + "step": 17465, + "train_speed(iter/s)": 0.449122 + }, + { + "acc": 0.64745283, + "epoch": 0.44317605276509386, + "grad_norm": 5.4375, + "learning_rate": 9.216421891532214e-06, + "loss": 1.62022858, + "memory(GiB)": 111.15, + "step": 17470, + "train_speed(iter/s)": 0.44919 + }, + { + "acc": 0.66475639, + "epoch": 0.4433028919330289, + "grad_norm": 5.65625, + "learning_rate": 9.215858197520178e-06, + "loss": 1.64729557, + "memory(GiB)": 111.15, + "step": 17475, + "train_speed(iter/s)": 0.449257 + }, + { + "acc": 0.65051727, + "epoch": 0.44342973110096395, + "grad_norm": 5.59375, + "learning_rate": 9.215294318075876e-06, + "loss": 1.63894806, + "memory(GiB)": 111.15, + "step": 17480, + "train_speed(iter/s)": 0.449326 + }, + { + "acc": 0.6567553, + "epoch": 0.44355657026889905, + "grad_norm": 5.71875, + "learning_rate": 9.21473025322411e-06, + "loss": 1.60424232, + "memory(GiB)": 111.15, + "step": 17485, + "train_speed(iter/s)": 0.449395 + }, + { + "acc": 0.64634748, + "epoch": 0.4436834094368341, + "grad_norm": 4.90625, + "learning_rate": 9.21416600298969e-06, + "loss": 1.67242088, + "memory(GiB)": 111.15, + "step": 17490, + "train_speed(iter/s)": 0.449464 + }, + { + "acc": 0.64441948, + "epoch": 0.44381024860476914, + "grad_norm": 5.21875, + "learning_rate": 9.213601567397434e-06, + "loss": 1.57439709, + "memory(GiB)": 111.15, + "step": 17495, + "train_speed(iter/s)": 0.449532 + }, + { + "acc": 0.63889122, + "epoch": 0.4439370877727042, + "grad_norm": 6.0625, + "learning_rate": 9.213036946472169e-06, + "loss": 1.71897106, + "memory(GiB)": 111.15, + "step": 17500, + "train_speed(iter/s)": 0.449601 + }, + { + "epoch": 0.4439370877727042, + "eval_acc": 0.6405183322561899, + "eval_loss": 1.623980164527893, + "eval_runtime": 112.9493, + "eval_samples_per_second": 56.397, + "eval_steps_per_second": 28.198, + "step": 17500 + }, + { + "acc": 0.63506665, + "epoch": 0.4440639269406393, + "grad_norm": 5.15625, + "learning_rate": 9.212472140238729e-06, + "loss": 1.64957829, + "memory(GiB)": 111.15, + "step": 17505, + "train_speed(iter/s)": 0.448317 + }, + { + "acc": 0.64288025, + "epoch": 0.44419076610857433, + "grad_norm": 5.75, + "learning_rate": 9.211907148721958e-06, + "loss": 1.65521431, + "memory(GiB)": 111.15, + "step": 17510, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.65054607, + "epoch": 0.4443176052765094, + "grad_norm": 4.78125, + "learning_rate": 9.211341971946705e-06, + "loss": 1.63157272, + "memory(GiB)": 111.15, + "step": 17515, + "train_speed(iter/s)": 0.448455 + }, + { + "acc": 0.66537852, + "epoch": 0.4444444444444444, + "grad_norm": 4.96875, + "learning_rate": 9.21077660993783e-06, + "loss": 1.615765, + "memory(GiB)": 111.15, + "step": 17520, + "train_speed(iter/s)": 0.448524 + }, + { + "acc": 0.65197239, + "epoch": 0.4445712836123795, + "grad_norm": 5.3125, + "learning_rate": 9.210211062720198e-06, + "loss": 1.68479404, + "memory(GiB)": 111.15, + "step": 17525, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.6300189, + "epoch": 0.44469812278031456, + "grad_norm": 4.75, + "learning_rate": 9.209645330318689e-06, + "loss": 1.71307297, + "memory(GiB)": 111.15, + "step": 17530, + "train_speed(iter/s)": 0.448661 + }, + { + "acc": 0.64983711, + "epoch": 0.4448249619482496, + "grad_norm": 5.9375, + "learning_rate": 9.209079412758183e-06, + "loss": 1.66355667, + "memory(GiB)": 111.15, + "step": 17535, + "train_speed(iter/s)": 0.44873 + }, + { + "acc": 0.64501028, + "epoch": 0.44495180111618465, + "grad_norm": 5.75, + "learning_rate": 9.208513310063572e-06, + "loss": 1.67153988, + "memory(GiB)": 111.15, + "step": 17540, + "train_speed(iter/s)": 0.448799 + }, + { + "acc": 0.65297213, + "epoch": 0.44507864028411975, + "grad_norm": 6.40625, + "learning_rate": 9.207947022259755e-06, + "loss": 1.63022537, + "memory(GiB)": 111.15, + "step": 17545, + "train_speed(iter/s)": 0.448868 + }, + { + "acc": 0.66059427, + "epoch": 0.4452054794520548, + "grad_norm": 6.1875, + "learning_rate": 9.207380549371642e-06, + "loss": 1.59212646, + "memory(GiB)": 111.15, + "step": 17550, + "train_speed(iter/s)": 0.448937 + }, + { + "acc": 0.64103365, + "epoch": 0.44533231861998984, + "grad_norm": 6.03125, + "learning_rate": 9.206813891424147e-06, + "loss": 1.69040985, + "memory(GiB)": 111.15, + "step": 17555, + "train_speed(iter/s)": 0.449005 + }, + { + "acc": 0.63632727, + "epoch": 0.4454591577879249, + "grad_norm": 4.96875, + "learning_rate": 9.206247048442196e-06, + "loss": 1.703442, + "memory(GiB)": 111.15, + "step": 17560, + "train_speed(iter/s)": 0.449074 + }, + { + "acc": 0.64868164, + "epoch": 0.44558599695586, + "grad_norm": 4.875, + "learning_rate": 9.20568002045072e-06, + "loss": 1.65853691, + "memory(GiB)": 111.15, + "step": 17565, + "train_speed(iter/s)": 0.449144 + }, + { + "acc": 0.63786397, + "epoch": 0.44571283612379503, + "grad_norm": 4.625, + "learning_rate": 9.20511280747466e-06, + "loss": 1.64694061, + "memory(GiB)": 111.15, + "step": 17570, + "train_speed(iter/s)": 0.449213 + }, + { + "acc": 0.65984659, + "epoch": 0.4458396752917301, + "grad_norm": 4.40625, + "learning_rate": 9.204545409538962e-06, + "loss": 1.57792387, + "memory(GiB)": 111.15, + "step": 17575, + "train_speed(iter/s)": 0.449282 + }, + { + "acc": 0.63484135, + "epoch": 0.4459665144596651, + "grad_norm": 5.0, + "learning_rate": 9.203977826668587e-06, + "loss": 1.74068489, + "memory(GiB)": 111.15, + "step": 17580, + "train_speed(iter/s)": 0.449351 + }, + { + "acc": 0.66680708, + "epoch": 0.4460933536276002, + "grad_norm": 5.15625, + "learning_rate": 9.203410058888498e-06, + "loss": 1.58136053, + "memory(GiB)": 111.15, + "step": 17585, + "train_speed(iter/s)": 0.449419 + }, + { + "acc": 0.6496191, + "epoch": 0.44622019279553526, + "grad_norm": 5.40625, + "learning_rate": 9.202842106223667e-06, + "loss": 1.65004292, + "memory(GiB)": 111.15, + "step": 17590, + "train_speed(iter/s)": 0.449487 + }, + { + "acc": 0.63693981, + "epoch": 0.4463470319634703, + "grad_norm": 5.5625, + "learning_rate": 9.202273968699075e-06, + "loss": 1.7075491, + "memory(GiB)": 111.15, + "step": 17595, + "train_speed(iter/s)": 0.449556 + }, + { + "acc": 0.66062627, + "epoch": 0.44647387113140535, + "grad_norm": 7.09375, + "learning_rate": 9.201705646339714e-06, + "loss": 1.65188828, + "memory(GiB)": 111.15, + "step": 17600, + "train_speed(iter/s)": 0.449626 + }, + { + "epoch": 0.44647387113140535, + "eval_acc": 0.6404506539520579, + "eval_loss": 1.6242213249206543, + "eval_runtime": 113.5679, + "eval_samples_per_second": 56.09, + "eval_steps_per_second": 28.045, + "step": 17600 + }, + { + "acc": 0.64620657, + "epoch": 0.44660071029934045, + "grad_norm": 4.625, + "learning_rate": 9.201137139170578e-06, + "loss": 1.59510727, + "memory(GiB)": 111.15, + "step": 17605, + "train_speed(iter/s)": 0.448341 + }, + { + "acc": 0.65623522, + "epoch": 0.4467275494672755, + "grad_norm": 4.875, + "learning_rate": 9.200568447216673e-06, + "loss": 1.60194702, + "memory(GiB)": 111.15, + "step": 17610, + "train_speed(iter/s)": 0.448409 + }, + { + "acc": 0.6442317, + "epoch": 0.44685438863521054, + "grad_norm": 4.71875, + "learning_rate": 9.199999570503015e-06, + "loss": 1.60527325, + "memory(GiB)": 111.15, + "step": 17615, + "train_speed(iter/s)": 0.448476 + }, + { + "acc": 0.62959371, + "epoch": 0.4469812278031456, + "grad_norm": 4.28125, + "learning_rate": 9.199430509054625e-06, + "loss": 1.78597984, + "memory(GiB)": 111.15, + "step": 17620, + "train_speed(iter/s)": 0.448544 + }, + { + "acc": 0.64660616, + "epoch": 0.4471080669710807, + "grad_norm": 5.71875, + "learning_rate": 9.19886126289653e-06, + "loss": 1.61263046, + "memory(GiB)": 111.15, + "step": 17625, + "train_speed(iter/s)": 0.448611 + }, + { + "acc": 0.65171008, + "epoch": 0.44723490613901573, + "grad_norm": 4.9375, + "learning_rate": 9.198291832053771e-06, + "loss": 1.67300797, + "memory(GiB)": 111.15, + "step": 17630, + "train_speed(iter/s)": 0.448679 + }, + { + "acc": 0.64988079, + "epoch": 0.4473617453069508, + "grad_norm": 5.59375, + "learning_rate": 9.197722216551395e-06, + "loss": 1.65984192, + "memory(GiB)": 111.15, + "step": 17635, + "train_speed(iter/s)": 0.448747 + }, + { + "acc": 0.65121088, + "epoch": 0.4474885844748858, + "grad_norm": 4.21875, + "learning_rate": 9.197152416414452e-06, + "loss": 1.5902524, + "memory(GiB)": 111.15, + "step": 17640, + "train_speed(iter/s)": 0.448815 + }, + { + "acc": 0.63410759, + "epoch": 0.4476154236428209, + "grad_norm": 5.125, + "learning_rate": 9.196582431668007e-06, + "loss": 1.74448967, + "memory(GiB)": 111.15, + "step": 17645, + "train_speed(iter/s)": 0.448883 + }, + { + "acc": 0.65439372, + "epoch": 0.44774226281075596, + "grad_norm": 5.96875, + "learning_rate": 9.196012262337131e-06, + "loss": 1.6173748, + "memory(GiB)": 111.15, + "step": 17650, + "train_speed(iter/s)": 0.44895 + }, + { + "acc": 0.63598108, + "epoch": 0.447869101978691, + "grad_norm": 5.25, + "learning_rate": 9.195441908446902e-06, + "loss": 1.73254776, + "memory(GiB)": 111.15, + "step": 17655, + "train_speed(iter/s)": 0.449018 + }, + { + "acc": 0.65106153, + "epoch": 0.44799594114662605, + "grad_norm": 5.46875, + "learning_rate": 9.194871370022407e-06, + "loss": 1.64211044, + "memory(GiB)": 111.15, + "step": 17660, + "train_speed(iter/s)": 0.449086 + }, + { + "acc": 0.65856009, + "epoch": 0.44812278031456115, + "grad_norm": 5.375, + "learning_rate": 9.194300647088739e-06, + "loss": 1.63055153, + "memory(GiB)": 111.15, + "step": 17665, + "train_speed(iter/s)": 0.449154 + }, + { + "acc": 0.63350105, + "epoch": 0.4482496194824962, + "grad_norm": 4.9375, + "learning_rate": 9.193729739671002e-06, + "loss": 1.66326599, + "memory(GiB)": 111.15, + "step": 17670, + "train_speed(iter/s)": 0.449222 + }, + { + "acc": 0.64212017, + "epoch": 0.44837645865043124, + "grad_norm": 4.375, + "learning_rate": 9.193158647794308e-06, + "loss": 1.64019852, + "memory(GiB)": 111.15, + "step": 17675, + "train_speed(iter/s)": 0.449289 + }, + { + "acc": 0.63937368, + "epoch": 0.4485032978183663, + "grad_norm": 5.46875, + "learning_rate": 9.192587371483777e-06, + "loss": 1.65612164, + "memory(GiB)": 111.15, + "step": 17680, + "train_speed(iter/s)": 0.449357 + }, + { + "acc": 0.63299065, + "epoch": 0.4486301369863014, + "grad_norm": 5.8125, + "learning_rate": 9.192015910764535e-06, + "loss": 1.72685509, + "memory(GiB)": 111.15, + "step": 17685, + "train_speed(iter/s)": 0.449425 + }, + { + "acc": 0.65312276, + "epoch": 0.44875697615423643, + "grad_norm": 5.75, + "learning_rate": 9.191444265661715e-06, + "loss": 1.65709629, + "memory(GiB)": 111.15, + "step": 17690, + "train_speed(iter/s)": 0.449492 + }, + { + "acc": 0.6427393, + "epoch": 0.4488838153221715, + "grad_norm": 4.96875, + "learning_rate": 9.190872436200464e-06, + "loss": 1.67179337, + "memory(GiB)": 111.15, + "step": 17695, + "train_speed(iter/s)": 0.44956 + }, + { + "acc": 0.63243065, + "epoch": 0.4490106544901065, + "grad_norm": 5.46875, + "learning_rate": 9.190300422405933e-06, + "loss": 1.65096474, + "memory(GiB)": 111.15, + "step": 17700, + "train_speed(iter/s)": 0.449628 + }, + { + "epoch": 0.4490106544901065, + "eval_acc": 0.640498279425336, + "eval_loss": 1.6239088773727417, + "eval_runtime": 112.7349, + "eval_samples_per_second": 56.504, + "eval_steps_per_second": 28.252, + "step": 17700 + }, + { + "acc": 0.64403028, + "epoch": 0.4491374936580416, + "grad_norm": 5.03125, + "learning_rate": 9.18972822430328e-06, + "loss": 1.66656342, + "memory(GiB)": 111.15, + "step": 17705, + "train_speed(iter/s)": 0.448359 + }, + { + "acc": 0.65383954, + "epoch": 0.44926433282597666, + "grad_norm": 4.84375, + "learning_rate": 9.189155841917675e-06, + "loss": 1.57767296, + "memory(GiB)": 111.15, + "step": 17710, + "train_speed(iter/s)": 0.448426 + }, + { + "acc": 0.64086895, + "epoch": 0.4493911719939117, + "grad_norm": 5.53125, + "learning_rate": 9.18858327527429e-06, + "loss": 1.6762289, + "memory(GiB)": 111.15, + "step": 17715, + "train_speed(iter/s)": 0.448493 + }, + { + "acc": 0.6326602, + "epoch": 0.44951801116184675, + "grad_norm": 5.15625, + "learning_rate": 9.188010524398314e-06, + "loss": 1.66197205, + "memory(GiB)": 111.15, + "step": 17720, + "train_speed(iter/s)": 0.44856 + }, + { + "acc": 0.67263527, + "epoch": 0.44964485032978185, + "grad_norm": 5.78125, + "learning_rate": 9.187437589314939e-06, + "loss": 1.51285524, + "memory(GiB)": 111.15, + "step": 17725, + "train_speed(iter/s)": 0.448627 + }, + { + "acc": 0.65184336, + "epoch": 0.4497716894977169, + "grad_norm": 5.28125, + "learning_rate": 9.186864470049358e-06, + "loss": 1.58560781, + "memory(GiB)": 111.15, + "step": 17730, + "train_speed(iter/s)": 0.448694 + }, + { + "acc": 0.6573977, + "epoch": 0.44989852866565194, + "grad_norm": 4.5, + "learning_rate": 9.186291166626789e-06, + "loss": 1.61542797, + "memory(GiB)": 111.15, + "step": 17735, + "train_speed(iter/s)": 0.448762 + }, + { + "acc": 0.65138731, + "epoch": 0.450025367833587, + "grad_norm": 5.4375, + "learning_rate": 9.185717679072444e-06, + "loss": 1.6294775, + "memory(GiB)": 111.15, + "step": 17740, + "train_speed(iter/s)": 0.448829 + }, + { + "acc": 0.64548292, + "epoch": 0.4501522070015221, + "grad_norm": 5.0, + "learning_rate": 9.185144007411547e-06, + "loss": 1.64064255, + "memory(GiB)": 111.15, + "step": 17745, + "train_speed(iter/s)": 0.448896 + }, + { + "acc": 0.62352839, + "epoch": 0.45027904616945713, + "grad_norm": 4.71875, + "learning_rate": 9.18457015166933e-06, + "loss": 1.63757572, + "memory(GiB)": 111.15, + "step": 17750, + "train_speed(iter/s)": 0.448963 + }, + { + "acc": 0.65318966, + "epoch": 0.4504058853373922, + "grad_norm": 6.3125, + "learning_rate": 9.183996111871034e-06, + "loss": 1.59947195, + "memory(GiB)": 111.15, + "step": 17755, + "train_speed(iter/s)": 0.449031 + }, + { + "acc": 0.66034698, + "epoch": 0.4505327245053272, + "grad_norm": 5.65625, + "learning_rate": 9.18342188804191e-06, + "loss": 1.67865868, + "memory(GiB)": 111.15, + "step": 17760, + "train_speed(iter/s)": 0.449098 + }, + { + "acc": 0.64079933, + "epoch": 0.4506595636732623, + "grad_norm": 5.5625, + "learning_rate": 9.182847480207215e-06, + "loss": 1.62811966, + "memory(GiB)": 111.15, + "step": 17765, + "train_speed(iter/s)": 0.449166 + }, + { + "acc": 0.65363417, + "epoch": 0.45078640284119736, + "grad_norm": 4.71875, + "learning_rate": 9.182272888392211e-06, + "loss": 1.59235191, + "memory(GiB)": 111.15, + "step": 17770, + "train_speed(iter/s)": 0.449233 + }, + { + "acc": 0.64311705, + "epoch": 0.4509132420091324, + "grad_norm": 5.65625, + "learning_rate": 9.181698112622175e-06, + "loss": 1.63919029, + "memory(GiB)": 111.15, + "step": 17775, + "train_speed(iter/s)": 0.449301 + }, + { + "acc": 0.62075005, + "epoch": 0.45104008117706745, + "grad_norm": 5.03125, + "learning_rate": 9.181123152922384e-06, + "loss": 1.67774334, + "memory(GiB)": 111.15, + "step": 17780, + "train_speed(iter/s)": 0.449368 + }, + { + "acc": 0.64064837, + "epoch": 0.45116692034500255, + "grad_norm": 4.9375, + "learning_rate": 9.18054800931813e-06, + "loss": 1.66456299, + "memory(GiB)": 111.15, + "step": 17785, + "train_speed(iter/s)": 0.449435 + }, + { + "acc": 0.64071608, + "epoch": 0.4512937595129376, + "grad_norm": 5.15625, + "learning_rate": 9.17997268183471e-06, + "loss": 1.60889549, + "memory(GiB)": 111.15, + "step": 17790, + "train_speed(iter/s)": 0.449503 + }, + { + "acc": 0.62419043, + "epoch": 0.45142059868087264, + "grad_norm": 4.96875, + "learning_rate": 9.17939717049743e-06, + "loss": 1.73013268, + "memory(GiB)": 111.15, + "step": 17795, + "train_speed(iter/s)": 0.44957 + }, + { + "acc": 0.65376396, + "epoch": 0.4515474378488077, + "grad_norm": 6.03125, + "learning_rate": 9.1788214753316e-06, + "loss": 1.65573158, + "memory(GiB)": 111.15, + "step": 17800, + "train_speed(iter/s)": 0.449637 + }, + { + "epoch": 0.4515474378488077, + "eval_acc": 0.6405325363447115, + "eval_loss": 1.6238375902175903, + "eval_runtime": 113.9741, + "eval_samples_per_second": 55.89, + "eval_steps_per_second": 27.945, + "step": 17800 + }, + { + "acc": 0.65548811, + "epoch": 0.4516742770167428, + "grad_norm": 5.71875, + "learning_rate": 9.178245596362546e-06, + "loss": 1.63439941, + "memory(GiB)": 111.15, + "step": 17805, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.64623051, + "epoch": 0.45180111618467783, + "grad_norm": 5.71875, + "learning_rate": 9.177669533615599e-06, + "loss": 1.63227196, + "memory(GiB)": 111.15, + "step": 17810, + "train_speed(iter/s)": 0.448429 + }, + { + "acc": 0.65017509, + "epoch": 0.4519279553526129, + "grad_norm": 4.96875, + "learning_rate": 9.17709328711609e-06, + "loss": 1.58975277, + "memory(GiB)": 111.15, + "step": 17815, + "train_speed(iter/s)": 0.448495 + }, + { + "acc": 0.63670368, + "epoch": 0.4520547945205479, + "grad_norm": 4.875, + "learning_rate": 9.17651685688937e-06, + "loss": 1.67137413, + "memory(GiB)": 111.15, + "step": 17820, + "train_speed(iter/s)": 0.448562 + }, + { + "acc": 0.64195786, + "epoch": 0.452181633688483, + "grad_norm": 5.71875, + "learning_rate": 9.175940242960792e-06, + "loss": 1.66914978, + "memory(GiB)": 111.15, + "step": 17825, + "train_speed(iter/s)": 0.448629 + }, + { + "acc": 0.64658833, + "epoch": 0.45230847285641806, + "grad_norm": 4.8125, + "learning_rate": 9.175363445355718e-06, + "loss": 1.69142265, + "memory(GiB)": 111.15, + "step": 17830, + "train_speed(iter/s)": 0.448696 + }, + { + "acc": 0.656566, + "epoch": 0.4524353120243531, + "grad_norm": 5.59375, + "learning_rate": 9.174786464099519e-06, + "loss": 1.62445297, + "memory(GiB)": 111.15, + "step": 17835, + "train_speed(iter/s)": 0.448762 + }, + { + "acc": 0.63754616, + "epoch": 0.45256215119228815, + "grad_norm": 5.40625, + "learning_rate": 9.17420929921757e-06, + "loss": 1.66258202, + "memory(GiB)": 111.15, + "step": 17840, + "train_speed(iter/s)": 0.448829 + }, + { + "acc": 0.66223087, + "epoch": 0.45268899036022325, + "grad_norm": 5.4375, + "learning_rate": 9.173631950735262e-06, + "loss": 1.57992363, + "memory(GiB)": 111.15, + "step": 17845, + "train_speed(iter/s)": 0.448896 + }, + { + "acc": 0.65069013, + "epoch": 0.4528158295281583, + "grad_norm": 5.4375, + "learning_rate": 9.173054418677985e-06, + "loss": 1.66575012, + "memory(GiB)": 111.15, + "step": 17850, + "train_speed(iter/s)": 0.448963 + }, + { + "acc": 0.65472178, + "epoch": 0.45294266869609334, + "grad_norm": 6.25, + "learning_rate": 9.172476703071145e-06, + "loss": 1.61576538, + "memory(GiB)": 111.15, + "step": 17855, + "train_speed(iter/s)": 0.44903 + }, + { + "acc": 0.65945024, + "epoch": 0.4530695078640284, + "grad_norm": 5.15625, + "learning_rate": 9.171898803940148e-06, + "loss": 1.56042118, + "memory(GiB)": 111.15, + "step": 17860, + "train_speed(iter/s)": 0.449097 + }, + { + "acc": 0.65726843, + "epoch": 0.4531963470319635, + "grad_norm": 4.5625, + "learning_rate": 9.17132072131042e-06, + "loss": 1.58314571, + "memory(GiB)": 111.15, + "step": 17865, + "train_speed(iter/s)": 0.449164 + }, + { + "acc": 0.64883142, + "epoch": 0.45332318619989853, + "grad_norm": 5.65625, + "learning_rate": 9.170742455207378e-06, + "loss": 1.63162174, + "memory(GiB)": 111.15, + "step": 17870, + "train_speed(iter/s)": 0.449231 + }, + { + "acc": 0.64885511, + "epoch": 0.4534500253678336, + "grad_norm": 6.125, + "learning_rate": 9.170164005656465e-06, + "loss": 1.62406502, + "memory(GiB)": 111.15, + "step": 17875, + "train_speed(iter/s)": 0.449297 + }, + { + "acc": 0.66215086, + "epoch": 0.4535768645357686, + "grad_norm": 5.375, + "learning_rate": 9.169585372683118e-06, + "loss": 1.50632133, + "memory(GiB)": 111.15, + "step": 17880, + "train_speed(iter/s)": 0.449363 + }, + { + "acc": 0.65495596, + "epoch": 0.4537037037037037, + "grad_norm": 5.53125, + "learning_rate": 9.169006556312794e-06, + "loss": 1.63736, + "memory(GiB)": 111.15, + "step": 17885, + "train_speed(iter/s)": 0.44943 + }, + { + "acc": 0.63862047, + "epoch": 0.45383054287163876, + "grad_norm": 5.5, + "learning_rate": 9.168427556570946e-06, + "loss": 1.68051605, + "memory(GiB)": 111.15, + "step": 17890, + "train_speed(iter/s)": 0.449496 + }, + { + "acc": 0.66267829, + "epoch": 0.4539573820395738, + "grad_norm": 5.5, + "learning_rate": 9.167848373483044e-06, + "loss": 1.61647301, + "memory(GiB)": 111.15, + "step": 17895, + "train_speed(iter/s)": 0.449563 + }, + { + "acc": 0.649401, + "epoch": 0.45408422120750885, + "grad_norm": 5.0, + "learning_rate": 9.167269007074561e-06, + "loss": 1.65486622, + "memory(GiB)": 111.15, + "step": 17900, + "train_speed(iter/s)": 0.449628 + }, + { + "epoch": 0.45408422120750885, + "eval_acc": 0.6406048100892476, + "eval_loss": 1.6238631010055542, + "eval_runtime": 113.4548, + "eval_samples_per_second": 56.146, + "eval_steps_per_second": 28.073, + "step": 17900 + }, + { + "acc": 0.64790087, + "epoch": 0.45421106037544395, + "grad_norm": 4.125, + "learning_rate": 9.166689457370983e-06, + "loss": 1.62117558, + "memory(GiB)": 111.15, + "step": 17905, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.62897086, + "epoch": 0.454337899543379, + "grad_norm": 5.1875, + "learning_rate": 9.166109724397801e-06, + "loss": 1.666782, + "memory(GiB)": 111.15, + "step": 17910, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.64613833, + "epoch": 0.45446473871131404, + "grad_norm": 6.0625, + "learning_rate": 9.165529808180511e-06, + "loss": 1.62342339, + "memory(GiB)": 111.15, + "step": 17915, + "train_speed(iter/s)": 0.4485 + }, + { + "acc": 0.64871936, + "epoch": 0.4545915778792491, + "grad_norm": 4.625, + "learning_rate": 9.164949708744622e-06, + "loss": 1.64257069, + "memory(GiB)": 111.15, + "step": 17920, + "train_speed(iter/s)": 0.448566 + }, + { + "acc": 0.64706926, + "epoch": 0.4547184170471842, + "grad_norm": 6.96875, + "learning_rate": 9.164369426115652e-06, + "loss": 1.58816538, + "memory(GiB)": 111.15, + "step": 17925, + "train_speed(iter/s)": 0.448632 + }, + { + "acc": 0.64860559, + "epoch": 0.45484525621511923, + "grad_norm": 5.125, + "learning_rate": 9.16378896031912e-06, + "loss": 1.67462597, + "memory(GiB)": 111.15, + "step": 17930, + "train_speed(iter/s)": 0.448698 + }, + { + "acc": 0.64833298, + "epoch": 0.4549720953830543, + "grad_norm": 5.5, + "learning_rate": 9.163208311380561e-06, + "loss": 1.62704239, + "memory(GiB)": 111.15, + "step": 17935, + "train_speed(iter/s)": 0.448763 + }, + { + "acc": 0.65435486, + "epoch": 0.4550989345509893, + "grad_norm": 4.59375, + "learning_rate": 9.162627479325511e-06, + "loss": 1.60184784, + "memory(GiB)": 111.15, + "step": 17940, + "train_speed(iter/s)": 0.448828 + }, + { + "acc": 0.65640984, + "epoch": 0.4552257737189244, + "grad_norm": 6.09375, + "learning_rate": 9.16204646417952e-06, + "loss": 1.61401539, + "memory(GiB)": 111.15, + "step": 17945, + "train_speed(iter/s)": 0.448894 + }, + { + "acc": 0.65014968, + "epoch": 0.45535261288685946, + "grad_norm": 4.78125, + "learning_rate": 9.161465265968145e-06, + "loss": 1.5764678, + "memory(GiB)": 111.15, + "step": 17950, + "train_speed(iter/s)": 0.448959 + }, + { + "acc": 0.66462374, + "epoch": 0.4554794520547945, + "grad_norm": 6.0625, + "learning_rate": 9.160883884716948e-06, + "loss": 1.5784709, + "memory(GiB)": 111.15, + "step": 17955, + "train_speed(iter/s)": 0.449025 + }, + { + "acc": 0.64806547, + "epoch": 0.45560629122272955, + "grad_norm": 4.6875, + "learning_rate": 9.1603023204515e-06, + "loss": 1.65666027, + "memory(GiB)": 111.15, + "step": 17960, + "train_speed(iter/s)": 0.449091 + }, + { + "acc": 0.63582096, + "epoch": 0.45573313039066465, + "grad_norm": 4.5, + "learning_rate": 9.15972057319738e-06, + "loss": 1.64318142, + "memory(GiB)": 111.15, + "step": 17965, + "train_speed(iter/s)": 0.449156 + }, + { + "acc": 0.65820394, + "epoch": 0.4558599695585997, + "grad_norm": 5.84375, + "learning_rate": 9.159138642980178e-06, + "loss": 1.64969425, + "memory(GiB)": 111.15, + "step": 17970, + "train_speed(iter/s)": 0.449222 + }, + { + "acc": 0.64330144, + "epoch": 0.45598680872653474, + "grad_norm": 6.5, + "learning_rate": 9.15855652982549e-06, + "loss": 1.67619934, + "memory(GiB)": 111.15, + "step": 17975, + "train_speed(iter/s)": 0.449288 + }, + { + "acc": 0.65490036, + "epoch": 0.4561136478944698, + "grad_norm": 7.40625, + "learning_rate": 9.15797423375892e-06, + "loss": 1.64396267, + "memory(GiB)": 111.15, + "step": 17980, + "train_speed(iter/s)": 0.449353 + }, + { + "acc": 0.64768448, + "epoch": 0.4562404870624049, + "grad_norm": 6.125, + "learning_rate": 9.157391754806079e-06, + "loss": 1.67707748, + "memory(GiB)": 111.15, + "step": 17985, + "train_speed(iter/s)": 0.449419 + }, + { + "acc": 0.64810991, + "epoch": 0.45636732623033993, + "grad_norm": 5.5625, + "learning_rate": 9.156809092992588e-06, + "loss": 1.61114922, + "memory(GiB)": 111.15, + "step": 17990, + "train_speed(iter/s)": 0.449484 + }, + { + "acc": 0.64348583, + "epoch": 0.456494165398275, + "grad_norm": 5.75, + "learning_rate": 9.156226248344072e-06, + "loss": 1.63893051, + "memory(GiB)": 111.15, + "step": 17995, + "train_speed(iter/s)": 0.44955 + }, + { + "acc": 0.66405864, + "epoch": 0.45662100456621, + "grad_norm": 4.71875, + "learning_rate": 9.15564322088617e-06, + "loss": 1.61182976, + "memory(GiB)": 111.15, + "step": 18000, + "train_speed(iter/s)": 0.449615 + }, + { + "epoch": 0.45662100456621, + "eval_acc": 0.6405985435796058, + "eval_loss": 1.623692274093628, + "eval_runtime": 114.6378, + "eval_samples_per_second": 55.566, + "eval_steps_per_second": 27.783, + "step": 18000 + }, + { + "acc": 0.64258127, + "epoch": 0.4567478437341451, + "grad_norm": 5.21875, + "learning_rate": 9.155060010644525e-06, + "loss": 1.64468613, + "memory(GiB)": 111.15, + "step": 18005, + "train_speed(iter/s)": 0.448346 + }, + { + "acc": 0.65395021, + "epoch": 0.45687468290208016, + "grad_norm": 6.15625, + "learning_rate": 9.154476617644792e-06, + "loss": 1.58812037, + "memory(GiB)": 111.15, + "step": 18010, + "train_speed(iter/s)": 0.448411 + }, + { + "acc": 0.64340105, + "epoch": 0.4570015220700152, + "grad_norm": 6.40625, + "learning_rate": 9.153893041912627e-06, + "loss": 1.569067, + "memory(GiB)": 111.15, + "step": 18015, + "train_speed(iter/s)": 0.448477 + }, + { + "acc": 0.66478567, + "epoch": 0.45712836123795025, + "grad_norm": 5.71875, + "learning_rate": 9.1533092834737e-06, + "loss": 1.5848959, + "memory(GiB)": 111.15, + "step": 18020, + "train_speed(iter/s)": 0.448543 + }, + { + "acc": 0.64590898, + "epoch": 0.45725520040588535, + "grad_norm": 5.15625, + "learning_rate": 9.152725342353688e-06, + "loss": 1.65465813, + "memory(GiB)": 111.15, + "step": 18025, + "train_speed(iter/s)": 0.448609 + }, + { + "acc": 0.63304882, + "epoch": 0.4573820395738204, + "grad_norm": 6.3125, + "learning_rate": 9.152141218578276e-06, + "loss": 1.69629402, + "memory(GiB)": 111.15, + "step": 18030, + "train_speed(iter/s)": 0.448675 + }, + { + "acc": 0.64024215, + "epoch": 0.45750887874175544, + "grad_norm": 5.15625, + "learning_rate": 9.151556912173154e-06, + "loss": 1.70532913, + "memory(GiB)": 111.15, + "step": 18035, + "train_speed(iter/s)": 0.44874 + }, + { + "acc": 0.64752245, + "epoch": 0.4576357179096905, + "grad_norm": 6.5625, + "learning_rate": 9.150972423164024e-06, + "loss": 1.64980774, + "memory(GiB)": 111.15, + "step": 18040, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.64653583, + "epoch": 0.4577625570776256, + "grad_norm": 6.21875, + "learning_rate": 9.150387751576594e-06, + "loss": 1.68800659, + "memory(GiB)": 111.15, + "step": 18045, + "train_speed(iter/s)": 0.448872 + }, + { + "acc": 0.64302607, + "epoch": 0.45788939624556063, + "grad_norm": 4.15625, + "learning_rate": 9.14980289743658e-06, + "loss": 1.66444874, + "memory(GiB)": 111.15, + "step": 18050, + "train_speed(iter/s)": 0.448937 + }, + { + "acc": 0.63830252, + "epoch": 0.4580162354134957, + "grad_norm": 5.5, + "learning_rate": 9.149217860769708e-06, + "loss": 1.68476601, + "memory(GiB)": 111.15, + "step": 18055, + "train_speed(iter/s)": 0.449003 + }, + { + "acc": 0.65752954, + "epoch": 0.4581430745814307, + "grad_norm": 5.875, + "learning_rate": 9.14863264160171e-06, + "loss": 1.57996426, + "memory(GiB)": 111.15, + "step": 18060, + "train_speed(iter/s)": 0.449069 + }, + { + "acc": 0.64375615, + "epoch": 0.4582699137493658, + "grad_norm": 5.40625, + "learning_rate": 9.148047239958325e-06, + "loss": 1.72266846, + "memory(GiB)": 111.15, + "step": 18065, + "train_speed(iter/s)": 0.449134 + }, + { + "acc": 0.64375401, + "epoch": 0.45839675291730086, + "grad_norm": 5.71875, + "learning_rate": 9.147461655865302e-06, + "loss": 1.664217, + "memory(GiB)": 111.15, + "step": 18070, + "train_speed(iter/s)": 0.4492 + }, + { + "acc": 0.65756102, + "epoch": 0.4585235920852359, + "grad_norm": 5.59375, + "learning_rate": 9.1468758893484e-06, + "loss": 1.70248566, + "memory(GiB)": 111.15, + "step": 18075, + "train_speed(iter/s)": 0.449267 + }, + { + "acc": 0.66085863, + "epoch": 0.45865043125317095, + "grad_norm": 4.71875, + "learning_rate": 9.14628994043338e-06, + "loss": 1.67144642, + "memory(GiB)": 111.15, + "step": 18080, + "train_speed(iter/s)": 0.449332 + }, + { + "acc": 0.65291948, + "epoch": 0.45877727042110605, + "grad_norm": 5.0625, + "learning_rate": 9.145703809146018e-06, + "loss": 1.67906094, + "memory(GiB)": 111.15, + "step": 18085, + "train_speed(iter/s)": 0.449398 + }, + { + "acc": 0.63214688, + "epoch": 0.4589041095890411, + "grad_norm": 5.3125, + "learning_rate": 9.145117495512092e-06, + "loss": 1.67810135, + "memory(GiB)": 111.15, + "step": 18090, + "train_speed(iter/s)": 0.449464 + }, + { + "acc": 0.64969735, + "epoch": 0.45903094875697614, + "grad_norm": 4.875, + "learning_rate": 9.144530999557393e-06, + "loss": 1.72802029, + "memory(GiB)": 111.15, + "step": 18095, + "train_speed(iter/s)": 0.44953 + }, + { + "acc": 0.65498438, + "epoch": 0.4591577879249112, + "grad_norm": 6.375, + "learning_rate": 9.143944321307718e-06, + "loss": 1.63395271, + "memory(GiB)": 111.15, + "step": 18100, + "train_speed(iter/s)": 0.449596 + }, + { + "epoch": 0.4591577879249112, + "eval_acc": 0.6405814151199181, + "eval_loss": 1.6230018138885498, + "eval_runtime": 113.2441, + "eval_samples_per_second": 56.25, + "eval_steps_per_second": 28.125, + "step": 18100 + }, + { + "acc": 0.6445003, + "epoch": 0.4592846270928463, + "grad_norm": 5.0, + "learning_rate": 9.143357460788868e-06, + "loss": 1.66909409, + "memory(GiB)": 111.15, + "step": 18105, + "train_speed(iter/s)": 0.448349 + }, + { + "acc": 0.6388299, + "epoch": 0.45941146626078133, + "grad_norm": 5.6875, + "learning_rate": 9.14277041802666e-06, + "loss": 1.7392643, + "memory(GiB)": 111.15, + "step": 18110, + "train_speed(iter/s)": 0.448414 + }, + { + "acc": 0.64023056, + "epoch": 0.4595383054287164, + "grad_norm": 5.5, + "learning_rate": 9.142183193046912e-06, + "loss": 1.64861126, + "memory(GiB)": 111.15, + "step": 18115, + "train_speed(iter/s)": 0.448479 + }, + { + "acc": 0.65999527, + "epoch": 0.4596651445966514, + "grad_norm": 4.90625, + "learning_rate": 9.141595785875453e-06, + "loss": 1.56845379, + "memory(GiB)": 111.15, + "step": 18120, + "train_speed(iter/s)": 0.448544 + }, + { + "acc": 0.6430191, + "epoch": 0.4597919837645865, + "grad_norm": 4.46875, + "learning_rate": 9.141008196538122e-06, + "loss": 1.64486713, + "memory(GiB)": 111.15, + "step": 18125, + "train_speed(iter/s)": 0.44861 + }, + { + "acc": 0.63748674, + "epoch": 0.45991882293252156, + "grad_norm": 5.40625, + "learning_rate": 9.14042042506076e-06, + "loss": 1.65775261, + "memory(GiB)": 111.15, + "step": 18130, + "train_speed(iter/s)": 0.448675 + }, + { + "acc": 0.64931374, + "epoch": 0.4600456621004566, + "grad_norm": 6.1875, + "learning_rate": 9.139832471469224e-06, + "loss": 1.620578, + "memory(GiB)": 111.15, + "step": 18135, + "train_speed(iter/s)": 0.448741 + }, + { + "acc": 0.66084557, + "epoch": 0.46017250126839165, + "grad_norm": 4.46875, + "learning_rate": 9.13924433578937e-06, + "loss": 1.60610352, + "memory(GiB)": 111.15, + "step": 18140, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.64961443, + "epoch": 0.46029934043632675, + "grad_norm": 4.8125, + "learning_rate": 9.138656018047074e-06, + "loss": 1.63161659, + "memory(GiB)": 111.15, + "step": 18145, + "train_speed(iter/s)": 0.448871 + }, + { + "acc": 0.6520113, + "epoch": 0.4604261796042618, + "grad_norm": 5.21875, + "learning_rate": 9.138067518268206e-06, + "loss": 1.61394463, + "memory(GiB)": 111.15, + "step": 18150, + "train_speed(iter/s)": 0.448935 + }, + { + "acc": 0.65558057, + "epoch": 0.46055301877219684, + "grad_norm": 4.90625, + "learning_rate": 9.137478836478654e-06, + "loss": 1.62159081, + "memory(GiB)": 111.15, + "step": 18155, + "train_speed(iter/s)": 0.449 + }, + { + "acc": 0.65343904, + "epoch": 0.4606798579401319, + "grad_norm": 5.3125, + "learning_rate": 9.13688997270431e-06, + "loss": 1.62685318, + "memory(GiB)": 111.15, + "step": 18160, + "train_speed(iter/s)": 0.449066 + }, + { + "acc": 0.67230778, + "epoch": 0.460806697108067, + "grad_norm": 5.25, + "learning_rate": 9.136300926971076e-06, + "loss": 1.56804619, + "memory(GiB)": 111.15, + "step": 18165, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.65672522, + "epoch": 0.46093353627600203, + "grad_norm": 4.625, + "learning_rate": 9.135711699304858e-06, + "loss": 1.69412727, + "memory(GiB)": 111.15, + "step": 18170, + "train_speed(iter/s)": 0.449196 + }, + { + "acc": 0.64287281, + "epoch": 0.4610603754439371, + "grad_norm": 5.34375, + "learning_rate": 9.135122289731575e-06, + "loss": 1.63418007, + "memory(GiB)": 111.15, + "step": 18175, + "train_speed(iter/s)": 0.449262 + }, + { + "acc": 0.64696617, + "epoch": 0.4611872146118721, + "grad_norm": 6.03125, + "learning_rate": 9.134532698277154e-06, + "loss": 1.59013243, + "memory(GiB)": 111.15, + "step": 18180, + "train_speed(iter/s)": 0.449328 + }, + { + "acc": 0.65078735, + "epoch": 0.4613140537798072, + "grad_norm": 6.90625, + "learning_rate": 9.133942924967524e-06, + "loss": 1.65923767, + "memory(GiB)": 111.15, + "step": 18185, + "train_speed(iter/s)": 0.449393 + }, + { + "acc": 0.66408453, + "epoch": 0.46144089294774226, + "grad_norm": 7.1875, + "learning_rate": 9.133352969828628e-06, + "loss": 1.57815218, + "memory(GiB)": 111.15, + "step": 18190, + "train_speed(iter/s)": 0.449458 + }, + { + "acc": 0.64419079, + "epoch": 0.4615677321156773, + "grad_norm": 5.0625, + "learning_rate": 9.132762832886416e-06, + "loss": 1.59385462, + "memory(GiB)": 111.15, + "step": 18195, + "train_speed(iter/s)": 0.449523 + }, + { + "acc": 0.64199886, + "epoch": 0.46169457128361235, + "grad_norm": 5.5625, + "learning_rate": 9.13217251416684e-06, + "loss": 1.70415878, + "memory(GiB)": 111.15, + "step": 18200, + "train_speed(iter/s)": 0.449589 + }, + { + "epoch": 0.46169457128361235, + "eval_acc": 0.6406808437395689, + "eval_loss": 1.6233675479888916, + "eval_runtime": 113.4279, + "eval_samples_per_second": 56.159, + "eval_steps_per_second": 28.08, + "step": 18200 + }, + { + "acc": 0.64497814, + "epoch": 0.46182141045154745, + "grad_norm": 6.09375, + "learning_rate": 9.131582013695867e-06, + "loss": 1.63758965, + "memory(GiB)": 111.15, + "step": 18205, + "train_speed(iter/s)": 0.448348 + }, + { + "acc": 0.65781593, + "epoch": 0.4619482496194825, + "grad_norm": 6.0625, + "learning_rate": 9.130991331499474e-06, + "loss": 1.69163628, + "memory(GiB)": 111.15, + "step": 18210, + "train_speed(iter/s)": 0.448413 + }, + { + "acc": 0.66803436, + "epoch": 0.46207508878741754, + "grad_norm": 5.9375, + "learning_rate": 9.130400467603637e-06, + "loss": 1.58227253, + "memory(GiB)": 111.15, + "step": 18215, + "train_speed(iter/s)": 0.448478 + }, + { + "acc": 0.63053923, + "epoch": 0.4622019279553526, + "grad_norm": 5.875, + "learning_rate": 9.129809422034349e-06, + "loss": 1.69475784, + "memory(GiB)": 111.15, + "step": 18220, + "train_speed(iter/s)": 0.448544 + }, + { + "acc": 0.65363598, + "epoch": 0.4623287671232877, + "grad_norm": 6.0625, + "learning_rate": 9.129218194817601e-06, + "loss": 1.68358421, + "memory(GiB)": 111.15, + "step": 18225, + "train_speed(iter/s)": 0.448609 + }, + { + "acc": 0.64460602, + "epoch": 0.46245560629122273, + "grad_norm": 5.21875, + "learning_rate": 9.128626785979404e-06, + "loss": 1.672369, + "memory(GiB)": 111.15, + "step": 18230, + "train_speed(iter/s)": 0.448675 + }, + { + "acc": 0.63654146, + "epoch": 0.4625824454591578, + "grad_norm": 5.4375, + "learning_rate": 9.128035195545766e-06, + "loss": 1.66236458, + "memory(GiB)": 111.15, + "step": 18235, + "train_speed(iter/s)": 0.44874 + }, + { + "acc": 0.63425946, + "epoch": 0.4627092846270928, + "grad_norm": 5.40625, + "learning_rate": 9.12744342354271e-06, + "loss": 1.74656162, + "memory(GiB)": 111.15, + "step": 18240, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.65481329, + "epoch": 0.4628361237950279, + "grad_norm": 5.875, + "learning_rate": 9.126851469996265e-06, + "loss": 1.62454433, + "memory(GiB)": 111.15, + "step": 18245, + "train_speed(iter/s)": 0.448871 + }, + { + "acc": 0.6440196, + "epoch": 0.46296296296296297, + "grad_norm": 5.5, + "learning_rate": 9.126259334932467e-06, + "loss": 1.63647861, + "memory(GiB)": 111.15, + "step": 18250, + "train_speed(iter/s)": 0.448937 + }, + { + "acc": 0.63573875, + "epoch": 0.463089802130898, + "grad_norm": 6.71875, + "learning_rate": 9.125667018377362e-06, + "loss": 1.7456625, + "memory(GiB)": 111.15, + "step": 18255, + "train_speed(iter/s)": 0.449004 + }, + { + "acc": 0.65065207, + "epoch": 0.46321664129883305, + "grad_norm": 6.28125, + "learning_rate": 9.125074520357002e-06, + "loss": 1.63900185, + "memory(GiB)": 111.15, + "step": 18260, + "train_speed(iter/s)": 0.449069 + }, + { + "acc": 0.6473959, + "epoch": 0.46334348046676815, + "grad_norm": 5.03125, + "learning_rate": 9.124481840897446e-06, + "loss": 1.60962448, + "memory(GiB)": 111.15, + "step": 18265, + "train_speed(iter/s)": 0.449135 + }, + { + "acc": 0.64412975, + "epoch": 0.4634703196347032, + "grad_norm": 4.875, + "learning_rate": 9.123888980024765e-06, + "loss": 1.6232233, + "memory(GiB)": 111.15, + "step": 18270, + "train_speed(iter/s)": 0.449199 + }, + { + "acc": 0.66579037, + "epoch": 0.46359715880263824, + "grad_norm": 4.5, + "learning_rate": 9.123295937765034e-06, + "loss": 1.58583975, + "memory(GiB)": 111.15, + "step": 18275, + "train_speed(iter/s)": 0.449265 + }, + { + "acc": 0.67146401, + "epoch": 0.4637239979705733, + "grad_norm": 5.8125, + "learning_rate": 9.122702714144339e-06, + "loss": 1.55856733, + "memory(GiB)": 111.15, + "step": 18280, + "train_speed(iter/s)": 0.44933 + }, + { + "acc": 0.64378691, + "epoch": 0.4638508371385084, + "grad_norm": 5.25, + "learning_rate": 9.122109309188772e-06, + "loss": 1.63967552, + "memory(GiB)": 111.15, + "step": 18285, + "train_speed(iter/s)": 0.449396 + }, + { + "acc": 0.66831989, + "epoch": 0.46397767630644343, + "grad_norm": 6.625, + "learning_rate": 9.121515722924435e-06, + "loss": 1.60099583, + "memory(GiB)": 111.15, + "step": 18290, + "train_speed(iter/s)": 0.449463 + }, + { + "acc": 0.65531921, + "epoch": 0.4641045154743785, + "grad_norm": 5.1875, + "learning_rate": 9.120921955377433e-06, + "loss": 1.5971077, + "memory(GiB)": 111.15, + "step": 18295, + "train_speed(iter/s)": 0.449529 + }, + { + "acc": 0.64050837, + "epoch": 0.4642313546423135, + "grad_norm": 5.0, + "learning_rate": 9.120328006573887e-06, + "loss": 1.62931023, + "memory(GiB)": 111.15, + "step": 18300, + "train_speed(iter/s)": 0.449594 + }, + { + "epoch": 0.4642313546423135, + "eval_acc": 0.6406165075739124, + "eval_loss": 1.6226986646652222, + "eval_runtime": 113.5654, + "eval_samples_per_second": 56.091, + "eval_steps_per_second": 28.046, + "step": 18300 + }, + { + "acc": 0.65506487, + "epoch": 0.4643581938102486, + "grad_norm": 5.21875, + "learning_rate": 9.119733876539916e-06, + "loss": 1.55712833, + "memory(GiB)": 111.15, + "step": 18305, + "train_speed(iter/s)": 0.448358 + }, + { + "acc": 0.64307866, + "epoch": 0.46448503297818367, + "grad_norm": 5.5, + "learning_rate": 9.119139565301658e-06, + "loss": 1.61205883, + "memory(GiB)": 111.15, + "step": 18310, + "train_speed(iter/s)": 0.448424 + }, + { + "acc": 0.65227137, + "epoch": 0.4646118721461187, + "grad_norm": 5.40625, + "learning_rate": 9.118545072885253e-06, + "loss": 1.6866869, + "memory(GiB)": 111.15, + "step": 18315, + "train_speed(iter/s)": 0.448489 + }, + { + "acc": 0.65983257, + "epoch": 0.46473871131405375, + "grad_norm": 6.4375, + "learning_rate": 9.117950399316845e-06, + "loss": 1.59319592, + "memory(GiB)": 111.15, + "step": 18320, + "train_speed(iter/s)": 0.448554 + }, + { + "acc": 0.66612692, + "epoch": 0.46486555048198885, + "grad_norm": 5.90625, + "learning_rate": 9.117355544622595e-06, + "loss": 1.59368172, + "memory(GiB)": 111.15, + "step": 18325, + "train_speed(iter/s)": 0.448619 + }, + { + "acc": 0.6472126, + "epoch": 0.4649923896499239, + "grad_norm": 5.03125, + "learning_rate": 9.116760508828664e-06, + "loss": 1.5690135, + "memory(GiB)": 111.15, + "step": 18330, + "train_speed(iter/s)": 0.448684 + }, + { + "acc": 0.66651535, + "epoch": 0.46511922881785894, + "grad_norm": 5.0625, + "learning_rate": 9.116165291961225e-06, + "loss": 1.62407017, + "memory(GiB)": 111.15, + "step": 18335, + "train_speed(iter/s)": 0.448749 + }, + { + "acc": 0.64161787, + "epoch": 0.465246067985794, + "grad_norm": 4.625, + "learning_rate": 9.11556989404646e-06, + "loss": 1.62849617, + "memory(GiB)": 111.15, + "step": 18340, + "train_speed(iter/s)": 0.448814 + }, + { + "acc": 0.63987713, + "epoch": 0.4653729071537291, + "grad_norm": 5.6875, + "learning_rate": 9.114974315110558e-06, + "loss": 1.72187824, + "memory(GiB)": 111.15, + "step": 18345, + "train_speed(iter/s)": 0.448879 + }, + { + "acc": 0.65916214, + "epoch": 0.46549974632166413, + "grad_norm": 5.90625, + "learning_rate": 9.114378555179712e-06, + "loss": 1.64385262, + "memory(GiB)": 111.15, + "step": 18350, + "train_speed(iter/s)": 0.448943 + }, + { + "acc": 0.66123972, + "epoch": 0.4656265854895992, + "grad_norm": 5.53125, + "learning_rate": 9.11378261428013e-06, + "loss": 1.60845757, + "memory(GiB)": 111.15, + "step": 18355, + "train_speed(iter/s)": 0.449008 + }, + { + "acc": 0.65182648, + "epoch": 0.4657534246575342, + "grad_norm": 5.25, + "learning_rate": 9.11318649243802e-06, + "loss": 1.6316618, + "memory(GiB)": 111.15, + "step": 18360, + "train_speed(iter/s)": 0.449074 + }, + { + "acc": 0.65041628, + "epoch": 0.4658802638254693, + "grad_norm": 5.5625, + "learning_rate": 9.112590189679604e-06, + "loss": 1.63320503, + "memory(GiB)": 111.15, + "step": 18365, + "train_speed(iter/s)": 0.449139 + }, + { + "acc": 0.64401913, + "epoch": 0.46600710299340437, + "grad_norm": 4.46875, + "learning_rate": 9.111993706031109e-06, + "loss": 1.6492672, + "memory(GiB)": 111.15, + "step": 18370, + "train_speed(iter/s)": 0.449204 + }, + { + "acc": 0.64830265, + "epoch": 0.4661339421613394, + "grad_norm": 5.03125, + "learning_rate": 9.111397041518774e-06, + "loss": 1.63535061, + "memory(GiB)": 111.15, + "step": 18375, + "train_speed(iter/s)": 0.449269 + }, + { + "acc": 0.64971189, + "epoch": 0.46626078132927445, + "grad_norm": 4.875, + "learning_rate": 9.11080019616884e-06, + "loss": 1.61010475, + "memory(GiB)": 111.15, + "step": 18380, + "train_speed(iter/s)": 0.449334 + }, + { + "acc": 0.65798349, + "epoch": 0.46638762049720955, + "grad_norm": 5.625, + "learning_rate": 9.11020317000756e-06, + "loss": 1.65622673, + "memory(GiB)": 111.15, + "step": 18385, + "train_speed(iter/s)": 0.449399 + }, + { + "acc": 0.62587147, + "epoch": 0.4665144596651446, + "grad_norm": 5.25, + "learning_rate": 9.109605963061194e-06, + "loss": 1.74048958, + "memory(GiB)": 111.15, + "step": 18390, + "train_speed(iter/s)": 0.449464 + }, + { + "acc": 0.65257325, + "epoch": 0.46664129883307964, + "grad_norm": 5.46875, + "learning_rate": 9.10900857535601e-06, + "loss": 1.61184502, + "memory(GiB)": 111.15, + "step": 18395, + "train_speed(iter/s)": 0.449529 + }, + { + "acc": 0.62558956, + "epoch": 0.4667681380010147, + "grad_norm": 5.9375, + "learning_rate": 9.108411006918283e-06, + "loss": 1.68565636, + "memory(GiB)": 111.15, + "step": 18400, + "train_speed(iter/s)": 0.449594 + }, + { + "epoch": 0.4667681380010147, + "eval_acc": 0.6407623083649131, + "eval_loss": 1.6228444576263428, + "eval_runtime": 113.5085, + "eval_samples_per_second": 56.119, + "eval_steps_per_second": 28.06, + "step": 18400 + }, + { + "acc": 0.65425553, + "epoch": 0.4668949771689498, + "grad_norm": 6.03125, + "learning_rate": 9.107813257774298e-06, + "loss": 1.66392746, + "memory(GiB)": 111.15, + "step": 18405, + "train_speed(iter/s)": 0.448366 + }, + { + "acc": 0.64056087, + "epoch": 0.46702181633688483, + "grad_norm": 5.3125, + "learning_rate": 9.107215327950345e-06, + "loss": 1.64805527, + "memory(GiB)": 111.15, + "step": 18410, + "train_speed(iter/s)": 0.44843 + }, + { + "acc": 0.6438386, + "epoch": 0.4671486555048199, + "grad_norm": 5.9375, + "learning_rate": 9.106617217472724e-06, + "loss": 1.6635006, + "memory(GiB)": 111.15, + "step": 18415, + "train_speed(iter/s)": 0.448495 + }, + { + "acc": 0.64418349, + "epoch": 0.4672754946727549, + "grad_norm": 5.21875, + "learning_rate": 9.106018926367744e-06, + "loss": 1.62999268, + "memory(GiB)": 111.15, + "step": 18420, + "train_speed(iter/s)": 0.44856 + }, + { + "acc": 0.64991632, + "epoch": 0.46740233384069, + "grad_norm": 5.5625, + "learning_rate": 9.10542045466172e-06, + "loss": 1.66267109, + "memory(GiB)": 111.15, + "step": 18425, + "train_speed(iter/s)": 0.448624 + }, + { + "acc": 0.65056028, + "epoch": 0.46752917300862507, + "grad_norm": 5.5625, + "learning_rate": 9.104821802380974e-06, + "loss": 1.67134666, + "memory(GiB)": 111.15, + "step": 18430, + "train_speed(iter/s)": 0.448688 + }, + { + "acc": 0.65150704, + "epoch": 0.4676560121765601, + "grad_norm": 4.84375, + "learning_rate": 9.104222969551838e-06, + "loss": 1.57996998, + "memory(GiB)": 111.15, + "step": 18435, + "train_speed(iter/s)": 0.448753 + }, + { + "acc": 0.65227833, + "epoch": 0.46778285134449515, + "grad_norm": 5.5, + "learning_rate": 9.103623956200654e-06, + "loss": 1.656633, + "memory(GiB)": 111.15, + "step": 18440, + "train_speed(iter/s)": 0.448817 + }, + { + "acc": 0.6445262, + "epoch": 0.46790969051243025, + "grad_norm": 5.375, + "learning_rate": 9.103024762353766e-06, + "loss": 1.69572487, + "memory(GiB)": 111.15, + "step": 18445, + "train_speed(iter/s)": 0.44888 + }, + { + "acc": 0.65994658, + "epoch": 0.4680365296803653, + "grad_norm": 4.71875, + "learning_rate": 9.102425388037527e-06, + "loss": 1.58760967, + "memory(GiB)": 111.15, + "step": 18450, + "train_speed(iter/s)": 0.448945 + }, + { + "acc": 0.64596262, + "epoch": 0.46816336884830034, + "grad_norm": 5.4375, + "learning_rate": 9.101825833278308e-06, + "loss": 1.64518871, + "memory(GiB)": 111.15, + "step": 18455, + "train_speed(iter/s)": 0.44901 + }, + { + "acc": 0.654251, + "epoch": 0.4682902080162354, + "grad_norm": 5.0625, + "learning_rate": 9.101226098102473e-06, + "loss": 1.58524752, + "memory(GiB)": 111.15, + "step": 18460, + "train_speed(iter/s)": 0.449074 + }, + { + "acc": 0.65029984, + "epoch": 0.4684170471841705, + "grad_norm": 6.8125, + "learning_rate": 9.100626182536405e-06, + "loss": 1.66436501, + "memory(GiB)": 111.15, + "step": 18465, + "train_speed(iter/s)": 0.449138 + }, + { + "acc": 0.65600443, + "epoch": 0.46854388635210553, + "grad_norm": 5.6875, + "learning_rate": 9.100026086606488e-06, + "loss": 1.59276485, + "memory(GiB)": 111.15, + "step": 18470, + "train_speed(iter/s)": 0.449203 + }, + { + "acc": 0.67464876, + "epoch": 0.4686707255200406, + "grad_norm": 5.09375, + "learning_rate": 9.09942581033912e-06, + "loss": 1.53188086, + "memory(GiB)": 111.15, + "step": 18475, + "train_speed(iter/s)": 0.449267 + }, + { + "acc": 0.66208801, + "epoch": 0.4687975646879756, + "grad_norm": 5.59375, + "learning_rate": 9.0988253537607e-06, + "loss": 1.54502659, + "memory(GiB)": 111.15, + "step": 18480, + "train_speed(iter/s)": 0.449331 + }, + { + "acc": 0.65728588, + "epoch": 0.4689244038559107, + "grad_norm": 4.5, + "learning_rate": 9.098224716897644e-06, + "loss": 1.62895718, + "memory(GiB)": 111.15, + "step": 18485, + "train_speed(iter/s)": 0.449396 + }, + { + "acc": 0.65967865, + "epoch": 0.46905124302384577, + "grad_norm": 5.25, + "learning_rate": 9.097623899776366e-06, + "loss": 1.62379494, + "memory(GiB)": 111.15, + "step": 18490, + "train_speed(iter/s)": 0.449461 + }, + { + "acc": 0.63681393, + "epoch": 0.4691780821917808, + "grad_norm": 5.90625, + "learning_rate": 9.097022902423294e-06, + "loss": 1.62050781, + "memory(GiB)": 111.15, + "step": 18495, + "train_speed(iter/s)": 0.449525 + }, + { + "acc": 0.64813433, + "epoch": 0.46930492135971585, + "grad_norm": 5.0, + "learning_rate": 9.096421724864864e-06, + "loss": 1.64656143, + "memory(GiB)": 111.15, + "step": 18500, + "train_speed(iter/s)": 0.449589 + }, + { + "epoch": 0.46930492135971585, + "eval_acc": 0.6407794368246008, + "eval_loss": 1.622450351715088, + "eval_runtime": 112.2195, + "eval_samples_per_second": 56.764, + "eval_steps_per_second": 28.382, + "step": 18500 + }, + { + "acc": 0.64854393, + "epoch": 0.46943176052765095, + "grad_norm": 4.78125, + "learning_rate": 9.095820367127517e-06, + "loss": 1.65303383, + "memory(GiB)": 111.15, + "step": 18505, + "train_speed(iter/s)": 0.448382 + }, + { + "acc": 0.64208736, + "epoch": 0.469558599695586, + "grad_norm": 4.71875, + "learning_rate": 9.095218829237703e-06, + "loss": 1.65377254, + "memory(GiB)": 111.15, + "step": 18510, + "train_speed(iter/s)": 0.448446 + }, + { + "acc": 0.63350401, + "epoch": 0.46968543886352104, + "grad_norm": 5.34375, + "learning_rate": 9.094617111221881e-06, + "loss": 1.67860489, + "memory(GiB)": 111.15, + "step": 18515, + "train_speed(iter/s)": 0.448511 + }, + { + "acc": 0.64017286, + "epoch": 0.4698122780314561, + "grad_norm": 4.625, + "learning_rate": 9.09401521310652e-06, + "loss": 1.60295906, + "memory(GiB)": 111.15, + "step": 18520, + "train_speed(iter/s)": 0.448576 + }, + { + "acc": 0.66613545, + "epoch": 0.4699391171993912, + "grad_norm": 5.09375, + "learning_rate": 9.093413134918088e-06, + "loss": 1.56817646, + "memory(GiB)": 111.15, + "step": 18525, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.65946789, + "epoch": 0.47006595636732623, + "grad_norm": 5.0, + "learning_rate": 9.09281087668307e-06, + "loss": 1.62598495, + "memory(GiB)": 111.15, + "step": 18530, + "train_speed(iter/s)": 0.448704 + }, + { + "acc": 0.66190076, + "epoch": 0.4701927955352613, + "grad_norm": 5.375, + "learning_rate": 9.09220843842796e-06, + "loss": 1.63116798, + "memory(GiB)": 111.15, + "step": 18535, + "train_speed(iter/s)": 0.448769 + }, + { + "acc": 0.64535542, + "epoch": 0.4703196347031963, + "grad_norm": 4.75, + "learning_rate": 9.09160582017925e-06, + "loss": 1.61513557, + "memory(GiB)": 111.15, + "step": 18540, + "train_speed(iter/s)": 0.448833 + }, + { + "acc": 0.62677269, + "epoch": 0.4704464738711314, + "grad_norm": 4.8125, + "learning_rate": 9.091003021963449e-06, + "loss": 1.71194382, + "memory(GiB)": 111.15, + "step": 18545, + "train_speed(iter/s)": 0.448897 + }, + { + "acc": 0.64866896, + "epoch": 0.47057331303906647, + "grad_norm": 5.3125, + "learning_rate": 9.09040004380707e-06, + "loss": 1.66419907, + "memory(GiB)": 111.15, + "step": 18550, + "train_speed(iter/s)": 0.44896 + }, + { + "acc": 0.64011097, + "epoch": 0.4707001522070015, + "grad_norm": 5.34375, + "learning_rate": 9.089796885736637e-06, + "loss": 1.62956238, + "memory(GiB)": 111.15, + "step": 18555, + "train_speed(iter/s)": 0.449024 + }, + { + "acc": 0.65847063, + "epoch": 0.47082699137493655, + "grad_norm": 4.59375, + "learning_rate": 9.089193547778674e-06, + "loss": 1.58198814, + "memory(GiB)": 111.15, + "step": 18560, + "train_speed(iter/s)": 0.449087 + }, + { + "acc": 0.64289827, + "epoch": 0.47095383054287165, + "grad_norm": 5.125, + "learning_rate": 9.088590029959724e-06, + "loss": 1.64502754, + "memory(GiB)": 111.15, + "step": 18565, + "train_speed(iter/s)": 0.449151 + }, + { + "acc": 0.63535042, + "epoch": 0.4710806697108067, + "grad_norm": 5.46875, + "learning_rate": 9.08798633230633e-06, + "loss": 1.72364941, + "memory(GiB)": 111.15, + "step": 18570, + "train_speed(iter/s)": 0.449215 + }, + { + "acc": 0.66117601, + "epoch": 0.47120750887874174, + "grad_norm": 4.5, + "learning_rate": 9.087382454845044e-06, + "loss": 1.66802788, + "memory(GiB)": 111.15, + "step": 18575, + "train_speed(iter/s)": 0.449278 + }, + { + "acc": 0.63896351, + "epoch": 0.4713343480466768, + "grad_norm": 4.875, + "learning_rate": 9.08677839760243e-06, + "loss": 1.69396667, + "memory(GiB)": 111.15, + "step": 18580, + "train_speed(iter/s)": 0.449341 + }, + { + "acc": 0.63585753, + "epoch": 0.4714611872146119, + "grad_norm": 5.28125, + "learning_rate": 9.086174160605055e-06, + "loss": 1.64885311, + "memory(GiB)": 111.15, + "step": 18585, + "train_speed(iter/s)": 0.449405 + }, + { + "acc": 0.63899016, + "epoch": 0.47158802638254693, + "grad_norm": 4.53125, + "learning_rate": 9.085569743879498e-06, + "loss": 1.67609367, + "memory(GiB)": 111.15, + "step": 18590, + "train_speed(iter/s)": 0.449468 + }, + { + "acc": 0.64270773, + "epoch": 0.471714865550482, + "grad_norm": 4.875, + "learning_rate": 9.084965147452342e-06, + "loss": 1.63265915, + "memory(GiB)": 111.15, + "step": 18595, + "train_speed(iter/s)": 0.449532 + }, + { + "acc": 0.64756985, + "epoch": 0.471841704718417, + "grad_norm": 4.6875, + "learning_rate": 9.084360371350182e-06, + "loss": 1.61219578, + "memory(GiB)": 111.15, + "step": 18600, + "train_speed(iter/s)": 0.449595 + }, + { + "epoch": 0.471841704718417, + "eval_acc": 0.6408095160708818, + "eval_loss": 1.6221330165863037, + "eval_runtime": 113.7313, + "eval_samples_per_second": 56.009, + "eval_steps_per_second": 28.005, + "step": 18600 + }, + { + "acc": 0.64520869, + "epoch": 0.4719685438863521, + "grad_norm": 4.875, + "learning_rate": 9.083755415599617e-06, + "loss": 1.62295189, + "memory(GiB)": 111.15, + "step": 18605, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.64881015, + "epoch": 0.47209538305428717, + "grad_norm": 5.3125, + "learning_rate": 9.083150280227255e-06, + "loss": 1.5992465, + "memory(GiB)": 111.15, + "step": 18610, + "train_speed(iter/s)": 0.448439 + }, + { + "acc": 0.65273237, + "epoch": 0.4722222222222222, + "grad_norm": 5.09375, + "learning_rate": 9.082544965259716e-06, + "loss": 1.61947994, + "memory(GiB)": 111.15, + "step": 18615, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.65438251, + "epoch": 0.47234906139015725, + "grad_norm": 5.0625, + "learning_rate": 9.081939470723619e-06, + "loss": 1.62616005, + "memory(GiB)": 111.15, + "step": 18620, + "train_speed(iter/s)": 0.448564 + }, + { + "acc": 0.64187794, + "epoch": 0.47247590055809235, + "grad_norm": 5.96875, + "learning_rate": 9.081333796645603e-06, + "loss": 1.67179718, + "memory(GiB)": 111.15, + "step": 18625, + "train_speed(iter/s)": 0.448627 + }, + { + "acc": 0.6504077, + "epoch": 0.4726027397260274, + "grad_norm": 6.09375, + "learning_rate": 9.080727943052304e-06, + "loss": 1.61995544, + "memory(GiB)": 111.15, + "step": 18630, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.64500012, + "epoch": 0.47272957889396244, + "grad_norm": 5.375, + "learning_rate": 9.080121909970369e-06, + "loss": 1.6666893, + "memory(GiB)": 111.15, + "step": 18635, + "train_speed(iter/s)": 0.448751 + }, + { + "acc": 0.6491581, + "epoch": 0.4728564180618975, + "grad_norm": 5.8125, + "learning_rate": 9.079515697426457e-06, + "loss": 1.58967876, + "memory(GiB)": 111.15, + "step": 18640, + "train_speed(iter/s)": 0.448814 + }, + { + "acc": 0.64640131, + "epoch": 0.4729832572298326, + "grad_norm": 5.28125, + "learning_rate": 9.078909305447231e-06, + "loss": 1.66195908, + "memory(GiB)": 111.15, + "step": 18645, + "train_speed(iter/s)": 0.448878 + }, + { + "acc": 0.66532059, + "epoch": 0.47311009639776763, + "grad_norm": 5.03125, + "learning_rate": 9.078302734059363e-06, + "loss": 1.55669117, + "memory(GiB)": 111.15, + "step": 18650, + "train_speed(iter/s)": 0.448941 + }, + { + "acc": 0.6552393, + "epoch": 0.4732369355657027, + "grad_norm": 5.46875, + "learning_rate": 9.077695983289531e-06, + "loss": 1.60450153, + "memory(GiB)": 111.15, + "step": 18655, + "train_speed(iter/s)": 0.449004 + }, + { + "acc": 0.63598161, + "epoch": 0.4733637747336377, + "grad_norm": 4.6875, + "learning_rate": 9.077089053164426e-06, + "loss": 1.649403, + "memory(GiB)": 111.15, + "step": 18660, + "train_speed(iter/s)": 0.449068 + }, + { + "acc": 0.65972185, + "epoch": 0.4734906139015728, + "grad_norm": 4.75, + "learning_rate": 9.076481943710742e-06, + "loss": 1.59884148, + "memory(GiB)": 111.15, + "step": 18665, + "train_speed(iter/s)": 0.44913 + }, + { + "acc": 0.64556623, + "epoch": 0.47361745306950787, + "grad_norm": 6.21875, + "learning_rate": 9.075874654955178e-06, + "loss": 1.63819695, + "memory(GiB)": 111.15, + "step": 18670, + "train_speed(iter/s)": 0.449194 + }, + { + "acc": 0.64022832, + "epoch": 0.4737442922374429, + "grad_norm": 4.46875, + "learning_rate": 9.075267186924453e-06, + "loss": 1.68326168, + "memory(GiB)": 111.15, + "step": 18675, + "train_speed(iter/s)": 0.449257 + }, + { + "acc": 0.63208704, + "epoch": 0.47387113140537795, + "grad_norm": 5.40625, + "learning_rate": 9.074659539645281e-06, + "loss": 1.63173866, + "memory(GiB)": 111.15, + "step": 18680, + "train_speed(iter/s)": 0.44932 + }, + { + "acc": 0.63972893, + "epoch": 0.47399797057331305, + "grad_norm": 5.40625, + "learning_rate": 9.074051713144392e-06, + "loss": 1.71313477, + "memory(GiB)": 111.15, + "step": 18685, + "train_speed(iter/s)": 0.449384 + }, + { + "acc": 0.6545918, + "epoch": 0.4741248097412481, + "grad_norm": 5.28125, + "learning_rate": 9.073443707448517e-06, + "loss": 1.64684696, + "memory(GiB)": 111.15, + "step": 18690, + "train_speed(iter/s)": 0.449447 + }, + { + "acc": 0.64129438, + "epoch": 0.47425164890918314, + "grad_norm": 5.9375, + "learning_rate": 9.072835522584402e-06, + "loss": 1.67804012, + "memory(GiB)": 111.15, + "step": 18695, + "train_speed(iter/s)": 0.44951 + }, + { + "acc": 0.66137505, + "epoch": 0.4743784880771182, + "grad_norm": 5.65625, + "learning_rate": 9.072227158578798e-06, + "loss": 1.59022255, + "memory(GiB)": 111.15, + "step": 18700, + "train_speed(iter/s)": 0.449572 + }, + { + "epoch": 0.4743784880771182, + "eval_acc": 0.640723873772443, + "eval_loss": 1.6221725940704346, + "eval_runtime": 114.8496, + "eval_samples_per_second": 55.464, + "eval_steps_per_second": 27.732, + "step": 18700 + }, + { + "acc": 0.63176355, + "epoch": 0.4745053272450533, + "grad_norm": 6.125, + "learning_rate": 9.07161861545846e-06, + "loss": 1.65353279, + "memory(GiB)": 111.15, + "step": 18705, + "train_speed(iter/s)": 0.448349 + }, + { + "acc": 0.64030285, + "epoch": 0.47463216641298833, + "grad_norm": 6.09375, + "learning_rate": 9.071009893250158e-06, + "loss": 1.70206947, + "memory(GiB)": 111.15, + "step": 18710, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.6406744, + "epoch": 0.4747590055809234, + "grad_norm": 5.46875, + "learning_rate": 9.070400991980666e-06, + "loss": 1.65643673, + "memory(GiB)": 111.15, + "step": 18715, + "train_speed(iter/s)": 0.448474 + }, + { + "acc": 0.6610714, + "epoch": 0.4748858447488584, + "grad_norm": 4.90625, + "learning_rate": 9.069791911676765e-06, + "loss": 1.63106213, + "memory(GiB)": 111.15, + "step": 18720, + "train_speed(iter/s)": 0.448538 + }, + { + "acc": 0.66428165, + "epoch": 0.4750126839167935, + "grad_norm": 5.375, + "learning_rate": 9.069182652365245e-06, + "loss": 1.61055508, + "memory(GiB)": 111.15, + "step": 18725, + "train_speed(iter/s)": 0.448601 + }, + { + "acc": 0.62881489, + "epoch": 0.47513952308472857, + "grad_norm": 5.6875, + "learning_rate": 9.068573214072904e-06, + "loss": 1.71944809, + "memory(GiB)": 111.15, + "step": 18730, + "train_speed(iter/s)": 0.448665 + }, + { + "acc": 0.63636727, + "epoch": 0.4752663622526636, + "grad_norm": 4.34375, + "learning_rate": 9.067963596826547e-06, + "loss": 1.66106853, + "memory(GiB)": 111.15, + "step": 18735, + "train_speed(iter/s)": 0.448728 + }, + { + "acc": 0.65077448, + "epoch": 0.47539320142059865, + "grad_norm": 5.03125, + "learning_rate": 9.067353800652991e-06, + "loss": 1.68044224, + "memory(GiB)": 111.15, + "step": 18740, + "train_speed(iter/s)": 0.448791 + }, + { + "acc": 0.65210552, + "epoch": 0.47552004058853375, + "grad_norm": 5.5625, + "learning_rate": 9.066743825579056e-06, + "loss": 1.64920654, + "memory(GiB)": 111.15, + "step": 18745, + "train_speed(iter/s)": 0.448854 + }, + { + "acc": 0.65511675, + "epoch": 0.4756468797564688, + "grad_norm": 6.03125, + "learning_rate": 9.06613367163157e-06, + "loss": 1.58560295, + "memory(GiB)": 111.15, + "step": 18750, + "train_speed(iter/s)": 0.448917 + }, + { + "acc": 0.64729409, + "epoch": 0.47577371892440384, + "grad_norm": 5.5, + "learning_rate": 9.06552333883737e-06, + "loss": 1.61739349, + "memory(GiB)": 111.15, + "step": 18755, + "train_speed(iter/s)": 0.448979 + }, + { + "acc": 0.64822044, + "epoch": 0.4759005580923389, + "grad_norm": 5.5625, + "learning_rate": 9.064912827223303e-06, + "loss": 1.65681229, + "memory(GiB)": 111.15, + "step": 18760, + "train_speed(iter/s)": 0.449043 + }, + { + "acc": 0.65030489, + "epoch": 0.476027397260274, + "grad_norm": 5.0, + "learning_rate": 9.06430213681622e-06, + "loss": 1.59055634, + "memory(GiB)": 111.15, + "step": 18765, + "train_speed(iter/s)": 0.449106 + }, + { + "acc": 0.62454462, + "epoch": 0.47615423642820903, + "grad_norm": 4.90625, + "learning_rate": 9.063691267642987e-06, + "loss": 1.71639156, + "memory(GiB)": 111.15, + "step": 18770, + "train_speed(iter/s)": 0.449168 + }, + { + "acc": 0.64751482, + "epoch": 0.4762810755961441, + "grad_norm": 5.46875, + "learning_rate": 9.063080219730467e-06, + "loss": 1.65420971, + "memory(GiB)": 111.15, + "step": 18775, + "train_speed(iter/s)": 0.449232 + }, + { + "acc": 0.64703283, + "epoch": 0.4764079147640791, + "grad_norm": 5.25, + "learning_rate": 9.062468993105538e-06, + "loss": 1.64552269, + "memory(GiB)": 111.15, + "step": 18780, + "train_speed(iter/s)": 0.449295 + }, + { + "acc": 0.66759686, + "epoch": 0.4765347539320142, + "grad_norm": 5.375, + "learning_rate": 9.061857587795084e-06, + "loss": 1.54307394, + "memory(GiB)": 111.15, + "step": 18785, + "train_speed(iter/s)": 0.449358 + }, + { + "acc": 0.64350672, + "epoch": 0.47666159309994927, + "grad_norm": 5.84375, + "learning_rate": 9.061246003826e-06, + "loss": 1.74264984, + "memory(GiB)": 111.15, + "step": 18790, + "train_speed(iter/s)": 0.449421 + }, + { + "acc": 0.64647236, + "epoch": 0.4767884322678843, + "grad_norm": 5.875, + "learning_rate": 9.060634241225184e-06, + "loss": 1.66083603, + "memory(GiB)": 111.15, + "step": 18795, + "train_speed(iter/s)": 0.449484 + }, + { + "acc": 0.65006104, + "epoch": 0.47691527143581935, + "grad_norm": 5.65625, + "learning_rate": 9.060022300019546e-06, + "loss": 1.63149872, + "memory(GiB)": 111.15, + "step": 18800, + "train_speed(iter/s)": 0.449547 + }, + { + "epoch": 0.47691527143581935, + "eval_acc": 0.6408207957882371, + "eval_loss": 1.6221325397491455, + "eval_runtime": 112.4919, + "eval_samples_per_second": 56.626, + "eval_steps_per_second": 28.313, + "step": 18800 + }, + { + "acc": 0.65572605, + "epoch": 0.47704211060375445, + "grad_norm": 5.78125, + "learning_rate": 9.059410180236e-06, + "loss": 1.61337185, + "memory(GiB)": 111.15, + "step": 18805, + "train_speed(iter/s)": 0.448355 + }, + { + "acc": 0.65356059, + "epoch": 0.4771689497716895, + "grad_norm": 4.46875, + "learning_rate": 9.058797881901469e-06, + "loss": 1.59716015, + "memory(GiB)": 111.15, + "step": 18810, + "train_speed(iter/s)": 0.448417 + }, + { + "acc": 0.66277456, + "epoch": 0.47729578893962454, + "grad_norm": 5.9375, + "learning_rate": 9.058185405042886e-06, + "loss": 1.54827862, + "memory(GiB)": 111.15, + "step": 18815, + "train_speed(iter/s)": 0.448479 + }, + { + "acc": 0.64740486, + "epoch": 0.4774226281075596, + "grad_norm": 4.90625, + "learning_rate": 9.05757274968719e-06, + "loss": 1.69446869, + "memory(GiB)": 111.15, + "step": 18820, + "train_speed(iter/s)": 0.448541 + }, + { + "acc": 0.64844704, + "epoch": 0.4775494672754947, + "grad_norm": 5.625, + "learning_rate": 9.056959915861331e-06, + "loss": 1.60191727, + "memory(GiB)": 111.15, + "step": 18825, + "train_speed(iter/s)": 0.448603 + }, + { + "acc": 0.6465189, + "epoch": 0.47767630644342973, + "grad_norm": 5.8125, + "learning_rate": 9.056346903592262e-06, + "loss": 1.65673962, + "memory(GiB)": 111.15, + "step": 18830, + "train_speed(iter/s)": 0.448666 + }, + { + "acc": 0.63962708, + "epoch": 0.4778031456113648, + "grad_norm": 4.375, + "learning_rate": 9.055733712906943e-06, + "loss": 1.6444313, + "memory(GiB)": 111.15, + "step": 18835, + "train_speed(iter/s)": 0.448728 + }, + { + "acc": 0.64170017, + "epoch": 0.4779299847792998, + "grad_norm": 5.34375, + "learning_rate": 9.05512034383235e-06, + "loss": 1.64773045, + "memory(GiB)": 111.15, + "step": 18840, + "train_speed(iter/s)": 0.448791 + }, + { + "acc": 0.63297062, + "epoch": 0.4780568239472349, + "grad_norm": 5.3125, + "learning_rate": 9.054506796395458e-06, + "loss": 1.67727013, + "memory(GiB)": 111.15, + "step": 18845, + "train_speed(iter/s)": 0.448854 + }, + { + "acc": 0.64635773, + "epoch": 0.47818366311516997, + "grad_norm": 5.4375, + "learning_rate": 9.053893070623256e-06, + "loss": 1.6804493, + "memory(GiB)": 111.15, + "step": 18850, + "train_speed(iter/s)": 0.448917 + }, + { + "acc": 0.65276842, + "epoch": 0.478310502283105, + "grad_norm": 5.78125, + "learning_rate": 9.053279166542738e-06, + "loss": 1.66035995, + "memory(GiB)": 111.15, + "step": 18855, + "train_speed(iter/s)": 0.44898 + }, + { + "acc": 0.66163511, + "epoch": 0.47843734145104005, + "grad_norm": 5.5625, + "learning_rate": 9.052665084180906e-06, + "loss": 1.60619774, + "memory(GiB)": 111.15, + "step": 18860, + "train_speed(iter/s)": 0.449042 + }, + { + "acc": 0.6435555, + "epoch": 0.47856418061897515, + "grad_norm": 4.625, + "learning_rate": 9.052050823564767e-06, + "loss": 1.65472984, + "memory(GiB)": 111.15, + "step": 18865, + "train_speed(iter/s)": 0.449105 + }, + { + "acc": 0.66019387, + "epoch": 0.4786910197869102, + "grad_norm": 4.75, + "learning_rate": 9.051436384721344e-06, + "loss": 1.62662754, + "memory(GiB)": 111.15, + "step": 18870, + "train_speed(iter/s)": 0.449168 + }, + { + "acc": 0.67041035, + "epoch": 0.47881785895484524, + "grad_norm": 5.125, + "learning_rate": 9.05082176767766e-06, + "loss": 1.55034599, + "memory(GiB)": 111.15, + "step": 18875, + "train_speed(iter/s)": 0.449231 + }, + { + "acc": 0.65229759, + "epoch": 0.4789446981227803, + "grad_norm": 4.34375, + "learning_rate": 9.050206972460749e-06, + "loss": 1.62118645, + "memory(GiB)": 111.15, + "step": 18880, + "train_speed(iter/s)": 0.449294 + }, + { + "acc": 0.65500956, + "epoch": 0.4790715372907154, + "grad_norm": 5.84375, + "learning_rate": 9.049591999097651e-06, + "loss": 1.66212234, + "memory(GiB)": 111.15, + "step": 18885, + "train_speed(iter/s)": 0.449356 + }, + { + "acc": 0.63629999, + "epoch": 0.47919837645865043, + "grad_norm": 5.5625, + "learning_rate": 9.048976847615418e-06, + "loss": 1.67770596, + "memory(GiB)": 111.15, + "step": 18890, + "train_speed(iter/s)": 0.449419 + }, + { + "acc": 0.6415987, + "epoch": 0.4793252156265855, + "grad_norm": 4.875, + "learning_rate": 9.048361518041107e-06, + "loss": 1.68103447, + "memory(GiB)": 111.15, + "step": 18895, + "train_speed(iter/s)": 0.449481 + }, + { + "acc": 0.65871015, + "epoch": 0.4794520547945205, + "grad_norm": 4.53125, + "learning_rate": 9.04774601040178e-06, + "loss": 1.56979551, + "memory(GiB)": 111.15, + "step": 18900, + "train_speed(iter/s)": 0.449544 + }, + { + "epoch": 0.4794520547945205, + "eval_acc": 0.6409156289674839, + "eval_loss": 1.6219755411148071, + "eval_runtime": 113.9931, + "eval_samples_per_second": 55.881, + "eval_steps_per_second": 27.94, + "step": 18900 + }, + { + "acc": 0.63968182, + "epoch": 0.4795788939624556, + "grad_norm": 6.375, + "learning_rate": 9.047130324724513e-06, + "loss": 1.64418335, + "memory(GiB)": 111.15, + "step": 18905, + "train_speed(iter/s)": 0.448343 + }, + { + "acc": 0.64582891, + "epoch": 0.47970573313039067, + "grad_norm": 7.28125, + "learning_rate": 9.046514461036385e-06, + "loss": 1.64974518, + "memory(GiB)": 111.15, + "step": 18910, + "train_speed(iter/s)": 0.448405 + }, + { + "acc": 0.64389315, + "epoch": 0.4798325722983257, + "grad_norm": 4.8125, + "learning_rate": 9.045898419364483e-06, + "loss": 1.64983215, + "memory(GiB)": 111.15, + "step": 18915, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.6542541, + "epoch": 0.47995941146626075, + "grad_norm": 4.25, + "learning_rate": 9.045282199735906e-06, + "loss": 1.60532722, + "memory(GiB)": 111.15, + "step": 18920, + "train_speed(iter/s)": 0.44853 + }, + { + "acc": 0.63222337, + "epoch": 0.48008625063419585, + "grad_norm": 5.34375, + "learning_rate": 9.044665802177756e-06, + "loss": 1.76689453, + "memory(GiB)": 111.15, + "step": 18925, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.64778614, + "epoch": 0.4802130898021309, + "grad_norm": 7.1875, + "learning_rate": 9.044049226717148e-06, + "loss": 1.61681767, + "memory(GiB)": 111.15, + "step": 18930, + "train_speed(iter/s)": 0.448655 + }, + { + "acc": 0.66843777, + "epoch": 0.48033992897006594, + "grad_norm": 4.5625, + "learning_rate": 9.043432473381198e-06, + "loss": 1.61745968, + "memory(GiB)": 111.15, + "step": 18935, + "train_speed(iter/s)": 0.448717 + }, + { + "acc": 0.64984846, + "epoch": 0.480466768138001, + "grad_norm": 5.59375, + "learning_rate": 9.042815542197037e-06, + "loss": 1.67952671, + "memory(GiB)": 111.15, + "step": 18940, + "train_speed(iter/s)": 0.448779 + }, + { + "acc": 0.64608345, + "epoch": 0.4805936073059361, + "grad_norm": 5.34375, + "learning_rate": 9.042198433191796e-06, + "loss": 1.59996204, + "memory(GiB)": 111.15, + "step": 18945, + "train_speed(iter/s)": 0.448842 + }, + { + "acc": 0.6438201, + "epoch": 0.48072044647387113, + "grad_norm": 6.34375, + "learning_rate": 9.041581146392621e-06, + "loss": 1.6727253, + "memory(GiB)": 111.15, + "step": 18950, + "train_speed(iter/s)": 0.448905 + }, + { + "acc": 0.63902388, + "epoch": 0.4808472856418062, + "grad_norm": 5.65625, + "learning_rate": 9.040963681826665e-06, + "loss": 1.61245079, + "memory(GiB)": 111.15, + "step": 18955, + "train_speed(iter/s)": 0.448968 + }, + { + "acc": 0.65387602, + "epoch": 0.4809741248097412, + "grad_norm": 5.0625, + "learning_rate": 9.040346039521085e-06, + "loss": 1.67941856, + "memory(GiB)": 111.15, + "step": 18960, + "train_speed(iter/s)": 0.449031 + }, + { + "acc": 0.65924988, + "epoch": 0.4811009639776763, + "grad_norm": 4.78125, + "learning_rate": 9.039728219503044e-06, + "loss": 1.58756809, + "memory(GiB)": 111.15, + "step": 18965, + "train_speed(iter/s)": 0.449093 + }, + { + "acc": 0.66051979, + "epoch": 0.48122780314561137, + "grad_norm": 4.75, + "learning_rate": 9.039110221799721e-06, + "loss": 1.58596115, + "memory(GiB)": 111.15, + "step": 18970, + "train_speed(iter/s)": 0.449156 + }, + { + "acc": 0.65079718, + "epoch": 0.4813546423135464, + "grad_norm": 5.46875, + "learning_rate": 9.038492046438298e-06, + "loss": 1.6492485, + "memory(GiB)": 111.15, + "step": 18975, + "train_speed(iter/s)": 0.449219 + }, + { + "acc": 0.64832649, + "epoch": 0.48148148148148145, + "grad_norm": 5.78125, + "learning_rate": 9.037873693445965e-06, + "loss": 1.65304794, + "memory(GiB)": 111.15, + "step": 18980, + "train_speed(iter/s)": 0.449281 + }, + { + "acc": 0.65419526, + "epoch": 0.48160832064941655, + "grad_norm": 4.875, + "learning_rate": 9.037255162849918e-06, + "loss": 1.63278275, + "memory(GiB)": 111.15, + "step": 18985, + "train_speed(iter/s)": 0.449344 + }, + { + "acc": 0.64988823, + "epoch": 0.4817351598173516, + "grad_norm": 5.4375, + "learning_rate": 9.036636454677363e-06, + "loss": 1.61742077, + "memory(GiB)": 111.15, + "step": 18990, + "train_speed(iter/s)": 0.449407 + }, + { + "acc": 0.6431056, + "epoch": 0.48186199898528664, + "grad_norm": 5.46875, + "learning_rate": 9.036017568955516e-06, + "loss": 1.66349716, + "memory(GiB)": 111.15, + "step": 18995, + "train_speed(iter/s)": 0.449469 + }, + { + "acc": 0.64743853, + "epoch": 0.4819888381532217, + "grad_norm": 4.34375, + "learning_rate": 9.035398505711597e-06, + "loss": 1.64596996, + "memory(GiB)": 111.15, + "step": 19000, + "train_speed(iter/s)": 0.449532 + }, + { + "epoch": 0.4819888381532217, + "eval_acc": 0.640792387611194, + "eval_loss": 1.6214004755020142, + "eval_runtime": 113.5192, + "eval_samples_per_second": 56.114, + "eval_steps_per_second": 28.057, + "step": 19000 + }, + { + "acc": 0.65405855, + "epoch": 0.4821156773211568, + "grad_norm": 5.8125, + "learning_rate": 9.034779264972834e-06, + "loss": 1.62203312, + "memory(GiB)": 111.15, + "step": 19005, + "train_speed(iter/s)": 0.448342 + }, + { + "acc": 0.65157089, + "epoch": 0.48224251648909183, + "grad_norm": 4.65625, + "learning_rate": 9.034159846766464e-06, + "loss": 1.63134403, + "memory(GiB)": 111.15, + "step": 19010, + "train_speed(iter/s)": 0.448404 + }, + { + "acc": 0.65551562, + "epoch": 0.4823693556570269, + "grad_norm": 5.65625, + "learning_rate": 9.033540251119734e-06, + "loss": 1.55701084, + "memory(GiB)": 111.15, + "step": 19015, + "train_speed(iter/s)": 0.448466 + }, + { + "acc": 0.65176058, + "epoch": 0.4824961948249619, + "grad_norm": 4.90625, + "learning_rate": 9.032920478059897e-06, + "loss": 1.59100113, + "memory(GiB)": 111.15, + "step": 19020, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.64677091, + "epoch": 0.482623033992897, + "grad_norm": 4.96875, + "learning_rate": 9.032300527614209e-06, + "loss": 1.71027184, + "memory(GiB)": 111.15, + "step": 19025, + "train_speed(iter/s)": 0.448591 + }, + { + "acc": 0.6445446, + "epoch": 0.48274987316083207, + "grad_norm": 5.5625, + "learning_rate": 9.031680399809941e-06, + "loss": 1.68353691, + "memory(GiB)": 111.15, + "step": 19030, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.63243742, + "epoch": 0.4828767123287671, + "grad_norm": 4.75, + "learning_rate": 9.031060094674371e-06, + "loss": 1.72041245, + "memory(GiB)": 111.15, + "step": 19035, + "train_speed(iter/s)": 0.448715 + }, + { + "acc": 0.6371913, + "epoch": 0.48300355149670215, + "grad_norm": 6.5625, + "learning_rate": 9.030439612234778e-06, + "loss": 1.64601326, + "memory(GiB)": 111.15, + "step": 19040, + "train_speed(iter/s)": 0.448777 + }, + { + "acc": 0.65773306, + "epoch": 0.48313039066463725, + "grad_norm": 4.78125, + "learning_rate": 9.029818952518458e-06, + "loss": 1.61323032, + "memory(GiB)": 111.15, + "step": 19045, + "train_speed(iter/s)": 0.44884 + }, + { + "acc": 0.66205664, + "epoch": 0.4832572298325723, + "grad_norm": 5.1875, + "learning_rate": 9.029198115552708e-06, + "loss": 1.56384802, + "memory(GiB)": 111.15, + "step": 19050, + "train_speed(iter/s)": 0.448902 + }, + { + "acc": 0.64410515, + "epoch": 0.48338406900050734, + "grad_norm": 5.25, + "learning_rate": 9.028577101364837e-06, + "loss": 1.6143074, + "memory(GiB)": 111.15, + "step": 19055, + "train_speed(iter/s)": 0.448964 + }, + { + "acc": 0.6591651, + "epoch": 0.4835109081684424, + "grad_norm": 5.46875, + "learning_rate": 9.027955909982157e-06, + "loss": 1.59567871, + "memory(GiB)": 111.15, + "step": 19060, + "train_speed(iter/s)": 0.449026 + }, + { + "acc": 0.67822781, + "epoch": 0.4836377473363775, + "grad_norm": 4.90625, + "learning_rate": 9.027334541431993e-06, + "loss": 1.52079248, + "memory(GiB)": 111.15, + "step": 19065, + "train_speed(iter/s)": 0.449087 + }, + { + "acc": 0.64846668, + "epoch": 0.48376458650431253, + "grad_norm": 6.53125, + "learning_rate": 9.026712995741676e-06, + "loss": 1.61672249, + "memory(GiB)": 111.15, + "step": 19070, + "train_speed(iter/s)": 0.449149 + }, + { + "acc": 0.62901874, + "epoch": 0.4838914256722476, + "grad_norm": 5.40625, + "learning_rate": 9.026091272938543e-06, + "loss": 1.69690247, + "memory(GiB)": 111.15, + "step": 19075, + "train_speed(iter/s)": 0.449212 + }, + { + "acc": 0.65588951, + "epoch": 0.4840182648401826, + "grad_norm": 4.875, + "learning_rate": 9.02546937304994e-06, + "loss": 1.60648041, + "memory(GiB)": 111.15, + "step": 19080, + "train_speed(iter/s)": 0.449274 + }, + { + "acc": 0.65000858, + "epoch": 0.4841451040081177, + "grad_norm": 5.40625, + "learning_rate": 9.02484729610322e-06, + "loss": 1.66610489, + "memory(GiB)": 111.15, + "step": 19085, + "train_speed(iter/s)": 0.449336 + }, + { + "acc": 0.63661933, + "epoch": 0.48427194317605277, + "grad_norm": 4.75, + "learning_rate": 9.02422504212575e-06, + "loss": 1.71079102, + "memory(GiB)": 111.15, + "step": 19090, + "train_speed(iter/s)": 0.449399 + }, + { + "acc": 0.65368338, + "epoch": 0.4843987823439878, + "grad_norm": 4.75, + "learning_rate": 9.023602611144893e-06, + "loss": 1.60765495, + "memory(GiB)": 111.15, + "step": 19095, + "train_speed(iter/s)": 0.449461 + }, + { + "acc": 0.64198837, + "epoch": 0.48452562151192285, + "grad_norm": 4.78125, + "learning_rate": 9.02298000318803e-06, + "loss": 1.6267786, + "memory(GiB)": 111.15, + "step": 19100, + "train_speed(iter/s)": 0.449523 + }, + { + "epoch": 0.48452562151192285, + "eval_acc": 0.6408809542807989, + "eval_loss": 1.6217284202575684, + "eval_runtime": 113.7076, + "eval_samples_per_second": 56.021, + "eval_steps_per_second": 28.01, + "step": 19100 + }, + { + "acc": 0.64203553, + "epoch": 0.48465246067985795, + "grad_norm": 5.15625, + "learning_rate": 9.022357218282546e-06, + "loss": 1.63061447, + "memory(GiB)": 111.15, + "step": 19105, + "train_speed(iter/s)": 0.448337 + }, + { + "acc": 0.64170532, + "epoch": 0.484779299847793, + "grad_norm": 5.53125, + "learning_rate": 9.021734256455832e-06, + "loss": 1.68958206, + "memory(GiB)": 111.15, + "step": 19110, + "train_speed(iter/s)": 0.448399 + }, + { + "acc": 0.67630024, + "epoch": 0.48490613901572804, + "grad_norm": 5.09375, + "learning_rate": 9.02111111773529e-06, + "loss": 1.56844578, + "memory(GiB)": 111.15, + "step": 19115, + "train_speed(iter/s)": 0.448461 + }, + { + "acc": 0.65351734, + "epoch": 0.4850329781836631, + "grad_norm": 4.8125, + "learning_rate": 9.020487802148328e-06, + "loss": 1.56021843, + "memory(GiB)": 111.15, + "step": 19120, + "train_speed(iter/s)": 0.448523 + }, + { + "acc": 0.64454651, + "epoch": 0.4851598173515982, + "grad_norm": 5.28125, + "learning_rate": 9.019864309722362e-06, + "loss": 1.63543339, + "memory(GiB)": 111.15, + "step": 19125, + "train_speed(iter/s)": 0.448584 + }, + { + "acc": 0.65549674, + "epoch": 0.48528665651953323, + "grad_norm": 4.34375, + "learning_rate": 9.019240640484816e-06, + "loss": 1.61127357, + "memory(GiB)": 111.15, + "step": 19130, + "train_speed(iter/s)": 0.448646 + }, + { + "acc": 0.64719925, + "epoch": 0.4854134956874683, + "grad_norm": 5.625, + "learning_rate": 9.018616794463124e-06, + "loss": 1.61882591, + "memory(GiB)": 111.15, + "step": 19135, + "train_speed(iter/s)": 0.448708 + }, + { + "acc": 0.63946924, + "epoch": 0.4855403348554033, + "grad_norm": 4.90625, + "learning_rate": 9.017992771684722e-06, + "loss": 1.66870384, + "memory(GiB)": 111.15, + "step": 19140, + "train_speed(iter/s)": 0.44877 + }, + { + "acc": 0.63684483, + "epoch": 0.4856671740233384, + "grad_norm": 5.71875, + "learning_rate": 9.017368572177058e-06, + "loss": 1.64602413, + "memory(GiB)": 111.15, + "step": 19145, + "train_speed(iter/s)": 0.448831 + }, + { + "acc": 0.64727974, + "epoch": 0.48579401319127347, + "grad_norm": 5.96875, + "learning_rate": 9.016744195967588e-06, + "loss": 1.62067356, + "memory(GiB)": 111.15, + "step": 19150, + "train_speed(iter/s)": 0.448892 + }, + { + "acc": 0.66369309, + "epoch": 0.4859208523592085, + "grad_norm": 5.6875, + "learning_rate": 9.016119643083777e-06, + "loss": 1.66478291, + "memory(GiB)": 111.15, + "step": 19155, + "train_speed(iter/s)": 0.448954 + }, + { + "acc": 0.65257587, + "epoch": 0.48604769152714355, + "grad_norm": 5.90625, + "learning_rate": 9.015494913553091e-06, + "loss": 1.64631138, + "memory(GiB)": 111.15, + "step": 19160, + "train_speed(iter/s)": 0.449016 + }, + { + "acc": 0.63958178, + "epoch": 0.48617453069507865, + "grad_norm": 4.84375, + "learning_rate": 9.014870007403012e-06, + "loss": 1.63401909, + "memory(GiB)": 111.15, + "step": 19165, + "train_speed(iter/s)": 0.449079 + }, + { + "acc": 0.64127951, + "epoch": 0.4863013698630137, + "grad_norm": 5.1875, + "learning_rate": 9.014244924661026e-06, + "loss": 1.57819738, + "memory(GiB)": 111.15, + "step": 19170, + "train_speed(iter/s)": 0.44914 + }, + { + "acc": 0.64438739, + "epoch": 0.48642820903094874, + "grad_norm": 5.125, + "learning_rate": 9.013619665354626e-06, + "loss": 1.64140167, + "memory(GiB)": 111.15, + "step": 19175, + "train_speed(iter/s)": 0.449202 + }, + { + "acc": 0.64124694, + "epoch": 0.4865550481988838, + "grad_norm": 5.03125, + "learning_rate": 9.01299422951131e-06, + "loss": 1.69079762, + "memory(GiB)": 111.15, + "step": 19180, + "train_speed(iter/s)": 0.449264 + }, + { + "acc": 0.63716879, + "epoch": 0.4866818873668189, + "grad_norm": 5.875, + "learning_rate": 9.012368617158593e-06, + "loss": 1.67937794, + "memory(GiB)": 111.15, + "step": 19185, + "train_speed(iter/s)": 0.449326 + }, + { + "acc": 0.63726978, + "epoch": 0.48680872653475393, + "grad_norm": 5.0625, + "learning_rate": 9.01174282832399e-06, + "loss": 1.63657322, + "memory(GiB)": 111.15, + "step": 19190, + "train_speed(iter/s)": 0.449388 + }, + { + "acc": 0.66072812, + "epoch": 0.486935565702689, + "grad_norm": 5.40625, + "learning_rate": 9.011116863035027e-06, + "loss": 1.55699387, + "memory(GiB)": 111.15, + "step": 19195, + "train_speed(iter/s)": 0.449449 + }, + { + "acc": 0.63863153, + "epoch": 0.487062404870624, + "grad_norm": 5.46875, + "learning_rate": 9.010490721319237e-06, + "loss": 1.66960392, + "memory(GiB)": 111.15, + "step": 19200, + "train_speed(iter/s)": 0.449511 + }, + { + "epoch": 0.487062404870624, + "eval_acc": 0.6409774585292836, + "eval_loss": 1.621353268623352, + "eval_runtime": 114.0373, + "eval_samples_per_second": 55.859, + "eval_steps_per_second": 27.929, + "step": 19200 + }, + { + "acc": 0.64599915, + "epoch": 0.4871892440385591, + "grad_norm": 5.8125, + "learning_rate": 9.009864403204157e-06, + "loss": 1.63188744, + "memory(GiB)": 111.15, + "step": 19205, + "train_speed(iter/s)": 0.448328 + }, + { + "acc": 0.65334759, + "epoch": 0.48731608320649417, + "grad_norm": 4.75, + "learning_rate": 9.00923790871734e-06, + "loss": 1.57228127, + "memory(GiB)": 111.15, + "step": 19210, + "train_speed(iter/s)": 0.44839 + }, + { + "acc": 0.64364328, + "epoch": 0.4874429223744292, + "grad_norm": 6.4375, + "learning_rate": 9.008611237886339e-06, + "loss": 1.68164902, + "memory(GiB)": 111.15, + "step": 19215, + "train_speed(iter/s)": 0.448452 + }, + { + "acc": 0.65107927, + "epoch": 0.48756976154236426, + "grad_norm": 5.46875, + "learning_rate": 9.00798439073872e-06, + "loss": 1.62828445, + "memory(GiB)": 111.15, + "step": 19220, + "train_speed(iter/s)": 0.448513 + }, + { + "acc": 0.64676628, + "epoch": 0.48769660071029935, + "grad_norm": 6.96875, + "learning_rate": 9.007357367302052e-06, + "loss": 1.64756699, + "memory(GiB)": 111.15, + "step": 19225, + "train_speed(iter/s)": 0.448575 + }, + { + "acc": 0.63618631, + "epoch": 0.4878234398782344, + "grad_norm": 5.375, + "learning_rate": 9.006730167603914e-06, + "loss": 1.70140572, + "memory(GiB)": 111.15, + "step": 19230, + "train_speed(iter/s)": 0.448636 + }, + { + "acc": 0.64493132, + "epoch": 0.48795027904616944, + "grad_norm": 5.6875, + "learning_rate": 9.006102791671896e-06, + "loss": 1.66279945, + "memory(GiB)": 111.15, + "step": 19235, + "train_speed(iter/s)": 0.448698 + }, + { + "acc": 0.65589643, + "epoch": 0.4880771182141045, + "grad_norm": 5.90625, + "learning_rate": 9.005475239533591e-06, + "loss": 1.63769798, + "memory(GiB)": 111.15, + "step": 19240, + "train_speed(iter/s)": 0.448761 + }, + { + "acc": 0.63749704, + "epoch": 0.4882039573820396, + "grad_norm": 5.65625, + "learning_rate": 9.0048475112166e-06, + "loss": 1.65136223, + "memory(GiB)": 111.15, + "step": 19245, + "train_speed(iter/s)": 0.448822 + }, + { + "acc": 0.64306965, + "epoch": 0.48833079654997463, + "grad_norm": 4.9375, + "learning_rate": 9.00421960674854e-06, + "loss": 1.64080315, + "memory(GiB)": 111.15, + "step": 19250, + "train_speed(iter/s)": 0.448884 + }, + { + "acc": 0.64099693, + "epoch": 0.4884576357179097, + "grad_norm": 5.4375, + "learning_rate": 9.003591526157021e-06, + "loss": 1.66535778, + "memory(GiB)": 111.15, + "step": 19255, + "train_speed(iter/s)": 0.448946 + }, + { + "acc": 0.64774833, + "epoch": 0.4885844748858447, + "grad_norm": 5.8125, + "learning_rate": 9.002963269469672e-06, + "loss": 1.66947594, + "memory(GiB)": 111.15, + "step": 19260, + "train_speed(iter/s)": 0.449008 + }, + { + "acc": 0.65985851, + "epoch": 0.4887113140537798, + "grad_norm": 5.03125, + "learning_rate": 9.002334836714126e-06, + "loss": 1.58462105, + "memory(GiB)": 111.15, + "step": 19265, + "train_speed(iter/s)": 0.44907 + }, + { + "acc": 0.64875956, + "epoch": 0.48883815322171487, + "grad_norm": 4.71875, + "learning_rate": 9.001706227918023e-06, + "loss": 1.60208797, + "memory(GiB)": 111.15, + "step": 19270, + "train_speed(iter/s)": 0.449132 + }, + { + "acc": 0.64401202, + "epoch": 0.4889649923896499, + "grad_norm": 5.0, + "learning_rate": 9.001077443109016e-06, + "loss": 1.69486084, + "memory(GiB)": 111.15, + "step": 19275, + "train_speed(iter/s)": 0.449194 + }, + { + "acc": 0.64017572, + "epoch": 0.48909183155758496, + "grad_norm": 4.5625, + "learning_rate": 9.00044848231476e-06, + "loss": 1.65506401, + "memory(GiB)": 111.15, + "step": 19280, + "train_speed(iter/s)": 0.449255 + }, + { + "acc": 0.66595573, + "epoch": 0.48921867072552006, + "grad_norm": 4.84375, + "learning_rate": 8.999819345562919e-06, + "loss": 1.56353426, + "memory(GiB)": 111.15, + "step": 19285, + "train_speed(iter/s)": 0.449317 + }, + { + "acc": 0.64263086, + "epoch": 0.4893455098934551, + "grad_norm": 5.5, + "learning_rate": 8.999190032881165e-06, + "loss": 1.65221329, + "memory(GiB)": 111.15, + "step": 19290, + "train_speed(iter/s)": 0.449379 + }, + { + "acc": 0.66589212, + "epoch": 0.48947234906139014, + "grad_norm": 5.1875, + "learning_rate": 8.998560544297176e-06, + "loss": 1.58555279, + "memory(GiB)": 111.15, + "step": 19295, + "train_speed(iter/s)": 0.449441 + }, + { + "acc": 0.64519072, + "epoch": 0.4895991882293252, + "grad_norm": 4.875, + "learning_rate": 8.997930879838646e-06, + "loss": 1.62003307, + "memory(GiB)": 111.15, + "step": 19300, + "train_speed(iter/s)": 0.449503 + }, + { + "epoch": 0.4895991882293252, + "eval_acc": 0.6409770407619741, + "eval_loss": 1.621379017829895, + "eval_runtime": 112.6403, + "eval_samples_per_second": 56.552, + "eval_steps_per_second": 28.276, + "step": 19300 + }, + { + "acc": 0.64992905, + "epoch": 0.4897260273972603, + "grad_norm": 5.0625, + "learning_rate": 8.997301039533264e-06, + "loss": 1.65375214, + "memory(GiB)": 111.15, + "step": 19305, + "train_speed(iter/s)": 0.448342 + }, + { + "acc": 0.64251685, + "epoch": 0.48985286656519533, + "grad_norm": 5.71875, + "learning_rate": 8.996671023408737e-06, + "loss": 1.6641201, + "memory(GiB)": 111.15, + "step": 19310, + "train_speed(iter/s)": 0.448403 + }, + { + "acc": 0.64107342, + "epoch": 0.4899797057331304, + "grad_norm": 5.0625, + "learning_rate": 8.996040831492772e-06, + "loss": 1.677528, + "memory(GiB)": 111.15, + "step": 19315, + "train_speed(iter/s)": 0.448465 + }, + { + "acc": 0.62711916, + "epoch": 0.4901065449010654, + "grad_norm": 4.5625, + "learning_rate": 8.995410463813093e-06, + "loss": 1.60708542, + "memory(GiB)": 111.15, + "step": 19320, + "train_speed(iter/s)": 0.448527 + }, + { + "acc": 0.65129037, + "epoch": 0.4902333840690005, + "grad_norm": 4.46875, + "learning_rate": 8.99477992039742e-06, + "loss": 1.5919631, + "memory(GiB)": 111.15, + "step": 19325, + "train_speed(iter/s)": 0.448588 + }, + { + "acc": 0.64937878, + "epoch": 0.49036022323693557, + "grad_norm": 6.28125, + "learning_rate": 8.994149201273495e-06, + "loss": 1.63562737, + "memory(GiB)": 111.15, + "step": 19330, + "train_speed(iter/s)": 0.44865 + }, + { + "acc": 0.63755674, + "epoch": 0.4904870624048706, + "grad_norm": 5.75, + "learning_rate": 8.993518306469052e-06, + "loss": 1.60786171, + "memory(GiB)": 111.15, + "step": 19335, + "train_speed(iter/s)": 0.448711 + }, + { + "acc": 0.62695413, + "epoch": 0.49061390157280566, + "grad_norm": 5.34375, + "learning_rate": 8.992887236011847e-06, + "loss": 1.7553627, + "memory(GiB)": 111.15, + "step": 19340, + "train_speed(iter/s)": 0.448773 + }, + { + "acc": 0.63381844, + "epoch": 0.49074074074074076, + "grad_norm": 4.84375, + "learning_rate": 8.992255989929632e-06, + "loss": 1.67682877, + "memory(GiB)": 111.15, + "step": 19345, + "train_speed(iter/s)": 0.448835 + }, + { + "acc": 0.66563301, + "epoch": 0.4908675799086758, + "grad_norm": 4.8125, + "learning_rate": 8.991624568250175e-06, + "loss": 1.58863144, + "memory(GiB)": 111.15, + "step": 19350, + "train_speed(iter/s)": 0.448896 + }, + { + "acc": 0.64343934, + "epoch": 0.49099441907661084, + "grad_norm": 6.0625, + "learning_rate": 8.99099297100125e-06, + "loss": 1.69840851, + "memory(GiB)": 111.15, + "step": 19355, + "train_speed(iter/s)": 0.448958 + }, + { + "acc": 0.65460687, + "epoch": 0.4911212582445459, + "grad_norm": 6.34375, + "learning_rate": 8.990361198210634e-06, + "loss": 1.62798386, + "memory(GiB)": 111.15, + "step": 19360, + "train_speed(iter/s)": 0.449019 + }, + { + "acc": 0.63621321, + "epoch": 0.491248097412481, + "grad_norm": 4.53125, + "learning_rate": 8.989729249906116e-06, + "loss": 1.73319359, + "memory(GiB)": 111.15, + "step": 19365, + "train_speed(iter/s)": 0.44908 + }, + { + "acc": 0.65269837, + "epoch": 0.49137493658041603, + "grad_norm": 5.0, + "learning_rate": 8.989097126115493e-06, + "loss": 1.60817852, + "memory(GiB)": 111.15, + "step": 19370, + "train_speed(iter/s)": 0.449141 + }, + { + "acc": 0.64523873, + "epoch": 0.4915017757483511, + "grad_norm": 5.625, + "learning_rate": 8.98846482686657e-06, + "loss": 1.69147663, + "memory(GiB)": 111.15, + "step": 19375, + "train_speed(iter/s)": 0.449203 + }, + { + "acc": 0.6513546, + "epoch": 0.4916286149162861, + "grad_norm": 5.90625, + "learning_rate": 8.987832352187156e-06, + "loss": 1.6666975, + "memory(GiB)": 111.15, + "step": 19380, + "train_speed(iter/s)": 0.449264 + }, + { + "acc": 0.65197001, + "epoch": 0.4917554540842212, + "grad_norm": 5.5, + "learning_rate": 8.987199702105071e-06, + "loss": 1.68130054, + "memory(GiB)": 111.15, + "step": 19385, + "train_speed(iter/s)": 0.449326 + }, + { + "acc": 0.63824501, + "epoch": 0.49188229325215627, + "grad_norm": 5.34375, + "learning_rate": 8.986566876648141e-06, + "loss": 1.68513718, + "memory(GiB)": 111.15, + "step": 19390, + "train_speed(iter/s)": 0.449388 + }, + { + "acc": 0.63807278, + "epoch": 0.4920091324200913, + "grad_norm": 5.0625, + "learning_rate": 8.985933875844202e-06, + "loss": 1.62893333, + "memory(GiB)": 111.15, + "step": 19395, + "train_speed(iter/s)": 0.449449 + }, + { + "acc": 0.67302475, + "epoch": 0.49213597158802636, + "grad_norm": 5.1875, + "learning_rate": 8.985300699721094e-06, + "loss": 1.58375063, + "memory(GiB)": 111.15, + "step": 19400, + "train_speed(iter/s)": 0.44951 + }, + { + "epoch": 0.49213597158802636, + "eval_acc": 0.6410580876200089, + "eval_loss": 1.620782494544983, + "eval_runtime": 114.4032, + "eval_samples_per_second": 55.68, + "eval_steps_per_second": 27.84, + "step": 19400 + }, + { + "acc": 0.6445715, + "epoch": 0.49226281075596146, + "grad_norm": 5.5625, + "learning_rate": 8.984667348306669e-06, + "loss": 1.66913948, + "memory(GiB)": 111.15, + "step": 19405, + "train_speed(iter/s)": 0.448336 + }, + { + "acc": 0.65802813, + "epoch": 0.4923896499238965, + "grad_norm": 5.21875, + "learning_rate": 8.984033821628782e-06, + "loss": 1.59751186, + "memory(GiB)": 111.15, + "step": 19410, + "train_speed(iter/s)": 0.448397 + }, + { + "acc": 0.66723785, + "epoch": 0.49251648909183154, + "grad_norm": 5.03125, + "learning_rate": 8.983400119715303e-06, + "loss": 1.58746223, + "memory(GiB)": 111.15, + "step": 19415, + "train_speed(iter/s)": 0.448458 + }, + { + "acc": 0.64559422, + "epoch": 0.4926433282597666, + "grad_norm": 5.59375, + "learning_rate": 8.982766242594099e-06, + "loss": 1.63232346, + "memory(GiB)": 111.15, + "step": 19420, + "train_speed(iter/s)": 0.448519 + }, + { + "acc": 0.66723385, + "epoch": 0.4927701674277017, + "grad_norm": 6.5625, + "learning_rate": 8.982132190293056e-06, + "loss": 1.6433773, + "memory(GiB)": 111.15, + "step": 19425, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.65465431, + "epoch": 0.49289700659563673, + "grad_norm": 5.34375, + "learning_rate": 8.98149796284006e-06, + "loss": 1.61076221, + "memory(GiB)": 111.15, + "step": 19430, + "train_speed(iter/s)": 0.448641 + }, + { + "acc": 0.65105076, + "epoch": 0.4930238457635718, + "grad_norm": 6.03125, + "learning_rate": 8.980863560263007e-06, + "loss": 1.62299786, + "memory(GiB)": 111.15, + "step": 19435, + "train_speed(iter/s)": 0.448702 + }, + { + "acc": 0.65418553, + "epoch": 0.4931506849315068, + "grad_norm": 5.40625, + "learning_rate": 8.980228982589802e-06, + "loss": 1.61417351, + "memory(GiB)": 111.15, + "step": 19440, + "train_speed(iter/s)": 0.448763 + }, + { + "acc": 0.66117258, + "epoch": 0.4932775240994419, + "grad_norm": 5.15625, + "learning_rate": 8.979594229848355e-06, + "loss": 1.65091095, + "memory(GiB)": 111.15, + "step": 19445, + "train_speed(iter/s)": 0.448825 + }, + { + "acc": 0.65229969, + "epoch": 0.49340436326737697, + "grad_norm": 5.125, + "learning_rate": 8.978959302066587e-06, + "loss": 1.67502575, + "memory(GiB)": 111.15, + "step": 19450, + "train_speed(iter/s)": 0.448886 + }, + { + "acc": 0.64084988, + "epoch": 0.493531202435312, + "grad_norm": 4.53125, + "learning_rate": 8.978324199272423e-06, + "loss": 1.59788513, + "memory(GiB)": 111.15, + "step": 19455, + "train_speed(iter/s)": 0.448948 + }, + { + "acc": 0.65719013, + "epoch": 0.49365804160324706, + "grad_norm": 5.6875, + "learning_rate": 8.977688921493799e-06, + "loss": 1.57431412, + "memory(GiB)": 111.15, + "step": 19460, + "train_speed(iter/s)": 0.449009 + }, + { + "acc": 0.657513, + "epoch": 0.49378488077118216, + "grad_norm": 4.75, + "learning_rate": 8.977053468758659e-06, + "loss": 1.56710997, + "memory(GiB)": 111.15, + "step": 19465, + "train_speed(iter/s)": 0.44907 + }, + { + "acc": 0.65799584, + "epoch": 0.4939117199391172, + "grad_norm": 5.5625, + "learning_rate": 8.976417841094949e-06, + "loss": 1.64296818, + "memory(GiB)": 111.15, + "step": 19470, + "train_speed(iter/s)": 0.449132 + }, + { + "acc": 0.65312796, + "epoch": 0.49403855910705224, + "grad_norm": 5.59375, + "learning_rate": 8.97578203853063e-06, + "loss": 1.63040695, + "memory(GiB)": 111.15, + "step": 19475, + "train_speed(iter/s)": 0.449193 + }, + { + "acc": 0.65051908, + "epoch": 0.4941653982749873, + "grad_norm": 4.6875, + "learning_rate": 8.975146061093667e-06, + "loss": 1.61847382, + "memory(GiB)": 111.15, + "step": 19480, + "train_speed(iter/s)": 0.449254 + }, + { + "acc": 0.63925037, + "epoch": 0.4942922374429224, + "grad_norm": 5.28125, + "learning_rate": 8.97450990881203e-06, + "loss": 1.72094994, + "memory(GiB)": 111.15, + "step": 19485, + "train_speed(iter/s)": 0.449315 + }, + { + "acc": 0.66057014, + "epoch": 0.49441907661085743, + "grad_norm": 6.46875, + "learning_rate": 8.973873581713705e-06, + "loss": 1.60698795, + "memory(GiB)": 111.15, + "step": 19490, + "train_speed(iter/s)": 0.449376 + }, + { + "acc": 0.65537696, + "epoch": 0.4945459157787925, + "grad_norm": 5.4375, + "learning_rate": 8.973237079826676e-06, + "loss": 1.63585701, + "memory(GiB)": 111.15, + "step": 19495, + "train_speed(iter/s)": 0.449436 + }, + { + "acc": 0.64741716, + "epoch": 0.4946727549467275, + "grad_norm": 6.53125, + "learning_rate": 8.972600403178941e-06, + "loss": 1.65512238, + "memory(GiB)": 111.15, + "step": 19500, + "train_speed(iter/s)": 0.449497 + }, + { + "epoch": 0.4946727549467275, + "eval_acc": 0.6410497322738198, + "eval_loss": 1.6204440593719482, + "eval_runtime": 113.1642, + "eval_samples_per_second": 56.29, + "eval_steps_per_second": 28.145, + "step": 19500 + }, + { + "acc": 0.64733772, + "epoch": 0.4947995941146626, + "grad_norm": 5.90625, + "learning_rate": 8.971963551798506e-06, + "loss": 1.59074268, + "memory(GiB)": 111.15, + "step": 19505, + "train_speed(iter/s)": 0.448342 + }, + { + "acc": 0.65378852, + "epoch": 0.49492643328259767, + "grad_norm": 5.28125, + "learning_rate": 8.971326525713378e-06, + "loss": 1.58030519, + "memory(GiB)": 111.15, + "step": 19510, + "train_speed(iter/s)": 0.448402 + }, + { + "acc": 0.64739265, + "epoch": 0.4950532724505327, + "grad_norm": 5.90625, + "learning_rate": 8.97068932495158e-06, + "loss": 1.66038361, + "memory(GiB)": 111.15, + "step": 19515, + "train_speed(iter/s)": 0.448462 + }, + { + "acc": 0.64897904, + "epoch": 0.49518011161846776, + "grad_norm": 7.625, + "learning_rate": 8.970051949541137e-06, + "loss": 1.65075798, + "memory(GiB)": 111.15, + "step": 19520, + "train_speed(iter/s)": 0.448523 + }, + { + "acc": 0.65034771, + "epoch": 0.49530695078640286, + "grad_norm": 5.375, + "learning_rate": 8.969414399510085e-06, + "loss": 1.6129776, + "memory(GiB)": 111.15, + "step": 19525, + "train_speed(iter/s)": 0.448583 + }, + { + "acc": 0.66213326, + "epoch": 0.4954337899543379, + "grad_norm": 4.75, + "learning_rate": 8.968776674886466e-06, + "loss": 1.61632652, + "memory(GiB)": 111.15, + "step": 19530, + "train_speed(iter/s)": 0.448644 + }, + { + "acc": 0.62931166, + "epoch": 0.49556062912227294, + "grad_norm": 7.78125, + "learning_rate": 8.968138775698328e-06, + "loss": 1.71390076, + "memory(GiB)": 111.15, + "step": 19535, + "train_speed(iter/s)": 0.448705 + }, + { + "acc": 0.66008778, + "epoch": 0.495687468290208, + "grad_norm": 5.65625, + "learning_rate": 8.96750070197373e-06, + "loss": 1.62999687, + "memory(GiB)": 111.15, + "step": 19540, + "train_speed(iter/s)": 0.448766 + }, + { + "acc": 0.64180026, + "epoch": 0.4958143074581431, + "grad_norm": 5.5, + "learning_rate": 8.966862453740738e-06, + "loss": 1.63300514, + "memory(GiB)": 111.15, + "step": 19545, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.6468627, + "epoch": 0.49594114662607813, + "grad_norm": 4.6875, + "learning_rate": 8.966224031027426e-06, + "loss": 1.67881184, + "memory(GiB)": 111.15, + "step": 19550, + "train_speed(iter/s)": 0.448888 + }, + { + "acc": 0.64234486, + "epoch": 0.4960679857940132, + "grad_norm": 5.53125, + "learning_rate": 8.965585433861871e-06, + "loss": 1.65521164, + "memory(GiB)": 111.15, + "step": 19555, + "train_speed(iter/s)": 0.448948 + }, + { + "acc": 0.6452858, + "epoch": 0.4961948249619482, + "grad_norm": 5.375, + "learning_rate": 8.964946662272167e-06, + "loss": 1.59792099, + "memory(GiB)": 111.15, + "step": 19560, + "train_speed(iter/s)": 0.449009 + }, + { + "acc": 0.65219803, + "epoch": 0.4963216641298833, + "grad_norm": 5.53125, + "learning_rate": 8.964307716286404e-06, + "loss": 1.59632511, + "memory(GiB)": 111.15, + "step": 19565, + "train_speed(iter/s)": 0.449069 + }, + { + "acc": 0.65406055, + "epoch": 0.49644850329781837, + "grad_norm": 6.09375, + "learning_rate": 8.963668595932689e-06, + "loss": 1.66682625, + "memory(GiB)": 111.15, + "step": 19570, + "train_speed(iter/s)": 0.44913 + }, + { + "acc": 0.64594746, + "epoch": 0.4965753424657534, + "grad_norm": 4.78125, + "learning_rate": 8.96302930123913e-06, + "loss": 1.63992786, + "memory(GiB)": 111.15, + "step": 19575, + "train_speed(iter/s)": 0.44919 + }, + { + "acc": 0.65579138, + "epoch": 0.49670218163368846, + "grad_norm": 5.125, + "learning_rate": 8.962389832233853e-06, + "loss": 1.61381836, + "memory(GiB)": 111.15, + "step": 19580, + "train_speed(iter/s)": 0.449251 + }, + { + "acc": 0.63740091, + "epoch": 0.49682902080162356, + "grad_norm": 7.875, + "learning_rate": 8.961750188944978e-06, + "loss": 1.71557102, + "memory(GiB)": 111.15, + "step": 19585, + "train_speed(iter/s)": 0.449312 + }, + { + "acc": 0.64190941, + "epoch": 0.4969558599695586, + "grad_norm": 5.5, + "learning_rate": 8.96111037140064e-06, + "loss": 1.64288044, + "memory(GiB)": 111.15, + "step": 19590, + "train_speed(iter/s)": 0.449372 + }, + { + "acc": 0.64492645, + "epoch": 0.49708269913749364, + "grad_norm": 4.59375, + "learning_rate": 8.960470379628986e-06, + "loss": 1.58975945, + "memory(GiB)": 111.15, + "step": 19595, + "train_speed(iter/s)": 0.449433 + }, + { + "acc": 0.64249325, + "epoch": 0.4972095383054287, + "grad_norm": 5.71875, + "learning_rate": 8.959830213658161e-06, + "loss": 1.62804146, + "memory(GiB)": 111.15, + "step": 19600, + "train_speed(iter/s)": 0.449493 + }, + { + "epoch": 0.4972095383054287, + "eval_acc": 0.6411128151375478, + "eval_loss": 1.6208021640777588, + "eval_runtime": 113.4802, + "eval_samples_per_second": 56.133, + "eval_steps_per_second": 28.067, + "step": 19600 + }, + { + "acc": 0.64347668, + "epoch": 0.4973363774733638, + "grad_norm": 5.40625, + "learning_rate": 8.959189873516324e-06, + "loss": 1.70902748, + "memory(GiB)": 111.15, + "step": 19605, + "train_speed(iter/s)": 0.448341 + }, + { + "acc": 0.6401154, + "epoch": 0.49746321664129883, + "grad_norm": 4.9375, + "learning_rate": 8.95854935923164e-06, + "loss": 1.71167965, + "memory(GiB)": 111.15, + "step": 19610, + "train_speed(iter/s)": 0.448401 + }, + { + "acc": 0.65336165, + "epoch": 0.4975900558092339, + "grad_norm": 6.875, + "learning_rate": 8.95790867083228e-06, + "loss": 1.67354736, + "memory(GiB)": 111.15, + "step": 19615, + "train_speed(iter/s)": 0.448462 + }, + { + "acc": 0.6361618, + "epoch": 0.4977168949771689, + "grad_norm": 6.4375, + "learning_rate": 8.957267808346428e-06, + "loss": 1.68470726, + "memory(GiB)": 111.15, + "step": 19620, + "train_speed(iter/s)": 0.448523 + }, + { + "acc": 0.65967007, + "epoch": 0.497843734145104, + "grad_norm": 4.84375, + "learning_rate": 8.95662677180227e-06, + "loss": 1.60451908, + "memory(GiB)": 111.15, + "step": 19625, + "train_speed(iter/s)": 0.448583 + }, + { + "acc": 0.6553791, + "epoch": 0.49797057331303907, + "grad_norm": 5.28125, + "learning_rate": 8.955985561228e-06, + "loss": 1.60309162, + "memory(GiB)": 111.15, + "step": 19630, + "train_speed(iter/s)": 0.448643 + }, + { + "acc": 0.65983362, + "epoch": 0.4980974124809741, + "grad_norm": 6.03125, + "learning_rate": 8.955344176651824e-06, + "loss": 1.63896217, + "memory(GiB)": 111.15, + "step": 19635, + "train_speed(iter/s)": 0.448704 + }, + { + "acc": 0.65563126, + "epoch": 0.49822425164890916, + "grad_norm": 6.46875, + "learning_rate": 8.954702618101952e-06, + "loss": 1.62643986, + "memory(GiB)": 111.15, + "step": 19640, + "train_speed(iter/s)": 0.448765 + }, + { + "acc": 0.63512735, + "epoch": 0.49835109081684426, + "grad_norm": 6.84375, + "learning_rate": 8.9540608856066e-06, + "loss": 1.65276241, + "memory(GiB)": 111.15, + "step": 19645, + "train_speed(iter/s)": 0.448825 + }, + { + "acc": 0.63559399, + "epoch": 0.4984779299847793, + "grad_norm": 6.09375, + "learning_rate": 8.953418979194e-06, + "loss": 1.72147312, + "memory(GiB)": 111.15, + "step": 19650, + "train_speed(iter/s)": 0.448885 + }, + { + "acc": 0.64839697, + "epoch": 0.49860476915271434, + "grad_norm": 5.28125, + "learning_rate": 8.95277689889238e-06, + "loss": 1.60835972, + "memory(GiB)": 111.15, + "step": 19655, + "train_speed(iter/s)": 0.448945 + }, + { + "acc": 0.64133368, + "epoch": 0.4987316083206494, + "grad_norm": 5.9375, + "learning_rate": 8.952134644729985e-06, + "loss": 1.67530327, + "memory(GiB)": 111.15, + "step": 19660, + "train_speed(iter/s)": 0.449006 + }, + { + "acc": 0.65316854, + "epoch": 0.4988584474885845, + "grad_norm": 5.78125, + "learning_rate": 8.951492216735062e-06, + "loss": 1.60997486, + "memory(GiB)": 111.15, + "step": 19665, + "train_speed(iter/s)": 0.449066 + }, + { + "acc": 0.63017092, + "epoch": 0.49898528665651953, + "grad_norm": 4.84375, + "learning_rate": 8.950849614935872e-06, + "loss": 1.66038017, + "memory(GiB)": 111.15, + "step": 19670, + "train_speed(iter/s)": 0.449126 + }, + { + "acc": 0.63417425, + "epoch": 0.4991121258244546, + "grad_norm": 5.9375, + "learning_rate": 8.950206839360674e-06, + "loss": 1.71010208, + "memory(GiB)": 111.15, + "step": 19675, + "train_speed(iter/s)": 0.449186 + }, + { + "acc": 0.64172668, + "epoch": 0.4992389649923896, + "grad_norm": 4.625, + "learning_rate": 8.949563890037745e-06, + "loss": 1.68182487, + "memory(GiB)": 111.15, + "step": 19680, + "train_speed(iter/s)": 0.449246 + }, + { + "acc": 0.63474569, + "epoch": 0.4993658041603247, + "grad_norm": 5.46875, + "learning_rate": 8.948920766995362e-06, + "loss": 1.6851017, + "memory(GiB)": 111.15, + "step": 19685, + "train_speed(iter/s)": 0.449307 + }, + { + "acc": 0.64971361, + "epoch": 0.49949264332825977, + "grad_norm": 5.625, + "learning_rate": 8.948277470261812e-06, + "loss": 1.62977314, + "memory(GiB)": 111.15, + "step": 19690, + "train_speed(iter/s)": 0.449367 + }, + { + "acc": 0.63511553, + "epoch": 0.4996194824961948, + "grad_norm": 5.46875, + "learning_rate": 8.94763399986539e-06, + "loss": 1.65221558, + "memory(GiB)": 111.15, + "step": 19695, + "train_speed(iter/s)": 0.449427 + }, + { + "acc": 0.65666561, + "epoch": 0.49974632166412986, + "grad_norm": 4.5625, + "learning_rate": 8.946990355834401e-06, + "loss": 1.59745445, + "memory(GiB)": 111.15, + "step": 19700, + "train_speed(iter/s)": 0.449488 + }, + { + "epoch": 0.49974632166412986, + "eval_acc": 0.641140387779972, + "eval_loss": 1.6207466125488281, + "eval_runtime": 113.0405, + "eval_samples_per_second": 56.351, + "eval_steps_per_second": 28.176, + "step": 19700 + }, + { + "acc": 0.65152073, + "epoch": 0.49987316083206496, + "grad_norm": 5.28125, + "learning_rate": 8.946346538197156e-06, + "loss": 1.60887604, + "memory(GiB)": 111.15, + "step": 19705, + "train_speed(iter/s)": 0.448346 + }, + { + "acc": 0.64312763, + "epoch": 0.5, + "grad_norm": 5.0625, + "learning_rate": 8.94570254698197e-06, + "loss": 1.66292133, + "memory(GiB)": 111.15, + "step": 19710, + "train_speed(iter/s)": 0.448406 + }, + { + "acc": 0.64098616, + "epoch": 0.5001268391679351, + "grad_norm": 5.15625, + "learning_rate": 8.945058382217168e-06, + "loss": 1.65659752, + "memory(GiB)": 111.15, + "step": 19715, + "train_speed(iter/s)": 0.448467 + }, + { + "acc": 0.63804965, + "epoch": 0.5002536783358701, + "grad_norm": 5.0625, + "learning_rate": 8.944414043931086e-06, + "loss": 1.72590218, + "memory(GiB)": 111.15, + "step": 19720, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.65472589, + "epoch": 0.5003805175038052, + "grad_norm": 4.21875, + "learning_rate": 8.943769532152065e-06, + "loss": 1.59081125, + "memory(GiB)": 111.15, + "step": 19725, + "train_speed(iter/s)": 0.448587 + }, + { + "acc": 0.64078412, + "epoch": 0.5005073566717403, + "grad_norm": 5.1875, + "learning_rate": 8.94312484690845e-06, + "loss": 1.63936729, + "memory(GiB)": 111.15, + "step": 19730, + "train_speed(iter/s)": 0.448648 + }, + { + "acc": 0.64954872, + "epoch": 0.5006341958396753, + "grad_norm": 5.46875, + "learning_rate": 8.9424799882286e-06, + "loss": 1.66516342, + "memory(GiB)": 111.15, + "step": 19735, + "train_speed(iter/s)": 0.448709 + }, + { + "acc": 0.65834932, + "epoch": 0.5007610350076104, + "grad_norm": 4.5625, + "learning_rate": 8.94183495614088e-06, + "loss": 1.6083374, + "memory(GiB)": 111.15, + "step": 19740, + "train_speed(iter/s)": 0.448768 + }, + { + "acc": 0.6696136, + "epoch": 0.5008878741755454, + "grad_norm": 5.21875, + "learning_rate": 8.941189750673658e-06, + "loss": 1.56559277, + "memory(GiB)": 111.15, + "step": 19745, + "train_speed(iter/s)": 0.448828 + }, + { + "acc": 0.639182, + "epoch": 0.5010147133434805, + "grad_norm": 4.90625, + "learning_rate": 8.940544371855315e-06, + "loss": 1.64966278, + "memory(GiB)": 111.15, + "step": 19750, + "train_speed(iter/s)": 0.448889 + }, + { + "acc": 0.65417347, + "epoch": 0.5011415525114156, + "grad_norm": 5.15625, + "learning_rate": 8.939898819714237e-06, + "loss": 1.5684557, + "memory(GiB)": 111.15, + "step": 19755, + "train_speed(iter/s)": 0.448949 + }, + { + "acc": 0.65219088, + "epoch": 0.5012683916793506, + "grad_norm": 4.78125, + "learning_rate": 8.93925309427882e-06, + "loss": 1.62795448, + "memory(GiB)": 111.15, + "step": 19760, + "train_speed(iter/s)": 0.44901 + }, + { + "acc": 0.64510913, + "epoch": 0.5013952308472857, + "grad_norm": 5.46875, + "learning_rate": 8.938607195577462e-06, + "loss": 1.68277283, + "memory(GiB)": 111.15, + "step": 19765, + "train_speed(iter/s)": 0.44907 + }, + { + "acc": 0.63969193, + "epoch": 0.5015220700152208, + "grad_norm": 4.78125, + "learning_rate": 8.937961123638577e-06, + "loss": 1.70045528, + "memory(GiB)": 111.15, + "step": 19770, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.64736881, + "epoch": 0.5016489091831557, + "grad_norm": 5.5625, + "learning_rate": 8.93731487849058e-06, + "loss": 1.65596733, + "memory(GiB)": 111.15, + "step": 19775, + "train_speed(iter/s)": 0.449191 + }, + { + "acc": 0.63354292, + "epoch": 0.5017757483510908, + "grad_norm": 5.9375, + "learning_rate": 8.936668460161895e-06, + "loss": 1.67505112, + "memory(GiB)": 111.15, + "step": 19780, + "train_speed(iter/s)": 0.449252 + }, + { + "acc": 0.65068741, + "epoch": 0.5019025875190258, + "grad_norm": 6.1875, + "learning_rate": 8.936021868680956e-06, + "loss": 1.72361488, + "memory(GiB)": 111.15, + "step": 19785, + "train_speed(iter/s)": 0.449311 + }, + { + "acc": 0.63680954, + "epoch": 0.5020294266869609, + "grad_norm": 5.90625, + "learning_rate": 8.935375104076201e-06, + "loss": 1.71470222, + "memory(GiB)": 111.15, + "step": 19790, + "train_speed(iter/s)": 0.449371 + }, + { + "acc": 0.64843993, + "epoch": 0.502156265854896, + "grad_norm": 4.875, + "learning_rate": 8.93472816637608e-06, + "loss": 1.6471674, + "memory(GiB)": 111.15, + "step": 19795, + "train_speed(iter/s)": 0.449431 + }, + { + "acc": 0.65148096, + "epoch": 0.502283105022831, + "grad_norm": 5.1875, + "learning_rate": 8.934081055609046e-06, + "loss": 1.63630924, + "memory(GiB)": 111.15, + "step": 19800, + "train_speed(iter/s)": 0.449491 + }, + { + "epoch": 0.502283105022831, + "eval_acc": 0.6411412233145909, + "eval_loss": 1.6205523014068604, + "eval_runtime": 113.4099, + "eval_samples_per_second": 56.168, + "eval_steps_per_second": 28.084, + "step": 19800 + }, + { + "acc": 0.65779028, + "epoch": 0.5024099441907661, + "grad_norm": 4.96875, + "learning_rate": 8.933433771803562e-06, + "loss": 1.61116714, + "memory(GiB)": 111.15, + "step": 19805, + "train_speed(iter/s)": 0.44835 + }, + { + "acc": 0.64732332, + "epoch": 0.5025367833587012, + "grad_norm": 4.90625, + "learning_rate": 8.932786314988099e-06, + "loss": 1.63017464, + "memory(GiB)": 111.15, + "step": 19810, + "train_speed(iter/s)": 0.448409 + }, + { + "acc": 0.65393696, + "epoch": 0.5026636225266362, + "grad_norm": 6.4375, + "learning_rate": 8.932138685191136e-06, + "loss": 1.64114685, + "memory(GiB)": 111.15, + "step": 19815, + "train_speed(iter/s)": 0.448469 + }, + { + "acc": 0.66011, + "epoch": 0.5027904616945713, + "grad_norm": 6.09375, + "learning_rate": 8.931490882441159e-06, + "loss": 1.61415825, + "memory(GiB)": 111.15, + "step": 19820, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.63887177, + "epoch": 0.5029173008625063, + "grad_norm": 5.78125, + "learning_rate": 8.930842906766659e-06, + "loss": 1.67127571, + "memory(GiB)": 111.15, + "step": 19825, + "train_speed(iter/s)": 0.448588 + }, + { + "acc": 0.66116543, + "epoch": 0.5030441400304414, + "grad_norm": 5.65625, + "learning_rate": 8.930194758196138e-06, + "loss": 1.58702316, + "memory(GiB)": 111.15, + "step": 19830, + "train_speed(iter/s)": 0.448647 + }, + { + "acc": 0.63884401, + "epoch": 0.5031709791983765, + "grad_norm": 4.71875, + "learning_rate": 8.929546436758105e-06, + "loss": 1.67797718, + "memory(GiB)": 111.15, + "step": 19835, + "train_speed(iter/s)": 0.448707 + }, + { + "acc": 0.64916, + "epoch": 0.5032978183663115, + "grad_norm": 4.875, + "learning_rate": 8.928897942481075e-06, + "loss": 1.61680603, + "memory(GiB)": 111.15, + "step": 19840, + "train_speed(iter/s)": 0.448766 + }, + { + "acc": 0.63736629, + "epoch": 0.5034246575342466, + "grad_norm": 6.90625, + "learning_rate": 8.928249275393572e-06, + "loss": 1.6150526, + "memory(GiB)": 111.15, + "step": 19845, + "train_speed(iter/s)": 0.448826 + }, + { + "acc": 0.64750786, + "epoch": 0.5035514967021817, + "grad_norm": 5.15625, + "learning_rate": 8.927600435524129e-06, + "loss": 1.5982069, + "memory(GiB)": 111.15, + "step": 19850, + "train_speed(iter/s)": 0.448886 + }, + { + "acc": 0.64708862, + "epoch": 0.5036783358701167, + "grad_norm": 5.625, + "learning_rate": 8.926951422901282e-06, + "loss": 1.67548809, + "memory(GiB)": 111.15, + "step": 19855, + "train_speed(iter/s)": 0.448946 + }, + { + "acc": 0.6586225, + "epoch": 0.5038051750380518, + "grad_norm": 7.125, + "learning_rate": 8.926302237553578e-06, + "loss": 1.61093044, + "memory(GiB)": 111.15, + "step": 19860, + "train_speed(iter/s)": 0.449006 + }, + { + "acc": 0.64287481, + "epoch": 0.5039320142059868, + "grad_norm": 5.75, + "learning_rate": 8.925652879509575e-06, + "loss": 1.61765022, + "memory(GiB)": 111.15, + "step": 19865, + "train_speed(iter/s)": 0.449067 + }, + { + "acc": 0.65285764, + "epoch": 0.5040588533739219, + "grad_norm": 5.46875, + "learning_rate": 8.925003348797829e-06, + "loss": 1.68288212, + "memory(GiB)": 111.15, + "step": 19870, + "train_speed(iter/s)": 0.449127 + }, + { + "acc": 0.65032539, + "epoch": 0.504185692541857, + "grad_norm": 5.0625, + "learning_rate": 8.924353645446912e-06, + "loss": 1.56781807, + "memory(GiB)": 111.15, + "step": 19875, + "train_speed(iter/s)": 0.449187 + }, + { + "acc": 0.64459152, + "epoch": 0.504312531709792, + "grad_norm": 5.0625, + "learning_rate": 8.923703769485403e-06, + "loss": 1.62718544, + "memory(GiB)": 111.15, + "step": 19880, + "train_speed(iter/s)": 0.449247 + }, + { + "acc": 0.65369778, + "epoch": 0.5044393708777271, + "grad_norm": 5.6875, + "learning_rate": 8.92305372094188e-06, + "loss": 1.65368176, + "memory(GiB)": 111.15, + "step": 19885, + "train_speed(iter/s)": 0.449307 + }, + { + "acc": 0.6608572, + "epoch": 0.5045662100456622, + "grad_norm": 5.5625, + "learning_rate": 8.922403499844943e-06, + "loss": 1.55294876, + "memory(GiB)": 111.15, + "step": 19890, + "train_speed(iter/s)": 0.449366 + }, + { + "acc": 0.64980078, + "epoch": 0.5046930492135971, + "grad_norm": 5.625, + "learning_rate": 8.921753106223186e-06, + "loss": 1.62799778, + "memory(GiB)": 111.15, + "step": 19895, + "train_speed(iter/s)": 0.449426 + }, + { + "acc": 0.6211278, + "epoch": 0.5048198883815322, + "grad_norm": 5.71875, + "learning_rate": 8.92110254010522e-06, + "loss": 1.78864441, + "memory(GiB)": 111.15, + "step": 19900, + "train_speed(iter/s)": 0.449486 + }, + { + "epoch": 0.5048198883815322, + "eval_acc": 0.6410509855757481, + "eval_loss": 1.6204253435134888, + "eval_runtime": 113.8356, + "eval_samples_per_second": 55.958, + "eval_steps_per_second": 27.979, + "step": 19900 + }, + { + "acc": 0.63189363, + "epoch": 0.5049467275494672, + "grad_norm": 4.375, + "learning_rate": 8.920451801519656e-06, + "loss": 1.72856541, + "memory(GiB)": 111.15, + "step": 19905, + "train_speed(iter/s)": 0.448346 + }, + { + "acc": 0.64705267, + "epoch": 0.5050735667174023, + "grad_norm": 6.46875, + "learning_rate": 8.919800890495118e-06, + "loss": 1.62633934, + "memory(GiB)": 111.15, + "step": 19910, + "train_speed(iter/s)": 0.448405 + }, + { + "acc": 0.64308343, + "epoch": 0.5052004058853374, + "grad_norm": 5.0625, + "learning_rate": 8.919149807060237e-06, + "loss": 1.67277737, + "memory(GiB)": 111.15, + "step": 19915, + "train_speed(iter/s)": 0.448464 + }, + { + "acc": 0.64374957, + "epoch": 0.5053272450532724, + "grad_norm": 5.59375, + "learning_rate": 8.918498551243649e-06, + "loss": 1.70391808, + "memory(GiB)": 111.15, + "step": 19920, + "train_speed(iter/s)": 0.448522 + }, + { + "acc": 0.64195366, + "epoch": 0.5054540842212075, + "grad_norm": 4.75, + "learning_rate": 8.917847123073999e-06, + "loss": 1.65286789, + "memory(GiB)": 111.15, + "step": 19925, + "train_speed(iter/s)": 0.448581 + }, + { + "acc": 0.64615202, + "epoch": 0.5055809233891426, + "grad_norm": 6.53125, + "learning_rate": 8.917195522579943e-06, + "loss": 1.62834053, + "memory(GiB)": 111.15, + "step": 19930, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.64980145, + "epoch": 0.5057077625570776, + "grad_norm": 5.40625, + "learning_rate": 8.916543749790139e-06, + "loss": 1.62496815, + "memory(GiB)": 111.15, + "step": 19935, + "train_speed(iter/s)": 0.448699 + }, + { + "acc": 0.64302568, + "epoch": 0.5058346017250127, + "grad_norm": 5.90625, + "learning_rate": 8.915891804733253e-06, + "loss": 1.6766943, + "memory(GiB)": 111.15, + "step": 19940, + "train_speed(iter/s)": 0.448758 + }, + { + "acc": 0.65832748, + "epoch": 0.5059614408929477, + "grad_norm": 5.21875, + "learning_rate": 8.915239687437963e-06, + "loss": 1.58243637, + "memory(GiB)": 111.15, + "step": 19945, + "train_speed(iter/s)": 0.448816 + }, + { + "acc": 0.6488071, + "epoch": 0.5060882800608828, + "grad_norm": 4.75, + "learning_rate": 8.91458739793295e-06, + "loss": 1.61363945, + "memory(GiB)": 111.15, + "step": 19950, + "train_speed(iter/s)": 0.448874 + }, + { + "acc": 0.64655004, + "epoch": 0.5062151192288179, + "grad_norm": 5.46875, + "learning_rate": 8.913934936246907e-06, + "loss": 1.67589722, + "memory(GiB)": 111.15, + "step": 19955, + "train_speed(iter/s)": 0.448933 + }, + { + "acc": 0.63078566, + "epoch": 0.5063419583967529, + "grad_norm": 5.1875, + "learning_rate": 8.91328230240853e-06, + "loss": 1.71206779, + "memory(GiB)": 111.15, + "step": 19960, + "train_speed(iter/s)": 0.448992 + }, + { + "acc": 0.65314703, + "epoch": 0.506468797564688, + "grad_norm": 5.46875, + "learning_rate": 8.912629496446528e-06, + "loss": 1.60752449, + "memory(GiB)": 111.15, + "step": 19965, + "train_speed(iter/s)": 0.44905 + }, + { + "acc": 0.65357242, + "epoch": 0.5065956367326231, + "grad_norm": 6.9375, + "learning_rate": 8.911976518389612e-06, + "loss": 1.55129414, + "memory(GiB)": 111.15, + "step": 19970, + "train_speed(iter/s)": 0.449109 + }, + { + "acc": 0.66091118, + "epoch": 0.5067224759005581, + "grad_norm": 5.53125, + "learning_rate": 8.9113233682665e-06, + "loss": 1.59045334, + "memory(GiB)": 111.15, + "step": 19975, + "train_speed(iter/s)": 0.449168 + }, + { + "acc": 0.64498329, + "epoch": 0.5068493150684932, + "grad_norm": 4.71875, + "learning_rate": 8.910670046105927e-06, + "loss": 1.62249432, + "memory(GiB)": 111.15, + "step": 19980, + "train_speed(iter/s)": 0.449227 + }, + { + "acc": 0.6498518, + "epoch": 0.5069761542364282, + "grad_norm": 5.53125, + "learning_rate": 8.910016551936623e-06, + "loss": 1.65905304, + "memory(GiB)": 111.15, + "step": 19985, + "train_speed(iter/s)": 0.449286 + }, + { + "acc": 0.64149647, + "epoch": 0.5071029934043633, + "grad_norm": 5.40625, + "learning_rate": 8.909362885787333e-06, + "loss": 1.67586403, + "memory(GiB)": 111.15, + "step": 19990, + "train_speed(iter/s)": 0.449345 + }, + { + "acc": 0.63777685, + "epoch": 0.5072298325722984, + "grad_norm": 5.0625, + "learning_rate": 8.908709047686813e-06, + "loss": 1.69009876, + "memory(GiB)": 111.15, + "step": 19995, + "train_speed(iter/s)": 0.449403 + }, + { + "acc": 0.65144787, + "epoch": 0.5073566717402334, + "grad_norm": 6.1875, + "learning_rate": 8.908055037663815e-06, + "loss": 1.64033413, + "memory(GiB)": 111.15, + "step": 20000, + "train_speed(iter/s)": 0.449463 + }, + { + "epoch": 0.5073566717402334, + "eval_acc": 0.6412749088536173, + "eval_loss": 1.6201322078704834, + "eval_runtime": 114.879, + "eval_samples_per_second": 55.45, + "eval_steps_per_second": 27.725, + "step": 20000 + }, + { + "acc": 0.65780358, + "epoch": 0.5074835109081685, + "grad_norm": 5.53125, + "learning_rate": 8.907400855747111e-06, + "loss": 1.63554573, + "memory(GiB)": 111.15, + "step": 20005, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.61898642, + "epoch": 0.5076103500761036, + "grad_norm": 4.875, + "learning_rate": 8.906746501965468e-06, + "loss": 1.73053017, + "memory(GiB)": 111.15, + "step": 20010, + "train_speed(iter/s)": 0.448378 + }, + { + "acc": 0.66952753, + "epoch": 0.5077371892440385, + "grad_norm": 6.1875, + "learning_rate": 8.906091976347675e-06, + "loss": 1.57055092, + "memory(GiB)": 111.15, + "step": 20015, + "train_speed(iter/s)": 0.448438 + }, + { + "acc": 0.63644347, + "epoch": 0.5078640284119736, + "grad_norm": 4.59375, + "learning_rate": 8.905437278922516e-06, + "loss": 1.64243927, + "memory(GiB)": 111.15, + "step": 20020, + "train_speed(iter/s)": 0.448497 + }, + { + "acc": 0.65560164, + "epoch": 0.5079908675799086, + "grad_norm": 5.96875, + "learning_rate": 8.90478240971879e-06, + "loss": 1.6263195, + "memory(GiB)": 111.15, + "step": 20025, + "train_speed(iter/s)": 0.448556 + }, + { + "acc": 0.65538573, + "epoch": 0.5081177067478437, + "grad_norm": 5.1875, + "learning_rate": 8.904127368765298e-06, + "loss": 1.63288651, + "memory(GiB)": 111.15, + "step": 20030, + "train_speed(iter/s)": 0.448615 + }, + { + "acc": 0.62948251, + "epoch": 0.5082445459157788, + "grad_norm": 6.125, + "learning_rate": 8.903472156090856e-06, + "loss": 1.6879715, + "memory(GiB)": 111.15, + "step": 20035, + "train_speed(iter/s)": 0.448674 + }, + { + "acc": 0.65240479, + "epoch": 0.5083713850837138, + "grad_norm": 4.875, + "learning_rate": 8.902816771724279e-06, + "loss": 1.6113575, + "memory(GiB)": 111.15, + "step": 20040, + "train_speed(iter/s)": 0.448734 + }, + { + "acc": 0.64111795, + "epoch": 0.5084982242516489, + "grad_norm": 4.875, + "learning_rate": 8.902161215694396e-06, + "loss": 1.62289848, + "memory(GiB)": 111.15, + "step": 20045, + "train_speed(iter/s)": 0.448793 + }, + { + "acc": 0.65775142, + "epoch": 0.508625063419584, + "grad_norm": 5.21875, + "learning_rate": 8.901505488030042e-06, + "loss": 1.62933159, + "memory(GiB)": 111.15, + "step": 20050, + "train_speed(iter/s)": 0.448852 + }, + { + "acc": 0.62770581, + "epoch": 0.508751902587519, + "grad_norm": 5.34375, + "learning_rate": 8.900849588760057e-06, + "loss": 1.63833771, + "memory(GiB)": 111.15, + "step": 20055, + "train_speed(iter/s)": 0.448911 + }, + { + "acc": 0.63344884, + "epoch": 0.5088787417554541, + "grad_norm": 5.1875, + "learning_rate": 8.90019351791329e-06, + "loss": 1.68736382, + "memory(GiB)": 111.15, + "step": 20060, + "train_speed(iter/s)": 0.44897 + }, + { + "acc": 0.64177279, + "epoch": 0.5090055809233891, + "grad_norm": 5.84375, + "learning_rate": 8.8995372755186e-06, + "loss": 1.63044662, + "memory(GiB)": 111.15, + "step": 20065, + "train_speed(iter/s)": 0.44903 + }, + { + "acc": 0.64139175, + "epoch": 0.5091324200913242, + "grad_norm": 8.1875, + "learning_rate": 8.89888086160485e-06, + "loss": 1.60529194, + "memory(GiB)": 111.15, + "step": 20070, + "train_speed(iter/s)": 0.449089 + }, + { + "acc": 0.64494324, + "epoch": 0.5092592592592593, + "grad_norm": 5.25, + "learning_rate": 8.898224276200913e-06, + "loss": 1.6623354, + "memory(GiB)": 111.15, + "step": 20075, + "train_speed(iter/s)": 0.449149 + }, + { + "acc": 0.65296164, + "epoch": 0.5093860984271943, + "grad_norm": 4.875, + "learning_rate": 8.897567519335669e-06, + "loss": 1.61010189, + "memory(GiB)": 111.15, + "step": 20080, + "train_speed(iter/s)": 0.449208 + }, + { + "acc": 0.64035749, + "epoch": 0.5095129375951294, + "grad_norm": 5.4375, + "learning_rate": 8.896910591038002e-06, + "loss": 1.64152527, + "memory(GiB)": 111.15, + "step": 20085, + "train_speed(iter/s)": 0.449268 + }, + { + "acc": 0.64090071, + "epoch": 0.5096397767630645, + "grad_norm": 6.28125, + "learning_rate": 8.896253491336809e-06, + "loss": 1.70509491, + "memory(GiB)": 111.15, + "step": 20090, + "train_speed(iter/s)": 0.449327 + }, + { + "acc": 0.6371511, + "epoch": 0.5097666159309995, + "grad_norm": 5.59375, + "learning_rate": 8.895596220260993e-06, + "loss": 1.66529884, + "memory(GiB)": 111.15, + "step": 20095, + "train_speed(iter/s)": 0.449386 + }, + { + "acc": 0.65029869, + "epoch": 0.5098934550989346, + "grad_norm": 6.5, + "learning_rate": 8.894938777839462e-06, + "loss": 1.66585808, + "memory(GiB)": 111.15, + "step": 20100, + "train_speed(iter/s)": 0.449445 + }, + { + "epoch": 0.5098934550989346, + "eval_acc": 0.6412314610534336, + "eval_loss": 1.6199264526367188, + "eval_runtime": 113.7293, + "eval_samples_per_second": 56.01, + "eval_steps_per_second": 28.005, + "step": 20100 + }, + { + "acc": 0.65497208, + "epoch": 0.5100202942668696, + "grad_norm": 5.46875, + "learning_rate": 8.894281164101134e-06, + "loss": 1.62883492, + "memory(GiB)": 111.15, + "step": 20105, + "train_speed(iter/s)": 0.448318 + }, + { + "acc": 0.62209892, + "epoch": 0.5101471334348047, + "grad_norm": 4.78125, + "learning_rate": 8.893623379074934e-06, + "loss": 1.64234848, + "memory(GiB)": 111.15, + "step": 20110, + "train_speed(iter/s)": 0.448377 + }, + { + "acc": 0.6445405, + "epoch": 0.5102739726027398, + "grad_norm": 5.28125, + "learning_rate": 8.892965422789793e-06, + "loss": 1.64053516, + "memory(GiB)": 111.15, + "step": 20115, + "train_speed(iter/s)": 0.448436 + }, + { + "acc": 0.6535265, + "epoch": 0.5104008117706748, + "grad_norm": 6.0625, + "learning_rate": 8.892307295274654e-06, + "loss": 1.67580452, + "memory(GiB)": 111.15, + "step": 20120, + "train_speed(iter/s)": 0.448495 + }, + { + "acc": 0.6519165, + "epoch": 0.5105276509386099, + "grad_norm": 6.15625, + "learning_rate": 8.89164899655846e-06, + "loss": 1.63578949, + "memory(GiB)": 111.15, + "step": 20125, + "train_speed(iter/s)": 0.448555 + }, + { + "acc": 0.67143373, + "epoch": 0.510654490106545, + "grad_norm": 5.1875, + "learning_rate": 8.89099052667017e-06, + "loss": 1.52573853, + "memory(GiB)": 111.15, + "step": 20130, + "train_speed(iter/s)": 0.448613 + }, + { + "acc": 0.6415709, + "epoch": 0.51078132927448, + "grad_norm": 4.8125, + "learning_rate": 8.890331885638744e-06, + "loss": 1.66831455, + "memory(GiB)": 111.15, + "step": 20135, + "train_speed(iter/s)": 0.448672 + }, + { + "acc": 0.65488882, + "epoch": 0.510908168442415, + "grad_norm": 4.84375, + "learning_rate": 8.889673073493151e-06, + "loss": 1.63171463, + "memory(GiB)": 111.15, + "step": 20140, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.65465717, + "epoch": 0.51103500761035, + "grad_norm": 5.71875, + "learning_rate": 8.889014090262371e-06, + "loss": 1.66564236, + "memory(GiB)": 111.15, + "step": 20145, + "train_speed(iter/s)": 0.44879 + }, + { + "acc": 0.64077549, + "epoch": 0.5111618467782851, + "grad_norm": 5.125, + "learning_rate": 8.888354935975388e-06, + "loss": 1.57877159, + "memory(GiB)": 111.15, + "step": 20150, + "train_speed(iter/s)": 0.448848 + }, + { + "acc": 0.65484886, + "epoch": 0.5112886859462202, + "grad_norm": 4.21875, + "learning_rate": 8.887695610661196e-06, + "loss": 1.60701046, + "memory(GiB)": 111.15, + "step": 20155, + "train_speed(iter/s)": 0.448907 + }, + { + "acc": 0.64016342, + "epoch": 0.5114155251141552, + "grad_norm": 5.375, + "learning_rate": 8.887036114348792e-06, + "loss": 1.7001709, + "memory(GiB)": 111.15, + "step": 20160, + "train_speed(iter/s)": 0.448965 + }, + { + "acc": 0.66147265, + "epoch": 0.5115423642820903, + "grad_norm": 5.09375, + "learning_rate": 8.886376447067186e-06, + "loss": 1.65595894, + "memory(GiB)": 111.15, + "step": 20165, + "train_speed(iter/s)": 0.449024 + }, + { + "acc": 0.64104271, + "epoch": 0.5116692034500254, + "grad_norm": 6.46875, + "learning_rate": 8.885716608845394e-06, + "loss": 1.65234413, + "memory(GiB)": 111.15, + "step": 20170, + "train_speed(iter/s)": 0.449083 + }, + { + "acc": 0.65104208, + "epoch": 0.5117960426179604, + "grad_norm": 4.21875, + "learning_rate": 8.885056599712436e-06, + "loss": 1.60193787, + "memory(GiB)": 111.15, + "step": 20175, + "train_speed(iter/s)": 0.449141 + }, + { + "acc": 0.64391069, + "epoch": 0.5119228817858955, + "grad_norm": 5.28125, + "learning_rate": 8.884396419697343e-06, + "loss": 1.68876915, + "memory(GiB)": 111.15, + "step": 20180, + "train_speed(iter/s)": 0.4492 + }, + { + "acc": 0.64758291, + "epoch": 0.5120497209538305, + "grad_norm": 5.6875, + "learning_rate": 8.883736068829151e-06, + "loss": 1.60949268, + "memory(GiB)": 111.15, + "step": 20185, + "train_speed(iter/s)": 0.44926 + }, + { + "acc": 0.64725008, + "epoch": 0.5121765601217656, + "grad_norm": 5.1875, + "learning_rate": 8.88307554713691e-06, + "loss": 1.63280983, + "memory(GiB)": 111.15, + "step": 20190, + "train_speed(iter/s)": 0.449318 + }, + { + "acc": 0.66268158, + "epoch": 0.5123033992897007, + "grad_norm": 5.28125, + "learning_rate": 8.882414854649667e-06, + "loss": 1.60372181, + "memory(GiB)": 111.15, + "step": 20195, + "train_speed(iter/s)": 0.449377 + }, + { + "acc": 0.65276198, + "epoch": 0.5124302384576357, + "grad_norm": 5.09375, + "learning_rate": 8.881753991396488e-06, + "loss": 1.66477737, + "memory(GiB)": 111.15, + "step": 20200, + "train_speed(iter/s)": 0.449436 + }, + { + "epoch": 0.5124302384576357, + "eval_acc": 0.6412690601112848, + "eval_loss": 1.6197140216827393, + "eval_runtime": 113.8311, + "eval_samples_per_second": 55.96, + "eval_steps_per_second": 27.98, + "step": 20200 + }, + { + "acc": 0.6475677, + "epoch": 0.5125570776255708, + "grad_norm": 4.90625, + "learning_rate": 8.881092957406436e-06, + "loss": 1.58843079, + "memory(GiB)": 111.15, + "step": 20205, + "train_speed(iter/s)": 0.448315 + }, + { + "acc": 0.65057421, + "epoch": 0.5126839167935059, + "grad_norm": 5.78125, + "learning_rate": 8.880431752708588e-06, + "loss": 1.58279285, + "memory(GiB)": 111.15, + "step": 20210, + "train_speed(iter/s)": 0.448372 + }, + { + "acc": 0.66456137, + "epoch": 0.5128107559614409, + "grad_norm": 5.53125, + "learning_rate": 8.879770377332026e-06, + "loss": 1.57497177, + "memory(GiB)": 111.15, + "step": 20215, + "train_speed(iter/s)": 0.448431 + }, + { + "acc": 0.64588451, + "epoch": 0.512937595129376, + "grad_norm": 6.34375, + "learning_rate": 8.879108831305842e-06, + "loss": 1.61518478, + "memory(GiB)": 111.15, + "step": 20220, + "train_speed(iter/s)": 0.448489 + }, + { + "acc": 0.64373875, + "epoch": 0.513064434297311, + "grad_norm": 5.34375, + "learning_rate": 8.878447114659131e-06, + "loss": 1.64181957, + "memory(GiB)": 111.15, + "step": 20225, + "train_speed(iter/s)": 0.448547 + }, + { + "acc": 0.66631756, + "epoch": 0.5131912734652461, + "grad_norm": 6.15625, + "learning_rate": 8.877785227421003e-06, + "loss": 1.58298378, + "memory(GiB)": 111.15, + "step": 20230, + "train_speed(iter/s)": 0.448606 + }, + { + "acc": 0.66338458, + "epoch": 0.5133181126331812, + "grad_norm": 5.15625, + "learning_rate": 8.877123169620565e-06, + "loss": 1.61580296, + "memory(GiB)": 111.15, + "step": 20235, + "train_speed(iter/s)": 0.448665 + }, + { + "acc": 0.64220743, + "epoch": 0.5134449518011162, + "grad_norm": 4.8125, + "learning_rate": 8.876460941286941e-06, + "loss": 1.63509293, + "memory(GiB)": 111.15, + "step": 20240, + "train_speed(iter/s)": 0.448723 + }, + { + "acc": 0.6343997, + "epoch": 0.5135717909690513, + "grad_norm": 5.46875, + "learning_rate": 8.875798542449256e-06, + "loss": 1.68862534, + "memory(GiB)": 111.15, + "step": 20245, + "train_speed(iter/s)": 0.448782 + }, + { + "acc": 0.6297996, + "epoch": 0.5136986301369864, + "grad_norm": 5.03125, + "learning_rate": 8.87513597313665e-06, + "loss": 1.68137665, + "memory(GiB)": 111.15, + "step": 20250, + "train_speed(iter/s)": 0.448841 + }, + { + "acc": 0.63260617, + "epoch": 0.5138254693049213, + "grad_norm": 5.4375, + "learning_rate": 8.87447323337826e-06, + "loss": 1.68320732, + "memory(GiB)": 111.15, + "step": 20255, + "train_speed(iter/s)": 0.448899 + }, + { + "acc": 0.65727234, + "epoch": 0.5139523084728564, + "grad_norm": 5.5625, + "learning_rate": 8.87381032320324e-06, + "loss": 1.58720598, + "memory(GiB)": 111.15, + "step": 20260, + "train_speed(iter/s)": 0.448957 + }, + { + "acc": 0.64046593, + "epoch": 0.5140791476407914, + "grad_norm": 5.375, + "learning_rate": 8.873147242640746e-06, + "loss": 1.66223335, + "memory(GiB)": 111.15, + "step": 20265, + "train_speed(iter/s)": 0.449015 + }, + { + "acc": 0.64339094, + "epoch": 0.5142059868087265, + "grad_norm": 4.46875, + "learning_rate": 8.872483991719944e-06, + "loss": 1.6780159, + "memory(GiB)": 111.15, + "step": 20270, + "train_speed(iter/s)": 0.449073 + }, + { + "acc": 0.62995806, + "epoch": 0.5143328259766616, + "grad_norm": 4.4375, + "learning_rate": 8.871820570470009e-06, + "loss": 1.67741699, + "memory(GiB)": 111.15, + "step": 20275, + "train_speed(iter/s)": 0.449132 + }, + { + "acc": 0.64295893, + "epoch": 0.5144596651445966, + "grad_norm": 7.0, + "learning_rate": 8.871156978920116e-06, + "loss": 1.64983196, + "memory(GiB)": 111.15, + "step": 20280, + "train_speed(iter/s)": 0.449191 + }, + { + "acc": 0.66206837, + "epoch": 0.5145865043125317, + "grad_norm": 4.9375, + "learning_rate": 8.870493217099456e-06, + "loss": 1.5314249, + "memory(GiB)": 111.15, + "step": 20285, + "train_speed(iter/s)": 0.449249 + }, + { + "acc": 0.65098386, + "epoch": 0.5147133434804668, + "grad_norm": 5.125, + "learning_rate": 8.869829285037224e-06, + "loss": 1.56703959, + "memory(GiB)": 111.15, + "step": 20290, + "train_speed(iter/s)": 0.449307 + }, + { + "acc": 0.64075227, + "epoch": 0.5148401826484018, + "grad_norm": 6.0625, + "learning_rate": 8.869165182762623e-06, + "loss": 1.67500153, + "memory(GiB)": 111.15, + "step": 20295, + "train_speed(iter/s)": 0.449366 + }, + { + "acc": 0.64885879, + "epoch": 0.5149670218163369, + "grad_norm": 5.28125, + "learning_rate": 8.868500910304863e-06, + "loss": 1.66765575, + "memory(GiB)": 111.15, + "step": 20300, + "train_speed(iter/s)": 0.449425 + }, + { + "epoch": 0.5149670218163369, + "eval_acc": 0.6413221165595859, + "eval_loss": 1.6198632717132568, + "eval_runtime": 114.1877, + "eval_samples_per_second": 55.785, + "eval_steps_per_second": 27.893, + "step": 20300 + }, + { + "acc": 0.64977808, + "epoch": 0.5150938609842719, + "grad_norm": 4.8125, + "learning_rate": 8.86783646769316e-06, + "loss": 1.61024303, + "memory(GiB)": 111.15, + "step": 20305, + "train_speed(iter/s)": 0.448304 + }, + { + "acc": 0.66161852, + "epoch": 0.515220700152207, + "grad_norm": 5.25, + "learning_rate": 8.867171854956742e-06, + "loss": 1.62311554, + "memory(GiB)": 111.15, + "step": 20310, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.64208775, + "epoch": 0.5153475393201421, + "grad_norm": 6.53125, + "learning_rate": 8.86650707212484e-06, + "loss": 1.63755093, + "memory(GiB)": 111.15, + "step": 20315, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.65008821, + "epoch": 0.5154743784880771, + "grad_norm": 5.09375, + "learning_rate": 8.865842119226693e-06, + "loss": 1.65770454, + "memory(GiB)": 111.15, + "step": 20320, + "train_speed(iter/s)": 0.448477 + }, + { + "acc": 0.63754449, + "epoch": 0.5156012176560122, + "grad_norm": 5.65625, + "learning_rate": 8.86517699629155e-06, + "loss": 1.71644325, + "memory(GiB)": 111.15, + "step": 20325, + "train_speed(iter/s)": 0.448535 + }, + { + "acc": 0.63472271, + "epoch": 0.5157280568239473, + "grad_norm": 4.59375, + "learning_rate": 8.864511703348666e-06, + "loss": 1.70413227, + "memory(GiB)": 111.15, + "step": 20330, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.63774271, + "epoch": 0.5158548959918823, + "grad_norm": 5.5, + "learning_rate": 8.863846240427306e-06, + "loss": 1.67435646, + "memory(GiB)": 111.15, + "step": 20335, + "train_speed(iter/s)": 0.448649 + }, + { + "acc": 0.66084795, + "epoch": 0.5159817351598174, + "grad_norm": 5.03125, + "learning_rate": 8.863180607556733e-06, + "loss": 1.53565702, + "memory(GiB)": 111.15, + "step": 20340, + "train_speed(iter/s)": 0.448707 + }, + { + "acc": 0.64474325, + "epoch": 0.5161085743277524, + "grad_norm": 6.0625, + "learning_rate": 8.862514804766234e-06, + "loss": 1.62035618, + "memory(GiB)": 111.15, + "step": 20345, + "train_speed(iter/s)": 0.448765 + }, + { + "acc": 0.63978577, + "epoch": 0.5162354134956875, + "grad_norm": 5.59375, + "learning_rate": 8.861848832085084e-06, + "loss": 1.64842911, + "memory(GiB)": 111.15, + "step": 20350, + "train_speed(iter/s)": 0.448823 + }, + { + "acc": 0.65127201, + "epoch": 0.5163622526636226, + "grad_norm": 5.40625, + "learning_rate": 8.861182689542585e-06, + "loss": 1.60889969, + "memory(GiB)": 111.15, + "step": 20355, + "train_speed(iter/s)": 0.44888 + }, + { + "acc": 0.65274248, + "epoch": 0.5164890918315576, + "grad_norm": 5.4375, + "learning_rate": 8.86051637716803e-06, + "loss": 1.58584824, + "memory(GiB)": 111.15, + "step": 20360, + "train_speed(iter/s)": 0.448938 + }, + { + "acc": 0.63952379, + "epoch": 0.5166159309994927, + "grad_norm": 4.96875, + "learning_rate": 8.859849894990728e-06, + "loss": 1.67411957, + "memory(GiB)": 111.15, + "step": 20365, + "train_speed(iter/s)": 0.448995 + }, + { + "acc": 0.64621215, + "epoch": 0.5167427701674278, + "grad_norm": 5.40625, + "learning_rate": 8.859183243039995e-06, + "loss": 1.67879143, + "memory(GiB)": 111.15, + "step": 20370, + "train_speed(iter/s)": 0.449052 + }, + { + "acc": 0.63781972, + "epoch": 0.5168696093353627, + "grad_norm": 5.84375, + "learning_rate": 8.858516421345154e-06, + "loss": 1.65565376, + "memory(GiB)": 111.15, + "step": 20375, + "train_speed(iter/s)": 0.44911 + }, + { + "acc": 0.64238071, + "epoch": 0.5169964485032978, + "grad_norm": 5.40625, + "learning_rate": 8.857849429935534e-06, + "loss": 1.68447151, + "memory(GiB)": 111.15, + "step": 20380, + "train_speed(iter/s)": 0.449168 + }, + { + "acc": 0.65431566, + "epoch": 0.5171232876712328, + "grad_norm": 6.125, + "learning_rate": 8.857182268840471e-06, + "loss": 1.66377468, + "memory(GiB)": 111.15, + "step": 20385, + "train_speed(iter/s)": 0.449225 + }, + { + "acc": 0.64510574, + "epoch": 0.5172501268391679, + "grad_norm": 5.03125, + "learning_rate": 8.856514938089312e-06, + "loss": 1.60667305, + "memory(GiB)": 111.15, + "step": 20390, + "train_speed(iter/s)": 0.449283 + }, + { + "acc": 0.64003325, + "epoch": 0.517376966007103, + "grad_norm": 5.34375, + "learning_rate": 8.855847437711407e-06, + "loss": 1.70861053, + "memory(GiB)": 111.15, + "step": 20395, + "train_speed(iter/s)": 0.44934 + }, + { + "acc": 0.6424902, + "epoch": 0.517503805175038, + "grad_norm": 5.75, + "learning_rate": 8.855179767736117e-06, + "loss": 1.63393021, + "memory(GiB)": 111.15, + "step": 20400, + "train_speed(iter/s)": 0.449398 + }, + { + "epoch": 0.517503805175038, + "eval_acc": 0.641213497059127, + "eval_loss": 1.6197670698165894, + "eval_runtime": 113.4695, + "eval_samples_per_second": 56.138, + "eval_steps_per_second": 28.069, + "step": 20400 + }, + { + "acc": 0.64450955, + "epoch": 0.5176306443429731, + "grad_norm": 5.09375, + "learning_rate": 8.854511928192808e-06, + "loss": 1.61953964, + "memory(GiB)": 111.15, + "step": 20405, + "train_speed(iter/s)": 0.448292 + }, + { + "acc": 0.63497038, + "epoch": 0.5177574835109082, + "grad_norm": 5.5, + "learning_rate": 8.853843919110856e-06, + "loss": 1.54750137, + "memory(GiB)": 111.15, + "step": 20410, + "train_speed(iter/s)": 0.44835 + }, + { + "acc": 0.64286399, + "epoch": 0.5178843226788432, + "grad_norm": 6.0625, + "learning_rate": 8.853175740519642e-06, + "loss": 1.68529148, + "memory(GiB)": 111.15, + "step": 20415, + "train_speed(iter/s)": 0.448408 + }, + { + "acc": 0.65844083, + "epoch": 0.5180111618467783, + "grad_norm": 5.5, + "learning_rate": 8.852507392448555e-06, + "loss": 1.59052219, + "memory(GiB)": 111.15, + "step": 20420, + "train_speed(iter/s)": 0.448467 + }, + { + "acc": 0.64411945, + "epoch": 0.5181380010147133, + "grad_norm": 6.09375, + "learning_rate": 8.851838874926994e-06, + "loss": 1.66908875, + "memory(GiB)": 111.15, + "step": 20425, + "train_speed(iter/s)": 0.448524 + }, + { + "acc": 0.64412246, + "epoch": 0.5182648401826484, + "grad_norm": 6.28125, + "learning_rate": 8.851170187984362e-06, + "loss": 1.62277298, + "memory(GiB)": 111.15, + "step": 20430, + "train_speed(iter/s)": 0.448582 + }, + { + "acc": 0.65265017, + "epoch": 0.5183916793505835, + "grad_norm": 5.125, + "learning_rate": 8.850501331650069e-06, + "loss": 1.58741951, + "memory(GiB)": 111.15, + "step": 20435, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.64619379, + "epoch": 0.5185185185185185, + "grad_norm": 5.25, + "learning_rate": 8.849832305953536e-06, + "loss": 1.6669632, + "memory(GiB)": 111.15, + "step": 20440, + "train_speed(iter/s)": 0.448698 + }, + { + "acc": 0.65018568, + "epoch": 0.5186453576864536, + "grad_norm": 4.5, + "learning_rate": 8.849163110924193e-06, + "loss": 1.6349102, + "memory(GiB)": 111.15, + "step": 20445, + "train_speed(iter/s)": 0.448756 + }, + { + "acc": 0.63906012, + "epoch": 0.5187721968543887, + "grad_norm": 4.90625, + "learning_rate": 8.84849374659147e-06, + "loss": 1.66595955, + "memory(GiB)": 111.15, + "step": 20450, + "train_speed(iter/s)": 0.448814 + }, + { + "acc": 0.64773979, + "epoch": 0.5188990360223237, + "grad_norm": 4.40625, + "learning_rate": 8.847824212984807e-06, + "loss": 1.56585531, + "memory(GiB)": 111.15, + "step": 20455, + "train_speed(iter/s)": 0.448872 + }, + { + "acc": 0.65745201, + "epoch": 0.5190258751902588, + "grad_norm": 5.1875, + "learning_rate": 8.84715451013366e-06, + "loss": 1.51215305, + "memory(GiB)": 111.15, + "step": 20460, + "train_speed(iter/s)": 0.44893 + }, + { + "acc": 0.65332522, + "epoch": 0.5191527143581938, + "grad_norm": 6.15625, + "learning_rate": 8.846484638067478e-06, + "loss": 1.6267086, + "memory(GiB)": 111.15, + "step": 20465, + "train_speed(iter/s)": 0.448988 + }, + { + "acc": 0.64451418, + "epoch": 0.5192795535261289, + "grad_norm": 4.96875, + "learning_rate": 8.84581459681573e-06, + "loss": 1.63981438, + "memory(GiB)": 111.15, + "step": 20470, + "train_speed(iter/s)": 0.449046 + }, + { + "acc": 0.66904049, + "epoch": 0.519406392694064, + "grad_norm": 5.9375, + "learning_rate": 8.845144386407884e-06, + "loss": 1.59844303, + "memory(GiB)": 111.15, + "step": 20475, + "train_speed(iter/s)": 0.449103 + }, + { + "acc": 0.64480515, + "epoch": 0.519533231861999, + "grad_norm": 4.78125, + "learning_rate": 8.844474006873422e-06, + "loss": 1.65803299, + "memory(GiB)": 111.15, + "step": 20480, + "train_speed(iter/s)": 0.449161 + }, + { + "acc": 0.65620909, + "epoch": 0.5196600710299341, + "grad_norm": 5.78125, + "learning_rate": 8.84380345824183e-06, + "loss": 1.58514194, + "memory(GiB)": 111.15, + "step": 20485, + "train_speed(iter/s)": 0.449219 + }, + { + "acc": 0.64652939, + "epoch": 0.5197869101978692, + "grad_norm": 5.03125, + "learning_rate": 8.843132740542599e-06, + "loss": 1.6645195, + "memory(GiB)": 111.15, + "step": 20490, + "train_speed(iter/s)": 0.449278 + }, + { + "acc": 0.65573711, + "epoch": 0.5199137493658041, + "grad_norm": 5.96875, + "learning_rate": 8.842461853805232e-06, + "loss": 1.55341616, + "memory(GiB)": 111.15, + "step": 20495, + "train_speed(iter/s)": 0.449336 + }, + { + "acc": 0.65555334, + "epoch": 0.5200405885337392, + "grad_norm": 5.21875, + "learning_rate": 8.841790798059237e-06, + "loss": 1.62602539, + "memory(GiB)": 111.15, + "step": 20500, + "train_speed(iter/s)": 0.449394 + }, + { + "epoch": 0.5200405885337392, + "eval_acc": 0.6412636291362619, + "eval_loss": 1.6192383766174316, + "eval_runtime": 113.696, + "eval_samples_per_second": 56.027, + "eval_steps_per_second": 28.013, + "step": 20500 + }, + { + "acc": 0.66065216, + "epoch": 0.5201674277016742, + "grad_norm": 6.28125, + "learning_rate": 8.84111957333413e-06, + "loss": 1.5499794, + "memory(GiB)": 111.15, + "step": 20505, + "train_speed(iter/s)": 0.44829 + }, + { + "acc": 0.63612781, + "epoch": 0.5202942668696093, + "grad_norm": 5.4375, + "learning_rate": 8.840448179659436e-06, + "loss": 1.75245819, + "memory(GiB)": 111.15, + "step": 20510, + "train_speed(iter/s)": 0.448348 + }, + { + "acc": 0.64651113, + "epoch": 0.5204211060375444, + "grad_norm": 4.9375, + "learning_rate": 8.839776617064683e-06, + "loss": 1.66760139, + "memory(GiB)": 111.15, + "step": 20515, + "train_speed(iter/s)": 0.448405 + }, + { + "acc": 0.64848423, + "epoch": 0.5205479452054794, + "grad_norm": 4.875, + "learning_rate": 8.839104885579413e-06, + "loss": 1.62671394, + "memory(GiB)": 111.15, + "step": 20520, + "train_speed(iter/s)": 0.448463 + }, + { + "acc": 0.63955612, + "epoch": 0.5206747843734145, + "grad_norm": 6.03125, + "learning_rate": 8.83843298523317e-06, + "loss": 1.65807877, + "memory(GiB)": 111.15, + "step": 20525, + "train_speed(iter/s)": 0.448521 + }, + { + "acc": 0.64260273, + "epoch": 0.5208016235413496, + "grad_norm": 6.03125, + "learning_rate": 8.837760916055505e-06, + "loss": 1.66262779, + "memory(GiB)": 111.15, + "step": 20530, + "train_speed(iter/s)": 0.448579 + }, + { + "acc": 0.63825531, + "epoch": 0.5209284627092846, + "grad_norm": 5.1875, + "learning_rate": 8.837088678075983e-06, + "loss": 1.70225029, + "memory(GiB)": 111.15, + "step": 20535, + "train_speed(iter/s)": 0.448637 + }, + { + "acc": 0.64030442, + "epoch": 0.5210553018772197, + "grad_norm": 4.625, + "learning_rate": 8.836416271324166e-06, + "loss": 1.65919514, + "memory(GiB)": 111.15, + "step": 20540, + "train_speed(iter/s)": 0.448695 + }, + { + "acc": 0.67224808, + "epoch": 0.5211821410451547, + "grad_norm": 4.25, + "learning_rate": 8.835743695829635e-06, + "loss": 1.53951759, + "memory(GiB)": 111.15, + "step": 20545, + "train_speed(iter/s)": 0.448753 + }, + { + "acc": 0.63773623, + "epoch": 0.5213089802130898, + "grad_norm": 6.3125, + "learning_rate": 8.835070951621971e-06, + "loss": 1.68298283, + "memory(GiB)": 111.15, + "step": 20550, + "train_speed(iter/s)": 0.448811 + }, + { + "acc": 0.6488215, + "epoch": 0.5214358193810249, + "grad_norm": 5.125, + "learning_rate": 8.834398038730765e-06, + "loss": 1.64031334, + "memory(GiB)": 111.15, + "step": 20555, + "train_speed(iter/s)": 0.448869 + }, + { + "acc": 0.6825655, + "epoch": 0.5215626585489599, + "grad_norm": 6.375, + "learning_rate": 8.833724957185612e-06, + "loss": 1.51895561, + "memory(GiB)": 111.15, + "step": 20560, + "train_speed(iter/s)": 0.448927 + }, + { + "acc": 0.64102654, + "epoch": 0.521689497716895, + "grad_norm": 6.1875, + "learning_rate": 8.83305170701612e-06, + "loss": 1.6964674, + "memory(GiB)": 111.15, + "step": 20565, + "train_speed(iter/s)": 0.448985 + }, + { + "acc": 0.65696597, + "epoch": 0.5218163368848301, + "grad_norm": 5.46875, + "learning_rate": 8.832378288251902e-06, + "loss": 1.64727287, + "memory(GiB)": 111.15, + "step": 20570, + "train_speed(iter/s)": 0.449043 + }, + { + "acc": 0.64583349, + "epoch": 0.5219431760527651, + "grad_norm": 5.59375, + "learning_rate": 8.831704700922574e-06, + "loss": 1.65551796, + "memory(GiB)": 111.15, + "step": 20575, + "train_speed(iter/s)": 0.449101 + }, + { + "acc": 0.64951196, + "epoch": 0.5220700152207002, + "grad_norm": 5.53125, + "learning_rate": 8.831030945057767e-06, + "loss": 1.62573948, + "memory(GiB)": 111.15, + "step": 20580, + "train_speed(iter/s)": 0.449158 + }, + { + "acc": 0.66495514, + "epoch": 0.5221968543886352, + "grad_norm": 4.78125, + "learning_rate": 8.830357020687115e-06, + "loss": 1.53906746, + "memory(GiB)": 111.15, + "step": 20585, + "train_speed(iter/s)": 0.449215 + }, + { + "acc": 0.64656038, + "epoch": 0.5223236935565703, + "grad_norm": 4.9375, + "learning_rate": 8.82968292784026e-06, + "loss": 1.64613762, + "memory(GiB)": 111.15, + "step": 20590, + "train_speed(iter/s)": 0.449273 + }, + { + "acc": 0.64068317, + "epoch": 0.5224505327245054, + "grad_norm": 6.34375, + "learning_rate": 8.82900866654685e-06, + "loss": 1.63355103, + "memory(GiB)": 111.15, + "step": 20595, + "train_speed(iter/s)": 0.449332 + }, + { + "acc": 0.6492631, + "epoch": 0.5225773718924404, + "grad_norm": 4.875, + "learning_rate": 8.828334236836546e-06, + "loss": 1.61788692, + "memory(GiB)": 111.15, + "step": 20600, + "train_speed(iter/s)": 0.449389 + }, + { + "epoch": 0.5225773718924404, + "eval_acc": 0.641443686846638, + "eval_loss": 1.6192506551742554, + "eval_runtime": 113.1557, + "eval_samples_per_second": 56.294, + "eval_steps_per_second": 28.147, + "step": 20600 + }, + { + "acc": 0.64619427, + "epoch": 0.5227042110603755, + "grad_norm": 4.96875, + "learning_rate": 8.827659638739007e-06, + "loss": 1.68572559, + "memory(GiB)": 111.15, + "step": 20605, + "train_speed(iter/s)": 0.448295 + }, + { + "acc": 0.6394949, + "epoch": 0.5228310502283106, + "grad_norm": 5.03125, + "learning_rate": 8.82698487228391e-06, + "loss": 1.60918732, + "memory(GiB)": 111.15, + "step": 20610, + "train_speed(iter/s)": 0.448353 + }, + { + "acc": 0.64383879, + "epoch": 0.5229578893962455, + "grad_norm": 5.15625, + "learning_rate": 8.826309937500932e-06, + "loss": 1.62483215, + "memory(GiB)": 111.15, + "step": 20615, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.64457626, + "epoch": 0.5230847285641806, + "grad_norm": 4.96875, + "learning_rate": 8.825634834419758e-06, + "loss": 1.66760273, + "memory(GiB)": 111.15, + "step": 20620, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.65588908, + "epoch": 0.5232115677321156, + "grad_norm": 6.0, + "learning_rate": 8.824959563070085e-06, + "loss": 1.58184528, + "memory(GiB)": 111.15, + "step": 20625, + "train_speed(iter/s)": 0.448525 + }, + { + "acc": 0.64928026, + "epoch": 0.5233384069000507, + "grad_norm": 5.09375, + "learning_rate": 8.824284123481614e-06, + "loss": 1.54839239, + "memory(GiB)": 111.15, + "step": 20630, + "train_speed(iter/s)": 0.448583 + }, + { + "acc": 0.64109669, + "epoch": 0.5234652460679858, + "grad_norm": 5.25, + "learning_rate": 8.823608515684053e-06, + "loss": 1.67186108, + "memory(GiB)": 111.15, + "step": 20635, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.63917036, + "epoch": 0.5235920852359208, + "grad_norm": 4.40625, + "learning_rate": 8.822932739707118e-06, + "loss": 1.70629234, + "memory(GiB)": 111.15, + "step": 20640, + "train_speed(iter/s)": 0.448698 + }, + { + "acc": 0.63644342, + "epoch": 0.5237189244038559, + "grad_norm": 5.46875, + "learning_rate": 8.822256795580532e-06, + "loss": 1.67135487, + "memory(GiB)": 111.15, + "step": 20645, + "train_speed(iter/s)": 0.448755 + }, + { + "acc": 0.65369778, + "epoch": 0.523845763571791, + "grad_norm": 5.78125, + "learning_rate": 8.821580683334027e-06, + "loss": 1.63749466, + "memory(GiB)": 111.15, + "step": 20650, + "train_speed(iter/s)": 0.448813 + }, + { + "acc": 0.63634567, + "epoch": 0.523972602739726, + "grad_norm": 6.09375, + "learning_rate": 8.820904402997343e-06, + "loss": 1.63799858, + "memory(GiB)": 111.15, + "step": 20655, + "train_speed(iter/s)": 0.44887 + }, + { + "acc": 0.64743538, + "epoch": 0.5240994419076611, + "grad_norm": 5.0, + "learning_rate": 8.820227954600222e-06, + "loss": 1.59456768, + "memory(GiB)": 111.15, + "step": 20660, + "train_speed(iter/s)": 0.448927 + }, + { + "acc": 0.63244963, + "epoch": 0.5242262810755961, + "grad_norm": 5.03125, + "learning_rate": 8.819551338172421e-06, + "loss": 1.71562481, + "memory(GiB)": 111.15, + "step": 20665, + "train_speed(iter/s)": 0.448984 + }, + { + "acc": 0.66015415, + "epoch": 0.5243531202435312, + "grad_norm": 4.71875, + "learning_rate": 8.8188745537437e-06, + "loss": 1.63960705, + "memory(GiB)": 111.15, + "step": 20670, + "train_speed(iter/s)": 0.449041 + }, + { + "acc": 0.64911795, + "epoch": 0.5244799594114663, + "grad_norm": 5.25, + "learning_rate": 8.818197601343822e-06, + "loss": 1.61926975, + "memory(GiB)": 111.15, + "step": 20675, + "train_speed(iter/s)": 0.449099 + }, + { + "acc": 0.64613681, + "epoch": 0.5246067985794013, + "grad_norm": 6.625, + "learning_rate": 8.81752048100257e-06, + "loss": 1.58071947, + "memory(GiB)": 111.15, + "step": 20680, + "train_speed(iter/s)": 0.449156 + }, + { + "acc": 0.63239841, + "epoch": 0.5247336377473364, + "grad_norm": 5.21875, + "learning_rate": 8.816843192749724e-06, + "loss": 1.71042824, + "memory(GiB)": 111.15, + "step": 20685, + "train_speed(iter/s)": 0.449214 + }, + { + "acc": 0.64500294, + "epoch": 0.5248604769152715, + "grad_norm": 5.78125, + "learning_rate": 8.816165736615072e-06, + "loss": 1.62676353, + "memory(GiB)": 111.15, + "step": 20690, + "train_speed(iter/s)": 0.449271 + }, + { + "acc": 0.64720793, + "epoch": 0.5249873160832065, + "grad_norm": 5.84375, + "learning_rate": 8.815488112628412e-06, + "loss": 1.64080124, + "memory(GiB)": 111.15, + "step": 20695, + "train_speed(iter/s)": 0.449329 + }, + { + "acc": 0.65860581, + "epoch": 0.5251141552511416, + "grad_norm": 5.40625, + "learning_rate": 8.814810320819551e-06, + "loss": 1.58190289, + "memory(GiB)": 111.15, + "step": 20700, + "train_speed(iter/s)": 0.449387 + }, + { + "epoch": 0.5251141552511416, + "eval_acc": 0.6413818572848383, + "eval_loss": 1.6190168857574463, + "eval_runtime": 113.0732, + "eval_samples_per_second": 56.335, + "eval_steps_per_second": 28.168, + "step": 20700 + }, + { + "acc": 0.65677853, + "epoch": 0.5252409944190766, + "grad_norm": 5.40625, + "learning_rate": 8.8141323612183e-06, + "loss": 1.62655773, + "memory(GiB)": 111.15, + "step": 20705, + "train_speed(iter/s)": 0.4483 + }, + { + "acc": 0.63819642, + "epoch": 0.5253678335870117, + "grad_norm": 6.46875, + "learning_rate": 8.813454233854479e-06, + "loss": 1.65494156, + "memory(GiB)": 111.15, + "step": 20710, + "train_speed(iter/s)": 0.448357 + }, + { + "acc": 0.64968767, + "epoch": 0.5254946727549468, + "grad_norm": 5.1875, + "learning_rate": 8.812775938757914e-06, + "loss": 1.61218185, + "memory(GiB)": 111.15, + "step": 20715, + "train_speed(iter/s)": 0.448415 + }, + { + "acc": 0.66025357, + "epoch": 0.5256215119228818, + "grad_norm": 5.9375, + "learning_rate": 8.812097475958442e-06, + "loss": 1.58391018, + "memory(GiB)": 111.15, + "step": 20720, + "train_speed(iter/s)": 0.448472 + }, + { + "acc": 0.64364376, + "epoch": 0.5257483510908169, + "grad_norm": 5.9375, + "learning_rate": 8.811418845485902e-06, + "loss": 1.67757301, + "memory(GiB)": 111.15, + "step": 20725, + "train_speed(iter/s)": 0.44853 + }, + { + "acc": 0.65662489, + "epoch": 0.525875190258752, + "grad_norm": 5.34375, + "learning_rate": 8.810740047370146e-06, + "loss": 1.66067619, + "memory(GiB)": 111.15, + "step": 20730, + "train_speed(iter/s)": 0.448587 + }, + { + "acc": 0.66195917, + "epoch": 0.526002029426687, + "grad_norm": 5.5, + "learning_rate": 8.810061081641026e-06, + "loss": 1.6368803, + "memory(GiB)": 111.15, + "step": 20735, + "train_speed(iter/s)": 0.448644 + }, + { + "acc": 0.65861826, + "epoch": 0.526128868594622, + "grad_norm": 4.375, + "learning_rate": 8.809381948328412e-06, + "loss": 1.60868912, + "memory(GiB)": 111.15, + "step": 20740, + "train_speed(iter/s)": 0.448701 + }, + { + "acc": 0.65172882, + "epoch": 0.526255707762557, + "grad_norm": 4.90625, + "learning_rate": 8.80870264746217e-06, + "loss": 1.65635357, + "memory(GiB)": 111.15, + "step": 20745, + "train_speed(iter/s)": 0.448759 + }, + { + "acc": 0.64303322, + "epoch": 0.5263825469304921, + "grad_norm": 4.71875, + "learning_rate": 8.808023179072183e-06, + "loss": 1.62728596, + "memory(GiB)": 111.15, + "step": 20750, + "train_speed(iter/s)": 0.448817 + }, + { + "acc": 0.63846941, + "epoch": 0.5265093860984272, + "grad_norm": 5.15625, + "learning_rate": 8.807343543188333e-06, + "loss": 1.71516266, + "memory(GiB)": 111.15, + "step": 20755, + "train_speed(iter/s)": 0.448874 + }, + { + "acc": 0.65823116, + "epoch": 0.5266362252663622, + "grad_norm": 5.84375, + "learning_rate": 8.806663739840515e-06, + "loss": 1.63009377, + "memory(GiB)": 111.15, + "step": 20760, + "train_speed(iter/s)": 0.448932 + }, + { + "acc": 0.64279265, + "epoch": 0.5267630644342973, + "grad_norm": 5.5625, + "learning_rate": 8.805983769058633e-06, + "loss": 1.67328529, + "memory(GiB)": 111.15, + "step": 20765, + "train_speed(iter/s)": 0.448989 + }, + { + "acc": 0.65562534, + "epoch": 0.5268899036022324, + "grad_norm": 5.78125, + "learning_rate": 8.80530363087259e-06, + "loss": 1.5953598, + "memory(GiB)": 111.15, + "step": 20770, + "train_speed(iter/s)": 0.449046 + }, + { + "acc": 0.65021629, + "epoch": 0.5270167427701674, + "grad_norm": 5.03125, + "learning_rate": 8.804623325312305e-06, + "loss": 1.67343712, + "memory(GiB)": 111.15, + "step": 20775, + "train_speed(iter/s)": 0.449104 + }, + { + "acc": 0.64371605, + "epoch": 0.5271435819381025, + "grad_norm": 5.03125, + "learning_rate": 8.8039428524077e-06, + "loss": 1.63680305, + "memory(GiB)": 111.15, + "step": 20780, + "train_speed(iter/s)": 0.449161 + }, + { + "acc": 0.65993819, + "epoch": 0.5272704211060375, + "grad_norm": 6.0, + "learning_rate": 8.803262212188703e-06, + "loss": 1.67922363, + "memory(GiB)": 111.15, + "step": 20785, + "train_speed(iter/s)": 0.449218 + }, + { + "acc": 0.65364218, + "epoch": 0.5273972602739726, + "grad_norm": 6.25, + "learning_rate": 8.802581404685255e-06, + "loss": 1.65515366, + "memory(GiB)": 111.15, + "step": 20790, + "train_speed(iter/s)": 0.449277 + }, + { + "acc": 0.6288456, + "epoch": 0.5275240994419077, + "grad_norm": 5.46875, + "learning_rate": 8.8019004299273e-06, + "loss": 1.69230194, + "memory(GiB)": 111.15, + "step": 20795, + "train_speed(iter/s)": 0.449334 + }, + { + "acc": 0.64219809, + "epoch": 0.5276509386098427, + "grad_norm": 6.5, + "learning_rate": 8.801219287944788e-06, + "loss": 1.65222931, + "memory(GiB)": 111.15, + "step": 20800, + "train_speed(iter/s)": 0.449391 + }, + { + "epoch": 0.5276509386098427, + "eval_acc": 0.641331307440394, + "eval_loss": 1.619089126586914, + "eval_runtime": 112.423, + "eval_samples_per_second": 56.661, + "eval_steps_per_second": 28.331, + "step": 20800 + }, + { + "acc": 0.65076442, + "epoch": 0.5277777777777778, + "grad_norm": 5.8125, + "learning_rate": 8.800537978767683e-06, + "loss": 1.6274025, + "memory(GiB)": 111.15, + "step": 20805, + "train_speed(iter/s)": 0.448316 + }, + { + "acc": 0.66007919, + "epoch": 0.5279046169457129, + "grad_norm": 6.34375, + "learning_rate": 8.79985650242595e-06, + "loss": 1.62142105, + "memory(GiB)": 111.15, + "step": 20810, + "train_speed(iter/s)": 0.448373 + }, + { + "acc": 0.63672967, + "epoch": 0.5280314561136479, + "grad_norm": 4.84375, + "learning_rate": 8.79917485894956e-06, + "loss": 1.67950211, + "memory(GiB)": 111.15, + "step": 20815, + "train_speed(iter/s)": 0.44843 + }, + { + "acc": 0.64817281, + "epoch": 0.528158295281583, + "grad_norm": 4.8125, + "learning_rate": 8.798493048368498e-06, + "loss": 1.67589493, + "memory(GiB)": 111.15, + "step": 20820, + "train_speed(iter/s)": 0.448486 + }, + { + "acc": 0.6432302, + "epoch": 0.528285134449518, + "grad_norm": 4.4375, + "learning_rate": 8.797811070712754e-06, + "loss": 1.59743347, + "memory(GiB)": 111.15, + "step": 20825, + "train_speed(iter/s)": 0.448543 + }, + { + "acc": 0.64535456, + "epoch": 0.5284119736174531, + "grad_norm": 5.15625, + "learning_rate": 8.797128926012323e-06, + "loss": 1.69040871, + "memory(GiB)": 111.15, + "step": 20830, + "train_speed(iter/s)": 0.448599 + }, + { + "acc": 0.64974117, + "epoch": 0.5285388127853882, + "grad_norm": 6.5, + "learning_rate": 8.796446614297208e-06, + "loss": 1.59495621, + "memory(GiB)": 111.15, + "step": 20835, + "train_speed(iter/s)": 0.448656 + }, + { + "acc": 0.64675612, + "epoch": 0.5286656519533232, + "grad_norm": 5.65625, + "learning_rate": 8.795764135597421e-06, + "loss": 1.64179955, + "memory(GiB)": 111.15, + "step": 20840, + "train_speed(iter/s)": 0.448713 + }, + { + "acc": 0.64897699, + "epoch": 0.5287924911212583, + "grad_norm": 4.65625, + "learning_rate": 8.79508148994298e-06, + "loss": 1.64623413, + "memory(GiB)": 111.15, + "step": 20845, + "train_speed(iter/s)": 0.448769 + }, + { + "acc": 0.63817172, + "epoch": 0.5289193302891934, + "grad_norm": 5.34375, + "learning_rate": 8.794398677363913e-06, + "loss": 1.65751839, + "memory(GiB)": 111.15, + "step": 20850, + "train_speed(iter/s)": 0.448826 + }, + { + "acc": 0.66346025, + "epoch": 0.5290461694571283, + "grad_norm": 5.125, + "learning_rate": 8.79371569789025e-06, + "loss": 1.58686733, + "memory(GiB)": 111.15, + "step": 20855, + "train_speed(iter/s)": 0.448883 + }, + { + "acc": 0.65006056, + "epoch": 0.5291730086250634, + "grad_norm": 7.09375, + "learning_rate": 8.793032551552034e-06, + "loss": 1.64545231, + "memory(GiB)": 111.15, + "step": 20860, + "train_speed(iter/s)": 0.448939 + }, + { + "acc": 0.64323597, + "epoch": 0.5292998477929984, + "grad_norm": 5.03125, + "learning_rate": 8.792349238379311e-06, + "loss": 1.66735725, + "memory(GiB)": 111.15, + "step": 20865, + "train_speed(iter/s)": 0.448996 + }, + { + "acc": 0.62644434, + "epoch": 0.5294266869609335, + "grad_norm": 5.625, + "learning_rate": 8.791665758402137e-06, + "loss": 1.71982765, + "memory(GiB)": 111.15, + "step": 20870, + "train_speed(iter/s)": 0.449053 + }, + { + "acc": 0.66253901, + "epoch": 0.5295535261288686, + "grad_norm": 5.25, + "learning_rate": 8.790982111650574e-06, + "loss": 1.61684151, + "memory(GiB)": 111.15, + "step": 20875, + "train_speed(iter/s)": 0.44911 + }, + { + "acc": 0.6539969, + "epoch": 0.5296803652968036, + "grad_norm": 5.03125, + "learning_rate": 8.790298298154694e-06, + "loss": 1.58747787, + "memory(GiB)": 111.15, + "step": 20880, + "train_speed(iter/s)": 0.449167 + }, + { + "acc": 0.64132042, + "epoch": 0.5298072044647387, + "grad_norm": 4.59375, + "learning_rate": 8.78961431794457e-06, + "loss": 1.65248795, + "memory(GiB)": 111.15, + "step": 20885, + "train_speed(iter/s)": 0.449223 + }, + { + "acc": 0.6528657, + "epoch": 0.5299340436326738, + "grad_norm": 4.96875, + "learning_rate": 8.78893017105029e-06, + "loss": 1.64326172, + "memory(GiB)": 111.15, + "step": 20890, + "train_speed(iter/s)": 0.44928 + }, + { + "acc": 0.64660087, + "epoch": 0.5300608828006088, + "grad_norm": 5.375, + "learning_rate": 8.788245857501944e-06, + "loss": 1.6061327, + "memory(GiB)": 111.15, + "step": 20895, + "train_speed(iter/s)": 0.449336 + }, + { + "acc": 0.64356489, + "epoch": 0.5301877219685439, + "grad_norm": 5.59375, + "learning_rate": 8.787561377329633e-06, + "loss": 1.68568916, + "memory(GiB)": 111.15, + "step": 20900, + "train_speed(iter/s)": 0.449392 + }, + { + "epoch": 0.5301877219685439, + "eval_acc": 0.6413041525652793, + "eval_loss": 1.6189017295837402, + "eval_runtime": 113.5019, + "eval_samples_per_second": 56.122, + "eval_steps_per_second": 28.061, + "step": 20900 + }, + { + "acc": 0.65849266, + "epoch": 0.5303145611364789, + "grad_norm": 5.34375, + "learning_rate": 8.786876730563462e-06, + "loss": 1.63563976, + "memory(GiB)": 111.15, + "step": 20905, + "train_speed(iter/s)": 0.448311 + }, + { + "acc": 0.64602413, + "epoch": 0.530441400304414, + "grad_norm": 6.3125, + "learning_rate": 8.786191917233545e-06, + "loss": 1.63195114, + "memory(GiB)": 111.15, + "step": 20910, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.65451756, + "epoch": 0.5305682394723491, + "grad_norm": 5.25, + "learning_rate": 8.785506937370003e-06, + "loss": 1.5705514, + "memory(GiB)": 111.15, + "step": 20915, + "train_speed(iter/s)": 0.448423 + }, + { + "acc": 0.65763016, + "epoch": 0.5306950786402841, + "grad_norm": 5.53125, + "learning_rate": 8.784821791002965e-06, + "loss": 1.62078648, + "memory(GiB)": 111.15, + "step": 20920, + "train_speed(iter/s)": 0.448479 + }, + { + "acc": 0.65717974, + "epoch": 0.5308219178082192, + "grad_norm": 5.71875, + "learning_rate": 8.784136478162567e-06, + "loss": 1.66754303, + "memory(GiB)": 111.15, + "step": 20925, + "train_speed(iter/s)": 0.448536 + }, + { + "acc": 0.65587173, + "epoch": 0.5309487569761543, + "grad_norm": 6.0625, + "learning_rate": 8.783450998878951e-06, + "loss": 1.59382124, + "memory(GiB)": 111.15, + "step": 20930, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.64687381, + "epoch": 0.5310755961440893, + "grad_norm": 5.0, + "learning_rate": 8.78276535318227e-06, + "loss": 1.64621277, + "memory(GiB)": 111.15, + "step": 20935, + "train_speed(iter/s)": 0.448648 + }, + { + "acc": 0.62829041, + "epoch": 0.5312024353120244, + "grad_norm": 6.25, + "learning_rate": 8.782079541102678e-06, + "loss": 1.72944145, + "memory(GiB)": 111.15, + "step": 20940, + "train_speed(iter/s)": 0.448705 + }, + { + "acc": 0.64178171, + "epoch": 0.5313292744799594, + "grad_norm": 4.71875, + "learning_rate": 8.781393562670342e-06, + "loss": 1.64785919, + "memory(GiB)": 111.15, + "step": 20945, + "train_speed(iter/s)": 0.448761 + }, + { + "acc": 0.6356853, + "epoch": 0.5314561136478945, + "grad_norm": 4.6875, + "learning_rate": 8.780707417915436e-06, + "loss": 1.6767683, + "memory(GiB)": 111.15, + "step": 20950, + "train_speed(iter/s)": 0.448818 + }, + { + "acc": 0.65264988, + "epoch": 0.5315829528158296, + "grad_norm": 4.9375, + "learning_rate": 8.780021106868138e-06, + "loss": 1.52395315, + "memory(GiB)": 111.15, + "step": 20955, + "train_speed(iter/s)": 0.448875 + }, + { + "acc": 0.64251995, + "epoch": 0.5317097919837646, + "grad_norm": 5.9375, + "learning_rate": 8.779334629558633e-06, + "loss": 1.66696548, + "memory(GiB)": 111.15, + "step": 20960, + "train_speed(iter/s)": 0.448931 + }, + { + "acc": 0.64771004, + "epoch": 0.5318366311516997, + "grad_norm": 6.78125, + "learning_rate": 8.77864798601712e-06, + "loss": 1.67555008, + "memory(GiB)": 111.15, + "step": 20965, + "train_speed(iter/s)": 0.448988 + }, + { + "acc": 0.65576429, + "epoch": 0.5319634703196348, + "grad_norm": 5.03125, + "learning_rate": 8.777961176273795e-06, + "loss": 1.67485046, + "memory(GiB)": 111.15, + "step": 20970, + "train_speed(iter/s)": 0.449044 + }, + { + "acc": 0.6507534, + "epoch": 0.5320903094875697, + "grad_norm": 7.09375, + "learning_rate": 8.777274200358873e-06, + "loss": 1.68157063, + "memory(GiB)": 111.15, + "step": 20975, + "train_speed(iter/s)": 0.449101 + }, + { + "acc": 0.66317434, + "epoch": 0.5322171486555048, + "grad_norm": 5.25, + "learning_rate": 8.776587058302566e-06, + "loss": 1.63404465, + "memory(GiB)": 111.15, + "step": 20980, + "train_speed(iter/s)": 0.449158 + }, + { + "acc": 0.65576854, + "epoch": 0.5323439878234398, + "grad_norm": 4.78125, + "learning_rate": 8.7758997501351e-06, + "loss": 1.58851452, + "memory(GiB)": 111.15, + "step": 20985, + "train_speed(iter/s)": 0.449215 + }, + { + "acc": 0.65709839, + "epoch": 0.5324708269913749, + "grad_norm": 5.78125, + "learning_rate": 8.775212275886705e-06, + "loss": 1.61322861, + "memory(GiB)": 111.15, + "step": 20990, + "train_speed(iter/s)": 0.449272 + }, + { + "acc": 0.63026285, + "epoch": 0.53259766615931, + "grad_norm": 6.28125, + "learning_rate": 8.774524635587617e-06, + "loss": 1.68695755, + "memory(GiB)": 111.15, + "step": 20995, + "train_speed(iter/s)": 0.449328 + }, + { + "acc": 0.64821205, + "epoch": 0.532724505327245, + "grad_norm": 5.21875, + "learning_rate": 8.773836829268084e-06, + "loss": 1.64276276, + "memory(GiB)": 111.15, + "step": 21000, + "train_speed(iter/s)": 0.449385 + }, + { + "epoch": 0.532724505327245, + "eval_acc": 0.6414378381043057, + "eval_loss": 1.6187134981155396, + "eval_runtime": 112.3554, + "eval_samples_per_second": 56.695, + "eval_steps_per_second": 28.348, + "step": 21000 + }, + { + "acc": 0.64413548, + "epoch": 0.5328513444951801, + "grad_norm": 5.21875, + "learning_rate": 8.77314885695836e-06, + "loss": 1.61260548, + "memory(GiB)": 111.15, + "step": 21005, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.6422503, + "epoch": 0.5329781836631152, + "grad_norm": 5.34375, + "learning_rate": 8.772460718688702e-06, + "loss": 1.60806236, + "memory(GiB)": 111.15, + "step": 21010, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.66154575, + "epoch": 0.5331050228310502, + "grad_norm": 5.78125, + "learning_rate": 8.771772414489379e-06, + "loss": 1.59863758, + "memory(GiB)": 111.15, + "step": 21015, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.64426212, + "epoch": 0.5332318619989853, + "grad_norm": 5.34375, + "learning_rate": 8.771083944390665e-06, + "loss": 1.67122746, + "memory(GiB)": 111.15, + "step": 21020, + "train_speed(iter/s)": 0.448489 + }, + { + "acc": 0.64007883, + "epoch": 0.5333587011669203, + "grad_norm": 4.78125, + "learning_rate": 8.770395308422842e-06, + "loss": 1.6791275, + "memory(GiB)": 111.15, + "step": 21025, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.65491676, + "epoch": 0.5334855403348554, + "grad_norm": 5.71875, + "learning_rate": 8.769706506616201e-06, + "loss": 1.6276022, + "memory(GiB)": 111.15, + "step": 21030, + "train_speed(iter/s)": 0.448601 + }, + { + "acc": 0.65747156, + "epoch": 0.5336123795027905, + "grad_norm": 4.21875, + "learning_rate": 8.769017539001037e-06, + "loss": 1.60681477, + "memory(GiB)": 111.15, + "step": 21035, + "train_speed(iter/s)": 0.448658 + }, + { + "acc": 0.65258904, + "epoch": 0.5337392186707255, + "grad_norm": 5.71875, + "learning_rate": 8.768328405607655e-06, + "loss": 1.6443924, + "memory(GiB)": 111.15, + "step": 21040, + "train_speed(iter/s)": 0.448714 + }, + { + "acc": 0.6404851, + "epoch": 0.5338660578386606, + "grad_norm": 5.0625, + "learning_rate": 8.767639106466364e-06, + "loss": 1.70296516, + "memory(GiB)": 111.15, + "step": 21045, + "train_speed(iter/s)": 0.448771 + }, + { + "acc": 0.64972425, + "epoch": 0.5339928970065957, + "grad_norm": 4.78125, + "learning_rate": 8.766949641607484e-06, + "loss": 1.66417522, + "memory(GiB)": 111.15, + "step": 21050, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.65165977, + "epoch": 0.5341197361745307, + "grad_norm": 5.40625, + "learning_rate": 8.76626001106134e-06, + "loss": 1.64909611, + "memory(GiB)": 111.15, + "step": 21055, + "train_speed(iter/s)": 0.448883 + }, + { + "acc": 0.65665426, + "epoch": 0.5342465753424658, + "grad_norm": 5.3125, + "learning_rate": 8.765570214858268e-06, + "loss": 1.55610275, + "memory(GiB)": 111.15, + "step": 21060, + "train_speed(iter/s)": 0.44894 + }, + { + "acc": 0.6453784, + "epoch": 0.5343734145104008, + "grad_norm": 5.53125, + "learning_rate": 8.764880253028604e-06, + "loss": 1.60476646, + "memory(GiB)": 111.15, + "step": 21065, + "train_speed(iter/s)": 0.448996 + }, + { + "acc": 0.64591665, + "epoch": 0.5345002536783359, + "grad_norm": 4.3125, + "learning_rate": 8.764190125602698e-06, + "loss": 1.61555367, + "memory(GiB)": 111.15, + "step": 21070, + "train_speed(iter/s)": 0.449052 + }, + { + "acc": 0.65049481, + "epoch": 0.534627092846271, + "grad_norm": 5.6875, + "learning_rate": 8.763499832610904e-06, + "loss": 1.67644615, + "memory(GiB)": 111.15, + "step": 21075, + "train_speed(iter/s)": 0.449109 + }, + { + "acc": 0.66562624, + "epoch": 0.534753932014206, + "grad_norm": 4.71875, + "learning_rate": 8.762809374083585e-06, + "loss": 1.5890379, + "memory(GiB)": 111.15, + "step": 21080, + "train_speed(iter/s)": 0.449165 + }, + { + "acc": 0.6356905, + "epoch": 0.5348807711821411, + "grad_norm": 5.84375, + "learning_rate": 8.76211875005111e-06, + "loss": 1.68113041, + "memory(GiB)": 111.15, + "step": 21085, + "train_speed(iter/s)": 0.449221 + }, + { + "acc": 0.67540054, + "epoch": 0.5350076103500762, + "grad_norm": 6.9375, + "learning_rate": 8.761427960543854e-06, + "loss": 1.54257078, + "memory(GiB)": 111.15, + "step": 21090, + "train_speed(iter/s)": 0.449278 + }, + { + "acc": 0.65144091, + "epoch": 0.5351344495180111, + "grad_norm": 5.53125, + "learning_rate": 8.760737005592205e-06, + "loss": 1.63055687, + "memory(GiB)": 111.15, + "step": 21095, + "train_speed(iter/s)": 0.449334 + }, + { + "acc": 0.67670503, + "epoch": 0.5352612886859462, + "grad_norm": 4.5625, + "learning_rate": 8.760045885226551e-06, + "loss": 1.60595703, + "memory(GiB)": 111.15, + "step": 21100, + "train_speed(iter/s)": 0.449391 + }, + { + "epoch": 0.5352612886859462, + "eval_acc": 0.641429064990807, + "eval_loss": 1.6188441514968872, + "eval_runtime": 112.9751, + "eval_samples_per_second": 56.384, + "eval_steps_per_second": 28.192, + "step": 21100 + }, + { + "acc": 0.6564713, + "epoch": 0.5353881278538812, + "grad_norm": 5.78125, + "learning_rate": 8.759354599477293e-06, + "loss": 1.65448074, + "memory(GiB)": 111.15, + "step": 21105, + "train_speed(iter/s)": 0.448325 + }, + { + "acc": 0.6599041, + "epoch": 0.5355149670218163, + "grad_norm": 5.09375, + "learning_rate": 8.758663148374833e-06, + "loss": 1.55637674, + "memory(GiB)": 111.15, + "step": 21110, + "train_speed(iter/s)": 0.448381 + }, + { + "acc": 0.6574172, + "epoch": 0.5356418061897514, + "grad_norm": 5.125, + "learning_rate": 8.757971531949587e-06, + "loss": 1.64465446, + "memory(GiB)": 111.15, + "step": 21115, + "train_speed(iter/s)": 0.448438 + }, + { + "acc": 0.65212374, + "epoch": 0.5357686453576864, + "grad_norm": 4.59375, + "learning_rate": 8.757279750231977e-06, + "loss": 1.61003189, + "memory(GiB)": 111.15, + "step": 21120, + "train_speed(iter/s)": 0.448493 + }, + { + "acc": 0.65529203, + "epoch": 0.5358954845256215, + "grad_norm": 5.0, + "learning_rate": 8.756587803252426e-06, + "loss": 1.56732788, + "memory(GiB)": 111.15, + "step": 21125, + "train_speed(iter/s)": 0.44855 + }, + { + "acc": 0.65529928, + "epoch": 0.5360223236935566, + "grad_norm": 5.375, + "learning_rate": 8.755895691041373e-06, + "loss": 1.63084736, + "memory(GiB)": 111.15, + "step": 21130, + "train_speed(iter/s)": 0.448606 + }, + { + "acc": 0.65080814, + "epoch": 0.5361491628614916, + "grad_norm": 4.71875, + "learning_rate": 8.755203413629257e-06, + "loss": 1.63112736, + "memory(GiB)": 111.15, + "step": 21135, + "train_speed(iter/s)": 0.448662 + }, + { + "acc": 0.65512714, + "epoch": 0.5362760020294267, + "grad_norm": 5.03125, + "learning_rate": 8.75451097104653e-06, + "loss": 1.54782982, + "memory(GiB)": 111.15, + "step": 21140, + "train_speed(iter/s)": 0.448718 + }, + { + "acc": 0.65637674, + "epoch": 0.5364028411973617, + "grad_norm": 5.15625, + "learning_rate": 8.75381836332365e-06, + "loss": 1.68503704, + "memory(GiB)": 111.15, + "step": 21145, + "train_speed(iter/s)": 0.448775 + }, + { + "acc": 0.62897263, + "epoch": 0.5365296803652968, + "grad_norm": 5.0625, + "learning_rate": 8.753125590491077e-06, + "loss": 1.65831604, + "memory(GiB)": 111.15, + "step": 21150, + "train_speed(iter/s)": 0.448832 + }, + { + "acc": 0.65479326, + "epoch": 0.5366565195332319, + "grad_norm": 4.6875, + "learning_rate": 8.752432652579284e-06, + "loss": 1.58161373, + "memory(GiB)": 111.15, + "step": 21155, + "train_speed(iter/s)": 0.448888 + }, + { + "acc": 0.63825378, + "epoch": 0.5367833587011669, + "grad_norm": 6.125, + "learning_rate": 8.751739549618749e-06, + "loss": 1.67486458, + "memory(GiB)": 111.15, + "step": 21160, + "train_speed(iter/s)": 0.448945 + }, + { + "acc": 0.65549507, + "epoch": 0.536910197869102, + "grad_norm": 5.9375, + "learning_rate": 8.751046281639958e-06, + "loss": 1.59011288, + "memory(GiB)": 111.15, + "step": 21165, + "train_speed(iter/s)": 0.449002 + }, + { + "acc": 0.65477495, + "epoch": 0.5370370370370371, + "grad_norm": 4.75, + "learning_rate": 8.750352848673405e-06, + "loss": 1.64748192, + "memory(GiB)": 111.15, + "step": 21170, + "train_speed(iter/s)": 0.449058 + }, + { + "acc": 0.64258213, + "epoch": 0.5371638762049721, + "grad_norm": 5.46875, + "learning_rate": 8.749659250749589e-06, + "loss": 1.70933609, + "memory(GiB)": 111.15, + "step": 21175, + "train_speed(iter/s)": 0.449115 + }, + { + "acc": 0.66172342, + "epoch": 0.5372907153729072, + "grad_norm": 6.59375, + "learning_rate": 8.748965487899019e-06, + "loss": 1.60262756, + "memory(GiB)": 111.15, + "step": 21180, + "train_speed(iter/s)": 0.449172 + }, + { + "acc": 0.65072265, + "epoch": 0.5374175545408422, + "grad_norm": 6.0, + "learning_rate": 8.748271560152208e-06, + "loss": 1.58556461, + "memory(GiB)": 111.15, + "step": 21185, + "train_speed(iter/s)": 0.449229 + }, + { + "acc": 0.64545364, + "epoch": 0.5375443937087773, + "grad_norm": 5.25, + "learning_rate": 8.74757746753968e-06, + "loss": 1.65436459, + "memory(GiB)": 111.15, + "step": 21190, + "train_speed(iter/s)": 0.449286 + }, + { + "acc": 0.65416293, + "epoch": 0.5376712328767124, + "grad_norm": 5.59375, + "learning_rate": 8.746883210091963e-06, + "loss": 1.55474968, + "memory(GiB)": 111.15, + "step": 21195, + "train_speed(iter/s)": 0.449342 + }, + { + "acc": 0.67353969, + "epoch": 0.5377980720446474, + "grad_norm": 4.8125, + "learning_rate": 8.746188787839593e-06, + "loss": 1.51669235, + "memory(GiB)": 111.15, + "step": 21200, + "train_speed(iter/s)": 0.449399 + }, + { + "epoch": 0.5377980720446474, + "eval_acc": 0.6414624863755636, + "eval_loss": 1.6185083389282227, + "eval_runtime": 113.4933, + "eval_samples_per_second": 56.127, + "eval_steps_per_second": 28.063, + "step": 21200 + }, + { + "acc": 0.65435314, + "epoch": 0.5379249112125825, + "grad_norm": 4.96875, + "learning_rate": 8.745494200813116e-06, + "loss": 1.63672543, + "memory(GiB)": 111.15, + "step": 21205, + "train_speed(iter/s)": 0.448333 + }, + { + "acc": 0.64973392, + "epoch": 0.5380517503805176, + "grad_norm": 7.1875, + "learning_rate": 8.74479944904308e-06, + "loss": 1.6382412, + "memory(GiB)": 111.15, + "step": 21210, + "train_speed(iter/s)": 0.448388 + }, + { + "acc": 0.64351807, + "epoch": 0.5381785895484525, + "grad_norm": 5.15625, + "learning_rate": 8.744104532560047e-06, + "loss": 1.63959942, + "memory(GiB)": 111.15, + "step": 21215, + "train_speed(iter/s)": 0.448444 + }, + { + "acc": 0.63166356, + "epoch": 0.5383054287163876, + "grad_norm": 5.125, + "learning_rate": 8.74340945139458e-06, + "loss": 1.72402325, + "memory(GiB)": 111.15, + "step": 21220, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.64511433, + "epoch": 0.5384322678843226, + "grad_norm": 7.625, + "learning_rate": 8.742714205577251e-06, + "loss": 1.62924709, + "memory(GiB)": 111.15, + "step": 21225, + "train_speed(iter/s)": 0.448557 + }, + { + "acc": 0.66363907, + "epoch": 0.5385591070522577, + "grad_norm": 4.59375, + "learning_rate": 8.742018795138642e-06, + "loss": 1.52479753, + "memory(GiB)": 111.15, + "step": 21230, + "train_speed(iter/s)": 0.448613 + }, + { + "acc": 0.64408321, + "epoch": 0.5386859462201928, + "grad_norm": 5.21875, + "learning_rate": 8.74132322010934e-06, + "loss": 1.63178291, + "memory(GiB)": 111.15, + "step": 21235, + "train_speed(iter/s)": 0.448669 + }, + { + "acc": 0.64753532, + "epoch": 0.5388127853881278, + "grad_norm": 5.03125, + "learning_rate": 8.740627480519937e-06, + "loss": 1.61229305, + "memory(GiB)": 111.15, + "step": 21240, + "train_speed(iter/s)": 0.448725 + }, + { + "acc": 0.63910685, + "epoch": 0.5389396245560629, + "grad_norm": 5.75, + "learning_rate": 8.739931576401037e-06, + "loss": 1.64049377, + "memory(GiB)": 111.15, + "step": 21245, + "train_speed(iter/s)": 0.448781 + }, + { + "acc": 0.64502535, + "epoch": 0.539066463723998, + "grad_norm": 5.03125, + "learning_rate": 8.73923550778325e-06, + "loss": 1.56924877, + "memory(GiB)": 111.15, + "step": 21250, + "train_speed(iter/s)": 0.448838 + }, + { + "acc": 0.659025, + "epoch": 0.539193302891933, + "grad_norm": 5.5625, + "learning_rate": 8.73853927469719e-06, + "loss": 1.62700233, + "memory(GiB)": 111.15, + "step": 21255, + "train_speed(iter/s)": 0.448894 + }, + { + "acc": 0.64575987, + "epoch": 0.5393201420598681, + "grad_norm": 6.34375, + "learning_rate": 8.73784287717348e-06, + "loss": 1.63756599, + "memory(GiB)": 111.15, + "step": 21260, + "train_speed(iter/s)": 0.44895 + }, + { + "acc": 0.61788235, + "epoch": 0.5394469812278031, + "grad_norm": 5.15625, + "learning_rate": 8.737146315242755e-06, + "loss": 1.73394165, + "memory(GiB)": 111.15, + "step": 21265, + "train_speed(iter/s)": 0.449006 + }, + { + "acc": 0.64003401, + "epoch": 0.5395738203957382, + "grad_norm": 4.96875, + "learning_rate": 8.73644958893565e-06, + "loss": 1.66983051, + "memory(GiB)": 111.15, + "step": 21270, + "train_speed(iter/s)": 0.449063 + }, + { + "acc": 0.65364151, + "epoch": 0.5397006595636733, + "grad_norm": 4.59375, + "learning_rate": 8.735752698282807e-06, + "loss": 1.59463224, + "memory(GiB)": 111.15, + "step": 21275, + "train_speed(iter/s)": 0.449119 + }, + { + "acc": 0.66295061, + "epoch": 0.5398274987316083, + "grad_norm": 4.6875, + "learning_rate": 8.735055643314883e-06, + "loss": 1.6527977, + "memory(GiB)": 111.15, + "step": 21280, + "train_speed(iter/s)": 0.449176 + }, + { + "acc": 0.63985553, + "epoch": 0.5399543378995434, + "grad_norm": 5.0, + "learning_rate": 8.734358424062536e-06, + "loss": 1.65722961, + "memory(GiB)": 111.15, + "step": 21285, + "train_speed(iter/s)": 0.449232 + }, + { + "acc": 0.65584946, + "epoch": 0.5400811770674785, + "grad_norm": 4.6875, + "learning_rate": 8.733661040556433e-06, + "loss": 1.61240749, + "memory(GiB)": 111.15, + "step": 21290, + "train_speed(iter/s)": 0.449288 + }, + { + "acc": 0.63166947, + "epoch": 0.5402080162354135, + "grad_norm": 4.96875, + "learning_rate": 8.732963492827248e-06, + "loss": 1.65951557, + "memory(GiB)": 111.15, + "step": 21295, + "train_speed(iter/s)": 0.449344 + }, + { + "acc": 0.6415226, + "epoch": 0.5403348554033486, + "grad_norm": 5.09375, + "learning_rate": 8.732265780905661e-06, + "loss": 1.66189098, + "memory(GiB)": 111.15, + "step": 21300, + "train_speed(iter/s)": 0.4494 + }, + { + "epoch": 0.5403348554033486, + "eval_acc": 0.6415080230122945, + "eval_loss": 1.6184227466583252, + "eval_runtime": 112.8171, + "eval_samples_per_second": 56.463, + "eval_steps_per_second": 28.232, + "step": 21300 + }, + { + "acc": 0.63509102, + "epoch": 0.5404616945712836, + "grad_norm": 5.0625, + "learning_rate": 8.731567904822362e-06, + "loss": 1.69000988, + "memory(GiB)": 111.15, + "step": 21305, + "train_speed(iter/s)": 0.448345 + }, + { + "acc": 0.65575619, + "epoch": 0.5405885337392187, + "grad_norm": 4.84375, + "learning_rate": 8.730869864608047e-06, + "loss": 1.53732662, + "memory(GiB)": 111.15, + "step": 21310, + "train_speed(iter/s)": 0.4484 + }, + { + "acc": 0.63919606, + "epoch": 0.5407153729071538, + "grad_norm": 5.875, + "learning_rate": 8.730171660293418e-06, + "loss": 1.71072845, + "memory(GiB)": 111.15, + "step": 21315, + "train_speed(iter/s)": 0.448456 + }, + { + "acc": 0.64673743, + "epoch": 0.5408422120750888, + "grad_norm": 4.375, + "learning_rate": 8.729473291909185e-06, + "loss": 1.61375656, + "memory(GiB)": 111.15, + "step": 21320, + "train_speed(iter/s)": 0.448511 + }, + { + "acc": 0.63442211, + "epoch": 0.5409690512430239, + "grad_norm": 5.59375, + "learning_rate": 8.728774759486065e-06, + "loss": 1.65525875, + "memory(GiB)": 111.15, + "step": 21325, + "train_speed(iter/s)": 0.448567 + }, + { + "acc": 0.64127774, + "epoch": 0.541095890410959, + "grad_norm": 4.875, + "learning_rate": 8.728076063054786e-06, + "loss": 1.61144962, + "memory(GiB)": 111.15, + "step": 21330, + "train_speed(iter/s)": 0.448622 + }, + { + "acc": 0.63859606, + "epoch": 0.541222729578894, + "grad_norm": 5.4375, + "learning_rate": 8.727377202646074e-06, + "loss": 1.72243271, + "memory(GiB)": 111.15, + "step": 21335, + "train_speed(iter/s)": 0.448678 + }, + { + "acc": 0.65193329, + "epoch": 0.541349568746829, + "grad_norm": 7.1875, + "learning_rate": 8.726678178290673e-06, + "loss": 1.70639458, + "memory(GiB)": 111.15, + "step": 21340, + "train_speed(iter/s)": 0.448734 + }, + { + "acc": 0.6582664, + "epoch": 0.541476407914764, + "grad_norm": 6.0, + "learning_rate": 8.725978990019326e-06, + "loss": 1.6251358, + "memory(GiB)": 111.15, + "step": 21345, + "train_speed(iter/s)": 0.44879 + }, + { + "acc": 0.66223326, + "epoch": 0.5416032470826991, + "grad_norm": 4.75, + "learning_rate": 8.72527963786279e-06, + "loss": 1.60527573, + "memory(GiB)": 111.15, + "step": 21350, + "train_speed(iter/s)": 0.448847 + }, + { + "acc": 0.64534087, + "epoch": 0.5417300862506342, + "grad_norm": 4.75, + "learning_rate": 8.72458012185182e-06, + "loss": 1.70588951, + "memory(GiB)": 111.15, + "step": 21355, + "train_speed(iter/s)": 0.448902 + }, + { + "acc": 0.65666637, + "epoch": 0.5418569254185692, + "grad_norm": 5.125, + "learning_rate": 8.72388044201719e-06, + "loss": 1.62027893, + "memory(GiB)": 111.15, + "step": 21360, + "train_speed(iter/s)": 0.448958 + }, + { + "acc": 0.66579733, + "epoch": 0.5419837645865043, + "grad_norm": 5.5625, + "learning_rate": 8.723180598389671e-06, + "loss": 1.52226791, + "memory(GiB)": 111.15, + "step": 21365, + "train_speed(iter/s)": 0.449014 + }, + { + "acc": 0.64282513, + "epoch": 0.5421106037544394, + "grad_norm": 4.21875, + "learning_rate": 8.722480591000046e-06, + "loss": 1.66130066, + "memory(GiB)": 111.15, + "step": 21370, + "train_speed(iter/s)": 0.449069 + }, + { + "acc": 0.66804643, + "epoch": 0.5422374429223744, + "grad_norm": 7.3125, + "learning_rate": 8.721780419879106e-06, + "loss": 1.52520046, + "memory(GiB)": 111.15, + "step": 21375, + "train_speed(iter/s)": 0.449125 + }, + { + "acc": 0.64686227, + "epoch": 0.5423642820903095, + "grad_norm": 4.65625, + "learning_rate": 8.721080085057646e-06, + "loss": 1.70411835, + "memory(GiB)": 111.15, + "step": 21380, + "train_speed(iter/s)": 0.44918 + }, + { + "acc": 0.64346323, + "epoch": 0.5424911212582445, + "grad_norm": 8.125, + "learning_rate": 8.72037958656647e-06, + "loss": 1.72639179, + "memory(GiB)": 111.15, + "step": 21385, + "train_speed(iter/s)": 0.449235 + }, + { + "acc": 0.63702135, + "epoch": 0.5426179604261796, + "grad_norm": 5.0, + "learning_rate": 8.71967892443639e-06, + "loss": 1.68252773, + "memory(GiB)": 111.15, + "step": 21390, + "train_speed(iter/s)": 0.44929 + }, + { + "acc": 0.65590219, + "epoch": 0.5427447995941147, + "grad_norm": 6.125, + "learning_rate": 8.718978098698226e-06, + "loss": 1.61833954, + "memory(GiB)": 111.15, + "step": 21395, + "train_speed(iter/s)": 0.449345 + }, + { + "acc": 0.64735394, + "epoch": 0.5428716387620497, + "grad_norm": 5.59375, + "learning_rate": 8.718277109382799e-06, + "loss": 1.65964622, + "memory(GiB)": 111.15, + "step": 21400, + "train_speed(iter/s)": 0.449401 + }, + { + "epoch": 0.5428716387620497, + "eval_acc": 0.6414520421928271, + "eval_loss": 1.6185967922210693, + "eval_runtime": 113.4324, + "eval_samples_per_second": 56.157, + "eval_steps_per_second": 28.078, + "step": 21400 + }, + { + "acc": 0.65197897, + "epoch": 0.5429984779299848, + "grad_norm": 5.90625, + "learning_rate": 8.717575956520942e-06, + "loss": 1.70419254, + "memory(GiB)": 111.15, + "step": 21405, + "train_speed(iter/s)": 0.448345 + }, + { + "acc": 0.65799093, + "epoch": 0.5431253170979199, + "grad_norm": 4.6875, + "learning_rate": 8.716874640143498e-06, + "loss": 1.56585445, + "memory(GiB)": 111.15, + "step": 21410, + "train_speed(iter/s)": 0.448401 + }, + { + "acc": 0.64862471, + "epoch": 0.5432521562658549, + "grad_norm": 5.375, + "learning_rate": 8.716173160281315e-06, + "loss": 1.69135952, + "memory(GiB)": 111.15, + "step": 21415, + "train_speed(iter/s)": 0.448457 + }, + { + "acc": 0.62319837, + "epoch": 0.54337899543379, + "grad_norm": 6.9375, + "learning_rate": 8.715471516965242e-06, + "loss": 1.72462196, + "memory(GiB)": 111.15, + "step": 21420, + "train_speed(iter/s)": 0.448512 + }, + { + "acc": 0.64214258, + "epoch": 0.543505834601725, + "grad_norm": 5.28125, + "learning_rate": 8.714769710226144e-06, + "loss": 1.67238026, + "memory(GiB)": 111.15, + "step": 21425, + "train_speed(iter/s)": 0.448568 + }, + { + "acc": 0.65179844, + "epoch": 0.5436326737696601, + "grad_norm": 5.53125, + "learning_rate": 8.714067740094888e-06, + "loss": 1.66757545, + "memory(GiB)": 111.15, + "step": 21430, + "train_speed(iter/s)": 0.448623 + }, + { + "acc": 0.66279392, + "epoch": 0.5437595129375952, + "grad_norm": 7.3125, + "learning_rate": 8.713365606602353e-06, + "loss": 1.61598549, + "memory(GiB)": 111.15, + "step": 21435, + "train_speed(iter/s)": 0.448679 + }, + { + "acc": 0.67257566, + "epoch": 0.5438863521055302, + "grad_norm": 5.90625, + "learning_rate": 8.71266330977942e-06, + "loss": 1.51615486, + "memory(GiB)": 111.15, + "step": 21440, + "train_speed(iter/s)": 0.448734 + }, + { + "acc": 0.67364702, + "epoch": 0.5440131912734653, + "grad_norm": 4.59375, + "learning_rate": 8.711960849656975e-06, + "loss": 1.52702885, + "memory(GiB)": 111.15, + "step": 21445, + "train_speed(iter/s)": 0.44879 + }, + { + "acc": 0.64109325, + "epoch": 0.5441400304414004, + "grad_norm": 6.1875, + "learning_rate": 8.711258226265922e-06, + "loss": 1.65407009, + "memory(GiB)": 111.15, + "step": 21450, + "train_speed(iter/s)": 0.448845 + }, + { + "acc": 0.65185413, + "epoch": 0.5442668696093353, + "grad_norm": 5.71875, + "learning_rate": 8.710555439637163e-06, + "loss": 1.68535461, + "memory(GiB)": 111.15, + "step": 21455, + "train_speed(iter/s)": 0.448901 + }, + { + "acc": 0.64048719, + "epoch": 0.5443937087772704, + "grad_norm": 6.5625, + "learning_rate": 8.709852489801608e-06, + "loss": 1.61397476, + "memory(GiB)": 111.15, + "step": 21460, + "train_speed(iter/s)": 0.448957 + }, + { + "acc": 0.64465995, + "epoch": 0.5445205479452054, + "grad_norm": 5.9375, + "learning_rate": 8.709149376790177e-06, + "loss": 1.65963593, + "memory(GiB)": 111.15, + "step": 21465, + "train_speed(iter/s)": 0.449013 + }, + { + "acc": 0.64795084, + "epoch": 0.5446473871131405, + "grad_norm": 5.59375, + "learning_rate": 8.708446100633796e-06, + "loss": 1.650387, + "memory(GiB)": 111.15, + "step": 21470, + "train_speed(iter/s)": 0.449068 + }, + { + "acc": 0.6789876, + "epoch": 0.5447742262810756, + "grad_norm": 6.5625, + "learning_rate": 8.707742661363401e-06, + "loss": 1.51767635, + "memory(GiB)": 111.15, + "step": 21475, + "train_speed(iter/s)": 0.449124 + }, + { + "acc": 0.63189445, + "epoch": 0.5449010654490106, + "grad_norm": 6.25, + "learning_rate": 8.707039059009927e-06, + "loss": 1.72502632, + "memory(GiB)": 111.15, + "step": 21480, + "train_speed(iter/s)": 0.449179 + }, + { + "acc": 0.65762272, + "epoch": 0.5450279046169457, + "grad_norm": 6.28125, + "learning_rate": 8.706335293604326e-06, + "loss": 1.62317352, + "memory(GiB)": 111.15, + "step": 21485, + "train_speed(iter/s)": 0.449235 + }, + { + "acc": 0.6510149, + "epoch": 0.5451547437848808, + "grad_norm": 5.28125, + "learning_rate": 8.705631365177552e-06, + "loss": 1.67762375, + "memory(GiB)": 111.15, + "step": 21490, + "train_speed(iter/s)": 0.44929 + }, + { + "acc": 0.65279026, + "epoch": 0.5452815829528158, + "grad_norm": 5.90625, + "learning_rate": 8.704927273760563e-06, + "loss": 1.6031023, + "memory(GiB)": 111.15, + "step": 21495, + "train_speed(iter/s)": 0.449345 + }, + { + "acc": 0.64373107, + "epoch": 0.5454084221207509, + "grad_norm": 5.15625, + "learning_rate": 8.704223019384334e-06, + "loss": 1.64749565, + "memory(GiB)": 111.15, + "step": 21500, + "train_speed(iter/s)": 0.4494 + }, + { + "epoch": 0.5454084221207509, + "eval_acc": 0.6414470289851137, + "eval_loss": 1.6185108423233032, + "eval_runtime": 112.6306, + "eval_samples_per_second": 56.557, + "eval_steps_per_second": 28.278, + "step": 21500 + }, + { + "acc": 0.65766225, + "epoch": 0.5455352612886859, + "grad_norm": 6.1875, + "learning_rate": 8.703518602079836e-06, + "loss": 1.59496307, + "memory(GiB)": 111.15, + "step": 21505, + "train_speed(iter/s)": 0.448357 + }, + { + "acc": 0.64293242, + "epoch": 0.545662100456621, + "grad_norm": 5.21875, + "learning_rate": 8.702814021878057e-06, + "loss": 1.65087471, + "memory(GiB)": 111.15, + "step": 21510, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.64187841, + "epoch": 0.5457889396245561, + "grad_norm": 5.15625, + "learning_rate": 8.702109278809985e-06, + "loss": 1.69321899, + "memory(GiB)": 111.15, + "step": 21515, + "train_speed(iter/s)": 0.448467 + }, + { + "acc": 0.65694437, + "epoch": 0.5459157787924911, + "grad_norm": 4.5625, + "learning_rate": 8.70140437290662e-06, + "loss": 1.62741871, + "memory(GiB)": 111.15, + "step": 21520, + "train_speed(iter/s)": 0.448522 + }, + { + "acc": 0.64183903, + "epoch": 0.5460426179604262, + "grad_norm": 5.0, + "learning_rate": 8.700699304198963e-06, + "loss": 1.71159515, + "memory(GiB)": 111.15, + "step": 21525, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.64283943, + "epoch": 0.5461694571283613, + "grad_norm": 5.0, + "learning_rate": 8.699994072718026e-06, + "loss": 1.6746933, + "memory(GiB)": 111.15, + "step": 21530, + "train_speed(iter/s)": 0.448632 + }, + { + "acc": 0.65460281, + "epoch": 0.5462962962962963, + "grad_norm": 5.6875, + "learning_rate": 8.699288678494833e-06, + "loss": 1.58495493, + "memory(GiB)": 111.15, + "step": 21535, + "train_speed(iter/s)": 0.448687 + }, + { + "acc": 0.63359537, + "epoch": 0.5464231354642314, + "grad_norm": 4.9375, + "learning_rate": 8.698583121560407e-06, + "loss": 1.71243229, + "memory(GiB)": 111.15, + "step": 21540, + "train_speed(iter/s)": 0.448742 + }, + { + "acc": 0.64332933, + "epoch": 0.5465499746321664, + "grad_norm": 5.375, + "learning_rate": 8.697877401945784e-06, + "loss": 1.65145378, + "memory(GiB)": 111.15, + "step": 21545, + "train_speed(iter/s)": 0.448797 + }, + { + "acc": 0.65854306, + "epoch": 0.5466768138001015, + "grad_norm": 5.15625, + "learning_rate": 8.697171519682002e-06, + "loss": 1.58043222, + "memory(GiB)": 111.15, + "step": 21550, + "train_speed(iter/s)": 0.448853 + }, + { + "acc": 0.65405903, + "epoch": 0.5468036529680366, + "grad_norm": 4.65625, + "learning_rate": 8.696465474800109e-06, + "loss": 1.60827484, + "memory(GiB)": 111.15, + "step": 21555, + "train_speed(iter/s)": 0.448907 + }, + { + "acc": 0.65037785, + "epoch": 0.5469304921359716, + "grad_norm": 5.1875, + "learning_rate": 8.695759267331162e-06, + "loss": 1.67558422, + "memory(GiB)": 111.15, + "step": 21560, + "train_speed(iter/s)": 0.448963 + }, + { + "acc": 0.64573255, + "epoch": 0.5470573313039067, + "grad_norm": 4.5625, + "learning_rate": 8.69505289730622e-06, + "loss": 1.65096264, + "memory(GiB)": 111.15, + "step": 21565, + "train_speed(iter/s)": 0.449018 + }, + { + "acc": 0.66038427, + "epoch": 0.5471841704718418, + "grad_norm": 5.90625, + "learning_rate": 8.694346364756356e-06, + "loss": 1.60301628, + "memory(GiB)": 111.15, + "step": 21570, + "train_speed(iter/s)": 0.449073 + }, + { + "acc": 0.64575472, + "epoch": 0.5473110096397767, + "grad_norm": 4.78125, + "learning_rate": 8.693639669712645e-06, + "loss": 1.65813103, + "memory(GiB)": 111.15, + "step": 21575, + "train_speed(iter/s)": 0.449127 + }, + { + "acc": 0.65283532, + "epoch": 0.5474378488077118, + "grad_norm": 6.0625, + "learning_rate": 8.692932812206171e-06, + "loss": 1.60422668, + "memory(GiB)": 111.15, + "step": 21580, + "train_speed(iter/s)": 0.449183 + }, + { + "acc": 0.64126329, + "epoch": 0.5475646879756468, + "grad_norm": 4.96875, + "learning_rate": 8.692225792268023e-06, + "loss": 1.63855457, + "memory(GiB)": 111.15, + "step": 21585, + "train_speed(iter/s)": 0.449238 + }, + { + "acc": 0.6487524, + "epoch": 0.5476915271435819, + "grad_norm": 5.71875, + "learning_rate": 8.691518609929302e-06, + "loss": 1.66655159, + "memory(GiB)": 111.15, + "step": 21590, + "train_speed(iter/s)": 0.449293 + }, + { + "acc": 0.64873476, + "epoch": 0.547818366311517, + "grad_norm": 4.78125, + "learning_rate": 8.690811265221108e-06, + "loss": 1.63797951, + "memory(GiB)": 111.15, + "step": 21595, + "train_speed(iter/s)": 0.449348 + }, + { + "acc": 0.66307068, + "epoch": 0.547945205479452, + "grad_norm": 5.40625, + "learning_rate": 8.690103758174558e-06, + "loss": 1.6313242, + "memory(GiB)": 111.15, + "step": 21600, + "train_speed(iter/s)": 0.449403 + }, + { + "epoch": 0.547945205479452, + "eval_acc": 0.6415247337046728, + "eval_loss": 1.6186153888702393, + "eval_runtime": 112.6072, + "eval_samples_per_second": 56.568, + "eval_steps_per_second": 28.284, + "step": 21600 + }, + { + "acc": 0.62818432, + "epoch": 0.5480720446473871, + "grad_norm": 5.875, + "learning_rate": 8.68939608882077e-06, + "loss": 1.66409111, + "memory(GiB)": 111.15, + "step": 21605, + "train_speed(iter/s)": 0.448365 + }, + { + "acc": 0.63791056, + "epoch": 0.5481988838153222, + "grad_norm": 6.0, + "learning_rate": 8.688688257190869e-06, + "loss": 1.66916542, + "memory(GiB)": 111.15, + "step": 21610, + "train_speed(iter/s)": 0.44842 + }, + { + "acc": 0.66799312, + "epoch": 0.5483257229832572, + "grad_norm": 5.90625, + "learning_rate": 8.68798026331599e-06, + "loss": 1.54684944, + "memory(GiB)": 111.15, + "step": 21615, + "train_speed(iter/s)": 0.448475 + }, + { + "acc": 0.64481955, + "epoch": 0.5484525621511923, + "grad_norm": 5.09375, + "learning_rate": 8.687272107227274e-06, + "loss": 1.68513298, + "memory(GiB)": 111.15, + "step": 21620, + "train_speed(iter/s)": 0.448529 + }, + { + "acc": 0.63986416, + "epoch": 0.5485794013191273, + "grad_norm": 4.3125, + "learning_rate": 8.686563788955867e-06, + "loss": 1.6848093, + "memory(GiB)": 111.15, + "step": 21625, + "train_speed(iter/s)": 0.448584 + }, + { + "acc": 0.65453272, + "epoch": 0.5487062404870624, + "grad_norm": 5.125, + "learning_rate": 8.685855308532926e-06, + "loss": 1.6016346, + "memory(GiB)": 111.15, + "step": 21630, + "train_speed(iter/s)": 0.448639 + }, + { + "acc": 0.65434923, + "epoch": 0.5488330796549975, + "grad_norm": 6.53125, + "learning_rate": 8.685146665989613e-06, + "loss": 1.63238029, + "memory(GiB)": 111.15, + "step": 21635, + "train_speed(iter/s)": 0.448694 + }, + { + "acc": 0.63961763, + "epoch": 0.5489599188229325, + "grad_norm": 5.5625, + "learning_rate": 8.684437861357095e-06, + "loss": 1.63574715, + "memory(GiB)": 111.15, + "step": 21640, + "train_speed(iter/s)": 0.448749 + }, + { + "acc": 0.64578218, + "epoch": 0.5490867579908676, + "grad_norm": 5.40625, + "learning_rate": 8.683728894666551e-06, + "loss": 1.59146423, + "memory(GiB)": 111.15, + "step": 21645, + "train_speed(iter/s)": 0.448803 + }, + { + "acc": 0.64896693, + "epoch": 0.5492135971588027, + "grad_norm": 5.4375, + "learning_rate": 8.683019765949163e-06, + "loss": 1.64683819, + "memory(GiB)": 111.15, + "step": 21650, + "train_speed(iter/s)": 0.448858 + }, + { + "acc": 0.63846321, + "epoch": 0.5493404363267377, + "grad_norm": 5.4375, + "learning_rate": 8.682310475236123e-06, + "loss": 1.6304924, + "memory(GiB)": 111.15, + "step": 21655, + "train_speed(iter/s)": 0.448913 + }, + { + "acc": 0.64225903, + "epoch": 0.5494672754946728, + "grad_norm": 5.65625, + "learning_rate": 8.681601022558628e-06, + "loss": 1.6579977, + "memory(GiB)": 111.15, + "step": 21660, + "train_speed(iter/s)": 0.448967 + }, + { + "acc": 0.64812336, + "epoch": 0.5495941146626078, + "grad_norm": 5.0625, + "learning_rate": 8.680891407947882e-06, + "loss": 1.61439323, + "memory(GiB)": 111.15, + "step": 21665, + "train_speed(iter/s)": 0.449022 + }, + { + "acc": 0.64019709, + "epoch": 0.5497209538305429, + "grad_norm": 5.21875, + "learning_rate": 8.680181631435098e-06, + "loss": 1.73579063, + "memory(GiB)": 111.15, + "step": 21670, + "train_speed(iter/s)": 0.449077 + }, + { + "acc": 0.65300231, + "epoch": 0.549847792998478, + "grad_norm": 6.46875, + "learning_rate": 8.679471693051495e-06, + "loss": 1.62649555, + "memory(GiB)": 111.15, + "step": 21675, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.65462399, + "epoch": 0.549974632166413, + "grad_norm": 4.8125, + "learning_rate": 8.678761592828301e-06, + "loss": 1.61235466, + "memory(GiB)": 111.15, + "step": 21680, + "train_speed(iter/s)": 0.449186 + }, + { + "acc": 0.65373678, + "epoch": 0.5501014713343481, + "grad_norm": 5.0625, + "learning_rate": 8.678051330796746e-06, + "loss": 1.59236908, + "memory(GiB)": 111.15, + "step": 21685, + "train_speed(iter/s)": 0.44924 + }, + { + "acc": 0.63811288, + "epoch": 0.5502283105022832, + "grad_norm": 5.34375, + "learning_rate": 8.677340906988072e-06, + "loss": 1.64484406, + "memory(GiB)": 111.15, + "step": 21690, + "train_speed(iter/s)": 0.449295 + }, + { + "acc": 0.64998856, + "epoch": 0.5503551496702181, + "grad_norm": 5.5625, + "learning_rate": 8.676630321433528e-06, + "loss": 1.6078846, + "memory(GiB)": 111.15, + "step": 21695, + "train_speed(iter/s)": 0.449349 + }, + { + "acc": 0.6526998, + "epoch": 0.5504819888381532, + "grad_norm": 5.1875, + "learning_rate": 8.675919574164366e-06, + "loss": 1.63919621, + "memory(GiB)": 111.15, + "step": 21700, + "train_speed(iter/s)": 0.449403 + }, + { + "epoch": 0.5504819888381532, + "eval_acc": 0.6414553843313029, + "eval_loss": 1.6181838512420654, + "eval_runtime": 113.2487, + "eval_samples_per_second": 56.248, + "eval_steps_per_second": 28.124, + "step": 21700 + }, + { + "acc": 0.64598684, + "epoch": 0.5506088280060882, + "grad_norm": 4.53125, + "learning_rate": 8.675208665211851e-06, + "loss": 1.650103, + "memory(GiB)": 111.15, + "step": 21705, + "train_speed(iter/s)": 0.448363 + }, + { + "acc": 0.64461842, + "epoch": 0.5507356671740233, + "grad_norm": 6.59375, + "learning_rate": 8.674497594607249e-06, + "loss": 1.63999214, + "memory(GiB)": 111.15, + "step": 21710, + "train_speed(iter/s)": 0.448418 + }, + { + "acc": 0.64075575, + "epoch": 0.5508625063419584, + "grad_norm": 6.25, + "learning_rate": 8.673786362381837e-06, + "loss": 1.6681221, + "memory(GiB)": 111.15, + "step": 21715, + "train_speed(iter/s)": 0.448472 + }, + { + "acc": 0.64351091, + "epoch": 0.5509893455098934, + "grad_norm": 4.53125, + "learning_rate": 8.673074968566899e-06, + "loss": 1.59686279, + "memory(GiB)": 111.15, + "step": 21720, + "train_speed(iter/s)": 0.448526 + }, + { + "acc": 0.6282083, + "epoch": 0.5511161846778285, + "grad_norm": 5.78125, + "learning_rate": 8.672363413193724e-06, + "loss": 1.71565056, + "memory(GiB)": 111.15, + "step": 21725, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.65201092, + "epoch": 0.5512430238457636, + "grad_norm": 4.28125, + "learning_rate": 8.671651696293613e-06, + "loss": 1.60431442, + "memory(GiB)": 111.15, + "step": 21730, + "train_speed(iter/s)": 0.448635 + }, + { + "acc": 0.64553738, + "epoch": 0.5513698630136986, + "grad_norm": 5.59375, + "learning_rate": 8.670939817897865e-06, + "loss": 1.69724274, + "memory(GiB)": 111.15, + "step": 21735, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.63552837, + "epoch": 0.5514967021816337, + "grad_norm": 4.40625, + "learning_rate": 8.670227778037796e-06, + "loss": 1.67347488, + "memory(GiB)": 111.15, + "step": 21740, + "train_speed(iter/s)": 0.448744 + }, + { + "acc": 0.65271053, + "epoch": 0.5516235413495687, + "grad_norm": 5.75, + "learning_rate": 8.669515576744722e-06, + "loss": 1.63084335, + "memory(GiB)": 111.15, + "step": 21745, + "train_speed(iter/s)": 0.448798 + }, + { + "acc": 0.64041672, + "epoch": 0.5517503805175038, + "grad_norm": 5.5625, + "learning_rate": 8.66880321404997e-06, + "loss": 1.66839828, + "memory(GiB)": 111.15, + "step": 21750, + "train_speed(iter/s)": 0.448853 + }, + { + "acc": 0.66144991, + "epoch": 0.5518772196854389, + "grad_norm": 6.625, + "learning_rate": 8.668090689984872e-06, + "loss": 1.61604652, + "memory(GiB)": 111.15, + "step": 21755, + "train_speed(iter/s)": 0.448907 + }, + { + "acc": 0.64550562, + "epoch": 0.5520040588533739, + "grad_norm": 5.625, + "learning_rate": 8.667378004580769e-06, + "loss": 1.59961491, + "memory(GiB)": 111.15, + "step": 21760, + "train_speed(iter/s)": 0.448962 + }, + { + "acc": 0.65925922, + "epoch": 0.552130898021309, + "grad_norm": 5.09375, + "learning_rate": 8.666665157869007e-06, + "loss": 1.59329557, + "memory(GiB)": 111.15, + "step": 21765, + "train_speed(iter/s)": 0.449017 + }, + { + "acc": 0.64432983, + "epoch": 0.5522577371892441, + "grad_norm": 5.09375, + "learning_rate": 8.665952149880942e-06, + "loss": 1.63972282, + "memory(GiB)": 111.15, + "step": 21770, + "train_speed(iter/s)": 0.449071 + }, + { + "acc": 0.66100316, + "epoch": 0.5523845763571791, + "grad_norm": 5.28125, + "learning_rate": 8.665238980647934e-06, + "loss": 1.55054169, + "memory(GiB)": 111.15, + "step": 21775, + "train_speed(iter/s)": 0.449125 + }, + { + "acc": 0.64277029, + "epoch": 0.5525114155251142, + "grad_norm": 5.6875, + "learning_rate": 8.66452565020135e-06, + "loss": 1.65826988, + "memory(GiB)": 111.15, + "step": 21780, + "train_speed(iter/s)": 0.44918 + }, + { + "acc": 0.6481616, + "epoch": 0.5526382546930492, + "grad_norm": 5.875, + "learning_rate": 8.663812158572568e-06, + "loss": 1.66529732, + "memory(GiB)": 111.15, + "step": 21785, + "train_speed(iter/s)": 0.449234 + }, + { + "acc": 0.6515821, + "epoch": 0.5527650938609843, + "grad_norm": 6.71875, + "learning_rate": 8.663098505792971e-06, + "loss": 1.63503036, + "memory(GiB)": 111.15, + "step": 21790, + "train_speed(iter/s)": 0.449289 + }, + { + "acc": 0.63934832, + "epoch": 0.5528919330289194, + "grad_norm": 4.8125, + "learning_rate": 8.662384691893947e-06, + "loss": 1.59750223, + "memory(GiB)": 111.15, + "step": 21795, + "train_speed(iter/s)": 0.449343 + }, + { + "acc": 0.65262384, + "epoch": 0.5530187721968544, + "grad_norm": 4.96875, + "learning_rate": 8.661670716906889e-06, + "loss": 1.55680695, + "memory(GiB)": 111.15, + "step": 21800, + "train_speed(iter/s)": 0.449398 + }, + { + "epoch": 0.5530187721968544, + "eval_acc": 0.6415147072892458, + "eval_loss": 1.617859125137329, + "eval_runtime": 112.4774, + "eval_samples_per_second": 56.634, + "eval_steps_per_second": 28.317, + "step": 21800 + }, + { + "acc": 0.65217867, + "epoch": 0.5531456113647895, + "grad_norm": 5.03125, + "learning_rate": 8.66095658086321e-06, + "loss": 1.69379807, + "memory(GiB)": 111.15, + "step": 21805, + "train_speed(iter/s)": 0.448371 + }, + { + "acc": 0.64295325, + "epoch": 0.5532724505327246, + "grad_norm": 4.75, + "learning_rate": 8.660242283794312e-06, + "loss": 1.61960831, + "memory(GiB)": 111.15, + "step": 21810, + "train_speed(iter/s)": 0.448424 + }, + { + "acc": 0.64025841, + "epoch": 0.5533992897006595, + "grad_norm": 5.4375, + "learning_rate": 8.659527825731617e-06, + "loss": 1.67250557, + "memory(GiB)": 111.15, + "step": 21815, + "train_speed(iter/s)": 0.448478 + }, + { + "acc": 0.64501858, + "epoch": 0.5535261288685946, + "grad_norm": 6.0, + "learning_rate": 8.65881320670655e-06, + "loss": 1.6179039, + "memory(GiB)": 111.15, + "step": 21820, + "train_speed(iter/s)": 0.448532 + }, + { + "acc": 0.65232639, + "epoch": 0.5536529680365296, + "grad_norm": 5.34375, + "learning_rate": 8.658098426750543e-06, + "loss": 1.65876198, + "memory(GiB)": 111.15, + "step": 21825, + "train_speed(iter/s)": 0.448586 + }, + { + "acc": 0.63486843, + "epoch": 0.5537798072044647, + "grad_norm": 4.84375, + "learning_rate": 8.657383485895034e-06, + "loss": 1.68514462, + "memory(GiB)": 111.15, + "step": 21830, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.66327477, + "epoch": 0.5539066463723998, + "grad_norm": 5.5625, + "learning_rate": 8.656668384171472e-06, + "loss": 1.65839043, + "memory(GiB)": 111.15, + "step": 21835, + "train_speed(iter/s)": 0.448693 + }, + { + "acc": 0.6686954, + "epoch": 0.5540334855403348, + "grad_norm": 6.78125, + "learning_rate": 8.655953121611307e-06, + "loss": 1.61207428, + "memory(GiB)": 111.15, + "step": 21840, + "train_speed(iter/s)": 0.448748 + }, + { + "acc": 0.64476252, + "epoch": 0.5541603247082699, + "grad_norm": 4.90625, + "learning_rate": 8.655237698246002e-06, + "loss": 1.62890472, + "memory(GiB)": 111.15, + "step": 21845, + "train_speed(iter/s)": 0.448801 + }, + { + "acc": 0.65166183, + "epoch": 0.554287163876205, + "grad_norm": 5.375, + "learning_rate": 8.654522114107024e-06, + "loss": 1.60706635, + "memory(GiB)": 111.15, + "step": 21850, + "train_speed(iter/s)": 0.448855 + }, + { + "acc": 0.65330381, + "epoch": 0.55441400304414, + "grad_norm": 5.1875, + "learning_rate": 8.653806369225846e-06, + "loss": 1.63259983, + "memory(GiB)": 111.15, + "step": 21855, + "train_speed(iter/s)": 0.448909 + }, + { + "acc": 0.64125605, + "epoch": 0.5545408422120751, + "grad_norm": 5.125, + "learning_rate": 8.65309046363395e-06, + "loss": 1.61428738, + "memory(GiB)": 111.15, + "step": 21860, + "train_speed(iter/s)": 0.448963 + }, + { + "acc": 0.64313545, + "epoch": 0.5546676813800101, + "grad_norm": 5.28125, + "learning_rate": 8.652374397362828e-06, + "loss": 1.6119875, + "memory(GiB)": 111.15, + "step": 21865, + "train_speed(iter/s)": 0.449016 + }, + { + "acc": 0.65544415, + "epoch": 0.5547945205479452, + "grad_norm": 4.625, + "learning_rate": 8.651658170443972e-06, + "loss": 1.53138046, + "memory(GiB)": 111.15, + "step": 21870, + "train_speed(iter/s)": 0.44907 + }, + { + "acc": 0.65568943, + "epoch": 0.5549213597158803, + "grad_norm": 5.25, + "learning_rate": 8.650941782908886e-06, + "loss": 1.60263939, + "memory(GiB)": 111.15, + "step": 21875, + "train_speed(iter/s)": 0.449124 + }, + { + "acc": 0.65980635, + "epoch": 0.5550481988838153, + "grad_norm": 4.25, + "learning_rate": 8.65022523478908e-06, + "loss": 1.58667479, + "memory(GiB)": 111.15, + "step": 21880, + "train_speed(iter/s)": 0.449178 + }, + { + "acc": 0.65078554, + "epoch": 0.5551750380517504, + "grad_norm": 5.0625, + "learning_rate": 8.649508526116073e-06, + "loss": 1.60160217, + "memory(GiB)": 111.15, + "step": 21885, + "train_speed(iter/s)": 0.449232 + }, + { + "acc": 0.64540858, + "epoch": 0.5553018772196855, + "grad_norm": 5.5, + "learning_rate": 8.648791656921384e-06, + "loss": 1.59335032, + "memory(GiB)": 111.15, + "step": 21890, + "train_speed(iter/s)": 0.449286 + }, + { + "acc": 0.65057983, + "epoch": 0.5554287163876205, + "grad_norm": 4.6875, + "learning_rate": 8.648074627236549e-06, + "loss": 1.62528992, + "memory(GiB)": 111.15, + "step": 21895, + "train_speed(iter/s)": 0.44934 + }, + { + "acc": 0.64766092, + "epoch": 0.5555555555555556, + "grad_norm": 6.9375, + "learning_rate": 8.647357437093104e-06, + "loss": 1.62623539, + "memory(GiB)": 111.15, + "step": 21900, + "train_speed(iter/s)": 0.449393 + }, + { + "epoch": 0.5555555555555556, + "eval_acc": 0.6415218093335066, + "eval_loss": 1.6180626153945923, + "eval_runtime": 114.0246, + "eval_samples_per_second": 55.865, + "eval_steps_per_second": 27.933, + "step": 21900 + }, + { + "acc": 0.65119672, + "epoch": 0.5556823947234906, + "grad_norm": 4.96875, + "learning_rate": 8.646640086522595e-06, + "loss": 1.65577888, + "memory(GiB)": 111.15, + "step": 21905, + "train_speed(iter/s)": 0.448357 + }, + { + "acc": 0.64759688, + "epoch": 0.5558092338914257, + "grad_norm": 5.71875, + "learning_rate": 8.645922575556575e-06, + "loss": 1.70955887, + "memory(GiB)": 111.15, + "step": 21910, + "train_speed(iter/s)": 0.448411 + }, + { + "acc": 0.64674435, + "epoch": 0.5559360730593608, + "grad_norm": 4.8125, + "learning_rate": 8.645204904226601e-06, + "loss": 1.64823723, + "memory(GiB)": 111.15, + "step": 21915, + "train_speed(iter/s)": 0.448465 + }, + { + "acc": 0.65998111, + "epoch": 0.5560629122272958, + "grad_norm": 6.40625, + "learning_rate": 8.64448707256424e-06, + "loss": 1.5726203, + "memory(GiB)": 111.15, + "step": 21920, + "train_speed(iter/s)": 0.448519 + }, + { + "acc": 0.65379362, + "epoch": 0.5561897513952309, + "grad_norm": 4.53125, + "learning_rate": 8.643769080601067e-06, + "loss": 1.58335419, + "memory(GiB)": 111.15, + "step": 21925, + "train_speed(iter/s)": 0.448573 + }, + { + "acc": 0.66494131, + "epoch": 0.556316590563166, + "grad_norm": 5.78125, + "learning_rate": 8.643050928368661e-06, + "loss": 1.56656313, + "memory(GiB)": 111.15, + "step": 21930, + "train_speed(iter/s)": 0.448627 + }, + { + "acc": 0.66048045, + "epoch": 0.556443429731101, + "grad_norm": 5.75, + "learning_rate": 8.642332615898611e-06, + "loss": 1.61121655, + "memory(GiB)": 111.15, + "step": 21935, + "train_speed(iter/s)": 0.448681 + }, + { + "acc": 0.65159473, + "epoch": 0.556570268899036, + "grad_norm": 5.28125, + "learning_rate": 8.64161414322251e-06, + "loss": 1.57150555, + "memory(GiB)": 111.15, + "step": 21940, + "train_speed(iter/s)": 0.448735 + }, + { + "acc": 0.66407423, + "epoch": 0.556697108066971, + "grad_norm": 5.40625, + "learning_rate": 8.64089551037196e-06, + "loss": 1.59064198, + "memory(GiB)": 111.15, + "step": 21945, + "train_speed(iter/s)": 0.448789 + }, + { + "acc": 0.65001612, + "epoch": 0.5568239472349061, + "grad_norm": 4.65625, + "learning_rate": 8.640176717378573e-06, + "loss": 1.62269287, + "memory(GiB)": 111.15, + "step": 21950, + "train_speed(iter/s)": 0.448843 + }, + { + "acc": 0.65759926, + "epoch": 0.5569507864028412, + "grad_norm": 5.90625, + "learning_rate": 8.639457764273957e-06, + "loss": 1.65928764, + "memory(GiB)": 111.15, + "step": 21955, + "train_speed(iter/s)": 0.448897 + }, + { + "acc": 0.6348681, + "epoch": 0.5570776255707762, + "grad_norm": 5.4375, + "learning_rate": 8.638738651089744e-06, + "loss": 1.65712051, + "memory(GiB)": 111.15, + "step": 21960, + "train_speed(iter/s)": 0.448951 + }, + { + "acc": 0.63771477, + "epoch": 0.5572044647387113, + "grad_norm": 5.4375, + "learning_rate": 8.638019377857555e-06, + "loss": 1.66657715, + "memory(GiB)": 111.15, + "step": 21965, + "train_speed(iter/s)": 0.449006 + }, + { + "acc": 0.6475069, + "epoch": 0.5573313039066464, + "grad_norm": 6.53125, + "learning_rate": 8.637299944609034e-06, + "loss": 1.69087906, + "memory(GiB)": 111.15, + "step": 21970, + "train_speed(iter/s)": 0.44906 + }, + { + "acc": 0.65433645, + "epoch": 0.5574581430745814, + "grad_norm": 6.46875, + "learning_rate": 8.636580351375821e-06, + "loss": 1.64633293, + "memory(GiB)": 111.15, + "step": 21975, + "train_speed(iter/s)": 0.449114 + }, + { + "acc": 0.63579054, + "epoch": 0.5575849822425165, + "grad_norm": 6.28125, + "learning_rate": 8.635860598189569e-06, + "loss": 1.7326685, + "memory(GiB)": 111.15, + "step": 21980, + "train_speed(iter/s)": 0.449169 + }, + { + "acc": 0.63710442, + "epoch": 0.5577118214104515, + "grad_norm": 4.5625, + "learning_rate": 8.635140685081936e-06, + "loss": 1.69505692, + "memory(GiB)": 111.15, + "step": 21985, + "train_speed(iter/s)": 0.449223 + }, + { + "acc": 0.65159044, + "epoch": 0.5578386605783866, + "grad_norm": 5.8125, + "learning_rate": 8.634420612084583e-06, + "loss": 1.62598457, + "memory(GiB)": 111.15, + "step": 21990, + "train_speed(iter/s)": 0.449278 + }, + { + "acc": 0.6497632, + "epoch": 0.5579654997463217, + "grad_norm": 5.125, + "learning_rate": 8.633700379229187e-06, + "loss": 1.65549564, + "memory(GiB)": 111.15, + "step": 21995, + "train_speed(iter/s)": 0.449332 + }, + { + "acc": 0.64530687, + "epoch": 0.5580923389142567, + "grad_norm": 5.71875, + "learning_rate": 8.632979986547423e-06, + "loss": 1.69569473, + "memory(GiB)": 111.15, + "step": 22000, + "train_speed(iter/s)": 0.449386 + }, + { + "epoch": 0.5580923389142567, + "eval_acc": 0.6415071874776755, + "eval_loss": 1.6175004243850708, + "eval_runtime": 112.5564, + "eval_samples_per_second": 56.594, + "eval_steps_per_second": 28.297, + "step": 22000 + }, + { + "acc": 0.6449892, + "epoch": 0.5582191780821918, + "grad_norm": 4.9375, + "learning_rate": 8.632259434070982e-06, + "loss": 1.65703793, + "memory(GiB)": 111.15, + "step": 22005, + "train_speed(iter/s)": 0.448368 + }, + { + "acc": 0.66050978, + "epoch": 0.5583460172501269, + "grad_norm": 5.25, + "learning_rate": 8.631538721831551e-06, + "loss": 1.62177315, + "memory(GiB)": 111.15, + "step": 22010, + "train_speed(iter/s)": 0.448421 + }, + { + "acc": 0.64068317, + "epoch": 0.5584728564180619, + "grad_norm": 4.6875, + "learning_rate": 8.630817849860835e-06, + "loss": 1.72173977, + "memory(GiB)": 111.15, + "step": 22015, + "train_speed(iter/s)": 0.448475 + }, + { + "acc": 0.64966583, + "epoch": 0.558599695585997, + "grad_norm": 6.125, + "learning_rate": 8.63009681819054e-06, + "loss": 1.5987977, + "memory(GiB)": 111.15, + "step": 22020, + "train_speed(iter/s)": 0.448529 + }, + { + "acc": 0.63976583, + "epoch": 0.558726534753932, + "grad_norm": 5.25, + "learning_rate": 8.629375626852378e-06, + "loss": 1.61269588, + "memory(GiB)": 111.15, + "step": 22025, + "train_speed(iter/s)": 0.448583 + }, + { + "acc": 0.64208803, + "epoch": 0.5588533739218671, + "grad_norm": 4.40625, + "learning_rate": 8.628654275878074e-06, + "loss": 1.65279236, + "memory(GiB)": 111.15, + "step": 22030, + "train_speed(iter/s)": 0.448636 + }, + { + "acc": 0.65324855, + "epoch": 0.5589802130898022, + "grad_norm": 4.75, + "learning_rate": 8.627932765299353e-06, + "loss": 1.61902008, + "memory(GiB)": 111.15, + "step": 22035, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.65356035, + "epoch": 0.5591070522577372, + "grad_norm": 5.15625, + "learning_rate": 8.627211095147952e-06, + "loss": 1.57886372, + "memory(GiB)": 111.15, + "step": 22040, + "train_speed(iter/s)": 0.448744 + }, + { + "acc": 0.6650207, + "epoch": 0.5592338914256723, + "grad_norm": 5.8125, + "learning_rate": 8.626489265455614e-06, + "loss": 1.70296974, + "memory(GiB)": 111.15, + "step": 22045, + "train_speed(iter/s)": 0.448797 + }, + { + "acc": 0.64016476, + "epoch": 0.5593607305936074, + "grad_norm": 5.28125, + "learning_rate": 8.625767276254084e-06, + "loss": 1.68468399, + "memory(GiB)": 111.15, + "step": 22050, + "train_speed(iter/s)": 0.44885 + }, + { + "acc": 0.64022017, + "epoch": 0.5594875697615423, + "grad_norm": 5.21875, + "learning_rate": 8.625045127575123e-06, + "loss": 1.69359055, + "memory(GiB)": 111.15, + "step": 22055, + "train_speed(iter/s)": 0.448903 + }, + { + "acc": 0.65500741, + "epoch": 0.5596144089294774, + "grad_norm": 8.3125, + "learning_rate": 8.624322819450493e-06, + "loss": 1.68929501, + "memory(GiB)": 111.15, + "step": 22060, + "train_speed(iter/s)": 0.448957 + }, + { + "acc": 0.65873671, + "epoch": 0.5597412480974124, + "grad_norm": 6.4375, + "learning_rate": 8.623600351911962e-06, + "loss": 1.56352463, + "memory(GiB)": 111.15, + "step": 22065, + "train_speed(iter/s)": 0.44901 + }, + { + "acc": 0.65598564, + "epoch": 0.5598680872653475, + "grad_norm": 4.875, + "learning_rate": 8.622877724991312e-06, + "loss": 1.64914055, + "memory(GiB)": 111.15, + "step": 22070, + "train_speed(iter/s)": 0.449063 + }, + { + "acc": 0.65805092, + "epoch": 0.5599949264332826, + "grad_norm": 4.78125, + "learning_rate": 8.622154938720323e-06, + "loss": 1.55282898, + "memory(GiB)": 111.15, + "step": 22075, + "train_speed(iter/s)": 0.449117 + }, + { + "acc": 0.64300566, + "epoch": 0.5601217656012176, + "grad_norm": 5.0625, + "learning_rate": 8.621431993130787e-06, + "loss": 1.64658241, + "memory(GiB)": 111.15, + "step": 22080, + "train_speed(iter/s)": 0.449171 + }, + { + "acc": 0.66366014, + "epoch": 0.5602486047691527, + "grad_norm": 4.65625, + "learning_rate": 8.620708888254506e-06, + "loss": 1.59407072, + "memory(GiB)": 111.15, + "step": 22085, + "train_speed(iter/s)": 0.449224 + }, + { + "acc": 0.64669399, + "epoch": 0.5603754439370878, + "grad_norm": 5.34375, + "learning_rate": 8.619985624123282e-06, + "loss": 1.58780365, + "memory(GiB)": 111.15, + "step": 22090, + "train_speed(iter/s)": 0.449278 + }, + { + "acc": 0.66421733, + "epoch": 0.5605022831050228, + "grad_norm": 4.65625, + "learning_rate": 8.619262200768928e-06, + "loss": 1.56993999, + "memory(GiB)": 111.15, + "step": 22095, + "train_speed(iter/s)": 0.449331 + }, + { + "acc": 0.66410046, + "epoch": 0.5606291222729579, + "grad_norm": 4.75, + "learning_rate": 8.618538618223262e-06, + "loss": 1.60599079, + "memory(GiB)": 111.15, + "step": 22100, + "train_speed(iter/s)": 0.449384 + }, + { + "epoch": 0.5606291222729579, + "eval_acc": 0.6416133003742778, + "eval_loss": 1.6177424192428589, + "eval_runtime": 113.3842, + "eval_samples_per_second": 56.181, + "eval_steps_per_second": 28.09, + "step": 22100 + }, + { + "acc": 0.66068826, + "epoch": 0.5607559614408929, + "grad_norm": 5.59375, + "learning_rate": 8.617814876518114e-06, + "loss": 1.5905159, + "memory(GiB)": 111.15, + "step": 22105, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.65260277, + "epoch": 0.560882800608828, + "grad_norm": 7.15625, + "learning_rate": 8.617090975685314e-06, + "loss": 1.67046165, + "memory(GiB)": 111.15, + "step": 22110, + "train_speed(iter/s)": 0.448415 + }, + { + "acc": 0.64956598, + "epoch": 0.5610096397767631, + "grad_norm": 5.4375, + "learning_rate": 8.616366915756704e-06, + "loss": 1.61902237, + "memory(GiB)": 111.15, + "step": 22115, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.64977179, + "epoch": 0.5611364789446981, + "grad_norm": 4.71875, + "learning_rate": 8.615642696764131e-06, + "loss": 1.64774284, + "memory(GiB)": 111.15, + "step": 22120, + "train_speed(iter/s)": 0.44852 + }, + { + "acc": 0.65350876, + "epoch": 0.5612633181126332, + "grad_norm": 4.9375, + "learning_rate": 8.614918318739452e-06, + "loss": 1.64072514, + "memory(GiB)": 111.15, + "step": 22125, + "train_speed(iter/s)": 0.448573 + }, + { + "acc": 0.65520549, + "epoch": 0.5613901572805683, + "grad_norm": 5.84375, + "learning_rate": 8.614193781714522e-06, + "loss": 1.64116211, + "memory(GiB)": 111.15, + "step": 22130, + "train_speed(iter/s)": 0.448625 + }, + { + "acc": 0.63991742, + "epoch": 0.5615169964485033, + "grad_norm": 5.03125, + "learning_rate": 8.613469085721215e-06, + "loss": 1.64511681, + "memory(GiB)": 111.15, + "step": 22135, + "train_speed(iter/s)": 0.448678 + }, + { + "acc": 0.6484818, + "epoch": 0.5616438356164384, + "grad_norm": 5.46875, + "learning_rate": 8.612744230791406e-06, + "loss": 1.67967453, + "memory(GiB)": 111.15, + "step": 22140, + "train_speed(iter/s)": 0.448732 + }, + { + "acc": 0.65529613, + "epoch": 0.5617706747843734, + "grad_norm": 5.84375, + "learning_rate": 8.612019216956975e-06, + "loss": 1.57116394, + "memory(GiB)": 111.15, + "step": 22145, + "train_speed(iter/s)": 0.448785 + }, + { + "acc": 0.64280081, + "epoch": 0.5618975139523085, + "grad_norm": 5.03125, + "learning_rate": 8.611294044249811e-06, + "loss": 1.62599049, + "memory(GiB)": 111.15, + "step": 22150, + "train_speed(iter/s)": 0.448837 + }, + { + "acc": 0.64896355, + "epoch": 0.5620243531202436, + "grad_norm": 5.4375, + "learning_rate": 8.610568712701814e-06, + "loss": 1.61730423, + "memory(GiB)": 111.15, + "step": 22155, + "train_speed(iter/s)": 0.44889 + }, + { + "acc": 0.64281354, + "epoch": 0.5621511922881786, + "grad_norm": 4.75, + "learning_rate": 8.609843222344883e-06, + "loss": 1.6481041, + "memory(GiB)": 111.15, + "step": 22160, + "train_speed(iter/s)": 0.448943 + }, + { + "acc": 0.65657406, + "epoch": 0.5622780314561137, + "grad_norm": 4.9375, + "learning_rate": 8.609117573210931e-06, + "loss": 1.52045975, + "memory(GiB)": 111.15, + "step": 22165, + "train_speed(iter/s)": 0.448997 + }, + { + "acc": 0.6487792, + "epoch": 0.5624048706240488, + "grad_norm": 5.21875, + "learning_rate": 8.608391765331875e-06, + "loss": 1.60518799, + "memory(GiB)": 111.15, + "step": 22170, + "train_speed(iter/s)": 0.449051 + }, + { + "acc": 0.65609574, + "epoch": 0.5625317097919837, + "grad_norm": 5.0, + "learning_rate": 8.607665798739638e-06, + "loss": 1.62755928, + "memory(GiB)": 111.15, + "step": 22175, + "train_speed(iter/s)": 0.449104 + }, + { + "acc": 0.65439491, + "epoch": 0.5626585489599188, + "grad_norm": 5.6875, + "learning_rate": 8.606939673466153e-06, + "loss": 1.61549797, + "memory(GiB)": 111.15, + "step": 22180, + "train_speed(iter/s)": 0.449158 + }, + { + "acc": 0.65275669, + "epoch": 0.5627853881278538, + "grad_norm": 6.0625, + "learning_rate": 8.606213389543356e-06, + "loss": 1.61544914, + "memory(GiB)": 111.15, + "step": 22185, + "train_speed(iter/s)": 0.44921 + }, + { + "acc": 0.64597597, + "epoch": 0.5629122272957889, + "grad_norm": 5.875, + "learning_rate": 8.605486947003194e-06, + "loss": 1.68403969, + "memory(GiB)": 111.15, + "step": 22190, + "train_speed(iter/s)": 0.449263 + }, + { + "acc": 0.63549681, + "epoch": 0.563039066463724, + "grad_norm": 4.6875, + "learning_rate": 8.60476034587762e-06, + "loss": 1.70922699, + "memory(GiB)": 111.15, + "step": 22195, + "train_speed(iter/s)": 0.449316 + }, + { + "acc": 0.65520959, + "epoch": 0.563165905631659, + "grad_norm": 5.0625, + "learning_rate": 8.604033586198592e-06, + "loss": 1.62609882, + "memory(GiB)": 111.15, + "step": 22200, + "train_speed(iter/s)": 0.449369 + }, + { + "epoch": 0.563165905631659, + "eval_acc": 0.6416575837090802, + "eval_loss": 1.6172444820404053, + "eval_runtime": 113.0106, + "eval_samples_per_second": 56.366, + "eval_steps_per_second": 28.183, + "step": 22200 + }, + { + "acc": 0.63160825, + "epoch": 0.5632927447995941, + "grad_norm": 5.59375, + "learning_rate": 8.603306667998074e-06, + "loss": 1.69084473, + "memory(GiB)": 111.15, + "step": 22205, + "train_speed(iter/s)": 0.448356 + }, + { + "acc": 0.64282274, + "epoch": 0.5634195839675292, + "grad_norm": 5.3125, + "learning_rate": 8.602579591308043e-06, + "loss": 1.66867104, + "memory(GiB)": 111.15, + "step": 22210, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.62530069, + "epoch": 0.5635464231354642, + "grad_norm": 5.25, + "learning_rate": 8.601852356160476e-06, + "loss": 1.70241528, + "memory(GiB)": 111.15, + "step": 22215, + "train_speed(iter/s)": 0.448464 + }, + { + "acc": 0.65207691, + "epoch": 0.5636732623033993, + "grad_norm": 5.125, + "learning_rate": 8.60112496258736e-06, + "loss": 1.61646423, + "memory(GiB)": 111.15, + "step": 22220, + "train_speed(iter/s)": 0.448518 + }, + { + "acc": 0.63869324, + "epoch": 0.5638001014713343, + "grad_norm": 4.84375, + "learning_rate": 8.600397410620693e-06, + "loss": 1.69360142, + "memory(GiB)": 111.15, + "step": 22225, + "train_speed(iter/s)": 0.448573 + }, + { + "acc": 0.64192963, + "epoch": 0.5639269406392694, + "grad_norm": 5.375, + "learning_rate": 8.599669700292472e-06, + "loss": 1.62948952, + "memory(GiB)": 111.15, + "step": 22230, + "train_speed(iter/s)": 0.448627 + }, + { + "acc": 0.66415005, + "epoch": 0.5640537798072045, + "grad_norm": 5.34375, + "learning_rate": 8.598941831634707e-06, + "loss": 1.66025219, + "memory(GiB)": 111.15, + "step": 22235, + "train_speed(iter/s)": 0.44868 + }, + { + "acc": 0.63441296, + "epoch": 0.5641806189751395, + "grad_norm": 4.875, + "learning_rate": 8.598213804679412e-06, + "loss": 1.67426872, + "memory(GiB)": 111.15, + "step": 22240, + "train_speed(iter/s)": 0.448735 + }, + { + "acc": 0.63767633, + "epoch": 0.5643074581430746, + "grad_norm": 5.09375, + "learning_rate": 8.597485619458609e-06, + "loss": 1.6960762, + "memory(GiB)": 111.15, + "step": 22245, + "train_speed(iter/s)": 0.448789 + }, + { + "acc": 0.64811869, + "epoch": 0.5644342973110097, + "grad_norm": 5.28125, + "learning_rate": 8.596757276004327e-06, + "loss": 1.6553627, + "memory(GiB)": 111.15, + "step": 22250, + "train_speed(iter/s)": 0.448843 + }, + { + "acc": 0.66147666, + "epoch": 0.5645611364789447, + "grad_norm": 6.09375, + "learning_rate": 8.5960287743486e-06, + "loss": 1.71235046, + "memory(GiB)": 111.15, + "step": 22255, + "train_speed(iter/s)": 0.448897 + }, + { + "acc": 0.65288815, + "epoch": 0.5646879756468798, + "grad_norm": 5.0625, + "learning_rate": 8.595300114523473e-06, + "loss": 1.61886406, + "memory(GiB)": 111.15, + "step": 22260, + "train_speed(iter/s)": 0.448951 + }, + { + "acc": 0.64970703, + "epoch": 0.5648148148148148, + "grad_norm": 5.0, + "learning_rate": 8.594571296560997e-06, + "loss": 1.6917141, + "memory(GiB)": 111.15, + "step": 22265, + "train_speed(iter/s)": 0.449005 + }, + { + "acc": 0.63839378, + "epoch": 0.5649416539827499, + "grad_norm": 4.6875, + "learning_rate": 8.593842320493224e-06, + "loss": 1.7114809, + "memory(GiB)": 111.15, + "step": 22270, + "train_speed(iter/s)": 0.449059 + }, + { + "acc": 0.64240112, + "epoch": 0.565068493150685, + "grad_norm": 5.375, + "learning_rate": 8.593113186352222e-06, + "loss": 1.72098961, + "memory(GiB)": 111.15, + "step": 22275, + "train_speed(iter/s)": 0.449113 + }, + { + "acc": 0.62562671, + "epoch": 0.56519533231862, + "grad_norm": 5.625, + "learning_rate": 8.592383894170059e-06, + "loss": 1.70323963, + "memory(GiB)": 111.15, + "step": 22280, + "train_speed(iter/s)": 0.449168 + }, + { + "acc": 0.65794206, + "epoch": 0.5653221714865551, + "grad_norm": 4.40625, + "learning_rate": 8.591654443978815e-06, + "loss": 1.60898476, + "memory(GiB)": 111.15, + "step": 22285, + "train_speed(iter/s)": 0.449222 + }, + { + "acc": 0.64694977, + "epoch": 0.5654490106544902, + "grad_norm": 5.78125, + "learning_rate": 8.590924835810572e-06, + "loss": 1.64140549, + "memory(GiB)": 111.15, + "step": 22290, + "train_speed(iter/s)": 0.449276 + }, + { + "acc": 0.65422006, + "epoch": 0.5655758498224251, + "grad_norm": 5.6875, + "learning_rate": 8.590195069697423e-06, + "loss": 1.5948658, + "memory(GiB)": 111.15, + "step": 22295, + "train_speed(iter/s)": 0.449329 + }, + { + "acc": 0.63341599, + "epoch": 0.5657026889903602, + "grad_norm": 4.71875, + "learning_rate": 8.589465145671465e-06, + "loss": 1.65612602, + "memory(GiB)": 111.15, + "step": 22300, + "train_speed(iter/s)": 0.449383 + }, + { + "epoch": 0.5657026889903602, + "eval_acc": 0.6416592547783181, + "eval_loss": 1.6173070669174194, + "eval_runtime": 113.426, + "eval_samples_per_second": 56.16, + "eval_steps_per_second": 28.08, + "step": 22300 + }, + { + "acc": 0.64246044, + "epoch": 0.5658295281582952, + "grad_norm": 5.03125, + "learning_rate": 8.588735063764803e-06, + "loss": 1.6381649, + "memory(GiB)": 111.15, + "step": 22305, + "train_speed(iter/s)": 0.448371 + }, + { + "acc": 0.65128722, + "epoch": 0.5659563673262303, + "grad_norm": 4.78125, + "learning_rate": 8.588004824009552e-06, + "loss": 1.69805794, + "memory(GiB)": 111.15, + "step": 22310, + "train_speed(iter/s)": 0.448424 + }, + { + "acc": 0.64704413, + "epoch": 0.5660832064941654, + "grad_norm": 5.6875, + "learning_rate": 8.58727442643783e-06, + "loss": 1.68568592, + "memory(GiB)": 111.15, + "step": 22315, + "train_speed(iter/s)": 0.448477 + }, + { + "acc": 0.64012465, + "epoch": 0.5662100456621004, + "grad_norm": 5.875, + "learning_rate": 8.586543871081764e-06, + "loss": 1.64785061, + "memory(GiB)": 111.15, + "step": 22320, + "train_speed(iter/s)": 0.44853 + }, + { + "acc": 0.66402073, + "epoch": 0.5663368848300355, + "grad_norm": 5.03125, + "learning_rate": 8.585813157973482e-06, + "loss": 1.58720055, + "memory(GiB)": 111.15, + "step": 22325, + "train_speed(iter/s)": 0.448584 + }, + { + "acc": 0.63483982, + "epoch": 0.5664637239979706, + "grad_norm": 5.875, + "learning_rate": 8.58508228714513e-06, + "loss": 1.62102203, + "memory(GiB)": 111.15, + "step": 22330, + "train_speed(iter/s)": 0.448637 + }, + { + "acc": 0.65540876, + "epoch": 0.5665905631659056, + "grad_norm": 5.625, + "learning_rate": 8.584351258628852e-06, + "loss": 1.60909958, + "memory(GiB)": 111.15, + "step": 22335, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.67111187, + "epoch": 0.5667174023338407, + "grad_norm": 4.96875, + "learning_rate": 8.583620072456803e-06, + "loss": 1.4935689, + "memory(GiB)": 111.15, + "step": 22340, + "train_speed(iter/s)": 0.448744 + }, + { + "acc": 0.65512905, + "epoch": 0.5668442415017757, + "grad_norm": 5.15625, + "learning_rate": 8.582888728661142e-06, + "loss": 1.62110195, + "memory(GiB)": 111.15, + "step": 22345, + "train_speed(iter/s)": 0.448797 + }, + { + "acc": 0.63930717, + "epoch": 0.5669710806697108, + "grad_norm": 5.5, + "learning_rate": 8.582157227274042e-06, + "loss": 1.69590759, + "memory(GiB)": 111.15, + "step": 22350, + "train_speed(iter/s)": 0.44885 + }, + { + "acc": 0.65268154, + "epoch": 0.5670979198376459, + "grad_norm": 5.75, + "learning_rate": 8.581425568327671e-06, + "loss": 1.6368, + "memory(GiB)": 111.15, + "step": 22355, + "train_speed(iter/s)": 0.448904 + }, + { + "acc": 0.63447061, + "epoch": 0.5672247590055809, + "grad_norm": 5.1875, + "learning_rate": 8.580693751854215e-06, + "loss": 1.6950861, + "memory(GiB)": 111.15, + "step": 22360, + "train_speed(iter/s)": 0.448958 + }, + { + "acc": 0.64718428, + "epoch": 0.567351598173516, + "grad_norm": 4.875, + "learning_rate": 8.57996177788586e-06, + "loss": 1.62124672, + "memory(GiB)": 111.15, + "step": 22365, + "train_speed(iter/s)": 0.449011 + }, + { + "acc": 0.64119024, + "epoch": 0.5674784373414511, + "grad_norm": 6.6875, + "learning_rate": 8.579229646454803e-06, + "loss": 1.66817055, + "memory(GiB)": 111.15, + "step": 22370, + "train_speed(iter/s)": 0.449065 + }, + { + "acc": 0.66411924, + "epoch": 0.5676052765093861, + "grad_norm": 8.0, + "learning_rate": 8.578497357593246e-06, + "loss": 1.5580328, + "memory(GiB)": 111.15, + "step": 22375, + "train_speed(iter/s)": 0.449118 + }, + { + "acc": 0.66866922, + "epoch": 0.5677321156773212, + "grad_norm": 4.78125, + "learning_rate": 8.5777649113334e-06, + "loss": 1.51795311, + "memory(GiB)": 111.15, + "step": 22380, + "train_speed(iter/s)": 0.449171 + }, + { + "acc": 0.65471077, + "epoch": 0.5678589548452562, + "grad_norm": 5.21875, + "learning_rate": 8.577032307707476e-06, + "loss": 1.60868187, + "memory(GiB)": 111.15, + "step": 22385, + "train_speed(iter/s)": 0.449225 + }, + { + "acc": 0.63154197, + "epoch": 0.5679857940131913, + "grad_norm": 5.375, + "learning_rate": 8.576299546747704e-06, + "loss": 1.72208195, + "memory(GiB)": 111.15, + "step": 22390, + "train_speed(iter/s)": 0.449279 + }, + { + "acc": 0.65218801, + "epoch": 0.5681126331811264, + "grad_norm": 5.5, + "learning_rate": 8.575566628486309e-06, + "loss": 1.66326485, + "memory(GiB)": 111.15, + "step": 22395, + "train_speed(iter/s)": 0.449333 + }, + { + "acc": 0.63140669, + "epoch": 0.5682394723490614, + "grad_norm": 5.21875, + "learning_rate": 8.574833552955532e-06, + "loss": 1.68630066, + "memory(GiB)": 111.15, + "step": 22400, + "train_speed(iter/s)": 0.449386 + }, + { + "epoch": 0.5682394723490614, + "eval_acc": 0.6417323640574731, + "eval_loss": 1.6172221899032593, + "eval_runtime": 113.3766, + "eval_samples_per_second": 56.184, + "eval_steps_per_second": 28.092, + "step": 22400 + }, + { + "acc": 0.64658957, + "epoch": 0.5683663115169965, + "grad_norm": 4.75, + "learning_rate": 8.574100320187612e-06, + "loss": 1.67430744, + "memory(GiB)": 111.15, + "step": 22405, + "train_speed(iter/s)": 0.448379 + }, + { + "acc": 0.64948206, + "epoch": 0.5684931506849316, + "grad_norm": 5.46875, + "learning_rate": 8.573366930214807e-06, + "loss": 1.68807449, + "memory(GiB)": 111.15, + "step": 22410, + "train_speed(iter/s)": 0.448432 + }, + { + "acc": 0.62832184, + "epoch": 0.5686199898528665, + "grad_norm": 6.34375, + "learning_rate": 8.572633383069366e-06, + "loss": 1.70685921, + "memory(GiB)": 111.15, + "step": 22415, + "train_speed(iter/s)": 0.448485 + }, + { + "acc": 0.63756914, + "epoch": 0.5687468290208016, + "grad_norm": 7.625, + "learning_rate": 8.571899678783561e-06, + "loss": 1.68096962, + "memory(GiB)": 111.15, + "step": 22420, + "train_speed(iter/s)": 0.448539 + }, + { + "acc": 0.64998541, + "epoch": 0.5688736681887366, + "grad_norm": 5.21875, + "learning_rate": 8.57116581738966e-06, + "loss": 1.65005436, + "memory(GiB)": 111.15, + "step": 22425, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.66172194, + "epoch": 0.5690005073566717, + "grad_norm": 4.53125, + "learning_rate": 8.570431798919941e-06, + "loss": 1.5977953, + "memory(GiB)": 111.15, + "step": 22430, + "train_speed(iter/s)": 0.448646 + }, + { + "acc": 0.65763259, + "epoch": 0.5691273465246068, + "grad_norm": 6.03125, + "learning_rate": 8.569697623406692e-06, + "loss": 1.60664749, + "memory(GiB)": 111.15, + "step": 22435, + "train_speed(iter/s)": 0.448699 + }, + { + "acc": 0.6450788, + "epoch": 0.5692541856925418, + "grad_norm": 4.78125, + "learning_rate": 8.568963290882204e-06, + "loss": 1.63901882, + "memory(GiB)": 111.15, + "step": 22440, + "train_speed(iter/s)": 0.448751 + }, + { + "acc": 0.66275063, + "epoch": 0.5693810248604769, + "grad_norm": 5.125, + "learning_rate": 8.568228801378775e-06, + "loss": 1.57805367, + "memory(GiB)": 111.15, + "step": 22445, + "train_speed(iter/s)": 0.448754 + }, + { + "acc": 0.6560482, + "epoch": 0.569507864028412, + "grad_norm": 4.84375, + "learning_rate": 8.567494154928713e-06, + "loss": 1.62301979, + "memory(GiB)": 111.15, + "step": 22450, + "train_speed(iter/s)": 0.448807 + }, + { + "acc": 0.6308424, + "epoch": 0.569634703196347, + "grad_norm": 4.75, + "learning_rate": 8.566759351564332e-06, + "loss": 1.66219082, + "memory(GiB)": 111.15, + "step": 22455, + "train_speed(iter/s)": 0.44886 + }, + { + "acc": 0.63788114, + "epoch": 0.5697615423642821, + "grad_norm": 5.34375, + "learning_rate": 8.566024391317947e-06, + "loss": 1.68257294, + "memory(GiB)": 111.15, + "step": 22460, + "train_speed(iter/s)": 0.448913 + }, + { + "acc": 0.64462752, + "epoch": 0.5698883815322171, + "grad_norm": 6.25, + "learning_rate": 8.565289274221891e-06, + "loss": 1.6257555, + "memory(GiB)": 111.15, + "step": 22465, + "train_speed(iter/s)": 0.448966 + }, + { + "acc": 0.6583847, + "epoch": 0.5700152207001522, + "grad_norm": 4.8125, + "learning_rate": 8.564554000308493e-06, + "loss": 1.65703354, + "memory(GiB)": 111.15, + "step": 22470, + "train_speed(iter/s)": 0.449019 + }, + { + "acc": 0.64397178, + "epoch": 0.5701420598680873, + "grad_norm": 5.125, + "learning_rate": 8.563818569610096e-06, + "loss": 1.60081577, + "memory(GiB)": 111.15, + "step": 22475, + "train_speed(iter/s)": 0.449072 + }, + { + "acc": 0.64409456, + "epoch": 0.5702688990360223, + "grad_norm": 4.59375, + "learning_rate": 8.563082982159048e-06, + "loss": 1.61242104, + "memory(GiB)": 111.15, + "step": 22480, + "train_speed(iter/s)": 0.449124 + }, + { + "acc": 0.6533731, + "epoch": 0.5703957382039574, + "grad_norm": 5.625, + "learning_rate": 8.562347237987701e-06, + "loss": 1.65443459, + "memory(GiB)": 111.15, + "step": 22485, + "train_speed(iter/s)": 0.449178 + }, + { + "acc": 0.63084102, + "epoch": 0.5705225773718925, + "grad_norm": 5.875, + "learning_rate": 8.561611337128418e-06, + "loss": 1.69784279, + "memory(GiB)": 111.15, + "step": 22490, + "train_speed(iter/s)": 0.449231 + }, + { + "acc": 0.65532846, + "epoch": 0.5706494165398275, + "grad_norm": 5.4375, + "learning_rate": 8.560875279613568e-06, + "loss": 1.5812603, + "memory(GiB)": 111.15, + "step": 22495, + "train_speed(iter/s)": 0.449283 + }, + { + "acc": 0.63838091, + "epoch": 0.5707762557077626, + "grad_norm": 4.9375, + "learning_rate": 8.560139065475523e-06, + "loss": 1.6905674, + "memory(GiB)": 111.15, + "step": 22500, + "train_speed(iter/s)": 0.449336 + }, + { + "epoch": 0.5707762557077626, + "eval_acc": 0.6416266689281803, + "eval_loss": 1.6171903610229492, + "eval_runtime": 113.2988, + "eval_samples_per_second": 56.223, + "eval_steps_per_second": 28.112, + "step": 22500 + }, + { + "acc": 0.66312513, + "epoch": 0.5709030948756976, + "grad_norm": 6.21875, + "learning_rate": 8.559402694746671e-06, + "loss": 1.63613052, + "memory(GiB)": 111.15, + "step": 22505, + "train_speed(iter/s)": 0.448334 + }, + { + "acc": 0.65852509, + "epoch": 0.5710299340436327, + "grad_norm": 4.6875, + "learning_rate": 8.558666167459393e-06, + "loss": 1.56096802, + "memory(GiB)": 111.15, + "step": 22510, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.64104824, + "epoch": 0.5711567732115678, + "grad_norm": 4.75, + "learning_rate": 8.55792948364609e-06, + "loss": 1.76535168, + "memory(GiB)": 111.15, + "step": 22515, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.65806174, + "epoch": 0.5712836123795028, + "grad_norm": 5.53125, + "learning_rate": 8.557192643339164e-06, + "loss": 1.68294926, + "memory(GiB)": 111.15, + "step": 22520, + "train_speed(iter/s)": 0.448492 + }, + { + "acc": 0.66040201, + "epoch": 0.5714104515474379, + "grad_norm": 4.46875, + "learning_rate": 8.556455646571022e-06, + "loss": 1.63673515, + "memory(GiB)": 111.15, + "step": 22525, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.65729427, + "epoch": 0.571537290715373, + "grad_norm": 5.28125, + "learning_rate": 8.555718493374084e-06, + "loss": 1.66201191, + "memory(GiB)": 111.15, + "step": 22530, + "train_speed(iter/s)": 0.448597 + }, + { + "acc": 0.65802045, + "epoch": 0.571664129883308, + "grad_norm": 5.3125, + "learning_rate": 8.55498118378077e-06, + "loss": 1.63256683, + "memory(GiB)": 111.15, + "step": 22535, + "train_speed(iter/s)": 0.44865 + }, + { + "acc": 0.66980476, + "epoch": 0.571790969051243, + "grad_norm": 7.15625, + "learning_rate": 8.554243717823512e-06, + "loss": 1.55046024, + "memory(GiB)": 111.15, + "step": 22540, + "train_speed(iter/s)": 0.448703 + }, + { + "acc": 0.6528841, + "epoch": 0.571917808219178, + "grad_norm": 5.0625, + "learning_rate": 8.553506095534747e-06, + "loss": 1.65310478, + "memory(GiB)": 111.15, + "step": 22545, + "train_speed(iter/s)": 0.448756 + }, + { + "acc": 0.64879627, + "epoch": 0.5720446473871131, + "grad_norm": 5.34375, + "learning_rate": 8.55276831694692e-06, + "loss": 1.6077652, + "memory(GiB)": 111.15, + "step": 22550, + "train_speed(iter/s)": 0.448809 + }, + { + "acc": 0.65907993, + "epoch": 0.5721714865550482, + "grad_norm": 5.09375, + "learning_rate": 8.552030382092477e-06, + "loss": 1.65524654, + "memory(GiB)": 111.15, + "step": 22555, + "train_speed(iter/s)": 0.448861 + }, + { + "acc": 0.65631819, + "epoch": 0.5722983257229832, + "grad_norm": 5.3125, + "learning_rate": 8.551292291003884e-06, + "loss": 1.63572521, + "memory(GiB)": 111.15, + "step": 22560, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.64630833, + "epoch": 0.5724251648909183, + "grad_norm": 6.1875, + "learning_rate": 8.550554043713597e-06, + "loss": 1.66465282, + "memory(GiB)": 111.15, + "step": 22565, + "train_speed(iter/s)": 0.448967 + }, + { + "acc": 0.65757027, + "epoch": 0.5725520040588534, + "grad_norm": 4.8125, + "learning_rate": 8.549815640254092e-06, + "loss": 1.58013744, + "memory(GiB)": 111.15, + "step": 22570, + "train_speed(iter/s)": 0.44902 + }, + { + "acc": 0.65324035, + "epoch": 0.5726788432267884, + "grad_norm": 5.71875, + "learning_rate": 8.549077080657846e-06, + "loss": 1.61754456, + "memory(GiB)": 111.15, + "step": 22575, + "train_speed(iter/s)": 0.449073 + }, + { + "acc": 0.63606825, + "epoch": 0.5728056823947235, + "grad_norm": 6.09375, + "learning_rate": 8.548338364957345e-06, + "loss": 1.73468494, + "memory(GiB)": 111.15, + "step": 22580, + "train_speed(iter/s)": 0.449125 + }, + { + "acc": 0.64174061, + "epoch": 0.5729325215626585, + "grad_norm": 5.34375, + "learning_rate": 8.54759949318508e-06, + "loss": 1.67128468, + "memory(GiB)": 111.15, + "step": 22585, + "train_speed(iter/s)": 0.449178 + }, + { + "acc": 0.64967022, + "epoch": 0.5730593607305936, + "grad_norm": 4.5, + "learning_rate": 8.546860465373552e-06, + "loss": 1.60665512, + "memory(GiB)": 111.15, + "step": 22590, + "train_speed(iter/s)": 0.44923 + }, + { + "acc": 0.65209365, + "epoch": 0.5731861998985287, + "grad_norm": 5.375, + "learning_rate": 8.546121281555265e-06, + "loss": 1.60904064, + "memory(GiB)": 111.15, + "step": 22595, + "train_speed(iter/s)": 0.449283 + }, + { + "acc": 0.65247617, + "epoch": 0.5733130390664637, + "grad_norm": 5.3125, + "learning_rate": 8.54538194176273e-06, + "loss": 1.58401756, + "memory(GiB)": 111.15, + "step": 22600, + "train_speed(iter/s)": 0.449335 + }, + { + "epoch": 0.5733130390664637, + "eval_acc": 0.6416859918861233, + "eval_loss": 1.6170921325683594, + "eval_runtime": 113.5197, + "eval_samples_per_second": 56.114, + "eval_steps_per_second": 28.057, + "step": 22600 + }, + { + "acc": 0.64860497, + "epoch": 0.5734398782343988, + "grad_norm": 5.5, + "learning_rate": 8.544642446028469e-06, + "loss": 1.62943077, + "memory(GiB)": 111.15, + "step": 22605, + "train_speed(iter/s)": 0.448336 + }, + { + "acc": 0.62641735, + "epoch": 0.5735667174023339, + "grad_norm": 5.90625, + "learning_rate": 8.543902794385008e-06, + "loss": 1.75010757, + "memory(GiB)": 111.15, + "step": 22610, + "train_speed(iter/s)": 0.448388 + }, + { + "acc": 0.63823576, + "epoch": 0.5736935565702689, + "grad_norm": 5.34375, + "learning_rate": 8.543162986864879e-06, + "loss": 1.70020714, + "memory(GiB)": 111.15, + "step": 22615, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.66741915, + "epoch": 0.573820395738204, + "grad_norm": 5.375, + "learning_rate": 8.542423023500623e-06, + "loss": 1.55416565, + "memory(GiB)": 111.15, + "step": 22620, + "train_speed(iter/s)": 0.448492 + }, + { + "acc": 0.64599051, + "epoch": 0.573947234906139, + "grad_norm": 5.84375, + "learning_rate": 8.541682904324786e-06, + "loss": 1.65907574, + "memory(GiB)": 111.15, + "step": 22625, + "train_speed(iter/s)": 0.448544 + }, + { + "acc": 0.64371023, + "epoch": 0.5740740740740741, + "grad_norm": 5.3125, + "learning_rate": 8.540942629369923e-06, + "loss": 1.66133575, + "memory(GiB)": 111.15, + "step": 22630, + "train_speed(iter/s)": 0.448597 + }, + { + "acc": 0.62952118, + "epoch": 0.5742009132420092, + "grad_norm": 5.125, + "learning_rate": 8.540202198668595e-06, + "loss": 1.65495567, + "memory(GiB)": 111.15, + "step": 22635, + "train_speed(iter/s)": 0.448649 + }, + { + "acc": 0.65301991, + "epoch": 0.5743277524099442, + "grad_norm": 4.9375, + "learning_rate": 8.539461612253368e-06, + "loss": 1.60597878, + "memory(GiB)": 111.15, + "step": 22640, + "train_speed(iter/s)": 0.448701 + }, + { + "acc": 0.65432053, + "epoch": 0.5744545915778793, + "grad_norm": 5.53125, + "learning_rate": 8.538720870156816e-06, + "loss": 1.61701031, + "memory(GiB)": 111.15, + "step": 22645, + "train_speed(iter/s)": 0.448754 + }, + { + "acc": 0.6460824, + "epoch": 0.5745814307458144, + "grad_norm": 5.0625, + "learning_rate": 8.53797997241152e-06, + "loss": 1.58248577, + "memory(GiB)": 111.15, + "step": 22650, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.65653992, + "epoch": 0.5747082699137493, + "grad_norm": 6.375, + "learning_rate": 8.537238919050071e-06, + "loss": 1.62921295, + "memory(GiB)": 111.15, + "step": 22655, + "train_speed(iter/s)": 0.448859 + }, + { + "acc": 0.66215358, + "epoch": 0.5748351090816844, + "grad_norm": 5.21875, + "learning_rate": 8.53649771010506e-06, + "loss": 1.58402061, + "memory(GiB)": 111.15, + "step": 22660, + "train_speed(iter/s)": 0.448912 + }, + { + "acc": 0.64714689, + "epoch": 0.5749619482496194, + "grad_norm": 5.1875, + "learning_rate": 8.535756345609092e-06, + "loss": 1.62751808, + "memory(GiB)": 111.15, + "step": 22665, + "train_speed(iter/s)": 0.448965 + }, + { + "acc": 0.65618377, + "epoch": 0.5750887874175545, + "grad_norm": 5.96875, + "learning_rate": 8.535014825594772e-06, + "loss": 1.57740288, + "memory(GiB)": 111.15, + "step": 22670, + "train_speed(iter/s)": 0.449016 + }, + { + "acc": 0.65600424, + "epoch": 0.5752156265854896, + "grad_norm": 4.78125, + "learning_rate": 8.534273150094718e-06, + "loss": 1.6085495, + "memory(GiB)": 111.15, + "step": 22675, + "train_speed(iter/s)": 0.449068 + }, + { + "acc": 0.64013004, + "epoch": 0.5753424657534246, + "grad_norm": 7.0625, + "learning_rate": 8.533531319141552e-06, + "loss": 1.59036884, + "memory(GiB)": 111.15, + "step": 22680, + "train_speed(iter/s)": 0.449121 + }, + { + "acc": 0.63325386, + "epoch": 0.5754693049213597, + "grad_norm": 5.96875, + "learning_rate": 8.532789332767902e-06, + "loss": 1.70432053, + "memory(GiB)": 111.15, + "step": 22685, + "train_speed(iter/s)": 0.449173 + }, + { + "acc": 0.64312506, + "epoch": 0.5755961440892948, + "grad_norm": 5.5, + "learning_rate": 8.532047191006405e-06, + "loss": 1.60400963, + "memory(GiB)": 111.15, + "step": 22690, + "train_speed(iter/s)": 0.449226 + }, + { + "acc": 0.6396183, + "epoch": 0.5757229832572298, + "grad_norm": 4.1875, + "learning_rate": 8.531304893889702e-06, + "loss": 1.63361092, + "memory(GiB)": 111.15, + "step": 22695, + "train_speed(iter/s)": 0.449278 + }, + { + "acc": 0.64858456, + "epoch": 0.5758498224251649, + "grad_norm": 4.96875, + "learning_rate": 8.530562441450445e-06, + "loss": 1.67291737, + "memory(GiB)": 111.15, + "step": 22700, + "train_speed(iter/s)": 0.449331 + }, + { + "epoch": 0.5758498224251649, + "eval_acc": 0.641664685753341, + "eval_loss": 1.6171623468399048, + "eval_runtime": 113.3625, + "eval_samples_per_second": 56.191, + "eval_steps_per_second": 28.096, + "step": 22700 + }, + { + "acc": 0.64676156, + "epoch": 0.5759766615930999, + "grad_norm": 5.375, + "learning_rate": 8.529819833721289e-06, + "loss": 1.66137619, + "memory(GiB)": 111.15, + "step": 22705, + "train_speed(iter/s)": 0.448337 + }, + { + "acc": 0.6483366, + "epoch": 0.576103500761035, + "grad_norm": 5.6875, + "learning_rate": 8.529077070734896e-06, + "loss": 1.6777647, + "memory(GiB)": 111.15, + "step": 22710, + "train_speed(iter/s)": 0.44839 + }, + { + "acc": 0.64880776, + "epoch": 0.5762303399289701, + "grad_norm": 5.59375, + "learning_rate": 8.528334152523938e-06, + "loss": 1.71902847, + "memory(GiB)": 111.15, + "step": 22715, + "train_speed(iter/s)": 0.448442 + }, + { + "acc": 0.66875343, + "epoch": 0.5763571790969051, + "grad_norm": 5.5, + "learning_rate": 8.52759107912109e-06, + "loss": 1.55564499, + "memory(GiB)": 111.15, + "step": 22720, + "train_speed(iter/s)": 0.448494 + }, + { + "acc": 0.65794525, + "epoch": 0.5764840182648402, + "grad_norm": 5.15625, + "learning_rate": 8.526847850559037e-06, + "loss": 1.61092987, + "memory(GiB)": 111.15, + "step": 22725, + "train_speed(iter/s)": 0.448547 + }, + { + "acc": 0.64561043, + "epoch": 0.5766108574327753, + "grad_norm": 4.53125, + "learning_rate": 8.526104466870472e-06, + "loss": 1.66713066, + "memory(GiB)": 111.15, + "step": 22730, + "train_speed(iter/s)": 0.448599 + }, + { + "acc": 0.65492029, + "epoch": 0.5767376966007103, + "grad_norm": 4.84375, + "learning_rate": 8.525360928088087e-06, + "loss": 1.59742489, + "memory(GiB)": 111.15, + "step": 22735, + "train_speed(iter/s)": 0.448651 + }, + { + "acc": 0.65408688, + "epoch": 0.5768645357686454, + "grad_norm": 5.0, + "learning_rate": 8.524617234244588e-06, + "loss": 1.62485962, + "memory(GiB)": 111.15, + "step": 22740, + "train_speed(iter/s)": 0.448703 + }, + { + "acc": 0.65393391, + "epoch": 0.5769913749365804, + "grad_norm": 5.4375, + "learning_rate": 8.523873385372687e-06, + "loss": 1.51788921, + "memory(GiB)": 111.15, + "step": 22745, + "train_speed(iter/s)": 0.448756 + }, + { + "acc": 0.64614258, + "epoch": 0.5771182141045155, + "grad_norm": 5.53125, + "learning_rate": 8.523129381505104e-06, + "loss": 1.68877754, + "memory(GiB)": 111.15, + "step": 22750, + "train_speed(iter/s)": 0.448808 + }, + { + "acc": 0.66538067, + "epoch": 0.5772450532724506, + "grad_norm": 5.46875, + "learning_rate": 8.522385222674559e-06, + "loss": 1.58131733, + "memory(GiB)": 111.15, + "step": 22755, + "train_speed(iter/s)": 0.448861 + }, + { + "acc": 0.65494986, + "epoch": 0.5773718924403856, + "grad_norm": 5.09375, + "learning_rate": 8.521640908913787e-06, + "loss": 1.59285774, + "memory(GiB)": 111.15, + "step": 22760, + "train_speed(iter/s)": 0.448913 + }, + { + "acc": 0.65601053, + "epoch": 0.5774987316083207, + "grad_norm": 4.75, + "learning_rate": 8.520896440255524e-06, + "loss": 1.63625393, + "memory(GiB)": 111.15, + "step": 22765, + "train_speed(iter/s)": 0.448966 + }, + { + "acc": 0.66091948, + "epoch": 0.5776255707762558, + "grad_norm": 5.84375, + "learning_rate": 8.520151816732517e-06, + "loss": 1.61985817, + "memory(GiB)": 111.15, + "step": 22770, + "train_speed(iter/s)": 0.449019 + }, + { + "acc": 0.63584509, + "epoch": 0.5777524099441907, + "grad_norm": 5.84375, + "learning_rate": 8.519407038377515e-06, + "loss": 1.70364723, + "memory(GiB)": 111.15, + "step": 22775, + "train_speed(iter/s)": 0.449072 + }, + { + "acc": 0.64105492, + "epoch": 0.5778792491121258, + "grad_norm": 4.875, + "learning_rate": 8.518662105223279e-06, + "loss": 1.67169838, + "memory(GiB)": 111.15, + "step": 22780, + "train_speed(iter/s)": 0.449124 + }, + { + "acc": 0.64951859, + "epoch": 0.5780060882800608, + "grad_norm": 5.53125, + "learning_rate": 8.517917017302574e-06, + "loss": 1.65342693, + "memory(GiB)": 111.15, + "step": 22785, + "train_speed(iter/s)": 0.449177 + }, + { + "acc": 0.65126319, + "epoch": 0.5781329274479959, + "grad_norm": 4.90625, + "learning_rate": 8.517171774648172e-06, + "loss": 1.68535767, + "memory(GiB)": 111.15, + "step": 22790, + "train_speed(iter/s)": 0.449229 + }, + { + "acc": 0.64107819, + "epoch": 0.578259766615931, + "grad_norm": 5.0, + "learning_rate": 8.516426377292854e-06, + "loss": 1.65018005, + "memory(GiB)": 111.15, + "step": 22795, + "train_speed(iter/s)": 0.449282 + }, + { + "acc": 0.65371814, + "epoch": 0.578386605783866, + "grad_norm": 4.90625, + "learning_rate": 8.515680825269404e-06, + "loss": 1.58780766, + "memory(GiB)": 111.15, + "step": 22800, + "train_speed(iter/s)": 0.449334 + }, + { + "epoch": 0.578386605783866, + "eval_acc": 0.641727768617069, + "eval_loss": 1.6169254779815674, + "eval_runtime": 113.9417, + "eval_samples_per_second": 55.906, + "eval_steps_per_second": 27.953, + "step": 22800 + }, + { + "acc": 0.65021534, + "epoch": 0.5785134449518011, + "grad_norm": 5.125, + "learning_rate": 8.514935118610613e-06, + "loss": 1.70228157, + "memory(GiB)": 111.15, + "step": 22805, + "train_speed(iter/s)": 0.44834 + }, + { + "acc": 0.64636803, + "epoch": 0.5786402841197362, + "grad_norm": 4.9375, + "learning_rate": 8.514189257349283e-06, + "loss": 1.65145149, + "memory(GiB)": 111.15, + "step": 22810, + "train_speed(iter/s)": 0.448391 + }, + { + "acc": 0.65868373, + "epoch": 0.5787671232876712, + "grad_norm": 5.21875, + "learning_rate": 8.51344324151822e-06, + "loss": 1.59472599, + "memory(GiB)": 111.15, + "step": 22815, + "train_speed(iter/s)": 0.448444 + }, + { + "acc": 0.64171066, + "epoch": 0.5788939624556063, + "grad_norm": 5.40625, + "learning_rate": 8.512697071150235e-06, + "loss": 1.63051186, + "memory(GiB)": 111.15, + "step": 22820, + "train_speed(iter/s)": 0.448496 + }, + { + "acc": 0.65031738, + "epoch": 0.5790208016235413, + "grad_norm": 5.46875, + "learning_rate": 8.511950746278152e-06, + "loss": 1.51984921, + "memory(GiB)": 111.15, + "step": 22825, + "train_speed(iter/s)": 0.448548 + }, + { + "acc": 0.65141535, + "epoch": 0.5791476407914764, + "grad_norm": 5.5, + "learning_rate": 8.511204266934797e-06, + "loss": 1.6476284, + "memory(GiB)": 111.15, + "step": 22830, + "train_speed(iter/s)": 0.4486 + }, + { + "acc": 0.63654189, + "epoch": 0.5792744799594115, + "grad_norm": 4.875, + "learning_rate": 8.510457633152998e-06, + "loss": 1.72171936, + "memory(GiB)": 111.15, + "step": 22835, + "train_speed(iter/s)": 0.448652 + }, + { + "acc": 0.6446506, + "epoch": 0.5794013191273465, + "grad_norm": 4.6875, + "learning_rate": 8.509710844965602e-06, + "loss": 1.62229767, + "memory(GiB)": 111.15, + "step": 22840, + "train_speed(iter/s)": 0.448704 + }, + { + "acc": 0.64561481, + "epoch": 0.5795281582952816, + "grad_norm": 6.125, + "learning_rate": 8.508963902405451e-06, + "loss": 1.7096138, + "memory(GiB)": 111.15, + "step": 22845, + "train_speed(iter/s)": 0.448756 + }, + { + "acc": 0.65416088, + "epoch": 0.5796549974632167, + "grad_norm": 5.625, + "learning_rate": 8.508216805505403e-06, + "loss": 1.57666492, + "memory(GiB)": 111.15, + "step": 22850, + "train_speed(iter/s)": 0.448808 + }, + { + "acc": 0.65341306, + "epoch": 0.5797818366311517, + "grad_norm": 4.46875, + "learning_rate": 8.507469554298318e-06, + "loss": 1.64896297, + "memory(GiB)": 111.15, + "step": 22855, + "train_speed(iter/s)": 0.44886 + }, + { + "acc": 0.6590559, + "epoch": 0.5799086757990868, + "grad_norm": 5.1875, + "learning_rate": 8.506722148817061e-06, + "loss": 1.60996857, + "memory(GiB)": 111.15, + "step": 22860, + "train_speed(iter/s)": 0.448913 + }, + { + "acc": 0.65924034, + "epoch": 0.5800355149670218, + "grad_norm": 4.96875, + "learning_rate": 8.505974589094505e-06, + "loss": 1.5777216, + "memory(GiB)": 111.15, + "step": 22865, + "train_speed(iter/s)": 0.448965 + }, + { + "acc": 0.66034298, + "epoch": 0.5801623541349569, + "grad_norm": 4.59375, + "learning_rate": 8.505226875163537e-06, + "loss": 1.65821972, + "memory(GiB)": 111.15, + "step": 22870, + "train_speed(iter/s)": 0.449017 + }, + { + "acc": 0.65001822, + "epoch": 0.580289193302892, + "grad_norm": 6.09375, + "learning_rate": 8.50447900705704e-06, + "loss": 1.62131176, + "memory(GiB)": 111.15, + "step": 22875, + "train_speed(iter/s)": 0.44907 + }, + { + "acc": 0.6546391, + "epoch": 0.580416032470827, + "grad_norm": 3.953125, + "learning_rate": 8.503730984807911e-06, + "loss": 1.55963974, + "memory(GiB)": 111.15, + "step": 22880, + "train_speed(iter/s)": 0.449121 + }, + { + "acc": 0.66047559, + "epoch": 0.5805428716387621, + "grad_norm": 5.125, + "learning_rate": 8.502982808449049e-06, + "loss": 1.59342995, + "memory(GiB)": 111.15, + "step": 22885, + "train_speed(iter/s)": 0.449173 + }, + { + "acc": 0.6405695, + "epoch": 0.5806697108066972, + "grad_norm": 4.84375, + "learning_rate": 8.502234478013363e-06, + "loss": 1.70039139, + "memory(GiB)": 111.15, + "step": 22890, + "train_speed(iter/s)": 0.449224 + }, + { + "acc": 0.64677019, + "epoch": 0.5807965499746321, + "grad_norm": 5.625, + "learning_rate": 8.501485993533769e-06, + "loss": 1.65048904, + "memory(GiB)": 111.15, + "step": 22895, + "train_speed(iter/s)": 0.449276 + }, + { + "acc": 0.64078865, + "epoch": 0.5809233891425672, + "grad_norm": 5.65625, + "learning_rate": 8.500737355043188e-06, + "loss": 1.66476669, + "memory(GiB)": 111.15, + "step": 22900, + "train_speed(iter/s)": 0.449329 + }, + { + "epoch": 0.5809233891425672, + "eval_acc": 0.6417674565114675, + "eval_loss": 1.616979956626892, + "eval_runtime": 113.1979, + "eval_samples_per_second": 56.273, + "eval_steps_per_second": 28.137, + "step": 22900 + }, + { + "acc": 0.65174408, + "epoch": 0.5810502283105022, + "grad_norm": 5.09375, + "learning_rate": 8.499988562574549e-06, + "loss": 1.65789757, + "memory(GiB)": 111.15, + "step": 22905, + "train_speed(iter/s)": 0.448344 + }, + { + "acc": 0.65864687, + "epoch": 0.5811770674784373, + "grad_norm": 6.84375, + "learning_rate": 8.499239616160787e-06, + "loss": 1.67414875, + "memory(GiB)": 111.15, + "step": 22910, + "train_speed(iter/s)": 0.448395 + }, + { + "acc": 0.65856905, + "epoch": 0.5813039066463724, + "grad_norm": 5.71875, + "learning_rate": 8.498490515834841e-06, + "loss": 1.65093994, + "memory(GiB)": 111.15, + "step": 22915, + "train_speed(iter/s)": 0.448447 + }, + { + "acc": 0.64424419, + "epoch": 0.5814307458143074, + "grad_norm": 4.65625, + "learning_rate": 8.497741261629664e-06, + "loss": 1.66695518, + "memory(GiB)": 111.15, + "step": 22920, + "train_speed(iter/s)": 0.448498 + }, + { + "acc": 0.63767805, + "epoch": 0.5815575849822425, + "grad_norm": 5.0, + "learning_rate": 8.496991853578212e-06, + "loss": 1.61589241, + "memory(GiB)": 111.15, + "step": 22925, + "train_speed(iter/s)": 0.448549 + }, + { + "acc": 0.64332085, + "epoch": 0.5816844241501776, + "grad_norm": 5.125, + "learning_rate": 8.496242291713444e-06, + "loss": 1.64475555, + "memory(GiB)": 111.15, + "step": 22930, + "train_speed(iter/s)": 0.448601 + }, + { + "acc": 0.65772643, + "epoch": 0.5818112633181126, + "grad_norm": 5.9375, + "learning_rate": 8.495492576068329e-06, + "loss": 1.63704109, + "memory(GiB)": 111.15, + "step": 22935, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.64648299, + "epoch": 0.5819381024860477, + "grad_norm": 4.71875, + "learning_rate": 8.494742706675844e-06, + "loss": 1.6729866, + "memory(GiB)": 111.15, + "step": 22940, + "train_speed(iter/s)": 0.448705 + }, + { + "acc": 0.63135757, + "epoch": 0.5820649416539827, + "grad_norm": 5.46875, + "learning_rate": 8.493992683568975e-06, + "loss": 1.72170258, + "memory(GiB)": 111.15, + "step": 22945, + "train_speed(iter/s)": 0.448756 + }, + { + "acc": 0.64865956, + "epoch": 0.5821917808219178, + "grad_norm": 7.25, + "learning_rate": 8.493242506780705e-06, + "loss": 1.62676964, + "memory(GiB)": 111.15, + "step": 22950, + "train_speed(iter/s)": 0.448807 + }, + { + "acc": 0.63747301, + "epoch": 0.5823186199898529, + "grad_norm": 6.84375, + "learning_rate": 8.492492176344035e-06, + "loss": 1.71799946, + "memory(GiB)": 111.15, + "step": 22955, + "train_speed(iter/s)": 0.448859 + }, + { + "acc": 0.6577549, + "epoch": 0.5824454591577879, + "grad_norm": 4.5, + "learning_rate": 8.491741692291967e-06, + "loss": 1.63643093, + "memory(GiB)": 111.15, + "step": 22960, + "train_speed(iter/s)": 0.44891 + }, + { + "acc": 0.63073153, + "epoch": 0.582572298325723, + "grad_norm": 6.0625, + "learning_rate": 8.490991054657507e-06, + "loss": 1.71229973, + "memory(GiB)": 111.15, + "step": 22965, + "train_speed(iter/s)": 0.448962 + }, + { + "acc": 0.64680843, + "epoch": 0.5826991374936581, + "grad_norm": 5.59375, + "learning_rate": 8.490240263473677e-06, + "loss": 1.68589077, + "memory(GiB)": 111.15, + "step": 22970, + "train_speed(iter/s)": 0.449013 + }, + { + "acc": 0.64029617, + "epoch": 0.5828259766615931, + "grad_norm": 7.03125, + "learning_rate": 8.489489318773496e-06, + "loss": 1.69028931, + "memory(GiB)": 111.15, + "step": 22975, + "train_speed(iter/s)": 0.449065 + }, + { + "acc": 0.64373875, + "epoch": 0.5829528158295282, + "grad_norm": 5.875, + "learning_rate": 8.488738220589996e-06, + "loss": 1.70428772, + "memory(GiB)": 111.15, + "step": 22980, + "train_speed(iter/s)": 0.449117 + }, + { + "acc": 0.65748262, + "epoch": 0.5830796549974632, + "grad_norm": 4.65625, + "learning_rate": 8.487986968956212e-06, + "loss": 1.6044136, + "memory(GiB)": 111.15, + "step": 22985, + "train_speed(iter/s)": 0.449167 + }, + { + "acc": 0.65411463, + "epoch": 0.5832064941653983, + "grad_norm": 4.9375, + "learning_rate": 8.487235563905191e-06, + "loss": 1.62674599, + "memory(GiB)": 111.15, + "step": 22990, + "train_speed(iter/s)": 0.449219 + }, + { + "acc": 0.6492919, + "epoch": 0.5833333333333334, + "grad_norm": 5.59375, + "learning_rate": 8.486484005469977e-06, + "loss": 1.61151657, + "memory(GiB)": 111.15, + "step": 22995, + "train_speed(iter/s)": 0.44927 + }, + { + "acc": 0.65252037, + "epoch": 0.5834601725012684, + "grad_norm": 5.09375, + "learning_rate": 8.485732293683633e-06, + "loss": 1.61574993, + "memory(GiB)": 111.15, + "step": 23000, + "train_speed(iter/s)": 0.449321 + }, + { + "epoch": 0.5834601725012684, + "eval_acc": 0.6418175885886024, + "eval_loss": 1.6167993545532227, + "eval_runtime": 113.4155, + "eval_samples_per_second": 56.165, + "eval_steps_per_second": 28.083, + "step": 23000 + }, + { + "acc": 0.64447079, + "epoch": 0.5835870116692035, + "grad_norm": 6.0625, + "learning_rate": 8.48498042857922e-06, + "loss": 1.65015812, + "memory(GiB)": 111.15, + "step": 23005, + "train_speed(iter/s)": 0.448339 + }, + { + "acc": 0.64177294, + "epoch": 0.5837138508371386, + "grad_norm": 5.1875, + "learning_rate": 8.484228410189807e-06, + "loss": 1.64903984, + "memory(GiB)": 111.15, + "step": 23010, + "train_speed(iter/s)": 0.44839 + }, + { + "acc": 0.63060074, + "epoch": 0.5838406900050735, + "grad_norm": 5.28125, + "learning_rate": 8.483476238548473e-06, + "loss": 1.7434576, + "memory(GiB)": 111.15, + "step": 23015, + "train_speed(iter/s)": 0.448441 + }, + { + "acc": 0.64391851, + "epoch": 0.5839675291730086, + "grad_norm": 4.6875, + "learning_rate": 8.482723913688301e-06, + "loss": 1.68703537, + "memory(GiB)": 111.15, + "step": 23020, + "train_speed(iter/s)": 0.448492 + }, + { + "acc": 0.65755997, + "epoch": 0.5840943683409436, + "grad_norm": 6.0, + "learning_rate": 8.481971435642382e-06, + "loss": 1.62821465, + "memory(GiB)": 111.15, + "step": 23025, + "train_speed(iter/s)": 0.448543 + }, + { + "acc": 0.65561604, + "epoch": 0.5842212075088787, + "grad_norm": 4.96875, + "learning_rate": 8.481218804443814e-06, + "loss": 1.56975269, + "memory(GiB)": 111.15, + "step": 23030, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.64319868, + "epoch": 0.5843480466768138, + "grad_norm": 5.125, + "learning_rate": 8.480466020125701e-06, + "loss": 1.66576729, + "memory(GiB)": 111.15, + "step": 23035, + "train_speed(iter/s)": 0.448645 + }, + { + "acc": 0.64515586, + "epoch": 0.5844748858447488, + "grad_norm": 4.53125, + "learning_rate": 8.479713082721153e-06, + "loss": 1.66706657, + "memory(GiB)": 111.15, + "step": 23040, + "train_speed(iter/s)": 0.448697 + }, + { + "acc": 0.66678553, + "epoch": 0.5846017250126839, + "grad_norm": 5.5625, + "learning_rate": 8.478959992263288e-06, + "loss": 1.57800655, + "memory(GiB)": 111.15, + "step": 23045, + "train_speed(iter/s)": 0.448748 + }, + { + "acc": 0.64039135, + "epoch": 0.584728564180619, + "grad_norm": 5.375, + "learning_rate": 8.478206748785229e-06, + "loss": 1.69505196, + "memory(GiB)": 111.15, + "step": 23050, + "train_speed(iter/s)": 0.4488 + }, + { + "acc": 0.65463028, + "epoch": 0.584855403348554, + "grad_norm": 4.625, + "learning_rate": 8.477453352320108e-06, + "loss": 1.60908527, + "memory(GiB)": 111.15, + "step": 23055, + "train_speed(iter/s)": 0.448851 + }, + { + "acc": 0.65811958, + "epoch": 0.5849822425164891, + "grad_norm": 4.625, + "learning_rate": 8.476699802901066e-06, + "loss": 1.62840118, + "memory(GiB)": 111.15, + "step": 23060, + "train_speed(iter/s)": 0.448902 + }, + { + "acc": 0.65275822, + "epoch": 0.5851090816844241, + "grad_norm": 5.09375, + "learning_rate": 8.47594610056124e-06, + "loss": 1.583358, + "memory(GiB)": 111.15, + "step": 23065, + "train_speed(iter/s)": 0.448953 + }, + { + "acc": 0.6436883, + "epoch": 0.5852359208523592, + "grad_norm": 5.46875, + "learning_rate": 8.475192245333787e-06, + "loss": 1.64192963, + "memory(GiB)": 111.15, + "step": 23070, + "train_speed(iter/s)": 0.449003 + }, + { + "acc": 0.65111942, + "epoch": 0.5853627600202943, + "grad_norm": 6.0625, + "learning_rate": 8.474438237251864e-06, + "loss": 1.58304501, + "memory(GiB)": 111.15, + "step": 23075, + "train_speed(iter/s)": 0.449054 + }, + { + "acc": 0.65524702, + "epoch": 0.5854895991882293, + "grad_norm": 4.96875, + "learning_rate": 8.473684076348635e-06, + "loss": 1.57415161, + "memory(GiB)": 111.15, + "step": 23080, + "train_speed(iter/s)": 0.449105 + }, + { + "acc": 0.65172296, + "epoch": 0.5856164383561644, + "grad_norm": 5.75, + "learning_rate": 8.472929762657272e-06, + "loss": 1.6312933, + "memory(GiB)": 111.15, + "step": 23085, + "train_speed(iter/s)": 0.449156 + }, + { + "acc": 0.64948559, + "epoch": 0.5857432775240995, + "grad_norm": 5.3125, + "learning_rate": 8.472175296210952e-06, + "loss": 1.59882069, + "memory(GiB)": 111.15, + "step": 23090, + "train_speed(iter/s)": 0.449207 + }, + { + "acc": 0.64895229, + "epoch": 0.5858701166920345, + "grad_norm": 5.59375, + "learning_rate": 8.471420677042858e-06, + "loss": 1.6470335, + "memory(GiB)": 111.15, + "step": 23095, + "train_speed(iter/s)": 0.449257 + }, + { + "acc": 0.65352297, + "epoch": 0.5859969558599696, + "grad_norm": 5.46875, + "learning_rate": 8.470665905186188e-06, + "loss": 1.59021416, + "memory(GiB)": 111.15, + "step": 23100, + "train_speed(iter/s)": 0.449308 + }, + { + "epoch": 0.5859969558599696, + "eval_acc": 0.6417348706613298, + "eval_loss": 1.6167936325073242, + "eval_runtime": 113.2567, + "eval_samples_per_second": 56.244, + "eval_steps_per_second": 28.122, + "step": 23100 + }, + { + "acc": 0.64987774, + "epoch": 0.5861237950279046, + "grad_norm": 4.5, + "learning_rate": 8.469910980674134e-06, + "loss": 1.65329247, + "memory(GiB)": 111.15, + "step": 23105, + "train_speed(iter/s)": 0.448332 + }, + { + "acc": 0.67062922, + "epoch": 0.5862506341958397, + "grad_norm": 5.0, + "learning_rate": 8.469155903539903e-06, + "loss": 1.51398582, + "memory(GiB)": 111.15, + "step": 23110, + "train_speed(iter/s)": 0.448383 + }, + { + "acc": 0.64584312, + "epoch": 0.5863774733637748, + "grad_norm": 6.28125, + "learning_rate": 8.468400673816705e-06, + "loss": 1.67449303, + "memory(GiB)": 111.15, + "step": 23115, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.64628758, + "epoch": 0.5865043125317098, + "grad_norm": 5.84375, + "learning_rate": 8.467645291537763e-06, + "loss": 1.70548992, + "memory(GiB)": 111.15, + "step": 23120, + "train_speed(iter/s)": 0.448484 + }, + { + "acc": 0.64169273, + "epoch": 0.5866311516996449, + "grad_norm": 5.21875, + "learning_rate": 8.466889756736298e-06, + "loss": 1.60538406, + "memory(GiB)": 111.15, + "step": 23125, + "train_speed(iter/s)": 0.448535 + }, + { + "acc": 0.64222288, + "epoch": 0.58675799086758, + "grad_norm": 4.84375, + "learning_rate": 8.466134069445544e-06, + "loss": 1.63615074, + "memory(GiB)": 111.15, + "step": 23130, + "train_speed(iter/s)": 0.448586 + }, + { + "acc": 0.6410542, + "epoch": 0.586884830035515, + "grad_norm": 5.0, + "learning_rate": 8.465378229698737e-06, + "loss": 1.68404961, + "memory(GiB)": 111.15, + "step": 23135, + "train_speed(iter/s)": 0.448637 + }, + { + "acc": 0.63822732, + "epoch": 0.58701166920345, + "grad_norm": 5.15625, + "learning_rate": 8.464622237529123e-06, + "loss": 1.67564049, + "memory(GiB)": 111.15, + "step": 23140, + "train_speed(iter/s)": 0.448687 + }, + { + "acc": 0.6611187, + "epoch": 0.587138508371385, + "grad_norm": 5.5625, + "learning_rate": 8.463866092969958e-06, + "loss": 1.5383131, + "memory(GiB)": 111.15, + "step": 23145, + "train_speed(iter/s)": 0.448738 + }, + { + "acc": 0.6561348, + "epoch": 0.5872653475393201, + "grad_norm": 5.375, + "learning_rate": 8.463109796054495e-06, + "loss": 1.64392185, + "memory(GiB)": 111.15, + "step": 23150, + "train_speed(iter/s)": 0.448788 + }, + { + "acc": 0.64802809, + "epoch": 0.5873921867072552, + "grad_norm": 6.0625, + "learning_rate": 8.462353346815999e-06, + "loss": 1.62696209, + "memory(GiB)": 111.15, + "step": 23155, + "train_speed(iter/s)": 0.448839 + }, + { + "acc": 0.65053215, + "epoch": 0.5875190258751902, + "grad_norm": 5.5, + "learning_rate": 8.461596745287747e-06, + "loss": 1.62008381, + "memory(GiB)": 111.15, + "step": 23160, + "train_speed(iter/s)": 0.44889 + }, + { + "acc": 0.6399344, + "epoch": 0.5876458650431253, + "grad_norm": 5.15625, + "learning_rate": 8.460839991503016e-06, + "loss": 1.70293407, + "memory(GiB)": 111.15, + "step": 23165, + "train_speed(iter/s)": 0.44894 + }, + { + "acc": 0.66283607, + "epoch": 0.5877727042110604, + "grad_norm": 6.46875, + "learning_rate": 8.46008308549509e-06, + "loss": 1.56987171, + "memory(GiB)": 111.15, + "step": 23170, + "train_speed(iter/s)": 0.448991 + }, + { + "acc": 0.64818206, + "epoch": 0.5878995433789954, + "grad_norm": 5.5, + "learning_rate": 8.459326027297261e-06, + "loss": 1.60361042, + "memory(GiB)": 111.15, + "step": 23175, + "train_speed(iter/s)": 0.449042 + }, + { + "acc": 0.65253243, + "epoch": 0.5880263825469305, + "grad_norm": 5.84375, + "learning_rate": 8.45856881694283e-06, + "loss": 1.6228056, + "memory(GiB)": 111.15, + "step": 23180, + "train_speed(iter/s)": 0.449093 + }, + { + "acc": 0.65476165, + "epoch": 0.5881532217148655, + "grad_norm": 5.53125, + "learning_rate": 8.4578114544651e-06, + "loss": 1.58376656, + "memory(GiB)": 111.15, + "step": 23185, + "train_speed(iter/s)": 0.449144 + }, + { + "acc": 0.65313787, + "epoch": 0.5882800608828006, + "grad_norm": 5.28125, + "learning_rate": 8.457053939897385e-06, + "loss": 1.65006256, + "memory(GiB)": 111.15, + "step": 23190, + "train_speed(iter/s)": 0.449194 + }, + { + "acc": 0.66425161, + "epoch": 0.5884069000507357, + "grad_norm": 4.6875, + "learning_rate": 8.456296273273e-06, + "loss": 1.59740944, + "memory(GiB)": 111.15, + "step": 23195, + "train_speed(iter/s)": 0.449245 + }, + { + "acc": 0.65103226, + "epoch": 0.5885337392186707, + "grad_norm": 5.59375, + "learning_rate": 8.455538454625276e-06, + "loss": 1.63570061, + "memory(GiB)": 111.15, + "step": 23200, + "train_speed(iter/s)": 0.449295 + }, + { + "epoch": 0.5885337392186707, + "eval_acc": 0.6418238550982442, + "eval_loss": 1.6165494918823242, + "eval_runtime": 115.0647, + "eval_samples_per_second": 55.36, + "eval_steps_per_second": 27.68, + "step": 23200 + }, + { + "acc": 0.63927102, + "epoch": 0.5886605783866058, + "grad_norm": 5.9375, + "learning_rate": 8.454780483987544e-06, + "loss": 1.65435066, + "memory(GiB)": 111.15, + "step": 23205, + "train_speed(iter/s)": 0.448308 + }, + { + "acc": 0.65784788, + "epoch": 0.5887874175545409, + "grad_norm": 7.90625, + "learning_rate": 8.45402236139314e-06, + "loss": 1.60503693, + "memory(GiB)": 111.15, + "step": 23210, + "train_speed(iter/s)": 0.448358 + }, + { + "acc": 0.6503089, + "epoch": 0.5889142567224759, + "grad_norm": 5.03125, + "learning_rate": 8.453264086875411e-06, + "loss": 1.61936226, + "memory(GiB)": 111.15, + "step": 23215, + "train_speed(iter/s)": 0.448408 + }, + { + "acc": 0.64707217, + "epoch": 0.589041095890411, + "grad_norm": 6.0625, + "learning_rate": 8.452505660467713e-06, + "loss": 1.57188129, + "memory(GiB)": 111.15, + "step": 23220, + "train_speed(iter/s)": 0.448458 + }, + { + "acc": 0.65684462, + "epoch": 0.589167935058346, + "grad_norm": 6.8125, + "learning_rate": 8.451747082203398e-06, + "loss": 1.63755074, + "memory(GiB)": 111.15, + "step": 23225, + "train_speed(iter/s)": 0.448509 + }, + { + "acc": 0.66022835, + "epoch": 0.5892947742262811, + "grad_norm": 6.875, + "learning_rate": 8.450988352115838e-06, + "loss": 1.57754545, + "memory(GiB)": 111.15, + "step": 23230, + "train_speed(iter/s)": 0.44856 + }, + { + "acc": 0.64977036, + "epoch": 0.5894216133942162, + "grad_norm": 4.84375, + "learning_rate": 8.450229470238401e-06, + "loss": 1.61089973, + "memory(GiB)": 111.15, + "step": 23235, + "train_speed(iter/s)": 0.448611 + }, + { + "acc": 0.64307156, + "epoch": 0.5895484525621512, + "grad_norm": 6.5, + "learning_rate": 8.44947043660447e-06, + "loss": 1.64613953, + "memory(GiB)": 111.15, + "step": 23240, + "train_speed(iter/s)": 0.448661 + }, + { + "acc": 0.66351995, + "epoch": 0.5896752917300863, + "grad_norm": 5.71875, + "learning_rate": 8.448711251247425e-06, + "loss": 1.58214035, + "memory(GiB)": 111.15, + "step": 23245, + "train_speed(iter/s)": 0.448712 + }, + { + "acc": 0.65131969, + "epoch": 0.5898021308980214, + "grad_norm": 5.78125, + "learning_rate": 8.447951914200665e-06, + "loss": 1.62277603, + "memory(GiB)": 111.15, + "step": 23250, + "train_speed(iter/s)": 0.448762 + }, + { + "acc": 0.66434689, + "epoch": 0.5899289700659563, + "grad_norm": 5.1875, + "learning_rate": 8.447192425497583e-06, + "loss": 1.58982592, + "memory(GiB)": 111.15, + "step": 23255, + "train_speed(iter/s)": 0.448812 + }, + { + "acc": 0.66494951, + "epoch": 0.5900558092338914, + "grad_norm": 5.9375, + "learning_rate": 8.44643278517159e-06, + "loss": 1.56908293, + "memory(GiB)": 111.15, + "step": 23260, + "train_speed(iter/s)": 0.448862 + }, + { + "acc": 0.63575554, + "epoch": 0.5901826484018264, + "grad_norm": 5.0625, + "learning_rate": 8.445672993256095e-06, + "loss": 1.72937813, + "memory(GiB)": 111.15, + "step": 23265, + "train_speed(iter/s)": 0.448912 + }, + { + "acc": 0.63868203, + "epoch": 0.5903094875697615, + "grad_norm": 5.40625, + "learning_rate": 8.444913049784517e-06, + "loss": 1.69368629, + "memory(GiB)": 111.15, + "step": 23270, + "train_speed(iter/s)": 0.448962 + }, + { + "acc": 0.66530075, + "epoch": 0.5904363267376966, + "grad_norm": 5.53125, + "learning_rate": 8.444152954790285e-06, + "loss": 1.58076134, + "memory(GiB)": 111.15, + "step": 23275, + "train_speed(iter/s)": 0.449012 + }, + { + "acc": 0.65249124, + "epoch": 0.5905631659056316, + "grad_norm": 5.40625, + "learning_rate": 8.443392708306827e-06, + "loss": 1.64800987, + "memory(GiB)": 111.15, + "step": 23280, + "train_speed(iter/s)": 0.449062 + }, + { + "acc": 0.6551228, + "epoch": 0.5906900050735667, + "grad_norm": 6.15625, + "learning_rate": 8.442632310367585e-06, + "loss": 1.62363167, + "memory(GiB)": 111.15, + "step": 23285, + "train_speed(iter/s)": 0.449112 + }, + { + "acc": 0.65323839, + "epoch": 0.5908168442415018, + "grad_norm": 5.3125, + "learning_rate": 8.441871761006001e-06, + "loss": 1.60642662, + "memory(GiB)": 111.15, + "step": 23290, + "train_speed(iter/s)": 0.449162 + }, + { + "acc": 0.64363985, + "epoch": 0.5909436834094368, + "grad_norm": 6.28125, + "learning_rate": 8.441111060255533e-06, + "loss": 1.68104668, + "memory(GiB)": 111.15, + "step": 23295, + "train_speed(iter/s)": 0.449213 + }, + { + "acc": 0.62153292, + "epoch": 0.5910705225773719, + "grad_norm": 5.8125, + "learning_rate": 8.440350208149637e-06, + "loss": 1.76574135, + "memory(GiB)": 111.15, + "step": 23300, + "train_speed(iter/s)": 0.449262 + }, + { + "epoch": 0.5910705225773719, + "eval_acc": 0.6418326282117428, + "eval_loss": 1.616451621055603, + "eval_runtime": 113.5846, + "eval_samples_per_second": 56.082, + "eval_steps_per_second": 28.041, + "step": 23300 + }, + { + "acc": 0.65538616, + "epoch": 0.5911973617453069, + "grad_norm": 6.4375, + "learning_rate": 8.43958920472178e-06, + "loss": 1.63556728, + "memory(GiB)": 111.15, + "step": 23305, + "train_speed(iter/s)": 0.448292 + }, + { + "acc": 0.65255718, + "epoch": 0.591324200913242, + "grad_norm": 4.5625, + "learning_rate": 8.43882805000543e-06, + "loss": 1.63698425, + "memory(GiB)": 111.15, + "step": 23310, + "train_speed(iter/s)": 0.448342 + }, + { + "acc": 0.63830509, + "epoch": 0.5914510400811771, + "grad_norm": 5.25, + "learning_rate": 8.43806674403407e-06, + "loss": 1.67597847, + "memory(GiB)": 111.15, + "step": 23315, + "train_speed(iter/s)": 0.448393 + }, + { + "acc": 0.66265259, + "epoch": 0.5915778792491121, + "grad_norm": 5.5625, + "learning_rate": 8.437305286841187e-06, + "loss": 1.55787048, + "memory(GiB)": 111.15, + "step": 23320, + "train_speed(iter/s)": 0.448443 + }, + { + "acc": 0.66816235, + "epoch": 0.5917047184170472, + "grad_norm": 4.96875, + "learning_rate": 8.436543678460269e-06, + "loss": 1.60376701, + "memory(GiB)": 111.15, + "step": 23325, + "train_speed(iter/s)": 0.448493 + }, + { + "acc": 0.65872583, + "epoch": 0.5918315575849823, + "grad_norm": 6.40625, + "learning_rate": 8.435781918924817e-06, + "loss": 1.61508293, + "memory(GiB)": 111.15, + "step": 23330, + "train_speed(iter/s)": 0.448543 + }, + { + "acc": 0.660042, + "epoch": 0.5919583967529173, + "grad_norm": 5.5, + "learning_rate": 8.435020008268335e-06, + "loss": 1.57884369, + "memory(GiB)": 111.15, + "step": 23335, + "train_speed(iter/s)": 0.448593 + }, + { + "acc": 0.65065513, + "epoch": 0.5920852359208524, + "grad_norm": 5.875, + "learning_rate": 8.43425794652434e-06, + "loss": 1.60144958, + "memory(GiB)": 111.15, + "step": 23340, + "train_speed(iter/s)": 0.448644 + }, + { + "acc": 0.6374424, + "epoch": 0.5922120750887874, + "grad_norm": 4.84375, + "learning_rate": 8.433495733726345e-06, + "loss": 1.67704544, + "memory(GiB)": 111.15, + "step": 23345, + "train_speed(iter/s)": 0.448695 + }, + { + "acc": 0.64767418, + "epoch": 0.5923389142567225, + "grad_norm": 4.5, + "learning_rate": 8.43273336990788e-06, + "loss": 1.69231625, + "memory(GiB)": 111.15, + "step": 23350, + "train_speed(iter/s)": 0.448745 + }, + { + "acc": 0.65411224, + "epoch": 0.5924657534246576, + "grad_norm": 5.25, + "learning_rate": 8.431970855102475e-06, + "loss": 1.62895432, + "memory(GiB)": 111.15, + "step": 23355, + "train_speed(iter/s)": 0.448796 + }, + { + "acc": 0.65874577, + "epoch": 0.5925925925925926, + "grad_norm": 4.21875, + "learning_rate": 8.43120818934367e-06, + "loss": 1.58886738, + "memory(GiB)": 111.15, + "step": 23360, + "train_speed(iter/s)": 0.448846 + }, + { + "acc": 0.65032244, + "epoch": 0.5927194317605277, + "grad_norm": 5.53125, + "learning_rate": 8.430445372665008e-06, + "loss": 1.67244072, + "memory(GiB)": 111.15, + "step": 23365, + "train_speed(iter/s)": 0.448897 + }, + { + "acc": 0.64135933, + "epoch": 0.5928462709284628, + "grad_norm": 4.8125, + "learning_rate": 8.429682405100042e-06, + "loss": 1.69656372, + "memory(GiB)": 111.15, + "step": 23370, + "train_speed(iter/s)": 0.448948 + }, + { + "acc": 0.6627584, + "epoch": 0.5929731100963977, + "grad_norm": 5.4375, + "learning_rate": 8.428919286682333e-06, + "loss": 1.59113445, + "memory(GiB)": 111.15, + "step": 23375, + "train_speed(iter/s)": 0.448999 + }, + { + "acc": 0.66017408, + "epoch": 0.5930999492643328, + "grad_norm": 5.0, + "learning_rate": 8.428156017445443e-06, + "loss": 1.59972477, + "memory(GiB)": 111.15, + "step": 23380, + "train_speed(iter/s)": 0.449049 + }, + { + "acc": 0.64495106, + "epoch": 0.5932267884322678, + "grad_norm": 5.625, + "learning_rate": 8.427392597422947e-06, + "loss": 1.64710007, + "memory(GiB)": 111.15, + "step": 23385, + "train_speed(iter/s)": 0.4491 + }, + { + "acc": 0.64679041, + "epoch": 0.5933536276002029, + "grad_norm": 5.0, + "learning_rate": 8.426629026648423e-06, + "loss": 1.63350277, + "memory(GiB)": 111.15, + "step": 23390, + "train_speed(iter/s)": 0.449151 + }, + { + "acc": 0.64052677, + "epoch": 0.593480466768138, + "grad_norm": 6.21875, + "learning_rate": 8.425865305155455e-06, + "loss": 1.67144451, + "memory(GiB)": 111.15, + "step": 23395, + "train_speed(iter/s)": 0.449201 + }, + { + "acc": 0.65852709, + "epoch": 0.593607305936073, + "grad_norm": 5.6875, + "learning_rate": 8.425101432977636e-06, + "loss": 1.64728832, + "memory(GiB)": 111.15, + "step": 23400, + "train_speed(iter/s)": 0.449251 + }, + { + "epoch": 0.593607305936073, + "eval_acc": 0.641776229624966, + "eval_loss": 1.6166644096374512, + "eval_runtime": 113.154, + "eval_samples_per_second": 56.295, + "eval_steps_per_second": 28.147, + "step": 23400 + }, + { + "acc": 0.62818718, + "epoch": 0.5937341451040081, + "grad_norm": 5.40625, + "learning_rate": 8.424337410148562e-06, + "loss": 1.68228779, + "memory(GiB)": 111.15, + "step": 23405, + "train_speed(iter/s)": 0.448289 + }, + { + "acc": 0.63048315, + "epoch": 0.5938609842719432, + "grad_norm": 5.875, + "learning_rate": 8.423573236701842e-06, + "loss": 1.68827209, + "memory(GiB)": 111.15, + "step": 23410, + "train_speed(iter/s)": 0.448339 + }, + { + "acc": 0.64381137, + "epoch": 0.5939878234398782, + "grad_norm": 5.0625, + "learning_rate": 8.422808912671086e-06, + "loss": 1.67398682, + "memory(GiB)": 111.15, + "step": 23415, + "train_speed(iter/s)": 0.448389 + }, + { + "acc": 0.66279402, + "epoch": 0.5941146626078133, + "grad_norm": 4.28125, + "learning_rate": 8.422044438089911e-06, + "loss": 1.61469784, + "memory(GiB)": 111.15, + "step": 23420, + "train_speed(iter/s)": 0.448439 + }, + { + "acc": 0.6581779, + "epoch": 0.5942415017757483, + "grad_norm": 7.0625, + "learning_rate": 8.421279812991944e-06, + "loss": 1.61584778, + "memory(GiB)": 111.15, + "step": 23425, + "train_speed(iter/s)": 0.448489 + }, + { + "acc": 0.6525785, + "epoch": 0.5943683409436834, + "grad_norm": 4.65625, + "learning_rate": 8.420515037410817e-06, + "loss": 1.62470512, + "memory(GiB)": 111.15, + "step": 23430, + "train_speed(iter/s)": 0.448539 + }, + { + "acc": 0.64000359, + "epoch": 0.5944951801116185, + "grad_norm": 4.90625, + "learning_rate": 8.419750111380166e-06, + "loss": 1.60711708, + "memory(GiB)": 111.15, + "step": 23435, + "train_speed(iter/s)": 0.448589 + }, + { + "acc": 0.64769611, + "epoch": 0.5946220192795535, + "grad_norm": 5.3125, + "learning_rate": 8.418985034933637e-06, + "loss": 1.66046066, + "memory(GiB)": 111.15, + "step": 23440, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.65740738, + "epoch": 0.5947488584474886, + "grad_norm": 5.34375, + "learning_rate": 8.418219808104882e-06, + "loss": 1.58712006, + "memory(GiB)": 111.15, + "step": 23445, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.65048394, + "epoch": 0.5948756976154237, + "grad_norm": 5.5625, + "learning_rate": 8.417454430927559e-06, + "loss": 1.63169708, + "memory(GiB)": 111.15, + "step": 23450, + "train_speed(iter/s)": 0.448741 + }, + { + "acc": 0.64746399, + "epoch": 0.5950025367833587, + "grad_norm": 5.125, + "learning_rate": 8.41668890343533e-06, + "loss": 1.66768456, + "memory(GiB)": 111.15, + "step": 23455, + "train_speed(iter/s)": 0.448791 + }, + { + "acc": 0.63737059, + "epoch": 0.5951293759512938, + "grad_norm": 4.4375, + "learning_rate": 8.41592322566187e-06, + "loss": 1.69460106, + "memory(GiB)": 111.15, + "step": 23460, + "train_speed(iter/s)": 0.448841 + }, + { + "acc": 0.6488472, + "epoch": 0.5952562151192288, + "grad_norm": 5.71875, + "learning_rate": 8.415157397640857e-06, + "loss": 1.55209961, + "memory(GiB)": 111.15, + "step": 23465, + "train_speed(iter/s)": 0.448891 + }, + { + "acc": 0.64201517, + "epoch": 0.5953830542871639, + "grad_norm": 5.375, + "learning_rate": 8.414391419405972e-06, + "loss": 1.63962631, + "memory(GiB)": 111.15, + "step": 23470, + "train_speed(iter/s)": 0.448941 + }, + { + "acc": 0.6491375, + "epoch": 0.595509893455099, + "grad_norm": 5.875, + "learning_rate": 8.413625290990909e-06, + "loss": 1.6684721, + "memory(GiB)": 111.15, + "step": 23475, + "train_speed(iter/s)": 0.448992 + }, + { + "acc": 0.66506882, + "epoch": 0.595636732623034, + "grad_norm": 5.25, + "learning_rate": 8.412859012429365e-06, + "loss": 1.56786804, + "memory(GiB)": 111.15, + "step": 23480, + "train_speed(iter/s)": 0.449042 + }, + { + "acc": 0.65743551, + "epoch": 0.5957635717909691, + "grad_norm": 5.3125, + "learning_rate": 8.412092583755043e-06, + "loss": 1.61064625, + "memory(GiB)": 111.15, + "step": 23485, + "train_speed(iter/s)": 0.449092 + }, + { + "acc": 0.64781294, + "epoch": 0.5958904109589042, + "grad_norm": 5.1875, + "learning_rate": 8.411326005001658e-06, + "loss": 1.64092655, + "memory(GiB)": 111.15, + "step": 23490, + "train_speed(iter/s)": 0.449142 + }, + { + "acc": 0.64041772, + "epoch": 0.5960172501268391, + "grad_norm": 5.40625, + "learning_rate": 8.410559276202922e-06, + "loss": 1.6614193, + "memory(GiB)": 111.15, + "step": 23495, + "train_speed(iter/s)": 0.449192 + }, + { + "acc": 0.63829336, + "epoch": 0.5961440892947742, + "grad_norm": 4.5, + "learning_rate": 8.409792397392565e-06, + "loss": 1.64132404, + "memory(GiB)": 111.15, + "step": 23500, + "train_speed(iter/s)": 0.449242 + }, + { + "epoch": 0.5961440892947742, + "eval_acc": 0.6418731516407602, + "eval_loss": 1.6161493062973022, + "eval_runtime": 113.7131, + "eval_samples_per_second": 56.018, + "eval_steps_per_second": 28.009, + "step": 23500 + }, + { + "acc": 0.64420033, + "epoch": 0.5962709284627092, + "grad_norm": 5.9375, + "learning_rate": 8.40902536860431e-06, + "loss": 1.63698349, + "memory(GiB)": 111.15, + "step": 23505, + "train_speed(iter/s)": 0.448278 + }, + { + "acc": 0.6546133, + "epoch": 0.5963977676306443, + "grad_norm": 5.375, + "learning_rate": 8.408258189871904e-06, + "loss": 1.59954939, + "memory(GiB)": 111.15, + "step": 23510, + "train_speed(iter/s)": 0.448327 + }, + { + "acc": 0.66496115, + "epoch": 0.5965246067985794, + "grad_norm": 5.0625, + "learning_rate": 8.407490861229084e-06, + "loss": 1.63849277, + "memory(GiB)": 111.15, + "step": 23515, + "train_speed(iter/s)": 0.448377 + }, + { + "acc": 0.6422358, + "epoch": 0.5966514459665144, + "grad_norm": 6.4375, + "learning_rate": 8.406723382709603e-06, + "loss": 1.6848402, + "memory(GiB)": 111.15, + "step": 23520, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.66246252, + "epoch": 0.5967782851344495, + "grad_norm": 5.03125, + "learning_rate": 8.405955754347216e-06, + "loss": 1.64211864, + "memory(GiB)": 111.15, + "step": 23525, + "train_speed(iter/s)": 0.448476 + }, + { + "acc": 0.65280604, + "epoch": 0.5969051243023846, + "grad_norm": 5.09375, + "learning_rate": 8.40518797617569e-06, + "loss": 1.62443275, + "memory(GiB)": 111.15, + "step": 23530, + "train_speed(iter/s)": 0.448525 + }, + { + "acc": 0.66929989, + "epoch": 0.5970319634703196, + "grad_norm": 5.25, + "learning_rate": 8.404420048228794e-06, + "loss": 1.48864632, + "memory(GiB)": 111.15, + "step": 23535, + "train_speed(iter/s)": 0.448575 + }, + { + "acc": 0.63921747, + "epoch": 0.5971588026382547, + "grad_norm": 4.65625, + "learning_rate": 8.403651970540305e-06, + "loss": 1.68685303, + "memory(GiB)": 111.15, + "step": 23540, + "train_speed(iter/s)": 0.448625 + }, + { + "acc": 0.6429018, + "epoch": 0.5972856418061897, + "grad_norm": 5.5, + "learning_rate": 8.402883743144005e-06, + "loss": 1.60751457, + "memory(GiB)": 111.15, + "step": 23545, + "train_speed(iter/s)": 0.448675 + }, + { + "acc": 0.64981928, + "epoch": 0.5974124809741248, + "grad_norm": 5.5625, + "learning_rate": 8.402115366073686e-06, + "loss": 1.66617432, + "memory(GiB)": 111.15, + "step": 23550, + "train_speed(iter/s)": 0.448725 + }, + { + "acc": 0.64550829, + "epoch": 0.5975393201420599, + "grad_norm": 5.40625, + "learning_rate": 8.401346839363143e-06, + "loss": 1.61635437, + "memory(GiB)": 111.15, + "step": 23555, + "train_speed(iter/s)": 0.448775 + }, + { + "acc": 0.65077085, + "epoch": 0.5976661593099949, + "grad_norm": 5.4375, + "learning_rate": 8.40057816304618e-06, + "loss": 1.63702354, + "memory(GiB)": 111.15, + "step": 23560, + "train_speed(iter/s)": 0.448825 + }, + { + "acc": 0.64409399, + "epoch": 0.59779299847793, + "grad_norm": 6.5, + "learning_rate": 8.399809337156608e-06, + "loss": 1.6966114, + "memory(GiB)": 111.15, + "step": 23565, + "train_speed(iter/s)": 0.448875 + }, + { + "acc": 0.64371099, + "epoch": 0.5979198376458651, + "grad_norm": 4.96875, + "learning_rate": 8.39904036172824e-06, + "loss": 1.62886677, + "memory(GiB)": 111.15, + "step": 23570, + "train_speed(iter/s)": 0.448925 + }, + { + "acc": 0.64710007, + "epoch": 0.5980466768138001, + "grad_norm": 5.125, + "learning_rate": 8.398271236794904e-06, + "loss": 1.62430553, + "memory(GiB)": 111.15, + "step": 23575, + "train_speed(iter/s)": 0.448975 + }, + { + "acc": 0.64346023, + "epoch": 0.5981735159817352, + "grad_norm": 5.875, + "learning_rate": 8.397501962390427e-06, + "loss": 1.62991676, + "memory(GiB)": 111.15, + "step": 23580, + "train_speed(iter/s)": 0.449025 + }, + { + "acc": 0.6354207, + "epoch": 0.5983003551496702, + "grad_norm": 4.875, + "learning_rate": 8.396732538548642e-06, + "loss": 1.64980659, + "memory(GiB)": 111.15, + "step": 23585, + "train_speed(iter/s)": 0.449074 + }, + { + "acc": 0.65266747, + "epoch": 0.5984271943176053, + "grad_norm": 5.0, + "learning_rate": 8.395962965303397e-06, + "loss": 1.69421768, + "memory(GiB)": 111.15, + "step": 23590, + "train_speed(iter/s)": 0.449124 + }, + { + "acc": 0.66890898, + "epoch": 0.5985540334855404, + "grad_norm": 4.78125, + "learning_rate": 8.395193242688537e-06, + "loss": 1.60606575, + "memory(GiB)": 111.15, + "step": 23595, + "train_speed(iter/s)": 0.449174 + }, + { + "acc": 0.65181241, + "epoch": 0.5986808726534754, + "grad_norm": 5.21875, + "learning_rate": 8.394423370737922e-06, + "loss": 1.55499516, + "memory(GiB)": 111.15, + "step": 23600, + "train_speed(iter/s)": 0.449224 + }, + { + "epoch": 0.5986808726534754, + "eval_acc": 0.6419082440947547, + "eval_loss": 1.6160775423049927, + "eval_runtime": 112.9358, + "eval_samples_per_second": 56.404, + "eval_steps_per_second": 28.202, + "step": 23600 + }, + { + "acc": 0.65439949, + "epoch": 0.5988077118214105, + "grad_norm": 5.21875, + "learning_rate": 8.393653349485412e-06, + "loss": 1.59696865, + "memory(GiB)": 111.15, + "step": 23605, + "train_speed(iter/s)": 0.448271 + }, + { + "acc": 0.64576721, + "epoch": 0.5989345509893456, + "grad_norm": 6.375, + "learning_rate": 8.392883178964874e-06, + "loss": 1.66987896, + "memory(GiB)": 111.15, + "step": 23610, + "train_speed(iter/s)": 0.448321 + }, + { + "acc": 0.6554678, + "epoch": 0.5990613901572805, + "grad_norm": 5.625, + "learning_rate": 8.392112859210186e-06, + "loss": 1.60680008, + "memory(GiB)": 111.15, + "step": 23615, + "train_speed(iter/s)": 0.44837 + }, + { + "acc": 0.64539347, + "epoch": 0.5991882293252156, + "grad_norm": 5.21875, + "learning_rate": 8.391342390255232e-06, + "loss": 1.66131287, + "memory(GiB)": 111.15, + "step": 23620, + "train_speed(iter/s)": 0.44842 + }, + { + "acc": 0.64740324, + "epoch": 0.5993150684931506, + "grad_norm": 4.65625, + "learning_rate": 8.390571772133896e-06, + "loss": 1.64869537, + "memory(GiB)": 111.15, + "step": 23625, + "train_speed(iter/s)": 0.448469 + }, + { + "acc": 0.66083355, + "epoch": 0.5994419076610857, + "grad_norm": 5.625, + "learning_rate": 8.389801004880077e-06, + "loss": 1.60963573, + "memory(GiB)": 111.15, + "step": 23630, + "train_speed(iter/s)": 0.448519 + }, + { + "acc": 0.64124022, + "epoch": 0.5995687468290208, + "grad_norm": 5.15625, + "learning_rate": 8.389030088527675e-06, + "loss": 1.68125591, + "memory(GiB)": 111.15, + "step": 23635, + "train_speed(iter/s)": 0.448568 + }, + { + "acc": 0.64289541, + "epoch": 0.5996955859969558, + "grad_norm": 5.34375, + "learning_rate": 8.388259023110598e-06, + "loss": 1.68053303, + "memory(GiB)": 111.15, + "step": 23640, + "train_speed(iter/s)": 0.448618 + }, + { + "acc": 0.64912472, + "epoch": 0.5998224251648909, + "grad_norm": 4.875, + "learning_rate": 8.387487808662765e-06, + "loss": 1.63533363, + "memory(GiB)": 111.15, + "step": 23645, + "train_speed(iter/s)": 0.448667 + }, + { + "acc": 0.63560953, + "epoch": 0.599949264332826, + "grad_norm": 5.21875, + "learning_rate": 8.38671644521809e-06, + "loss": 1.67629261, + "memory(GiB)": 111.15, + "step": 23650, + "train_speed(iter/s)": 0.448717 + }, + { + "acc": 0.65077744, + "epoch": 0.600076103500761, + "grad_norm": 5.03125, + "learning_rate": 8.385944932810508e-06, + "loss": 1.64791584, + "memory(GiB)": 111.15, + "step": 23655, + "train_speed(iter/s)": 0.448766 + }, + { + "acc": 0.6605206, + "epoch": 0.6002029426686961, + "grad_norm": 5.40625, + "learning_rate": 8.385173271473948e-06, + "loss": 1.61464653, + "memory(GiB)": 111.15, + "step": 23660, + "train_speed(iter/s)": 0.448816 + }, + { + "acc": 0.64481888, + "epoch": 0.6003297818366311, + "grad_norm": 4.78125, + "learning_rate": 8.384401461242355e-06, + "loss": 1.68187256, + "memory(GiB)": 111.15, + "step": 23665, + "train_speed(iter/s)": 0.448865 + }, + { + "acc": 0.64898453, + "epoch": 0.6004566210045662, + "grad_norm": 6.0625, + "learning_rate": 8.383629502149678e-06, + "loss": 1.61226654, + "memory(GiB)": 111.15, + "step": 23670, + "train_speed(iter/s)": 0.448915 + }, + { + "acc": 0.64739227, + "epoch": 0.6005834601725013, + "grad_norm": 5.59375, + "learning_rate": 8.382857394229865e-06, + "loss": 1.61824398, + "memory(GiB)": 111.15, + "step": 23675, + "train_speed(iter/s)": 0.448964 + }, + { + "acc": 0.64616699, + "epoch": 0.6007102993404363, + "grad_norm": 5.3125, + "learning_rate": 8.382085137516883e-06, + "loss": 1.65122662, + "memory(GiB)": 111.15, + "step": 23680, + "train_speed(iter/s)": 0.449014 + }, + { + "acc": 0.66652431, + "epoch": 0.6008371385083714, + "grad_norm": 4.84375, + "learning_rate": 8.381312732044696e-06, + "loss": 1.60495186, + "memory(GiB)": 111.15, + "step": 23685, + "train_speed(iter/s)": 0.449064 + }, + { + "acc": 0.66459036, + "epoch": 0.6009639776763065, + "grad_norm": 5.9375, + "learning_rate": 8.380540177847278e-06, + "loss": 1.58745747, + "memory(GiB)": 111.15, + "step": 23690, + "train_speed(iter/s)": 0.449114 + }, + { + "acc": 0.63450336, + "epoch": 0.6010908168442415, + "grad_norm": 5.09375, + "learning_rate": 8.37976747495861e-06, + "loss": 1.71009655, + "memory(GiB)": 111.15, + "step": 23695, + "train_speed(iter/s)": 0.449164 + }, + { + "acc": 0.64483576, + "epoch": 0.6012176560121766, + "grad_norm": 6.59375, + "learning_rate": 8.378994623412679e-06, + "loss": 1.58282604, + "memory(GiB)": 111.15, + "step": 23700, + "train_speed(iter/s)": 0.449214 + }, + { + "epoch": 0.6012176560121766, + "eval_acc": 0.6419801000719813, + "eval_loss": 1.6157214641571045, + "eval_runtime": 113.2671, + "eval_samples_per_second": 56.239, + "eval_steps_per_second": 28.119, + "step": 23700 + }, + { + "acc": 0.63603334, + "epoch": 0.6013444951801116, + "grad_norm": 5.1875, + "learning_rate": 8.378221623243478e-06, + "loss": 1.64662209, + "memory(GiB)": 111.15, + "step": 23705, + "train_speed(iter/s)": 0.448263 + }, + { + "acc": 0.64660902, + "epoch": 0.6014713343480467, + "grad_norm": 4.75, + "learning_rate": 8.377448474485008e-06, + "loss": 1.57717714, + "memory(GiB)": 111.15, + "step": 23710, + "train_speed(iter/s)": 0.448312 + }, + { + "acc": 0.65618019, + "epoch": 0.6015981735159818, + "grad_norm": 4.625, + "learning_rate": 8.376675177171273e-06, + "loss": 1.63086014, + "memory(GiB)": 111.15, + "step": 23715, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.65910282, + "epoch": 0.6017250126839168, + "grad_norm": 5.0625, + "learning_rate": 8.375901731336292e-06, + "loss": 1.60130844, + "memory(GiB)": 111.15, + "step": 23720, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.65640688, + "epoch": 0.6018518518518519, + "grad_norm": 6.28125, + "learning_rate": 8.375128137014076e-06, + "loss": 1.61374168, + "memory(GiB)": 111.15, + "step": 23725, + "train_speed(iter/s)": 0.448462 + }, + { + "acc": 0.63872781, + "epoch": 0.601978691019787, + "grad_norm": 6.78125, + "learning_rate": 8.374354394238658e-06, + "loss": 1.73616867, + "memory(GiB)": 111.15, + "step": 23730, + "train_speed(iter/s)": 0.448512 + }, + { + "acc": 0.65630341, + "epoch": 0.602105530187722, + "grad_norm": 5.375, + "learning_rate": 8.373580503044068e-06, + "loss": 1.58801384, + "memory(GiB)": 111.15, + "step": 23735, + "train_speed(iter/s)": 0.448562 + }, + { + "acc": 0.65598059, + "epoch": 0.602232369355657, + "grad_norm": 6.90625, + "learning_rate": 8.372806463464347e-06, + "loss": 1.66198902, + "memory(GiB)": 111.15, + "step": 23740, + "train_speed(iter/s)": 0.448612 + }, + { + "acc": 0.66888008, + "epoch": 0.602359208523592, + "grad_norm": 5.4375, + "learning_rate": 8.372032275533538e-06, + "loss": 1.51733227, + "memory(GiB)": 111.15, + "step": 23745, + "train_speed(iter/s)": 0.448662 + }, + { + "acc": 0.64712029, + "epoch": 0.6024860476915271, + "grad_norm": 5.6875, + "learning_rate": 8.371257939285692e-06, + "loss": 1.6361578, + "memory(GiB)": 111.15, + "step": 23750, + "train_speed(iter/s)": 0.448713 + }, + { + "acc": 0.6643734, + "epoch": 0.6026128868594622, + "grad_norm": 5.53125, + "learning_rate": 8.370483454754873e-06, + "loss": 1.612043, + "memory(GiB)": 111.15, + "step": 23755, + "train_speed(iter/s)": 0.448763 + }, + { + "acc": 0.65706902, + "epoch": 0.6027397260273972, + "grad_norm": 4.90625, + "learning_rate": 8.369708821975144e-06, + "loss": 1.61559105, + "memory(GiB)": 111.15, + "step": 23760, + "train_speed(iter/s)": 0.448813 + }, + { + "acc": 0.66564808, + "epoch": 0.6028665651953323, + "grad_norm": 5.71875, + "learning_rate": 8.368934040980576e-06, + "loss": 1.56224689, + "memory(GiB)": 111.15, + "step": 23765, + "train_speed(iter/s)": 0.448863 + }, + { + "acc": 0.65637512, + "epoch": 0.6029934043632674, + "grad_norm": 5.875, + "learning_rate": 8.368159111805246e-06, + "loss": 1.68899155, + "memory(GiB)": 111.15, + "step": 23770, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.66560898, + "epoch": 0.6031202435312024, + "grad_norm": 4.59375, + "learning_rate": 8.367384034483242e-06, + "loss": 1.5762455, + "memory(GiB)": 111.15, + "step": 23775, + "train_speed(iter/s)": 0.448964 + }, + { + "acc": 0.65194826, + "epoch": 0.6032470826991375, + "grad_norm": 5.4375, + "learning_rate": 8.366608809048653e-06, + "loss": 1.67635307, + "memory(GiB)": 111.15, + "step": 23780, + "train_speed(iter/s)": 0.449014 + }, + { + "acc": 0.6362474, + "epoch": 0.6033739218670725, + "grad_norm": 6.15625, + "learning_rate": 8.365833435535579e-06, + "loss": 1.65092773, + "memory(GiB)": 111.15, + "step": 23785, + "train_speed(iter/s)": 0.449064 + }, + { + "acc": 0.65501471, + "epoch": 0.6035007610350076, + "grad_norm": 5.46875, + "learning_rate": 8.365057913978123e-06, + "loss": 1.68895664, + "memory(GiB)": 111.15, + "step": 23790, + "train_speed(iter/s)": 0.449113 + }, + { + "acc": 0.63755245, + "epoch": 0.6036276002029427, + "grad_norm": 6.0, + "learning_rate": 8.364282244410394e-06, + "loss": 1.72748241, + "memory(GiB)": 111.15, + "step": 23795, + "train_speed(iter/s)": 0.449163 + }, + { + "acc": 0.65550599, + "epoch": 0.6037544393708777, + "grad_norm": 4.4375, + "learning_rate": 8.363506426866513e-06, + "loss": 1.62264557, + "memory(GiB)": 111.15, + "step": 23800, + "train_speed(iter/s)": 0.449213 + }, + { + "epoch": 0.6037544393708777, + "eval_acc": 0.6420277255452594, + "eval_loss": 1.6157273054122925, + "eval_runtime": 112.6587, + "eval_samples_per_second": 56.542, + "eval_steps_per_second": 28.271, + "step": 23800 + }, + { + "acc": 0.64704466, + "epoch": 0.6038812785388128, + "grad_norm": 4.625, + "learning_rate": 8.362730461380602e-06, + "loss": 1.62869205, + "memory(GiB)": 111.15, + "step": 23805, + "train_speed(iter/s)": 0.448271 + }, + { + "acc": 0.6419909, + "epoch": 0.6040081177067479, + "grad_norm": 5.625, + "learning_rate": 8.361954347986793e-06, + "loss": 1.63328896, + "memory(GiB)": 111.15, + "step": 23810, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.6649404, + "epoch": 0.6041349568746829, + "grad_norm": 4.375, + "learning_rate": 8.36117808671922e-06, + "loss": 1.5370882, + "memory(GiB)": 111.15, + "step": 23815, + "train_speed(iter/s)": 0.44837 + }, + { + "acc": 0.64838963, + "epoch": 0.604261796042618, + "grad_norm": 4.90625, + "learning_rate": 8.36040167761203e-06, + "loss": 1.62437916, + "memory(GiB)": 111.15, + "step": 23820, + "train_speed(iter/s)": 0.44842 + }, + { + "acc": 0.64471788, + "epoch": 0.604388635210553, + "grad_norm": 5.5625, + "learning_rate": 8.359625120699368e-06, + "loss": 1.69578857, + "memory(GiB)": 111.15, + "step": 23825, + "train_speed(iter/s)": 0.44847 + }, + { + "acc": 0.65368986, + "epoch": 0.6045154743784881, + "grad_norm": 5.15625, + "learning_rate": 8.358848416015397e-06, + "loss": 1.58141546, + "memory(GiB)": 111.15, + "step": 23830, + "train_speed(iter/s)": 0.448519 + }, + { + "acc": 0.63312416, + "epoch": 0.6046423135464232, + "grad_norm": 4.625, + "learning_rate": 8.358071563594274e-06, + "loss": 1.68102283, + "memory(GiB)": 111.15, + "step": 23835, + "train_speed(iter/s)": 0.448569 + }, + { + "acc": 0.64522429, + "epoch": 0.6047691527143582, + "grad_norm": 5.25, + "learning_rate": 8.357294563470173e-06, + "loss": 1.63712883, + "memory(GiB)": 111.15, + "step": 23840, + "train_speed(iter/s)": 0.448619 + }, + { + "acc": 0.66373386, + "epoch": 0.6048959918822933, + "grad_norm": 5.5625, + "learning_rate": 8.356517415677267e-06, + "loss": 1.56209164, + "memory(GiB)": 111.15, + "step": 23845, + "train_speed(iter/s)": 0.448668 + }, + { + "acc": 0.63260403, + "epoch": 0.6050228310502284, + "grad_norm": 5.125, + "learning_rate": 8.355740120249739e-06, + "loss": 1.72288589, + "memory(GiB)": 111.15, + "step": 23850, + "train_speed(iter/s)": 0.448718 + }, + { + "acc": 0.63627009, + "epoch": 0.6051496702181633, + "grad_norm": 5.53125, + "learning_rate": 8.354962677221779e-06, + "loss": 1.6465107, + "memory(GiB)": 111.15, + "step": 23855, + "train_speed(iter/s)": 0.448767 + }, + { + "acc": 0.64391503, + "epoch": 0.6052765093860984, + "grad_norm": 5.21875, + "learning_rate": 8.35418508662758e-06, + "loss": 1.6186821, + "memory(GiB)": 111.15, + "step": 23860, + "train_speed(iter/s)": 0.448818 + }, + { + "acc": 0.64188638, + "epoch": 0.6054033485540334, + "grad_norm": 5.5, + "learning_rate": 8.353407348501346e-06, + "loss": 1.65964394, + "memory(GiB)": 111.15, + "step": 23865, + "train_speed(iter/s)": 0.448867 + }, + { + "acc": 0.64982538, + "epoch": 0.6055301877219685, + "grad_norm": 5.34375, + "learning_rate": 8.352629462877286e-06, + "loss": 1.63497505, + "memory(GiB)": 111.15, + "step": 23870, + "train_speed(iter/s)": 0.448917 + }, + { + "acc": 0.64089413, + "epoch": 0.6056570268899036, + "grad_norm": 5.46875, + "learning_rate": 8.351851429789613e-06, + "loss": 1.66134377, + "memory(GiB)": 111.15, + "step": 23875, + "train_speed(iter/s)": 0.448967 + }, + { + "acc": 0.63387203, + "epoch": 0.6057838660578386, + "grad_norm": 5.125, + "learning_rate": 8.35107324927255e-06, + "loss": 1.62418346, + "memory(GiB)": 111.15, + "step": 23880, + "train_speed(iter/s)": 0.449018 + }, + { + "acc": 0.64221153, + "epoch": 0.6059107052257737, + "grad_norm": 6.3125, + "learning_rate": 8.350294921360323e-06, + "loss": 1.68231888, + "memory(GiB)": 111.15, + "step": 23885, + "train_speed(iter/s)": 0.449068 + }, + { + "acc": 0.64223852, + "epoch": 0.6060375443937088, + "grad_norm": 5.03125, + "learning_rate": 8.349516446087168e-06, + "loss": 1.68805752, + "memory(GiB)": 111.15, + "step": 23890, + "train_speed(iter/s)": 0.449117 + }, + { + "acc": 0.63160515, + "epoch": 0.6061643835616438, + "grad_norm": 5.46875, + "learning_rate": 8.348737823487325e-06, + "loss": 1.6857523, + "memory(GiB)": 111.15, + "step": 23895, + "train_speed(iter/s)": 0.449167 + }, + { + "acc": 0.64534397, + "epoch": 0.6062912227295789, + "grad_norm": 5.46875, + "learning_rate": 8.347959053595042e-06, + "loss": 1.69602242, + "memory(GiB)": 111.15, + "step": 23900, + "train_speed(iter/s)": 0.449217 + }, + { + "epoch": 0.6062912227295789, + "eval_acc": 0.6419638071469125, + "eval_loss": 1.6155890226364136, + "eval_runtime": 113.5631, + "eval_samples_per_second": 56.092, + "eval_steps_per_second": 28.046, + "step": 23900 + }, + { + "acc": 0.63994799, + "epoch": 0.6064180618975139, + "grad_norm": 5.90625, + "learning_rate": 8.347180136444572e-06, + "loss": 1.69585075, + "memory(GiB)": 111.15, + "step": 23905, + "train_speed(iter/s)": 0.448272 + }, + { + "acc": 0.65492764, + "epoch": 0.606544901065449, + "grad_norm": 5.4375, + "learning_rate": 8.346401072070174e-06, + "loss": 1.63015194, + "memory(GiB)": 111.15, + "step": 23910, + "train_speed(iter/s)": 0.448321 + }, + { + "acc": 0.6284873, + "epoch": 0.6066717402333841, + "grad_norm": 5.625, + "learning_rate": 8.345621860506119e-06, + "loss": 1.69242439, + "memory(GiB)": 111.15, + "step": 23915, + "train_speed(iter/s)": 0.44837 + }, + { + "acc": 0.65389061, + "epoch": 0.6067985794013191, + "grad_norm": 5.6875, + "learning_rate": 8.344842501786675e-06, + "loss": 1.5634449, + "memory(GiB)": 111.15, + "step": 23920, + "train_speed(iter/s)": 0.44842 + }, + { + "acc": 0.63611941, + "epoch": 0.6069254185692542, + "grad_norm": 5.96875, + "learning_rate": 8.344062995946125e-06, + "loss": 1.71707325, + "memory(GiB)": 111.15, + "step": 23925, + "train_speed(iter/s)": 0.448469 + }, + { + "acc": 0.64097881, + "epoch": 0.6070522577371893, + "grad_norm": 5.25, + "learning_rate": 8.343283343018755e-06, + "loss": 1.66664677, + "memory(GiB)": 111.15, + "step": 23930, + "train_speed(iter/s)": 0.448518 + }, + { + "acc": 0.6466095, + "epoch": 0.6071790969051243, + "grad_norm": 5.28125, + "learning_rate": 8.342503543038855e-06, + "loss": 1.65252228, + "memory(GiB)": 111.15, + "step": 23935, + "train_speed(iter/s)": 0.448567 + }, + { + "acc": 0.64600081, + "epoch": 0.6073059360730594, + "grad_norm": 5.53125, + "learning_rate": 8.341723596040728e-06, + "loss": 1.70074825, + "memory(GiB)": 111.15, + "step": 23940, + "train_speed(iter/s)": 0.448616 + }, + { + "acc": 0.63149948, + "epoch": 0.6074327752409944, + "grad_norm": 5.09375, + "learning_rate": 8.340943502058675e-06, + "loss": 1.70064774, + "memory(GiB)": 111.15, + "step": 23945, + "train_speed(iter/s)": 0.448665 + }, + { + "acc": 0.64725795, + "epoch": 0.6075596144089295, + "grad_norm": 5.4375, + "learning_rate": 8.340163261127014e-06, + "loss": 1.68268623, + "memory(GiB)": 111.15, + "step": 23950, + "train_speed(iter/s)": 0.448715 + }, + { + "acc": 0.63979397, + "epoch": 0.6076864535768646, + "grad_norm": 5.375, + "learning_rate": 8.339382873280058e-06, + "loss": 1.67447395, + "memory(GiB)": 111.15, + "step": 23955, + "train_speed(iter/s)": 0.448764 + }, + { + "acc": 0.64614072, + "epoch": 0.6078132927447996, + "grad_norm": 4.8125, + "learning_rate": 8.338602338552136e-06, + "loss": 1.68541336, + "memory(GiB)": 111.15, + "step": 23960, + "train_speed(iter/s)": 0.448813 + }, + { + "acc": 0.64770303, + "epoch": 0.6079401319127347, + "grad_norm": 5.15625, + "learning_rate": 8.337821656977574e-06, + "loss": 1.6513298, + "memory(GiB)": 111.15, + "step": 23965, + "train_speed(iter/s)": 0.448863 + }, + { + "acc": 0.65554428, + "epoch": 0.6080669710806698, + "grad_norm": 5.0625, + "learning_rate": 8.337040828590715e-06, + "loss": 1.62814636, + "memory(GiB)": 111.15, + "step": 23970, + "train_speed(iter/s)": 0.448912 + }, + { + "acc": 0.64036188, + "epoch": 0.6081938102486047, + "grad_norm": 5.1875, + "learning_rate": 8.336259853425901e-06, + "loss": 1.63228836, + "memory(GiB)": 111.15, + "step": 23975, + "train_speed(iter/s)": 0.448962 + }, + { + "acc": 0.64361305, + "epoch": 0.6083206494165398, + "grad_norm": 5.1875, + "learning_rate": 8.335478731517484e-06, + "loss": 1.63008537, + "memory(GiB)": 111.15, + "step": 23980, + "train_speed(iter/s)": 0.449011 + }, + { + "acc": 0.63859744, + "epoch": 0.6084474885844748, + "grad_norm": 5.8125, + "learning_rate": 8.33469746289982e-06, + "loss": 1.66584663, + "memory(GiB)": 111.15, + "step": 23985, + "train_speed(iter/s)": 0.449061 + }, + { + "acc": 0.63884811, + "epoch": 0.6085743277524099, + "grad_norm": 5.46875, + "learning_rate": 8.333916047607274e-06, + "loss": 1.66823769, + "memory(GiB)": 111.15, + "step": 23990, + "train_speed(iter/s)": 0.44911 + }, + { + "acc": 0.65851912, + "epoch": 0.608701166920345, + "grad_norm": 6.65625, + "learning_rate": 8.333134485674214e-06, + "loss": 1.63489323, + "memory(GiB)": 111.15, + "step": 23995, + "train_speed(iter/s)": 0.44916 + }, + { + "acc": 0.63658485, + "epoch": 0.60882800608828, + "grad_norm": 5.34375, + "learning_rate": 8.33235277713502e-06, + "loss": 1.64930916, + "memory(GiB)": 111.15, + "step": 24000, + "train_speed(iter/s)": 0.44921 + }, + { + "epoch": 0.60882800608828, + "eval_acc": 0.6420373341933769, + "eval_loss": 1.6158993244171143, + "eval_runtime": 114.2562, + "eval_samples_per_second": 55.752, + "eval_steps_per_second": 27.876, + "step": 24000 + }, + { + "acc": 0.65650396, + "epoch": 0.6089548452562151, + "grad_norm": 6.8125, + "learning_rate": 8.33157092202407e-06, + "loss": 1.68729725, + "memory(GiB)": 111.15, + "step": 24005, + "train_speed(iter/s)": 0.448263 + }, + { + "acc": 0.64971142, + "epoch": 0.6090816844241502, + "grad_norm": 4.1875, + "learning_rate": 8.33078892037576e-06, + "loss": 1.64300747, + "memory(GiB)": 111.15, + "step": 24010, + "train_speed(iter/s)": 0.448312 + }, + { + "acc": 0.64831991, + "epoch": 0.6092085235920852, + "grad_norm": 6.1875, + "learning_rate": 8.33000677222448e-06, + "loss": 1.65981102, + "memory(GiB)": 111.15, + "step": 24015, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.64808822, + "epoch": 0.6093353627600203, + "grad_norm": 4.78125, + "learning_rate": 8.329224477604635e-06, + "loss": 1.62695045, + "memory(GiB)": 111.15, + "step": 24020, + "train_speed(iter/s)": 0.448411 + }, + { + "acc": 0.64259205, + "epoch": 0.6094622019279553, + "grad_norm": 5.4375, + "learning_rate": 8.328442036550633e-06, + "loss": 1.68496475, + "memory(GiB)": 111.15, + "step": 24025, + "train_speed(iter/s)": 0.448461 + }, + { + "acc": 0.65025859, + "epoch": 0.6095890410958904, + "grad_norm": 5.53125, + "learning_rate": 8.327659449096892e-06, + "loss": 1.5814291, + "memory(GiB)": 111.15, + "step": 24030, + "train_speed(iter/s)": 0.44851 + }, + { + "acc": 0.67203693, + "epoch": 0.6097158802638255, + "grad_norm": 6.0, + "learning_rate": 8.32687671527783e-06, + "loss": 1.56463842, + "memory(GiB)": 111.15, + "step": 24035, + "train_speed(iter/s)": 0.44856 + }, + { + "acc": 0.65068455, + "epoch": 0.6098427194317605, + "grad_norm": 5.34375, + "learning_rate": 8.326093835127878e-06, + "loss": 1.67745209, + "memory(GiB)": 111.15, + "step": 24040, + "train_speed(iter/s)": 0.448609 + }, + { + "acc": 0.64283671, + "epoch": 0.6099695585996956, + "grad_norm": 6.15625, + "learning_rate": 8.325310808681466e-06, + "loss": 1.64439754, + "memory(GiB)": 111.15, + "step": 24045, + "train_speed(iter/s)": 0.448659 + }, + { + "acc": 0.64999614, + "epoch": 0.6100963977676307, + "grad_norm": 5.71875, + "learning_rate": 8.32452763597304e-06, + "loss": 1.57233887, + "memory(GiB)": 111.15, + "step": 24050, + "train_speed(iter/s)": 0.448708 + }, + { + "acc": 0.63814769, + "epoch": 0.6102232369355657, + "grad_norm": 5.40625, + "learning_rate": 8.323744317037048e-06, + "loss": 1.57949219, + "memory(GiB)": 111.15, + "step": 24055, + "train_speed(iter/s)": 0.448758 + }, + { + "acc": 0.64196286, + "epoch": 0.6103500761035008, + "grad_norm": 4.59375, + "learning_rate": 8.322960851907937e-06, + "loss": 1.65290565, + "memory(GiB)": 111.15, + "step": 24060, + "train_speed(iter/s)": 0.448807 + }, + { + "acc": 0.63780303, + "epoch": 0.6104769152714358, + "grad_norm": 6.71875, + "learning_rate": 8.322177240620175e-06, + "loss": 1.68860435, + "memory(GiB)": 111.15, + "step": 24065, + "train_speed(iter/s)": 0.448857 + }, + { + "acc": 0.66063881, + "epoch": 0.6106037544393709, + "grad_norm": 5.59375, + "learning_rate": 8.321393483208224e-06, + "loss": 1.58395901, + "memory(GiB)": 111.15, + "step": 24070, + "train_speed(iter/s)": 0.448906 + }, + { + "acc": 0.65795722, + "epoch": 0.610730593607306, + "grad_norm": 5.21875, + "learning_rate": 8.32060957970656e-06, + "loss": 1.57965889, + "memory(GiB)": 111.15, + "step": 24075, + "train_speed(iter/s)": 0.448956 + }, + { + "acc": 0.64166441, + "epoch": 0.610857432775241, + "grad_norm": 4.90625, + "learning_rate": 8.319825530149661e-06, + "loss": 1.69460754, + "memory(GiB)": 111.15, + "step": 24080, + "train_speed(iter/s)": 0.449004 + }, + { + "acc": 0.64727421, + "epoch": 0.6109842719431761, + "grad_norm": 5.5, + "learning_rate": 8.319041334572012e-06, + "loss": 1.67683792, + "memory(GiB)": 111.15, + "step": 24085, + "train_speed(iter/s)": 0.449054 + }, + { + "acc": 0.64247022, + "epoch": 0.6111111111111112, + "grad_norm": 5.5625, + "learning_rate": 8.318256993008108e-06, + "loss": 1.67788887, + "memory(GiB)": 111.15, + "step": 24090, + "train_speed(iter/s)": 0.449103 + }, + { + "acc": 0.63213425, + "epoch": 0.6112379502790461, + "grad_norm": 5.09375, + "learning_rate": 8.317472505492446e-06, + "loss": 1.71761703, + "memory(GiB)": 111.15, + "step": 24095, + "train_speed(iter/s)": 0.449152 + }, + { + "acc": 0.66539721, + "epoch": 0.6113647894469812, + "grad_norm": 4.9375, + "learning_rate": 8.31668787205953e-06, + "loss": 1.54201641, + "memory(GiB)": 111.15, + "step": 24100, + "train_speed(iter/s)": 0.449202 + }, + { + "epoch": 0.6113647894469812, + "eval_acc": 0.6420168635952136, + "eval_loss": 1.6157805919647217, + "eval_runtime": 113.9775, + "eval_samples_per_second": 55.888, + "eval_steps_per_second": 27.944, + "step": 24100 + }, + { + "acc": 0.66001291, + "epoch": 0.6114916286149162, + "grad_norm": 4.59375, + "learning_rate": 8.315903092743876e-06, + "loss": 1.61317825, + "memory(GiB)": 111.15, + "step": 24105, + "train_speed(iter/s)": 0.448261 + }, + { + "acc": 0.66748753, + "epoch": 0.6116184677828513, + "grad_norm": 5.0625, + "learning_rate": 8.315118167579999e-06, + "loss": 1.62678337, + "memory(GiB)": 111.15, + "step": 24110, + "train_speed(iter/s)": 0.44831 + }, + { + "acc": 0.656071, + "epoch": 0.6117453069507864, + "grad_norm": 5.71875, + "learning_rate": 8.314333096602423e-06, + "loss": 1.58352661, + "memory(GiB)": 111.15, + "step": 24115, + "train_speed(iter/s)": 0.448359 + }, + { + "acc": 0.65408249, + "epoch": 0.6118721461187214, + "grad_norm": 5.3125, + "learning_rate": 8.313547879845682e-06, + "loss": 1.64884853, + "memory(GiB)": 111.15, + "step": 24120, + "train_speed(iter/s)": 0.448409 + }, + { + "acc": 0.63324099, + "epoch": 0.6119989852866565, + "grad_norm": 6.65625, + "learning_rate": 8.312762517344308e-06, + "loss": 1.69926701, + "memory(GiB)": 111.15, + "step": 24125, + "train_speed(iter/s)": 0.448458 + }, + { + "acc": 0.64043751, + "epoch": 0.6121258244545916, + "grad_norm": 5.90625, + "learning_rate": 8.311977009132851e-06, + "loss": 1.69693871, + "memory(GiB)": 111.15, + "step": 24130, + "train_speed(iter/s)": 0.448508 + }, + { + "acc": 0.65016484, + "epoch": 0.6122526636225266, + "grad_norm": 5.03125, + "learning_rate": 8.311191355245858e-06, + "loss": 1.65546303, + "memory(GiB)": 111.15, + "step": 24135, + "train_speed(iter/s)": 0.448557 + }, + { + "acc": 0.66069188, + "epoch": 0.6123795027904617, + "grad_norm": 4.53125, + "learning_rate": 8.310405555717884e-06, + "loss": 1.64879761, + "memory(GiB)": 111.15, + "step": 24140, + "train_speed(iter/s)": 0.448607 + }, + { + "acc": 0.6424634, + "epoch": 0.6125063419583967, + "grad_norm": 4.78125, + "learning_rate": 8.309619610583495e-06, + "loss": 1.68864899, + "memory(GiB)": 111.15, + "step": 24145, + "train_speed(iter/s)": 0.448655 + }, + { + "acc": 0.67724562, + "epoch": 0.6126331811263318, + "grad_norm": 5.125, + "learning_rate": 8.30883351987726e-06, + "loss": 1.58632383, + "memory(GiB)": 111.15, + "step": 24150, + "train_speed(iter/s)": 0.448705 + }, + { + "acc": 0.64657207, + "epoch": 0.6127600202942669, + "grad_norm": 6.875, + "learning_rate": 8.30804728363375e-06, + "loss": 1.6429966, + "memory(GiB)": 111.15, + "step": 24155, + "train_speed(iter/s)": 0.448754 + }, + { + "acc": 0.65886908, + "epoch": 0.6128868594622019, + "grad_norm": 6.90625, + "learning_rate": 8.307260901887556e-06, + "loss": 1.62525978, + "memory(GiB)": 111.15, + "step": 24160, + "train_speed(iter/s)": 0.448803 + }, + { + "acc": 0.65235424, + "epoch": 0.613013698630137, + "grad_norm": 4.71875, + "learning_rate": 8.306474374673259e-06, + "loss": 1.56023836, + "memory(GiB)": 111.15, + "step": 24165, + "train_speed(iter/s)": 0.448853 + }, + { + "acc": 0.65424767, + "epoch": 0.6131405377980721, + "grad_norm": 5.65625, + "learning_rate": 8.305687702025457e-06, + "loss": 1.6178524, + "memory(GiB)": 111.15, + "step": 24170, + "train_speed(iter/s)": 0.448902 + }, + { + "acc": 0.65057411, + "epoch": 0.6132673769660071, + "grad_norm": 5.375, + "learning_rate": 8.304900883978753e-06, + "loss": 1.60652695, + "memory(GiB)": 111.15, + "step": 24175, + "train_speed(iter/s)": 0.448952 + }, + { + "acc": 0.63827314, + "epoch": 0.6133942161339422, + "grad_norm": 5.0, + "learning_rate": 8.304113920567751e-06, + "loss": 1.68926163, + "memory(GiB)": 111.15, + "step": 24180, + "train_speed(iter/s)": 0.449001 + }, + { + "acc": 0.64826517, + "epoch": 0.6135210553018772, + "grad_norm": 4.90625, + "learning_rate": 8.303326811827066e-06, + "loss": 1.62731571, + "memory(GiB)": 111.15, + "step": 24185, + "train_speed(iter/s)": 0.449051 + }, + { + "acc": 0.64603887, + "epoch": 0.6136478944698123, + "grad_norm": 4.5625, + "learning_rate": 8.302539557791322e-06, + "loss": 1.64456406, + "memory(GiB)": 111.15, + "step": 24190, + "train_speed(iter/s)": 0.4491 + }, + { + "acc": 0.6352911, + "epoch": 0.6137747336377474, + "grad_norm": 4.90625, + "learning_rate": 8.301752158495141e-06, + "loss": 1.67170372, + "memory(GiB)": 111.15, + "step": 24195, + "train_speed(iter/s)": 0.449149 + }, + { + "acc": 0.64600053, + "epoch": 0.6139015728056824, + "grad_norm": 5.28125, + "learning_rate": 8.300964613973159e-06, + "loss": 1.64609451, + "memory(GiB)": 111.15, + "step": 24200, + "train_speed(iter/s)": 0.449199 + }, + { + "epoch": 0.6139015728056824, + "eval_acc": 0.6420356631241392, + "eval_loss": 1.6154577732086182, + "eval_runtime": 114.2357, + "eval_samples_per_second": 55.762, + "eval_steps_per_second": 27.881, + "step": 24200 + }, + { + "acc": 0.64543533, + "epoch": 0.6140284119736175, + "grad_norm": 5.0, + "learning_rate": 8.300176924260017e-06, + "loss": 1.62836208, + "memory(GiB)": 111.15, + "step": 24205, + "train_speed(iter/s)": 0.448259 + }, + { + "acc": 0.64406996, + "epoch": 0.6141552511415526, + "grad_norm": 5.25, + "learning_rate": 8.299389089390359e-06, + "loss": 1.64323082, + "memory(GiB)": 111.15, + "step": 24210, + "train_speed(iter/s)": 0.448308 + }, + { + "acc": 0.65264387, + "epoch": 0.6142820903094875, + "grad_norm": 4.625, + "learning_rate": 8.298601109398838e-06, + "loss": 1.61092548, + "memory(GiB)": 111.15, + "step": 24215, + "train_speed(iter/s)": 0.448356 + }, + { + "acc": 0.64476175, + "epoch": 0.6144089294774226, + "grad_norm": 4.34375, + "learning_rate": 8.297812984320113e-06, + "loss": 1.64497414, + "memory(GiB)": 111.15, + "step": 24220, + "train_speed(iter/s)": 0.448405 + }, + { + "acc": 0.64652634, + "epoch": 0.6145357686453576, + "grad_norm": 5.25, + "learning_rate": 8.297024714188851e-06, + "loss": 1.64502296, + "memory(GiB)": 111.15, + "step": 24225, + "train_speed(iter/s)": 0.448453 + }, + { + "acc": 0.64738159, + "epoch": 0.6146626078132927, + "grad_norm": 6.09375, + "learning_rate": 8.296236299039719e-06, + "loss": 1.67708092, + "memory(GiB)": 111.15, + "step": 24230, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.64694719, + "epoch": 0.6147894469812278, + "grad_norm": 5.0, + "learning_rate": 8.295447738907401e-06, + "loss": 1.57938786, + "memory(GiB)": 111.15, + "step": 24235, + "train_speed(iter/s)": 0.44855 + }, + { + "acc": 0.65093474, + "epoch": 0.6149162861491628, + "grad_norm": 4.71875, + "learning_rate": 8.294659033826576e-06, + "loss": 1.659021, + "memory(GiB)": 111.15, + "step": 24240, + "train_speed(iter/s)": 0.448598 + }, + { + "acc": 0.63814721, + "epoch": 0.6150431253170979, + "grad_norm": 5.96875, + "learning_rate": 8.293870183831937e-06, + "loss": 1.66376114, + "memory(GiB)": 111.15, + "step": 24245, + "train_speed(iter/s)": 0.448647 + }, + { + "acc": 0.65392866, + "epoch": 0.615169964485033, + "grad_norm": 5.9375, + "learning_rate": 8.293081188958183e-06, + "loss": 1.58644428, + "memory(GiB)": 111.15, + "step": 24250, + "train_speed(iter/s)": 0.448695 + }, + { + "acc": 0.65939302, + "epoch": 0.615296803652968, + "grad_norm": 5.28125, + "learning_rate": 8.292292049240014e-06, + "loss": 1.6149128, + "memory(GiB)": 111.15, + "step": 24255, + "train_speed(iter/s)": 0.448743 + }, + { + "acc": 0.63794746, + "epoch": 0.6154236428209031, + "grad_norm": 5.1875, + "learning_rate": 8.291502764712143e-06, + "loss": 1.66460171, + "memory(GiB)": 111.15, + "step": 24260, + "train_speed(iter/s)": 0.448792 + }, + { + "acc": 0.65251427, + "epoch": 0.6155504819888381, + "grad_norm": 5.28125, + "learning_rate": 8.290713335409284e-06, + "loss": 1.62063026, + "memory(GiB)": 111.15, + "step": 24265, + "train_speed(iter/s)": 0.448841 + }, + { + "acc": 0.65412669, + "epoch": 0.6156773211567732, + "grad_norm": 4.8125, + "learning_rate": 8.28992376136616e-06, + "loss": 1.60557213, + "memory(GiB)": 111.15, + "step": 24270, + "train_speed(iter/s)": 0.448889 + }, + { + "acc": 0.64345584, + "epoch": 0.6158041603247083, + "grad_norm": 5.15625, + "learning_rate": 8.289134042617502e-06, + "loss": 1.68845482, + "memory(GiB)": 111.15, + "step": 24275, + "train_speed(iter/s)": 0.448937 + }, + { + "acc": 0.65022907, + "epoch": 0.6159309994926433, + "grad_norm": 5.25, + "learning_rate": 8.288344179198043e-06, + "loss": 1.62153625, + "memory(GiB)": 111.15, + "step": 24280, + "train_speed(iter/s)": 0.448985 + }, + { + "acc": 0.64560814, + "epoch": 0.6160578386605784, + "grad_norm": 5.25, + "learning_rate": 8.287554171142525e-06, + "loss": 1.64183178, + "memory(GiB)": 111.15, + "step": 24285, + "train_speed(iter/s)": 0.449034 + }, + { + "acc": 0.63403606, + "epoch": 0.6161846778285135, + "grad_norm": 4.9375, + "learning_rate": 8.2867640184857e-06, + "loss": 1.68773251, + "memory(GiB)": 111.15, + "step": 24290, + "train_speed(iter/s)": 0.449082 + }, + { + "acc": 0.63215141, + "epoch": 0.6163115169964485, + "grad_norm": 5.28125, + "learning_rate": 8.285973721262315e-06, + "loss": 1.6968544, + "memory(GiB)": 111.15, + "step": 24295, + "train_speed(iter/s)": 0.449131 + }, + { + "acc": 0.64754543, + "epoch": 0.6164383561643836, + "grad_norm": 4.875, + "learning_rate": 8.285183279507135e-06, + "loss": 1.61360664, + "memory(GiB)": 111.15, + "step": 24300, + "train_speed(iter/s)": 0.449179 + }, + { + "epoch": 0.6164383561643836, + "eval_acc": 0.6421121145417699, + "eval_loss": 1.615362524986267, + "eval_runtime": 114.7306, + "eval_samples_per_second": 55.521, + "eval_steps_per_second": 27.761, + "step": 24300 + }, + { + "acc": 0.63545208, + "epoch": 0.6165651953323186, + "grad_norm": 5.75, + "learning_rate": 8.28439269325493e-06, + "loss": 1.59305143, + "memory(GiB)": 111.15, + "step": 24305, + "train_speed(iter/s)": 0.448239 + }, + { + "acc": 0.64200306, + "epoch": 0.6166920345002537, + "grad_norm": 5.15625, + "learning_rate": 8.28360196254047e-06, + "loss": 1.70084648, + "memory(GiB)": 111.15, + "step": 24310, + "train_speed(iter/s)": 0.448288 + }, + { + "acc": 0.64844432, + "epoch": 0.6168188736681888, + "grad_norm": 5.75, + "learning_rate": 8.282811087398535e-06, + "loss": 1.63865623, + "memory(GiB)": 111.15, + "step": 24315, + "train_speed(iter/s)": 0.448336 + }, + { + "acc": 0.64279213, + "epoch": 0.6169457128361238, + "grad_norm": 6.3125, + "learning_rate": 8.282020067863911e-06, + "loss": 1.64153843, + "memory(GiB)": 111.15, + "step": 24320, + "train_speed(iter/s)": 0.448385 + }, + { + "acc": 0.66817284, + "epoch": 0.6170725520040589, + "grad_norm": 5.5625, + "learning_rate": 8.281228903971391e-06, + "loss": 1.57785645, + "memory(GiB)": 111.15, + "step": 24325, + "train_speed(iter/s)": 0.448434 + }, + { + "acc": 0.65765181, + "epoch": 0.617199391171994, + "grad_norm": 7.03125, + "learning_rate": 8.280437595755774e-06, + "loss": 1.53902664, + "memory(GiB)": 111.15, + "step": 24330, + "train_speed(iter/s)": 0.448483 + }, + { + "acc": 0.64277325, + "epoch": 0.617326230339929, + "grad_norm": 5.21875, + "learning_rate": 8.279646143251867e-06, + "loss": 1.55683289, + "memory(GiB)": 111.15, + "step": 24335, + "train_speed(iter/s)": 0.448532 + }, + { + "acc": 0.65317426, + "epoch": 0.617453069507864, + "grad_norm": 5.4375, + "learning_rate": 8.278854546494479e-06, + "loss": 1.65138187, + "memory(GiB)": 111.15, + "step": 24340, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.64930458, + "epoch": 0.617579908675799, + "grad_norm": 5.625, + "learning_rate": 8.27806280551843e-06, + "loss": 1.56441154, + "memory(GiB)": 111.15, + "step": 24345, + "train_speed(iter/s)": 0.448629 + }, + { + "acc": 0.64230733, + "epoch": 0.6177067478437341, + "grad_norm": 5.53125, + "learning_rate": 8.277270920358542e-06, + "loss": 1.61448326, + "memory(GiB)": 111.15, + "step": 24350, + "train_speed(iter/s)": 0.448678 + }, + { + "acc": 0.67030349, + "epoch": 0.6178335870116692, + "grad_norm": 5.0625, + "learning_rate": 8.276478891049649e-06, + "loss": 1.62374649, + "memory(GiB)": 111.15, + "step": 24355, + "train_speed(iter/s)": 0.448727 + }, + { + "acc": 0.66116638, + "epoch": 0.6179604261796042, + "grad_norm": 5.625, + "learning_rate": 8.275686717626584e-06, + "loss": 1.57728577, + "memory(GiB)": 111.15, + "step": 24360, + "train_speed(iter/s)": 0.448775 + }, + { + "acc": 0.65010004, + "epoch": 0.6180872653475393, + "grad_norm": 4.34375, + "learning_rate": 8.274894400124191e-06, + "loss": 1.61490211, + "memory(GiB)": 111.15, + "step": 24365, + "train_speed(iter/s)": 0.448824 + }, + { + "acc": 0.64460096, + "epoch": 0.6182141045154744, + "grad_norm": 5.25, + "learning_rate": 8.274101938577324e-06, + "loss": 1.6625782, + "memory(GiB)": 111.15, + "step": 24370, + "train_speed(iter/s)": 0.448873 + }, + { + "acc": 0.6566783, + "epoch": 0.6183409436834094, + "grad_norm": 6.65625, + "learning_rate": 8.273309333020834e-06, + "loss": 1.58700638, + "memory(GiB)": 111.15, + "step": 24375, + "train_speed(iter/s)": 0.448921 + }, + { + "acc": 0.65058284, + "epoch": 0.6184677828513445, + "grad_norm": 4.6875, + "learning_rate": 8.272516583489587e-06, + "loss": 1.6653183, + "memory(GiB)": 111.15, + "step": 24380, + "train_speed(iter/s)": 0.44897 + }, + { + "acc": 0.65957437, + "epoch": 0.6185946220192795, + "grad_norm": 5.5625, + "learning_rate": 8.271723690018448e-06, + "loss": 1.58668842, + "memory(GiB)": 111.15, + "step": 24385, + "train_speed(iter/s)": 0.449019 + }, + { + "acc": 0.63386369, + "epoch": 0.6187214611872146, + "grad_norm": 5.125, + "learning_rate": 8.270930652642295e-06, + "loss": 1.62891521, + "memory(GiB)": 111.15, + "step": 24390, + "train_speed(iter/s)": 0.449067 + }, + { + "acc": 0.63957849, + "epoch": 0.6188483003551497, + "grad_norm": 5.09375, + "learning_rate": 8.270137471396007e-06, + "loss": 1.65154495, + "memory(GiB)": 111.15, + "step": 24395, + "train_speed(iter/s)": 0.449116 + }, + { + "acc": 0.65384665, + "epoch": 0.6189751395230847, + "grad_norm": 4.46875, + "learning_rate": 8.269344146314475e-06, + "loss": 1.62977219, + "memory(GiB)": 111.15, + "step": 24400, + "train_speed(iter/s)": 0.449165 + }, + { + "epoch": 0.6189751395230847, + "eval_acc": 0.6421284074668386, + "eval_loss": 1.6152629852294922, + "eval_runtime": 113.4468, + "eval_samples_per_second": 56.15, + "eval_steps_per_second": 28.075, + "step": 24400 + }, + { + "acc": 0.64319849, + "epoch": 0.6191019786910198, + "grad_norm": 5.53125, + "learning_rate": 8.26855067743259e-06, + "loss": 1.68319054, + "memory(GiB)": 111.15, + "step": 24405, + "train_speed(iter/s)": 0.44824 + }, + { + "acc": 0.65194054, + "epoch": 0.6192288178589549, + "grad_norm": 5.0625, + "learning_rate": 8.267757064785254e-06, + "loss": 1.62091312, + "memory(GiB)": 111.15, + "step": 24410, + "train_speed(iter/s)": 0.448289 + }, + { + "acc": 0.65029297, + "epoch": 0.6193556570268899, + "grad_norm": 5.78125, + "learning_rate": 8.26696330840737e-06, + "loss": 1.63800316, + "memory(GiB)": 111.15, + "step": 24415, + "train_speed(iter/s)": 0.448338 + }, + { + "acc": 0.66730137, + "epoch": 0.619482496194825, + "grad_norm": 5.21875, + "learning_rate": 8.266169408333856e-06, + "loss": 1.58976936, + "memory(GiB)": 111.15, + "step": 24420, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.65076752, + "epoch": 0.61960933536276, + "grad_norm": 5.375, + "learning_rate": 8.265375364599629e-06, + "loss": 1.62822857, + "memory(GiB)": 111.15, + "step": 24425, + "train_speed(iter/s)": 0.448436 + }, + { + "acc": 0.65382295, + "epoch": 0.6197361745306951, + "grad_norm": 6.9375, + "learning_rate": 8.264581177239615e-06, + "loss": 1.6179678, + "memory(GiB)": 111.15, + "step": 24430, + "train_speed(iter/s)": 0.448486 + }, + { + "acc": 0.66210556, + "epoch": 0.6198630136986302, + "grad_norm": 5.65625, + "learning_rate": 8.263786846288745e-06, + "loss": 1.5630352, + "memory(GiB)": 111.15, + "step": 24435, + "train_speed(iter/s)": 0.448535 + }, + { + "acc": 0.65581779, + "epoch": 0.6199898528665652, + "grad_norm": 6.21875, + "learning_rate": 8.262992371781956e-06, + "loss": 1.60925026, + "memory(GiB)": 111.15, + "step": 24440, + "train_speed(iter/s)": 0.448585 + }, + { + "acc": 0.62915201, + "epoch": 0.6201166920345003, + "grad_norm": 4.96875, + "learning_rate": 8.262197753754195e-06, + "loss": 1.69937286, + "memory(GiB)": 111.15, + "step": 24445, + "train_speed(iter/s)": 0.448635 + }, + { + "acc": 0.6406374, + "epoch": 0.6202435312024354, + "grad_norm": 5.65625, + "learning_rate": 8.261402992240414e-06, + "loss": 1.69906044, + "memory(GiB)": 111.15, + "step": 24450, + "train_speed(iter/s)": 0.448684 + }, + { + "acc": 0.64919839, + "epoch": 0.6203703703703703, + "grad_norm": 6.28125, + "learning_rate": 8.260608087275566e-06, + "loss": 1.67895451, + "memory(GiB)": 111.15, + "step": 24455, + "train_speed(iter/s)": 0.448733 + }, + { + "acc": 0.65254116, + "epoch": 0.6204972095383054, + "grad_norm": 4.78125, + "learning_rate": 8.259813038894617e-06, + "loss": 1.65591583, + "memory(GiB)": 111.15, + "step": 24460, + "train_speed(iter/s)": 0.448782 + }, + { + "acc": 0.65182519, + "epoch": 0.6206240487062404, + "grad_norm": 4.90625, + "learning_rate": 8.259017847132538e-06, + "loss": 1.60837059, + "memory(GiB)": 111.15, + "step": 24465, + "train_speed(iter/s)": 0.448831 + }, + { + "acc": 0.64760332, + "epoch": 0.6207508878741755, + "grad_norm": 5.875, + "learning_rate": 8.258222512024303e-06, + "loss": 1.68212013, + "memory(GiB)": 111.15, + "step": 24470, + "train_speed(iter/s)": 0.448881 + }, + { + "acc": 0.65729432, + "epoch": 0.6208777270421106, + "grad_norm": 4.625, + "learning_rate": 8.257427033604894e-06, + "loss": 1.54912157, + "memory(GiB)": 111.15, + "step": 24475, + "train_speed(iter/s)": 0.44893 + }, + { + "acc": 0.63597445, + "epoch": 0.6210045662100456, + "grad_norm": 5.09375, + "learning_rate": 8.256631411909305e-06, + "loss": 1.68532867, + "memory(GiB)": 111.15, + "step": 24480, + "train_speed(iter/s)": 0.448979 + }, + { + "acc": 0.64554701, + "epoch": 0.6211314053779807, + "grad_norm": 5.5, + "learning_rate": 8.25583564697252e-06, + "loss": 1.54473982, + "memory(GiB)": 111.15, + "step": 24485, + "train_speed(iter/s)": 0.449027 + }, + { + "acc": 0.65068507, + "epoch": 0.6212582445459158, + "grad_norm": 5.40625, + "learning_rate": 8.255039738829552e-06, + "loss": 1.63337288, + "memory(GiB)": 111.15, + "step": 24490, + "train_speed(iter/s)": 0.449076 + }, + { + "acc": 0.64778237, + "epoch": 0.6213850837138508, + "grad_norm": 5.5625, + "learning_rate": 8.254243687515402e-06, + "loss": 1.61975174, + "memory(GiB)": 111.15, + "step": 24495, + "train_speed(iter/s)": 0.449125 + }, + { + "acc": 0.64725237, + "epoch": 0.6215119228817859, + "grad_norm": 5.65625, + "learning_rate": 8.253447493065085e-06, + "loss": 1.65691872, + "memory(GiB)": 111.15, + "step": 24500, + "train_speed(iter/s)": 0.449174 + }, + { + "epoch": 0.6215119228817859, + "eval_acc": 0.6420816175281795, + "eval_loss": 1.615227460861206, + "eval_runtime": 112.0447, + "eval_samples_per_second": 56.852, + "eval_steps_per_second": 28.426, + "step": 24500 + }, + { + "acc": 0.64350023, + "epoch": 0.6216387620497209, + "grad_norm": 5.625, + "learning_rate": 8.252651155513622e-06, + "loss": 1.65579357, + "memory(GiB)": 111.15, + "step": 24505, + "train_speed(iter/s)": 0.448265 + }, + { + "acc": 0.63125229, + "epoch": 0.621765601217656, + "grad_norm": 5.3125, + "learning_rate": 8.251854674896039e-06, + "loss": 1.72449112, + "memory(GiB)": 111.15, + "step": 24510, + "train_speed(iter/s)": 0.448313 + }, + { + "acc": 0.64305048, + "epoch": 0.6218924403855911, + "grad_norm": 5.3125, + "learning_rate": 8.251058051247368e-06, + "loss": 1.65257721, + "memory(GiB)": 111.15, + "step": 24515, + "train_speed(iter/s)": 0.448361 + }, + { + "acc": 0.65995722, + "epoch": 0.6220192795535261, + "grad_norm": 5.375, + "learning_rate": 8.250261284602651e-06, + "loss": 1.59693956, + "memory(GiB)": 111.15, + "step": 24520, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.65359645, + "epoch": 0.6221461187214612, + "grad_norm": 5.21875, + "learning_rate": 8.249464374996932e-06, + "loss": 1.63125591, + "memory(GiB)": 111.15, + "step": 24525, + "train_speed(iter/s)": 0.448458 + }, + { + "acc": 0.64554834, + "epoch": 0.6222729578893963, + "grad_norm": 5.3125, + "learning_rate": 8.24866732246526e-06, + "loss": 1.66055832, + "memory(GiB)": 111.15, + "step": 24530, + "train_speed(iter/s)": 0.448507 + }, + { + "acc": 0.64384832, + "epoch": 0.6223997970573313, + "grad_norm": 6.1875, + "learning_rate": 8.247870127042695e-06, + "loss": 1.62322044, + "memory(GiB)": 111.15, + "step": 24535, + "train_speed(iter/s)": 0.448556 + }, + { + "acc": 0.63472824, + "epoch": 0.6225266362252664, + "grad_norm": 5.875, + "learning_rate": 8.247072788764302e-06, + "loss": 1.63904152, + "memory(GiB)": 111.15, + "step": 24540, + "train_speed(iter/s)": 0.448605 + }, + { + "acc": 0.63686657, + "epoch": 0.6226534753932014, + "grad_norm": 4.96875, + "learning_rate": 8.246275307665147e-06, + "loss": 1.67584839, + "memory(GiB)": 111.15, + "step": 24545, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.65850101, + "epoch": 0.6227803145611365, + "grad_norm": 5.40625, + "learning_rate": 8.245477683780316e-06, + "loss": 1.55479746, + "memory(GiB)": 111.15, + "step": 24550, + "train_speed(iter/s)": 0.448702 + }, + { + "acc": 0.65098324, + "epoch": 0.6229071537290716, + "grad_norm": 5.3125, + "learning_rate": 8.244679917144883e-06, + "loss": 1.63133011, + "memory(GiB)": 111.15, + "step": 24555, + "train_speed(iter/s)": 0.448751 + }, + { + "acc": 0.63788128, + "epoch": 0.6230339928970066, + "grad_norm": 5.59375, + "learning_rate": 8.243882007793941e-06, + "loss": 1.71023464, + "memory(GiB)": 111.15, + "step": 24560, + "train_speed(iter/s)": 0.448799 + }, + { + "acc": 0.65203543, + "epoch": 0.6231608320649417, + "grad_norm": 5.53125, + "learning_rate": 8.243083955762588e-06, + "loss": 1.62124748, + "memory(GiB)": 111.15, + "step": 24565, + "train_speed(iter/s)": 0.448848 + }, + { + "acc": 0.65106983, + "epoch": 0.6232876712328768, + "grad_norm": 5.875, + "learning_rate": 8.24228576108592e-06, + "loss": 1.63046703, + "memory(GiB)": 111.15, + "step": 24570, + "train_speed(iter/s)": 0.448897 + }, + { + "acc": 0.65519886, + "epoch": 0.6234145104008117, + "grad_norm": 5.875, + "learning_rate": 8.24148742379905e-06, + "loss": 1.56613674, + "memory(GiB)": 111.15, + "step": 24575, + "train_speed(iter/s)": 0.448946 + }, + { + "acc": 0.64988289, + "epoch": 0.6235413495687468, + "grad_norm": 5.1875, + "learning_rate": 8.240688943937092e-06, + "loss": 1.63868408, + "memory(GiB)": 111.15, + "step": 24580, + "train_speed(iter/s)": 0.448994 + }, + { + "acc": 0.65359159, + "epoch": 0.6236681887366818, + "grad_norm": 5.875, + "learning_rate": 8.239890321535163e-06, + "loss": 1.63639545, + "memory(GiB)": 111.15, + "step": 24585, + "train_speed(iter/s)": 0.449043 + }, + { + "acc": 0.64441786, + "epoch": 0.6237950279046169, + "grad_norm": 5.5, + "learning_rate": 8.239091556628395e-06, + "loss": 1.66438198, + "memory(GiB)": 111.15, + "step": 24590, + "train_speed(iter/s)": 0.449091 + }, + { + "acc": 0.64209185, + "epoch": 0.623921867072552, + "grad_norm": 5.5, + "learning_rate": 8.238292649251918e-06, + "loss": 1.66698036, + "memory(GiB)": 111.15, + "step": 24595, + "train_speed(iter/s)": 0.44914 + }, + { + "acc": 0.65982246, + "epoch": 0.624048706240487, + "grad_norm": 4.6875, + "learning_rate": 8.237493599440871e-06, + "loss": 1.60147877, + "memory(GiB)": 111.15, + "step": 24600, + "train_speed(iter/s)": 0.449189 + }, + { + "epoch": 0.624048706240487, + "eval_acc": 0.6421455359265265, + "eval_loss": 1.6150586605072021, + "eval_runtime": 114.561, + "eval_samples_per_second": 55.604, + "eval_steps_per_second": 27.802, + "step": 24600 + }, + { + "acc": 0.63105469, + "epoch": 0.6241755454084221, + "grad_norm": 5.4375, + "learning_rate": 8.236694407230402e-06, + "loss": 1.75528755, + "memory(GiB)": 111.15, + "step": 24605, + "train_speed(iter/s)": 0.448262 + }, + { + "acc": 0.66373043, + "epoch": 0.6243023845763572, + "grad_norm": 5.3125, + "learning_rate": 8.235895072655664e-06, + "loss": 1.52104988, + "memory(GiB)": 111.15, + "step": 24610, + "train_speed(iter/s)": 0.44831 + }, + { + "acc": 0.64703484, + "epoch": 0.6244292237442922, + "grad_norm": 5.78125, + "learning_rate": 8.235095595751809e-06, + "loss": 1.66627674, + "memory(GiB)": 111.15, + "step": 24615, + "train_speed(iter/s)": 0.448359 + }, + { + "acc": 0.66195703, + "epoch": 0.6245560629122273, + "grad_norm": 5.40625, + "learning_rate": 8.23429597655401e-06, + "loss": 1.58856335, + "memory(GiB)": 111.15, + "step": 24620, + "train_speed(iter/s)": 0.448406 + }, + { + "acc": 0.62617679, + "epoch": 0.6246829020801623, + "grad_norm": 5.59375, + "learning_rate": 8.233496215097433e-06, + "loss": 1.6893734, + "memory(GiB)": 111.15, + "step": 24625, + "train_speed(iter/s)": 0.448454 + }, + { + "acc": 0.64955873, + "epoch": 0.6248097412480974, + "grad_norm": 5.375, + "learning_rate": 8.232696311417256e-06, + "loss": 1.58007202, + "memory(GiB)": 111.15, + "step": 24630, + "train_speed(iter/s)": 0.448502 + }, + { + "acc": 0.64064445, + "epoch": 0.6249365804160325, + "grad_norm": 6.1875, + "learning_rate": 8.231896265548662e-06, + "loss": 1.67542267, + "memory(GiB)": 111.15, + "step": 24635, + "train_speed(iter/s)": 0.448549 + }, + { + "acc": 0.64979734, + "epoch": 0.6250634195839675, + "grad_norm": 5.09375, + "learning_rate": 8.231096077526841e-06, + "loss": 1.66190338, + "memory(GiB)": 111.15, + "step": 24640, + "train_speed(iter/s)": 0.448597 + }, + { + "acc": 0.6658618, + "epoch": 0.6251902587519026, + "grad_norm": 5.125, + "learning_rate": 8.230295747386988e-06, + "loss": 1.6023468, + "memory(GiB)": 111.15, + "step": 24645, + "train_speed(iter/s)": 0.448645 + }, + { + "acc": 0.65771971, + "epoch": 0.6253170979198377, + "grad_norm": 5.375, + "learning_rate": 8.229495275164307e-06, + "loss": 1.54776373, + "memory(GiB)": 111.15, + "step": 24650, + "train_speed(iter/s)": 0.448693 + }, + { + "acc": 0.65152493, + "epoch": 0.6254439370877727, + "grad_norm": 5.875, + "learning_rate": 8.228694660894003e-06, + "loss": 1.68634052, + "memory(GiB)": 111.15, + "step": 24655, + "train_speed(iter/s)": 0.448741 + }, + { + "acc": 0.65895205, + "epoch": 0.6255707762557078, + "grad_norm": 5.4375, + "learning_rate": 8.227893904611295e-06, + "loss": 1.57311916, + "memory(GiB)": 111.15, + "step": 24660, + "train_speed(iter/s)": 0.448789 + }, + { + "acc": 0.65322676, + "epoch": 0.6256976154236428, + "grad_norm": 4.65625, + "learning_rate": 8.2270930063514e-06, + "loss": 1.60488739, + "memory(GiB)": 111.15, + "step": 24665, + "train_speed(iter/s)": 0.448837 + }, + { + "acc": 0.64790897, + "epoch": 0.6258244545915779, + "grad_norm": 5.3125, + "learning_rate": 8.226291966149549e-06, + "loss": 1.69035683, + "memory(GiB)": 111.15, + "step": 24670, + "train_speed(iter/s)": 0.448885 + }, + { + "acc": 0.65753794, + "epoch": 0.625951293759513, + "grad_norm": 5.21875, + "learning_rate": 8.225490784040971e-06, + "loss": 1.57754707, + "memory(GiB)": 111.15, + "step": 24675, + "train_speed(iter/s)": 0.448932 + }, + { + "acc": 0.6441834, + "epoch": 0.626078132927448, + "grad_norm": 5.9375, + "learning_rate": 8.224689460060908e-06, + "loss": 1.65963898, + "memory(GiB)": 111.15, + "step": 24680, + "train_speed(iter/s)": 0.44898 + }, + { + "acc": 0.66411753, + "epoch": 0.6262049720953831, + "grad_norm": 5.15625, + "learning_rate": 8.223887994244604e-06, + "loss": 1.58518963, + "memory(GiB)": 111.15, + "step": 24685, + "train_speed(iter/s)": 0.449028 + }, + { + "acc": 0.65664158, + "epoch": 0.6263318112633182, + "grad_norm": 5.21875, + "learning_rate": 8.223086386627314e-06, + "loss": 1.59518318, + "memory(GiB)": 111.15, + "step": 24690, + "train_speed(iter/s)": 0.449076 + }, + { + "acc": 0.66993484, + "epoch": 0.6264586504312532, + "grad_norm": 5.5625, + "learning_rate": 8.222284637244296e-06, + "loss": 1.57162247, + "memory(GiB)": 111.15, + "step": 24695, + "train_speed(iter/s)": 0.449124 + }, + { + "acc": 0.65790186, + "epoch": 0.6265854895991883, + "grad_norm": 4.5, + "learning_rate": 8.221482746130811e-06, + "loss": 1.5776968, + "memory(GiB)": 111.15, + "step": 24700, + "train_speed(iter/s)": 0.449173 + }, + { + "epoch": 0.6265854895991883, + "eval_acc": 0.6421304963033859, + "eval_loss": 1.614924669265747, + "eval_runtime": 114.1418, + "eval_samples_per_second": 55.808, + "eval_steps_per_second": 27.904, + "step": 24700 + }, + { + "acc": 0.64822178, + "epoch": 0.6267123287671232, + "grad_norm": 5.3125, + "learning_rate": 8.220680713322131e-06, + "loss": 1.64028664, + "memory(GiB)": 111.15, + "step": 24705, + "train_speed(iter/s)": 0.448253 + }, + { + "acc": 0.64937296, + "epoch": 0.6268391679350583, + "grad_norm": 4.90625, + "learning_rate": 8.219878538853537e-06, + "loss": 1.64805965, + "memory(GiB)": 111.15, + "step": 24710, + "train_speed(iter/s)": 0.4483 + }, + { + "acc": 0.63956256, + "epoch": 0.6269660071029934, + "grad_norm": 4.75, + "learning_rate": 8.219076222760307e-06, + "loss": 1.68880081, + "memory(GiB)": 111.15, + "step": 24715, + "train_speed(iter/s)": 0.448347 + }, + { + "acc": 0.63150253, + "epoch": 0.6270928462709284, + "grad_norm": 4.78125, + "learning_rate": 8.218273765077734e-06, + "loss": 1.64123192, + "memory(GiB)": 111.15, + "step": 24720, + "train_speed(iter/s)": 0.448394 + }, + { + "acc": 0.64997406, + "epoch": 0.6272196854388635, + "grad_norm": 5.125, + "learning_rate": 8.21747116584111e-06, + "loss": 1.61530857, + "memory(GiB)": 111.15, + "step": 24725, + "train_speed(iter/s)": 0.448442 + }, + { + "acc": 0.6392477, + "epoch": 0.6273465246067986, + "grad_norm": 4.75, + "learning_rate": 8.21666842508574e-06, + "loss": 1.59267149, + "memory(GiB)": 111.15, + "step": 24730, + "train_speed(iter/s)": 0.448489 + }, + { + "acc": 0.64898, + "epoch": 0.6274733637747336, + "grad_norm": 6.1875, + "learning_rate": 8.215865542846932e-06, + "loss": 1.72334042, + "memory(GiB)": 111.15, + "step": 24735, + "train_speed(iter/s)": 0.448537 + }, + { + "acc": 0.63713641, + "epoch": 0.6276002029426687, + "grad_norm": 5.46875, + "learning_rate": 8.215062519160002e-06, + "loss": 1.66728516, + "memory(GiB)": 111.15, + "step": 24740, + "train_speed(iter/s)": 0.448585 + }, + { + "acc": 0.65354347, + "epoch": 0.6277270421106037, + "grad_norm": 5.46875, + "learning_rate": 8.214259354060263e-06, + "loss": 1.65049953, + "memory(GiB)": 111.15, + "step": 24745, + "train_speed(iter/s)": 0.448632 + }, + { + "acc": 0.64345884, + "epoch": 0.6278538812785388, + "grad_norm": 6.28125, + "learning_rate": 8.21345604758305e-06, + "loss": 1.6189003, + "memory(GiB)": 111.15, + "step": 24750, + "train_speed(iter/s)": 0.448679 + }, + { + "acc": 0.65353503, + "epoch": 0.6279807204464739, + "grad_norm": 4.71875, + "learning_rate": 8.212652599763693e-06, + "loss": 1.59570761, + "memory(GiB)": 111.15, + "step": 24755, + "train_speed(iter/s)": 0.448727 + }, + { + "acc": 0.64034443, + "epoch": 0.6281075596144089, + "grad_norm": 6.96875, + "learning_rate": 8.211849010637532e-06, + "loss": 1.67312393, + "memory(GiB)": 111.15, + "step": 24760, + "train_speed(iter/s)": 0.448775 + }, + { + "acc": 0.63686047, + "epoch": 0.628234398782344, + "grad_norm": 5.1875, + "learning_rate": 8.211045280239908e-06, + "loss": 1.60019341, + "memory(GiB)": 111.15, + "step": 24765, + "train_speed(iter/s)": 0.448822 + }, + { + "acc": 0.64352894, + "epoch": 0.6283612379502791, + "grad_norm": 5.40625, + "learning_rate": 8.210241408606182e-06, + "loss": 1.58856125, + "memory(GiB)": 111.15, + "step": 24770, + "train_speed(iter/s)": 0.44887 + }, + { + "acc": 0.65654764, + "epoch": 0.6284880771182141, + "grad_norm": 5.65625, + "learning_rate": 8.2094373957717e-06, + "loss": 1.65181389, + "memory(GiB)": 111.15, + "step": 24775, + "train_speed(iter/s)": 0.448917 + }, + { + "acc": 0.64211187, + "epoch": 0.6286149162861492, + "grad_norm": 5.34375, + "learning_rate": 8.208633241771836e-06, + "loss": 1.65398216, + "memory(GiB)": 111.15, + "step": 24780, + "train_speed(iter/s)": 0.448965 + }, + { + "acc": 0.65252814, + "epoch": 0.6287417554540842, + "grad_norm": 5.28125, + "learning_rate": 8.207828946641956e-06, + "loss": 1.63055496, + "memory(GiB)": 111.15, + "step": 24785, + "train_speed(iter/s)": 0.449012 + }, + { + "acc": 0.66257081, + "epoch": 0.6288685946220193, + "grad_norm": 4.53125, + "learning_rate": 8.207024510417436e-06, + "loss": 1.62453251, + "memory(GiB)": 111.15, + "step": 24790, + "train_speed(iter/s)": 0.44906 + }, + { + "acc": 0.64790735, + "epoch": 0.6289954337899544, + "grad_norm": 5.90625, + "learning_rate": 8.20621993313366e-06, + "loss": 1.66902485, + "memory(GiB)": 111.15, + "step": 24795, + "train_speed(iter/s)": 0.449107 + }, + { + "acc": 0.65533366, + "epoch": 0.6291222729578894, + "grad_norm": 5.21875, + "learning_rate": 8.205415214826018e-06, + "loss": 1.76335373, + "memory(GiB)": 111.15, + "step": 24800, + "train_speed(iter/s)": 0.449155 + }, + { + "epoch": 0.6291222729578894, + "eval_acc": 0.6421630821535237, + "eval_loss": 1.6146864891052246, + "eval_runtime": 113.0313, + "eval_samples_per_second": 56.356, + "eval_steps_per_second": 28.178, + "step": 24800 + }, + { + "acc": 0.64455194, + "epoch": 0.6292491121258245, + "grad_norm": 4.90625, + "learning_rate": 8.204610355529901e-06, + "loss": 1.64725723, + "memory(GiB)": 111.15, + "step": 24805, + "train_speed(iter/s)": 0.448248 + }, + { + "acc": 0.6361043, + "epoch": 0.6293759512937596, + "grad_norm": 4.84375, + "learning_rate": 8.203805355280715e-06, + "loss": 1.64709473, + "memory(GiB)": 111.15, + "step": 24810, + "train_speed(iter/s)": 0.448296 + }, + { + "acc": 0.64780111, + "epoch": 0.6295027904616946, + "grad_norm": 4.8125, + "learning_rate": 8.203000214113865e-06, + "loss": 1.68123569, + "memory(GiB)": 111.15, + "step": 24815, + "train_speed(iter/s)": 0.448343 + }, + { + "acc": 0.65505977, + "epoch": 0.6296296296296297, + "grad_norm": 5.03125, + "learning_rate": 8.202194932064767e-06, + "loss": 1.61248169, + "memory(GiB)": 111.15, + "step": 24820, + "train_speed(iter/s)": 0.448391 + }, + { + "acc": 0.64617968, + "epoch": 0.6297564687975646, + "grad_norm": 6.09375, + "learning_rate": 8.201389509168836e-06, + "loss": 1.66658134, + "memory(GiB)": 111.15, + "step": 24825, + "train_speed(iter/s)": 0.448438 + }, + { + "acc": 0.62642918, + "epoch": 0.6298833079654997, + "grad_norm": 6.09375, + "learning_rate": 8.200583945461502e-06, + "loss": 1.6690279, + "memory(GiB)": 111.15, + "step": 24830, + "train_speed(iter/s)": 0.448486 + }, + { + "acc": 0.65200763, + "epoch": 0.6300101471334348, + "grad_norm": 5.40625, + "learning_rate": 8.199778240978197e-06, + "loss": 1.58610277, + "memory(GiB)": 111.15, + "step": 24835, + "train_speed(iter/s)": 0.448533 + }, + { + "acc": 0.62820039, + "epoch": 0.6301369863013698, + "grad_norm": 5.8125, + "learning_rate": 8.19897239575436e-06, + "loss": 1.69679184, + "memory(GiB)": 111.15, + "step": 24840, + "train_speed(iter/s)": 0.448582 + }, + { + "acc": 0.63861303, + "epoch": 0.6302638254693049, + "grad_norm": 4.59375, + "learning_rate": 8.198166409825434e-06, + "loss": 1.69984512, + "memory(GiB)": 111.15, + "step": 24845, + "train_speed(iter/s)": 0.44863 + }, + { + "acc": 0.64915924, + "epoch": 0.63039066463724, + "grad_norm": 6.0625, + "learning_rate": 8.19736028322687e-06, + "loss": 1.61700306, + "memory(GiB)": 111.15, + "step": 24850, + "train_speed(iter/s)": 0.448678 + }, + { + "acc": 0.64389763, + "epoch": 0.630517503805175, + "grad_norm": 5.0, + "learning_rate": 8.196554015994126e-06, + "loss": 1.65045681, + "memory(GiB)": 111.15, + "step": 24855, + "train_speed(iter/s)": 0.448726 + }, + { + "acc": 0.65448847, + "epoch": 0.6306443429731101, + "grad_norm": 5.75, + "learning_rate": 8.195747608162665e-06, + "loss": 1.64016132, + "memory(GiB)": 111.15, + "step": 24860, + "train_speed(iter/s)": 0.448774 + }, + { + "acc": 0.64733167, + "epoch": 0.6307711821410451, + "grad_norm": 4.8125, + "learning_rate": 8.194941059767957e-06, + "loss": 1.632831, + "memory(GiB)": 111.15, + "step": 24865, + "train_speed(iter/s)": 0.448822 + }, + { + "acc": 0.63943119, + "epoch": 0.6308980213089802, + "grad_norm": 5.46875, + "learning_rate": 8.194134370845474e-06, + "loss": 1.63394527, + "memory(GiB)": 111.15, + "step": 24870, + "train_speed(iter/s)": 0.448871 + }, + { + "acc": 0.6489212, + "epoch": 0.6310248604769153, + "grad_norm": 6.15625, + "learning_rate": 8.193327541430703e-06, + "loss": 1.70312233, + "memory(GiB)": 111.15, + "step": 24875, + "train_speed(iter/s)": 0.448918 + }, + { + "acc": 0.64567485, + "epoch": 0.6311516996448503, + "grad_norm": 5.3125, + "learning_rate": 8.192520571559128e-06, + "loss": 1.64940224, + "memory(GiB)": 111.15, + "step": 24880, + "train_speed(iter/s)": 0.448966 + }, + { + "acc": 0.65717254, + "epoch": 0.6312785388127854, + "grad_norm": 6.03125, + "learning_rate": 8.191713461266246e-06, + "loss": 1.65217781, + "memory(GiB)": 111.15, + "step": 24885, + "train_speed(iter/s)": 0.449014 + }, + { + "acc": 0.64579906, + "epoch": 0.6314053779807205, + "grad_norm": 5.28125, + "learning_rate": 8.190906210587555e-06, + "loss": 1.65746593, + "memory(GiB)": 111.15, + "step": 24890, + "train_speed(iter/s)": 0.449062 + }, + { + "acc": 0.65276146, + "epoch": 0.6315322171486555, + "grad_norm": 4.96875, + "learning_rate": 8.190098819558562e-06, + "loss": 1.57652779, + "memory(GiB)": 111.15, + "step": 24895, + "train_speed(iter/s)": 0.44911 + }, + { + "acc": 0.64430218, + "epoch": 0.6316590563165906, + "grad_norm": 4.5, + "learning_rate": 8.189291288214782e-06, + "loss": 1.6269062, + "memory(GiB)": 111.15, + "step": 24900, + "train_speed(iter/s)": 0.449158 + }, + { + "epoch": 0.6316590563165906, + "eval_acc": 0.64215054913424, + "eval_loss": 1.6153156757354736, + "eval_runtime": 113.9187, + "eval_samples_per_second": 55.917, + "eval_steps_per_second": 27.959, + "step": 24900 + }, + { + "acc": 0.64274473, + "epoch": 0.6317858954845256, + "grad_norm": 4.3125, + "learning_rate": 8.18848361659173e-06, + "loss": 1.63923054, + "memory(GiB)": 111.15, + "step": 24905, + "train_speed(iter/s)": 0.448247 + }, + { + "acc": 0.66719494, + "epoch": 0.6319127346524607, + "grad_norm": 5.4375, + "learning_rate": 8.187675804724935e-06, + "loss": 1.60533104, + "memory(GiB)": 111.15, + "step": 24910, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.64727278, + "epoch": 0.6320395738203958, + "grad_norm": 5.125, + "learning_rate": 8.186867852649925e-06, + "loss": 1.62624588, + "memory(GiB)": 111.15, + "step": 24915, + "train_speed(iter/s)": 0.448341 + }, + { + "acc": 0.65123062, + "epoch": 0.6321664129883308, + "grad_norm": 5.25, + "learning_rate": 8.186059760402238e-06, + "loss": 1.65888672, + "memory(GiB)": 111.15, + "step": 24920, + "train_speed(iter/s)": 0.448388 + }, + { + "acc": 0.65739236, + "epoch": 0.6322932521562659, + "grad_norm": 5.34375, + "learning_rate": 8.185251528017419e-06, + "loss": 1.62798195, + "memory(GiB)": 111.15, + "step": 24925, + "train_speed(iter/s)": 0.448435 + }, + { + "acc": 0.64424782, + "epoch": 0.632420091324201, + "grad_norm": 5.5625, + "learning_rate": 8.184443155531016e-06, + "loss": 1.63872452, + "memory(GiB)": 111.15, + "step": 24930, + "train_speed(iter/s)": 0.448483 + }, + { + "acc": 0.65128002, + "epoch": 0.632546930492136, + "grad_norm": 5.0625, + "learning_rate": 8.183634642978586e-06, + "loss": 1.65350571, + "memory(GiB)": 111.15, + "step": 24935, + "train_speed(iter/s)": 0.44853 + }, + { + "acc": 0.64950018, + "epoch": 0.632673769660071, + "grad_norm": 6.125, + "learning_rate": 8.18282599039569e-06, + "loss": 1.62527161, + "memory(GiB)": 111.15, + "step": 24940, + "train_speed(iter/s)": 0.448577 + }, + { + "acc": 0.64705515, + "epoch": 0.632800608828006, + "grad_norm": 5.90625, + "learning_rate": 8.182017197817898e-06, + "loss": 1.68726654, + "memory(GiB)": 111.15, + "step": 24945, + "train_speed(iter/s)": 0.448625 + }, + { + "acc": 0.65741043, + "epoch": 0.6329274479959411, + "grad_norm": 5.21875, + "learning_rate": 8.181208265280782e-06, + "loss": 1.62450294, + "memory(GiB)": 111.15, + "step": 24950, + "train_speed(iter/s)": 0.448671 + }, + { + "acc": 0.63850965, + "epoch": 0.6330542871638762, + "grad_norm": 5.28125, + "learning_rate": 8.180399192819923e-06, + "loss": 1.71565781, + "memory(GiB)": 111.15, + "step": 24955, + "train_speed(iter/s)": 0.448718 + }, + { + "acc": 0.64313908, + "epoch": 0.6331811263318112, + "grad_norm": 6.1875, + "learning_rate": 8.17958998047091e-06, + "loss": 1.64225006, + "memory(GiB)": 111.15, + "step": 24960, + "train_speed(iter/s)": 0.448766 + }, + { + "acc": 0.65503531, + "epoch": 0.6333079654997463, + "grad_norm": 6.3125, + "learning_rate": 8.178780628269332e-06, + "loss": 1.57361002, + "memory(GiB)": 111.15, + "step": 24965, + "train_speed(iter/s)": 0.448813 + }, + { + "acc": 0.64624314, + "epoch": 0.6334348046676814, + "grad_norm": 5.03125, + "learning_rate": 8.177971136250788e-06, + "loss": 1.70225239, + "memory(GiB)": 111.15, + "step": 24970, + "train_speed(iter/s)": 0.44886 + }, + { + "acc": 0.64975262, + "epoch": 0.6335616438356164, + "grad_norm": 4.84375, + "learning_rate": 8.177161504450887e-06, + "loss": 1.63090858, + "memory(GiB)": 111.15, + "step": 24975, + "train_speed(iter/s)": 0.448907 + }, + { + "acc": 0.63992052, + "epoch": 0.6336884830035515, + "grad_norm": 5.03125, + "learning_rate": 8.176351732905239e-06, + "loss": 1.71269741, + "memory(GiB)": 111.15, + "step": 24980, + "train_speed(iter/s)": 0.448954 + }, + { + "acc": 0.65621042, + "epoch": 0.6338153221714865, + "grad_norm": 5.5625, + "learning_rate": 8.175541821649459e-06, + "loss": 1.64829102, + "memory(GiB)": 111.15, + "step": 24985, + "train_speed(iter/s)": 0.449001 + }, + { + "acc": 0.67137823, + "epoch": 0.6339421613394216, + "grad_norm": 5.6875, + "learning_rate": 8.174731770719173e-06, + "loss": 1.5600666, + "memory(GiB)": 111.15, + "step": 24990, + "train_speed(iter/s)": 0.449048 + }, + { + "acc": 0.647223, + "epoch": 0.6340690005073567, + "grad_norm": 4.9375, + "learning_rate": 8.173921580150008e-06, + "loss": 1.639254, + "memory(GiB)": 111.15, + "step": 24995, + "train_speed(iter/s)": 0.449096 + }, + { + "acc": 0.64199047, + "epoch": 0.6341958396752917, + "grad_norm": 5.71875, + "learning_rate": 8.173111249977602e-06, + "loss": 1.69705372, + "memory(GiB)": 111.15, + "step": 25000, + "train_speed(iter/s)": 0.449143 + }, + { + "epoch": 0.6341958396752917, + "eval_acc": 0.6420661601377295, + "eval_loss": 1.6147788763046265, + "eval_runtime": 113.9107, + "eval_samples_per_second": 55.921, + "eval_steps_per_second": 27.961, + "step": 25000 + }, + { + "acc": 0.66102285, + "epoch": 0.6343226788432268, + "grad_norm": 5.40625, + "learning_rate": 8.172300780237596e-06, + "loss": 1.56731091, + "memory(GiB)": 111.15, + "step": 25005, + "train_speed(iter/s)": 0.448236 + }, + { + "acc": 0.64904962, + "epoch": 0.6344495180111619, + "grad_norm": 5.5, + "learning_rate": 8.171490170965639e-06, + "loss": 1.59657936, + "memory(GiB)": 111.15, + "step": 25010, + "train_speed(iter/s)": 0.448284 + }, + { + "acc": 0.66357908, + "epoch": 0.6345763571790969, + "grad_norm": 5.875, + "learning_rate": 8.170679422197385e-06, + "loss": 1.55304871, + "memory(GiB)": 111.15, + "step": 25015, + "train_speed(iter/s)": 0.448331 + }, + { + "acc": 0.64722466, + "epoch": 0.634703196347032, + "grad_norm": 5.6875, + "learning_rate": 8.169868533968493e-06, + "loss": 1.62319908, + "memory(GiB)": 111.15, + "step": 25020, + "train_speed(iter/s)": 0.448379 + }, + { + "acc": 0.63708782, + "epoch": 0.634830035514967, + "grad_norm": 4.65625, + "learning_rate": 8.16905750631463e-06, + "loss": 1.66490021, + "memory(GiB)": 111.15, + "step": 25025, + "train_speed(iter/s)": 0.448426 + }, + { + "acc": 0.66393471, + "epoch": 0.6349568746829021, + "grad_norm": 6.34375, + "learning_rate": 8.168246339271471e-06, + "loss": 1.61631336, + "memory(GiB)": 111.15, + "step": 25030, + "train_speed(iter/s)": 0.448473 + }, + { + "acc": 0.63253698, + "epoch": 0.6350837138508372, + "grad_norm": 4.9375, + "learning_rate": 8.16743503287469e-06, + "loss": 1.64849377, + "memory(GiB)": 111.15, + "step": 25035, + "train_speed(iter/s)": 0.44852 + }, + { + "acc": 0.64529724, + "epoch": 0.6352105530187722, + "grad_norm": 5.6875, + "learning_rate": 8.166623587159978e-06, + "loss": 1.57249537, + "memory(GiB)": 111.15, + "step": 25040, + "train_speed(iter/s)": 0.448568 + }, + { + "acc": 0.65153241, + "epoch": 0.6353373921867073, + "grad_norm": 5.59375, + "learning_rate": 8.16581200216302e-06, + "loss": 1.61163254, + "memory(GiB)": 111.15, + "step": 25045, + "train_speed(iter/s)": 0.448615 + }, + { + "acc": 0.64610701, + "epoch": 0.6354642313546424, + "grad_norm": 4.75, + "learning_rate": 8.165000277919517e-06, + "loss": 1.66350708, + "memory(GiB)": 111.15, + "step": 25050, + "train_speed(iter/s)": 0.448663 + }, + { + "acc": 0.65383902, + "epoch": 0.6355910705225774, + "grad_norm": 5.5625, + "learning_rate": 8.16418841446517e-06, + "loss": 1.60010834, + "memory(GiB)": 111.15, + "step": 25055, + "train_speed(iter/s)": 0.44871 + }, + { + "acc": 0.64665141, + "epoch": 0.6357179096905125, + "grad_norm": 4.8125, + "learning_rate": 8.163376411835691e-06, + "loss": 1.5940299, + "memory(GiB)": 111.15, + "step": 25060, + "train_speed(iter/s)": 0.448758 + }, + { + "acc": 0.66407509, + "epoch": 0.6358447488584474, + "grad_norm": 4.5, + "learning_rate": 8.162564270066793e-06, + "loss": 1.5231823, + "memory(GiB)": 111.15, + "step": 25065, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.64939265, + "epoch": 0.6359715880263825, + "grad_norm": 5.6875, + "learning_rate": 8.1617519891942e-06, + "loss": 1.64153442, + "memory(GiB)": 111.15, + "step": 25070, + "train_speed(iter/s)": 0.448853 + }, + { + "acc": 0.6522121, + "epoch": 0.6360984271943176, + "grad_norm": 5.5625, + "learning_rate": 8.160939569253637e-06, + "loss": 1.58080626, + "memory(GiB)": 111.15, + "step": 25075, + "train_speed(iter/s)": 0.448901 + }, + { + "acc": 0.65004058, + "epoch": 0.6362252663622526, + "grad_norm": 5.15625, + "learning_rate": 8.160127010280838e-06, + "loss": 1.63753471, + "memory(GiB)": 111.15, + "step": 25080, + "train_speed(iter/s)": 0.448948 + }, + { + "acc": 0.64070101, + "epoch": 0.6363521055301877, + "grad_norm": 6.5625, + "learning_rate": 8.159314312311546e-06, + "loss": 1.69272919, + "memory(GiB)": 111.15, + "step": 25085, + "train_speed(iter/s)": 0.448996 + }, + { + "acc": 0.63848009, + "epoch": 0.6364789446981228, + "grad_norm": 5.25, + "learning_rate": 8.158501475381505e-06, + "loss": 1.67859955, + "memory(GiB)": 111.15, + "step": 25090, + "train_speed(iter/s)": 0.449043 + }, + { + "acc": 0.64950032, + "epoch": 0.6366057838660578, + "grad_norm": 5.03125, + "learning_rate": 8.157688499526466e-06, + "loss": 1.56578875, + "memory(GiB)": 111.15, + "step": 25095, + "train_speed(iter/s)": 0.44909 + }, + { + "acc": 0.65237565, + "epoch": 0.6367326230339929, + "grad_norm": 5.09375, + "learning_rate": 8.15687538478219e-06, + "loss": 1.64658051, + "memory(GiB)": 111.15, + "step": 25100, + "train_speed(iter/s)": 0.449137 + }, + { + "epoch": 0.6367326230339929, + "eval_acc": 0.6422332670615125, + "eval_loss": 1.6146001815795898, + "eval_runtime": 112.4745, + "eval_samples_per_second": 56.635, + "eval_steps_per_second": 28.318, + "step": 25100 + }, + { + "acc": 0.63391418, + "epoch": 0.6368594622019279, + "grad_norm": 5.59375, + "learning_rate": 8.156062131184439e-06, + "loss": 1.68686485, + "memory(GiB)": 111.15, + "step": 25105, + "train_speed(iter/s)": 0.448246 + }, + { + "acc": 0.6507432, + "epoch": 0.636986301369863, + "grad_norm": 5.84375, + "learning_rate": 8.155248738768986e-06, + "loss": 1.65502472, + "memory(GiB)": 111.15, + "step": 25110, + "train_speed(iter/s)": 0.448293 + }, + { + "acc": 0.66088276, + "epoch": 0.6371131405377981, + "grad_norm": 5.3125, + "learning_rate": 8.154435207571606e-06, + "loss": 1.55605249, + "memory(GiB)": 111.15, + "step": 25115, + "train_speed(iter/s)": 0.44834 + }, + { + "acc": 0.64329348, + "epoch": 0.6372399797057331, + "grad_norm": 7.1875, + "learning_rate": 8.153621537628083e-06, + "loss": 1.59533501, + "memory(GiB)": 111.15, + "step": 25120, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.6480298, + "epoch": 0.6373668188736682, + "grad_norm": 6.21875, + "learning_rate": 8.152807728974203e-06, + "loss": 1.59587269, + "memory(GiB)": 111.15, + "step": 25125, + "train_speed(iter/s)": 0.448434 + }, + { + "acc": 0.66501017, + "epoch": 0.6374936580416033, + "grad_norm": 6.34375, + "learning_rate": 8.151993781645765e-06, + "loss": 1.60118752, + "memory(GiB)": 111.15, + "step": 25130, + "train_speed(iter/s)": 0.448481 + }, + { + "acc": 0.63164458, + "epoch": 0.6376204972095383, + "grad_norm": 7.4375, + "learning_rate": 8.151179695678565e-06, + "loss": 1.68535767, + "memory(GiB)": 111.15, + "step": 25135, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.663834, + "epoch": 0.6377473363774734, + "grad_norm": 5.03125, + "learning_rate": 8.150365471108414e-06, + "loss": 1.56416922, + "memory(GiB)": 111.15, + "step": 25140, + "train_speed(iter/s)": 0.448574 + }, + { + "acc": 0.63921051, + "epoch": 0.6378741755454084, + "grad_norm": 4.90625, + "learning_rate": 8.149551107971125e-06, + "loss": 1.69376488, + "memory(GiB)": 111.15, + "step": 25145, + "train_speed(iter/s)": 0.448621 + }, + { + "acc": 0.63716726, + "epoch": 0.6380010147133435, + "grad_norm": 5.25, + "learning_rate": 8.148736606302517e-06, + "loss": 1.69583054, + "memory(GiB)": 111.15, + "step": 25150, + "train_speed(iter/s)": 0.448668 + }, + { + "acc": 0.63453145, + "epoch": 0.6381278538812786, + "grad_norm": 5.53125, + "learning_rate": 8.147921966138412e-06, + "loss": 1.68752117, + "memory(GiB)": 111.15, + "step": 25155, + "train_speed(iter/s)": 0.448714 + }, + { + "acc": 0.67142248, + "epoch": 0.6382546930492136, + "grad_norm": 4.46875, + "learning_rate": 8.147107187514647e-06, + "loss": 1.55762539, + "memory(GiB)": 111.15, + "step": 25160, + "train_speed(iter/s)": 0.448761 + }, + { + "acc": 0.65803041, + "epoch": 0.6383815322171487, + "grad_norm": 4.5625, + "learning_rate": 8.146292270467056e-06, + "loss": 1.52209463, + "memory(GiB)": 111.15, + "step": 25165, + "train_speed(iter/s)": 0.448807 + }, + { + "acc": 0.65684171, + "epoch": 0.6385083713850838, + "grad_norm": 5.375, + "learning_rate": 8.145477215031486e-06, + "loss": 1.60488758, + "memory(GiB)": 111.15, + "step": 25170, + "train_speed(iter/s)": 0.448854 + }, + { + "acc": 0.6499114, + "epoch": 0.6386352105530188, + "grad_norm": 6.0, + "learning_rate": 8.144662021243782e-06, + "loss": 1.56386786, + "memory(GiB)": 111.15, + "step": 25175, + "train_speed(iter/s)": 0.4489 + }, + { + "acc": 0.6411643, + "epoch": 0.6387620497209539, + "grad_norm": 5.0, + "learning_rate": 8.143846689139805e-06, + "loss": 1.66774864, + "memory(GiB)": 111.15, + "step": 25180, + "train_speed(iter/s)": 0.448947 + }, + { + "acc": 0.64126363, + "epoch": 0.6388888888888888, + "grad_norm": 4.65625, + "learning_rate": 8.143031218755411e-06, + "loss": 1.66498585, + "memory(GiB)": 111.15, + "step": 25185, + "train_speed(iter/s)": 0.448993 + }, + { + "acc": 0.62882071, + "epoch": 0.6390157280568239, + "grad_norm": 5.65625, + "learning_rate": 8.142215610126474e-06, + "loss": 1.70018044, + "memory(GiB)": 111.15, + "step": 25190, + "train_speed(iter/s)": 0.44904 + }, + { + "acc": 0.66710491, + "epoch": 0.639142567224759, + "grad_norm": 5.34375, + "learning_rate": 8.141399863288863e-06, + "loss": 1.60788517, + "memory(GiB)": 111.15, + "step": 25195, + "train_speed(iter/s)": 0.449086 + }, + { + "acc": 0.64333487, + "epoch": 0.639269406392694, + "grad_norm": 5.03125, + "learning_rate": 8.140583978278463e-06, + "loss": 1.63184509, + "memory(GiB)": 111.15, + "step": 25200, + "train_speed(iter/s)": 0.449133 + }, + { + "epoch": 0.639269406392694, + "eval_acc": 0.6422750437924583, + "eval_loss": 1.6144720315933228, + "eval_runtime": 114.5331, + "eval_samples_per_second": 55.617, + "eval_steps_per_second": 27.809, + "step": 25200 + }, + { + "acc": 0.66849899, + "epoch": 0.6393962455606291, + "grad_norm": 4.5625, + "learning_rate": 8.139767955131157e-06, + "loss": 1.47731323, + "memory(GiB)": 111.15, + "step": 25205, + "train_speed(iter/s)": 0.448228 + }, + { + "acc": 0.66981125, + "epoch": 0.6395230847285642, + "grad_norm": 5.03125, + "learning_rate": 8.138951793882838e-06, + "loss": 1.56193562, + "memory(GiB)": 111.15, + "step": 25210, + "train_speed(iter/s)": 0.448274 + }, + { + "acc": 0.66646719, + "epoch": 0.6396499238964992, + "grad_norm": 5.28125, + "learning_rate": 8.138135494569405e-06, + "loss": 1.58883667, + "memory(GiB)": 111.15, + "step": 25215, + "train_speed(iter/s)": 0.448321 + }, + { + "acc": 0.65175028, + "epoch": 0.6397767630644343, + "grad_norm": 5.59375, + "learning_rate": 8.137319057226763e-06, + "loss": 1.57053032, + "memory(GiB)": 111.15, + "step": 25220, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.65277653, + "epoch": 0.6399036022323693, + "grad_norm": 4.5, + "learning_rate": 8.136502481890821e-06, + "loss": 1.63260155, + "memory(GiB)": 111.15, + "step": 25225, + "train_speed(iter/s)": 0.448413 + }, + { + "acc": 0.65161858, + "epoch": 0.6400304414003044, + "grad_norm": 6.46875, + "learning_rate": 8.135685768597496e-06, + "loss": 1.62139053, + "memory(GiB)": 111.15, + "step": 25230, + "train_speed(iter/s)": 0.44846 + }, + { + "acc": 0.6460598, + "epoch": 0.6401572805682395, + "grad_norm": 4.84375, + "learning_rate": 8.134868917382713e-06, + "loss": 1.57758598, + "memory(GiB)": 111.15, + "step": 25235, + "train_speed(iter/s)": 0.448506 + }, + { + "acc": 0.64111748, + "epoch": 0.6402841197361745, + "grad_norm": 5.03125, + "learning_rate": 8.134051928282396e-06, + "loss": 1.64872055, + "memory(GiB)": 111.15, + "step": 25240, + "train_speed(iter/s)": 0.448552 + }, + { + "acc": 0.65597315, + "epoch": 0.6404109589041096, + "grad_norm": 5.0, + "learning_rate": 8.133234801332484e-06, + "loss": 1.6460474, + "memory(GiB)": 111.15, + "step": 25245, + "train_speed(iter/s)": 0.448599 + }, + { + "acc": 0.65668058, + "epoch": 0.6405377980720447, + "grad_norm": 5.125, + "learning_rate": 8.132417536568918e-06, + "loss": 1.56980152, + "memory(GiB)": 111.15, + "step": 25250, + "train_speed(iter/s)": 0.448645 + }, + { + "acc": 0.62802076, + "epoch": 0.6406646372399797, + "grad_norm": 5.0625, + "learning_rate": 8.131600134027641e-06, + "loss": 1.70628281, + "memory(GiB)": 111.15, + "step": 25255, + "train_speed(iter/s)": 0.448641 + }, + { + "acc": 0.64581118, + "epoch": 0.6407914764079148, + "grad_norm": 4.84375, + "learning_rate": 8.13078259374461e-06, + "loss": 1.62270813, + "memory(GiB)": 111.15, + "step": 25260, + "train_speed(iter/s)": 0.448687 + }, + { + "acc": 0.63924856, + "epoch": 0.6409183155758498, + "grad_norm": 5.84375, + "learning_rate": 8.129964915755781e-06, + "loss": 1.65237961, + "memory(GiB)": 111.15, + "step": 25265, + "train_speed(iter/s)": 0.448734 + }, + { + "acc": 0.66460958, + "epoch": 0.6410451547437849, + "grad_norm": 4.4375, + "learning_rate": 8.129147100097122e-06, + "loss": 1.54580479, + "memory(GiB)": 111.15, + "step": 25270, + "train_speed(iter/s)": 0.44878 + }, + { + "acc": 0.65604186, + "epoch": 0.64117199391172, + "grad_norm": 5.34375, + "learning_rate": 8.128329146804604e-06, + "loss": 1.6217186, + "memory(GiB)": 111.15, + "step": 25275, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.65051589, + "epoch": 0.641298833079655, + "grad_norm": 5.21875, + "learning_rate": 8.127511055914201e-06, + "loss": 1.59140387, + "memory(GiB)": 111.15, + "step": 25280, + "train_speed(iter/s)": 0.448874 + }, + { + "acc": 0.66991301, + "epoch": 0.6414256722475901, + "grad_norm": 5.6875, + "learning_rate": 8.1266928274619e-06, + "loss": 1.59676704, + "memory(GiB)": 111.15, + "step": 25285, + "train_speed(iter/s)": 0.448921 + }, + { + "acc": 0.65021882, + "epoch": 0.6415525114155252, + "grad_norm": 5.0, + "learning_rate": 8.125874461483687e-06, + "loss": 1.62409973, + "memory(GiB)": 111.15, + "step": 25290, + "train_speed(iter/s)": 0.448967 + }, + { + "acc": 0.62484035, + "epoch": 0.6416793505834602, + "grad_norm": 5.21875, + "learning_rate": 8.12505595801556e-06, + "loss": 1.70603123, + "memory(GiB)": 111.15, + "step": 25295, + "train_speed(iter/s)": 0.449014 + }, + { + "acc": 0.6335062, + "epoch": 0.6418061897513953, + "grad_norm": 7.71875, + "learning_rate": 8.12423731709352e-06, + "loss": 1.72106781, + "memory(GiB)": 111.15, + "step": 25300, + "train_speed(iter/s)": 0.44906 + }, + { + "epoch": 0.6418061897513953, + "eval_acc": 0.6422846524405758, + "eval_loss": 1.6143816709518433, + "eval_runtime": 113.8136, + "eval_samples_per_second": 55.969, + "eval_steps_per_second": 27.984, + "step": 25300 + }, + { + "acc": 0.64162898, + "epoch": 0.6419330289193302, + "grad_norm": 4.9375, + "learning_rate": 8.123418538753573e-06, + "loss": 1.59354877, + "memory(GiB)": 111.15, + "step": 25305, + "train_speed(iter/s)": 0.448165 + }, + { + "acc": 0.65647416, + "epoch": 0.6420598680872653, + "grad_norm": 6.0625, + "learning_rate": 8.122599623031735e-06, + "loss": 1.58887157, + "memory(GiB)": 111.15, + "step": 25310, + "train_speed(iter/s)": 0.448211 + }, + { + "acc": 0.64369383, + "epoch": 0.6421867072552004, + "grad_norm": 4.625, + "learning_rate": 8.121780569964024e-06, + "loss": 1.60995808, + "memory(GiB)": 111.15, + "step": 25315, + "train_speed(iter/s)": 0.448257 + }, + { + "acc": 0.65754333, + "epoch": 0.6423135464231354, + "grad_norm": 5.25, + "learning_rate": 8.120961379586466e-06, + "loss": 1.61866016, + "memory(GiB)": 111.15, + "step": 25320, + "train_speed(iter/s)": 0.448304 + }, + { + "acc": 0.64092803, + "epoch": 0.6424403855910705, + "grad_norm": 5.3125, + "learning_rate": 8.120142051935092e-06, + "loss": 1.69172668, + "memory(GiB)": 111.15, + "step": 25325, + "train_speed(iter/s)": 0.44835 + }, + { + "acc": 0.62302303, + "epoch": 0.6425672247590056, + "grad_norm": 5.90625, + "learning_rate": 8.11932258704594e-06, + "loss": 1.75324974, + "memory(GiB)": 111.15, + "step": 25330, + "train_speed(iter/s)": 0.448395 + }, + { + "acc": 0.62923098, + "epoch": 0.6426940639269406, + "grad_norm": 7.09375, + "learning_rate": 8.118502984955053e-06, + "loss": 1.74558086, + "memory(GiB)": 111.15, + "step": 25335, + "train_speed(iter/s)": 0.448441 + }, + { + "acc": 0.64373865, + "epoch": 0.6428209030948757, + "grad_norm": 5.40625, + "learning_rate": 8.117683245698483e-06, + "loss": 1.65544014, + "memory(GiB)": 111.15, + "step": 25340, + "train_speed(iter/s)": 0.448488 + }, + { + "acc": 0.65083966, + "epoch": 0.6429477422628107, + "grad_norm": 5.375, + "learning_rate": 8.116863369312283e-06, + "loss": 1.66050835, + "memory(GiB)": 111.15, + "step": 25345, + "train_speed(iter/s)": 0.448534 + }, + { + "acc": 0.65403724, + "epoch": 0.6430745814307458, + "grad_norm": 5.4375, + "learning_rate": 8.116043355832518e-06, + "loss": 1.65070534, + "memory(GiB)": 111.15, + "step": 25350, + "train_speed(iter/s)": 0.448581 + }, + { + "acc": 0.64204597, + "epoch": 0.6432014205986809, + "grad_norm": 4.34375, + "learning_rate": 8.115223205295253e-06, + "loss": 1.63095093, + "memory(GiB)": 111.15, + "step": 25355, + "train_speed(iter/s)": 0.448627 + }, + { + "acc": 0.65409422, + "epoch": 0.6433282597666159, + "grad_norm": 6.125, + "learning_rate": 8.114402917736563e-06, + "loss": 1.64258766, + "memory(GiB)": 111.15, + "step": 25360, + "train_speed(iter/s)": 0.448674 + }, + { + "acc": 0.6532855, + "epoch": 0.643455098934551, + "grad_norm": 5.78125, + "learning_rate": 8.113582493192529e-06, + "loss": 1.59929485, + "memory(GiB)": 111.15, + "step": 25365, + "train_speed(iter/s)": 0.44872 + }, + { + "acc": 0.66132126, + "epoch": 0.6435819381024861, + "grad_norm": 5.5625, + "learning_rate": 8.112761931699235e-06, + "loss": 1.65522804, + "memory(GiB)": 111.15, + "step": 25370, + "train_speed(iter/s)": 0.448767 + }, + { + "acc": 0.6460021, + "epoch": 0.6437087772704211, + "grad_norm": 5.4375, + "learning_rate": 8.111941233292772e-06, + "loss": 1.67019958, + "memory(GiB)": 111.15, + "step": 25375, + "train_speed(iter/s)": 0.448814 + }, + { + "acc": 0.66555166, + "epoch": 0.6438356164383562, + "grad_norm": 6.125, + "learning_rate": 8.111120398009243e-06, + "loss": 1.59681482, + "memory(GiB)": 111.15, + "step": 25380, + "train_speed(iter/s)": 0.44886 + }, + { + "acc": 0.63871188, + "epoch": 0.6439624556062912, + "grad_norm": 5.15625, + "learning_rate": 8.110299425884745e-06, + "loss": 1.67221985, + "memory(GiB)": 111.15, + "step": 25385, + "train_speed(iter/s)": 0.448906 + }, + { + "acc": 0.64048529, + "epoch": 0.6440892947742263, + "grad_norm": 4.875, + "learning_rate": 8.109478316955394e-06, + "loss": 1.71570053, + "memory(GiB)": 111.15, + "step": 25390, + "train_speed(iter/s)": 0.448952 + }, + { + "acc": 0.64274487, + "epoch": 0.6442161339421614, + "grad_norm": 4.96875, + "learning_rate": 8.108657071257304e-06, + "loss": 1.67503834, + "memory(GiB)": 111.15, + "step": 25395, + "train_speed(iter/s)": 0.448999 + }, + { + "acc": 0.64238253, + "epoch": 0.6443429731100964, + "grad_norm": 5.5, + "learning_rate": 8.107835688826598e-06, + "loss": 1.67036705, + "memory(GiB)": 111.15, + "step": 25400, + "train_speed(iter/s)": 0.449045 + }, + { + "epoch": 0.6443429731100964, + "eval_acc": 0.6422240761807044, + "eval_loss": 1.6145817041397095, + "eval_runtime": 113.4903, + "eval_samples_per_second": 56.128, + "eval_steps_per_second": 28.064, + "step": 25400 + }, + { + "acc": 0.66519203, + "epoch": 0.6444698122780315, + "grad_norm": 4.6875, + "learning_rate": 8.1070141696994e-06, + "loss": 1.60821724, + "memory(GiB)": 111.15, + "step": 25405, + "train_speed(iter/s)": 0.448156 + }, + { + "acc": 0.65472312, + "epoch": 0.6445966514459666, + "grad_norm": 5.375, + "learning_rate": 8.106192513911849e-06, + "loss": 1.60691814, + "memory(GiB)": 111.15, + "step": 25410, + "train_speed(iter/s)": 0.448202 + }, + { + "acc": 0.64731035, + "epoch": 0.6447234906139016, + "grad_norm": 4.90625, + "learning_rate": 8.105370721500083e-06, + "loss": 1.7259182, + "memory(GiB)": 111.15, + "step": 25415, + "train_speed(iter/s)": 0.448248 + }, + { + "acc": 0.65094666, + "epoch": 0.6448503297818367, + "grad_norm": 5.0, + "learning_rate": 8.104548792500246e-06, + "loss": 1.59813118, + "memory(GiB)": 111.15, + "step": 25420, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.64835467, + "epoch": 0.6449771689497716, + "grad_norm": 5.40625, + "learning_rate": 8.103726726948495e-06, + "loss": 1.66234684, + "memory(GiB)": 111.15, + "step": 25425, + "train_speed(iter/s)": 0.448341 + }, + { + "acc": 0.65083532, + "epoch": 0.6451040081177067, + "grad_norm": 4.59375, + "learning_rate": 8.102904524880985e-06, + "loss": 1.66054497, + "memory(GiB)": 111.15, + "step": 25430, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.63968124, + "epoch": 0.6452308472856418, + "grad_norm": 5.0, + "learning_rate": 8.10208218633388e-06, + "loss": 1.65318108, + "memory(GiB)": 111.15, + "step": 25435, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.64405375, + "epoch": 0.6453576864535768, + "grad_norm": 6.0, + "learning_rate": 8.10125971134335e-06, + "loss": 1.67444611, + "memory(GiB)": 111.15, + "step": 25440, + "train_speed(iter/s)": 0.44848 + }, + { + "acc": 0.64455309, + "epoch": 0.6454845256215119, + "grad_norm": 5.4375, + "learning_rate": 8.100437099945572e-06, + "loss": 1.61682014, + "memory(GiB)": 111.15, + "step": 25445, + "train_speed(iter/s)": 0.448526 + }, + { + "acc": 0.65455694, + "epoch": 0.645611364789447, + "grad_norm": 4.59375, + "learning_rate": 8.099614352176727e-06, + "loss": 1.56520061, + "memory(GiB)": 111.15, + "step": 25450, + "train_speed(iter/s)": 0.448572 + }, + { + "acc": 0.65275202, + "epoch": 0.645738203957382, + "grad_norm": 5.40625, + "learning_rate": 8.098791468073007e-06, + "loss": 1.64330025, + "memory(GiB)": 111.15, + "step": 25455, + "train_speed(iter/s)": 0.448618 + }, + { + "acc": 0.65771189, + "epoch": 0.6458650431253171, + "grad_norm": 6.34375, + "learning_rate": 8.097968447670601e-06, + "loss": 1.59385395, + "memory(GiB)": 111.15, + "step": 25460, + "train_speed(iter/s)": 0.448665 + }, + { + "acc": 0.65540133, + "epoch": 0.6459918822932521, + "grad_norm": 6.125, + "learning_rate": 8.09714529100571e-06, + "loss": 1.61375885, + "memory(GiB)": 111.15, + "step": 25465, + "train_speed(iter/s)": 0.448711 + }, + { + "acc": 0.65138292, + "epoch": 0.6461187214611872, + "grad_norm": 5.96875, + "learning_rate": 8.096321998114545e-06, + "loss": 1.63932648, + "memory(GiB)": 111.15, + "step": 25470, + "train_speed(iter/s)": 0.448757 + }, + { + "acc": 0.6624577, + "epoch": 0.6462455606291223, + "grad_norm": 6.28125, + "learning_rate": 8.09549856903331e-06, + "loss": 1.55449047, + "memory(GiB)": 111.15, + "step": 25475, + "train_speed(iter/s)": 0.448803 + }, + { + "acc": 0.63620672, + "epoch": 0.6463723997970573, + "grad_norm": 5.0, + "learning_rate": 8.094675003798232e-06, + "loss": 1.67134361, + "memory(GiB)": 111.15, + "step": 25480, + "train_speed(iter/s)": 0.448849 + }, + { + "acc": 0.66873422, + "epoch": 0.6464992389649924, + "grad_norm": 6.25, + "learning_rate": 8.093851302445528e-06, + "loss": 1.5526825, + "memory(GiB)": 111.15, + "step": 25485, + "train_speed(iter/s)": 0.448895 + }, + { + "acc": 0.65198755, + "epoch": 0.6466260781329275, + "grad_norm": 5.03125, + "learning_rate": 8.093027465011431e-06, + "loss": 1.5978735, + "memory(GiB)": 111.15, + "step": 25490, + "train_speed(iter/s)": 0.448942 + }, + { + "acc": 0.65398464, + "epoch": 0.6467529173008625, + "grad_norm": 4.96875, + "learning_rate": 8.092203491532178e-06, + "loss": 1.60472717, + "memory(GiB)": 111.15, + "step": 25495, + "train_speed(iter/s)": 0.448988 + }, + { + "acc": 0.65068512, + "epoch": 0.6468797564687976, + "grad_norm": 5.21875, + "learning_rate": 8.091379382044009e-06, + "loss": 1.66947441, + "memory(GiB)": 111.15, + "step": 25500, + "train_speed(iter/s)": 0.449034 + }, + { + "epoch": 0.6468797564687976, + "eval_acc": 0.6422520665904381, + "eval_loss": 1.614400863647461, + "eval_runtime": 113.6548, + "eval_samples_per_second": 56.047, + "eval_steps_per_second": 28.023, + "step": 25500 + }, + { + "acc": 0.65688868, + "epoch": 0.6470065956367326, + "grad_norm": 6.0625, + "learning_rate": 8.090555136583172e-06, + "loss": 1.59176197, + "memory(GiB)": 111.15, + "step": 25505, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.64921751, + "epoch": 0.6471334348046677, + "grad_norm": 5.25, + "learning_rate": 8.089730755185921e-06, + "loss": 1.62365608, + "memory(GiB)": 111.15, + "step": 25510, + "train_speed(iter/s)": 0.448193 + }, + { + "acc": 0.64324703, + "epoch": 0.6472602739726028, + "grad_norm": 5.3125, + "learning_rate": 8.088906237888517e-06, + "loss": 1.6427784, + "memory(GiB)": 111.15, + "step": 25515, + "train_speed(iter/s)": 0.448238 + }, + { + "acc": 0.64541678, + "epoch": 0.6473871131405378, + "grad_norm": 4.84375, + "learning_rate": 8.088081584727228e-06, + "loss": 1.60653896, + "memory(GiB)": 111.15, + "step": 25520, + "train_speed(iter/s)": 0.448284 + }, + { + "acc": 0.63972807, + "epoch": 0.6475139523084729, + "grad_norm": 6.40625, + "learning_rate": 8.08725679573832e-06, + "loss": 1.62352257, + "memory(GiB)": 111.15, + "step": 25525, + "train_speed(iter/s)": 0.44833 + }, + { + "acc": 0.65983524, + "epoch": 0.647640791476408, + "grad_norm": 5.0625, + "learning_rate": 8.086431870958078e-06, + "loss": 1.57216702, + "memory(GiB)": 111.15, + "step": 25530, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.64312534, + "epoch": 0.647767630644343, + "grad_norm": 4.375, + "learning_rate": 8.085606810422781e-06, + "loss": 1.65218964, + "memory(GiB)": 111.15, + "step": 25535, + "train_speed(iter/s)": 0.448422 + }, + { + "acc": 0.64134121, + "epoch": 0.647894469812278, + "grad_norm": 5.65625, + "learning_rate": 8.08478161416872e-06, + "loss": 1.69889946, + "memory(GiB)": 111.15, + "step": 25540, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.65057154, + "epoch": 0.648021308980213, + "grad_norm": 5.03125, + "learning_rate": 8.083956282232192e-06, + "loss": 1.66884556, + "memory(GiB)": 111.15, + "step": 25545, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.64467421, + "epoch": 0.6481481481481481, + "grad_norm": 5.875, + "learning_rate": 8.083130814649498e-06, + "loss": 1.59348049, + "memory(GiB)": 111.15, + "step": 25550, + "train_speed(iter/s)": 0.448559 + }, + { + "acc": 0.67037735, + "epoch": 0.6482749873160832, + "grad_norm": 5.875, + "learning_rate": 8.082305211456943e-06, + "loss": 1.5313797, + "memory(GiB)": 111.15, + "step": 25555, + "train_speed(iter/s)": 0.448606 + }, + { + "acc": 0.64469061, + "epoch": 0.6484018264840182, + "grad_norm": 4.84375, + "learning_rate": 8.081479472690846e-06, + "loss": 1.66899948, + "memory(GiB)": 111.15, + "step": 25560, + "train_speed(iter/s)": 0.448651 + }, + { + "acc": 0.64559317, + "epoch": 0.6485286656519533, + "grad_norm": 4.4375, + "learning_rate": 8.080653598387522e-06, + "loss": 1.56335545, + "memory(GiB)": 111.15, + "step": 25565, + "train_speed(iter/s)": 0.448697 + }, + { + "acc": 0.66758184, + "epoch": 0.6486555048198884, + "grad_norm": 5.8125, + "learning_rate": 8.0798275885833e-06, + "loss": 1.57728138, + "memory(GiB)": 111.15, + "step": 25570, + "train_speed(iter/s)": 0.448743 + }, + { + "acc": 0.65168467, + "epoch": 0.6487823439878234, + "grad_norm": 5.40625, + "learning_rate": 8.07900144331451e-06, + "loss": 1.56445971, + "memory(GiB)": 111.15, + "step": 25575, + "train_speed(iter/s)": 0.448788 + }, + { + "acc": 0.65222054, + "epoch": 0.6489091831557585, + "grad_norm": 5.9375, + "learning_rate": 8.07817516261749e-06, + "loss": 1.61443195, + "memory(GiB)": 111.15, + "step": 25580, + "train_speed(iter/s)": 0.448834 + }, + { + "acc": 0.65001545, + "epoch": 0.6490360223236935, + "grad_norm": 4.84375, + "learning_rate": 8.077348746528583e-06, + "loss": 1.59643841, + "memory(GiB)": 111.15, + "step": 25585, + "train_speed(iter/s)": 0.44888 + }, + { + "acc": 0.65492501, + "epoch": 0.6491628614916286, + "grad_norm": 4.84375, + "learning_rate": 8.076522195084139e-06, + "loss": 1.57444572, + "memory(GiB)": 111.15, + "step": 25590, + "train_speed(iter/s)": 0.448926 + }, + { + "acc": 0.65835314, + "epoch": 0.6492897006595637, + "grad_norm": 4.6875, + "learning_rate": 8.075695508320512e-06, + "loss": 1.64909401, + "memory(GiB)": 111.15, + "step": 25595, + "train_speed(iter/s)": 0.448972 + }, + { + "acc": 0.64032898, + "epoch": 0.6494165398274987, + "grad_norm": 6.34375, + "learning_rate": 8.074868686274065e-06, + "loss": 1.71772232, + "memory(GiB)": 111.15, + "step": 25600, + "train_speed(iter/s)": 0.449018 + }, + { + "epoch": 0.6494165398274987, + "eval_acc": 0.6423159849887851, + "eval_loss": 1.6143243312835693, + "eval_runtime": 113.9109, + "eval_samples_per_second": 55.921, + "eval_steps_per_second": 27.96, + "step": 25600 + }, + { + "acc": 0.66195016, + "epoch": 0.6495433789954338, + "grad_norm": 5.625, + "learning_rate": 8.074041728981166e-06, + "loss": 1.56921539, + "memory(GiB)": 111.15, + "step": 25605, + "train_speed(iter/s)": 0.448133 + }, + { + "acc": 0.65274968, + "epoch": 0.6496702181633689, + "grad_norm": 5.0, + "learning_rate": 8.073214636478186e-06, + "loss": 1.59697227, + "memory(GiB)": 111.15, + "step": 25610, + "train_speed(iter/s)": 0.44818 + }, + { + "acc": 0.65663395, + "epoch": 0.6497970573313039, + "grad_norm": 7.375, + "learning_rate": 8.072387408801506e-06, + "loss": 1.6655735, + "memory(GiB)": 111.15, + "step": 25615, + "train_speed(iter/s)": 0.448226 + }, + { + "acc": 0.64439588, + "epoch": 0.649923896499239, + "grad_norm": 4.71875, + "learning_rate": 8.07156004598751e-06, + "loss": 1.63851318, + "memory(GiB)": 111.15, + "step": 25620, + "train_speed(iter/s)": 0.448272 + }, + { + "acc": 0.65031786, + "epoch": 0.650050735667174, + "grad_norm": 5.1875, + "learning_rate": 8.07073254807259e-06, + "loss": 1.60623093, + "memory(GiB)": 111.15, + "step": 25625, + "train_speed(iter/s)": 0.448318 + }, + { + "acc": 0.64127941, + "epoch": 0.6501775748351091, + "grad_norm": 4.6875, + "learning_rate": 8.069904915093144e-06, + "loss": 1.65174904, + "memory(GiB)": 111.15, + "step": 25630, + "train_speed(iter/s)": 0.448364 + }, + { + "acc": 0.63821106, + "epoch": 0.6503044140030442, + "grad_norm": 4.4375, + "learning_rate": 8.069077147085571e-06, + "loss": 1.61451721, + "memory(GiB)": 111.15, + "step": 25635, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.65962601, + "epoch": 0.6504312531709792, + "grad_norm": 5.5, + "learning_rate": 8.068249244086283e-06, + "loss": 1.60972652, + "memory(GiB)": 111.15, + "step": 25640, + "train_speed(iter/s)": 0.448456 + }, + { + "acc": 0.63471899, + "epoch": 0.6505580923389143, + "grad_norm": 5.5, + "learning_rate": 8.067421206131696e-06, + "loss": 1.7025568, + "memory(GiB)": 111.15, + "step": 25645, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.63853784, + "epoch": 0.6506849315068494, + "grad_norm": 5.625, + "learning_rate": 8.06659303325823e-06, + "loss": 1.63815269, + "memory(GiB)": 111.15, + "step": 25650, + "train_speed(iter/s)": 0.448548 + }, + { + "acc": 0.64889979, + "epoch": 0.6508117706747844, + "grad_norm": 5.15625, + "learning_rate": 8.06576472550231e-06, + "loss": 1.67945194, + "memory(GiB)": 111.15, + "step": 25655, + "train_speed(iter/s)": 0.448593 + }, + { + "acc": 0.64380674, + "epoch": 0.6509386098427195, + "grad_norm": 5.34375, + "learning_rate": 8.064936282900368e-06, + "loss": 1.63129807, + "memory(GiB)": 111.15, + "step": 25660, + "train_speed(iter/s)": 0.448639 + }, + { + "acc": 0.6471302, + "epoch": 0.6510654490106544, + "grad_norm": 5.9375, + "learning_rate": 8.064107705488846e-06, + "loss": 1.60866394, + "memory(GiB)": 111.15, + "step": 25665, + "train_speed(iter/s)": 0.448685 + }, + { + "acc": 0.65062513, + "epoch": 0.6511922881785895, + "grad_norm": 5.6875, + "learning_rate": 8.063278993304188e-06, + "loss": 1.57072515, + "memory(GiB)": 111.15, + "step": 25670, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.64598055, + "epoch": 0.6513191273465246, + "grad_norm": 6.3125, + "learning_rate": 8.06245014638284e-06, + "loss": 1.6252985, + "memory(GiB)": 111.15, + "step": 25675, + "train_speed(iter/s)": 0.448777 + }, + { + "acc": 0.66055613, + "epoch": 0.6514459665144596, + "grad_norm": 6.34375, + "learning_rate": 8.061621164761266e-06, + "loss": 1.54991283, + "memory(GiB)": 111.15, + "step": 25680, + "train_speed(iter/s)": 0.448823 + }, + { + "acc": 0.64463511, + "epoch": 0.6515728056823947, + "grad_norm": 5.4375, + "learning_rate": 8.06079204847592e-06, + "loss": 1.60334682, + "memory(GiB)": 111.15, + "step": 25685, + "train_speed(iter/s)": 0.448868 + }, + { + "acc": 0.64333115, + "epoch": 0.6516996448503298, + "grad_norm": 4.65625, + "learning_rate": 8.059962797563277e-06, + "loss": 1.58990459, + "memory(GiB)": 111.15, + "step": 25690, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.65165014, + "epoch": 0.6518264840182648, + "grad_norm": 4.46875, + "learning_rate": 8.059133412059808e-06, + "loss": 1.59367733, + "memory(GiB)": 111.15, + "step": 25695, + "train_speed(iter/s)": 0.44896 + }, + { + "acc": 0.65217199, + "epoch": 0.6519533231861999, + "grad_norm": 5.28125, + "learning_rate": 8.058303892001993e-06, + "loss": 1.6195612, + "memory(GiB)": 111.15, + "step": 25700, + "train_speed(iter/s)": 0.449006 + }, + { + "epoch": 0.6519533231861999, + "eval_acc": 0.6422708661193637, + "eval_loss": 1.6139781475067139, + "eval_runtime": 113.9052, + "eval_samples_per_second": 55.924, + "eval_steps_per_second": 27.962, + "step": 25700 + }, + { + "acc": 0.64697795, + "epoch": 0.6520801623541349, + "grad_norm": 5.5625, + "learning_rate": 8.057474237426318e-06, + "loss": 1.5952507, + "memory(GiB)": 111.15, + "step": 25705, + "train_speed(iter/s)": 0.448125 + }, + { + "acc": 0.63727083, + "epoch": 0.65220700152207, + "grad_norm": 5.15625, + "learning_rate": 8.056644448369275e-06, + "loss": 1.7091753, + "memory(GiB)": 111.15, + "step": 25710, + "train_speed(iter/s)": 0.448171 + }, + { + "acc": 0.63986816, + "epoch": 0.6523338406900051, + "grad_norm": 4.78125, + "learning_rate": 8.055814524867364e-06, + "loss": 1.69963837, + "memory(GiB)": 111.15, + "step": 25715, + "train_speed(iter/s)": 0.448218 + }, + { + "acc": 0.6435585, + "epoch": 0.6524606798579401, + "grad_norm": 5.6875, + "learning_rate": 8.054984466957085e-06, + "loss": 1.69991112, + "memory(GiB)": 111.15, + "step": 25720, + "train_speed(iter/s)": 0.448265 + }, + { + "acc": 0.6611845, + "epoch": 0.6525875190258752, + "grad_norm": 5.0, + "learning_rate": 8.05415427467495e-06, + "loss": 1.49158974, + "memory(GiB)": 111.15, + "step": 25725, + "train_speed(iter/s)": 0.448311 + }, + { + "acc": 0.65234733, + "epoch": 0.6527143581938103, + "grad_norm": 4.96875, + "learning_rate": 8.053323948057477e-06, + "loss": 1.64783707, + "memory(GiB)": 111.15, + "step": 25730, + "train_speed(iter/s)": 0.448357 + }, + { + "acc": 0.63713412, + "epoch": 0.6528411973617453, + "grad_norm": 5.21875, + "learning_rate": 8.052493487141183e-06, + "loss": 1.66128464, + "memory(GiB)": 111.15, + "step": 25735, + "train_speed(iter/s)": 0.448404 + }, + { + "acc": 0.64546366, + "epoch": 0.6529680365296804, + "grad_norm": 5.09375, + "learning_rate": 8.051662891962594e-06, + "loss": 1.62852325, + "memory(GiB)": 111.15, + "step": 25740, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.65301733, + "epoch": 0.6530948756976154, + "grad_norm": 4.75, + "learning_rate": 8.05083216255825e-06, + "loss": 1.63335819, + "memory(GiB)": 111.15, + "step": 25745, + "train_speed(iter/s)": 0.448496 + }, + { + "acc": 0.65266104, + "epoch": 0.6532217148655505, + "grad_norm": 4.84375, + "learning_rate": 8.050001298964685e-06, + "loss": 1.67435913, + "memory(GiB)": 111.15, + "step": 25750, + "train_speed(iter/s)": 0.448543 + }, + { + "acc": 0.65955648, + "epoch": 0.6533485540334856, + "grad_norm": 5.46875, + "learning_rate": 8.049170301218445e-06, + "loss": 1.66154099, + "memory(GiB)": 111.15, + "step": 25755, + "train_speed(iter/s)": 0.448589 + }, + { + "acc": 0.66046896, + "epoch": 0.6534753932014206, + "grad_norm": 6.09375, + "learning_rate": 8.048339169356085e-06, + "loss": 1.60851879, + "memory(GiB)": 111.15, + "step": 25760, + "train_speed(iter/s)": 0.448636 + }, + { + "acc": 0.66982365, + "epoch": 0.6536022323693557, + "grad_norm": 4.9375, + "learning_rate": 8.047507903414155e-06, + "loss": 1.54360275, + "memory(GiB)": 111.15, + "step": 25765, + "train_speed(iter/s)": 0.448682 + }, + { + "acc": 0.66849079, + "epoch": 0.6537290715372908, + "grad_norm": 5.0, + "learning_rate": 8.046676503429222e-06, + "loss": 1.55382681, + "memory(GiB)": 111.15, + "step": 25770, + "train_speed(iter/s)": 0.448729 + }, + { + "acc": 0.65166116, + "epoch": 0.6538559107052258, + "grad_norm": 5.5, + "learning_rate": 8.045844969437855e-06, + "loss": 1.61276741, + "memory(GiB)": 111.15, + "step": 25775, + "train_speed(iter/s)": 0.448775 + }, + { + "acc": 0.63842721, + "epoch": 0.6539827498731609, + "grad_norm": 5.4375, + "learning_rate": 8.045013301476625e-06, + "loss": 1.62988701, + "memory(GiB)": 111.15, + "step": 25780, + "train_speed(iter/s)": 0.448821 + }, + { + "acc": 0.65300555, + "epoch": 0.6541095890410958, + "grad_norm": 4.46875, + "learning_rate": 8.044181499582117e-06, + "loss": 1.63524551, + "memory(GiB)": 111.15, + "step": 25785, + "train_speed(iter/s)": 0.448868 + }, + { + "acc": 0.65854082, + "epoch": 0.6542364282090309, + "grad_norm": 5.46875, + "learning_rate": 8.043349563790917e-06, + "loss": 1.64527779, + "memory(GiB)": 111.15, + "step": 25790, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.65147352, + "epoch": 0.654363267376966, + "grad_norm": 5.25, + "learning_rate": 8.042517494139612e-06, + "loss": 1.66203117, + "memory(GiB)": 111.15, + "step": 25795, + "train_speed(iter/s)": 0.448961 + }, + { + "acc": 0.65452523, + "epoch": 0.654490106544901, + "grad_norm": 4.84375, + "learning_rate": 8.041685290664806e-06, + "loss": 1.57043362, + "memory(GiB)": 111.15, + "step": 25800, + "train_speed(iter/s)": 0.449007 + }, + { + "epoch": 0.654490106544901, + "eval_acc": 0.6423293535426877, + "eval_loss": 1.6140364408493042, + "eval_runtime": 113.4341, + "eval_samples_per_second": 56.156, + "eval_steps_per_second": 28.078, + "step": 25800 + }, + { + "acc": 0.64233599, + "epoch": 0.6546169457128361, + "grad_norm": 4.9375, + "learning_rate": 8.0408529534031e-06, + "loss": 1.59457197, + "memory(GiB)": 111.15, + "step": 25805, + "train_speed(iter/s)": 0.448134 + }, + { + "acc": 0.6470438, + "epoch": 0.6547437848807712, + "grad_norm": 4.8125, + "learning_rate": 8.040020482391105e-06, + "loss": 1.58700294, + "memory(GiB)": 111.15, + "step": 25810, + "train_speed(iter/s)": 0.44818 + }, + { + "acc": 0.68016114, + "epoch": 0.6548706240487062, + "grad_norm": 5.09375, + "learning_rate": 8.039187877665435e-06, + "loss": 1.50465231, + "memory(GiB)": 111.15, + "step": 25815, + "train_speed(iter/s)": 0.448226 + }, + { + "acc": 0.65454764, + "epoch": 0.6549974632166413, + "grad_norm": 6.96875, + "learning_rate": 8.038355139262716e-06, + "loss": 1.59597073, + "memory(GiB)": 111.15, + "step": 25820, + "train_speed(iter/s)": 0.448272 + }, + { + "acc": 0.66752234, + "epoch": 0.6551243023845763, + "grad_norm": 5.6875, + "learning_rate": 8.037522267219571e-06, + "loss": 1.58674889, + "memory(GiB)": 111.15, + "step": 25825, + "train_speed(iter/s)": 0.448318 + }, + { + "acc": 0.6382863, + "epoch": 0.6552511415525114, + "grad_norm": 5.34375, + "learning_rate": 8.036689261572636e-06, + "loss": 1.63378983, + "memory(GiB)": 111.15, + "step": 25830, + "train_speed(iter/s)": 0.448365 + }, + { + "acc": 0.64015913, + "epoch": 0.6553779807204465, + "grad_norm": 6.5, + "learning_rate": 8.035856122358548e-06, + "loss": 1.69332047, + "memory(GiB)": 111.15, + "step": 25835, + "train_speed(iter/s)": 0.448411 + }, + { + "acc": 0.64647026, + "epoch": 0.6555048198883815, + "grad_norm": 4.90625, + "learning_rate": 8.035022849613954e-06, + "loss": 1.65143471, + "memory(GiB)": 111.15, + "step": 25840, + "train_speed(iter/s)": 0.448458 + }, + { + "acc": 0.6527626, + "epoch": 0.6556316590563166, + "grad_norm": 5.875, + "learning_rate": 8.034189443375505e-06, + "loss": 1.65209656, + "memory(GiB)": 111.15, + "step": 25845, + "train_speed(iter/s)": 0.448504 + }, + { + "acc": 0.65280967, + "epoch": 0.6557584982242517, + "grad_norm": 5.25, + "learning_rate": 8.033355903679858e-06, + "loss": 1.59163551, + "memory(GiB)": 111.15, + "step": 25850, + "train_speed(iter/s)": 0.44855 + }, + { + "acc": 0.659128, + "epoch": 0.6558853373921867, + "grad_norm": 5.875, + "learning_rate": 8.032522230563676e-06, + "loss": 1.66483002, + "memory(GiB)": 111.15, + "step": 25855, + "train_speed(iter/s)": 0.448597 + }, + { + "acc": 0.64467721, + "epoch": 0.6560121765601218, + "grad_norm": 6.09375, + "learning_rate": 8.031688424063625e-06, + "loss": 1.61489983, + "memory(GiB)": 111.15, + "step": 25860, + "train_speed(iter/s)": 0.448643 + }, + { + "acc": 0.64759479, + "epoch": 0.6561390157280568, + "grad_norm": 5.5625, + "learning_rate": 8.030854484216381e-06, + "loss": 1.65349693, + "memory(GiB)": 111.15, + "step": 25865, + "train_speed(iter/s)": 0.448689 + }, + { + "acc": 0.64728861, + "epoch": 0.6562658548959919, + "grad_norm": 4.59375, + "learning_rate": 8.030020411058627e-06, + "loss": 1.66470413, + "memory(GiB)": 111.15, + "step": 25870, + "train_speed(iter/s)": 0.448735 + }, + { + "acc": 0.65999928, + "epoch": 0.656392694063927, + "grad_norm": 5.65625, + "learning_rate": 8.029186204627049e-06, + "loss": 1.59315844, + "memory(GiB)": 111.15, + "step": 25875, + "train_speed(iter/s)": 0.448781 + }, + { + "acc": 0.64395471, + "epoch": 0.656519533231862, + "grad_norm": 6.90625, + "learning_rate": 8.028351864958335e-06, + "loss": 1.61719456, + "memory(GiB)": 111.15, + "step": 25880, + "train_speed(iter/s)": 0.448828 + }, + { + "acc": 0.66763391, + "epoch": 0.6566463723997971, + "grad_norm": 5.4375, + "learning_rate": 8.027517392089185e-06, + "loss": 1.5679944, + "memory(GiB)": 111.15, + "step": 25885, + "train_speed(iter/s)": 0.448873 + }, + { + "acc": 0.64639921, + "epoch": 0.6567732115677322, + "grad_norm": 5.875, + "learning_rate": 8.026682786056304e-06, + "loss": 1.65703278, + "memory(GiB)": 111.15, + "step": 25890, + "train_speed(iter/s)": 0.448919 + }, + { + "acc": 0.62948933, + "epoch": 0.6569000507356672, + "grad_norm": 4.53125, + "learning_rate": 8.025848046896401e-06, + "loss": 1.71914291, + "memory(GiB)": 111.15, + "step": 25895, + "train_speed(iter/s)": 0.448965 + }, + { + "acc": 0.651717, + "epoch": 0.6570268899036023, + "grad_norm": 5.4375, + "learning_rate": 8.02501317464619e-06, + "loss": 1.64855537, + "memory(GiB)": 111.15, + "step": 25900, + "train_speed(iter/s)": 0.449011 + }, + { + "epoch": 0.6570268899036023, + "eval_acc": 0.6422796392328622, + "eval_loss": 1.61420738697052, + "eval_runtime": 114.0496, + "eval_samples_per_second": 55.853, + "eval_steps_per_second": 27.926, + "step": 25900 + }, + { + "acc": 0.65029802, + "epoch": 0.6571537290715372, + "grad_norm": 6.84375, + "learning_rate": 8.024178169342396e-06, + "loss": 1.69455376, + "memory(GiB)": 111.15, + "step": 25905, + "train_speed(iter/s)": 0.448136 + }, + { + "acc": 0.65199099, + "epoch": 0.6572805682394723, + "grad_norm": 5.34375, + "learning_rate": 8.023343031021744e-06, + "loss": 1.62866764, + "memory(GiB)": 111.15, + "step": 25910, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.64826579, + "epoch": 0.6574074074074074, + "grad_norm": 4.75, + "learning_rate": 8.022507759720966e-06, + "loss": 1.61901627, + "memory(GiB)": 111.15, + "step": 25915, + "train_speed(iter/s)": 0.448227 + }, + { + "acc": 0.65236654, + "epoch": 0.6575342465753424, + "grad_norm": 5.3125, + "learning_rate": 8.021672355476802e-06, + "loss": 1.67185726, + "memory(GiB)": 111.15, + "step": 25920, + "train_speed(iter/s)": 0.448273 + }, + { + "acc": 0.6442853, + "epoch": 0.6576610857432775, + "grad_norm": 5.03125, + "learning_rate": 8.020836818325997e-06, + "loss": 1.60283241, + "memory(GiB)": 111.15, + "step": 25925, + "train_speed(iter/s)": 0.448318 + }, + { + "acc": 0.65298371, + "epoch": 0.6577879249112126, + "grad_norm": 7.1875, + "learning_rate": 8.020001148305304e-06, + "loss": 1.59266806, + "memory(GiB)": 111.15, + "step": 25930, + "train_speed(iter/s)": 0.448364 + }, + { + "acc": 0.67037945, + "epoch": 0.6579147640791476, + "grad_norm": 5.78125, + "learning_rate": 8.019165345451475e-06, + "loss": 1.57959309, + "memory(GiB)": 111.15, + "step": 25935, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.6629024, + "epoch": 0.6580416032470827, + "grad_norm": 6.625, + "learning_rate": 8.018329409801276e-06, + "loss": 1.61035194, + "memory(GiB)": 111.15, + "step": 25940, + "train_speed(iter/s)": 0.448456 + }, + { + "acc": 0.65397105, + "epoch": 0.6581684424150177, + "grad_norm": 5.03125, + "learning_rate": 8.017493341391471e-06, + "loss": 1.59367628, + "memory(GiB)": 111.15, + "step": 25945, + "train_speed(iter/s)": 0.448502 + }, + { + "acc": 0.64936762, + "epoch": 0.6582952815829528, + "grad_norm": 5.125, + "learning_rate": 8.016657140258839e-06, + "loss": 1.57363319, + "memory(GiB)": 111.15, + "step": 25950, + "train_speed(iter/s)": 0.448548 + }, + { + "acc": 0.65408335, + "epoch": 0.6584221207508879, + "grad_norm": 5.78125, + "learning_rate": 8.015820806440157e-06, + "loss": 1.63827991, + "memory(GiB)": 111.15, + "step": 25955, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.62544222, + "epoch": 0.6585489599188229, + "grad_norm": 5.25, + "learning_rate": 8.014984339972211e-06, + "loss": 1.61633873, + "memory(GiB)": 111.15, + "step": 25960, + "train_speed(iter/s)": 0.44864 + }, + { + "acc": 0.65522795, + "epoch": 0.658675799086758, + "grad_norm": 5.03125, + "learning_rate": 8.014147740891793e-06, + "loss": 1.63047123, + "memory(GiB)": 111.15, + "step": 25965, + "train_speed(iter/s)": 0.448685 + }, + { + "acc": 0.65415854, + "epoch": 0.6588026382546931, + "grad_norm": 6.71875, + "learning_rate": 8.0133110092357e-06, + "loss": 1.58265972, + "memory(GiB)": 111.15, + "step": 25970, + "train_speed(iter/s)": 0.448732 + }, + { + "acc": 0.65792809, + "epoch": 0.6589294774226281, + "grad_norm": 5.3125, + "learning_rate": 8.012474145040737e-06, + "loss": 1.70031681, + "memory(GiB)": 111.15, + "step": 25975, + "train_speed(iter/s)": 0.448777 + }, + { + "acc": 0.650741, + "epoch": 0.6590563165905632, + "grad_norm": 5.15625, + "learning_rate": 8.01163714834371e-06, + "loss": 1.63954201, + "memory(GiB)": 111.15, + "step": 25980, + "train_speed(iter/s)": 0.448823 + }, + { + "acc": 0.64753981, + "epoch": 0.6591831557584982, + "grad_norm": 6.25, + "learning_rate": 8.010800019181433e-06, + "loss": 1.68878403, + "memory(GiB)": 111.15, + "step": 25985, + "train_speed(iter/s)": 0.448869 + }, + { + "acc": 0.64774389, + "epoch": 0.6593099949264333, + "grad_norm": 5.15625, + "learning_rate": 8.009962757590732e-06, + "loss": 1.6271656, + "memory(GiB)": 111.15, + "step": 25990, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.65323577, + "epoch": 0.6594368340943684, + "grad_norm": 4.78125, + "learning_rate": 8.00912536360843e-06, + "loss": 1.62697315, + "memory(GiB)": 111.15, + "step": 25995, + "train_speed(iter/s)": 0.44896 + }, + { + "acc": 0.65433664, + "epoch": 0.6595636732623034, + "grad_norm": 6.0625, + "learning_rate": 8.008287837271359e-06, + "loss": 1.5888093, + "memory(GiB)": 111.15, + "step": 26000, + "train_speed(iter/s)": 0.449005 + }, + { + "epoch": 0.6595636732623034, + "eval_acc": 0.6422558264962231, + "eval_loss": 1.613673448562622, + "eval_runtime": 113.0658, + "eval_samples_per_second": 56.339, + "eval_steps_per_second": 28.169, + "step": 26000 + }, + { + "acc": 0.65878649, + "epoch": 0.6596905124302385, + "grad_norm": 5.5625, + "learning_rate": 8.007450178616356e-06, + "loss": 1.56436157, + "memory(GiB)": 111.15, + "step": 26005, + "train_speed(iter/s)": 0.448141 + }, + { + "acc": 0.64537811, + "epoch": 0.6598173515981736, + "grad_norm": 5.09375, + "learning_rate": 8.00661238768027e-06, + "loss": 1.62054863, + "memory(GiB)": 111.15, + "step": 26010, + "train_speed(iter/s)": 0.448187 + }, + { + "acc": 0.64779925, + "epoch": 0.6599441907661086, + "grad_norm": 5.75, + "learning_rate": 8.005774464499947e-06, + "loss": 1.62803993, + "memory(GiB)": 111.15, + "step": 26015, + "train_speed(iter/s)": 0.448233 + }, + { + "acc": 0.63961244, + "epoch": 0.6600710299340437, + "grad_norm": 4.90625, + "learning_rate": 8.004936409112243e-06, + "loss": 1.62620773, + "memory(GiB)": 111.15, + "step": 26020, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.64943361, + "epoch": 0.6601978691019786, + "grad_norm": 4.9375, + "learning_rate": 8.004098221554018e-06, + "loss": 1.64074631, + "memory(GiB)": 111.15, + "step": 26025, + "train_speed(iter/s)": 0.448325 + }, + { + "acc": 0.65786648, + "epoch": 0.6603247082699137, + "grad_norm": 4.40625, + "learning_rate": 8.003259901862143e-06, + "loss": 1.61132622, + "memory(GiB)": 111.15, + "step": 26030, + "train_speed(iter/s)": 0.44837 + }, + { + "acc": 0.64131036, + "epoch": 0.6604515474378488, + "grad_norm": 5.125, + "learning_rate": 8.002421450073488e-06, + "loss": 1.61688519, + "memory(GiB)": 111.15, + "step": 26035, + "train_speed(iter/s)": 0.448415 + }, + { + "acc": 0.65050335, + "epoch": 0.6605783866057838, + "grad_norm": 5.78125, + "learning_rate": 8.001582866224932e-06, + "loss": 1.59697428, + "memory(GiB)": 111.15, + "step": 26040, + "train_speed(iter/s)": 0.44846 + }, + { + "acc": 0.65930552, + "epoch": 0.6607052257737189, + "grad_norm": 5.46875, + "learning_rate": 8.000744150353362e-06, + "loss": 1.65636311, + "memory(GiB)": 111.15, + "step": 26045, + "train_speed(iter/s)": 0.448506 + }, + { + "acc": 0.65093937, + "epoch": 0.660832064941654, + "grad_norm": 5.375, + "learning_rate": 7.999905302495667e-06, + "loss": 1.65364189, + "memory(GiB)": 111.15, + "step": 26050, + "train_speed(iter/s)": 0.448551 + }, + { + "acc": 0.65470271, + "epoch": 0.660958904109589, + "grad_norm": 5.40625, + "learning_rate": 7.999066322688743e-06, + "loss": 1.60443764, + "memory(GiB)": 111.15, + "step": 26055, + "train_speed(iter/s)": 0.448597 + }, + { + "acc": 0.6540926, + "epoch": 0.6610857432775241, + "grad_norm": 6.0, + "learning_rate": 7.998227210969491e-06, + "loss": 1.67997704, + "memory(GiB)": 111.15, + "step": 26060, + "train_speed(iter/s)": 0.448642 + }, + { + "acc": 0.65886192, + "epoch": 0.6612125824454591, + "grad_norm": 5.375, + "learning_rate": 7.997387967374821e-06, + "loss": 1.62640514, + "memory(GiB)": 111.15, + "step": 26065, + "train_speed(iter/s)": 0.448687 + }, + { + "acc": 0.63969588, + "epoch": 0.6613394216133942, + "grad_norm": 6.34375, + "learning_rate": 7.996548591941647e-06, + "loss": 1.68916245, + "memory(GiB)": 111.15, + "step": 26070, + "train_speed(iter/s)": 0.448733 + }, + { + "acc": 0.64784498, + "epoch": 0.6614662607813293, + "grad_norm": 6.59375, + "learning_rate": 7.995709084706884e-06, + "loss": 1.6145565, + "memory(GiB)": 111.15, + "step": 26075, + "train_speed(iter/s)": 0.448778 + }, + { + "acc": 0.63520079, + "epoch": 0.6615930999492643, + "grad_norm": 4.9375, + "learning_rate": 7.994869445707463e-06, + "loss": 1.66923103, + "memory(GiB)": 111.15, + "step": 26080, + "train_speed(iter/s)": 0.448823 + }, + { + "acc": 0.64625702, + "epoch": 0.6617199391171994, + "grad_norm": 5.09375, + "learning_rate": 7.994029674980313e-06, + "loss": 1.60941048, + "memory(GiB)": 111.15, + "step": 26085, + "train_speed(iter/s)": 0.448868 + }, + { + "acc": 0.66076159, + "epoch": 0.6618467782851345, + "grad_norm": 4.90625, + "learning_rate": 7.99318977256237e-06, + "loss": 1.58558264, + "memory(GiB)": 111.15, + "step": 26090, + "train_speed(iter/s)": 0.448914 + }, + { + "acc": 0.6447669, + "epoch": 0.6619736174530695, + "grad_norm": 4.28125, + "learning_rate": 7.992349738490576e-06, + "loss": 1.64601479, + "memory(GiB)": 111.15, + "step": 26095, + "train_speed(iter/s)": 0.448959 + }, + { + "acc": 0.66230564, + "epoch": 0.6621004566210046, + "grad_norm": 5.9375, + "learning_rate": 7.991509572801883e-06, + "loss": 1.59785137, + "memory(GiB)": 111.15, + "step": 26100, + "train_speed(iter/s)": 0.449004 + }, + { + "epoch": 0.6621004566210046, + "eval_acc": 0.6422600041693177, + "eval_loss": 1.6139650344848633, + "eval_runtime": 114.2115, + "eval_samples_per_second": 55.774, + "eval_steps_per_second": 27.887, + "step": 26100 + }, + { + "acc": 0.64322872, + "epoch": 0.6622272957889396, + "grad_norm": 4.46875, + "learning_rate": 7.990669275533241e-06, + "loss": 1.65127258, + "memory(GiB)": 111.15, + "step": 26105, + "train_speed(iter/s)": 0.448133 + }, + { + "acc": 0.63818436, + "epoch": 0.6623541349568747, + "grad_norm": 5.34375, + "learning_rate": 7.989828846721613e-06, + "loss": 1.66477165, + "memory(GiB)": 111.15, + "step": 26110, + "train_speed(iter/s)": 0.448177 + }, + { + "acc": 0.6419591, + "epoch": 0.6624809741248098, + "grad_norm": 4.9375, + "learning_rate": 7.98898828640396e-06, + "loss": 1.62112045, + "memory(GiB)": 111.15, + "step": 26115, + "train_speed(iter/s)": 0.448222 + }, + { + "acc": 0.64322033, + "epoch": 0.6626078132927448, + "grad_norm": 5.6875, + "learning_rate": 7.988147594617262e-06, + "loss": 1.65018234, + "memory(GiB)": 111.15, + "step": 26120, + "train_speed(iter/s)": 0.448266 + }, + { + "acc": 0.65331054, + "epoch": 0.6627346524606799, + "grad_norm": 4.90625, + "learning_rate": 7.987306771398489e-06, + "loss": 1.57978554, + "memory(GiB)": 111.15, + "step": 26125, + "train_speed(iter/s)": 0.448311 + }, + { + "acc": 0.6595263, + "epoch": 0.662861491628615, + "grad_norm": 5.59375, + "learning_rate": 7.986465816784628e-06, + "loss": 1.63917027, + "memory(GiB)": 111.15, + "step": 26130, + "train_speed(iter/s)": 0.448356 + }, + { + "acc": 0.64352474, + "epoch": 0.66298833079655, + "grad_norm": 4.6875, + "learning_rate": 7.985624730812667e-06, + "loss": 1.64071693, + "memory(GiB)": 111.15, + "step": 26135, + "train_speed(iter/s)": 0.448401 + }, + { + "acc": 0.64296012, + "epoch": 0.663115169964485, + "grad_norm": 6.09375, + "learning_rate": 7.984783513519601e-06, + "loss": 1.66208496, + "memory(GiB)": 111.15, + "step": 26140, + "train_speed(iter/s)": 0.448446 + }, + { + "acc": 0.64663854, + "epoch": 0.66324200913242, + "grad_norm": 5.71875, + "learning_rate": 7.98394216494243e-06, + "loss": 1.71035271, + "memory(GiB)": 111.15, + "step": 26145, + "train_speed(iter/s)": 0.44849 + }, + { + "acc": 0.6500186, + "epoch": 0.6633688483003551, + "grad_norm": 5.65625, + "learning_rate": 7.983100685118157e-06, + "loss": 1.663764, + "memory(GiB)": 111.15, + "step": 26150, + "train_speed(iter/s)": 0.448535 + }, + { + "acc": 0.65422521, + "epoch": 0.6634956874682902, + "grad_norm": 4.65625, + "learning_rate": 7.9822590740838e-06, + "loss": 1.56751575, + "memory(GiB)": 111.15, + "step": 26155, + "train_speed(iter/s)": 0.44858 + }, + { + "acc": 0.65565071, + "epoch": 0.6636225266362252, + "grad_norm": 4.875, + "learning_rate": 7.981417331876373e-06, + "loss": 1.58145561, + "memory(GiB)": 111.15, + "step": 26160, + "train_speed(iter/s)": 0.448625 + }, + { + "acc": 0.64795103, + "epoch": 0.6637493658041603, + "grad_norm": 4.8125, + "learning_rate": 7.980575458532901e-06, + "loss": 1.67585697, + "memory(GiB)": 111.15, + "step": 26165, + "train_speed(iter/s)": 0.44867 + }, + { + "acc": 0.64814091, + "epoch": 0.6638762049720954, + "grad_norm": 4.96875, + "learning_rate": 7.979733454090415e-06, + "loss": 1.63946495, + "memory(GiB)": 111.15, + "step": 26170, + "train_speed(iter/s)": 0.448715 + }, + { + "acc": 0.63238964, + "epoch": 0.6640030441400304, + "grad_norm": 5.6875, + "learning_rate": 7.978891318585947e-06, + "loss": 1.58985157, + "memory(GiB)": 111.15, + "step": 26175, + "train_speed(iter/s)": 0.44876 + }, + { + "acc": 0.65480785, + "epoch": 0.6641298833079655, + "grad_norm": 6.03125, + "learning_rate": 7.978049052056537e-06, + "loss": 1.65804634, + "memory(GiB)": 111.15, + "step": 26180, + "train_speed(iter/s)": 0.448804 + }, + { + "acc": 0.64617519, + "epoch": 0.6642567224759005, + "grad_norm": 5.75, + "learning_rate": 7.977206654539235e-06, + "loss": 1.64199295, + "memory(GiB)": 111.15, + "step": 26185, + "train_speed(iter/s)": 0.44885 + }, + { + "acc": 0.64133253, + "epoch": 0.6643835616438356, + "grad_norm": 4.9375, + "learning_rate": 7.976364126071092e-06, + "loss": 1.66103058, + "memory(GiB)": 111.15, + "step": 26190, + "train_speed(iter/s)": 0.448895 + }, + { + "acc": 0.62299881, + "epoch": 0.6645104008117707, + "grad_norm": 5.5, + "learning_rate": 7.975521466689166e-06, + "loss": 1.7413538, + "memory(GiB)": 111.15, + "step": 26195, + "train_speed(iter/s)": 0.448939 + }, + { + "acc": 0.64067678, + "epoch": 0.6646372399797057, + "grad_norm": 6.21875, + "learning_rate": 7.974678676430523e-06, + "loss": 1.62885437, + "memory(GiB)": 111.15, + "step": 26200, + "train_speed(iter/s)": 0.448984 + }, + { + "epoch": 0.6646372399797057, + "eval_acc": 0.6423928541737252, + "eval_loss": 1.61385977268219, + "eval_runtime": 113.1833, + "eval_samples_per_second": 56.28, + "eval_steps_per_second": 28.14, + "step": 26200 + }, + { + "acc": 0.66259947, + "epoch": 0.6647640791476408, + "grad_norm": 4.84375, + "learning_rate": 7.97383575533223e-06, + "loss": 1.56478291, + "memory(GiB)": 111.15, + "step": 26205, + "train_speed(iter/s)": 0.448125 + }, + { + "acc": 0.63975019, + "epoch": 0.6648909183155759, + "grad_norm": 5.125, + "learning_rate": 7.972992703431362e-06, + "loss": 1.59772549, + "memory(GiB)": 111.15, + "step": 26210, + "train_speed(iter/s)": 0.44817 + }, + { + "acc": 0.64407434, + "epoch": 0.6650177574835109, + "grad_norm": 5.75, + "learning_rate": 7.972149520765e-06, + "loss": 1.66438007, + "memory(GiB)": 111.15, + "step": 26215, + "train_speed(iter/s)": 0.448214 + }, + { + "acc": 0.64254656, + "epoch": 0.665144596651446, + "grad_norm": 5.125, + "learning_rate": 7.971306207370236e-06, + "loss": 1.64627686, + "memory(GiB)": 111.15, + "step": 26220, + "train_speed(iter/s)": 0.448259 + }, + { + "acc": 0.6550066, + "epoch": 0.665271435819381, + "grad_norm": 5.1875, + "learning_rate": 7.970462763284157e-06, + "loss": 1.62422829, + "memory(GiB)": 111.15, + "step": 26225, + "train_speed(iter/s)": 0.448304 + }, + { + "acc": 0.64689207, + "epoch": 0.6653982749873161, + "grad_norm": 5.71875, + "learning_rate": 7.969619188543865e-06, + "loss": 1.67792549, + "memory(GiB)": 111.15, + "step": 26230, + "train_speed(iter/s)": 0.448349 + }, + { + "acc": 0.648768, + "epoch": 0.6655251141552512, + "grad_norm": 5.9375, + "learning_rate": 7.968775483186462e-06, + "loss": 1.63606644, + "memory(GiB)": 111.15, + "step": 26235, + "train_speed(iter/s)": 0.448394 + }, + { + "acc": 0.66898346, + "epoch": 0.6656519533231862, + "grad_norm": 5.0, + "learning_rate": 7.967931647249058e-06, + "loss": 1.55273266, + "memory(GiB)": 111.15, + "step": 26240, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.63613982, + "epoch": 0.6657787924911213, + "grad_norm": 6.09375, + "learning_rate": 7.967087680768768e-06, + "loss": 1.62639103, + "memory(GiB)": 111.15, + "step": 26245, + "train_speed(iter/s)": 0.448485 + }, + { + "acc": 0.65079765, + "epoch": 0.6659056316590564, + "grad_norm": 5.1875, + "learning_rate": 7.966243583782718e-06, + "loss": 1.66153755, + "memory(GiB)": 111.15, + "step": 26250, + "train_speed(iter/s)": 0.448529 + }, + { + "acc": 0.64202251, + "epoch": 0.6660324708269914, + "grad_norm": 4.96875, + "learning_rate": 7.96539935632803e-06, + "loss": 1.65456772, + "memory(GiB)": 111.15, + "step": 26255, + "train_speed(iter/s)": 0.448574 + }, + { + "acc": 0.6477788, + "epoch": 0.6661593099949265, + "grad_norm": 5.53125, + "learning_rate": 7.964554998441839e-06, + "loss": 1.62010269, + "memory(GiB)": 111.15, + "step": 26260, + "train_speed(iter/s)": 0.44862 + }, + { + "acc": 0.66394501, + "epoch": 0.6662861491628614, + "grad_norm": 5.4375, + "learning_rate": 7.963710510161282e-06, + "loss": 1.60302467, + "memory(GiB)": 111.15, + "step": 26265, + "train_speed(iter/s)": 0.448665 + }, + { + "acc": 0.66264076, + "epoch": 0.6664129883307965, + "grad_norm": 5.125, + "learning_rate": 7.962865891523508e-06, + "loss": 1.58006821, + "memory(GiB)": 111.15, + "step": 26270, + "train_speed(iter/s)": 0.448709 + }, + { + "acc": 0.64388304, + "epoch": 0.6665398274987316, + "grad_norm": 6.0625, + "learning_rate": 7.96202114256566e-06, + "loss": 1.72715912, + "memory(GiB)": 111.15, + "step": 26275, + "train_speed(iter/s)": 0.448753 + }, + { + "acc": 0.64347038, + "epoch": 0.6666666666666666, + "grad_norm": 5.0625, + "learning_rate": 7.961176263324902e-06, + "loss": 1.6466835, + "memory(GiB)": 111.15, + "step": 26280, + "train_speed(iter/s)": 0.448798 + }, + { + "acc": 0.64383249, + "epoch": 0.6667935058346017, + "grad_norm": 5.09375, + "learning_rate": 7.960331253838387e-06, + "loss": 1.61689262, + "memory(GiB)": 111.15, + "step": 26285, + "train_speed(iter/s)": 0.448843 + }, + { + "acc": 0.6438941, + "epoch": 0.6669203450025368, + "grad_norm": 4.375, + "learning_rate": 7.95948611414329e-06, + "loss": 1.61037636, + "memory(GiB)": 111.15, + "step": 26290, + "train_speed(iter/s)": 0.448888 + }, + { + "acc": 0.64680853, + "epoch": 0.6670471841704718, + "grad_norm": 5.59375, + "learning_rate": 7.958640844276776e-06, + "loss": 1.69050045, + "memory(GiB)": 111.15, + "step": 26295, + "train_speed(iter/s)": 0.448932 + }, + { + "acc": 0.63537717, + "epoch": 0.6671740233384069, + "grad_norm": 5.15625, + "learning_rate": 7.957795444276033e-06, + "loss": 1.67635078, + "memory(GiB)": 111.15, + "step": 26300, + "train_speed(iter/s)": 0.448977 + }, + { + "epoch": 0.6671740233384069, + "eval_acc": 0.6423264291715215, + "eval_loss": 1.614088535308838, + "eval_runtime": 114.6096, + "eval_samples_per_second": 55.58, + "eval_steps_per_second": 27.79, + "step": 26300 + }, + { + "acc": 0.64826775, + "epoch": 0.6673008625063419, + "grad_norm": 5.90625, + "learning_rate": 7.956949914178239e-06, + "loss": 1.60827026, + "memory(GiB)": 111.15, + "step": 26305, + "train_speed(iter/s)": 0.448111 + }, + { + "acc": 0.6521327, + "epoch": 0.667427701674277, + "grad_norm": 5.5, + "learning_rate": 7.956104254020587e-06, + "loss": 1.57084045, + "memory(GiB)": 111.15, + "step": 26310, + "train_speed(iter/s)": 0.448156 + }, + { + "acc": 0.65590334, + "epoch": 0.6675545408422121, + "grad_norm": 5.03125, + "learning_rate": 7.95525846384027e-06, + "loss": 1.62026176, + "memory(GiB)": 111.15, + "step": 26315, + "train_speed(iter/s)": 0.448201 + }, + { + "acc": 0.65590239, + "epoch": 0.6676813800101471, + "grad_norm": 4.09375, + "learning_rate": 7.954412543674493e-06, + "loss": 1.57253399, + "memory(GiB)": 111.15, + "step": 26320, + "train_speed(iter/s)": 0.448245 + }, + { + "acc": 0.64606524, + "epoch": 0.6678082191780822, + "grad_norm": 5.875, + "learning_rate": 7.95356649356046e-06, + "loss": 1.6988472, + "memory(GiB)": 111.15, + "step": 26325, + "train_speed(iter/s)": 0.44829 + }, + { + "acc": 0.63687487, + "epoch": 0.6679350583460173, + "grad_norm": 6.28125, + "learning_rate": 7.952720313535387e-06, + "loss": 1.67288494, + "memory(GiB)": 111.15, + "step": 26330, + "train_speed(iter/s)": 0.448335 + }, + { + "acc": 0.64349661, + "epoch": 0.6680618975139523, + "grad_norm": 5.0625, + "learning_rate": 7.951874003636492e-06, + "loss": 1.70083122, + "memory(GiB)": 111.15, + "step": 26335, + "train_speed(iter/s)": 0.44838 + }, + { + "acc": 0.65932612, + "epoch": 0.6681887366818874, + "grad_norm": 4.4375, + "learning_rate": 7.951027563901e-06, + "loss": 1.53188019, + "memory(GiB)": 111.15, + "step": 26340, + "train_speed(iter/s)": 0.448424 + }, + { + "acc": 0.63367157, + "epoch": 0.6683155758498224, + "grad_norm": 6.875, + "learning_rate": 7.950180994366138e-06, + "loss": 1.69134884, + "memory(GiB)": 111.15, + "step": 26345, + "train_speed(iter/s)": 0.448469 + }, + { + "acc": 0.67218637, + "epoch": 0.6684424150177575, + "grad_norm": 5.4375, + "learning_rate": 7.949334295069147e-06, + "loss": 1.53318348, + "memory(GiB)": 111.15, + "step": 26350, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.64261093, + "epoch": 0.6685692541856926, + "grad_norm": 6.15625, + "learning_rate": 7.948487466047263e-06, + "loss": 1.68601704, + "memory(GiB)": 111.15, + "step": 26355, + "train_speed(iter/s)": 0.448559 + }, + { + "acc": 0.64814944, + "epoch": 0.6686960933536276, + "grad_norm": 6.53125, + "learning_rate": 7.947640507337737e-06, + "loss": 1.62900391, + "memory(GiB)": 111.15, + "step": 26360, + "train_speed(iter/s)": 0.448604 + }, + { + "acc": 0.63021793, + "epoch": 0.6688229325215627, + "grad_norm": 5.40625, + "learning_rate": 7.946793418977821e-06, + "loss": 1.62759018, + "memory(GiB)": 111.15, + "step": 26365, + "train_speed(iter/s)": 0.448648 + }, + { + "acc": 0.63412657, + "epoch": 0.6689497716894978, + "grad_norm": 5.03125, + "learning_rate": 7.945946201004775e-06, + "loss": 1.64692326, + "memory(GiB)": 111.15, + "step": 26370, + "train_speed(iter/s)": 0.448693 + }, + { + "acc": 0.63767157, + "epoch": 0.6690766108574328, + "grad_norm": 4.875, + "learning_rate": 7.945098853455862e-06, + "loss": 1.62972565, + "memory(GiB)": 111.15, + "step": 26375, + "train_speed(iter/s)": 0.448738 + }, + { + "acc": 0.66610022, + "epoch": 0.6692034500253679, + "grad_norm": 4.25, + "learning_rate": 7.944251376368352e-06, + "loss": 1.58844299, + "memory(GiB)": 111.15, + "step": 26380, + "train_speed(iter/s)": 0.448783 + }, + { + "acc": 0.64082756, + "epoch": 0.6693302891933028, + "grad_norm": 5.28125, + "learning_rate": 7.943403769779523e-06, + "loss": 1.68669739, + "memory(GiB)": 111.15, + "step": 26385, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.64510326, + "epoch": 0.6694571283612379, + "grad_norm": 6.15625, + "learning_rate": 7.942556033726654e-06, + "loss": 1.70049896, + "memory(GiB)": 111.15, + "step": 26390, + "train_speed(iter/s)": 0.448872 + }, + { + "acc": 0.66431866, + "epoch": 0.669583967529173, + "grad_norm": 5.0, + "learning_rate": 7.941708168247033e-06, + "loss": 1.63441162, + "memory(GiB)": 111.15, + "step": 26395, + "train_speed(iter/s)": 0.448917 + }, + { + "acc": 0.65343018, + "epoch": 0.669710806697108, + "grad_norm": 5.28125, + "learning_rate": 7.940860173377952e-06, + "loss": 1.64482536, + "memory(GiB)": 111.15, + "step": 26400, + "train_speed(iter/s)": 0.448961 + }, + { + "epoch": 0.669710806697108, + "eval_acc": 0.6423489886062321, + "eval_loss": 1.6135584115982056, + "eval_runtime": 113.5281, + "eval_samples_per_second": 56.109, + "eval_steps_per_second": 28.055, + "step": 26400 + }, + { + "acc": 0.65042486, + "epoch": 0.6698376458650431, + "grad_norm": 5.625, + "learning_rate": 7.940012049156711e-06, + "loss": 1.63586922, + "memory(GiB)": 111.15, + "step": 26405, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.65657606, + "epoch": 0.6699644850329782, + "grad_norm": 4.59375, + "learning_rate": 7.939163795620614e-06, + "loss": 1.6177639, + "memory(GiB)": 111.15, + "step": 26410, + "train_speed(iter/s)": 0.448152 + }, + { + "acc": 0.6422039, + "epoch": 0.6700913242009132, + "grad_norm": 5.71875, + "learning_rate": 7.938315412806971e-06, + "loss": 1.66981678, + "memory(GiB)": 111.15, + "step": 26415, + "train_speed(iter/s)": 0.448197 + }, + { + "acc": 0.66614051, + "epoch": 0.6702181633688483, + "grad_norm": 5.21875, + "learning_rate": 7.937466900753098e-06, + "loss": 1.5737752, + "memory(GiB)": 111.15, + "step": 26420, + "train_speed(iter/s)": 0.448242 + }, + { + "acc": 0.64501114, + "epoch": 0.6703450025367833, + "grad_norm": 6.34375, + "learning_rate": 7.936618259496316e-06, + "loss": 1.65295448, + "memory(GiB)": 111.15, + "step": 26425, + "train_speed(iter/s)": 0.448287 + }, + { + "acc": 0.64946413, + "epoch": 0.6704718417047184, + "grad_norm": 5.3125, + "learning_rate": 7.935769489073952e-06, + "loss": 1.63194351, + "memory(GiB)": 111.15, + "step": 26430, + "train_speed(iter/s)": 0.448332 + }, + { + "acc": 0.6600729, + "epoch": 0.6705986808726535, + "grad_norm": 5.25, + "learning_rate": 7.934920589523336e-06, + "loss": 1.52198572, + "memory(GiB)": 111.15, + "step": 26435, + "train_speed(iter/s)": 0.448377 + }, + { + "acc": 0.64333868, + "epoch": 0.6707255200405885, + "grad_norm": 5.15625, + "learning_rate": 7.934071560881812e-06, + "loss": 1.68957062, + "memory(GiB)": 111.15, + "step": 26440, + "train_speed(iter/s)": 0.448422 + }, + { + "acc": 0.65825977, + "epoch": 0.6708523592085236, + "grad_norm": 5.125, + "learning_rate": 7.93322240318672e-06, + "loss": 1.57723999, + "memory(GiB)": 111.15, + "step": 26445, + "train_speed(iter/s)": 0.448467 + }, + { + "acc": 0.65111303, + "epoch": 0.6709791983764587, + "grad_norm": 4.9375, + "learning_rate": 7.93237311647541e-06, + "loss": 1.65557365, + "memory(GiB)": 111.15, + "step": 26450, + "train_speed(iter/s)": 0.448512 + }, + { + "acc": 0.66194887, + "epoch": 0.6711060375443937, + "grad_norm": 5.25, + "learning_rate": 7.93152370078524e-06, + "loss": 1.52688656, + "memory(GiB)": 111.15, + "step": 26455, + "train_speed(iter/s)": 0.448556 + }, + { + "acc": 0.65077381, + "epoch": 0.6712328767123288, + "grad_norm": 4.6875, + "learning_rate": 7.930674156153569e-06, + "loss": 1.59578094, + "memory(GiB)": 111.15, + "step": 26460, + "train_speed(iter/s)": 0.448602 + }, + { + "acc": 0.64809856, + "epoch": 0.6713597158802638, + "grad_norm": 4.96875, + "learning_rate": 7.929824482617763e-06, + "loss": 1.65049095, + "memory(GiB)": 111.15, + "step": 26465, + "train_speed(iter/s)": 0.448646 + }, + { + "acc": 0.64169469, + "epoch": 0.6714865550481989, + "grad_norm": 6.5625, + "learning_rate": 7.928974680215196e-06, + "loss": 1.71370316, + "memory(GiB)": 111.15, + "step": 26470, + "train_speed(iter/s)": 0.448692 + }, + { + "acc": 0.63013306, + "epoch": 0.671613394216134, + "grad_norm": 5.15625, + "learning_rate": 7.928124748983244e-06, + "loss": 1.67696419, + "memory(GiB)": 111.15, + "step": 26475, + "train_speed(iter/s)": 0.448737 + }, + { + "acc": 0.63988948, + "epoch": 0.671740233384069, + "grad_norm": 6.125, + "learning_rate": 7.927274688959294e-06, + "loss": 1.70424595, + "memory(GiB)": 111.15, + "step": 26480, + "train_speed(iter/s)": 0.448782 + }, + { + "acc": 0.64109745, + "epoch": 0.6718670725520041, + "grad_norm": 6.3125, + "learning_rate": 7.926424500180734e-06, + "loss": 1.69590225, + "memory(GiB)": 111.15, + "step": 26485, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.67884693, + "epoch": 0.6719939117199392, + "grad_norm": 5.59375, + "learning_rate": 7.92557418268496e-06, + "loss": 1.56500874, + "memory(GiB)": 111.15, + "step": 26490, + "train_speed(iter/s)": 0.448872 + }, + { + "acc": 0.65196943, + "epoch": 0.6721207508878742, + "grad_norm": 4.9375, + "learning_rate": 7.92472373650937e-06, + "loss": 1.60157528, + "memory(GiB)": 111.15, + "step": 26495, + "train_speed(iter/s)": 0.448917 + }, + { + "acc": 0.6288106, + "epoch": 0.6722475900558093, + "grad_norm": 4.71875, + "learning_rate": 7.923873161691373e-06, + "loss": 1.71914501, + "memory(GiB)": 111.15, + "step": 26500, + "train_speed(iter/s)": 0.448961 + }, + { + "epoch": 0.6722475900558093, + "eval_acc": 0.6423748901794185, + "eval_loss": 1.613545298576355, + "eval_runtime": 113.8464, + "eval_samples_per_second": 55.953, + "eval_steps_per_second": 27.976, + "step": 26500 + }, + { + "acc": 0.6656064, + "epoch": 0.6723744292237442, + "grad_norm": 5.3125, + "learning_rate": 7.923022458268379e-06, + "loss": 1.57511806, + "memory(GiB)": 111.15, + "step": 26505, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.6608655, + "epoch": 0.6725012683916793, + "grad_norm": 5.34375, + "learning_rate": 7.922171626277809e-06, + "loss": 1.63352432, + "memory(GiB)": 111.15, + "step": 26510, + "train_speed(iter/s)": 0.448152 + }, + { + "acc": 0.64728012, + "epoch": 0.6726281075596144, + "grad_norm": 4.3125, + "learning_rate": 7.921320665757081e-06, + "loss": 1.60305367, + "memory(GiB)": 111.15, + "step": 26515, + "train_speed(iter/s)": 0.448196 + }, + { + "acc": 0.65221214, + "epoch": 0.6727549467275494, + "grad_norm": 5.65625, + "learning_rate": 7.920469576743631e-06, + "loss": 1.58747787, + "memory(GiB)": 111.15, + "step": 26520, + "train_speed(iter/s)": 0.448241 + }, + { + "acc": 0.65201778, + "epoch": 0.6728817858954845, + "grad_norm": 5.90625, + "learning_rate": 7.919618359274888e-06, + "loss": 1.64219704, + "memory(GiB)": 111.15, + "step": 26525, + "train_speed(iter/s)": 0.448286 + }, + { + "acc": 0.6585989, + "epoch": 0.6730086250634196, + "grad_norm": 5.96875, + "learning_rate": 7.918767013388295e-06, + "loss": 1.65853882, + "memory(GiB)": 111.15, + "step": 26530, + "train_speed(iter/s)": 0.44833 + }, + { + "acc": 0.63881874, + "epoch": 0.6731354642313546, + "grad_norm": 5.46875, + "learning_rate": 7.917915539121297e-06, + "loss": 1.66516647, + "memory(GiB)": 111.15, + "step": 26535, + "train_speed(iter/s)": 0.448375 + }, + { + "acc": 0.62412648, + "epoch": 0.6732623033992897, + "grad_norm": 4.9375, + "learning_rate": 7.917063936511347e-06, + "loss": 1.7445303, + "memory(GiB)": 111.15, + "step": 26540, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.63924494, + "epoch": 0.6733891425672247, + "grad_norm": 6.03125, + "learning_rate": 7.9162122055959e-06, + "loss": 1.67609749, + "memory(GiB)": 111.15, + "step": 26545, + "train_speed(iter/s)": 0.448464 + }, + { + "acc": 0.65716681, + "epoch": 0.6735159817351598, + "grad_norm": 5.4375, + "learning_rate": 7.91536034641242e-06, + "loss": 1.61700439, + "memory(GiB)": 111.15, + "step": 26550, + "train_speed(iter/s)": 0.448508 + }, + { + "acc": 0.64531746, + "epoch": 0.6736428209030949, + "grad_norm": 8.4375, + "learning_rate": 7.914508358998376e-06, + "loss": 1.67286625, + "memory(GiB)": 111.15, + "step": 26555, + "train_speed(iter/s)": 0.448553 + }, + { + "acc": 0.66955552, + "epoch": 0.6737696600710299, + "grad_norm": 5.34375, + "learning_rate": 7.913656243391243e-06, + "loss": 1.5655899, + "memory(GiB)": 111.15, + "step": 26560, + "train_speed(iter/s)": 0.448598 + }, + { + "acc": 0.64335737, + "epoch": 0.673896499238965, + "grad_norm": 5.09375, + "learning_rate": 7.9128039996285e-06, + "loss": 1.61546822, + "memory(GiB)": 111.15, + "step": 26565, + "train_speed(iter/s)": 0.448643 + }, + { + "acc": 0.63856525, + "epoch": 0.6740233384069001, + "grad_norm": 4.8125, + "learning_rate": 7.911951627747633e-06, + "loss": 1.66797123, + "memory(GiB)": 111.15, + "step": 26570, + "train_speed(iter/s)": 0.448687 + }, + { + "acc": 0.63949847, + "epoch": 0.6741501775748351, + "grad_norm": 5.5625, + "learning_rate": 7.91109912778613e-06, + "loss": 1.68856239, + "memory(GiB)": 111.15, + "step": 26575, + "train_speed(iter/s)": 0.448732 + }, + { + "acc": 0.64839354, + "epoch": 0.6742770167427702, + "grad_norm": 4.84375, + "learning_rate": 7.910246499781492e-06, + "loss": 1.61389961, + "memory(GiB)": 111.15, + "step": 26580, + "train_speed(iter/s)": 0.448776 + }, + { + "acc": 0.64089952, + "epoch": 0.6744038559107052, + "grad_norm": 5.0625, + "learning_rate": 7.90939374377122e-06, + "loss": 1.683288, + "memory(GiB)": 111.15, + "step": 26585, + "train_speed(iter/s)": 0.448821 + }, + { + "acc": 0.64676743, + "epoch": 0.6745306950786403, + "grad_norm": 4.8125, + "learning_rate": 7.908540859792821e-06, + "loss": 1.58397045, + "memory(GiB)": 111.15, + "step": 26590, + "train_speed(iter/s)": 0.448865 + }, + { + "acc": 0.65355453, + "epoch": 0.6746575342465754, + "grad_norm": 5.40625, + "learning_rate": 7.907687847883809e-06, + "loss": 1.5775692, + "memory(GiB)": 111.15, + "step": 26595, + "train_speed(iter/s)": 0.44891 + }, + { + "acc": 0.64727859, + "epoch": 0.6747843734145104, + "grad_norm": 5.6875, + "learning_rate": 7.906834708081703e-06, + "loss": 1.62429276, + "memory(GiB)": 111.15, + "step": 26600, + "train_speed(iter/s)": 0.448954 + }, + { + "epoch": 0.6747843734145104, + "eval_acc": 0.6424104004007224, + "eval_loss": 1.6134413480758667, + "eval_runtime": 113.9327, + "eval_samples_per_second": 55.91, + "eval_steps_per_second": 27.955, + "step": 26600 + }, + { + "acc": 0.65623484, + "epoch": 0.6749112125824455, + "grad_norm": 5.5, + "learning_rate": 7.90598144042403e-06, + "loss": 1.61602821, + "memory(GiB)": 111.15, + "step": 26605, + "train_speed(iter/s)": 0.448103 + }, + { + "acc": 0.65157695, + "epoch": 0.6750380517503806, + "grad_norm": 5.90625, + "learning_rate": 7.905128044948318e-06, + "loss": 1.64370155, + "memory(GiB)": 111.15, + "step": 26610, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.63859291, + "epoch": 0.6751648909183156, + "grad_norm": 4.5625, + "learning_rate": 7.904274521692104e-06, + "loss": 1.67388058, + "memory(GiB)": 111.15, + "step": 26615, + "train_speed(iter/s)": 0.448191 + }, + { + "acc": 0.64978991, + "epoch": 0.6752917300862507, + "grad_norm": 5.0625, + "learning_rate": 7.90342087069293e-06, + "loss": 1.6262394, + "memory(GiB)": 111.15, + "step": 26620, + "train_speed(iter/s)": 0.448235 + }, + { + "acc": 0.64784198, + "epoch": 0.6754185692541856, + "grad_norm": 5.28125, + "learning_rate": 7.902567091988343e-06, + "loss": 1.5625349, + "memory(GiB)": 111.15, + "step": 26625, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.6521122, + "epoch": 0.6755454084221207, + "grad_norm": 5.8125, + "learning_rate": 7.901713185615898e-06, + "loss": 1.64546432, + "memory(GiB)": 111.15, + "step": 26630, + "train_speed(iter/s)": 0.448323 + }, + { + "acc": 0.64371786, + "epoch": 0.6756722475900558, + "grad_norm": 5.8125, + "learning_rate": 7.90085915161315e-06, + "loss": 1.63861961, + "memory(GiB)": 111.15, + "step": 26635, + "train_speed(iter/s)": 0.448368 + }, + { + "acc": 0.64883447, + "epoch": 0.6757990867579908, + "grad_norm": 4.84375, + "learning_rate": 7.900004990017667e-06, + "loss": 1.66120853, + "memory(GiB)": 111.15, + "step": 26640, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.63555164, + "epoch": 0.6759259259259259, + "grad_norm": 6.3125, + "learning_rate": 7.899150700867014e-06, + "loss": 1.71150475, + "memory(GiB)": 111.15, + "step": 26645, + "train_speed(iter/s)": 0.448456 + }, + { + "acc": 0.64265041, + "epoch": 0.676052765093861, + "grad_norm": 5.5625, + "learning_rate": 7.898296284198772e-06, + "loss": 1.60275421, + "memory(GiB)": 111.15, + "step": 26650, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.65254798, + "epoch": 0.676179604261796, + "grad_norm": 5.8125, + "learning_rate": 7.897441740050518e-06, + "loss": 1.65432892, + "memory(GiB)": 111.15, + "step": 26655, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.62976899, + "epoch": 0.6763064434297311, + "grad_norm": 4.84375, + "learning_rate": 7.89658706845984e-06, + "loss": 1.65171547, + "memory(GiB)": 111.15, + "step": 26660, + "train_speed(iter/s)": 0.44859 + }, + { + "acc": 0.64569702, + "epoch": 0.6764332825976661, + "grad_norm": 5.40625, + "learning_rate": 7.89573226946433e-06, + "loss": 1.63700142, + "memory(GiB)": 111.15, + "step": 26665, + "train_speed(iter/s)": 0.448634 + }, + { + "acc": 0.64668894, + "epoch": 0.6765601217656012, + "grad_norm": 6.53125, + "learning_rate": 7.89487734310159e-06, + "loss": 1.65378914, + "memory(GiB)": 111.15, + "step": 26670, + "train_speed(iter/s)": 0.448679 + }, + { + "acc": 0.64808354, + "epoch": 0.6766869609335363, + "grad_norm": 4.90625, + "learning_rate": 7.894022289409216e-06, + "loss": 1.62835331, + "memory(GiB)": 111.15, + "step": 26675, + "train_speed(iter/s)": 0.448723 + }, + { + "acc": 0.63534608, + "epoch": 0.6768138001014713, + "grad_norm": 4.53125, + "learning_rate": 7.893167108424822e-06, + "loss": 1.68508072, + "memory(GiB)": 111.15, + "step": 26680, + "train_speed(iter/s)": 0.448767 + }, + { + "acc": 0.6540844, + "epoch": 0.6769406392694064, + "grad_norm": 5.25, + "learning_rate": 7.89231180018602e-06, + "loss": 1.55906868, + "memory(GiB)": 111.15, + "step": 26685, + "train_speed(iter/s)": 0.448812 + }, + { + "acc": 0.65094805, + "epoch": 0.6770674784373415, + "grad_norm": 5.3125, + "learning_rate": 7.891456364730434e-06, + "loss": 1.65201187, + "memory(GiB)": 111.15, + "step": 26690, + "train_speed(iter/s)": 0.448856 + }, + { + "acc": 0.64199886, + "epoch": 0.6771943176052765, + "grad_norm": 5.0625, + "learning_rate": 7.890600802095686e-06, + "loss": 1.70462055, + "memory(GiB)": 111.15, + "step": 26695, + "train_speed(iter/s)": 0.448901 + }, + { + "acc": 0.64379416, + "epoch": 0.6773211567732116, + "grad_norm": 5.875, + "learning_rate": 7.889745112319411e-06, + "loss": 1.5928628, + "memory(GiB)": 111.15, + "step": 26700, + "train_speed(iter/s)": 0.448946 + }, + { + "epoch": 0.6773211567732116, + "eval_acc": 0.642385334362155, + "eval_loss": 1.6135154962539673, + "eval_runtime": 113.3143, + "eval_samples_per_second": 56.215, + "eval_steps_per_second": 28.108, + "step": 26700 + }, + { + "acc": 0.64770875, + "epoch": 0.6774479959411466, + "grad_norm": 4.875, + "learning_rate": 7.888889295439244e-06, + "loss": 1.63203945, + "memory(GiB)": 111.15, + "step": 26705, + "train_speed(iter/s)": 0.448102 + }, + { + "acc": 0.64732485, + "epoch": 0.6775748351090817, + "grad_norm": 5.46875, + "learning_rate": 7.888033351492827e-06, + "loss": 1.65415421, + "memory(GiB)": 111.15, + "step": 26710, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.63932056, + "epoch": 0.6777016742770168, + "grad_norm": 4.96875, + "learning_rate": 7.887177280517808e-06, + "loss": 1.70096416, + "memory(GiB)": 111.15, + "step": 26715, + "train_speed(iter/s)": 0.448191 + }, + { + "acc": 0.65437183, + "epoch": 0.6778285134449518, + "grad_norm": 5.625, + "learning_rate": 7.886321082551845e-06, + "loss": 1.57246151, + "memory(GiB)": 111.15, + "step": 26720, + "train_speed(iter/s)": 0.448236 + }, + { + "acc": 0.66651793, + "epoch": 0.6779553526128869, + "grad_norm": 5.34375, + "learning_rate": 7.88546475763259e-06, + "loss": 1.58562231, + "memory(GiB)": 111.15, + "step": 26725, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.65972404, + "epoch": 0.678082191780822, + "grad_norm": 4.34375, + "learning_rate": 7.884608305797716e-06, + "loss": 1.58518448, + "memory(GiB)": 111.15, + "step": 26730, + "train_speed(iter/s)": 0.448323 + }, + { + "acc": 0.63886676, + "epoch": 0.678209030948757, + "grad_norm": 4.8125, + "learning_rate": 7.883751727084888e-06, + "loss": 1.66441574, + "memory(GiB)": 111.15, + "step": 26735, + "train_speed(iter/s)": 0.448368 + }, + { + "acc": 0.6452435, + "epoch": 0.678335870116692, + "grad_norm": 5.0, + "learning_rate": 7.882895021531784e-06, + "loss": 1.68922691, + "memory(GiB)": 111.15, + "step": 26740, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.64843884, + "epoch": 0.678462709284627, + "grad_norm": 5.96875, + "learning_rate": 7.882038189176085e-06, + "loss": 1.64935036, + "memory(GiB)": 111.15, + "step": 26745, + "train_speed(iter/s)": 0.448457 + }, + { + "acc": 0.65742407, + "epoch": 0.6785895484525621, + "grad_norm": 5.25, + "learning_rate": 7.881181230055481e-06, + "loss": 1.61824284, + "memory(GiB)": 111.15, + "step": 26750, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.65716667, + "epoch": 0.6787163876204972, + "grad_norm": 8.9375, + "learning_rate": 7.880324144207663e-06, + "loss": 1.61386299, + "memory(GiB)": 111.15, + "step": 26755, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.65877094, + "epoch": 0.6788432267884322, + "grad_norm": 5.15625, + "learning_rate": 7.879466931670328e-06, + "loss": 1.67288265, + "memory(GiB)": 111.15, + "step": 26760, + "train_speed(iter/s)": 0.44859 + }, + { + "acc": 0.63891597, + "epoch": 0.6789700659563673, + "grad_norm": 6.28125, + "learning_rate": 7.878609592481182e-06, + "loss": 1.74146385, + "memory(GiB)": 111.15, + "step": 26765, + "train_speed(iter/s)": 0.448634 + }, + { + "acc": 0.65627451, + "epoch": 0.6790969051243024, + "grad_norm": 5.34375, + "learning_rate": 7.877752126677933e-06, + "loss": 1.63038044, + "memory(GiB)": 111.15, + "step": 26770, + "train_speed(iter/s)": 0.448678 + }, + { + "acc": 0.66876287, + "epoch": 0.6792237442922374, + "grad_norm": 5.0625, + "learning_rate": 7.876894534298298e-06, + "loss": 1.49921303, + "memory(GiB)": 111.15, + "step": 26775, + "train_speed(iter/s)": 0.448723 + }, + { + "acc": 0.65520697, + "epoch": 0.6793505834601725, + "grad_norm": 5.53125, + "learning_rate": 7.87603681538e-06, + "loss": 1.5803936, + "memory(GiB)": 111.15, + "step": 26780, + "train_speed(iter/s)": 0.448767 + }, + { + "acc": 0.6327745, + "epoch": 0.6794774226281075, + "grad_norm": 5.53125, + "learning_rate": 7.875178969960757e-06, + "loss": 1.74032516, + "memory(GiB)": 111.15, + "step": 26785, + "train_speed(iter/s)": 0.448811 + }, + { + "acc": 0.64561715, + "epoch": 0.6796042617960426, + "grad_norm": 4.71875, + "learning_rate": 7.87432099807831e-06, + "loss": 1.66211128, + "memory(GiB)": 111.15, + "step": 26790, + "train_speed(iter/s)": 0.448855 + }, + { + "acc": 0.64776425, + "epoch": 0.6797311009639777, + "grad_norm": 5.3125, + "learning_rate": 7.87346289977039e-06, + "loss": 1.65970154, + "memory(GiB)": 111.15, + "step": 26795, + "train_speed(iter/s)": 0.448899 + }, + { + "acc": 0.66038513, + "epoch": 0.6798579401319127, + "grad_norm": 5.53125, + "learning_rate": 7.872604675074745e-06, + "loss": 1.6227478, + "memory(GiB)": 111.15, + "step": 26800, + "train_speed(iter/s)": 0.448944 + }, + { + "epoch": 0.6798579401319127, + "eval_acc": 0.6423523307447079, + "eval_loss": 1.6134400367736816, + "eval_runtime": 113.7079, + "eval_samples_per_second": 56.021, + "eval_steps_per_second": 28.01, + "step": 26800 + }, + { + "acc": 0.63957696, + "epoch": 0.6799847792998478, + "grad_norm": 5.59375, + "learning_rate": 7.871746324029119e-06, + "loss": 1.61455307, + "memory(GiB)": 111.15, + "step": 26805, + "train_speed(iter/s)": 0.4481 + }, + { + "acc": 0.65515013, + "epoch": 0.6801116184677829, + "grad_norm": 5.34375, + "learning_rate": 7.87088784667127e-06, + "loss": 1.60473118, + "memory(GiB)": 111.15, + "step": 26810, + "train_speed(iter/s)": 0.448144 + }, + { + "acc": 0.62987185, + "epoch": 0.6802384576357179, + "grad_norm": 5.1875, + "learning_rate": 7.870029243038955e-06, + "loss": 1.71166191, + "memory(GiB)": 111.15, + "step": 26815, + "train_speed(iter/s)": 0.448188 + }, + { + "acc": 0.63865252, + "epoch": 0.680365296803653, + "grad_norm": 5.5, + "learning_rate": 7.869170513169941e-06, + "loss": 1.66307049, + "memory(GiB)": 111.15, + "step": 26820, + "train_speed(iter/s)": 0.448232 + }, + { + "acc": 0.62583017, + "epoch": 0.680492135971588, + "grad_norm": 5.09375, + "learning_rate": 7.868311657101996e-06, + "loss": 1.6472187, + "memory(GiB)": 111.15, + "step": 26825, + "train_speed(iter/s)": 0.448277 + }, + { + "acc": 0.64459772, + "epoch": 0.6806189751395231, + "grad_norm": 6.8125, + "learning_rate": 7.8674526748729e-06, + "loss": 1.66756802, + "memory(GiB)": 111.15, + "step": 26830, + "train_speed(iter/s)": 0.448321 + }, + { + "acc": 0.67718077, + "epoch": 0.6807458143074582, + "grad_norm": 6.1875, + "learning_rate": 7.866593566520432e-06, + "loss": 1.53802967, + "memory(GiB)": 111.15, + "step": 26835, + "train_speed(iter/s)": 0.448365 + }, + { + "acc": 0.65185323, + "epoch": 0.6808726534753932, + "grad_norm": 5.09375, + "learning_rate": 7.865734332082382e-06, + "loss": 1.68677731, + "memory(GiB)": 111.15, + "step": 26840, + "train_speed(iter/s)": 0.448409 + }, + { + "acc": 0.66086502, + "epoch": 0.6809994926433283, + "grad_norm": 6.21875, + "learning_rate": 7.86487497159654e-06, + "loss": 1.56068201, + "memory(GiB)": 111.15, + "step": 26845, + "train_speed(iter/s)": 0.448453 + }, + { + "acc": 0.64579043, + "epoch": 0.6811263318112634, + "grad_norm": 5.125, + "learning_rate": 7.864015485100706e-06, + "loss": 1.6240345, + "memory(GiB)": 111.15, + "step": 26850, + "train_speed(iter/s)": 0.448497 + }, + { + "acc": 0.64868155, + "epoch": 0.6812531709791984, + "grad_norm": 5.21875, + "learning_rate": 7.863155872632685e-06, + "loss": 1.56953449, + "memory(GiB)": 111.15, + "step": 26855, + "train_speed(iter/s)": 0.448541 + }, + { + "acc": 0.65022974, + "epoch": 0.6813800101471335, + "grad_norm": 5.90625, + "learning_rate": 7.862296134230287e-06, + "loss": 1.59436302, + "memory(GiB)": 111.15, + "step": 26860, + "train_speed(iter/s)": 0.448585 + }, + { + "acc": 0.64992528, + "epoch": 0.6815068493150684, + "grad_norm": 5.625, + "learning_rate": 7.861436269931322e-06, + "loss": 1.61498337, + "memory(GiB)": 111.15, + "step": 26865, + "train_speed(iter/s)": 0.44863 + }, + { + "acc": 0.66036196, + "epoch": 0.6816336884830035, + "grad_norm": 5.03125, + "learning_rate": 7.860576279773617e-06, + "loss": 1.57613449, + "memory(GiB)": 111.15, + "step": 26870, + "train_speed(iter/s)": 0.448674 + }, + { + "acc": 0.65069032, + "epoch": 0.6817605276509386, + "grad_norm": 5.9375, + "learning_rate": 7.859716163794995e-06, + "loss": 1.59818935, + "memory(GiB)": 111.15, + "step": 26875, + "train_speed(iter/s)": 0.448718 + }, + { + "acc": 0.65535021, + "epoch": 0.6818873668188736, + "grad_norm": 3.953125, + "learning_rate": 7.858855922033289e-06, + "loss": 1.60890255, + "memory(GiB)": 111.15, + "step": 26880, + "train_speed(iter/s)": 0.448762 + }, + { + "acc": 0.64086466, + "epoch": 0.6820142059868087, + "grad_norm": 5.625, + "learning_rate": 7.857995554526334e-06, + "loss": 1.69791813, + "memory(GiB)": 111.15, + "step": 26885, + "train_speed(iter/s)": 0.448805 + }, + { + "acc": 0.64264765, + "epoch": 0.6821410451547438, + "grad_norm": 5.21875, + "learning_rate": 7.857135061311977e-06, + "loss": 1.65953407, + "memory(GiB)": 111.15, + "step": 26890, + "train_speed(iter/s)": 0.448849 + }, + { + "acc": 0.64729986, + "epoch": 0.6822678843226788, + "grad_norm": 5.34375, + "learning_rate": 7.856274442428062e-06, + "loss": 1.69646416, + "memory(GiB)": 111.15, + "step": 26895, + "train_speed(iter/s)": 0.448893 + }, + { + "acc": 0.63907433, + "epoch": 0.6823947234906139, + "grad_norm": 5.625, + "learning_rate": 7.855413697912446e-06, + "loss": 1.64591579, + "memory(GiB)": 111.15, + "step": 26900, + "train_speed(iter/s)": 0.448938 + }, + { + "epoch": 0.6823947234906139, + "eval_acc": 0.6423581794870402, + "eval_loss": 1.6130332946777344, + "eval_runtime": 113.5839, + "eval_samples_per_second": 56.082, + "eval_steps_per_second": 28.041, + "step": 26900 + }, + { + "acc": 0.64576931, + "epoch": 0.6825215626585489, + "grad_norm": 5.0625, + "learning_rate": 7.854552827802987e-06, + "loss": 1.63153954, + "memory(GiB)": 111.15, + "step": 26905, + "train_speed(iter/s)": 0.448099 + }, + { + "acc": 0.66078296, + "epoch": 0.682648401826484, + "grad_norm": 7.03125, + "learning_rate": 7.853691832137547e-06, + "loss": 1.62020683, + "memory(GiB)": 111.15, + "step": 26910, + "train_speed(iter/s)": 0.448143 + }, + { + "acc": 0.66204648, + "epoch": 0.6827752409944191, + "grad_norm": 4.71875, + "learning_rate": 7.852830710954003e-06, + "loss": 1.59321594, + "memory(GiB)": 111.15, + "step": 26915, + "train_speed(iter/s)": 0.448186 + }, + { + "acc": 0.6484396, + "epoch": 0.6829020801623541, + "grad_norm": 5.65625, + "learning_rate": 7.851969464290226e-06, + "loss": 1.65079956, + "memory(GiB)": 111.15, + "step": 26920, + "train_speed(iter/s)": 0.448231 + }, + { + "acc": 0.65205269, + "epoch": 0.6830289193302892, + "grad_norm": 6.03125, + "learning_rate": 7.851108092184099e-06, + "loss": 1.61463928, + "memory(GiB)": 111.15, + "step": 26925, + "train_speed(iter/s)": 0.448274 + }, + { + "acc": 0.65546541, + "epoch": 0.6831557584982243, + "grad_norm": 5.09375, + "learning_rate": 7.850246594673508e-06, + "loss": 1.62870216, + "memory(GiB)": 111.15, + "step": 26930, + "train_speed(iter/s)": 0.448319 + }, + { + "acc": 0.64578185, + "epoch": 0.6832825976661593, + "grad_norm": 6.375, + "learning_rate": 7.849384971796346e-06, + "loss": 1.69458237, + "memory(GiB)": 111.15, + "step": 26935, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.6530632, + "epoch": 0.6834094368340944, + "grad_norm": 5.3125, + "learning_rate": 7.848523223590514e-06, + "loss": 1.63932724, + "memory(GiB)": 111.15, + "step": 26940, + "train_speed(iter/s)": 0.448406 + }, + { + "acc": 0.64247379, + "epoch": 0.6835362760020294, + "grad_norm": 6.1875, + "learning_rate": 7.84766135009391e-06, + "loss": 1.66448975, + "memory(GiB)": 111.15, + "step": 26945, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.64360042, + "epoch": 0.6836631151699645, + "grad_norm": 4.625, + "learning_rate": 7.846799351344447e-06, + "loss": 1.66257973, + "memory(GiB)": 111.15, + "step": 26950, + "train_speed(iter/s)": 0.448494 + }, + { + "acc": 0.66403136, + "epoch": 0.6837899543378996, + "grad_norm": 5.3125, + "learning_rate": 7.845937227380038e-06, + "loss": 1.60891495, + "memory(GiB)": 111.15, + "step": 26955, + "train_speed(iter/s)": 0.448538 + }, + { + "acc": 0.63398666, + "epoch": 0.6839167935058346, + "grad_norm": 5.03125, + "learning_rate": 7.845074978238604e-06, + "loss": 1.68343029, + "memory(GiB)": 111.15, + "step": 26960, + "train_speed(iter/s)": 0.448582 + }, + { + "acc": 0.65639954, + "epoch": 0.6840436326737697, + "grad_norm": 5.34375, + "learning_rate": 7.84421260395807e-06, + "loss": 1.65233269, + "memory(GiB)": 111.15, + "step": 26965, + "train_speed(iter/s)": 0.448626 + }, + { + "acc": 0.64712787, + "epoch": 0.6841704718417048, + "grad_norm": 5.40625, + "learning_rate": 7.84335010457637e-06, + "loss": 1.62664108, + "memory(GiB)": 111.15, + "step": 26970, + "train_speed(iter/s)": 0.44867 + }, + { + "acc": 0.65759664, + "epoch": 0.6842973110096398, + "grad_norm": 6.15625, + "learning_rate": 7.842487480131435e-06, + "loss": 1.54200048, + "memory(GiB)": 111.15, + "step": 26975, + "train_speed(iter/s)": 0.448714 + }, + { + "acc": 0.64696722, + "epoch": 0.6844241501775749, + "grad_norm": 5.78125, + "learning_rate": 7.84162473066121e-06, + "loss": 1.7127739, + "memory(GiB)": 111.15, + "step": 26980, + "train_speed(iter/s)": 0.448757 + }, + { + "acc": 0.65742331, + "epoch": 0.6845509893455098, + "grad_norm": 5.09375, + "learning_rate": 7.840761856203642e-06, + "loss": 1.51608372, + "memory(GiB)": 111.15, + "step": 26985, + "train_speed(iter/s)": 0.448801 + }, + { + "acc": 0.639674, + "epoch": 0.6846778285134449, + "grad_norm": 4.8125, + "learning_rate": 7.839898856796685e-06, + "loss": 1.6932888, + "memory(GiB)": 111.15, + "step": 26990, + "train_speed(iter/s)": 0.448845 + }, + { + "acc": 0.63679085, + "epoch": 0.68480466768138, + "grad_norm": 5.53125, + "learning_rate": 7.839035732478297e-06, + "loss": 1.67074261, + "memory(GiB)": 111.15, + "step": 26995, + "train_speed(iter/s)": 0.448889 + }, + { + "acc": 0.64877386, + "epoch": 0.684931506849315, + "grad_norm": 5.34375, + "learning_rate": 7.838172483286441e-06, + "loss": 1.66660309, + "memory(GiB)": 111.15, + "step": 27000, + "train_speed(iter/s)": 0.448933 + }, + { + "epoch": 0.684931506849315, + "eval_acc": 0.6423155672214755, + "eval_loss": 1.6131644248962402, + "eval_runtime": 113.7174, + "eval_samples_per_second": 56.016, + "eval_steps_per_second": 28.008, + "step": 27000 + }, + { + "acc": 0.66580801, + "epoch": 0.6850583460172501, + "grad_norm": 4.78125, + "learning_rate": 7.83730910925909e-06, + "loss": 1.59387274, + "memory(GiB)": 111.15, + "step": 27005, + "train_speed(iter/s)": 0.448095 + }, + { + "acc": 0.64784317, + "epoch": 0.6851851851851852, + "grad_norm": 4.8125, + "learning_rate": 7.836445610434215e-06, + "loss": 1.61885738, + "memory(GiB)": 111.15, + "step": 27010, + "train_speed(iter/s)": 0.448139 + }, + { + "acc": 0.64943123, + "epoch": 0.6853120243531202, + "grad_norm": 5.59375, + "learning_rate": 7.835581986849799e-06, + "loss": 1.65475311, + "memory(GiB)": 111.15, + "step": 27015, + "train_speed(iter/s)": 0.448183 + }, + { + "acc": 0.64305878, + "epoch": 0.6854388635210553, + "grad_norm": 5.25, + "learning_rate": 7.834718238543827e-06, + "loss": 1.64459915, + "memory(GiB)": 111.15, + "step": 27020, + "train_speed(iter/s)": 0.448226 + }, + { + "acc": 0.63771305, + "epoch": 0.6855657026889903, + "grad_norm": 4.78125, + "learning_rate": 7.833854365554289e-06, + "loss": 1.66105194, + "memory(GiB)": 111.15, + "step": 27025, + "train_speed(iter/s)": 0.44827 + }, + { + "acc": 0.64167442, + "epoch": 0.6856925418569254, + "grad_norm": 4.5625, + "learning_rate": 7.832990367919186e-06, + "loss": 1.6878067, + "memory(GiB)": 111.15, + "step": 27030, + "train_speed(iter/s)": 0.448314 + }, + { + "acc": 0.654426, + "epoch": 0.6858193810248605, + "grad_norm": 6.375, + "learning_rate": 7.832126245676518e-06, + "loss": 1.61333237, + "memory(GiB)": 111.15, + "step": 27035, + "train_speed(iter/s)": 0.448358 + }, + { + "acc": 0.64580555, + "epoch": 0.6859462201927955, + "grad_norm": 6.375, + "learning_rate": 7.831261998864293e-06, + "loss": 1.61481285, + "memory(GiB)": 111.15, + "step": 27040, + "train_speed(iter/s)": 0.448402 + }, + { + "acc": 0.6532577, + "epoch": 0.6860730593607306, + "grad_norm": 5.1875, + "learning_rate": 7.830397627520526e-06, + "loss": 1.66595478, + "memory(GiB)": 111.15, + "step": 27045, + "train_speed(iter/s)": 0.448445 + }, + { + "acc": 0.63900938, + "epoch": 0.6861998985286657, + "grad_norm": 5.40625, + "learning_rate": 7.82953313168323e-06, + "loss": 1.65289955, + "memory(GiB)": 111.15, + "step": 27050, + "train_speed(iter/s)": 0.448489 + }, + { + "acc": 0.6438591, + "epoch": 0.6863267376966007, + "grad_norm": 5.65625, + "learning_rate": 7.828668511390439e-06, + "loss": 1.59794827, + "memory(GiB)": 111.15, + "step": 27055, + "train_speed(iter/s)": 0.448532 + }, + { + "acc": 0.64391241, + "epoch": 0.6864535768645358, + "grad_norm": 5.0, + "learning_rate": 7.827803766680176e-06, + "loss": 1.62259674, + "memory(GiB)": 111.15, + "step": 27060, + "train_speed(iter/s)": 0.448576 + }, + { + "acc": 0.64617891, + "epoch": 0.6865804160324708, + "grad_norm": 6.15625, + "learning_rate": 7.826938897590477e-06, + "loss": 1.62988701, + "memory(GiB)": 111.15, + "step": 27065, + "train_speed(iter/s)": 0.448619 + }, + { + "acc": 0.66058416, + "epoch": 0.6867072552004059, + "grad_norm": 5.0625, + "learning_rate": 7.826073904159384e-06, + "loss": 1.56282501, + "memory(GiB)": 111.15, + "step": 27070, + "train_speed(iter/s)": 0.448662 + }, + { + "acc": 0.65731859, + "epoch": 0.686834094368341, + "grad_norm": 4.96875, + "learning_rate": 7.825208786424944e-06, + "loss": 1.63133621, + "memory(GiB)": 111.15, + "step": 27075, + "train_speed(iter/s)": 0.448705 + }, + { + "acc": 0.62951365, + "epoch": 0.686960933536276, + "grad_norm": 6.28125, + "learning_rate": 7.824343544425207e-06, + "loss": 1.71893883, + "memory(GiB)": 111.15, + "step": 27080, + "train_speed(iter/s)": 0.448749 + }, + { + "acc": 0.65072365, + "epoch": 0.6870877727042111, + "grad_norm": 4.53125, + "learning_rate": 7.823478178198234e-06, + "loss": 1.61609783, + "memory(GiB)": 111.15, + "step": 27085, + "train_speed(iter/s)": 0.448792 + }, + { + "acc": 0.65851603, + "epoch": 0.6872146118721462, + "grad_norm": 4.9375, + "learning_rate": 7.822612687782083e-06, + "loss": 1.57316074, + "memory(GiB)": 111.15, + "step": 27090, + "train_speed(iter/s)": 0.448836 + }, + { + "acc": 0.65838842, + "epoch": 0.6873414510400812, + "grad_norm": 5.53125, + "learning_rate": 7.821747073214823e-06, + "loss": 1.55784245, + "memory(GiB)": 111.15, + "step": 27095, + "train_speed(iter/s)": 0.448879 + }, + { + "acc": 0.65382214, + "epoch": 0.6874682902080163, + "grad_norm": 4.75, + "learning_rate": 7.820881334534529e-06, + "loss": 1.57903633, + "memory(GiB)": 111.15, + "step": 27100, + "train_speed(iter/s)": 0.448923 + }, + { + "epoch": 0.6874682902080163, + "eval_acc": 0.6423468997696848, + "eval_loss": 1.6130595207214355, + "eval_runtime": 113.2734, + "eval_samples_per_second": 56.236, + "eval_steps_per_second": 28.118, + "step": 27100 + }, + { + "acc": 0.6600296, + "epoch": 0.6875951293759512, + "grad_norm": 4.96875, + "learning_rate": 7.820015471779278e-06, + "loss": 1.5554739, + "memory(GiB)": 111.15, + "step": 27105, + "train_speed(iter/s)": 0.448092 + }, + { + "acc": 0.66325703, + "epoch": 0.6877219685438863, + "grad_norm": 5.53125, + "learning_rate": 7.819149484987159e-06, + "loss": 1.59647732, + "memory(GiB)": 111.15, + "step": 27110, + "train_speed(iter/s)": 0.448135 + }, + { + "acc": 0.66047325, + "epoch": 0.6878488077118214, + "grad_norm": 4.625, + "learning_rate": 7.818283374196259e-06, + "loss": 1.58700552, + "memory(GiB)": 111.15, + "step": 27115, + "train_speed(iter/s)": 0.448179 + }, + { + "acc": 0.65124817, + "epoch": 0.6879756468797564, + "grad_norm": 5.40625, + "learning_rate": 7.817417139444671e-06, + "loss": 1.66482372, + "memory(GiB)": 111.15, + "step": 27120, + "train_speed(iter/s)": 0.448223 + }, + { + "acc": 0.63304548, + "epoch": 0.6881024860476915, + "grad_norm": 5.84375, + "learning_rate": 7.8165507807705e-06, + "loss": 1.7174902, + "memory(GiB)": 111.15, + "step": 27125, + "train_speed(iter/s)": 0.448266 + }, + { + "acc": 0.63934455, + "epoch": 0.6882293252156266, + "grad_norm": 5.34375, + "learning_rate": 7.81568429821185e-06, + "loss": 1.69177895, + "memory(GiB)": 111.15, + "step": 27130, + "train_speed(iter/s)": 0.44831 + }, + { + "acc": 0.65429296, + "epoch": 0.6883561643835616, + "grad_norm": 5.59375, + "learning_rate": 7.814817691806834e-06, + "loss": 1.67952614, + "memory(GiB)": 111.15, + "step": 27135, + "train_speed(iter/s)": 0.448353 + }, + { + "acc": 0.65168266, + "epoch": 0.6884830035514967, + "grad_norm": 5.3125, + "learning_rate": 7.813950961593569e-06, + "loss": 1.63514843, + "memory(GiB)": 111.15, + "step": 27140, + "train_speed(iter/s)": 0.448396 + }, + { + "acc": 0.63961034, + "epoch": 0.6886098427194317, + "grad_norm": 4.46875, + "learning_rate": 7.813084107610175e-06, + "loss": 1.62346745, + "memory(GiB)": 111.15, + "step": 27145, + "train_speed(iter/s)": 0.44844 + }, + { + "acc": 0.65446038, + "epoch": 0.6887366818873668, + "grad_norm": 5.03125, + "learning_rate": 7.812217129894785e-06, + "loss": 1.52882929, + "memory(GiB)": 111.15, + "step": 27150, + "train_speed(iter/s)": 0.448483 + }, + { + "acc": 0.65641894, + "epoch": 0.6888635210553019, + "grad_norm": 5.15625, + "learning_rate": 7.811350028485531e-06, + "loss": 1.58613987, + "memory(GiB)": 111.15, + "step": 27155, + "train_speed(iter/s)": 0.448526 + }, + { + "acc": 0.65418739, + "epoch": 0.6889903602232369, + "grad_norm": 5.84375, + "learning_rate": 7.810482803420549e-06, + "loss": 1.58340473, + "memory(GiB)": 111.15, + "step": 27160, + "train_speed(iter/s)": 0.44857 + }, + { + "acc": 0.64442277, + "epoch": 0.689117199391172, + "grad_norm": 5.4375, + "learning_rate": 7.809615454737984e-06, + "loss": 1.64776516, + "memory(GiB)": 111.15, + "step": 27165, + "train_speed(iter/s)": 0.448613 + }, + { + "acc": 0.6372869, + "epoch": 0.6892440385591071, + "grad_norm": 4.84375, + "learning_rate": 7.808747982475991e-06, + "loss": 1.66596298, + "memory(GiB)": 111.15, + "step": 27170, + "train_speed(iter/s)": 0.448657 + }, + { + "acc": 0.63553753, + "epoch": 0.6893708777270421, + "grad_norm": 6.71875, + "learning_rate": 7.807880386672718e-06, + "loss": 1.68006096, + "memory(GiB)": 111.15, + "step": 27175, + "train_speed(iter/s)": 0.448701 + }, + { + "acc": 0.63246098, + "epoch": 0.6894977168949772, + "grad_norm": 5.59375, + "learning_rate": 7.807012667366332e-06, + "loss": 1.69895325, + "memory(GiB)": 111.15, + "step": 27180, + "train_speed(iter/s)": 0.448744 + }, + { + "acc": 0.64890976, + "epoch": 0.6896245560629122, + "grad_norm": 6.4375, + "learning_rate": 7.806144824594994e-06, + "loss": 1.64491673, + "memory(GiB)": 111.15, + "step": 27185, + "train_speed(iter/s)": 0.448788 + }, + { + "acc": 0.64255495, + "epoch": 0.6897513952308473, + "grad_norm": 5.46875, + "learning_rate": 7.805276858396879e-06, + "loss": 1.64829903, + "memory(GiB)": 111.15, + "step": 27190, + "train_speed(iter/s)": 0.448832 + }, + { + "acc": 0.65521183, + "epoch": 0.6898782343987824, + "grad_norm": 5.125, + "learning_rate": 7.804408768810164e-06, + "loss": 1.61088448, + "memory(GiB)": 111.15, + "step": 27195, + "train_speed(iter/s)": 0.448875 + }, + { + "acc": 0.6437953, + "epoch": 0.6900050735667174, + "grad_norm": 5.46875, + "learning_rate": 7.80354055587303e-06, + "loss": 1.72474499, + "memory(GiB)": 111.15, + "step": 27200, + "train_speed(iter/s)": 0.448919 + }, + { + "epoch": 0.6900050735667174, + "eval_acc": 0.6424300354642669, + "eval_loss": 1.6128261089324951, + "eval_runtime": 114.06, + "eval_samples_per_second": 55.848, + "eval_steps_per_second": 27.924, + "step": 27200 + }, + { + "acc": 0.64334044, + "epoch": 0.6901319127346525, + "grad_norm": 5.28125, + "learning_rate": 7.802672219623665e-06, + "loss": 1.63186874, + "memory(GiB)": 111.15, + "step": 27205, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.66239381, + "epoch": 0.6902587519025876, + "grad_norm": 5.84375, + "learning_rate": 7.801803760100264e-06, + "loss": 1.6322197, + "memory(GiB)": 111.15, + "step": 27210, + "train_speed(iter/s)": 0.448128 + }, + { + "acc": 0.64245625, + "epoch": 0.6903855910705226, + "grad_norm": 6.71875, + "learning_rate": 7.800935177341022e-06, + "loss": 1.59839554, + "memory(GiB)": 111.15, + "step": 27215, + "train_speed(iter/s)": 0.448172 + }, + { + "acc": 0.66313972, + "epoch": 0.6905124302384577, + "grad_norm": 5.15625, + "learning_rate": 7.800066471384149e-06, + "loss": 1.61726418, + "memory(GiB)": 111.15, + "step": 27220, + "train_speed(iter/s)": 0.448214 + }, + { + "acc": 0.6621943, + "epoch": 0.6906392694063926, + "grad_norm": 5.09375, + "learning_rate": 7.799197642267848e-06, + "loss": 1.62828674, + "memory(GiB)": 111.15, + "step": 27225, + "train_speed(iter/s)": 0.448258 + }, + { + "acc": 0.64925089, + "epoch": 0.6907661085743277, + "grad_norm": 4.78125, + "learning_rate": 7.79832869003034e-06, + "loss": 1.60706367, + "memory(GiB)": 111.15, + "step": 27230, + "train_speed(iter/s)": 0.448301 + }, + { + "acc": 0.65675955, + "epoch": 0.6908929477422628, + "grad_norm": 5.5625, + "learning_rate": 7.797459614709842e-06, + "loss": 1.64934464, + "memory(GiB)": 111.15, + "step": 27235, + "train_speed(iter/s)": 0.448344 + }, + { + "acc": 0.64912891, + "epoch": 0.6910197869101978, + "grad_norm": 5.53125, + "learning_rate": 7.796590416344578e-06, + "loss": 1.61923523, + "memory(GiB)": 111.15, + "step": 27240, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.64007111, + "epoch": 0.6911466260781329, + "grad_norm": 5.40625, + "learning_rate": 7.795721094972783e-06, + "loss": 1.70615559, + "memory(GiB)": 111.15, + "step": 27245, + "train_speed(iter/s)": 0.44843 + }, + { + "acc": 0.63446789, + "epoch": 0.691273465246068, + "grad_norm": 5.6875, + "learning_rate": 7.794851650632693e-06, + "loss": 1.73550949, + "memory(GiB)": 111.15, + "step": 27250, + "train_speed(iter/s)": 0.448474 + }, + { + "acc": 0.65848227, + "epoch": 0.691400304414003, + "grad_norm": 4.6875, + "learning_rate": 7.793982083362548e-06, + "loss": 1.51315727, + "memory(GiB)": 111.15, + "step": 27255, + "train_speed(iter/s)": 0.448517 + }, + { + "acc": 0.66465425, + "epoch": 0.6915271435819381, + "grad_norm": 6.25, + "learning_rate": 7.7931123932006e-06, + "loss": 1.59119492, + "memory(GiB)": 111.15, + "step": 27260, + "train_speed(iter/s)": 0.44856 + }, + { + "acc": 0.64187422, + "epoch": 0.6916539827498731, + "grad_norm": 5.03125, + "learning_rate": 7.792242580185095e-06, + "loss": 1.61752701, + "memory(GiB)": 111.15, + "step": 27265, + "train_speed(iter/s)": 0.448603 + }, + { + "acc": 0.65111341, + "epoch": 0.6917808219178082, + "grad_norm": 7.84375, + "learning_rate": 7.791372644354295e-06, + "loss": 1.58319283, + "memory(GiB)": 111.15, + "step": 27270, + "train_speed(iter/s)": 0.448647 + }, + { + "acc": 0.64268608, + "epoch": 0.6919076610857433, + "grad_norm": 5.46875, + "learning_rate": 7.790502585746464e-06, + "loss": 1.65260468, + "memory(GiB)": 111.15, + "step": 27275, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.65899243, + "epoch": 0.6920345002536783, + "grad_norm": 5.3125, + "learning_rate": 7.789632404399872e-06, + "loss": 1.57038841, + "memory(GiB)": 111.15, + "step": 27280, + "train_speed(iter/s)": 0.448733 + }, + { + "acc": 0.64838166, + "epoch": 0.6921613394216134, + "grad_norm": 6.0625, + "learning_rate": 7.788762100352791e-06, + "loss": 1.62661648, + "memory(GiB)": 111.15, + "step": 27285, + "train_speed(iter/s)": 0.448776 + }, + { + "acc": 0.66353436, + "epoch": 0.6922881785895485, + "grad_norm": 5.09375, + "learning_rate": 7.787891673643501e-06, + "loss": 1.59445438, + "memory(GiB)": 111.15, + "step": 27290, + "train_speed(iter/s)": 0.448818 + }, + { + "acc": 0.64245696, + "epoch": 0.6924150177574835, + "grad_norm": 5.96875, + "learning_rate": 7.78702112431029e-06, + "loss": 1.65724201, + "memory(GiB)": 111.15, + "step": 27295, + "train_speed(iter/s)": 0.448862 + }, + { + "acc": 0.65120373, + "epoch": 0.6925418569254186, + "grad_norm": 5.34375, + "learning_rate": 7.786150452391446e-06, + "loss": 1.59369106, + "memory(GiB)": 111.15, + "step": 27300, + "train_speed(iter/s)": 0.448905 + }, + { + "epoch": 0.6925418569254186, + "eval_acc": 0.6423878409660118, + "eval_loss": 1.612882375717163, + "eval_runtime": 114.6265, + "eval_samples_per_second": 55.572, + "eval_steps_per_second": 27.786, + "step": 27300 + }, + { + "acc": 0.66089926, + "epoch": 0.6926686960933536, + "grad_norm": 5.34375, + "learning_rate": 7.785279657925265e-06, + "loss": 1.57625294, + "memory(GiB)": 111.15, + "step": 27305, + "train_speed(iter/s)": 0.44807 + }, + { + "acc": 0.6380579, + "epoch": 0.6927955352612887, + "grad_norm": 4.875, + "learning_rate": 7.784408740950051e-06, + "loss": 1.65558128, + "memory(GiB)": 111.15, + "step": 27310, + "train_speed(iter/s)": 0.448113 + }, + { + "acc": 0.66649084, + "epoch": 0.6929223744292238, + "grad_norm": 5.71875, + "learning_rate": 7.783537701504109e-06, + "loss": 1.60381908, + "memory(GiB)": 111.15, + "step": 27315, + "train_speed(iter/s)": 0.448156 + }, + { + "acc": 0.65498495, + "epoch": 0.6930492135971588, + "grad_norm": 5.03125, + "learning_rate": 7.782666539625749e-06, + "loss": 1.6545229, + "memory(GiB)": 111.15, + "step": 27320, + "train_speed(iter/s)": 0.448199 + }, + { + "acc": 0.6586226, + "epoch": 0.6931760527650939, + "grad_norm": 5.78125, + "learning_rate": 7.781795255353293e-06, + "loss": 1.57397137, + "memory(GiB)": 111.15, + "step": 27325, + "train_speed(iter/s)": 0.448242 + }, + { + "acc": 0.66413288, + "epoch": 0.693302891933029, + "grad_norm": 5.4375, + "learning_rate": 7.780923848725061e-06, + "loss": 1.58822441, + "memory(GiB)": 111.15, + "step": 27330, + "train_speed(iter/s)": 0.448285 + }, + { + "acc": 0.64076443, + "epoch": 0.693429731100964, + "grad_norm": 4.90625, + "learning_rate": 7.780052319779382e-06, + "loss": 1.61288414, + "memory(GiB)": 111.15, + "step": 27335, + "train_speed(iter/s)": 0.448328 + }, + { + "acc": 0.64450464, + "epoch": 0.693556570268899, + "grad_norm": 4.9375, + "learning_rate": 7.779180668554591e-06, + "loss": 1.69561844, + "memory(GiB)": 111.15, + "step": 27340, + "train_speed(iter/s)": 0.448371 + }, + { + "acc": 0.66133704, + "epoch": 0.693683409436834, + "grad_norm": 5.125, + "learning_rate": 7.778308895089024e-06, + "loss": 1.58780937, + "memory(GiB)": 111.15, + "step": 27345, + "train_speed(iter/s)": 0.448415 + }, + { + "acc": 0.66322408, + "epoch": 0.6938102486047691, + "grad_norm": 5.3125, + "learning_rate": 7.77743699942103e-06, + "loss": 1.59558506, + "memory(GiB)": 111.15, + "step": 27350, + "train_speed(iter/s)": 0.448458 + }, + { + "acc": 0.66435289, + "epoch": 0.6939370877727042, + "grad_norm": 5.71875, + "learning_rate": 7.776564981588955e-06, + "loss": 1.59251385, + "memory(GiB)": 111.15, + "step": 27355, + "train_speed(iter/s)": 0.448502 + }, + { + "acc": 0.64788933, + "epoch": 0.6940639269406392, + "grad_norm": 6.1875, + "learning_rate": 7.775692841631154e-06, + "loss": 1.72122841, + "memory(GiB)": 111.15, + "step": 27360, + "train_speed(iter/s)": 0.448545 + }, + { + "acc": 0.64346743, + "epoch": 0.6941907661085743, + "grad_norm": 4.15625, + "learning_rate": 7.774820579585993e-06, + "loss": 1.64558296, + "memory(GiB)": 111.15, + "step": 27365, + "train_speed(iter/s)": 0.448588 + }, + { + "acc": 0.65252762, + "epoch": 0.6943176052765094, + "grad_norm": 5.0625, + "learning_rate": 7.773948195491831e-06, + "loss": 1.58605566, + "memory(GiB)": 111.15, + "step": 27370, + "train_speed(iter/s)": 0.448632 + }, + { + "acc": 0.64322672, + "epoch": 0.6944444444444444, + "grad_norm": 5.5625, + "learning_rate": 7.773075689387044e-06, + "loss": 1.70017815, + "memory(GiB)": 111.15, + "step": 27375, + "train_speed(iter/s)": 0.448676 + }, + { + "acc": 0.6585835, + "epoch": 0.6945712836123795, + "grad_norm": 6.28125, + "learning_rate": 7.772203061310008e-06, + "loss": 1.60125523, + "memory(GiB)": 111.15, + "step": 27380, + "train_speed(iter/s)": 0.448719 + }, + { + "acc": 0.64209185, + "epoch": 0.6946981227803145, + "grad_norm": 5.15625, + "learning_rate": 7.771330311299104e-06, + "loss": 1.72223206, + "memory(GiB)": 111.15, + "step": 27385, + "train_speed(iter/s)": 0.448763 + }, + { + "acc": 0.64728494, + "epoch": 0.6948249619482496, + "grad_norm": 4.5625, + "learning_rate": 7.770457439392719e-06, + "loss": 1.66091709, + "memory(GiB)": 111.15, + "step": 27390, + "train_speed(iter/s)": 0.448806 + }, + { + "acc": 0.63640733, + "epoch": 0.6949518011161847, + "grad_norm": 4.71875, + "learning_rate": 7.769584445629247e-06, + "loss": 1.73308144, + "memory(GiB)": 111.15, + "step": 27395, + "train_speed(iter/s)": 0.448848 + }, + { + "acc": 0.64614906, + "epoch": 0.6950786402841197, + "grad_norm": 6.375, + "learning_rate": 7.768711330047087e-06, + "loss": 1.61098747, + "memory(GiB)": 111.15, + "step": 27400, + "train_speed(iter/s)": 0.448892 + }, + { + "epoch": 0.6950786402841197, + "eval_acc": 0.6423740546447996, + "eval_loss": 1.6128815412521362, + "eval_runtime": 113.7269, + "eval_samples_per_second": 56.011, + "eval_steps_per_second": 28.006, + "step": 27400 + }, + { + "acc": 0.65466757, + "epoch": 0.6952054794520548, + "grad_norm": 5.53125, + "learning_rate": 7.767838092684638e-06, + "loss": 1.65315247, + "memory(GiB)": 111.15, + "step": 27405, + "train_speed(iter/s)": 0.448067 + }, + { + "acc": 0.66523895, + "epoch": 0.6953323186199899, + "grad_norm": 4.09375, + "learning_rate": 7.766964733580316e-06, + "loss": 1.53610964, + "memory(GiB)": 111.15, + "step": 27410, + "train_speed(iter/s)": 0.448109 + }, + { + "acc": 0.65682898, + "epoch": 0.6954591577879249, + "grad_norm": 4.8125, + "learning_rate": 7.76609125277253e-06, + "loss": 1.61778183, + "memory(GiB)": 111.15, + "step": 27415, + "train_speed(iter/s)": 0.448153 + }, + { + "acc": 0.64641399, + "epoch": 0.69558599695586, + "grad_norm": 7.5, + "learning_rate": 7.7652176502997e-06, + "loss": 1.60407753, + "memory(GiB)": 111.15, + "step": 27420, + "train_speed(iter/s)": 0.448195 + }, + { + "acc": 0.66008792, + "epoch": 0.695712836123795, + "grad_norm": 5.34375, + "learning_rate": 7.764343926200254e-06, + "loss": 1.55034218, + "memory(GiB)": 111.15, + "step": 27425, + "train_speed(iter/s)": 0.448238 + }, + { + "acc": 0.64974041, + "epoch": 0.6958396752917301, + "grad_norm": 5.03125, + "learning_rate": 7.763470080512617e-06, + "loss": 1.6393898, + "memory(GiB)": 111.15, + "step": 27430, + "train_speed(iter/s)": 0.448281 + }, + { + "acc": 0.64724865, + "epoch": 0.6959665144596652, + "grad_norm": 6.46875, + "learning_rate": 7.762596113275229e-06, + "loss": 1.62379246, + "memory(GiB)": 111.15, + "step": 27435, + "train_speed(iter/s)": 0.448324 + }, + { + "acc": 0.64981194, + "epoch": 0.6960933536276002, + "grad_norm": 6.59375, + "learning_rate": 7.761722024526533e-06, + "loss": 1.66894455, + "memory(GiB)": 111.15, + "step": 27440, + "train_speed(iter/s)": 0.448368 + }, + { + "acc": 0.65776367, + "epoch": 0.6962201927955353, + "grad_norm": 5.09375, + "learning_rate": 7.760847814304969e-06, + "loss": 1.62574844, + "memory(GiB)": 111.15, + "step": 27445, + "train_speed(iter/s)": 0.448411 + }, + { + "acc": 0.64897532, + "epoch": 0.6963470319634704, + "grad_norm": 8.5625, + "learning_rate": 7.759973482648992e-06, + "loss": 1.6074688, + "memory(GiB)": 111.15, + "step": 27450, + "train_speed(iter/s)": 0.448454 + }, + { + "acc": 0.64874029, + "epoch": 0.6964738711314054, + "grad_norm": 4.65625, + "learning_rate": 7.75909902959706e-06, + "loss": 1.67528343, + "memory(GiB)": 111.15, + "step": 27455, + "train_speed(iter/s)": 0.448497 + }, + { + "acc": 0.63151455, + "epoch": 0.6966007102993405, + "grad_norm": 5.125, + "learning_rate": 7.758224455187632e-06, + "loss": 1.69375534, + "memory(GiB)": 111.15, + "step": 27460, + "train_speed(iter/s)": 0.44854 + }, + { + "acc": 0.64292421, + "epoch": 0.6967275494672754, + "grad_norm": 5.4375, + "learning_rate": 7.75734975945918e-06, + "loss": 1.70451317, + "memory(GiB)": 111.15, + "step": 27465, + "train_speed(iter/s)": 0.448583 + }, + { + "acc": 0.64648781, + "epoch": 0.6968543886352105, + "grad_norm": 6.1875, + "learning_rate": 7.756474942450174e-06, + "loss": 1.65056171, + "memory(GiB)": 111.15, + "step": 27470, + "train_speed(iter/s)": 0.448626 + }, + { + "acc": 0.6621109, + "epoch": 0.6969812278031456, + "grad_norm": 6.21875, + "learning_rate": 7.755600004199094e-06, + "loss": 1.57999783, + "memory(GiB)": 111.15, + "step": 27475, + "train_speed(iter/s)": 0.448669 + }, + { + "acc": 0.65615435, + "epoch": 0.6971080669710806, + "grad_norm": 5.84375, + "learning_rate": 7.754724944744423e-06, + "loss": 1.63501034, + "memory(GiB)": 111.15, + "step": 27480, + "train_speed(iter/s)": 0.448712 + }, + { + "acc": 0.65436749, + "epoch": 0.6972349061390157, + "grad_norm": 5.6875, + "learning_rate": 7.753849764124648e-06, + "loss": 1.6761858, + "memory(GiB)": 111.15, + "step": 27485, + "train_speed(iter/s)": 0.448755 + }, + { + "acc": 0.64499607, + "epoch": 0.6973617453069508, + "grad_norm": 5.4375, + "learning_rate": 7.752974462378268e-06, + "loss": 1.68089008, + "memory(GiB)": 111.15, + "step": 27490, + "train_speed(iter/s)": 0.448798 + }, + { + "acc": 0.64656434, + "epoch": 0.6974885844748858, + "grad_norm": 4.71875, + "learning_rate": 7.752099039543778e-06, + "loss": 1.64931545, + "memory(GiB)": 111.15, + "step": 27495, + "train_speed(iter/s)": 0.448841 + }, + { + "acc": 0.64883447, + "epoch": 0.6976154236428209, + "grad_norm": 5.21875, + "learning_rate": 7.751223495659685e-06, + "loss": 1.59872494, + "memory(GiB)": 111.15, + "step": 27500, + "train_speed(iter/s)": 0.448885 + }, + { + "epoch": 0.6976154236428209, + "eval_acc": 0.6424300354642669, + "eval_loss": 1.613012671470642, + "eval_runtime": 113.4648, + "eval_samples_per_second": 56.141, + "eval_steps_per_second": 28.07, + "step": 27500 + }, + { + "acc": 0.65321102, + "epoch": 0.6977422628107559, + "grad_norm": 4.8125, + "learning_rate": 7.7503478307645e-06, + "loss": 1.62668114, + "memory(GiB)": 111.15, + "step": 27505, + "train_speed(iter/s)": 0.448064 + }, + { + "acc": 0.6624752, + "epoch": 0.697869101978691, + "grad_norm": 4.96875, + "learning_rate": 7.74947204489674e-06, + "loss": 1.58785324, + "memory(GiB)": 111.15, + "step": 27510, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.64432735, + "epoch": 0.6979959411466261, + "grad_norm": 5.625, + "learning_rate": 7.748596138094922e-06, + "loss": 1.69139557, + "memory(GiB)": 111.15, + "step": 27515, + "train_speed(iter/s)": 0.44815 + }, + { + "acc": 0.62034159, + "epoch": 0.6981227803145611, + "grad_norm": 5.8125, + "learning_rate": 7.747720110397573e-06, + "loss": 1.68548203, + "memory(GiB)": 111.15, + "step": 27520, + "train_speed(iter/s)": 0.448193 + }, + { + "acc": 0.65265794, + "epoch": 0.6982496194824962, + "grad_norm": 5.25, + "learning_rate": 7.746843961843226e-06, + "loss": 1.68045044, + "memory(GiB)": 111.15, + "step": 27525, + "train_speed(iter/s)": 0.448236 + }, + { + "acc": 0.66274924, + "epoch": 0.6983764586504313, + "grad_norm": 5.1875, + "learning_rate": 7.74596769247042e-06, + "loss": 1.58748817, + "memory(GiB)": 111.15, + "step": 27530, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.65341997, + "epoch": 0.6985032978183663, + "grad_norm": 5.59375, + "learning_rate": 7.745091302317694e-06, + "loss": 1.61613235, + "memory(GiB)": 111.15, + "step": 27535, + "train_speed(iter/s)": 0.448323 + }, + { + "acc": 0.65679498, + "epoch": 0.6986301369863014, + "grad_norm": 5.65625, + "learning_rate": 7.744214791423597e-06, + "loss": 1.65667725, + "memory(GiB)": 111.15, + "step": 27540, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.63040023, + "epoch": 0.6987569761542364, + "grad_norm": 5.34375, + "learning_rate": 7.74333815982668e-06, + "loss": 1.670965, + "memory(GiB)": 111.15, + "step": 27545, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.64400167, + "epoch": 0.6988838153221715, + "grad_norm": 4.875, + "learning_rate": 7.742461407565504e-06, + "loss": 1.61446953, + "memory(GiB)": 111.15, + "step": 27550, + "train_speed(iter/s)": 0.448454 + }, + { + "acc": 0.65236917, + "epoch": 0.6990106544901066, + "grad_norm": 4.84375, + "learning_rate": 7.741584534678632e-06, + "loss": 1.64871922, + "memory(GiB)": 111.15, + "step": 27555, + "train_speed(iter/s)": 0.448497 + }, + { + "acc": 0.65684319, + "epoch": 0.6991374936580416, + "grad_norm": 4.78125, + "learning_rate": 7.74070754120463e-06, + "loss": 1.6314106, + "memory(GiB)": 111.15, + "step": 27560, + "train_speed(iter/s)": 0.44854 + }, + { + "acc": 0.65072422, + "epoch": 0.6992643328259767, + "grad_norm": 5.3125, + "learning_rate": 7.739830427182073e-06, + "loss": 1.66485519, + "memory(GiB)": 111.15, + "step": 27565, + "train_speed(iter/s)": 0.448584 + }, + { + "acc": 0.65733271, + "epoch": 0.6993911719939118, + "grad_norm": 5.71875, + "learning_rate": 7.738953192649544e-06, + "loss": 1.63601112, + "memory(GiB)": 111.15, + "step": 27570, + "train_speed(iter/s)": 0.448627 + }, + { + "acc": 0.64282942, + "epoch": 0.6995180111618468, + "grad_norm": 5.1875, + "learning_rate": 7.738075837645625e-06, + "loss": 1.63813763, + "memory(GiB)": 111.15, + "step": 27575, + "train_speed(iter/s)": 0.448671 + }, + { + "acc": 0.65217743, + "epoch": 0.6996448503297819, + "grad_norm": 5.90625, + "learning_rate": 7.737198362208904e-06, + "loss": 1.65303917, + "memory(GiB)": 111.15, + "step": 27580, + "train_speed(iter/s)": 0.448714 + }, + { + "acc": 0.65220785, + "epoch": 0.6997716894977168, + "grad_norm": 5.71875, + "learning_rate": 7.736320766377978e-06, + "loss": 1.689048, + "memory(GiB)": 111.15, + "step": 27585, + "train_speed(iter/s)": 0.448757 + }, + { + "acc": 0.63506594, + "epoch": 0.6998985286656519, + "grad_norm": 5.09375, + "learning_rate": 7.735443050191452e-06, + "loss": 1.59329052, + "memory(GiB)": 111.15, + "step": 27590, + "train_speed(iter/s)": 0.4488 + }, + { + "acc": 0.64802408, + "epoch": 0.700025367833587, + "grad_norm": 4.75, + "learning_rate": 7.734565213687923e-06, + "loss": 1.58202076, + "memory(GiB)": 111.15, + "step": 27595, + "train_speed(iter/s)": 0.448843 + }, + { + "acc": 0.64828405, + "epoch": 0.700152207001522, + "grad_norm": 4.71875, + "learning_rate": 7.733687256906009e-06, + "loss": 1.65334358, + "memory(GiB)": 111.15, + "step": 27600, + "train_speed(iter/s)": 0.448887 + }, + { + "epoch": 0.700152207001522, + "eval_acc": 0.6423769790159658, + "eval_loss": 1.612640380859375, + "eval_runtime": 112.9447, + "eval_samples_per_second": 56.399, + "eval_steps_per_second": 28.2, + "step": 27600 + }, + { + "acc": 0.65701203, + "epoch": 0.7002790461694571, + "grad_norm": 4.65625, + "learning_rate": 7.732809179884324e-06, + "loss": 1.61926041, + "memory(GiB)": 111.15, + "step": 27605, + "train_speed(iter/s)": 0.448073 + }, + { + "acc": 0.65285063, + "epoch": 0.7004058853373922, + "grad_norm": 5.71875, + "learning_rate": 7.73193098266149e-06, + "loss": 1.60494995, + "memory(GiB)": 111.15, + "step": 27610, + "train_speed(iter/s)": 0.448117 + }, + { + "acc": 0.63638062, + "epoch": 0.7005327245053272, + "grad_norm": 6.40625, + "learning_rate": 7.731052665276135e-06, + "loss": 1.68978424, + "memory(GiB)": 111.15, + "step": 27615, + "train_speed(iter/s)": 0.44816 + }, + { + "acc": 0.6664186, + "epoch": 0.7006595636732623, + "grad_norm": 5.96875, + "learning_rate": 7.730174227766892e-06, + "loss": 1.61118469, + "memory(GiB)": 111.15, + "step": 27620, + "train_speed(iter/s)": 0.448203 + }, + { + "acc": 0.65710602, + "epoch": 0.7007864028411973, + "grad_norm": 6.09375, + "learning_rate": 7.729295670172394e-06, + "loss": 1.61265182, + "memory(GiB)": 111.15, + "step": 27625, + "train_speed(iter/s)": 0.448246 + }, + { + "acc": 0.65566854, + "epoch": 0.7009132420091324, + "grad_norm": 5.875, + "learning_rate": 7.728416992531287e-06, + "loss": 1.62226486, + "memory(GiB)": 111.15, + "step": 27630, + "train_speed(iter/s)": 0.448289 + }, + { + "acc": 0.64823942, + "epoch": 0.7010400811770675, + "grad_norm": 6.03125, + "learning_rate": 7.72753819488222e-06, + "loss": 1.65574074, + "memory(GiB)": 111.15, + "step": 27635, + "train_speed(iter/s)": 0.448332 + }, + { + "acc": 0.65435743, + "epoch": 0.7011669203450025, + "grad_norm": 4.875, + "learning_rate": 7.726659277263848e-06, + "loss": 1.59657249, + "memory(GiB)": 111.15, + "step": 27640, + "train_speed(iter/s)": 0.448375 + }, + { + "acc": 0.64979696, + "epoch": 0.7012937595129376, + "grad_norm": 8.625, + "learning_rate": 7.725780239714824e-06, + "loss": 1.66448498, + "memory(GiB)": 111.15, + "step": 27645, + "train_speed(iter/s)": 0.448418 + }, + { + "acc": 0.64664135, + "epoch": 0.7014205986808727, + "grad_norm": 5.34375, + "learning_rate": 7.724901082273817e-06, + "loss": 1.60405827, + "memory(GiB)": 111.15, + "step": 27650, + "train_speed(iter/s)": 0.44846 + }, + { + "acc": 0.65254378, + "epoch": 0.7015474378488077, + "grad_norm": 5.875, + "learning_rate": 7.724021804979493e-06, + "loss": 1.63061714, + "memory(GiB)": 111.15, + "step": 27655, + "train_speed(iter/s)": 0.448503 + }, + { + "acc": 0.65203714, + "epoch": 0.7016742770167428, + "grad_norm": 4.9375, + "learning_rate": 7.723142407870532e-06, + "loss": 1.63688908, + "memory(GiB)": 111.15, + "step": 27660, + "train_speed(iter/s)": 0.448546 + }, + { + "acc": 0.66981001, + "epoch": 0.7018011161846778, + "grad_norm": 5.15625, + "learning_rate": 7.722262890985605e-06, + "loss": 1.65540314, + "memory(GiB)": 111.15, + "step": 27665, + "train_speed(iter/s)": 0.448589 + }, + { + "acc": 0.64271383, + "epoch": 0.7019279553526129, + "grad_norm": 5.3125, + "learning_rate": 7.721383254363407e-06, + "loss": 1.65850601, + "memory(GiB)": 111.15, + "step": 27670, + "train_speed(iter/s)": 0.448632 + }, + { + "acc": 0.6683012, + "epoch": 0.702054794520548, + "grad_norm": 5.03125, + "learning_rate": 7.72050349804262e-06, + "loss": 1.54100981, + "memory(GiB)": 111.15, + "step": 27675, + "train_speed(iter/s)": 0.448674 + }, + { + "acc": 0.65183206, + "epoch": 0.702181633688483, + "grad_norm": 4.90625, + "learning_rate": 7.719623622061943e-06, + "loss": 1.59268093, + "memory(GiB)": 111.15, + "step": 27680, + "train_speed(iter/s)": 0.448717 + }, + { + "acc": 0.63978047, + "epoch": 0.7023084728564181, + "grad_norm": 5.34375, + "learning_rate": 7.718743626460076e-06, + "loss": 1.68547707, + "memory(GiB)": 111.15, + "step": 27685, + "train_speed(iter/s)": 0.44876 + }, + { + "acc": 0.66569538, + "epoch": 0.7024353120243532, + "grad_norm": 5.0, + "learning_rate": 7.71786351127573e-06, + "loss": 1.56865406, + "memory(GiB)": 111.15, + "step": 27690, + "train_speed(iter/s)": 0.448802 + }, + { + "acc": 0.64319062, + "epoch": 0.7025621511922882, + "grad_norm": 5.5, + "learning_rate": 7.71698327654761e-06, + "loss": 1.62170734, + "memory(GiB)": 111.15, + "step": 27695, + "train_speed(iter/s)": 0.448845 + }, + { + "acc": 0.64652147, + "epoch": 0.7026889903602233, + "grad_norm": 5.65625, + "learning_rate": 7.716102922314435e-06, + "loss": 1.62374401, + "memory(GiB)": 111.15, + "step": 27700, + "train_speed(iter/s)": 0.448887 + }, + { + "epoch": 0.7026889903602233, + "eval_acc": 0.6424655456855708, + "eval_loss": 1.6128194332122803, + "eval_runtime": 113.0536, + "eval_samples_per_second": 56.345, + "eval_steps_per_second": 28.172, + "step": 27700 + }, + { + "acc": 0.62849774, + "epoch": 0.7028158295281582, + "grad_norm": 5.4375, + "learning_rate": 7.715222448614926e-06, + "loss": 1.71739712, + "memory(GiB)": 111.15, + "step": 27705, + "train_speed(iter/s)": 0.448076 + }, + { + "acc": 0.65747905, + "epoch": 0.7029426686960933, + "grad_norm": 6.4375, + "learning_rate": 7.714341855487812e-06, + "loss": 1.63962307, + "memory(GiB)": 111.15, + "step": 27710, + "train_speed(iter/s)": 0.448119 + }, + { + "acc": 0.64330645, + "epoch": 0.7030695078640284, + "grad_norm": 4.875, + "learning_rate": 7.713461142971824e-06, + "loss": 1.62593231, + "memory(GiB)": 111.15, + "step": 27715, + "train_speed(iter/s)": 0.448161 + }, + { + "acc": 0.6624094, + "epoch": 0.7031963470319634, + "grad_norm": 5.25, + "learning_rate": 7.712580311105701e-06, + "loss": 1.60742302, + "memory(GiB)": 111.15, + "step": 27720, + "train_speed(iter/s)": 0.448204 + }, + { + "acc": 0.65381608, + "epoch": 0.7033231861998985, + "grad_norm": 4.96875, + "learning_rate": 7.711699359928184e-06, + "loss": 1.54453707, + "memory(GiB)": 111.15, + "step": 27725, + "train_speed(iter/s)": 0.448247 + }, + { + "acc": 0.65976238, + "epoch": 0.7034500253678336, + "grad_norm": 6.15625, + "learning_rate": 7.710818289478024e-06, + "loss": 1.60643024, + "memory(GiB)": 111.15, + "step": 27730, + "train_speed(iter/s)": 0.44829 + }, + { + "acc": 0.64964495, + "epoch": 0.7035768645357686, + "grad_norm": 5.15625, + "learning_rate": 7.709937099793971e-06, + "loss": 1.63145523, + "memory(GiB)": 111.15, + "step": 27735, + "train_speed(iter/s)": 0.448333 + }, + { + "acc": 0.6379662, + "epoch": 0.7037037037037037, + "grad_norm": 6.0, + "learning_rate": 7.709055790914787e-06, + "loss": 1.64742413, + "memory(GiB)": 111.15, + "step": 27740, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.65004673, + "epoch": 0.7038305428716387, + "grad_norm": 5.59375, + "learning_rate": 7.708174362879234e-06, + "loss": 1.63490238, + "memory(GiB)": 111.15, + "step": 27745, + "train_speed(iter/s)": 0.448419 + }, + { + "acc": 0.64052091, + "epoch": 0.7039573820395738, + "grad_norm": 5.21875, + "learning_rate": 7.70729281572608e-06, + "loss": 1.68704548, + "memory(GiB)": 111.15, + "step": 27750, + "train_speed(iter/s)": 0.448461 + }, + { + "acc": 0.63486443, + "epoch": 0.7040842212075089, + "grad_norm": 5.09375, + "learning_rate": 7.706411149494102e-06, + "loss": 1.69436321, + "memory(GiB)": 111.15, + "step": 27755, + "train_speed(iter/s)": 0.448504 + }, + { + "acc": 0.65157528, + "epoch": 0.7042110603754439, + "grad_norm": 5.84375, + "learning_rate": 7.705529364222079e-06, + "loss": 1.67440033, + "memory(GiB)": 111.15, + "step": 27760, + "train_speed(iter/s)": 0.448548 + }, + { + "acc": 0.65212708, + "epoch": 0.704337899543379, + "grad_norm": 5.8125, + "learning_rate": 7.704647459948793e-06, + "loss": 1.59622679, + "memory(GiB)": 111.15, + "step": 27765, + "train_speed(iter/s)": 0.448591 + }, + { + "acc": 0.6448103, + "epoch": 0.7044647387113141, + "grad_norm": 5.15625, + "learning_rate": 7.703765436713038e-06, + "loss": 1.62515888, + "memory(GiB)": 111.15, + "step": 27770, + "train_speed(iter/s)": 0.448634 + }, + { + "acc": 0.66771221, + "epoch": 0.7045915778792491, + "grad_norm": 5.71875, + "learning_rate": 7.702883294553607e-06, + "loss": 1.57454319, + "memory(GiB)": 111.15, + "step": 27775, + "train_speed(iter/s)": 0.448676 + }, + { + "acc": 0.65548706, + "epoch": 0.7047184170471842, + "grad_norm": 4.9375, + "learning_rate": 7.702001033509302e-06, + "loss": 1.55681686, + "memory(GiB)": 111.15, + "step": 27780, + "train_speed(iter/s)": 0.448719 + }, + { + "acc": 0.63695726, + "epoch": 0.7048452562151192, + "grad_norm": 4.46875, + "learning_rate": 7.701118653618927e-06, + "loss": 1.65198708, + "memory(GiB)": 111.15, + "step": 27785, + "train_speed(iter/s)": 0.448762 + }, + { + "acc": 0.63980789, + "epoch": 0.7049720953830543, + "grad_norm": 5.5, + "learning_rate": 7.700236154921294e-06, + "loss": 1.62068043, + "memory(GiB)": 111.15, + "step": 27790, + "train_speed(iter/s)": 0.448805 + }, + { + "acc": 0.64066372, + "epoch": 0.7050989345509894, + "grad_norm": 4.78125, + "learning_rate": 7.699353537455222e-06, + "loss": 1.66257553, + "memory(GiB)": 111.15, + "step": 27795, + "train_speed(iter/s)": 0.448847 + }, + { + "acc": 0.64247928, + "epoch": 0.7052257737189244, + "grad_norm": 4.9375, + "learning_rate": 7.698470801259526e-06, + "loss": 1.69301281, + "memory(GiB)": 111.15, + "step": 27800, + "train_speed(iter/s)": 0.44889 + }, + { + "epoch": 0.7052257737189244, + "eval_acc": 0.642561632166746, + "eval_loss": 1.6123883724212646, + "eval_runtime": 113.381, + "eval_samples_per_second": 56.182, + "eval_steps_per_second": 28.091, + "step": 27800 + }, + { + "acc": 0.65677185, + "epoch": 0.7053526128868595, + "grad_norm": 5.3125, + "learning_rate": 7.697587946373037e-06, + "loss": 1.64699707, + "memory(GiB)": 111.15, + "step": 27805, + "train_speed(iter/s)": 0.44808 + }, + { + "acc": 0.64685812, + "epoch": 0.7054794520547946, + "grad_norm": 4.75, + "learning_rate": 7.696704972834589e-06, + "loss": 1.62361298, + "memory(GiB)": 111.15, + "step": 27810, + "train_speed(iter/s)": 0.448122 + }, + { + "acc": 0.65574903, + "epoch": 0.7056062912227296, + "grad_norm": 5.21875, + "learning_rate": 7.695821880683012e-06, + "loss": 1.63915443, + "memory(GiB)": 111.15, + "step": 27815, + "train_speed(iter/s)": 0.448164 + }, + { + "acc": 0.64726295, + "epoch": 0.7057331303906647, + "grad_norm": 5.96875, + "learning_rate": 7.694938669957156e-06, + "loss": 1.60624313, + "memory(GiB)": 111.15, + "step": 27820, + "train_speed(iter/s)": 0.448207 + }, + { + "acc": 0.65158129, + "epoch": 0.7058599695585996, + "grad_norm": 5.65625, + "learning_rate": 7.694055340695862e-06, + "loss": 1.60712566, + "memory(GiB)": 111.15, + "step": 27825, + "train_speed(iter/s)": 0.448249 + }, + { + "acc": 0.65932217, + "epoch": 0.7059868087265347, + "grad_norm": 6.46875, + "learning_rate": 7.693171892937991e-06, + "loss": 1.60267754, + "memory(GiB)": 111.15, + "step": 27830, + "train_speed(iter/s)": 0.448292 + }, + { + "acc": 0.64953151, + "epoch": 0.7061136478944698, + "grad_norm": 4.96875, + "learning_rate": 7.692288326722393e-06, + "loss": 1.66166725, + "memory(GiB)": 111.15, + "step": 27835, + "train_speed(iter/s)": 0.448334 + }, + { + "acc": 0.6479702, + "epoch": 0.7062404870624048, + "grad_norm": 5.6875, + "learning_rate": 7.691404642087933e-06, + "loss": 1.59075546, + "memory(GiB)": 111.15, + "step": 27840, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.66386023, + "epoch": 0.7063673262303399, + "grad_norm": 5.96875, + "learning_rate": 7.690520839073484e-06, + "loss": 1.60875931, + "memory(GiB)": 111.15, + "step": 27845, + "train_speed(iter/s)": 0.448418 + }, + { + "acc": 0.64657288, + "epoch": 0.706494165398275, + "grad_norm": 5.875, + "learning_rate": 7.689636917717913e-06, + "loss": 1.64208755, + "memory(GiB)": 111.15, + "step": 27850, + "train_speed(iter/s)": 0.448461 + }, + { + "acc": 0.65228872, + "epoch": 0.70662100456621, + "grad_norm": 5.125, + "learning_rate": 7.688752878060103e-06, + "loss": 1.64844742, + "memory(GiB)": 111.15, + "step": 27855, + "train_speed(iter/s)": 0.448503 + }, + { + "acc": 0.64434052, + "epoch": 0.7067478437341451, + "grad_norm": 4.78125, + "learning_rate": 7.687868720138939e-06, + "loss": 1.60672302, + "memory(GiB)": 111.15, + "step": 27860, + "train_speed(iter/s)": 0.448546 + }, + { + "acc": 0.65063162, + "epoch": 0.7068746829020801, + "grad_norm": 5.84375, + "learning_rate": 7.686984443993304e-06, + "loss": 1.6598362, + "memory(GiB)": 111.15, + "step": 27865, + "train_speed(iter/s)": 0.448589 + }, + { + "acc": 0.64606781, + "epoch": 0.7070015220700152, + "grad_norm": 5.28125, + "learning_rate": 7.6861000496621e-06, + "loss": 1.65954094, + "memory(GiB)": 111.15, + "step": 27870, + "train_speed(iter/s)": 0.448631 + }, + { + "acc": 0.65370607, + "epoch": 0.7071283612379503, + "grad_norm": 5.5625, + "learning_rate": 7.685215537184223e-06, + "loss": 1.68233528, + "memory(GiB)": 111.15, + "step": 27875, + "train_speed(iter/s)": 0.448674 + }, + { + "acc": 0.6516818, + "epoch": 0.7072552004058853, + "grad_norm": 5.71875, + "learning_rate": 7.684330906598577e-06, + "loss": 1.69651756, + "memory(GiB)": 111.15, + "step": 27880, + "train_speed(iter/s)": 0.448717 + }, + { + "acc": 0.66169562, + "epoch": 0.7073820395738204, + "grad_norm": 5.1875, + "learning_rate": 7.683446157944075e-06, + "loss": 1.60417862, + "memory(GiB)": 111.15, + "step": 27885, + "train_speed(iter/s)": 0.448759 + }, + { + "acc": 0.64415512, + "epoch": 0.7075088787417555, + "grad_norm": 5.4375, + "learning_rate": 7.682561291259628e-06, + "loss": 1.68880672, + "memory(GiB)": 111.15, + "step": 27890, + "train_speed(iter/s)": 0.448801 + }, + { + "acc": 0.64884696, + "epoch": 0.7076357179096905, + "grad_norm": 5.21875, + "learning_rate": 7.681676306584159e-06, + "loss": 1.60524559, + "memory(GiB)": 111.15, + "step": 27895, + "train_speed(iter/s)": 0.448843 + }, + { + "acc": 0.64571161, + "epoch": 0.7077625570776256, + "grad_norm": 5.78125, + "learning_rate": 7.680791203956594e-06, + "loss": 1.59074707, + "memory(GiB)": 111.15, + "step": 27900, + "train_speed(iter/s)": 0.448886 + }, + { + "epoch": 0.7077625570776256, + "eval_acc": 0.642483091912568, + "eval_loss": 1.6126298904418945, + "eval_runtime": 114.4726, + "eval_samples_per_second": 55.646, + "eval_steps_per_second": 27.823, + "step": 27900 + }, + { + "acc": 0.64305725, + "epoch": 0.7078893962455606, + "grad_norm": 5.53125, + "learning_rate": 7.679905983415861e-06, + "loss": 1.67776833, + "memory(GiB)": 111.15, + "step": 27905, + "train_speed(iter/s)": 0.44807 + }, + { + "acc": 0.65947766, + "epoch": 0.7080162354134957, + "grad_norm": 4.96875, + "learning_rate": 7.6790206450009e-06, + "loss": 1.64223366, + "memory(GiB)": 111.15, + "step": 27910, + "train_speed(iter/s)": 0.448112 + }, + { + "acc": 0.63835707, + "epoch": 0.7081430745814308, + "grad_norm": 5.125, + "learning_rate": 7.678135188750648e-06, + "loss": 1.6498661, + "memory(GiB)": 111.15, + "step": 27915, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.64710779, + "epoch": 0.7082699137493658, + "grad_norm": 5.21875, + "learning_rate": 7.677249614704057e-06, + "loss": 1.61492062, + "memory(GiB)": 111.15, + "step": 27920, + "train_speed(iter/s)": 0.448196 + }, + { + "acc": 0.66124034, + "epoch": 0.7083967529173009, + "grad_norm": 4.9375, + "learning_rate": 7.676363922900073e-06, + "loss": 1.58523006, + "memory(GiB)": 111.15, + "step": 27925, + "train_speed(iter/s)": 0.448238 + }, + { + "acc": 0.63368821, + "epoch": 0.708523592085236, + "grad_norm": 5.71875, + "learning_rate": 7.675478113377653e-06, + "loss": 1.70831585, + "memory(GiB)": 111.15, + "step": 27930, + "train_speed(iter/s)": 0.44828 + }, + { + "acc": 0.6531642, + "epoch": 0.708650431253171, + "grad_norm": 5.96875, + "learning_rate": 7.674592186175762e-06, + "loss": 1.57435207, + "memory(GiB)": 111.15, + "step": 27935, + "train_speed(iter/s)": 0.448322 + }, + { + "acc": 0.65839481, + "epoch": 0.708777270421106, + "grad_norm": 4.8125, + "learning_rate": 7.673706141333365e-06, + "loss": 1.54583817, + "memory(GiB)": 111.15, + "step": 27940, + "train_speed(iter/s)": 0.448364 + }, + { + "acc": 0.64479866, + "epoch": 0.708904109589041, + "grad_norm": 5.0625, + "learning_rate": 7.672819978889435e-06, + "loss": 1.65067425, + "memory(GiB)": 111.15, + "step": 27945, + "train_speed(iter/s)": 0.448407 + }, + { + "acc": 0.6499074, + "epoch": 0.7090309487569761, + "grad_norm": 5.9375, + "learning_rate": 7.67193369888295e-06, + "loss": 1.67257614, + "memory(GiB)": 111.15, + "step": 27950, + "train_speed(iter/s)": 0.448449 + }, + { + "acc": 0.66732044, + "epoch": 0.7091577879249112, + "grad_norm": 5.21875, + "learning_rate": 7.67104730135289e-06, + "loss": 1.56769314, + "memory(GiB)": 111.15, + "step": 27955, + "train_speed(iter/s)": 0.448491 + }, + { + "acc": 0.6578423, + "epoch": 0.7092846270928462, + "grad_norm": 5.46875, + "learning_rate": 7.670160786338246e-06, + "loss": 1.65622368, + "memory(GiB)": 111.15, + "step": 27960, + "train_speed(iter/s)": 0.448533 + }, + { + "acc": 0.6560832, + "epoch": 0.7094114662607813, + "grad_norm": 5.375, + "learning_rate": 7.669274153878006e-06, + "loss": 1.60550156, + "memory(GiB)": 111.15, + "step": 27965, + "train_speed(iter/s)": 0.448575 + }, + { + "acc": 0.64362154, + "epoch": 0.7095383054287164, + "grad_norm": 4.9375, + "learning_rate": 7.668387404011176e-06, + "loss": 1.67298965, + "memory(GiB)": 111.15, + "step": 27970, + "train_speed(iter/s)": 0.448617 + }, + { + "acc": 0.64360952, + "epoch": 0.7096651445966514, + "grad_norm": 5.5625, + "learning_rate": 7.667500536776748e-06, + "loss": 1.69110069, + "memory(GiB)": 111.15, + "step": 27975, + "train_speed(iter/s)": 0.448659 + }, + { + "acc": 0.64224939, + "epoch": 0.7097919837645865, + "grad_norm": 4.6875, + "learning_rate": 7.666613552213742e-06, + "loss": 1.63325043, + "memory(GiB)": 111.15, + "step": 27980, + "train_speed(iter/s)": 0.448701 + }, + { + "acc": 0.65375295, + "epoch": 0.7099188229325215, + "grad_norm": 5.625, + "learning_rate": 7.665726450361165e-06, + "loss": 1.59965725, + "memory(GiB)": 111.15, + "step": 27985, + "train_speed(iter/s)": 0.448743 + }, + { + "acc": 0.65635514, + "epoch": 0.7100456621004566, + "grad_norm": 4.5, + "learning_rate": 7.664839231258036e-06, + "loss": 1.62208977, + "memory(GiB)": 111.15, + "step": 27990, + "train_speed(iter/s)": 0.448784 + }, + { + "acc": 0.65990105, + "epoch": 0.7101725012683917, + "grad_norm": 5.75, + "learning_rate": 7.663951894943383e-06, + "loss": 1.64975014, + "memory(GiB)": 111.15, + "step": 27995, + "train_speed(iter/s)": 0.448827 + }, + { + "acc": 0.64638233, + "epoch": 0.7102993404363267, + "grad_norm": 5.84375, + "learning_rate": 7.66306444145623e-06, + "loss": 1.64090233, + "memory(GiB)": 111.15, + "step": 28000, + "train_speed(iter/s)": 0.448868 + }, + { + "epoch": 0.7102993404363267, + "eval_acc": 0.6424709766605937, + "eval_loss": 1.6125950813293457, + "eval_runtime": 114.4806, + "eval_samples_per_second": 55.643, + "eval_steps_per_second": 27.821, + "step": 28000 + }, + { + "acc": 0.63968534, + "epoch": 0.7104261796042618, + "grad_norm": 5.84375, + "learning_rate": 7.662176870835614e-06, + "loss": 1.66435966, + "memory(GiB)": 111.15, + "step": 28005, + "train_speed(iter/s)": 0.448055 + }, + { + "acc": 0.66160889, + "epoch": 0.7105530187721969, + "grad_norm": 5.0, + "learning_rate": 7.661289183120572e-06, + "loss": 1.59238281, + "memory(GiB)": 111.15, + "step": 28010, + "train_speed(iter/s)": 0.448097 + }, + { + "acc": 0.65534248, + "epoch": 0.7106798579401319, + "grad_norm": 5.71875, + "learning_rate": 7.66040137835015e-06, + "loss": 1.65107002, + "memory(GiB)": 111.15, + "step": 28015, + "train_speed(iter/s)": 0.448138 + }, + { + "acc": 0.62199397, + "epoch": 0.710806697108067, + "grad_norm": 4.84375, + "learning_rate": 7.659513456563399e-06, + "loss": 1.71907349, + "memory(GiB)": 111.15, + "step": 28020, + "train_speed(iter/s)": 0.44818 + }, + { + "acc": 0.66341181, + "epoch": 0.710933536276002, + "grad_norm": 5.25, + "learning_rate": 7.658625417799372e-06, + "loss": 1.61953506, + "memory(GiB)": 111.15, + "step": 28025, + "train_speed(iter/s)": 0.448222 + }, + { + "acc": 0.65338049, + "epoch": 0.7110603754439371, + "grad_norm": 4.9375, + "learning_rate": 7.657737262097128e-06, + "loss": 1.59846611, + "memory(GiB)": 111.15, + "step": 28030, + "train_speed(iter/s)": 0.448263 + }, + { + "acc": 0.65676594, + "epoch": 0.7111872146118722, + "grad_norm": 4.96875, + "learning_rate": 7.656848989495733e-06, + "loss": 1.59982052, + "memory(GiB)": 111.15, + "step": 28035, + "train_speed(iter/s)": 0.448304 + }, + { + "acc": 0.66353145, + "epoch": 0.7113140537798072, + "grad_norm": 5.4375, + "learning_rate": 7.65596060003426e-06, + "loss": 1.65233459, + "memory(GiB)": 111.15, + "step": 28040, + "train_speed(iter/s)": 0.448346 + }, + { + "acc": 0.63607554, + "epoch": 0.7114408929477423, + "grad_norm": 7.59375, + "learning_rate": 7.655072093751779e-06, + "loss": 1.71882477, + "memory(GiB)": 111.15, + "step": 28045, + "train_speed(iter/s)": 0.448388 + }, + { + "acc": 0.65272174, + "epoch": 0.7115677321156774, + "grad_norm": 6.03125, + "learning_rate": 7.654183470687375e-06, + "loss": 1.65077343, + "memory(GiB)": 111.15, + "step": 28050, + "train_speed(iter/s)": 0.44843 + }, + { + "acc": 0.66287422, + "epoch": 0.7116945712836124, + "grad_norm": 5.40625, + "learning_rate": 7.653294730880131e-06, + "loss": 1.6063942, + "memory(GiB)": 111.15, + "step": 28055, + "train_speed(iter/s)": 0.448472 + }, + { + "acc": 0.65929871, + "epoch": 0.7118214104515475, + "grad_norm": 5.625, + "learning_rate": 7.65240587436914e-06, + "loss": 1.5426405, + "memory(GiB)": 111.15, + "step": 28060, + "train_speed(iter/s)": 0.448514 + }, + { + "acc": 0.64724331, + "epoch": 0.7119482496194824, + "grad_norm": 4.90625, + "learning_rate": 7.651516901193494e-06, + "loss": 1.68215637, + "memory(GiB)": 111.15, + "step": 28065, + "train_speed(iter/s)": 0.448556 + }, + { + "acc": 0.66206102, + "epoch": 0.7120750887874175, + "grad_norm": 5.5, + "learning_rate": 7.650627811392298e-06, + "loss": 1.55513611, + "memory(GiB)": 111.15, + "step": 28070, + "train_speed(iter/s)": 0.448597 + }, + { + "acc": 0.63644247, + "epoch": 0.7122019279553526, + "grad_norm": 5.1875, + "learning_rate": 7.649738605004658e-06, + "loss": 1.64522476, + "memory(GiB)": 111.15, + "step": 28075, + "train_speed(iter/s)": 0.448639 + }, + { + "acc": 0.66124306, + "epoch": 0.7123287671232876, + "grad_norm": 4.9375, + "learning_rate": 7.648849282069682e-06, + "loss": 1.50599651, + "memory(GiB)": 111.15, + "step": 28080, + "train_speed(iter/s)": 0.448681 + }, + { + "acc": 0.65081091, + "epoch": 0.7124556062912227, + "grad_norm": 4.84375, + "learning_rate": 7.647959842626489e-06, + "loss": 1.653162, + "memory(GiB)": 111.15, + "step": 28085, + "train_speed(iter/s)": 0.448723 + }, + { + "acc": 0.64328694, + "epoch": 0.7125824454591578, + "grad_norm": 5.21875, + "learning_rate": 7.6470702867142e-06, + "loss": 1.63627319, + "memory(GiB)": 111.15, + "step": 28090, + "train_speed(iter/s)": 0.448765 + }, + { + "acc": 0.63722239, + "epoch": 0.7127092846270928, + "grad_norm": 5.09375, + "learning_rate": 7.646180614371941e-06, + "loss": 1.66458702, + "memory(GiB)": 111.15, + "step": 28095, + "train_speed(iter/s)": 0.448807 + }, + { + "acc": 0.64690638, + "epoch": 0.7128361237950279, + "grad_norm": 6.8125, + "learning_rate": 7.645290825638845e-06, + "loss": 1.66203022, + "memory(GiB)": 111.15, + "step": 28100, + "train_speed(iter/s)": 0.448848 + }, + { + "epoch": 0.7128361237950279, + "eval_acc": 0.6425566189590325, + "eval_loss": 1.6122413873672485, + "eval_runtime": 113.1054, + "eval_samples_per_second": 56.319, + "eval_steps_per_second": 28.16, + "step": 28100 + }, + { + "acc": 0.64906578, + "epoch": 0.7129629629629629, + "grad_norm": 6.4375, + "learning_rate": 7.644400920554048e-06, + "loss": 1.63195, + "memory(GiB)": 111.15, + "step": 28105, + "train_speed(iter/s)": 0.448048 + }, + { + "acc": 0.65014777, + "epoch": 0.713089802130898, + "grad_norm": 4.71875, + "learning_rate": 7.64351089915669e-06, + "loss": 1.67783909, + "memory(GiB)": 111.15, + "step": 28110, + "train_speed(iter/s)": 0.44809 + }, + { + "acc": 0.67413301, + "epoch": 0.7132166412988331, + "grad_norm": 5.28125, + "learning_rate": 7.642620761485921e-06, + "loss": 1.53314724, + "memory(GiB)": 111.15, + "step": 28115, + "train_speed(iter/s)": 0.448132 + }, + { + "acc": 0.65261507, + "epoch": 0.7133434804667681, + "grad_norm": 5.15625, + "learning_rate": 7.641730507580896e-06, + "loss": 1.62534885, + "memory(GiB)": 111.15, + "step": 28120, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.6371027, + "epoch": 0.7134703196347032, + "grad_norm": 4.90625, + "learning_rate": 7.640840137480763e-06, + "loss": 1.69881382, + "memory(GiB)": 111.15, + "step": 28125, + "train_speed(iter/s)": 0.448215 + }, + { + "acc": 0.64986377, + "epoch": 0.7135971588026383, + "grad_norm": 5.21875, + "learning_rate": 7.639949651224697e-06, + "loss": 1.58542967, + "memory(GiB)": 111.15, + "step": 28130, + "train_speed(iter/s)": 0.448258 + }, + { + "acc": 0.63611469, + "epoch": 0.7137239979705733, + "grad_norm": 5.375, + "learning_rate": 7.639059048851853e-06, + "loss": 1.69374542, + "memory(GiB)": 111.15, + "step": 28135, + "train_speed(iter/s)": 0.4483 + }, + { + "acc": 0.66080828, + "epoch": 0.7138508371385084, + "grad_norm": 5.0625, + "learning_rate": 7.638168330401412e-06, + "loss": 1.56179647, + "memory(GiB)": 111.15, + "step": 28140, + "train_speed(iter/s)": 0.448342 + }, + { + "acc": 0.64395189, + "epoch": 0.7139776763064434, + "grad_norm": 6.4375, + "learning_rate": 7.637277495912548e-06, + "loss": 1.63141632, + "memory(GiB)": 111.15, + "step": 28145, + "train_speed(iter/s)": 0.448384 + }, + { + "acc": 0.65748024, + "epoch": 0.7141045154743785, + "grad_norm": 4.90625, + "learning_rate": 7.636386545424447e-06, + "loss": 1.54546852, + "memory(GiB)": 111.15, + "step": 28150, + "train_speed(iter/s)": 0.448426 + }, + { + "acc": 0.64391646, + "epoch": 0.7142313546423136, + "grad_norm": 5.4375, + "learning_rate": 7.635495478976294e-06, + "loss": 1.63982773, + "memory(GiB)": 111.15, + "step": 28155, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.64258366, + "epoch": 0.7143581938102486, + "grad_norm": 6.09375, + "learning_rate": 7.634604296607285e-06, + "loss": 1.74375839, + "memory(GiB)": 111.15, + "step": 28160, + "train_speed(iter/s)": 0.44851 + }, + { + "acc": 0.65149736, + "epoch": 0.7144850329781837, + "grad_norm": 5.09375, + "learning_rate": 7.633712998356612e-06, + "loss": 1.55160713, + "memory(GiB)": 111.15, + "step": 28165, + "train_speed(iter/s)": 0.448552 + }, + { + "acc": 0.64023647, + "epoch": 0.7146118721461188, + "grad_norm": 5.875, + "learning_rate": 7.632821584263486e-06, + "loss": 1.68535461, + "memory(GiB)": 111.15, + "step": 28170, + "train_speed(iter/s)": 0.448594 + }, + { + "acc": 0.65281858, + "epoch": 0.7147387113140538, + "grad_norm": 5.15625, + "learning_rate": 7.631930054367112e-06, + "loss": 1.63781471, + "memory(GiB)": 111.15, + "step": 28175, + "train_speed(iter/s)": 0.448637 + }, + { + "acc": 0.65719037, + "epoch": 0.7148655504819889, + "grad_norm": 4.78125, + "learning_rate": 7.631038408706703e-06, + "loss": 1.58153877, + "memory(GiB)": 111.15, + "step": 28180, + "train_speed(iter/s)": 0.448679 + }, + { + "acc": 0.64706383, + "epoch": 0.7149923896499238, + "grad_norm": 5.375, + "learning_rate": 7.630146647321476e-06, + "loss": 1.60407104, + "memory(GiB)": 111.15, + "step": 28185, + "train_speed(iter/s)": 0.448721 + }, + { + "acc": 0.64396887, + "epoch": 0.7151192288178589, + "grad_norm": 4.75, + "learning_rate": 7.62925477025066e-06, + "loss": 1.68766193, + "memory(GiB)": 111.15, + "step": 28190, + "train_speed(iter/s)": 0.448763 + }, + { + "acc": 0.64180174, + "epoch": 0.715246067985794, + "grad_norm": 5.34375, + "learning_rate": 7.628362777533479e-06, + "loss": 1.70164719, + "memory(GiB)": 111.15, + "step": 28195, + "train_speed(iter/s)": 0.448805 + }, + { + "acc": 0.64566021, + "epoch": 0.715372907153729, + "grad_norm": 5.46875, + "learning_rate": 7.627470669209169e-06, + "loss": 1.64249153, + "memory(GiB)": 111.15, + "step": 28200, + "train_speed(iter/s)": 0.448847 + }, + { + "epoch": 0.715372907153729, + "eval_acc": 0.6425633032359838, + "eval_loss": 1.6124693155288696, + "eval_runtime": 114.8212, + "eval_samples_per_second": 55.478, + "eval_steps_per_second": 27.739, + "step": 28200 + }, + { + "acc": 0.65479898, + "epoch": 0.7154997463216641, + "grad_norm": 5.125, + "learning_rate": 7.626578445316968e-06, + "loss": 1.58418045, + "memory(GiB)": 111.15, + "step": 28205, + "train_speed(iter/s)": 0.448038 + }, + { + "acc": 0.6456512, + "epoch": 0.7156265854895992, + "grad_norm": 5.375, + "learning_rate": 7.62568610589612e-06, + "loss": 1.63546104, + "memory(GiB)": 111.15, + "step": 28210, + "train_speed(iter/s)": 0.448079 + }, + { + "acc": 0.64035544, + "epoch": 0.7157534246575342, + "grad_norm": 7.15625, + "learning_rate": 7.624793650985873e-06, + "loss": 1.68172932, + "memory(GiB)": 111.15, + "step": 28215, + "train_speed(iter/s)": 0.448121 + }, + { + "acc": 0.63746405, + "epoch": 0.7158802638254693, + "grad_norm": 5.25, + "learning_rate": 7.6239010806254835e-06, + "loss": 1.64677353, + "memory(GiB)": 111.15, + "step": 28220, + "train_speed(iter/s)": 0.448163 + }, + { + "acc": 0.64205542, + "epoch": 0.7160071029934043, + "grad_norm": 5.53125, + "learning_rate": 7.6230083948542084e-06, + "loss": 1.66778622, + "memory(GiB)": 111.15, + "step": 28225, + "train_speed(iter/s)": 0.448204 + }, + { + "acc": 0.66347704, + "epoch": 0.7161339421613394, + "grad_norm": 5.21875, + "learning_rate": 7.622115593711314e-06, + "loss": 1.5568841, + "memory(GiB)": 111.15, + "step": 28230, + "train_speed(iter/s)": 0.448245 + }, + { + "acc": 0.64962559, + "epoch": 0.7162607813292745, + "grad_norm": 4.6875, + "learning_rate": 7.62122267723607e-06, + "loss": 1.60519352, + "memory(GiB)": 111.15, + "step": 28235, + "train_speed(iter/s)": 0.448287 + }, + { + "acc": 0.65822978, + "epoch": 0.7163876204972095, + "grad_norm": 4.625, + "learning_rate": 7.620329645467748e-06, + "loss": 1.54508667, + "memory(GiB)": 111.15, + "step": 28240, + "train_speed(iter/s)": 0.448328 + }, + { + "acc": 0.63690476, + "epoch": 0.7165144596651446, + "grad_norm": 5.46875, + "learning_rate": 7.61943649844563e-06, + "loss": 1.73837929, + "memory(GiB)": 111.15, + "step": 28245, + "train_speed(iter/s)": 0.44837 + }, + { + "acc": 0.64081907, + "epoch": 0.7166412988330797, + "grad_norm": 5.0625, + "learning_rate": 7.618543236209001e-06, + "loss": 1.66744194, + "memory(GiB)": 111.15, + "step": 28250, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.64322939, + "epoch": 0.7167681380010147, + "grad_norm": 5.15625, + "learning_rate": 7.617649858797147e-06, + "loss": 1.65141373, + "memory(GiB)": 111.15, + "step": 28255, + "train_speed(iter/s)": 0.448453 + }, + { + "acc": 0.65029678, + "epoch": 0.7168949771689498, + "grad_norm": 5.1875, + "learning_rate": 7.616756366249367e-06, + "loss": 1.64547672, + "memory(GiB)": 111.15, + "step": 28260, + "train_speed(iter/s)": 0.448495 + }, + { + "acc": 0.63830051, + "epoch": 0.7170218163368848, + "grad_norm": 4.59375, + "learning_rate": 7.6158627586049586e-06, + "loss": 1.68375931, + "memory(GiB)": 111.15, + "step": 28265, + "train_speed(iter/s)": 0.448537 + }, + { + "acc": 0.65948009, + "epoch": 0.7171486555048199, + "grad_norm": 5.53125, + "learning_rate": 7.614969035903228e-06, + "loss": 1.58789034, + "memory(GiB)": 111.15, + "step": 28270, + "train_speed(iter/s)": 0.448578 + }, + { + "acc": 0.65647483, + "epoch": 0.717275494672755, + "grad_norm": 5.09375, + "learning_rate": 7.614075198183482e-06, + "loss": 1.54205141, + "memory(GiB)": 111.15, + "step": 28275, + "train_speed(iter/s)": 0.44862 + }, + { + "acc": 0.64715729, + "epoch": 0.71740233384069, + "grad_norm": 7.34375, + "learning_rate": 7.6131812454850406e-06, + "loss": 1.63954468, + "memory(GiB)": 111.15, + "step": 28280, + "train_speed(iter/s)": 0.448662 + }, + { + "acc": 0.65084963, + "epoch": 0.7175291730086251, + "grad_norm": 5.28125, + "learning_rate": 7.612287177847219e-06, + "loss": 1.64505577, + "memory(GiB)": 111.15, + "step": 28285, + "train_speed(iter/s)": 0.448703 + }, + { + "acc": 0.65606089, + "epoch": 0.7176560121765602, + "grad_norm": 5.0, + "learning_rate": 7.611392995309345e-06, + "loss": 1.60830936, + "memory(GiB)": 111.15, + "step": 28290, + "train_speed(iter/s)": 0.448745 + }, + { + "acc": 0.65350919, + "epoch": 0.7177828513444952, + "grad_norm": 6.15625, + "learning_rate": 7.610498697910748e-06, + "loss": 1.59084845, + "memory(GiB)": 111.15, + "step": 28295, + "train_speed(iter/s)": 0.448787 + }, + { + "acc": 0.65306807, + "epoch": 0.7179096905124303, + "grad_norm": 4.78125, + "learning_rate": 7.609604285690762e-06, + "loss": 1.60806999, + "memory(GiB)": 111.15, + "step": 28300, + "train_speed(iter/s)": 0.448829 + }, + { + "epoch": 0.7179096905124303, + "eval_acc": 0.6425374016627975, + "eval_loss": 1.6123080253601074, + "eval_runtime": 113.7948, + "eval_samples_per_second": 55.978, + "eval_steps_per_second": 27.989, + "step": 28300 + }, + { + "acc": 0.66678743, + "epoch": 0.7180365296803652, + "grad_norm": 5.6875, + "learning_rate": 7.608709758688731e-06, + "loss": 1.57819805, + "memory(GiB)": 111.15, + "step": 28305, + "train_speed(iter/s)": 0.448029 + }, + { + "acc": 0.64170251, + "epoch": 0.7181633688483003, + "grad_norm": 5.28125, + "learning_rate": 7.607815116943995e-06, + "loss": 1.60779572, + "memory(GiB)": 111.15, + "step": 28310, + "train_speed(iter/s)": 0.448071 + }, + { + "acc": 0.65149889, + "epoch": 0.7182902080162354, + "grad_norm": 4.90625, + "learning_rate": 7.606920360495908e-06, + "loss": 1.57017336, + "memory(GiB)": 111.15, + "step": 28315, + "train_speed(iter/s)": 0.448112 + }, + { + "acc": 0.63614807, + "epoch": 0.7184170471841704, + "grad_norm": 4.84375, + "learning_rate": 7.6060254893838255e-06, + "loss": 1.68340302, + "memory(GiB)": 111.15, + "step": 28320, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.63547826, + "epoch": 0.7185438863521055, + "grad_norm": 5.09375, + "learning_rate": 7.6051305036471065e-06, + "loss": 1.65224915, + "memory(GiB)": 111.15, + "step": 28325, + "train_speed(iter/s)": 0.448196 + }, + { + "acc": 0.6584866, + "epoch": 0.7186707255200406, + "grad_norm": 5.5625, + "learning_rate": 7.604235403325117e-06, + "loss": 1.56153555, + "memory(GiB)": 111.15, + "step": 28330, + "train_speed(iter/s)": 0.448237 + }, + { + "acc": 0.64270077, + "epoch": 0.7187975646879756, + "grad_norm": 5.46875, + "learning_rate": 7.603340188457227e-06, + "loss": 1.64957943, + "memory(GiB)": 111.15, + "step": 28335, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.64315095, + "epoch": 0.7189244038559107, + "grad_norm": 5.03125, + "learning_rate": 7.602444859082814e-06, + "loss": 1.65805264, + "memory(GiB)": 111.15, + "step": 28340, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.64658608, + "epoch": 0.7190512430238457, + "grad_norm": 5.6875, + "learning_rate": 7.601549415241254e-06, + "loss": 1.64475327, + "memory(GiB)": 111.15, + "step": 28345, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.65261307, + "epoch": 0.7191780821917808, + "grad_norm": 6.40625, + "learning_rate": 7.6006538569719375e-06, + "loss": 1.60732746, + "memory(GiB)": 111.15, + "step": 28350, + "train_speed(iter/s)": 0.448404 + }, + { + "acc": 0.64100513, + "epoch": 0.7193049213597159, + "grad_norm": 5.4375, + "learning_rate": 7.599758184314252e-06, + "loss": 1.64776955, + "memory(GiB)": 111.15, + "step": 28355, + "train_speed(iter/s)": 0.448445 + }, + { + "acc": 0.64474869, + "epoch": 0.7194317605276509, + "grad_norm": 5.75, + "learning_rate": 7.598862397307596e-06, + "loss": 1.65211258, + "memory(GiB)": 111.15, + "step": 28360, + "train_speed(iter/s)": 0.448487 + }, + { + "acc": 0.68190455, + "epoch": 0.719558599695586, + "grad_norm": 5.125, + "learning_rate": 7.597966495991368e-06, + "loss": 1.5192728, + "memory(GiB)": 111.15, + "step": 28365, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.65800877, + "epoch": 0.7196854388635211, + "grad_norm": 5.21875, + "learning_rate": 7.597070480404974e-06, + "loss": 1.52145252, + "memory(GiB)": 111.15, + "step": 28370, + "train_speed(iter/s)": 0.44857 + }, + { + "acc": 0.64715986, + "epoch": 0.7198122780314561, + "grad_norm": 4.8125, + "learning_rate": 7.596174350587826e-06, + "loss": 1.62580719, + "memory(GiB)": 111.15, + "step": 28375, + "train_speed(iter/s)": 0.448612 + }, + { + "acc": 0.63876915, + "epoch": 0.7199391171993912, + "grad_norm": 4.8125, + "learning_rate": 7.595278106579339e-06, + "loss": 1.6495079, + "memory(GiB)": 111.15, + "step": 28380, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.65382843, + "epoch": 0.7200659563673262, + "grad_norm": 5.625, + "learning_rate": 7.594381748418933e-06, + "loss": 1.59350777, + "memory(GiB)": 111.15, + "step": 28385, + "train_speed(iter/s)": 0.448695 + }, + { + "acc": 0.6486764, + "epoch": 0.7201927955352613, + "grad_norm": 4.875, + "learning_rate": 7.593485276146035e-06, + "loss": 1.59397182, + "memory(GiB)": 111.15, + "step": 28390, + "train_speed(iter/s)": 0.448736 + }, + { + "acc": 0.6717227, + "epoch": 0.7203196347031964, + "grad_norm": 8.125, + "learning_rate": 7.592588689800077e-06, + "loss": 1.61176147, + "memory(GiB)": 111.15, + "step": 28395, + "train_speed(iter/s)": 0.448778 + }, + { + "acc": 0.63578296, + "epoch": 0.7204464738711314, + "grad_norm": 5.40625, + "learning_rate": 7.591691989420491e-06, + "loss": 1.6722311, + "memory(GiB)": 111.15, + "step": 28400, + "train_speed(iter/s)": 0.44882 + }, + { + "epoch": 0.7204464738711314, + "eval_acc": 0.6424601147105479, + "eval_loss": 1.6121509075164795, + "eval_runtime": 113.4385, + "eval_samples_per_second": 56.154, + "eval_steps_per_second": 28.077, + "step": 28400 + }, + { + "acc": 0.66399651, + "epoch": 0.7205733130390665, + "grad_norm": 5.4375, + "learning_rate": 7.590795175046721e-06, + "loss": 1.57033453, + "memory(GiB)": 111.15, + "step": 28405, + "train_speed(iter/s)": 0.448026 + }, + { + "acc": 0.64412985, + "epoch": 0.7207001522070016, + "grad_norm": 5.875, + "learning_rate": 7.5898982467182125e-06, + "loss": 1.59988861, + "memory(GiB)": 111.15, + "step": 28410, + "train_speed(iter/s)": 0.448067 + }, + { + "acc": 0.64197264, + "epoch": 0.7208269913749366, + "grad_norm": 6.1875, + "learning_rate": 7.589001204474416e-06, + "loss": 1.6344862, + "memory(GiB)": 111.15, + "step": 28415, + "train_speed(iter/s)": 0.448109 + }, + { + "acc": 0.6310739, + "epoch": 0.7209538305428717, + "grad_norm": 4.96875, + "learning_rate": 7.588104048354787e-06, + "loss": 1.64502983, + "memory(GiB)": 111.15, + "step": 28420, + "train_speed(iter/s)": 0.44815 + }, + { + "acc": 0.64789295, + "epoch": 0.7210806697108066, + "grad_norm": 5.625, + "learning_rate": 7.587206778398788e-06, + "loss": 1.62225685, + "memory(GiB)": 111.15, + "step": 28425, + "train_speed(iter/s)": 0.448192 + }, + { + "acc": 0.66442051, + "epoch": 0.7212075088787417, + "grad_norm": 6.96875, + "learning_rate": 7.586309394645882e-06, + "loss": 1.54275627, + "memory(GiB)": 111.15, + "step": 28430, + "train_speed(iter/s)": 0.448234 + }, + { + "acc": 0.66298304, + "epoch": 0.7213343480466768, + "grad_norm": 5.59375, + "learning_rate": 7.585411897135544e-06, + "loss": 1.54224701, + "memory(GiB)": 111.15, + "step": 28435, + "train_speed(iter/s)": 0.448275 + }, + { + "acc": 0.64935732, + "epoch": 0.7214611872146118, + "grad_norm": 6.375, + "learning_rate": 7.584514285907245e-06, + "loss": 1.55328608, + "memory(GiB)": 111.15, + "step": 28440, + "train_speed(iter/s)": 0.448317 + }, + { + "acc": 0.66264, + "epoch": 0.7215880263825469, + "grad_norm": 5.53125, + "learning_rate": 7.583616561000471e-06, + "loss": 1.57036123, + "memory(GiB)": 111.15, + "step": 28445, + "train_speed(iter/s)": 0.448359 + }, + { + "acc": 0.63986869, + "epoch": 0.721714865550482, + "grad_norm": 4.78125, + "learning_rate": 7.582718722454705e-06, + "loss": 1.67435951, + "memory(GiB)": 111.15, + "step": 28450, + "train_speed(iter/s)": 0.4484 + }, + { + "acc": 0.64611959, + "epoch": 0.721841704718417, + "grad_norm": 5.28125, + "learning_rate": 7.581820770309438e-06, + "loss": 1.66436386, + "memory(GiB)": 111.15, + "step": 28455, + "train_speed(iter/s)": 0.448441 + }, + { + "acc": 0.67084503, + "epoch": 0.7219685438863521, + "grad_norm": 5.4375, + "learning_rate": 7.580922704604168e-06, + "loss": 1.56290293, + "memory(GiB)": 111.15, + "step": 28460, + "train_speed(iter/s)": 0.448482 + }, + { + "acc": 0.63449221, + "epoch": 0.7220953830542871, + "grad_norm": 5.21875, + "learning_rate": 7.5800245253783935e-06, + "loss": 1.64534607, + "memory(GiB)": 111.15, + "step": 28465, + "train_speed(iter/s)": 0.448524 + }, + { + "acc": 0.63529181, + "epoch": 0.7222222222222222, + "grad_norm": 5.375, + "learning_rate": 7.579126232671621e-06, + "loss": 1.63519974, + "memory(GiB)": 111.15, + "step": 28470, + "train_speed(iter/s)": 0.448566 + }, + { + "acc": 0.65685205, + "epoch": 0.7223490613901573, + "grad_norm": 6.53125, + "learning_rate": 7.578227826523361e-06, + "loss": 1.68714848, + "memory(GiB)": 111.15, + "step": 28475, + "train_speed(iter/s)": 0.448607 + }, + { + "acc": 0.66871223, + "epoch": 0.7224759005580923, + "grad_norm": 4.9375, + "learning_rate": 7.577329306973132e-06, + "loss": 1.59263325, + "memory(GiB)": 111.15, + "step": 28480, + "train_speed(iter/s)": 0.448648 + }, + { + "acc": 0.66050787, + "epoch": 0.7226027397260274, + "grad_norm": 5.3125, + "learning_rate": 7.576430674060452e-06, + "loss": 1.5566246, + "memory(GiB)": 111.15, + "step": 28485, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.63799086, + "epoch": 0.7227295788939625, + "grad_norm": 5.40625, + "learning_rate": 7.575531927824849e-06, + "loss": 1.68423576, + "memory(GiB)": 111.15, + "step": 28490, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.65548592, + "epoch": 0.7228564180618975, + "grad_norm": 5.78125, + "learning_rate": 7.574633068305852e-06, + "loss": 1.66707268, + "memory(GiB)": 111.15, + "step": 28495, + "train_speed(iter/s)": 0.448772 + }, + { + "acc": 0.64841318, + "epoch": 0.7229832572298326, + "grad_norm": 6.3125, + "learning_rate": 7.5737340955429995e-06, + "loss": 1.66877174, + "memory(GiB)": 111.15, + "step": 28500, + "train_speed(iter/s)": 0.448814 + }, + { + "epoch": 0.7229832572298326, + "eval_acc": 0.6425607966321271, + "eval_loss": 1.6121740341186523, + "eval_runtime": 113.36, + "eval_samples_per_second": 56.193, + "eval_steps_per_second": 28.096, + "step": 28500 + }, + { + "acc": 0.65799932, + "epoch": 0.7231100963977676, + "grad_norm": 5.125, + "learning_rate": 7.572835009575828e-06, + "loss": 1.5571414, + "memory(GiB)": 111.15, + "step": 28505, + "train_speed(iter/s)": 0.448023 + }, + { + "acc": 0.65211072, + "epoch": 0.7232369355657027, + "grad_norm": 6.46875, + "learning_rate": 7.571935810443886e-06, + "loss": 1.63289719, + "memory(GiB)": 111.15, + "step": 28510, + "train_speed(iter/s)": 0.448065 + }, + { + "acc": 0.66095533, + "epoch": 0.7233637747336378, + "grad_norm": 4.625, + "learning_rate": 7.571036498186727e-06, + "loss": 1.57812681, + "memory(GiB)": 111.15, + "step": 28515, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.6423419, + "epoch": 0.7234906139015728, + "grad_norm": 5.96875, + "learning_rate": 7.570137072843902e-06, + "loss": 1.66683178, + "memory(GiB)": 111.15, + "step": 28520, + "train_speed(iter/s)": 0.448148 + }, + { + "acc": 0.64330883, + "epoch": 0.7236174530695079, + "grad_norm": 5.75, + "learning_rate": 7.569237534454974e-06, + "loss": 1.65316982, + "memory(GiB)": 111.15, + "step": 28525, + "train_speed(iter/s)": 0.44819 + }, + { + "acc": 0.63790483, + "epoch": 0.723744292237443, + "grad_norm": 5.25, + "learning_rate": 7.568337883059509e-06, + "loss": 1.66292934, + "memory(GiB)": 111.15, + "step": 28530, + "train_speed(iter/s)": 0.448231 + }, + { + "acc": 0.64760857, + "epoch": 0.723871131405378, + "grad_norm": 5.71875, + "learning_rate": 7.567438118697077e-06, + "loss": 1.66241817, + "memory(GiB)": 111.15, + "step": 28535, + "train_speed(iter/s)": 0.448271 + }, + { + "acc": 0.64841928, + "epoch": 0.723997970573313, + "grad_norm": 5.90625, + "learning_rate": 7.566538241407253e-06, + "loss": 1.58126583, + "memory(GiB)": 111.15, + "step": 28540, + "train_speed(iter/s)": 0.448313 + }, + { + "acc": 0.64039202, + "epoch": 0.724124809741248, + "grad_norm": 4.90625, + "learning_rate": 7.565638251229617e-06, + "loss": 1.6442791, + "memory(GiB)": 111.15, + "step": 28545, + "train_speed(iter/s)": 0.448354 + }, + { + "acc": 0.66028471, + "epoch": 0.7242516489091831, + "grad_norm": 5.15625, + "learning_rate": 7.5647381482037585e-06, + "loss": 1.56645098, + "memory(GiB)": 111.15, + "step": 28550, + "train_speed(iter/s)": 0.448395 + }, + { + "acc": 0.65618877, + "epoch": 0.7243784880771182, + "grad_norm": 4.71875, + "learning_rate": 7.563837932369264e-06, + "loss": 1.62528687, + "memory(GiB)": 111.15, + "step": 28555, + "train_speed(iter/s)": 0.448436 + }, + { + "acc": 0.64975863, + "epoch": 0.7245053272450532, + "grad_norm": 5.375, + "learning_rate": 7.562937603765732e-06, + "loss": 1.58730507, + "memory(GiB)": 111.15, + "step": 28560, + "train_speed(iter/s)": 0.448477 + }, + { + "acc": 0.64938073, + "epoch": 0.7246321664129883, + "grad_norm": 5.1875, + "learning_rate": 7.562037162432761e-06, + "loss": 1.61337662, + "memory(GiB)": 111.15, + "step": 28565, + "train_speed(iter/s)": 0.448518 + }, + { + "acc": 0.63914542, + "epoch": 0.7247590055809234, + "grad_norm": 5.4375, + "learning_rate": 7.561136608409956e-06, + "loss": 1.59878969, + "memory(GiB)": 111.15, + "step": 28570, + "train_speed(iter/s)": 0.448559 + }, + { + "acc": 0.636871, + "epoch": 0.7248858447488584, + "grad_norm": 5.125, + "learning_rate": 7.560235941736929e-06, + "loss": 1.72807541, + "memory(GiB)": 111.15, + "step": 28575, + "train_speed(iter/s)": 0.4486 + }, + { + "acc": 0.64130535, + "epoch": 0.7250126839167935, + "grad_norm": 5.5625, + "learning_rate": 7.559335162453294e-06, + "loss": 1.67029533, + "memory(GiB)": 111.15, + "step": 28580, + "train_speed(iter/s)": 0.448641 + }, + { + "acc": 0.65966034, + "epoch": 0.7251395230847285, + "grad_norm": 6.75, + "learning_rate": 7.558434270598672e-06, + "loss": 1.59847355, + "memory(GiB)": 111.15, + "step": 28585, + "train_speed(iter/s)": 0.448682 + }, + { + "acc": 0.64996829, + "epoch": 0.7252663622526636, + "grad_norm": 5.46875, + "learning_rate": 7.5575332662126885e-06, + "loss": 1.62838821, + "memory(GiB)": 111.15, + "step": 28590, + "train_speed(iter/s)": 0.448724 + }, + { + "acc": 0.65180893, + "epoch": 0.7253932014205987, + "grad_norm": 5.09375, + "learning_rate": 7.556632149334975e-06, + "loss": 1.56949768, + "memory(GiB)": 111.15, + "step": 28595, + "train_speed(iter/s)": 0.448765 + }, + { + "acc": 0.65328817, + "epoch": 0.7255200405885337, + "grad_norm": 4.96875, + "learning_rate": 7.555730920005163e-06, + "loss": 1.56504173, + "memory(GiB)": 111.15, + "step": 28600, + "train_speed(iter/s)": 0.448805 + }, + { + "epoch": 0.7255200405885337, + "eval_acc": 0.6426138530804282, + "eval_loss": 1.6121642589569092, + "eval_runtime": 113.5938, + "eval_samples_per_second": 56.077, + "eval_steps_per_second": 28.039, + "step": 28600 + }, + { + "acc": 0.61114445, + "epoch": 0.7256468797564688, + "grad_norm": 6.0625, + "learning_rate": 7.554829578262894e-06, + "loss": 1.71723938, + "memory(GiB)": 111.15, + "step": 28605, + "train_speed(iter/s)": 0.448016 + }, + { + "acc": 0.65944786, + "epoch": 0.7257737189244039, + "grad_norm": 7.03125, + "learning_rate": 7.5539281241478155e-06, + "loss": 1.64690819, + "memory(GiB)": 111.15, + "step": 28610, + "train_speed(iter/s)": 0.448057 + }, + { + "acc": 0.66694994, + "epoch": 0.7259005580923389, + "grad_norm": 4.875, + "learning_rate": 7.5530265576995756e-06, + "loss": 1.56313114, + "memory(GiB)": 111.15, + "step": 28615, + "train_speed(iter/s)": 0.448099 + }, + { + "acc": 0.64592142, + "epoch": 0.726027397260274, + "grad_norm": 4.84375, + "learning_rate": 7.552124878957829e-06, + "loss": 1.62813339, + "memory(GiB)": 111.15, + "step": 28620, + "train_speed(iter/s)": 0.44814 + }, + { + "acc": 0.64444695, + "epoch": 0.726154236428209, + "grad_norm": 5.46875, + "learning_rate": 7.551223087962234e-06, + "loss": 1.65249825, + "memory(GiB)": 111.15, + "step": 28625, + "train_speed(iter/s)": 0.44818 + }, + { + "acc": 0.66640916, + "epoch": 0.7262810755961441, + "grad_norm": 4.875, + "learning_rate": 7.55032118475246e-06, + "loss": 1.55896797, + "memory(GiB)": 111.15, + "step": 28630, + "train_speed(iter/s)": 0.448221 + }, + { + "acc": 0.64687228, + "epoch": 0.7264079147640792, + "grad_norm": 4.40625, + "learning_rate": 7.549419169368171e-06, + "loss": 1.64751015, + "memory(GiB)": 111.15, + "step": 28635, + "train_speed(iter/s)": 0.448262 + }, + { + "acc": 0.63368607, + "epoch": 0.7265347539320142, + "grad_norm": 5.375, + "learning_rate": 7.548517041849048e-06, + "loss": 1.67703037, + "memory(GiB)": 111.15, + "step": 28640, + "train_speed(iter/s)": 0.448304 + }, + { + "acc": 0.65365939, + "epoch": 0.7266615930999493, + "grad_norm": 5.03125, + "learning_rate": 7.547614802234764e-06, + "loss": 1.60639496, + "memory(GiB)": 111.15, + "step": 28645, + "train_speed(iter/s)": 0.448345 + }, + { + "acc": 0.66404791, + "epoch": 0.7267884322678844, + "grad_norm": 4.59375, + "learning_rate": 7.546712450565008e-06, + "loss": 1.66870461, + "memory(GiB)": 111.15, + "step": 28650, + "train_speed(iter/s)": 0.448386 + }, + { + "acc": 0.66565199, + "epoch": 0.7269152714358194, + "grad_norm": 4.4375, + "learning_rate": 7.545809986879469e-06, + "loss": 1.5006628, + "memory(GiB)": 111.15, + "step": 28655, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.65227356, + "epoch": 0.7270421106037545, + "grad_norm": 6.28125, + "learning_rate": 7.5449074112178385e-06, + "loss": 1.63367996, + "memory(GiB)": 111.15, + "step": 28660, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.65384545, + "epoch": 0.7271689497716894, + "grad_norm": 6.09375, + "learning_rate": 7.54400472361982e-06, + "loss": 1.6893116, + "memory(GiB)": 111.15, + "step": 28665, + "train_speed(iter/s)": 0.448509 + }, + { + "acc": 0.63634276, + "epoch": 0.7272957889396245, + "grad_norm": 5.3125, + "learning_rate": 7.543101924125115e-06, + "loss": 1.64826126, + "memory(GiB)": 111.15, + "step": 28670, + "train_speed(iter/s)": 0.44855 + }, + { + "acc": 0.64760866, + "epoch": 0.7274226281075596, + "grad_norm": 6.0625, + "learning_rate": 7.542199012773432e-06, + "loss": 1.70191269, + "memory(GiB)": 111.15, + "step": 28675, + "train_speed(iter/s)": 0.448592 + }, + { + "acc": 0.65282965, + "epoch": 0.7275494672754946, + "grad_norm": 5.53125, + "learning_rate": 7.541295989604488e-06, + "loss": 1.65911636, + "memory(GiB)": 111.15, + "step": 28680, + "train_speed(iter/s)": 0.448633 + }, + { + "acc": 0.66410999, + "epoch": 0.7276763064434297, + "grad_norm": 5.75, + "learning_rate": 7.540392854657999e-06, + "loss": 1.55660191, + "memory(GiB)": 111.15, + "step": 28685, + "train_speed(iter/s)": 0.448673 + }, + { + "acc": 0.6553587, + "epoch": 0.7278031456113648, + "grad_norm": 6.53125, + "learning_rate": 7.539489607973691e-06, + "loss": 1.61225548, + "memory(GiB)": 111.15, + "step": 28690, + "train_speed(iter/s)": 0.448715 + }, + { + "acc": 0.64829378, + "epoch": 0.7279299847792998, + "grad_norm": 6.1875, + "learning_rate": 7.5385862495912905e-06, + "loss": 1.69893894, + "memory(GiB)": 111.15, + "step": 28695, + "train_speed(iter/s)": 0.448756 + }, + { + "acc": 0.64756513, + "epoch": 0.7280568239472349, + "grad_norm": 6.5625, + "learning_rate": 7.537682779550537e-06, + "loss": 1.65732765, + "memory(GiB)": 111.15, + "step": 28700, + "train_speed(iter/s)": 0.448797 + }, + { + "epoch": 0.7280568239472349, + "eval_acc": 0.6425060691145882, + "eval_loss": 1.6122620105743408, + "eval_runtime": 114.3034, + "eval_samples_per_second": 55.729, + "eval_steps_per_second": 27.864, + "step": 28700 + }, + { + "acc": 0.63390226, + "epoch": 0.7281836631151699, + "grad_norm": 5.96875, + "learning_rate": 7.536779197891159e-06, + "loss": 1.67917728, + "memory(GiB)": 111.15, + "step": 28705, + "train_speed(iter/s)": 0.448005 + }, + { + "acc": 0.65839176, + "epoch": 0.728310502283105, + "grad_norm": 5.125, + "learning_rate": 7.535875504652912e-06, + "loss": 1.59147072, + "memory(GiB)": 111.15, + "step": 28710, + "train_speed(iter/s)": 0.448046 + }, + { + "acc": 0.64969254, + "epoch": 0.7284373414510401, + "grad_norm": 5.40625, + "learning_rate": 7.534971699875534e-06, + "loss": 1.64550285, + "memory(GiB)": 111.15, + "step": 28715, + "train_speed(iter/s)": 0.448087 + }, + { + "acc": 0.6503583, + "epoch": 0.7285641806189751, + "grad_norm": 5.0625, + "learning_rate": 7.534067783598784e-06, + "loss": 1.57519283, + "memory(GiB)": 111.15, + "step": 28720, + "train_speed(iter/s)": 0.448128 + }, + { + "acc": 0.63005486, + "epoch": 0.7286910197869102, + "grad_norm": 4.9375, + "learning_rate": 7.533163755862419e-06, + "loss": 1.6385788, + "memory(GiB)": 111.15, + "step": 28725, + "train_speed(iter/s)": 0.448169 + }, + { + "acc": 0.64552927, + "epoch": 0.7288178589548453, + "grad_norm": 5.3125, + "learning_rate": 7.5322596167062035e-06, + "loss": 1.63371735, + "memory(GiB)": 111.15, + "step": 28730, + "train_speed(iter/s)": 0.44821 + }, + { + "acc": 0.65613108, + "epoch": 0.7289446981227803, + "grad_norm": 5.25, + "learning_rate": 7.5313553661699035e-06, + "loss": 1.59916019, + "memory(GiB)": 111.15, + "step": 28735, + "train_speed(iter/s)": 0.448251 + }, + { + "acc": 0.66204209, + "epoch": 0.7290715372907154, + "grad_norm": 5.25, + "learning_rate": 7.530451004293292e-06, + "loss": 1.5459074, + "memory(GiB)": 111.15, + "step": 28740, + "train_speed(iter/s)": 0.448291 + }, + { + "acc": 0.64189644, + "epoch": 0.7291983764586504, + "grad_norm": 5.25, + "learning_rate": 7.5295465311161485e-06, + "loss": 1.62375145, + "memory(GiB)": 111.15, + "step": 28745, + "train_speed(iter/s)": 0.448332 + }, + { + "acc": 0.65440636, + "epoch": 0.7293252156265855, + "grad_norm": 5.1875, + "learning_rate": 7.5286419466782546e-06, + "loss": 1.57633038, + "memory(GiB)": 111.15, + "step": 28750, + "train_speed(iter/s)": 0.448374 + }, + { + "acc": 0.65195608, + "epoch": 0.7294520547945206, + "grad_norm": 10.625, + "learning_rate": 7.527737251019399e-06, + "loss": 1.63205013, + "memory(GiB)": 111.15, + "step": 28755, + "train_speed(iter/s)": 0.448415 + }, + { + "acc": 0.63041224, + "epoch": 0.7295788939624556, + "grad_norm": 5.5625, + "learning_rate": 7.526832444179373e-06, + "loss": 1.69073105, + "memory(GiB)": 111.15, + "step": 28760, + "train_speed(iter/s)": 0.448456 + }, + { + "acc": 0.65096483, + "epoch": 0.7297057331303907, + "grad_norm": 5.78125, + "learning_rate": 7.525927526197974e-06, + "loss": 1.64330444, + "memory(GiB)": 111.15, + "step": 28765, + "train_speed(iter/s)": 0.448497 + }, + { + "acc": 0.63903933, + "epoch": 0.7298325722983258, + "grad_norm": 4.84375, + "learning_rate": 7.5250224971150065e-06, + "loss": 1.68053226, + "memory(GiB)": 111.15, + "step": 28770, + "train_speed(iter/s)": 0.448538 + }, + { + "acc": 0.65885854, + "epoch": 0.7299594114662608, + "grad_norm": 4.71875, + "learning_rate": 7.524117356970275e-06, + "loss": 1.61209373, + "memory(GiB)": 111.15, + "step": 28775, + "train_speed(iter/s)": 0.448579 + }, + { + "acc": 0.6475595, + "epoch": 0.7300862506341959, + "grad_norm": 5.53125, + "learning_rate": 7.523212105803594e-06, + "loss": 1.72454376, + "memory(GiB)": 111.15, + "step": 28780, + "train_speed(iter/s)": 0.44862 + }, + { + "acc": 0.65413156, + "epoch": 0.7302130898021308, + "grad_norm": 6.0625, + "learning_rate": 7.522306743654777e-06, + "loss": 1.6006443, + "memory(GiB)": 111.15, + "step": 28785, + "train_speed(iter/s)": 0.448661 + }, + { + "acc": 0.65561514, + "epoch": 0.7303399289700659, + "grad_norm": 5.09375, + "learning_rate": 7.521401270563651e-06, + "loss": 1.64150925, + "memory(GiB)": 111.15, + "step": 28790, + "train_speed(iter/s)": 0.448702 + }, + { + "acc": 0.64167252, + "epoch": 0.730466768138001, + "grad_norm": 5.09375, + "learning_rate": 7.520495686570037e-06, + "loss": 1.61857758, + "memory(GiB)": 111.15, + "step": 28795, + "train_speed(iter/s)": 0.448743 + }, + { + "acc": 0.67071457, + "epoch": 0.730593607305936, + "grad_norm": 5.78125, + "learning_rate": 7.5195899917137716e-06, + "loss": 1.57240067, + "memory(GiB)": 111.15, + "step": 28800, + "train_speed(iter/s)": 0.448784 + }, + { + "epoch": 0.730593607305936, + "eval_acc": 0.6425641387706027, + "eval_loss": 1.6117708683013916, + "eval_runtime": 113.217, + "eval_samples_per_second": 56.264, + "eval_steps_per_second": 28.132, + "step": 28800 + }, + { + "acc": 0.63813877, + "epoch": 0.7307204464738711, + "grad_norm": 6.28125, + "learning_rate": 7.518684186034688e-06, + "loss": 1.6111351, + "memory(GiB)": 111.15, + "step": 28805, + "train_speed(iter/s)": 0.448003 + }, + { + "acc": 0.65908251, + "epoch": 0.7308472856418062, + "grad_norm": 4.21875, + "learning_rate": 7.51777826957263e-06, + "loss": 1.60618057, + "memory(GiB)": 111.15, + "step": 28810, + "train_speed(iter/s)": 0.448043 + }, + { + "acc": 0.64603691, + "epoch": 0.7309741248097412, + "grad_norm": 5.4375, + "learning_rate": 7.516872242367441e-06, + "loss": 1.62057095, + "memory(GiB)": 111.15, + "step": 28815, + "train_speed(iter/s)": 0.448084 + }, + { + "acc": 0.64925823, + "epoch": 0.7311009639776763, + "grad_norm": 5.59375, + "learning_rate": 7.5159661044589745e-06, + "loss": 1.6780262, + "memory(GiB)": 111.15, + "step": 28820, + "train_speed(iter/s)": 0.448124 + }, + { + "acc": 0.65560055, + "epoch": 0.7312278031456113, + "grad_norm": 6.8125, + "learning_rate": 7.515059855887087e-06, + "loss": 1.58768425, + "memory(GiB)": 111.15, + "step": 28825, + "train_speed(iter/s)": 0.448165 + }, + { + "acc": 0.65048513, + "epoch": 0.7313546423135464, + "grad_norm": 6.0625, + "learning_rate": 7.514153496691636e-06, + "loss": 1.5634676, + "memory(GiB)": 111.15, + "step": 28830, + "train_speed(iter/s)": 0.448205 + }, + { + "acc": 0.65686302, + "epoch": 0.7314814814814815, + "grad_norm": 5.125, + "learning_rate": 7.513247026912491e-06, + "loss": 1.52791634, + "memory(GiB)": 111.15, + "step": 28835, + "train_speed(iter/s)": 0.448246 + }, + { + "acc": 0.65599265, + "epoch": 0.7316083206494165, + "grad_norm": 6.1875, + "learning_rate": 7.512340446589521e-06, + "loss": 1.61348724, + "memory(GiB)": 111.15, + "step": 28840, + "train_speed(iter/s)": 0.448287 + }, + { + "acc": 0.66059461, + "epoch": 0.7317351598173516, + "grad_norm": 4.46875, + "learning_rate": 7.5114337557625985e-06, + "loss": 1.58489389, + "memory(GiB)": 111.15, + "step": 28845, + "train_speed(iter/s)": 0.448327 + }, + { + "acc": 0.66765442, + "epoch": 0.7318619989852867, + "grad_norm": 5.59375, + "learning_rate": 7.510526954471611e-06, + "loss": 1.55083761, + "memory(GiB)": 111.15, + "step": 28850, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.64093938, + "epoch": 0.7319888381532217, + "grad_norm": 6.0625, + "learning_rate": 7.509620042756436e-06, + "loss": 1.66986313, + "memory(GiB)": 111.15, + "step": 28855, + "train_speed(iter/s)": 0.448407 + }, + { + "acc": 0.63565588, + "epoch": 0.7321156773211568, + "grad_norm": 5.40625, + "learning_rate": 7.508713020656968e-06, + "loss": 1.70329933, + "memory(GiB)": 111.15, + "step": 28860, + "train_speed(iter/s)": 0.448448 + }, + { + "acc": 0.64402924, + "epoch": 0.7322425164890918, + "grad_norm": 5.40625, + "learning_rate": 7.5078058882131e-06, + "loss": 1.67820892, + "memory(GiB)": 111.15, + "step": 28865, + "train_speed(iter/s)": 0.448488 + }, + { + "acc": 0.64130983, + "epoch": 0.7323693556570269, + "grad_norm": 5.09375, + "learning_rate": 7.506898645464733e-06, + "loss": 1.66814671, + "memory(GiB)": 111.15, + "step": 28870, + "train_speed(iter/s)": 0.448528 + }, + { + "acc": 0.65724506, + "epoch": 0.732496194824962, + "grad_norm": 5.46875, + "learning_rate": 7.505991292451772e-06, + "loss": 1.57002106, + "memory(GiB)": 111.15, + "step": 28875, + "train_speed(iter/s)": 0.448569 + }, + { + "acc": 0.64488659, + "epoch": 0.732623033992897, + "grad_norm": 4.4375, + "learning_rate": 7.505083829214125e-06, + "loss": 1.68583183, + "memory(GiB)": 111.15, + "step": 28880, + "train_speed(iter/s)": 0.44861 + }, + { + "acc": 0.6449666, + "epoch": 0.7327498731608321, + "grad_norm": 4.46875, + "learning_rate": 7.5041762557917065e-06, + "loss": 1.69428749, + "memory(GiB)": 111.15, + "step": 28885, + "train_speed(iter/s)": 0.44865 + }, + { + "acc": 0.646942, + "epoch": 0.7328767123287672, + "grad_norm": 4.8125, + "learning_rate": 7.5032685722244355e-06, + "loss": 1.703228, + "memory(GiB)": 111.15, + "step": 28890, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.63872261, + "epoch": 0.7330035514967022, + "grad_norm": 5.625, + "learning_rate": 7.502360778552238e-06, + "loss": 1.63986874, + "memory(GiB)": 111.15, + "step": 28895, + "train_speed(iter/s)": 0.448731 + }, + { + "acc": 0.63908596, + "epoch": 0.7331303906646373, + "grad_norm": 5.21875, + "learning_rate": 7.5014528748150405e-06, + "loss": 1.65397987, + "memory(GiB)": 111.15, + "step": 28900, + "train_speed(iter/s)": 0.448771 + }, + { + "epoch": 0.7331303906646373, + "eval_acc": 0.642639336886305, + "eval_loss": 1.611669659614563, + "eval_runtime": 114.3263, + "eval_samples_per_second": 55.718, + "eval_steps_per_second": 27.859, + "step": 28900 + }, + { + "acc": 0.64146466, + "epoch": 0.7332572298325722, + "grad_norm": 5.75, + "learning_rate": 7.5005448610527765e-06, + "loss": 1.61881313, + "memory(GiB)": 111.15, + "step": 28905, + "train_speed(iter/s)": 0.447985 + }, + { + "acc": 0.66403112, + "epoch": 0.7333840690005073, + "grad_norm": 5.0625, + "learning_rate": 7.499636737305386e-06, + "loss": 1.54060307, + "memory(GiB)": 111.15, + "step": 28910, + "train_speed(iter/s)": 0.448026 + }, + { + "acc": 0.6755578, + "epoch": 0.7335109081684424, + "grad_norm": 5.8125, + "learning_rate": 7.498728503612811e-06, + "loss": 1.55732183, + "memory(GiB)": 111.15, + "step": 28915, + "train_speed(iter/s)": 0.448066 + }, + { + "acc": 0.65447311, + "epoch": 0.7336377473363774, + "grad_norm": 6.1875, + "learning_rate": 7.497820160015002e-06, + "loss": 1.63697777, + "memory(GiB)": 111.15, + "step": 28920, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.63372879, + "epoch": 0.7337645865043125, + "grad_norm": 4.65625, + "learning_rate": 7.496911706551908e-06, + "loss": 1.66858215, + "memory(GiB)": 111.15, + "step": 28925, + "train_speed(iter/s)": 0.448148 + }, + { + "acc": 0.66220264, + "epoch": 0.7338914256722476, + "grad_norm": 5.59375, + "learning_rate": 7.496003143263492e-06, + "loss": 1.58180332, + "memory(GiB)": 111.15, + "step": 28930, + "train_speed(iter/s)": 0.448189 + }, + { + "acc": 0.64142022, + "epoch": 0.7340182648401826, + "grad_norm": 5.0, + "learning_rate": 7.495094470189712e-06, + "loss": 1.69146061, + "memory(GiB)": 111.15, + "step": 28935, + "train_speed(iter/s)": 0.448229 + }, + { + "acc": 0.63209128, + "epoch": 0.7341451040081177, + "grad_norm": 6.53125, + "learning_rate": 7.4941856873705376e-06, + "loss": 1.75993156, + "memory(GiB)": 111.15, + "step": 28940, + "train_speed(iter/s)": 0.44827 + }, + { + "acc": 0.6440074, + "epoch": 0.7342719431760527, + "grad_norm": 5.0, + "learning_rate": 7.493276794845941e-06, + "loss": 1.63935852, + "memory(GiB)": 111.15, + "step": 28945, + "train_speed(iter/s)": 0.448311 + }, + { + "acc": 0.6530499, + "epoch": 0.7343987823439878, + "grad_norm": 5.625, + "learning_rate": 7.4923677926559005e-06, + "loss": 1.56926403, + "memory(GiB)": 111.15, + "step": 28950, + "train_speed(iter/s)": 0.448352 + }, + { + "acc": 0.65737772, + "epoch": 0.7345256215119229, + "grad_norm": 5.15625, + "learning_rate": 7.491458680840396e-06, + "loss": 1.61017513, + "memory(GiB)": 111.15, + "step": 28955, + "train_speed(iter/s)": 0.448393 + }, + { + "acc": 0.64373407, + "epoch": 0.7346524606798579, + "grad_norm": 5.25, + "learning_rate": 7.490549459439415e-06, + "loss": 1.63315201, + "memory(GiB)": 111.15, + "step": 28960, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.64362264, + "epoch": 0.734779299847793, + "grad_norm": 4.125, + "learning_rate": 7.48964012849295e-06, + "loss": 1.62404308, + "memory(GiB)": 111.15, + "step": 28965, + "train_speed(iter/s)": 0.448475 + }, + { + "acc": 0.64284124, + "epoch": 0.7349061390157281, + "grad_norm": 5.53125, + "learning_rate": 7.488730688040995e-06, + "loss": 1.6445118, + "memory(GiB)": 111.15, + "step": 28970, + "train_speed(iter/s)": 0.448515 + }, + { + "acc": 0.64734693, + "epoch": 0.7350329781836631, + "grad_norm": 5.875, + "learning_rate": 7.487821138123554e-06, + "loss": 1.68403282, + "memory(GiB)": 111.15, + "step": 28975, + "train_speed(iter/s)": 0.448556 + }, + { + "acc": 0.67993135, + "epoch": 0.7351598173515982, + "grad_norm": 5.3125, + "learning_rate": 7.486911478780633e-06, + "loss": 1.53025417, + "memory(GiB)": 111.15, + "step": 28980, + "train_speed(iter/s)": 0.448597 + }, + { + "acc": 0.65746927, + "epoch": 0.7352866565195332, + "grad_norm": 5.09375, + "learning_rate": 7.4860017100522395e-06, + "loss": 1.59587278, + "memory(GiB)": 111.15, + "step": 28985, + "train_speed(iter/s)": 0.448637 + }, + { + "acc": 0.63873568, + "epoch": 0.7354134956874683, + "grad_norm": 6.4375, + "learning_rate": 7.485091831978394e-06, + "loss": 1.66829185, + "memory(GiB)": 111.15, + "step": 28990, + "train_speed(iter/s)": 0.448678 + }, + { + "acc": 0.65307674, + "epoch": 0.7355403348554034, + "grad_norm": 5.03125, + "learning_rate": 7.484181844599113e-06, + "loss": 1.67677422, + "memory(GiB)": 111.15, + "step": 28995, + "train_speed(iter/s)": 0.448719 + }, + { + "acc": 0.65499344, + "epoch": 0.7356671740233384, + "grad_norm": 5.71875, + "learning_rate": 7.483271747954425e-06, + "loss": 1.57612371, + "memory(GiB)": 111.15, + "step": 29000, + "train_speed(iter/s)": 0.44876 + }, + { + "epoch": 0.7356671740233384, + "eval_acc": 0.6425340595243217, + "eval_loss": 1.6114765405654907, + "eval_runtime": 114.3653, + "eval_samples_per_second": 55.699, + "eval_steps_per_second": 27.849, + "step": 29000 + }, + { + "acc": 0.64585381, + "epoch": 0.7357940131912735, + "grad_norm": 5.9375, + "learning_rate": 7.482361542084356e-06, + "loss": 1.62298241, + "memory(GiB)": 111.15, + "step": 29005, + "train_speed(iter/s)": 0.447976 + }, + { + "acc": 0.64721518, + "epoch": 0.7359208523592086, + "grad_norm": 4.34375, + "learning_rate": 7.481451227028946e-06, + "loss": 1.5576539, + "memory(GiB)": 111.15, + "step": 29010, + "train_speed(iter/s)": 0.448016 + }, + { + "acc": 0.66630187, + "epoch": 0.7360476915271436, + "grad_norm": 5.9375, + "learning_rate": 7.4805408028282316e-06, + "loss": 1.57110777, + "memory(GiB)": 111.15, + "step": 29015, + "train_speed(iter/s)": 0.448057 + }, + { + "acc": 0.65921688, + "epoch": 0.7361745306950787, + "grad_norm": 4.65625, + "learning_rate": 7.479630269522257e-06, + "loss": 1.60866165, + "memory(GiB)": 111.15, + "step": 29020, + "train_speed(iter/s)": 0.448097 + }, + { + "acc": 0.63960524, + "epoch": 0.7363013698630136, + "grad_norm": 5.59375, + "learning_rate": 7.478719627151073e-06, + "loss": 1.74650211, + "memory(GiB)": 111.15, + "step": 29025, + "train_speed(iter/s)": 0.448138 + }, + { + "acc": 0.64255228, + "epoch": 0.7364282090309487, + "grad_norm": 5.90625, + "learning_rate": 7.4778088757547325e-06, + "loss": 1.66784306, + "memory(GiB)": 111.15, + "step": 29030, + "train_speed(iter/s)": 0.448178 + }, + { + "acc": 0.65021229, + "epoch": 0.7365550481988838, + "grad_norm": 5.78125, + "learning_rate": 7.476898015373296e-06, + "loss": 1.67145271, + "memory(GiB)": 111.15, + "step": 29035, + "train_speed(iter/s)": 0.448218 + }, + { + "acc": 0.65701065, + "epoch": 0.7366818873668188, + "grad_norm": 5.125, + "learning_rate": 7.4759870460468256e-06, + "loss": 1.57578926, + "memory(GiB)": 111.15, + "step": 29040, + "train_speed(iter/s)": 0.448259 + }, + { + "acc": 0.64523315, + "epoch": 0.7368087265347539, + "grad_norm": 6.15625, + "learning_rate": 7.475075967815391e-06, + "loss": 1.61252975, + "memory(GiB)": 111.15, + "step": 29045, + "train_speed(iter/s)": 0.448299 + }, + { + "acc": 0.64708314, + "epoch": 0.736935565702689, + "grad_norm": 5.125, + "learning_rate": 7.474164780719064e-06, + "loss": 1.61332169, + "memory(GiB)": 111.15, + "step": 29050, + "train_speed(iter/s)": 0.44834 + }, + { + "acc": 0.64389343, + "epoch": 0.737062404870624, + "grad_norm": 6.03125, + "learning_rate": 7.473253484797924e-06, + "loss": 1.64670906, + "memory(GiB)": 111.15, + "step": 29055, + "train_speed(iter/s)": 0.44838 + }, + { + "acc": 0.66633196, + "epoch": 0.7371892440385591, + "grad_norm": 6.375, + "learning_rate": 7.4723420800920545e-06, + "loss": 1.59822044, + "memory(GiB)": 111.15, + "step": 29060, + "train_speed(iter/s)": 0.448421 + }, + { + "acc": 0.65356307, + "epoch": 0.7373160832064941, + "grad_norm": 5.125, + "learning_rate": 7.47143056664154e-06, + "loss": 1.60590496, + "memory(GiB)": 111.15, + "step": 29065, + "train_speed(iter/s)": 0.448462 + }, + { + "acc": 0.65235147, + "epoch": 0.7374429223744292, + "grad_norm": 5.5625, + "learning_rate": 7.470518944486476e-06, + "loss": 1.61864319, + "memory(GiB)": 111.15, + "step": 29070, + "train_speed(iter/s)": 0.448502 + }, + { + "acc": 0.64632368, + "epoch": 0.7375697615423643, + "grad_norm": 5.5625, + "learning_rate": 7.469607213666958e-06, + "loss": 1.6498661, + "memory(GiB)": 111.15, + "step": 29075, + "train_speed(iter/s)": 0.448543 + }, + { + "acc": 0.64858084, + "epoch": 0.7376966007102993, + "grad_norm": 5.875, + "learning_rate": 7.468695374223092e-06, + "loss": 1.61862907, + "memory(GiB)": 111.15, + "step": 29080, + "train_speed(iter/s)": 0.448583 + }, + { + "acc": 0.65040374, + "epoch": 0.7378234398782344, + "grad_norm": 4.6875, + "learning_rate": 7.4677834261949765e-06, + "loss": 1.63179932, + "memory(GiB)": 111.15, + "step": 29085, + "train_speed(iter/s)": 0.448624 + }, + { + "acc": 0.66974878, + "epoch": 0.7379502790461695, + "grad_norm": 5.5, + "learning_rate": 7.466871369622731e-06, + "loss": 1.53445778, + "memory(GiB)": 111.15, + "step": 29090, + "train_speed(iter/s)": 0.448665 + }, + { + "acc": 0.64431472, + "epoch": 0.7380771182141045, + "grad_norm": 5.9375, + "learning_rate": 7.465959204546469e-06, + "loss": 1.66033211, + "memory(GiB)": 111.15, + "step": 29095, + "train_speed(iter/s)": 0.448705 + }, + { + "acc": 0.64191065, + "epoch": 0.7382039573820396, + "grad_norm": 5.65625, + "learning_rate": 7.465046931006311e-06, + "loss": 1.66242256, + "memory(GiB)": 111.15, + "step": 29100, + "train_speed(iter/s)": 0.448746 + }, + { + "epoch": 0.7382039573820396, + "eval_acc": 0.6426297282381875, + "eval_loss": 1.6115998029708862, + "eval_runtime": 114.1449, + "eval_samples_per_second": 55.806, + "eval_steps_per_second": 27.903, + "step": 29100 + }, + { + "acc": 0.65629659, + "epoch": 0.7383307965499746, + "grad_norm": 5.28125, + "learning_rate": 7.464134549042383e-06, + "loss": 1.57267895, + "memory(GiB)": 111.15, + "step": 29105, + "train_speed(iter/s)": 0.447966 + }, + { + "acc": 0.65339622, + "epoch": 0.7384576357179097, + "grad_norm": 4.9375, + "learning_rate": 7.463222058694817e-06, + "loss": 1.64722328, + "memory(GiB)": 111.15, + "step": 29110, + "train_speed(iter/s)": 0.448006 + }, + { + "acc": 0.64759622, + "epoch": 0.7385844748858448, + "grad_norm": 4.6875, + "learning_rate": 7.462309460003747e-06, + "loss": 1.65577736, + "memory(GiB)": 111.15, + "step": 29115, + "train_speed(iter/s)": 0.448046 + }, + { + "acc": 0.64903812, + "epoch": 0.7387113140537798, + "grad_norm": 6.46875, + "learning_rate": 7.461396753009314e-06, + "loss": 1.64622383, + "memory(GiB)": 111.15, + "step": 29120, + "train_speed(iter/s)": 0.448087 + }, + { + "acc": 0.66136274, + "epoch": 0.7388381532217149, + "grad_norm": 5.03125, + "learning_rate": 7.460483937751662e-06, + "loss": 1.5687767, + "memory(GiB)": 111.15, + "step": 29125, + "train_speed(iter/s)": 0.448127 + }, + { + "acc": 0.64324884, + "epoch": 0.73896499238965, + "grad_norm": 4.90625, + "learning_rate": 7.45957101427094e-06, + "loss": 1.57087421, + "memory(GiB)": 111.15, + "step": 29130, + "train_speed(iter/s)": 0.448167 + }, + { + "acc": 0.64440498, + "epoch": 0.739091831557585, + "grad_norm": 4.90625, + "learning_rate": 7.458657982607303e-06, + "loss": 1.64942818, + "memory(GiB)": 111.15, + "step": 29135, + "train_speed(iter/s)": 0.448208 + }, + { + "acc": 0.64200516, + "epoch": 0.73921867072552, + "grad_norm": 6.1875, + "learning_rate": 7.457744842800913e-06, + "loss": 1.66835403, + "memory(GiB)": 111.15, + "step": 29140, + "train_speed(iter/s)": 0.448248 + }, + { + "acc": 0.63269348, + "epoch": 0.739345509893455, + "grad_norm": 5.5, + "learning_rate": 7.45683159489193e-06, + "loss": 1.71690826, + "memory(GiB)": 111.15, + "step": 29145, + "train_speed(iter/s)": 0.448288 + }, + { + "acc": 0.64072227, + "epoch": 0.7394723490613901, + "grad_norm": 5.90625, + "learning_rate": 7.455918238920526e-06, + "loss": 1.63409367, + "memory(GiB)": 111.15, + "step": 29150, + "train_speed(iter/s)": 0.448329 + }, + { + "acc": 0.64474344, + "epoch": 0.7395991882293252, + "grad_norm": 5.40625, + "learning_rate": 7.455004774926873e-06, + "loss": 1.65958672, + "memory(GiB)": 111.15, + "step": 29155, + "train_speed(iter/s)": 0.448369 + }, + { + "acc": 0.65025768, + "epoch": 0.7397260273972602, + "grad_norm": 7.0, + "learning_rate": 7.454091202951148e-06, + "loss": 1.60298309, + "memory(GiB)": 111.15, + "step": 29160, + "train_speed(iter/s)": 0.44841 + }, + { + "acc": 0.64391394, + "epoch": 0.7398528665651953, + "grad_norm": 5.875, + "learning_rate": 7.453177523033536e-06, + "loss": 1.62446022, + "memory(GiB)": 111.15, + "step": 29165, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.64198828, + "epoch": 0.7399797057331304, + "grad_norm": 6.34375, + "learning_rate": 7.452263735214223e-06, + "loss": 1.64020386, + "memory(GiB)": 111.15, + "step": 29170, + "train_speed(iter/s)": 0.448491 + }, + { + "acc": 0.65269241, + "epoch": 0.7401065449010654, + "grad_norm": 4.90625, + "learning_rate": 7.451349839533404e-06, + "loss": 1.65752068, + "memory(GiB)": 111.15, + "step": 29175, + "train_speed(iter/s)": 0.448531 + }, + { + "acc": 0.64224358, + "epoch": 0.7402333840690005, + "grad_norm": 5.375, + "learning_rate": 7.450435836031273e-06, + "loss": 1.59026737, + "memory(GiB)": 111.15, + "step": 29180, + "train_speed(iter/s)": 0.448572 + }, + { + "acc": 0.65218239, + "epoch": 0.7403602232369355, + "grad_norm": 5.21875, + "learning_rate": 7.449521724748034e-06, + "loss": 1.59969654, + "memory(GiB)": 111.15, + "step": 29185, + "train_speed(iter/s)": 0.448612 + }, + { + "acc": 0.6520628, + "epoch": 0.7404870624048706, + "grad_norm": 5.75, + "learning_rate": 7.4486075057238936e-06, + "loss": 1.57663403, + "memory(GiB)": 111.15, + "step": 29190, + "train_speed(iter/s)": 0.448653 + }, + { + "acc": 0.65496516, + "epoch": 0.7406139015728057, + "grad_norm": 5.40625, + "learning_rate": 7.447693178999062e-06, + "loss": 1.65670128, + "memory(GiB)": 111.15, + "step": 29195, + "train_speed(iter/s)": 0.448693 + }, + { + "acc": 0.64273338, + "epoch": 0.7407407407407407, + "grad_norm": 5.75, + "learning_rate": 7.446778744613759e-06, + "loss": 1.64350891, + "memory(GiB)": 111.15, + "step": 29200, + "train_speed(iter/s)": 0.448733 + }, + { + "epoch": 0.7407407407407407, + "eval_acc": 0.6426380835843767, + "eval_loss": 1.6114305257797241, + "eval_runtime": 113.0035, + "eval_samples_per_second": 56.37, + "eval_steps_per_second": 28.185, + "step": 29200 + }, + { + "acc": 0.66264858, + "epoch": 0.7408675799086758, + "grad_norm": 5.125, + "learning_rate": 7.445864202608198e-06, + "loss": 1.56495304, + "memory(GiB)": 111.15, + "step": 29205, + "train_speed(iter/s)": 0.447964 + }, + { + "acc": 0.6416831, + "epoch": 0.7409944190766109, + "grad_norm": 5.90625, + "learning_rate": 7.444949553022613e-06, + "loss": 1.61845398, + "memory(GiB)": 111.15, + "step": 29210, + "train_speed(iter/s)": 0.448004 + }, + { + "acc": 0.65138731, + "epoch": 0.7411212582445459, + "grad_norm": 5.96875, + "learning_rate": 7.444034795897229e-06, + "loss": 1.59268589, + "memory(GiB)": 111.15, + "step": 29215, + "train_speed(iter/s)": 0.448044 + }, + { + "acc": 0.66728354, + "epoch": 0.741248097412481, + "grad_norm": 5.25, + "learning_rate": 7.443119931272285e-06, + "loss": 1.59935894, + "memory(GiB)": 111.15, + "step": 29220, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.6589582, + "epoch": 0.741374936580416, + "grad_norm": 5.53125, + "learning_rate": 7.442204959188016e-06, + "loss": 1.60165405, + "memory(GiB)": 111.15, + "step": 29225, + "train_speed(iter/s)": 0.448125 + }, + { + "acc": 0.6432857, + "epoch": 0.7415017757483511, + "grad_norm": 5.21875, + "learning_rate": 7.4412898796846724e-06, + "loss": 1.65292816, + "memory(GiB)": 111.15, + "step": 29230, + "train_speed(iter/s)": 0.448165 + }, + { + "acc": 0.6655982, + "epoch": 0.7416286149162862, + "grad_norm": 5.03125, + "learning_rate": 7.440374692802497e-06, + "loss": 1.58186235, + "memory(GiB)": 111.15, + "step": 29235, + "train_speed(iter/s)": 0.448206 + }, + { + "acc": 0.66162276, + "epoch": 0.7417554540842212, + "grad_norm": 5.96875, + "learning_rate": 7.439459398581747e-06, + "loss": 1.55382938, + "memory(GiB)": 111.15, + "step": 29240, + "train_speed(iter/s)": 0.448246 + }, + { + "acc": 0.65515318, + "epoch": 0.7418822932521563, + "grad_norm": 6.71875, + "learning_rate": 7.438543997062684e-06, + "loss": 1.53960314, + "memory(GiB)": 111.15, + "step": 29245, + "train_speed(iter/s)": 0.448286 + }, + { + "acc": 0.63221388, + "epoch": 0.7420091324200914, + "grad_norm": 5.25, + "learning_rate": 7.437628488285568e-06, + "loss": 1.61046371, + "memory(GiB)": 111.15, + "step": 29250, + "train_speed(iter/s)": 0.448327 + }, + { + "acc": 0.64691267, + "epoch": 0.7421359715880264, + "grad_norm": 4.65625, + "learning_rate": 7.4367128722906665e-06, + "loss": 1.60231438, + "memory(GiB)": 111.15, + "step": 29255, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.64576454, + "epoch": 0.7422628107559615, + "grad_norm": 5.40625, + "learning_rate": 7.435797149118255e-06, + "loss": 1.64953461, + "memory(GiB)": 111.15, + "step": 29260, + "train_speed(iter/s)": 0.448407 + }, + { + "acc": 0.64568243, + "epoch": 0.7423896499238964, + "grad_norm": 5.15625, + "learning_rate": 7.434881318808609e-06, + "loss": 1.63254986, + "memory(GiB)": 111.15, + "step": 29265, + "train_speed(iter/s)": 0.448448 + }, + { + "acc": 0.64882431, + "epoch": 0.7425164890918315, + "grad_norm": 4.96875, + "learning_rate": 7.433965381402013e-06, + "loss": 1.59137335, + "memory(GiB)": 111.15, + "step": 29270, + "train_speed(iter/s)": 0.448489 + }, + { + "acc": 0.65075922, + "epoch": 0.7426433282597666, + "grad_norm": 6.0625, + "learning_rate": 7.4330493369387514e-06, + "loss": 1.59557304, + "memory(GiB)": 111.15, + "step": 29275, + "train_speed(iter/s)": 0.448529 + }, + { + "acc": 0.63262119, + "epoch": 0.7427701674277016, + "grad_norm": 4.96875, + "learning_rate": 7.432133185459117e-06, + "loss": 1.60346889, + "memory(GiB)": 111.15, + "step": 29280, + "train_speed(iter/s)": 0.448569 + }, + { + "acc": 0.64509044, + "epoch": 0.7428970065956367, + "grad_norm": 5.53125, + "learning_rate": 7.431216927003406e-06, + "loss": 1.59330788, + "memory(GiB)": 111.15, + "step": 29285, + "train_speed(iter/s)": 0.44861 + }, + { + "acc": 0.65697455, + "epoch": 0.7430238457635718, + "grad_norm": 5.84375, + "learning_rate": 7.430300561611922e-06, + "loss": 1.56828794, + "memory(GiB)": 111.15, + "step": 29290, + "train_speed(iter/s)": 0.44865 + }, + { + "acc": 0.65387545, + "epoch": 0.7431506849315068, + "grad_norm": 4.84375, + "learning_rate": 7.429384089324967e-06, + "loss": 1.55713673, + "memory(GiB)": 111.15, + "step": 29295, + "train_speed(iter/s)": 0.44869 + }, + { + "acc": 0.64134035, + "epoch": 0.7432775240994419, + "grad_norm": 5.34375, + "learning_rate": 7.428467510182854e-06, + "loss": 1.63156509, + "memory(GiB)": 111.15, + "step": 29300, + "train_speed(iter/s)": 0.448731 + }, + { + "epoch": 0.7432775240994419, + "eval_acc": 0.6426790247807035, + "eval_loss": 1.6120020151138306, + "eval_runtime": 112.1652, + "eval_samples_per_second": 56.791, + "eval_steps_per_second": 28.396, + "step": 29300 + }, + { + "acc": 0.66511407, + "epoch": 0.7434043632673769, + "grad_norm": 5.3125, + "learning_rate": 7.427550824225896e-06, + "loss": 1.53677807, + "memory(GiB)": 111.15, + "step": 29305, + "train_speed(iter/s)": 0.44797 + }, + { + "acc": 0.65208721, + "epoch": 0.743531202435312, + "grad_norm": 6.125, + "learning_rate": 7.426634031494417e-06, + "loss": 1.64579105, + "memory(GiB)": 111.15, + "step": 29310, + "train_speed(iter/s)": 0.44801 + }, + { + "acc": 0.66410961, + "epoch": 0.7436580416032471, + "grad_norm": 5.625, + "learning_rate": 7.425717132028738e-06, + "loss": 1.54909515, + "memory(GiB)": 111.15, + "step": 29315, + "train_speed(iter/s)": 0.44805 + }, + { + "acc": 0.62959862, + "epoch": 0.7437848807711821, + "grad_norm": 5.125, + "learning_rate": 7.42480012586919e-06, + "loss": 1.70022736, + "memory(GiB)": 111.15, + "step": 29320, + "train_speed(iter/s)": 0.44809 + }, + { + "acc": 0.64820552, + "epoch": 0.7439117199391172, + "grad_norm": 5.59375, + "learning_rate": 7.423883013056106e-06, + "loss": 1.5664731, + "memory(GiB)": 111.15, + "step": 29325, + "train_speed(iter/s)": 0.44813 + }, + { + "acc": 0.64884439, + "epoch": 0.7440385591070523, + "grad_norm": 5.6875, + "learning_rate": 7.422965793629825e-06, + "loss": 1.60878658, + "memory(GiB)": 111.15, + "step": 29330, + "train_speed(iter/s)": 0.448169 + }, + { + "acc": 0.64543371, + "epoch": 0.7441653982749873, + "grad_norm": 6.5, + "learning_rate": 7.422048467630691e-06, + "loss": 1.67448158, + "memory(GiB)": 111.15, + "step": 29335, + "train_speed(iter/s)": 0.448209 + }, + { + "acc": 0.65076547, + "epoch": 0.7442922374429224, + "grad_norm": 5.21875, + "learning_rate": 7.421131035099052e-06, + "loss": 1.55163946, + "memory(GiB)": 111.15, + "step": 29340, + "train_speed(iter/s)": 0.448249 + }, + { + "acc": 0.65343723, + "epoch": 0.7444190766108574, + "grad_norm": 5.28125, + "learning_rate": 7.42021349607526e-06, + "loss": 1.59713478, + "memory(GiB)": 111.15, + "step": 29345, + "train_speed(iter/s)": 0.448289 + }, + { + "acc": 0.66814728, + "epoch": 0.7445459157787925, + "grad_norm": 4.375, + "learning_rate": 7.419295850599673e-06, + "loss": 1.49626923, + "memory(GiB)": 111.15, + "step": 29350, + "train_speed(iter/s)": 0.448329 + }, + { + "acc": 0.65309067, + "epoch": 0.7446727549467276, + "grad_norm": 5.9375, + "learning_rate": 7.418378098712653e-06, + "loss": 1.6198103, + "memory(GiB)": 111.15, + "step": 29355, + "train_speed(iter/s)": 0.448369 + }, + { + "acc": 0.63620911, + "epoch": 0.7447995941146626, + "grad_norm": 4.9375, + "learning_rate": 7.417460240454568e-06, + "loss": 1.75131035, + "memory(GiB)": 111.15, + "step": 29360, + "train_speed(iter/s)": 0.448409 + }, + { + "acc": 0.66232438, + "epoch": 0.7449264332825977, + "grad_norm": 5.375, + "learning_rate": 7.4165422758657865e-06, + "loss": 1.60025024, + "memory(GiB)": 111.15, + "step": 29365, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.65401545, + "epoch": 0.7450532724505328, + "grad_norm": 5.5, + "learning_rate": 7.415624204986689e-06, + "loss": 1.58204355, + "memory(GiB)": 111.15, + "step": 29370, + "train_speed(iter/s)": 0.44849 + }, + { + "acc": 0.64803619, + "epoch": 0.7451801116184678, + "grad_norm": 4.84375, + "learning_rate": 7.4147060278576525e-06, + "loss": 1.63326607, + "memory(GiB)": 111.15, + "step": 29375, + "train_speed(iter/s)": 0.44853 + }, + { + "acc": 0.65079765, + "epoch": 0.7453069507864029, + "grad_norm": 5.78125, + "learning_rate": 7.413787744519064e-06, + "loss": 1.61018982, + "memory(GiB)": 111.15, + "step": 29380, + "train_speed(iter/s)": 0.44857 + }, + { + "acc": 0.64109211, + "epoch": 0.7454337899543378, + "grad_norm": 5.15625, + "learning_rate": 7.412869355011314e-06, + "loss": 1.61877098, + "memory(GiB)": 111.15, + "step": 29385, + "train_speed(iter/s)": 0.44861 + }, + { + "acc": 0.64149733, + "epoch": 0.7455606291222729, + "grad_norm": 5.34375, + "learning_rate": 7.411950859374797e-06, + "loss": 1.67174187, + "memory(GiB)": 111.15, + "step": 29390, + "train_speed(iter/s)": 0.44865 + }, + { + "acc": 0.66227632, + "epoch": 0.745687468290208, + "grad_norm": 5.8125, + "learning_rate": 7.411032257649913e-06, + "loss": 1.65630989, + "memory(GiB)": 111.15, + "step": 29395, + "train_speed(iter/s)": 0.448691 + }, + { + "acc": 0.65377064, + "epoch": 0.745814307458143, + "grad_norm": 5.28125, + "learning_rate": 7.410113549877065e-06, + "loss": 1.59004078, + "memory(GiB)": 111.15, + "step": 29400, + "train_speed(iter/s)": 0.44873 + }, + { + "epoch": 0.745814307458143, + "eval_acc": 0.6426802780826318, + "eval_loss": 1.6115682125091553, + "eval_runtime": 113.8122, + "eval_samples_per_second": 55.969, + "eval_steps_per_second": 27.985, + "step": 29400 + }, + { + "acc": 0.63829079, + "epoch": 0.7459411466260781, + "grad_norm": 4.6875, + "learning_rate": 7.409194736096663e-06, + "loss": 1.63111629, + "memory(GiB)": 111.15, + "step": 29405, + "train_speed(iter/s)": 0.447962 + }, + { + "acc": 0.66006489, + "epoch": 0.7460679857940132, + "grad_norm": 5.5625, + "learning_rate": 7.408275816349121e-06, + "loss": 1.60469952, + "memory(GiB)": 111.15, + "step": 29410, + "train_speed(iter/s)": 0.448001 + }, + { + "acc": 0.64852786, + "epoch": 0.7461948249619482, + "grad_norm": 4.875, + "learning_rate": 7.4073567906748555e-06, + "loss": 1.60312099, + "memory(GiB)": 111.15, + "step": 29415, + "train_speed(iter/s)": 0.448041 + }, + { + "acc": 0.65310087, + "epoch": 0.7463216641298833, + "grad_norm": 4.9375, + "learning_rate": 7.406437659114291e-06, + "loss": 1.59351635, + "memory(GiB)": 111.15, + "step": 29420, + "train_speed(iter/s)": 0.448081 + }, + { + "acc": 0.64683132, + "epoch": 0.7464485032978183, + "grad_norm": 5.3125, + "learning_rate": 7.405518421707854e-06, + "loss": 1.65713463, + "memory(GiB)": 111.15, + "step": 29425, + "train_speed(iter/s)": 0.44812 + }, + { + "acc": 0.6504765, + "epoch": 0.7465753424657534, + "grad_norm": 5.4375, + "learning_rate": 7.404599078495977e-06, + "loss": 1.56935759, + "memory(GiB)": 111.15, + "step": 29430, + "train_speed(iter/s)": 0.44816 + }, + { + "acc": 0.65037112, + "epoch": 0.7467021816336885, + "grad_norm": 4.53125, + "learning_rate": 7.403679629519096e-06, + "loss": 1.66003571, + "memory(GiB)": 111.15, + "step": 29435, + "train_speed(iter/s)": 0.4482 + }, + { + "acc": 0.63711114, + "epoch": 0.7468290208016235, + "grad_norm": 5.1875, + "learning_rate": 7.402760074817654e-06, + "loss": 1.74275894, + "memory(GiB)": 111.15, + "step": 29440, + "train_speed(iter/s)": 0.44824 + }, + { + "acc": 0.67355967, + "epoch": 0.7469558599695586, + "grad_norm": 5.09375, + "learning_rate": 7.4018404144320955e-06, + "loss": 1.58881054, + "memory(GiB)": 111.15, + "step": 29445, + "train_speed(iter/s)": 0.44828 + }, + { + "acc": 0.65925965, + "epoch": 0.7470826991374937, + "grad_norm": 5.8125, + "learning_rate": 7.4009206484028735e-06, + "loss": 1.61961365, + "memory(GiB)": 111.15, + "step": 29450, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.65247531, + "epoch": 0.7472095383054287, + "grad_norm": 5.5, + "learning_rate": 7.400000776770441e-06, + "loss": 1.59361467, + "memory(GiB)": 111.15, + "step": 29455, + "train_speed(iter/s)": 0.44836 + }, + { + "acc": 0.64123402, + "epoch": 0.7473363774733638, + "grad_norm": 6.0625, + "learning_rate": 7.39908079957526e-06, + "loss": 1.63577023, + "memory(GiB)": 111.15, + "step": 29460, + "train_speed(iter/s)": 0.4484 + }, + { + "acc": 0.64594517, + "epoch": 0.7474632166412988, + "grad_norm": 7.15625, + "learning_rate": 7.398160716857794e-06, + "loss": 1.62455482, + "memory(GiB)": 111.15, + "step": 29465, + "train_speed(iter/s)": 0.448439 + }, + { + "acc": 0.65376158, + "epoch": 0.7475900558092339, + "grad_norm": 5.28125, + "learning_rate": 7.397240528658513e-06, + "loss": 1.59006748, + "memory(GiB)": 111.15, + "step": 29470, + "train_speed(iter/s)": 0.448479 + }, + { + "acc": 0.6520299, + "epoch": 0.747716894977169, + "grad_norm": 5.3125, + "learning_rate": 7.39632023501789e-06, + "loss": 1.62408829, + "memory(GiB)": 111.15, + "step": 29475, + "train_speed(iter/s)": 0.448519 + }, + { + "acc": 0.63823261, + "epoch": 0.747843734145104, + "grad_norm": 4.90625, + "learning_rate": 7.3953998359764036e-06, + "loss": 1.69240799, + "memory(GiB)": 111.15, + "step": 29480, + "train_speed(iter/s)": 0.448559 + }, + { + "acc": 0.64778252, + "epoch": 0.7479705733130391, + "grad_norm": 5.96875, + "learning_rate": 7.394479331574539e-06, + "loss": 1.64523525, + "memory(GiB)": 111.15, + "step": 29485, + "train_speed(iter/s)": 0.448599 + }, + { + "acc": 0.63866835, + "epoch": 0.7480974124809742, + "grad_norm": 4.875, + "learning_rate": 7.393558721852783e-06, + "loss": 1.69294891, + "memory(GiB)": 111.15, + "step": 29490, + "train_speed(iter/s)": 0.448639 + }, + { + "acc": 0.63603315, + "epoch": 0.7482242516489092, + "grad_norm": 5.75, + "learning_rate": 7.392638006851627e-06, + "loss": 1.6570961, + "memory(GiB)": 111.15, + "step": 29495, + "train_speed(iter/s)": 0.448678 + }, + { + "acc": 0.66273909, + "epoch": 0.7483510908168443, + "grad_norm": 5.40625, + "learning_rate": 7.391717186611569e-06, + "loss": 1.60712452, + "memory(GiB)": 111.15, + "step": 29500, + "train_speed(iter/s)": 0.448718 + }, + { + "epoch": 0.7483510908168443, + "eval_acc": 0.6425424148705109, + "eval_loss": 1.6116019487380981, + "eval_runtime": 114.9076, + "eval_samples_per_second": 55.436, + "eval_steps_per_second": 27.718, + "step": 29500 + }, + { + "acc": 0.65913205, + "epoch": 0.7484779299847792, + "grad_norm": 5.4375, + "learning_rate": 7.39079626117311e-06, + "loss": 1.58531246, + "memory(GiB)": 111.15, + "step": 29505, + "train_speed(iter/s)": 0.447944 + }, + { + "acc": 0.63766146, + "epoch": 0.7486047691527143, + "grad_norm": 5.65625, + "learning_rate": 7.3898752305767595e-06, + "loss": 1.6749897, + "memory(GiB)": 111.15, + "step": 29510, + "train_speed(iter/s)": 0.447984 + }, + { + "acc": 0.63315835, + "epoch": 0.7487316083206494, + "grad_norm": 4.96875, + "learning_rate": 7.3889540948630245e-06, + "loss": 1.65557556, + "memory(GiB)": 111.15, + "step": 29515, + "train_speed(iter/s)": 0.448023 + }, + { + "acc": 0.64826269, + "epoch": 0.7488584474885844, + "grad_norm": 6.25, + "learning_rate": 7.388032854072424e-06, + "loss": 1.60160694, + "memory(GiB)": 111.15, + "step": 29520, + "train_speed(iter/s)": 0.448063 + }, + { + "acc": 0.6403389, + "epoch": 0.7489852866565195, + "grad_norm": 5.59375, + "learning_rate": 7.387111508245476e-06, + "loss": 1.67040653, + "memory(GiB)": 111.15, + "step": 29525, + "train_speed(iter/s)": 0.448103 + }, + { + "acc": 0.6607872, + "epoch": 0.7491121258244546, + "grad_norm": 5.3125, + "learning_rate": 7.386190057422706e-06, + "loss": 1.51414986, + "memory(GiB)": 111.15, + "step": 29530, + "train_speed(iter/s)": 0.448142 + }, + { + "acc": 0.64716215, + "epoch": 0.7492389649923896, + "grad_norm": 5.0, + "learning_rate": 7.385268501644645e-06, + "loss": 1.61074123, + "memory(GiB)": 111.15, + "step": 29535, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.65388522, + "epoch": 0.7493658041603247, + "grad_norm": 5.21875, + "learning_rate": 7.384346840951824e-06, + "loss": 1.66741581, + "memory(GiB)": 111.15, + "step": 29540, + "train_speed(iter/s)": 0.448221 + }, + { + "acc": 0.6549139, + "epoch": 0.7494926433282597, + "grad_norm": 5.8125, + "learning_rate": 7.383425075384785e-06, + "loss": 1.57834387, + "memory(GiB)": 111.15, + "step": 29545, + "train_speed(iter/s)": 0.448261 + }, + { + "acc": 0.64871311, + "epoch": 0.7496194824961948, + "grad_norm": 5.21875, + "learning_rate": 7.382503204984069e-06, + "loss": 1.58654814, + "memory(GiB)": 111.15, + "step": 29550, + "train_speed(iter/s)": 0.4483 + }, + { + "acc": 0.64162226, + "epoch": 0.7497463216641299, + "grad_norm": 5.34375, + "learning_rate": 7.381581229790226e-06, + "loss": 1.58373375, + "memory(GiB)": 111.15, + "step": 29555, + "train_speed(iter/s)": 0.44834 + }, + { + "acc": 0.65455599, + "epoch": 0.7498731608320649, + "grad_norm": 5.59375, + "learning_rate": 7.380659149843806e-06, + "loss": 1.58460407, + "memory(GiB)": 111.15, + "step": 29560, + "train_speed(iter/s)": 0.44838 + }, + { + "acc": 0.64732928, + "epoch": 0.75, + "grad_norm": 4.8125, + "learning_rate": 7.379736965185369e-06, + "loss": 1.59611406, + "memory(GiB)": 111.15, + "step": 29565, + "train_speed(iter/s)": 0.44842 + }, + { + "acc": 0.6345233, + "epoch": 0.7501268391679351, + "grad_norm": 4.5625, + "learning_rate": 7.378814675855475e-06, + "loss": 1.70942955, + "memory(GiB)": 111.15, + "step": 29570, + "train_speed(iter/s)": 0.448459 + }, + { + "acc": 0.66387906, + "epoch": 0.7502536783358701, + "grad_norm": 5.875, + "learning_rate": 7.37789228189469e-06, + "loss": 1.59051838, + "memory(GiB)": 111.15, + "step": 29575, + "train_speed(iter/s)": 0.448499 + }, + { + "acc": 0.65322037, + "epoch": 0.7503805175038052, + "grad_norm": 4.625, + "learning_rate": 7.376969783343588e-06, + "loss": 1.60753441, + "memory(GiB)": 111.15, + "step": 29580, + "train_speed(iter/s)": 0.448539 + }, + { + "acc": 0.65544038, + "epoch": 0.7505073566717403, + "grad_norm": 5.625, + "learning_rate": 7.37604718024274e-06, + "loss": 1.56431484, + "memory(GiB)": 111.15, + "step": 29585, + "train_speed(iter/s)": 0.448542 + }, + { + "acc": 0.65853043, + "epoch": 0.7506341958396753, + "grad_norm": 6.125, + "learning_rate": 7.375124472632732e-06, + "loss": 1.65765018, + "memory(GiB)": 111.15, + "step": 29590, + "train_speed(iter/s)": 0.448582 + }, + { + "acc": 0.63557653, + "epoch": 0.7507610350076104, + "grad_norm": 4.9375, + "learning_rate": 7.374201660554142e-06, + "loss": 1.69352913, + "memory(GiB)": 111.15, + "step": 29595, + "train_speed(iter/s)": 0.448622 + }, + { + "acc": 0.64573288, + "epoch": 0.7508878741755454, + "grad_norm": 6.125, + "learning_rate": 7.373278744047565e-06, + "loss": 1.65005512, + "memory(GiB)": 111.15, + "step": 29600, + "train_speed(iter/s)": 0.448662 + }, + { + "epoch": 0.7508878741755454, + "eval_acc": 0.6426522876728982, + "eval_loss": 1.611423373222351, + "eval_runtime": 114.5153, + "eval_samples_per_second": 55.626, + "eval_steps_per_second": 27.813, + "step": 29600 + }, + { + "acc": 0.64399376, + "epoch": 0.7510147133434805, + "grad_norm": 6.6875, + "learning_rate": 7.372355723153593e-06, + "loss": 1.70458298, + "memory(GiB)": 111.15, + "step": 29605, + "train_speed(iter/s)": 0.447894 + }, + { + "acc": 0.63826246, + "epoch": 0.7511415525114156, + "grad_norm": 5.875, + "learning_rate": 7.371432597912824e-06, + "loss": 1.66380424, + "memory(GiB)": 111.15, + "step": 29610, + "train_speed(iter/s)": 0.447934 + }, + { + "acc": 0.65049663, + "epoch": 0.7512683916793506, + "grad_norm": 5.4375, + "learning_rate": 7.3705093683658616e-06, + "loss": 1.64381676, + "memory(GiB)": 111.15, + "step": 29615, + "train_speed(iter/s)": 0.447974 + }, + { + "acc": 0.65327501, + "epoch": 0.7513952308472857, + "grad_norm": 5.65625, + "learning_rate": 7.369586034553313e-06, + "loss": 1.64494114, + "memory(GiB)": 111.15, + "step": 29620, + "train_speed(iter/s)": 0.448014 + }, + { + "acc": 0.64212008, + "epoch": 0.7515220700152208, + "grad_norm": 5.75, + "learning_rate": 7.368662596515792e-06, + "loss": 1.71837444, + "memory(GiB)": 111.15, + "step": 29625, + "train_speed(iter/s)": 0.448053 + }, + { + "acc": 0.64545407, + "epoch": 0.7516489091831557, + "grad_norm": 5.90625, + "learning_rate": 7.367739054293914e-06, + "loss": 1.66051254, + "memory(GiB)": 111.15, + "step": 29630, + "train_speed(iter/s)": 0.448094 + }, + { + "acc": 0.64345474, + "epoch": 0.7517757483510908, + "grad_norm": 4.4375, + "learning_rate": 7.366815407928302e-06, + "loss": 1.61324348, + "memory(GiB)": 111.15, + "step": 29635, + "train_speed(iter/s)": 0.448134 + }, + { + "acc": 0.65344048, + "epoch": 0.7519025875190258, + "grad_norm": 4.875, + "learning_rate": 7.365891657459582e-06, + "loss": 1.60489197, + "memory(GiB)": 111.15, + "step": 29640, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.63664408, + "epoch": 0.7520294266869609, + "grad_norm": 5.53125, + "learning_rate": 7.3649678029283825e-06, + "loss": 1.68509712, + "memory(GiB)": 111.15, + "step": 29645, + "train_speed(iter/s)": 0.448214 + }, + { + "acc": 0.64011197, + "epoch": 0.752156265854896, + "grad_norm": 5.8125, + "learning_rate": 7.364043844375342e-06, + "loss": 1.68060188, + "memory(GiB)": 111.15, + "step": 29650, + "train_speed(iter/s)": 0.448254 + }, + { + "acc": 0.64152231, + "epoch": 0.752283105022831, + "grad_norm": 5.1875, + "learning_rate": 7.363119781841095e-06, + "loss": 1.65603142, + "memory(GiB)": 111.15, + "step": 29655, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.63718853, + "epoch": 0.7524099441907661, + "grad_norm": 4.96875, + "learning_rate": 7.362195615366293e-06, + "loss": 1.63750381, + "memory(GiB)": 111.15, + "step": 29660, + "train_speed(iter/s)": 0.448334 + }, + { + "acc": 0.6599865, + "epoch": 0.7525367833587012, + "grad_norm": 6.40625, + "learning_rate": 7.361271344991579e-06, + "loss": 1.57443857, + "memory(GiB)": 111.15, + "step": 29665, + "train_speed(iter/s)": 0.448374 + }, + { + "acc": 0.64805613, + "epoch": 0.7526636225266362, + "grad_norm": 4.96875, + "learning_rate": 7.36034697075761e-06, + "loss": 1.6416853, + "memory(GiB)": 111.15, + "step": 29670, + "train_speed(iter/s)": 0.448414 + }, + { + "acc": 0.65326385, + "epoch": 0.7527904616945713, + "grad_norm": 5.875, + "learning_rate": 7.359422492705043e-06, + "loss": 1.59165592, + "memory(GiB)": 111.15, + "step": 29675, + "train_speed(iter/s)": 0.448454 + }, + { + "acc": 0.65794597, + "epoch": 0.7529173008625063, + "grad_norm": 5.8125, + "learning_rate": 7.3584979108745405e-06, + "loss": 1.59165134, + "memory(GiB)": 111.15, + "step": 29680, + "train_speed(iter/s)": 0.448494 + }, + { + "acc": 0.65152845, + "epoch": 0.7530441400304414, + "grad_norm": 5.15625, + "learning_rate": 7.357573225306771e-06, + "loss": 1.61260643, + "memory(GiB)": 111.15, + "step": 29685, + "train_speed(iter/s)": 0.448534 + }, + { + "acc": 0.64984012, + "epoch": 0.7531709791983765, + "grad_norm": 6.21875, + "learning_rate": 7.356648436042404e-06, + "loss": 1.62704163, + "memory(GiB)": 111.15, + "step": 29690, + "train_speed(iter/s)": 0.448574 + }, + { + "acc": 0.64312568, + "epoch": 0.7532978183663115, + "grad_norm": 6.125, + "learning_rate": 7.355723543122118e-06, + "loss": 1.67074623, + "memory(GiB)": 111.15, + "step": 29695, + "train_speed(iter/s)": 0.448614 + }, + { + "acc": 0.66907072, + "epoch": 0.7534246575342466, + "grad_norm": 4.8125, + "learning_rate": 7.354798546586592e-06, + "loss": 1.52832975, + "memory(GiB)": 111.15, + "step": 29700, + "train_speed(iter/s)": 0.448654 + }, + { + "epoch": 0.7534246575342466, + "eval_acc": 0.6426936466365345, + "eval_loss": 1.6112762689590454, + "eval_runtime": 112.7957, + "eval_samples_per_second": 56.474, + "eval_steps_per_second": 28.237, + "step": 29700 + }, + { + "acc": 0.65637279, + "epoch": 0.7535514967021817, + "grad_norm": 5.4375, + "learning_rate": 7.353873446476512e-06, + "loss": 1.56464252, + "memory(GiB)": 111.15, + "step": 29705, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.65018182, + "epoch": 0.7536783358701167, + "grad_norm": 5.6875, + "learning_rate": 7.3529482428325705e-06, + "loss": 1.65220795, + "memory(GiB)": 111.15, + "step": 29710, + "train_speed(iter/s)": 0.447939 + }, + { + "acc": 0.64791107, + "epoch": 0.7538051750380518, + "grad_norm": 5.84375, + "learning_rate": 7.35202293569546e-06, + "loss": 1.64511318, + "memory(GiB)": 111.15, + "step": 29715, + "train_speed(iter/s)": 0.447979 + }, + { + "acc": 0.64803314, + "epoch": 0.7539320142059868, + "grad_norm": 4.5625, + "learning_rate": 7.351097525105878e-06, + "loss": 1.60299511, + "memory(GiB)": 111.15, + "step": 29720, + "train_speed(iter/s)": 0.448018 + }, + { + "acc": 0.64695387, + "epoch": 0.7540588533739219, + "grad_norm": 4.6875, + "learning_rate": 7.35017201110453e-06, + "loss": 1.55142221, + "memory(GiB)": 111.15, + "step": 29725, + "train_speed(iter/s)": 0.448057 + }, + { + "acc": 0.63772345, + "epoch": 0.754185692541857, + "grad_norm": 5.0, + "learning_rate": 7.349246393732126e-06, + "loss": 1.67036209, + "memory(GiB)": 111.15, + "step": 29730, + "train_speed(iter/s)": 0.448097 + }, + { + "acc": 0.64900789, + "epoch": 0.754312531709792, + "grad_norm": 5.71875, + "learning_rate": 7.3483206730293755e-06, + "loss": 1.63202858, + "memory(GiB)": 111.15, + "step": 29735, + "train_speed(iter/s)": 0.448136 + }, + { + "acc": 0.65556798, + "epoch": 0.7544393708777271, + "grad_norm": 5.25, + "learning_rate": 7.347394849036998e-06, + "loss": 1.61921959, + "memory(GiB)": 111.15, + "step": 29740, + "train_speed(iter/s)": 0.448176 + }, + { + "acc": 0.65656414, + "epoch": 0.7545662100456622, + "grad_norm": 4.96875, + "learning_rate": 7.346468921795714e-06, + "loss": 1.5492382, + "memory(GiB)": 111.15, + "step": 29745, + "train_speed(iter/s)": 0.448215 + }, + { + "acc": 0.6481123, + "epoch": 0.7546930492135971, + "grad_norm": 4.90625, + "learning_rate": 7.345542891346251e-06, + "loss": 1.64267769, + "memory(GiB)": 111.15, + "step": 29750, + "train_speed(iter/s)": 0.448255 + }, + { + "acc": 0.64709363, + "epoch": 0.7548198883815322, + "grad_norm": 5.3125, + "learning_rate": 7.344616757729341e-06, + "loss": 1.62819481, + "memory(GiB)": 111.15, + "step": 29755, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.65700607, + "epoch": 0.7549467275494672, + "grad_norm": 5.34375, + "learning_rate": 7.343690520985716e-06, + "loss": 1.64271355, + "memory(GiB)": 111.15, + "step": 29760, + "train_speed(iter/s)": 0.448334 + }, + { + "acc": 0.6434, + "epoch": 0.7550735667174023, + "grad_norm": 5.21875, + "learning_rate": 7.342764181156119e-06, + "loss": 1.61232777, + "memory(GiB)": 111.15, + "step": 29765, + "train_speed(iter/s)": 0.448374 + }, + { + "acc": 0.65307589, + "epoch": 0.7552004058853374, + "grad_norm": 4.875, + "learning_rate": 7.341837738281293e-06, + "loss": 1.61520882, + "memory(GiB)": 111.15, + "step": 29770, + "train_speed(iter/s)": 0.448413 + }, + { + "acc": 0.65553608, + "epoch": 0.7553272450532724, + "grad_norm": 7.875, + "learning_rate": 7.3409111924019885e-06, + "loss": 1.66322956, + "memory(GiB)": 111.15, + "step": 29775, + "train_speed(iter/s)": 0.448453 + }, + { + "acc": 0.64997621, + "epoch": 0.7554540842212075, + "grad_norm": 6.0625, + "learning_rate": 7.3399845435589574e-06, + "loss": 1.65032883, + "memory(GiB)": 111.15, + "step": 29780, + "train_speed(iter/s)": 0.448492 + }, + { + "acc": 0.6641778, + "epoch": 0.7555809233891426, + "grad_norm": 5.25, + "learning_rate": 7.33905779179296e-06, + "loss": 1.58568811, + "memory(GiB)": 111.15, + "step": 29785, + "train_speed(iter/s)": 0.448532 + }, + { + "acc": 0.64975443, + "epoch": 0.7557077625570776, + "grad_norm": 4.90625, + "learning_rate": 7.338130937144756e-06, + "loss": 1.6420784, + "memory(GiB)": 111.15, + "step": 29790, + "train_speed(iter/s)": 0.448571 + }, + { + "acc": 0.66121421, + "epoch": 0.7558346017250127, + "grad_norm": 4.75, + "learning_rate": 7.3372039796551156e-06, + "loss": 1.58178139, + "memory(GiB)": 111.15, + "step": 29795, + "train_speed(iter/s)": 0.448611 + }, + { + "acc": 0.64968777, + "epoch": 0.7559614408929477, + "grad_norm": 5.03125, + "learning_rate": 7.33627691936481e-06, + "loss": 1.64864197, + "memory(GiB)": 111.15, + "step": 29800, + "train_speed(iter/s)": 0.448651 + }, + { + "epoch": 0.7559614408929477, + "eval_acc": 0.6427467030848356, + "eval_loss": 1.6113356351852417, + "eval_runtime": 114.2308, + "eval_samples_per_second": 55.764, + "eval_steps_per_second": 27.882, + "step": 29800 + }, + { + "acc": 0.67684326, + "epoch": 0.7560882800608828, + "grad_norm": 4.1875, + "learning_rate": 7.335349756314614e-06, + "loss": 1.55186005, + "memory(GiB)": 111.15, + "step": 29805, + "train_speed(iter/s)": 0.447889 + }, + { + "acc": 0.64452791, + "epoch": 0.7562151192288179, + "grad_norm": 4.875, + "learning_rate": 7.33442249054531e-06, + "loss": 1.68445702, + "memory(GiB)": 111.15, + "step": 29810, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.6431283, + "epoch": 0.7563419583967529, + "grad_norm": 4.375, + "learning_rate": 7.33349512209768e-06, + "loss": 1.65035229, + "memory(GiB)": 111.15, + "step": 29815, + "train_speed(iter/s)": 0.447967 + }, + { + "acc": 0.64354343, + "epoch": 0.756468797564688, + "grad_norm": 5.5625, + "learning_rate": 7.332567651012518e-06, + "loss": 1.63578873, + "memory(GiB)": 111.15, + "step": 29820, + "train_speed(iter/s)": 0.448007 + }, + { + "acc": 0.65441008, + "epoch": 0.7565956367326231, + "grad_norm": 5.1875, + "learning_rate": 7.331640077330616e-06, + "loss": 1.61273174, + "memory(GiB)": 111.15, + "step": 29825, + "train_speed(iter/s)": 0.448046 + }, + { + "acc": 0.65694013, + "epoch": 0.7567224759005581, + "grad_norm": 5.53125, + "learning_rate": 7.330712401092773e-06, + "loss": 1.65111542, + "memory(GiB)": 111.15, + "step": 29830, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.64831715, + "epoch": 0.7568493150684932, + "grad_norm": 4.8125, + "learning_rate": 7.329784622339794e-06, + "loss": 1.61761227, + "memory(GiB)": 111.15, + "step": 29835, + "train_speed(iter/s)": 0.448124 + }, + { + "acc": 0.64092751, + "epoch": 0.7569761542364282, + "grad_norm": 6.125, + "learning_rate": 7.328856741112484e-06, + "loss": 1.66966381, + "memory(GiB)": 111.15, + "step": 29840, + "train_speed(iter/s)": 0.448164 + }, + { + "acc": 0.64216509, + "epoch": 0.7571029934043633, + "grad_norm": 5.90625, + "learning_rate": 7.327928757451659e-06, + "loss": 1.61722794, + "memory(GiB)": 111.15, + "step": 29845, + "train_speed(iter/s)": 0.448202 + }, + { + "acc": 0.66727724, + "epoch": 0.7572298325722984, + "grad_norm": 4.5, + "learning_rate": 7.3270006713981325e-06, + "loss": 1.54002466, + "memory(GiB)": 111.15, + "step": 29850, + "train_speed(iter/s)": 0.448242 + }, + { + "acc": 0.64765081, + "epoch": 0.7573566717402334, + "grad_norm": 5.1875, + "learning_rate": 7.326072482992728e-06, + "loss": 1.6719841, + "memory(GiB)": 111.15, + "step": 29855, + "train_speed(iter/s)": 0.448281 + }, + { + "acc": 0.63983617, + "epoch": 0.7574835109081685, + "grad_norm": 5.59375, + "learning_rate": 7.325144192276269e-06, + "loss": 1.66524811, + "memory(GiB)": 111.15, + "step": 29860, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.64262218, + "epoch": 0.7576103500761036, + "grad_norm": 5.90625, + "learning_rate": 7.324215799289588e-06, + "loss": 1.69362106, + "memory(GiB)": 111.15, + "step": 29865, + "train_speed(iter/s)": 0.44836 + }, + { + "acc": 0.66570525, + "epoch": 0.7577371892440385, + "grad_norm": 7.1875, + "learning_rate": 7.3232873040735194e-06, + "loss": 1.59484787, + "memory(GiB)": 111.15, + "step": 29870, + "train_speed(iter/s)": 0.448399 + }, + { + "acc": 0.64402337, + "epoch": 0.7578640284119736, + "grad_norm": 5.15625, + "learning_rate": 7.322358706668901e-06, + "loss": 1.6546608, + "memory(GiB)": 111.15, + "step": 29875, + "train_speed(iter/s)": 0.448438 + }, + { + "acc": 0.65233955, + "epoch": 0.7579908675799086, + "grad_norm": 4.75, + "learning_rate": 7.321430007116582e-06, + "loss": 1.63348522, + "memory(GiB)": 111.15, + "step": 29880, + "train_speed(iter/s)": 0.448477 + }, + { + "acc": 0.63426504, + "epoch": 0.7581177067478437, + "grad_norm": 6.25, + "learning_rate": 7.320501205457403e-06, + "loss": 1.69893074, + "memory(GiB)": 111.15, + "step": 29885, + "train_speed(iter/s)": 0.448516 + }, + { + "acc": 0.64417114, + "epoch": 0.7582445459157788, + "grad_norm": 6.0, + "learning_rate": 7.319572301732224e-06, + "loss": 1.60525341, + "memory(GiB)": 111.15, + "step": 29890, + "train_speed(iter/s)": 0.448556 + }, + { + "acc": 0.63854198, + "epoch": 0.7583713850837138, + "grad_norm": 4.53125, + "learning_rate": 7.3186432959818956e-06, + "loss": 1.65383415, + "memory(GiB)": 111.15, + "step": 29895, + "train_speed(iter/s)": 0.448595 + }, + { + "acc": 0.65765052, + "epoch": 0.7584982242516489, + "grad_norm": 5.34375, + "learning_rate": 7.317714188247285e-06, + "loss": 1.60192642, + "memory(GiB)": 111.15, + "step": 29900, + "train_speed(iter/s)": 0.448635 + }, + { + "epoch": 0.7584982242516489, + "eval_acc": 0.6426982420769385, + "eval_loss": 1.611272931098938, + "eval_runtime": 114.8233, + "eval_samples_per_second": 55.477, + "eval_steps_per_second": 27.738, + "step": 29900 + }, + { + "acc": 0.64634085, + "epoch": 0.758625063419584, + "grad_norm": 5.0625, + "learning_rate": 7.316784978569256e-06, + "loss": 1.65383759, + "memory(GiB)": 111.15, + "step": 29905, + "train_speed(iter/s)": 0.447872 + }, + { + "acc": 0.65898113, + "epoch": 0.758751902587519, + "grad_norm": 5.03125, + "learning_rate": 7.31585566698868e-06, + "loss": 1.49823742, + "memory(GiB)": 111.15, + "step": 29910, + "train_speed(iter/s)": 0.447911 + }, + { + "acc": 0.66311164, + "epoch": 0.7588787417554541, + "grad_norm": 5.78125, + "learning_rate": 7.314926253546433e-06, + "loss": 1.51500111, + "memory(GiB)": 111.15, + "step": 29915, + "train_speed(iter/s)": 0.44795 + }, + { + "acc": 0.65015774, + "epoch": 0.7590055809233891, + "grad_norm": 5.125, + "learning_rate": 7.313996738283393e-06, + "loss": 1.64286995, + "memory(GiB)": 111.15, + "step": 29920, + "train_speed(iter/s)": 0.447989 + }, + { + "acc": 0.65285101, + "epoch": 0.7591324200913242, + "grad_norm": 6.28125, + "learning_rate": 7.3130671212404455e-06, + "loss": 1.6490242, + "memory(GiB)": 111.15, + "step": 29925, + "train_speed(iter/s)": 0.448029 + }, + { + "acc": 0.65334873, + "epoch": 0.7592592592592593, + "grad_norm": 4.9375, + "learning_rate": 7.312137402458479e-06, + "loss": 1.6387249, + "memory(GiB)": 111.15, + "step": 29930, + "train_speed(iter/s)": 0.448068 + }, + { + "acc": 0.64803734, + "epoch": 0.7593860984271943, + "grad_norm": 4.75, + "learning_rate": 7.3112075819783864e-06, + "loss": 1.62522545, + "memory(GiB)": 111.15, + "step": 29935, + "train_speed(iter/s)": 0.448108 + }, + { + "acc": 0.6336606, + "epoch": 0.7595129375951294, + "grad_norm": 6.78125, + "learning_rate": 7.310277659841066e-06, + "loss": 1.74649258, + "memory(GiB)": 111.15, + "step": 29940, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.63511086, + "epoch": 0.7596397767630645, + "grad_norm": 4.90625, + "learning_rate": 7.309347636087418e-06, + "loss": 1.65225697, + "memory(GiB)": 111.15, + "step": 29945, + "train_speed(iter/s)": 0.448186 + }, + { + "acc": 0.63393641, + "epoch": 0.7597666159309995, + "grad_norm": 4.375, + "learning_rate": 7.308417510758353e-06, + "loss": 1.68019085, + "memory(GiB)": 111.15, + "step": 29950, + "train_speed(iter/s)": 0.448226 + }, + { + "acc": 0.63289709, + "epoch": 0.7598934550989346, + "grad_norm": 5.03125, + "learning_rate": 7.307487283894777e-06, + "loss": 1.72885895, + "memory(GiB)": 111.15, + "step": 29955, + "train_speed(iter/s)": 0.448265 + }, + { + "acc": 0.66571217, + "epoch": 0.7600202942668696, + "grad_norm": 5.0, + "learning_rate": 7.30655695553761e-06, + "loss": 1.52079039, + "memory(GiB)": 111.15, + "step": 29960, + "train_speed(iter/s)": 0.448305 + }, + { + "acc": 0.65741158, + "epoch": 0.7601471334348047, + "grad_norm": 5.5625, + "learning_rate": 7.305626525727769e-06, + "loss": 1.60763683, + "memory(GiB)": 111.15, + "step": 29965, + "train_speed(iter/s)": 0.448344 + }, + { + "acc": 0.64417782, + "epoch": 0.7602739726027398, + "grad_norm": 4.875, + "learning_rate": 7.30469599450618e-06, + "loss": 1.71989708, + "memory(GiB)": 111.15, + "step": 29970, + "train_speed(iter/s)": 0.448384 + }, + { + "acc": 0.65053682, + "epoch": 0.7604008117706748, + "grad_norm": 4.6875, + "learning_rate": 7.30376536191377e-06, + "loss": 1.62371521, + "memory(GiB)": 111.15, + "step": 29975, + "train_speed(iter/s)": 0.448423 + }, + { + "acc": 0.65328941, + "epoch": 0.7605276509386099, + "grad_norm": 4.5, + "learning_rate": 7.302834627991477e-06, + "loss": 1.63394432, + "memory(GiB)": 111.15, + "step": 29980, + "train_speed(iter/s)": 0.448462 + }, + { + "acc": 0.64684954, + "epoch": 0.760654490106545, + "grad_norm": 4.8125, + "learning_rate": 7.301903792780233e-06, + "loss": 1.67160091, + "memory(GiB)": 111.15, + "step": 29985, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.65974007, + "epoch": 0.76078132927448, + "grad_norm": 5.125, + "learning_rate": 7.300972856320984e-06, + "loss": 1.58969059, + "memory(GiB)": 111.15, + "step": 29990, + "train_speed(iter/s)": 0.448541 + }, + { + "acc": 0.64870405, + "epoch": 0.760908168442415, + "grad_norm": 5.5625, + "learning_rate": 7.3000418186546754e-06, + "loss": 1.63670921, + "memory(GiB)": 111.15, + "step": 29995, + "train_speed(iter/s)": 0.448581 + }, + { + "acc": 0.66252594, + "epoch": 0.76103500761035, + "grad_norm": 4.4375, + "learning_rate": 7.299110679822258e-06, + "loss": 1.54388256, + "memory(GiB)": 111.15, + "step": 30000, + "train_speed(iter/s)": 0.448621 + }, + { + "epoch": 0.76103500761035, + "eval_acc": 0.6426669095287292, + "eval_loss": 1.6111167669296265, + "eval_runtime": 112.9886, + "eval_samples_per_second": 56.377, + "eval_steps_per_second": 28.189, + "step": 30000 + }, + { + "acc": 0.64246559, + "epoch": 0.7611618467782851, + "grad_norm": 5.4375, + "learning_rate": 7.298179439864689e-06, + "loss": 1.66643944, + "memory(GiB)": 111.15, + "step": 30005, + "train_speed(iter/s)": 0.447873 + }, + { + "acc": 0.65696039, + "epoch": 0.7612886859462202, + "grad_norm": 4.78125, + "learning_rate": 7.297248098822926e-06, + "loss": 1.596068, + "memory(GiB)": 111.15, + "step": 30010, + "train_speed(iter/s)": 0.447912 + }, + { + "acc": 0.65419993, + "epoch": 0.7614155251141552, + "grad_norm": 4.625, + "learning_rate": 7.296316656737936e-06, + "loss": 1.62100105, + "memory(GiB)": 111.15, + "step": 30015, + "train_speed(iter/s)": 0.44795 + }, + { + "acc": 0.67120819, + "epoch": 0.7615423642820903, + "grad_norm": 4.96875, + "learning_rate": 7.295385113650689e-06, + "loss": 1.59611034, + "memory(GiB)": 111.15, + "step": 30020, + "train_speed(iter/s)": 0.447989 + }, + { + "acc": 0.6441031, + "epoch": 0.7616692034500254, + "grad_norm": 5.15625, + "learning_rate": 7.294453469602154e-06, + "loss": 1.61740303, + "memory(GiB)": 111.15, + "step": 30025, + "train_speed(iter/s)": 0.448028 + }, + { + "acc": 0.67082419, + "epoch": 0.7617960426179604, + "grad_norm": 4.75, + "learning_rate": 7.293521724633313e-06, + "loss": 1.5952342, + "memory(GiB)": 111.15, + "step": 30030, + "train_speed(iter/s)": 0.448067 + }, + { + "acc": 0.6465045, + "epoch": 0.7619228817858955, + "grad_norm": 5.09375, + "learning_rate": 7.2925898787851455e-06, + "loss": 1.63666, + "memory(GiB)": 111.15, + "step": 30035, + "train_speed(iter/s)": 0.448106 + }, + { + "acc": 0.65152845, + "epoch": 0.7620497209538305, + "grad_norm": 5.15625, + "learning_rate": 7.2916579320986415e-06, + "loss": 1.62414341, + "memory(GiB)": 111.15, + "step": 30040, + "train_speed(iter/s)": 0.448145 + }, + { + "acc": 0.65448036, + "epoch": 0.7621765601217656, + "grad_norm": 5.21875, + "learning_rate": 7.290725884614787e-06, + "loss": 1.66457996, + "memory(GiB)": 111.15, + "step": 30045, + "train_speed(iter/s)": 0.448184 + }, + { + "acc": 0.6500072, + "epoch": 0.7623033992897007, + "grad_norm": 4.90625, + "learning_rate": 7.2897937363745844e-06, + "loss": 1.60955658, + "memory(GiB)": 111.15, + "step": 30050, + "train_speed(iter/s)": 0.448223 + }, + { + "acc": 0.64587259, + "epoch": 0.7624302384576357, + "grad_norm": 6.21875, + "learning_rate": 7.2888614874190276e-06, + "loss": 1.63615265, + "memory(GiB)": 111.15, + "step": 30055, + "train_speed(iter/s)": 0.448262 + }, + { + "acc": 0.65153112, + "epoch": 0.7625570776255708, + "grad_norm": 5.59375, + "learning_rate": 7.287929137789124e-06, + "loss": 1.61883278, + "memory(GiB)": 111.15, + "step": 30060, + "train_speed(iter/s)": 0.448301 + }, + { + "acc": 0.66019835, + "epoch": 0.7626839167935059, + "grad_norm": 5.90625, + "learning_rate": 7.286996687525882e-06, + "loss": 1.63256721, + "memory(GiB)": 111.15, + "step": 30065, + "train_speed(iter/s)": 0.44834 + }, + { + "acc": 0.64206676, + "epoch": 0.7628107559614409, + "grad_norm": 4.8125, + "learning_rate": 7.2860641366703155e-06, + "loss": 1.6178894, + "memory(GiB)": 111.15, + "step": 30070, + "train_speed(iter/s)": 0.448379 + }, + { + "acc": 0.64384236, + "epoch": 0.762937595129376, + "grad_norm": 5.1875, + "learning_rate": 7.285131485263441e-06, + "loss": 1.64949455, + "memory(GiB)": 111.15, + "step": 30075, + "train_speed(iter/s)": 0.448417 + }, + { + "acc": 0.65056772, + "epoch": 0.763064434297311, + "grad_norm": 5.1875, + "learning_rate": 7.2841987333462815e-06, + "loss": 1.70610847, + "memory(GiB)": 111.15, + "step": 30080, + "train_speed(iter/s)": 0.448456 + }, + { + "acc": 0.65302486, + "epoch": 0.7631912734652461, + "grad_norm": 5.75, + "learning_rate": 7.283265880959863e-06, + "loss": 1.62368565, + "memory(GiB)": 111.15, + "step": 30085, + "train_speed(iter/s)": 0.448495 + }, + { + "acc": 0.64530873, + "epoch": 0.7633181126331812, + "grad_norm": 4.9375, + "learning_rate": 7.282332928145219e-06, + "loss": 1.62245865, + "memory(GiB)": 111.15, + "step": 30090, + "train_speed(iter/s)": 0.448535 + }, + { + "acc": 0.65086517, + "epoch": 0.7634449518011162, + "grad_norm": 5.25, + "learning_rate": 7.281399874943381e-06, + "loss": 1.5647274, + "memory(GiB)": 111.15, + "step": 30095, + "train_speed(iter/s)": 0.448573 + }, + { + "acc": 0.64655824, + "epoch": 0.7635717909690513, + "grad_norm": 5.8125, + "learning_rate": 7.280466721395393e-06, + "loss": 1.68896599, + "memory(GiB)": 111.15, + "step": 30100, + "train_speed(iter/s)": 0.448612 + }, + { + "epoch": 0.7635717909690513, + "eval_acc": 0.64278179553883, + "eval_loss": 1.6111350059509277, + "eval_runtime": 113.8017, + "eval_samples_per_second": 55.975, + "eval_steps_per_second": 27.987, + "step": 30100 + }, + { + "acc": 0.63785639, + "epoch": 0.7636986301369864, + "grad_norm": 4.78125, + "learning_rate": 7.279533467542295e-06, + "loss": 1.65344791, + "memory(GiB)": 111.15, + "step": 30105, + "train_speed(iter/s)": 0.447862 + }, + { + "acc": 0.64260335, + "epoch": 0.7638254693049213, + "grad_norm": 5.53125, + "learning_rate": 7.2786001134251385e-06, + "loss": 1.66452274, + "memory(GiB)": 111.15, + "step": 30110, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.65048285, + "epoch": 0.7639523084728564, + "grad_norm": 5.84375, + "learning_rate": 7.2776666590849744e-06, + "loss": 1.6166687, + "memory(GiB)": 111.15, + "step": 30115, + "train_speed(iter/s)": 0.447939 + }, + { + "acc": 0.6577178, + "epoch": 0.7640791476407914, + "grad_norm": 4.8125, + "learning_rate": 7.276733104562863e-06, + "loss": 1.60323982, + "memory(GiB)": 111.15, + "step": 30120, + "train_speed(iter/s)": 0.447978 + }, + { + "acc": 0.65243826, + "epoch": 0.7642059868087265, + "grad_norm": 4.84375, + "learning_rate": 7.275799449899865e-06, + "loss": 1.63850365, + "memory(GiB)": 111.15, + "step": 30125, + "train_speed(iter/s)": 0.448017 + }, + { + "acc": 0.65368385, + "epoch": 0.7643328259766616, + "grad_norm": 5.375, + "learning_rate": 7.274865695137046e-06, + "loss": 1.66626396, + "memory(GiB)": 111.15, + "step": 30130, + "train_speed(iter/s)": 0.448056 + }, + { + "acc": 0.6532546, + "epoch": 0.7644596651445966, + "grad_norm": 5.46875, + "learning_rate": 7.273931840315477e-06, + "loss": 1.61202908, + "memory(GiB)": 111.15, + "step": 30135, + "train_speed(iter/s)": 0.448095 + }, + { + "acc": 0.66442995, + "epoch": 0.7645865043125317, + "grad_norm": 4.8125, + "learning_rate": 7.272997885476234e-06, + "loss": 1.5684701, + "memory(GiB)": 111.15, + "step": 30140, + "train_speed(iter/s)": 0.448134 + }, + { + "acc": 0.63842745, + "epoch": 0.7647133434804668, + "grad_norm": 4.75, + "learning_rate": 7.272063830660395e-06, + "loss": 1.69974861, + "memory(GiB)": 111.15, + "step": 30145, + "train_speed(iter/s)": 0.448173 + }, + { + "acc": 0.65501518, + "epoch": 0.7648401826484018, + "grad_norm": 5.75, + "learning_rate": 7.271129675909046e-06, + "loss": 1.62374744, + "memory(GiB)": 111.15, + "step": 30150, + "train_speed(iter/s)": 0.448212 + }, + { + "acc": 0.66463337, + "epoch": 0.7649670218163369, + "grad_norm": 5.84375, + "learning_rate": 7.270195421263271e-06, + "loss": 1.53539562, + "memory(GiB)": 111.15, + "step": 30155, + "train_speed(iter/s)": 0.448251 + }, + { + "acc": 0.65181675, + "epoch": 0.7650938609842719, + "grad_norm": 5.125, + "learning_rate": 7.269261066764169e-06, + "loss": 1.61901569, + "memory(GiB)": 111.15, + "step": 30160, + "train_speed(iter/s)": 0.44829 + }, + { + "acc": 0.63298106, + "epoch": 0.765220700152207, + "grad_norm": 6.1875, + "learning_rate": 7.268326612452832e-06, + "loss": 1.63864021, + "memory(GiB)": 111.15, + "step": 30165, + "train_speed(iter/s)": 0.448328 + }, + { + "acc": 0.65076232, + "epoch": 0.7653475393201421, + "grad_norm": 5.71875, + "learning_rate": 7.267392058370364e-06, + "loss": 1.6069519, + "memory(GiB)": 111.15, + "step": 30170, + "train_speed(iter/s)": 0.448367 + }, + { + "acc": 0.65491314, + "epoch": 0.7654743784880771, + "grad_norm": 5.46875, + "learning_rate": 7.2664574045578685e-06, + "loss": 1.61631947, + "memory(GiB)": 111.15, + "step": 30175, + "train_speed(iter/s)": 0.448406 + }, + { + "acc": 0.65142212, + "epoch": 0.7656012176560122, + "grad_norm": 4.96875, + "learning_rate": 7.26552265105646e-06, + "loss": 1.65182323, + "memory(GiB)": 111.15, + "step": 30180, + "train_speed(iter/s)": 0.448445 + }, + { + "acc": 0.64304886, + "epoch": 0.7657280568239473, + "grad_norm": 5.0625, + "learning_rate": 7.264587797907248e-06, + "loss": 1.62409077, + "memory(GiB)": 111.15, + "step": 30185, + "train_speed(iter/s)": 0.448484 + }, + { + "acc": 0.65312738, + "epoch": 0.7658548959918823, + "grad_norm": 4.9375, + "learning_rate": 7.263652845151354e-06, + "loss": 1.5696826, + "memory(GiB)": 111.15, + "step": 30190, + "train_speed(iter/s)": 0.448522 + }, + { + "acc": 0.65120454, + "epoch": 0.7659817351598174, + "grad_norm": 5.71875, + "learning_rate": 7.262717792829903e-06, + "loss": 1.62431145, + "memory(GiB)": 111.15, + "step": 30195, + "train_speed(iter/s)": 0.448561 + }, + { + "acc": 0.6600009, + "epoch": 0.7661085743277524, + "grad_norm": 5.625, + "learning_rate": 7.261782640984021e-06, + "loss": 1.59343166, + "memory(GiB)": 111.15, + "step": 30200, + "train_speed(iter/s)": 0.448599 + }, + { + "epoch": 0.7661085743277524, + "eval_acc": 0.6426836202211075, + "eval_loss": 1.6109641790390015, + "eval_runtime": 112.9517, + "eval_samples_per_second": 56.396, + "eval_steps_per_second": 28.198, + "step": 30200 + }, + { + "acc": 0.6364152, + "epoch": 0.7662354134956875, + "grad_norm": 5.09375, + "learning_rate": 7.26084738965484e-06, + "loss": 1.69018402, + "memory(GiB)": 111.15, + "step": 30205, + "train_speed(iter/s)": 0.447857 + }, + { + "acc": 0.64831028, + "epoch": 0.7663622526636226, + "grad_norm": 5.5625, + "learning_rate": 7.2599120388834964e-06, + "loss": 1.65325546, + "memory(GiB)": 111.15, + "step": 30210, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.65875854, + "epoch": 0.7664890918315576, + "grad_norm": 5.5625, + "learning_rate": 7.258976588711133e-06, + "loss": 1.62519493, + "memory(GiB)": 111.15, + "step": 30215, + "train_speed(iter/s)": 0.447934 + }, + { + "acc": 0.6614954, + "epoch": 0.7666159309994927, + "grad_norm": 5.59375, + "learning_rate": 7.258041039178891e-06, + "loss": 1.57508144, + "memory(GiB)": 111.15, + "step": 30220, + "train_speed(iter/s)": 0.447973 + }, + { + "acc": 0.63735447, + "epoch": 0.7667427701674278, + "grad_norm": 5.78125, + "learning_rate": 7.257105390327925e-06, + "loss": 1.65613213, + "memory(GiB)": 111.15, + "step": 30225, + "train_speed(iter/s)": 0.448012 + }, + { + "acc": 0.65610299, + "epoch": 0.7668696093353627, + "grad_norm": 5.21875, + "learning_rate": 7.256169642199386e-06, + "loss": 1.5868084, + "memory(GiB)": 111.15, + "step": 30230, + "train_speed(iter/s)": 0.44805 + }, + { + "acc": 0.6431921, + "epoch": 0.7669964485032978, + "grad_norm": 5.25, + "learning_rate": 7.255233794834432e-06, + "loss": 1.65678825, + "memory(GiB)": 111.15, + "step": 30235, + "train_speed(iter/s)": 0.448089 + }, + { + "acc": 0.66883497, + "epoch": 0.7671232876712328, + "grad_norm": 5.75, + "learning_rate": 7.254297848274229e-06, + "loss": 1.53180084, + "memory(GiB)": 111.15, + "step": 30240, + "train_speed(iter/s)": 0.448127 + }, + { + "acc": 0.65240412, + "epoch": 0.7672501268391679, + "grad_norm": 6.15625, + "learning_rate": 7.25336180255994e-06, + "loss": 1.63762321, + "memory(GiB)": 111.15, + "step": 30245, + "train_speed(iter/s)": 0.448166 + }, + { + "acc": 0.64969745, + "epoch": 0.767376966007103, + "grad_norm": 4.65625, + "learning_rate": 7.25242565773274e-06, + "loss": 1.67601967, + "memory(GiB)": 111.15, + "step": 30250, + "train_speed(iter/s)": 0.448204 + }, + { + "acc": 0.64600344, + "epoch": 0.767503805175038, + "grad_norm": 5.0, + "learning_rate": 7.251489413833801e-06, + "loss": 1.63121719, + "memory(GiB)": 111.15, + "step": 30255, + "train_speed(iter/s)": 0.448243 + }, + { + "acc": 0.64322805, + "epoch": 0.7676306443429731, + "grad_norm": 5.4375, + "learning_rate": 7.250553070904307e-06, + "loss": 1.67427368, + "memory(GiB)": 111.15, + "step": 30260, + "train_speed(iter/s)": 0.448281 + }, + { + "acc": 0.65524769, + "epoch": 0.7677574835109082, + "grad_norm": 5.75, + "learning_rate": 7.2496166289854404e-06, + "loss": 1.58608332, + "memory(GiB)": 111.15, + "step": 30265, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.65058756, + "epoch": 0.7678843226788432, + "grad_norm": 5.21875, + "learning_rate": 7.24868008811839e-06, + "loss": 1.64924107, + "memory(GiB)": 111.15, + "step": 30270, + "train_speed(iter/s)": 0.448358 + }, + { + "acc": 0.65252647, + "epoch": 0.7680111618467783, + "grad_norm": 5.40625, + "learning_rate": 7.247743448344351e-06, + "loss": 1.62771244, + "memory(GiB)": 111.15, + "step": 30275, + "train_speed(iter/s)": 0.448397 + }, + { + "acc": 0.63750067, + "epoch": 0.7681380010147133, + "grad_norm": 4.8125, + "learning_rate": 7.246806709704519e-06, + "loss": 1.69692993, + "memory(GiB)": 111.15, + "step": 30280, + "train_speed(iter/s)": 0.448435 + }, + { + "acc": 0.64564075, + "epoch": 0.7682648401826484, + "grad_norm": 5.09375, + "learning_rate": 7.245869872240098e-06, + "loss": 1.69884911, + "memory(GiB)": 111.15, + "step": 30285, + "train_speed(iter/s)": 0.448474 + }, + { + "acc": 0.65779438, + "epoch": 0.7683916793505835, + "grad_norm": 5.46875, + "learning_rate": 7.244932935992292e-06, + "loss": 1.6648819, + "memory(GiB)": 111.15, + "step": 30290, + "train_speed(iter/s)": 0.448512 + }, + { + "acc": 0.65220356, + "epoch": 0.7685185185185185, + "grad_norm": 6.46875, + "learning_rate": 7.243995901002312e-06, + "loss": 1.68145103, + "memory(GiB)": 111.15, + "step": 30295, + "train_speed(iter/s)": 0.448551 + }, + { + "acc": 0.64849854, + "epoch": 0.7686453576864536, + "grad_norm": 5.6875, + "learning_rate": 7.243058767311374e-06, + "loss": 1.63159981, + "memory(GiB)": 111.15, + "step": 30300, + "train_speed(iter/s)": 0.44859 + }, + { + "epoch": 0.7686453576864536, + "eval_acc": 0.6426606430190873, + "eval_loss": 1.6111055612564087, + "eval_runtime": 114.4022, + "eval_samples_per_second": 55.681, + "eval_steps_per_second": 27.84, + "step": 30300 + }, + { + "acc": 0.66738143, + "epoch": 0.7687721968543887, + "grad_norm": 6.09375, + "learning_rate": 7.2421215349606955e-06, + "loss": 1.57330503, + "memory(GiB)": 111.15, + "step": 30305, + "train_speed(iter/s)": 0.44784 + }, + { + "acc": 0.65920448, + "epoch": 0.7688990360223237, + "grad_norm": 5.6875, + "learning_rate": 7.241184203991505e-06, + "loss": 1.52713652, + "memory(GiB)": 111.15, + "step": 30310, + "train_speed(iter/s)": 0.447878 + }, + { + "acc": 0.64299231, + "epoch": 0.7690258751902588, + "grad_norm": 5.65625, + "learning_rate": 7.240246774445024e-06, + "loss": 1.62029419, + "memory(GiB)": 111.15, + "step": 30315, + "train_speed(iter/s)": 0.447917 + }, + { + "acc": 0.65016432, + "epoch": 0.7691527143581938, + "grad_norm": 5.28125, + "learning_rate": 7.23930924636249e-06, + "loss": 1.62468147, + "memory(GiB)": 111.15, + "step": 30320, + "train_speed(iter/s)": 0.447956 + }, + { + "acc": 0.6469203, + "epoch": 0.7692795535261289, + "grad_norm": 5.09375, + "learning_rate": 7.238371619785134e-06, + "loss": 1.63639526, + "memory(GiB)": 111.15, + "step": 30325, + "train_speed(iter/s)": 0.447994 + }, + { + "acc": 0.64485559, + "epoch": 0.769406392694064, + "grad_norm": 5.3125, + "learning_rate": 7.237433894754205e-06, + "loss": 1.63096466, + "memory(GiB)": 111.15, + "step": 30330, + "train_speed(iter/s)": 0.448033 + }, + { + "acc": 0.64584804, + "epoch": 0.769533231861999, + "grad_norm": 5.53125, + "learning_rate": 7.23649607131094e-06, + "loss": 1.6668087, + "memory(GiB)": 111.15, + "step": 30335, + "train_speed(iter/s)": 0.448073 + }, + { + "acc": 0.65014143, + "epoch": 0.7696600710299341, + "grad_norm": 5.0, + "learning_rate": 7.235558149496595e-06, + "loss": 1.59731741, + "memory(GiB)": 111.15, + "step": 30340, + "train_speed(iter/s)": 0.448112 + }, + { + "acc": 0.66149254, + "epoch": 0.7697869101978692, + "grad_norm": 5.09375, + "learning_rate": 7.23462012935242e-06, + "loss": 1.59674988, + "memory(GiB)": 111.15, + "step": 30345, + "train_speed(iter/s)": 0.448151 + }, + { + "acc": 0.64201016, + "epoch": 0.7699137493658041, + "grad_norm": 5.125, + "learning_rate": 7.233682010919676e-06, + "loss": 1.6620018, + "memory(GiB)": 111.15, + "step": 30350, + "train_speed(iter/s)": 0.44819 + }, + { + "acc": 0.665131, + "epoch": 0.7700405885337392, + "grad_norm": 4.5, + "learning_rate": 7.2327437942396236e-06, + "loss": 1.59963169, + "memory(GiB)": 111.15, + "step": 30355, + "train_speed(iter/s)": 0.448229 + }, + { + "acc": 0.66392565, + "epoch": 0.7701674277016742, + "grad_norm": 5.75, + "learning_rate": 7.231805479353532e-06, + "loss": 1.56770182, + "memory(GiB)": 111.15, + "step": 30360, + "train_speed(iter/s)": 0.448268 + }, + { + "acc": 0.6601572, + "epoch": 0.7702942668696093, + "grad_norm": 4.96875, + "learning_rate": 7.2308670663026705e-06, + "loss": 1.64786453, + "memory(GiB)": 111.15, + "step": 30365, + "train_speed(iter/s)": 0.448307 + }, + { + "acc": 0.64040294, + "epoch": 0.7704211060375444, + "grad_norm": 4.625, + "learning_rate": 7.229928555128315e-06, + "loss": 1.67627544, + "memory(GiB)": 111.15, + "step": 30370, + "train_speed(iter/s)": 0.448346 + }, + { + "acc": 0.65747218, + "epoch": 0.7705479452054794, + "grad_norm": 5.5625, + "learning_rate": 7.228989945871745e-06, + "loss": 1.65121117, + "memory(GiB)": 111.15, + "step": 30375, + "train_speed(iter/s)": 0.448385 + }, + { + "acc": 0.64784069, + "epoch": 0.7706747843734145, + "grad_norm": 4.78125, + "learning_rate": 7.2280512385742475e-06, + "loss": 1.60855274, + "memory(GiB)": 111.15, + "step": 30380, + "train_speed(iter/s)": 0.448423 + }, + { + "acc": 0.65265765, + "epoch": 0.7708016235413496, + "grad_norm": 5.3125, + "learning_rate": 7.227112433277107e-06, + "loss": 1.60414352, + "memory(GiB)": 111.15, + "step": 30385, + "train_speed(iter/s)": 0.448462 + }, + { + "acc": 0.64870348, + "epoch": 0.7709284627092846, + "grad_norm": 5.78125, + "learning_rate": 7.2261735300216195e-06, + "loss": 1.64852905, + "memory(GiB)": 111.15, + "step": 30390, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.64650574, + "epoch": 0.7710553018772197, + "grad_norm": 5.875, + "learning_rate": 7.22523452884908e-06, + "loss": 1.66050949, + "memory(GiB)": 111.15, + "step": 30395, + "train_speed(iter/s)": 0.44854 + }, + { + "acc": 0.63345776, + "epoch": 0.7711821410451547, + "grad_norm": 5.40625, + "learning_rate": 7.224295429800792e-06, + "loss": 1.698559, + "memory(GiB)": 111.15, + "step": 30400, + "train_speed(iter/s)": 0.448578 + }, + { + "epoch": 0.7711821410451547, + "eval_acc": 0.6427107750962222, + "eval_loss": 1.6110895872116089, + "eval_runtime": 113.7505, + "eval_samples_per_second": 56.0, + "eval_steps_per_second": 28.0, + "step": 30400 + }, + { + "acc": 0.64281955, + "epoch": 0.7713089802130898, + "grad_norm": 5.0, + "learning_rate": 7.22335623291806e-06, + "loss": 1.67297974, + "memory(GiB)": 111.15, + "step": 30405, + "train_speed(iter/s)": 0.447835 + }, + { + "acc": 0.64722748, + "epoch": 0.7714358193810249, + "grad_norm": 4.59375, + "learning_rate": 7.222416938242194e-06, + "loss": 1.6641983, + "memory(GiB)": 111.15, + "step": 30410, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.66258779, + "epoch": 0.7715626585489599, + "grad_norm": 6.1875, + "learning_rate": 7.221477545814509e-06, + "loss": 1.53363323, + "memory(GiB)": 111.15, + "step": 30415, + "train_speed(iter/s)": 0.447913 + }, + { + "acc": 0.65373149, + "epoch": 0.771689497716895, + "grad_norm": 4.875, + "learning_rate": 7.220538055676323e-06, + "loss": 1.6405941, + "memory(GiB)": 111.15, + "step": 30420, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.64540758, + "epoch": 0.7718163368848301, + "grad_norm": 6.84375, + "learning_rate": 7.21959846786896e-06, + "loss": 1.59779587, + "memory(GiB)": 111.15, + "step": 30425, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.66007209, + "epoch": 0.7719431760527651, + "grad_norm": 5.71875, + "learning_rate": 7.218658782433746e-06, + "loss": 1.58770237, + "memory(GiB)": 111.15, + "step": 30430, + "train_speed(iter/s)": 0.448029 + }, + { + "acc": 0.65856233, + "epoch": 0.7720700152207002, + "grad_norm": 5.5, + "learning_rate": 7.217718999412013e-06, + "loss": 1.56408529, + "memory(GiB)": 111.15, + "step": 30435, + "train_speed(iter/s)": 0.448068 + }, + { + "acc": 0.6506496, + "epoch": 0.7721968543886352, + "grad_norm": 5.78125, + "learning_rate": 7.216779118845097e-06, + "loss": 1.64257507, + "memory(GiB)": 111.15, + "step": 30440, + "train_speed(iter/s)": 0.448106 + }, + { + "acc": 0.63714762, + "epoch": 0.7723236935565703, + "grad_norm": 5.59375, + "learning_rate": 7.215839140774339e-06, + "loss": 1.62282982, + "memory(GiB)": 111.15, + "step": 30445, + "train_speed(iter/s)": 0.448145 + }, + { + "acc": 0.65486498, + "epoch": 0.7724505327245054, + "grad_norm": 5.03125, + "learning_rate": 7.214899065241082e-06, + "loss": 1.66951561, + "memory(GiB)": 111.15, + "step": 30450, + "train_speed(iter/s)": 0.448184 + }, + { + "acc": 0.66500282, + "epoch": 0.7725773718924404, + "grad_norm": 6.65625, + "learning_rate": 7.213958892286674e-06, + "loss": 1.59347134, + "memory(GiB)": 111.15, + "step": 30455, + "train_speed(iter/s)": 0.448222 + }, + { + "acc": 0.64602823, + "epoch": 0.7727042110603755, + "grad_norm": 5.0625, + "learning_rate": 7.213018621952472e-06, + "loss": 1.5939188, + "memory(GiB)": 111.15, + "step": 30460, + "train_speed(iter/s)": 0.448261 + }, + { + "acc": 0.65890102, + "epoch": 0.7728310502283106, + "grad_norm": 5.3125, + "learning_rate": 7.212078254279828e-06, + "loss": 1.56410618, + "memory(GiB)": 111.15, + "step": 30465, + "train_speed(iter/s)": 0.4483 + }, + { + "acc": 0.63967867, + "epoch": 0.7729578893962455, + "grad_norm": 4.96875, + "learning_rate": 7.211137789310109e-06, + "loss": 1.66014519, + "memory(GiB)": 111.15, + "step": 30470, + "train_speed(iter/s)": 0.448339 + }, + { + "acc": 0.6699645, + "epoch": 0.7730847285641806, + "grad_norm": 5.03125, + "learning_rate": 7.2101972270846756e-06, + "loss": 1.55592957, + "memory(GiB)": 111.15, + "step": 30475, + "train_speed(iter/s)": 0.448377 + }, + { + "acc": 0.6681859, + "epoch": 0.7732115677321156, + "grad_norm": 4.75, + "learning_rate": 7.2092565676449e-06, + "loss": 1.51437483, + "memory(GiB)": 111.15, + "step": 30480, + "train_speed(iter/s)": 0.448416 + }, + { + "acc": 0.66867652, + "epoch": 0.7733384069000507, + "grad_norm": 6.03125, + "learning_rate": 7.208315811032158e-06, + "loss": 1.52819538, + "memory(GiB)": 111.15, + "step": 30485, + "train_speed(iter/s)": 0.448454 + }, + { + "acc": 0.65450106, + "epoch": 0.7734652460679858, + "grad_norm": 5.84375, + "learning_rate": 7.207374957287828e-06, + "loss": 1.60962296, + "memory(GiB)": 111.15, + "step": 30490, + "train_speed(iter/s)": 0.448493 + }, + { + "acc": 0.64216127, + "epoch": 0.7735920852359208, + "grad_norm": 5.8125, + "learning_rate": 7.2064340064532914e-06, + "loss": 1.65317688, + "memory(GiB)": 111.15, + "step": 30495, + "train_speed(iter/s)": 0.448532 + }, + { + "acc": 0.62308502, + "epoch": 0.7737189244038559, + "grad_norm": 5.21875, + "learning_rate": 7.205492958569936e-06, + "loss": 1.69645977, + "memory(GiB)": 111.15, + "step": 30500, + "train_speed(iter/s)": 0.448571 + }, + { + "epoch": 0.7737189244038559, + "eval_acc": 0.6427266502539817, + "eval_loss": 1.6109801530838013, + "eval_runtime": 113.7777, + "eval_samples_per_second": 55.986, + "eval_steps_per_second": 27.993, + "step": 30500 + }, + { + "acc": 0.6566534, + "epoch": 0.773845763571791, + "grad_norm": 4.78125, + "learning_rate": 7.204551813679154e-06, + "loss": 1.7154108, + "memory(GiB)": 111.15, + "step": 30505, + "train_speed(iter/s)": 0.44783 + }, + { + "acc": 0.63998847, + "epoch": 0.773972602739726, + "grad_norm": 5.59375, + "learning_rate": 7.2036105718223405e-06, + "loss": 1.68291721, + "memory(GiB)": 111.15, + "step": 30510, + "train_speed(iter/s)": 0.447869 + }, + { + "acc": 0.64690113, + "epoch": 0.7740994419076611, + "grad_norm": 5.65625, + "learning_rate": 7.202669233040896e-06, + "loss": 1.67814255, + "memory(GiB)": 111.15, + "step": 30515, + "train_speed(iter/s)": 0.447907 + }, + { + "acc": 0.65066261, + "epoch": 0.7742262810755961, + "grad_norm": 5.21875, + "learning_rate": 7.201727797376223e-06, + "loss": 1.67535095, + "memory(GiB)": 111.15, + "step": 30520, + "train_speed(iter/s)": 0.447945 + }, + { + "acc": 0.65824919, + "epoch": 0.7743531202435312, + "grad_norm": 5.1875, + "learning_rate": 7.200786264869732e-06, + "loss": 1.66223335, + "memory(GiB)": 111.15, + "step": 30525, + "train_speed(iter/s)": 0.447984 + }, + { + "acc": 0.6458724, + "epoch": 0.7744799594114663, + "grad_norm": 4.65625, + "learning_rate": 7.199844635562836e-06, + "loss": 1.59313221, + "memory(GiB)": 111.15, + "step": 30530, + "train_speed(iter/s)": 0.448022 + }, + { + "acc": 0.64443059, + "epoch": 0.7746067985794013, + "grad_norm": 4.375, + "learning_rate": 7.19890290949695e-06, + "loss": 1.66826172, + "memory(GiB)": 111.15, + "step": 30535, + "train_speed(iter/s)": 0.448061 + }, + { + "acc": 0.63160334, + "epoch": 0.7747336377473364, + "grad_norm": 5.125, + "learning_rate": 7.197961086713498e-06, + "loss": 1.65784283, + "memory(GiB)": 111.15, + "step": 30540, + "train_speed(iter/s)": 0.448099 + }, + { + "acc": 0.6425684, + "epoch": 0.7748604769152715, + "grad_norm": 5.84375, + "learning_rate": 7.197019167253904e-06, + "loss": 1.63749237, + "memory(GiB)": 111.15, + "step": 30545, + "train_speed(iter/s)": 0.448137 + }, + { + "acc": 0.65910654, + "epoch": 0.7749873160832065, + "grad_norm": 4.1875, + "learning_rate": 7.196077151159597e-06, + "loss": 1.60639668, + "memory(GiB)": 111.15, + "step": 30550, + "train_speed(iter/s)": 0.448175 + }, + { + "acc": 0.6543437, + "epoch": 0.7751141552511416, + "grad_norm": 5.21875, + "learning_rate": 7.195135038472013e-06, + "loss": 1.64946938, + "memory(GiB)": 111.15, + "step": 30555, + "train_speed(iter/s)": 0.448213 + }, + { + "acc": 0.6716464, + "epoch": 0.7752409944190766, + "grad_norm": 5.28125, + "learning_rate": 7.194192829232589e-06, + "loss": 1.61036377, + "memory(GiB)": 111.15, + "step": 30560, + "train_speed(iter/s)": 0.448251 + }, + { + "acc": 0.64679594, + "epoch": 0.7753678335870117, + "grad_norm": 6.5625, + "learning_rate": 7.1932505234827686e-06, + "loss": 1.62901974, + "memory(GiB)": 111.15, + "step": 30565, + "train_speed(iter/s)": 0.44829 + }, + { + "acc": 0.64860187, + "epoch": 0.7754946727549468, + "grad_norm": 5.3125, + "learning_rate": 7.192308121263998e-06, + "loss": 1.68528786, + "memory(GiB)": 111.15, + "step": 30570, + "train_speed(iter/s)": 0.448328 + }, + { + "acc": 0.66241741, + "epoch": 0.7756215119228818, + "grad_norm": 5.28125, + "learning_rate": 7.191365622617728e-06, + "loss": 1.59852791, + "memory(GiB)": 111.15, + "step": 30575, + "train_speed(iter/s)": 0.448366 + }, + { + "acc": 0.64901543, + "epoch": 0.7757483510908169, + "grad_norm": 4.75, + "learning_rate": 7.190423027585414e-06, + "loss": 1.66021786, + "memory(GiB)": 111.15, + "step": 30580, + "train_speed(iter/s)": 0.448404 + }, + { + "acc": 0.65245819, + "epoch": 0.775875190258752, + "grad_norm": 7.0625, + "learning_rate": 7.189480336208516e-06, + "loss": 1.61332855, + "memory(GiB)": 111.15, + "step": 30585, + "train_speed(iter/s)": 0.448442 + }, + { + "acc": 0.64749832, + "epoch": 0.776002029426687, + "grad_norm": 5.71875, + "learning_rate": 7.188537548528498e-06, + "loss": 1.61437302, + "memory(GiB)": 111.15, + "step": 30590, + "train_speed(iter/s)": 0.448481 + }, + { + "acc": 0.645225, + "epoch": 0.776128868594622, + "grad_norm": 5.28125, + "learning_rate": 7.187594664586826e-06, + "loss": 1.63506222, + "memory(GiB)": 111.15, + "step": 30595, + "train_speed(iter/s)": 0.448519 + }, + { + "acc": 0.64895506, + "epoch": 0.776255707762557, + "grad_norm": 4.71875, + "learning_rate": 7.186651684424975e-06, + "loss": 1.68785706, + "memory(GiB)": 111.15, + "step": 30600, + "train_speed(iter/s)": 0.448558 + }, + { + "epoch": 0.776255707762557, + "eval_acc": 0.6427826310734489, + "eval_loss": 1.6108895540237427, + "eval_runtime": 113.6054, + "eval_samples_per_second": 56.071, + "eval_steps_per_second": 28.036, + "step": 30600 + }, + { + "acc": 0.64584837, + "epoch": 0.7763825469304921, + "grad_norm": 7.0625, + "learning_rate": 7.185708608084418e-06, + "loss": 1.66216373, + "memory(GiB)": 111.15, + "step": 30605, + "train_speed(iter/s)": 0.447821 + }, + { + "acc": 0.65720558, + "epoch": 0.7765093860984272, + "grad_norm": 4.75, + "learning_rate": 7.184765435606642e-06, + "loss": 1.61251087, + "memory(GiB)": 111.15, + "step": 30610, + "train_speed(iter/s)": 0.447859 + }, + { + "acc": 0.66171293, + "epoch": 0.7766362252663622, + "grad_norm": 5.46875, + "learning_rate": 7.183822167033124e-06, + "loss": 1.64326477, + "memory(GiB)": 111.15, + "step": 30615, + "train_speed(iter/s)": 0.447897 + }, + { + "acc": 0.64777317, + "epoch": 0.7767630644342973, + "grad_norm": 5.5, + "learning_rate": 7.18287880240536e-06, + "loss": 1.62285957, + "memory(GiB)": 111.15, + "step": 30620, + "train_speed(iter/s)": 0.447936 + }, + { + "acc": 0.67557473, + "epoch": 0.7768899036022324, + "grad_norm": 6.1875, + "learning_rate": 7.1819353417648386e-06, + "loss": 1.56881609, + "memory(GiB)": 111.15, + "step": 30625, + "train_speed(iter/s)": 0.447974 + }, + { + "acc": 0.65477772, + "epoch": 0.7770167427701674, + "grad_norm": 5.15625, + "learning_rate": 7.180991785153059e-06, + "loss": 1.6469408, + "memory(GiB)": 111.15, + "step": 30630, + "train_speed(iter/s)": 0.448013 + }, + { + "acc": 0.65404816, + "epoch": 0.7771435819381025, + "grad_norm": 4.71875, + "learning_rate": 7.180048132611524e-06, + "loss": 1.57132282, + "memory(GiB)": 111.15, + "step": 30635, + "train_speed(iter/s)": 0.448051 + }, + { + "acc": 0.64483247, + "epoch": 0.7772704211060375, + "grad_norm": 8.0, + "learning_rate": 7.17910438418174e-06, + "loss": 1.59463215, + "memory(GiB)": 111.15, + "step": 30640, + "train_speed(iter/s)": 0.44809 + }, + { + "acc": 0.64406567, + "epoch": 0.7773972602739726, + "grad_norm": 5.5625, + "learning_rate": 7.178160539905214e-06, + "loss": 1.67310295, + "memory(GiB)": 111.15, + "step": 30645, + "train_speed(iter/s)": 0.448128 + }, + { + "acc": 0.65861564, + "epoch": 0.7775240994419077, + "grad_norm": 5.375, + "learning_rate": 7.1772165998234645e-06, + "loss": 1.58935957, + "memory(GiB)": 111.15, + "step": 30650, + "train_speed(iter/s)": 0.448166 + }, + { + "acc": 0.64089804, + "epoch": 0.7776509386098427, + "grad_norm": 4.71875, + "learning_rate": 7.176272563978007e-06, + "loss": 1.66168137, + "memory(GiB)": 111.15, + "step": 30655, + "train_speed(iter/s)": 0.448205 + }, + { + "acc": 0.64520826, + "epoch": 0.7777777777777778, + "grad_norm": 5.53125, + "learning_rate": 7.175328432410367e-06, + "loss": 1.63162804, + "memory(GiB)": 111.15, + "step": 30660, + "train_speed(iter/s)": 0.448243 + }, + { + "acc": 0.64064164, + "epoch": 0.7779046169457129, + "grad_norm": 5.71875, + "learning_rate": 7.17438420516207e-06, + "loss": 1.6610733, + "memory(GiB)": 111.15, + "step": 30665, + "train_speed(iter/s)": 0.448281 + }, + { + "acc": 0.6598896, + "epoch": 0.7780314561136479, + "grad_norm": 6.03125, + "learning_rate": 7.173439882274647e-06, + "loss": 1.58247375, + "memory(GiB)": 111.15, + "step": 30670, + "train_speed(iter/s)": 0.448319 + }, + { + "acc": 0.64924078, + "epoch": 0.778158295281583, + "grad_norm": 5.59375, + "learning_rate": 7.172495463789635e-06, + "loss": 1.6560524, + "memory(GiB)": 111.15, + "step": 30675, + "train_speed(iter/s)": 0.448358 + }, + { + "acc": 0.64951935, + "epoch": 0.778285134449518, + "grad_norm": 5.375, + "learning_rate": 7.171550949748574e-06, + "loss": 1.64118195, + "memory(GiB)": 111.15, + "step": 30680, + "train_speed(iter/s)": 0.448397 + }, + { + "acc": 0.65147104, + "epoch": 0.7784119736174531, + "grad_norm": 4.90625, + "learning_rate": 7.170606340193003e-06, + "loss": 1.60523338, + "memory(GiB)": 111.15, + "step": 30685, + "train_speed(iter/s)": 0.448435 + }, + { + "acc": 0.64675589, + "epoch": 0.7785388127853882, + "grad_norm": 5.4375, + "learning_rate": 7.1696616351644786e-06, + "loss": 1.6931942, + "memory(GiB)": 111.15, + "step": 30690, + "train_speed(iter/s)": 0.448474 + }, + { + "acc": 0.66326504, + "epoch": 0.7786656519533232, + "grad_norm": 5.5, + "learning_rate": 7.168716834704546e-06, + "loss": 1.62249146, + "memory(GiB)": 111.15, + "step": 30695, + "train_speed(iter/s)": 0.448512 + }, + { + "acc": 0.64857044, + "epoch": 0.7787924911212583, + "grad_norm": 5.125, + "learning_rate": 7.167771938854766e-06, + "loss": 1.65923405, + "memory(GiB)": 111.15, + "step": 30700, + "train_speed(iter/s)": 0.448551 + }, + { + "epoch": 0.7787924911212583, + "eval_acc": 0.6427930752561853, + "eval_loss": 1.6107133626937866, + "eval_runtime": 113.385, + "eval_samples_per_second": 56.18, + "eval_steps_per_second": 28.09, + "step": 30700 + }, + { + "acc": 0.64862461, + "epoch": 0.7789193302891934, + "grad_norm": 6.5625, + "learning_rate": 7.166826947656696e-06, + "loss": 1.69295807, + "memory(GiB)": 111.15, + "step": 30705, + "train_speed(iter/s)": 0.447818 + }, + { + "acc": 0.65006351, + "epoch": 0.7790461694571283, + "grad_norm": 4.6875, + "learning_rate": 7.165881861151904e-06, + "loss": 1.64642735, + "memory(GiB)": 111.15, + "step": 30710, + "train_speed(iter/s)": 0.447856 + }, + { + "acc": 0.6562645, + "epoch": 0.7791730086250634, + "grad_norm": 7.34375, + "learning_rate": 7.164936679381957e-06, + "loss": 1.626408, + "memory(GiB)": 111.15, + "step": 30715, + "train_speed(iter/s)": 0.447895 + }, + { + "acc": 0.65120454, + "epoch": 0.7792998477929984, + "grad_norm": 5.46875, + "learning_rate": 7.16399140238843e-06, + "loss": 1.68466225, + "memory(GiB)": 111.15, + "step": 30720, + "train_speed(iter/s)": 0.447934 + }, + { + "acc": 0.64600229, + "epoch": 0.7794266869609335, + "grad_norm": 4.75, + "learning_rate": 7.163046030212899e-06, + "loss": 1.65435352, + "memory(GiB)": 111.15, + "step": 30725, + "train_speed(iter/s)": 0.447972 + }, + { + "acc": 0.65804095, + "epoch": 0.7795535261288686, + "grad_norm": 5.09375, + "learning_rate": 7.1621005628969475e-06, + "loss": 1.56945915, + "memory(GiB)": 111.15, + "step": 30730, + "train_speed(iter/s)": 0.44801 + }, + { + "acc": 0.64205022, + "epoch": 0.7796803652968036, + "grad_norm": 6.03125, + "learning_rate": 7.161155000482159e-06, + "loss": 1.70678787, + "memory(GiB)": 111.15, + "step": 30735, + "train_speed(iter/s)": 0.448049 + }, + { + "acc": 0.63220692, + "epoch": 0.7798072044647387, + "grad_norm": 5.375, + "learning_rate": 7.160209343010125e-06, + "loss": 1.64246101, + "memory(GiB)": 111.15, + "step": 30740, + "train_speed(iter/s)": 0.448088 + }, + { + "acc": 0.63886237, + "epoch": 0.7799340436326738, + "grad_norm": 5.25, + "learning_rate": 7.1592635905224386e-06, + "loss": 1.70061092, + "memory(GiB)": 111.15, + "step": 30745, + "train_speed(iter/s)": 0.448126 + }, + { + "acc": 0.64097061, + "epoch": 0.7800608828006088, + "grad_norm": 5.875, + "learning_rate": 7.1583177430606995e-06, + "loss": 1.69046097, + "memory(GiB)": 111.15, + "step": 30750, + "train_speed(iter/s)": 0.448164 + }, + { + "acc": 0.64286985, + "epoch": 0.7801877219685439, + "grad_norm": 5.46875, + "learning_rate": 7.1573718006665095e-06, + "loss": 1.66496773, + "memory(GiB)": 111.15, + "step": 30755, + "train_speed(iter/s)": 0.448202 + }, + { + "acc": 0.63579836, + "epoch": 0.7803145611364789, + "grad_norm": 7.0, + "learning_rate": 7.156425763381477e-06, + "loss": 1.69531326, + "memory(GiB)": 111.15, + "step": 30760, + "train_speed(iter/s)": 0.44824 + }, + { + "acc": 0.65979061, + "epoch": 0.780441400304414, + "grad_norm": 5.5, + "learning_rate": 7.155479631247211e-06, + "loss": 1.60385361, + "memory(GiB)": 111.15, + "step": 30765, + "train_speed(iter/s)": 0.448278 + }, + { + "acc": 0.63736191, + "epoch": 0.7805682394723491, + "grad_norm": 5.09375, + "learning_rate": 7.154533404305327e-06, + "loss": 1.68179321, + "memory(GiB)": 111.15, + "step": 30770, + "train_speed(iter/s)": 0.448316 + }, + { + "acc": 0.6632256, + "epoch": 0.7806950786402841, + "grad_norm": 5.71875, + "learning_rate": 7.153587082597445e-06, + "loss": 1.50969067, + "memory(GiB)": 111.15, + "step": 30775, + "train_speed(iter/s)": 0.448355 + }, + { + "acc": 0.66346331, + "epoch": 0.7808219178082192, + "grad_norm": 5.09375, + "learning_rate": 7.152640666165187e-06, + "loss": 1.51963634, + "memory(GiB)": 111.15, + "step": 30780, + "train_speed(iter/s)": 0.448393 + }, + { + "acc": 0.65551224, + "epoch": 0.7809487569761543, + "grad_norm": 4.90625, + "learning_rate": 7.151694155050184e-06, + "loss": 1.57005634, + "memory(GiB)": 111.15, + "step": 30785, + "train_speed(iter/s)": 0.448431 + }, + { + "acc": 0.64438848, + "epoch": 0.7810755961440893, + "grad_norm": 4.84375, + "learning_rate": 7.150747549294064e-06, + "loss": 1.68739796, + "memory(GiB)": 111.15, + "step": 30790, + "train_speed(iter/s)": 0.44847 + }, + { + "acc": 0.65118885, + "epoch": 0.7812024353120244, + "grad_norm": 6.125, + "learning_rate": 7.149800848938464e-06, + "loss": 1.60922604, + "memory(GiB)": 111.15, + "step": 30795, + "train_speed(iter/s)": 0.448508 + }, + { + "acc": 0.63533068, + "epoch": 0.7813292744799594, + "grad_norm": 5.5, + "learning_rate": 7.1488540540250254e-06, + "loss": 1.67451897, + "memory(GiB)": 111.15, + "step": 30800, + "train_speed(iter/s)": 0.448546 + }, + { + "epoch": 0.7813292744799594, + "eval_acc": 0.6427663381483801, + "eval_loss": 1.6108155250549316, + "eval_runtime": 114.2335, + "eval_samples_per_second": 55.763, + "eval_steps_per_second": 27.881, + "step": 30800 + }, + { + "acc": 0.64741917, + "epoch": 0.7814561136478945, + "grad_norm": 5.0, + "learning_rate": 7.14790716459539e-06, + "loss": 1.62743549, + "memory(GiB)": 111.15, + "step": 30805, + "train_speed(iter/s)": 0.44781 + }, + { + "acc": 0.65753269, + "epoch": 0.7815829528158296, + "grad_norm": 5.0, + "learning_rate": 7.146960180691209e-06, + "loss": 1.57152958, + "memory(GiB)": 111.15, + "step": 30810, + "train_speed(iter/s)": 0.447847 + }, + { + "acc": 0.6423707, + "epoch": 0.7817097919837646, + "grad_norm": 5.59375, + "learning_rate": 7.146013102354133e-06, + "loss": 1.6205471, + "memory(GiB)": 111.15, + "step": 30815, + "train_speed(iter/s)": 0.447886 + }, + { + "acc": 0.66003194, + "epoch": 0.7818366311516997, + "grad_norm": 5.15625, + "learning_rate": 7.145065929625821e-06, + "loss": 1.57598858, + "memory(GiB)": 111.15, + "step": 30820, + "train_speed(iter/s)": 0.447924 + }, + { + "acc": 0.65117698, + "epoch": 0.7819634703196348, + "grad_norm": 5.125, + "learning_rate": 7.1441186625479304e-06, + "loss": 1.65818958, + "memory(GiB)": 111.15, + "step": 30825, + "train_speed(iter/s)": 0.447962 + }, + { + "acc": 0.63875875, + "epoch": 0.7820903094875697, + "grad_norm": 6.15625, + "learning_rate": 7.143171301162131e-06, + "loss": 1.67396126, + "memory(GiB)": 111.15, + "step": 30830, + "train_speed(iter/s)": 0.448 + }, + { + "acc": 0.62688546, + "epoch": 0.7822171486555048, + "grad_norm": 5.71875, + "learning_rate": 7.142223845510086e-06, + "loss": 1.74163609, + "memory(GiB)": 111.15, + "step": 30835, + "train_speed(iter/s)": 0.448038 + }, + { + "acc": 0.63917665, + "epoch": 0.7823439878234398, + "grad_norm": 6.0625, + "learning_rate": 7.1412762956334746e-06, + "loss": 1.66056442, + "memory(GiB)": 111.15, + "step": 30840, + "train_speed(iter/s)": 0.448077 + }, + { + "acc": 0.65063796, + "epoch": 0.7824708269913749, + "grad_norm": 4.59375, + "learning_rate": 7.140328651573969e-06, + "loss": 1.62079926, + "memory(GiB)": 111.15, + "step": 30845, + "train_speed(iter/s)": 0.448115 + }, + { + "acc": 0.64573078, + "epoch": 0.78259766615931, + "grad_norm": 5.15625, + "learning_rate": 7.139380913373255e-06, + "loss": 1.64585686, + "memory(GiB)": 111.15, + "step": 30850, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.6448422, + "epoch": 0.782724505327245, + "grad_norm": 4.9375, + "learning_rate": 7.138433081073017e-06, + "loss": 1.68308144, + "memory(GiB)": 111.15, + "step": 30855, + "train_speed(iter/s)": 0.448192 + }, + { + "acc": 0.66168218, + "epoch": 0.7828513444951801, + "grad_norm": 5.15625, + "learning_rate": 7.137485154714945e-06, + "loss": 1.58274536, + "memory(GiB)": 111.15, + "step": 30860, + "train_speed(iter/s)": 0.448231 + }, + { + "acc": 0.66205597, + "epoch": 0.7829781836631152, + "grad_norm": 5.0625, + "learning_rate": 7.1365371343407304e-06, + "loss": 1.61109924, + "memory(GiB)": 111.15, + "step": 30865, + "train_speed(iter/s)": 0.448269 + }, + { + "acc": 0.66781082, + "epoch": 0.7831050228310502, + "grad_norm": 5.6875, + "learning_rate": 7.135589019992076e-06, + "loss": 1.61279774, + "memory(GiB)": 111.15, + "step": 30870, + "train_speed(iter/s)": 0.448308 + }, + { + "acc": 0.66033773, + "epoch": 0.7832318619989853, + "grad_norm": 5.9375, + "learning_rate": 7.134640811710681e-06, + "loss": 1.60158863, + "memory(GiB)": 111.15, + "step": 30875, + "train_speed(iter/s)": 0.448346 + }, + { + "acc": 0.63533506, + "epoch": 0.7833587011669203, + "grad_norm": 6.0, + "learning_rate": 7.133692509538253e-06, + "loss": 1.62528992, + "memory(GiB)": 111.15, + "step": 30880, + "train_speed(iter/s)": 0.448384 + }, + { + "acc": 0.64281855, + "epoch": 0.7834855403348554, + "grad_norm": 5.46875, + "learning_rate": 7.132744113516502e-06, + "loss": 1.70961037, + "memory(GiB)": 111.15, + "step": 30885, + "train_speed(iter/s)": 0.448423 + }, + { + "acc": 0.6659873, + "epoch": 0.7836123795027905, + "grad_norm": 5.4375, + "learning_rate": 7.1317956236871436e-06, + "loss": 1.59920311, + "memory(GiB)": 111.15, + "step": 30890, + "train_speed(iter/s)": 0.448461 + }, + { + "acc": 0.66405911, + "epoch": 0.7837392186707255, + "grad_norm": 5.15625, + "learning_rate": 7.130847040091893e-06, + "loss": 1.60594406, + "memory(GiB)": 111.15, + "step": 30895, + "train_speed(iter/s)": 0.448499 + }, + { + "acc": 0.64323874, + "epoch": 0.7838660578386606, + "grad_norm": 5.25, + "learning_rate": 7.1298983627724795e-06, + "loss": 1.62966881, + "memory(GiB)": 111.15, + "step": 30900, + "train_speed(iter/s)": 0.448537 + }, + { + "epoch": 0.7838660578386606, + "eval_acc": 0.642777200098426, + "eval_loss": 1.610750436782837, + "eval_runtime": 111.3462, + "eval_samples_per_second": 57.209, + "eval_steps_per_second": 28.604, + "step": 30900 + }, + { + "acc": 0.63616338, + "epoch": 0.7839928970065957, + "grad_norm": 6.4375, + "learning_rate": 7.128949591770624e-06, + "loss": 1.70839043, + "memory(GiB)": 111.15, + "step": 30905, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.66298161, + "epoch": 0.7841197361745307, + "grad_norm": 5.3125, + "learning_rate": 7.128000727128063e-06, + "loss": 1.6531353, + "memory(GiB)": 111.15, + "step": 30910, + "train_speed(iter/s)": 0.44786 + }, + { + "acc": 0.64129581, + "epoch": 0.7842465753424658, + "grad_norm": 5.59375, + "learning_rate": 7.127051768886527e-06, + "loss": 1.66708755, + "memory(GiB)": 111.15, + "step": 30915, + "train_speed(iter/s)": 0.447899 + }, + { + "acc": 0.66942554, + "epoch": 0.7843734145104008, + "grad_norm": 5.8125, + "learning_rate": 7.126102717087758e-06, + "loss": 1.58191547, + "memory(GiB)": 111.15, + "step": 30920, + "train_speed(iter/s)": 0.447937 + }, + { + "acc": 0.64453368, + "epoch": 0.7845002536783359, + "grad_norm": 4.9375, + "learning_rate": 7.1251535717735e-06, + "loss": 1.67702789, + "memory(GiB)": 111.15, + "step": 30925, + "train_speed(iter/s)": 0.447975 + }, + { + "acc": 0.63794889, + "epoch": 0.784627092846271, + "grad_norm": 6.96875, + "learning_rate": 7.1242043329854995e-06, + "loss": 1.68617096, + "memory(GiB)": 111.15, + "step": 30930, + "train_speed(iter/s)": 0.448013 + }, + { + "acc": 0.64637618, + "epoch": 0.784753932014206, + "grad_norm": 5.5, + "learning_rate": 7.123255000765508e-06, + "loss": 1.62841606, + "memory(GiB)": 111.15, + "step": 30935, + "train_speed(iter/s)": 0.448052 + }, + { + "acc": 0.64883261, + "epoch": 0.7848807711821411, + "grad_norm": 6.1875, + "learning_rate": 7.122305575155283e-06, + "loss": 1.63535442, + "memory(GiB)": 111.15, + "step": 30940, + "train_speed(iter/s)": 0.44809 + }, + { + "acc": 0.65482521, + "epoch": 0.7850076103500762, + "grad_norm": 5.03125, + "learning_rate": 7.121356056196582e-06, + "loss": 1.6417881, + "memory(GiB)": 111.15, + "step": 30945, + "train_speed(iter/s)": 0.448128 + }, + { + "acc": 0.67526751, + "epoch": 0.7851344495180111, + "grad_norm": 4.625, + "learning_rate": 7.1204064439311715e-06, + "loss": 1.57394161, + "memory(GiB)": 111.15, + "step": 30950, + "train_speed(iter/s)": 0.448167 + }, + { + "acc": 0.65488586, + "epoch": 0.7852612886859462, + "grad_norm": 6.0625, + "learning_rate": 7.119456738400818e-06, + "loss": 1.57184706, + "memory(GiB)": 111.15, + "step": 30955, + "train_speed(iter/s)": 0.448205 + }, + { + "acc": 0.6617012, + "epoch": 0.7853881278538812, + "grad_norm": 5.21875, + "learning_rate": 7.118506939647295e-06, + "loss": 1.51508751, + "memory(GiB)": 111.15, + "step": 30960, + "train_speed(iter/s)": 0.448243 + }, + { + "acc": 0.64925838, + "epoch": 0.7855149670218163, + "grad_norm": 5.78125, + "learning_rate": 7.1175570477123776e-06, + "loss": 1.63446712, + "memory(GiB)": 111.15, + "step": 30965, + "train_speed(iter/s)": 0.448282 + }, + { + "acc": 0.64819326, + "epoch": 0.7856418061897514, + "grad_norm": 5.5, + "learning_rate": 7.116607062637848e-06, + "loss": 1.64832993, + "memory(GiB)": 111.15, + "step": 30970, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.65542517, + "epoch": 0.7857686453576864, + "grad_norm": 5.8125, + "learning_rate": 7.115656984465489e-06, + "loss": 1.58983278, + "memory(GiB)": 111.15, + "step": 30975, + "train_speed(iter/s)": 0.448358 + }, + { + "acc": 0.64042444, + "epoch": 0.7858954845256215, + "grad_norm": 5.71875, + "learning_rate": 7.114706813237091e-06, + "loss": 1.71252022, + "memory(GiB)": 111.15, + "step": 30980, + "train_speed(iter/s)": 0.448396 + }, + { + "acc": 0.65176029, + "epoch": 0.7860223236935566, + "grad_norm": 4.59375, + "learning_rate": 7.1137565489944445e-06, + "loss": 1.62491436, + "memory(GiB)": 111.15, + "step": 30985, + "train_speed(iter/s)": 0.448435 + }, + { + "acc": 0.65771918, + "epoch": 0.7861491628614916, + "grad_norm": 5.21875, + "learning_rate": 7.112806191779349e-06, + "loss": 1.59344769, + "memory(GiB)": 111.15, + "step": 30990, + "train_speed(iter/s)": 0.448473 + }, + { + "acc": 0.64596863, + "epoch": 0.7862760020294267, + "grad_norm": 5.21875, + "learning_rate": 7.111855741633603e-06, + "loss": 1.65545101, + "memory(GiB)": 111.15, + "step": 30995, + "train_speed(iter/s)": 0.448511 + }, + { + "acc": 0.65507011, + "epoch": 0.7864028411973617, + "grad_norm": 6.59375, + "learning_rate": 7.1109051985990145e-06, + "loss": 1.63063011, + "memory(GiB)": 111.15, + "step": 31000, + "train_speed(iter/s)": 0.448549 + }, + { + "epoch": 0.7864028411973617, + "eval_acc": 0.642796417394661, + "eval_loss": 1.6107975244522095, + "eval_runtime": 114.2363, + "eval_samples_per_second": 55.762, + "eval_steps_per_second": 27.881, + "step": 31000 + }, + { + "acc": 0.64518003, + "epoch": 0.7865296803652968, + "grad_norm": 4.625, + "learning_rate": 7.109954562717389e-06, + "loss": 1.64898396, + "memory(GiB)": 111.15, + "step": 31005, + "train_speed(iter/s)": 0.447818 + }, + { + "acc": 0.66206203, + "epoch": 0.7866565195332319, + "grad_norm": 6.09375, + "learning_rate": 7.109003834030543e-06, + "loss": 1.62006245, + "memory(GiB)": 111.15, + "step": 31010, + "train_speed(iter/s)": 0.447856 + }, + { + "acc": 0.66016941, + "epoch": 0.7867833587011669, + "grad_norm": 7.0, + "learning_rate": 7.108053012580291e-06, + "loss": 1.57476673, + "memory(GiB)": 111.15, + "step": 31015, + "train_speed(iter/s)": 0.447894 + }, + { + "acc": 0.64697161, + "epoch": 0.786910197869102, + "grad_norm": 4.9375, + "learning_rate": 7.107102098408457e-06, + "loss": 1.57776089, + "memory(GiB)": 111.15, + "step": 31020, + "train_speed(iter/s)": 0.447932 + }, + { + "acc": 0.65201693, + "epoch": 0.7870370370370371, + "grad_norm": 5.1875, + "learning_rate": 7.106151091556865e-06, + "loss": 1.63846264, + "memory(GiB)": 111.15, + "step": 31025, + "train_speed(iter/s)": 0.447969 + }, + { + "acc": 0.65150242, + "epoch": 0.7871638762049721, + "grad_norm": 5.09375, + "learning_rate": 7.105199992067344e-06, + "loss": 1.62783928, + "memory(GiB)": 111.15, + "step": 31030, + "train_speed(iter/s)": 0.448007 + }, + { + "acc": 0.64858766, + "epoch": 0.7872907153729072, + "grad_norm": 5.03125, + "learning_rate": 7.1042487999817275e-06, + "loss": 1.6289053, + "memory(GiB)": 111.15, + "step": 31035, + "train_speed(iter/s)": 0.448045 + }, + { + "acc": 0.65656414, + "epoch": 0.7874175545408422, + "grad_norm": 5.59375, + "learning_rate": 7.103297515341857e-06, + "loss": 1.59551468, + "memory(GiB)": 111.15, + "step": 31040, + "train_speed(iter/s)": 0.448083 + }, + { + "acc": 0.65150142, + "epoch": 0.7875443937087773, + "grad_norm": 6.09375, + "learning_rate": 7.1023461381895685e-06, + "loss": 1.61672211, + "memory(GiB)": 111.15, + "step": 31045, + "train_speed(iter/s)": 0.448121 + }, + { + "acc": 0.64974575, + "epoch": 0.7876712328767124, + "grad_norm": 5.03125, + "learning_rate": 7.1013946685667125e-06, + "loss": 1.63908215, + "memory(GiB)": 111.15, + "step": 31050, + "train_speed(iter/s)": 0.448159 + }, + { + "acc": 0.66036777, + "epoch": 0.7877980720446474, + "grad_norm": 6.0625, + "learning_rate": 7.100443106515135e-06, + "loss": 1.59898205, + "memory(GiB)": 111.15, + "step": 31055, + "train_speed(iter/s)": 0.448197 + }, + { + "acc": 0.66473131, + "epoch": 0.7879249112125825, + "grad_norm": 6.28125, + "learning_rate": 7.099491452076693e-06, + "loss": 1.62502632, + "memory(GiB)": 111.15, + "step": 31060, + "train_speed(iter/s)": 0.448235 + }, + { + "acc": 0.64832411, + "epoch": 0.7880517503805176, + "grad_norm": 7.84375, + "learning_rate": 7.098539705293242e-06, + "loss": 1.65329247, + "memory(GiB)": 111.15, + "step": 31065, + "train_speed(iter/s)": 0.448273 + }, + { + "acc": 0.64531212, + "epoch": 0.7881785895484525, + "grad_norm": 6.375, + "learning_rate": 7.097587866206647e-06, + "loss": 1.62400188, + "memory(GiB)": 111.15, + "step": 31070, + "train_speed(iter/s)": 0.448311 + }, + { + "acc": 0.66134143, + "epoch": 0.7883054287163876, + "grad_norm": 6.625, + "learning_rate": 7.096635934858772e-06, + "loss": 1.61105042, + "memory(GiB)": 111.15, + "step": 31075, + "train_speed(iter/s)": 0.448348 + }, + { + "acc": 0.65088539, + "epoch": 0.7884322678843226, + "grad_norm": 5.25, + "learning_rate": 7.095683911291488e-06, + "loss": 1.64476032, + "memory(GiB)": 111.15, + "step": 31080, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.64607506, + "epoch": 0.7885591070522577, + "grad_norm": 5.6875, + "learning_rate": 7.0947317955466686e-06, + "loss": 1.73367271, + "memory(GiB)": 111.15, + "step": 31085, + "train_speed(iter/s)": 0.448425 + }, + { + "acc": 0.64504452, + "epoch": 0.7886859462201928, + "grad_norm": 5.25, + "learning_rate": 7.093779587666193e-06, + "loss": 1.64341621, + "memory(GiB)": 111.15, + "step": 31090, + "train_speed(iter/s)": 0.448463 + }, + { + "acc": 0.65902405, + "epoch": 0.7888127853881278, + "grad_norm": 5.0625, + "learning_rate": 7.092827287691943e-06, + "loss": 1.56710167, + "memory(GiB)": 111.15, + "step": 31095, + "train_speed(iter/s)": 0.448501 + }, + { + "acc": 0.65478497, + "epoch": 0.7889396245560629, + "grad_norm": 5.40625, + "learning_rate": 7.091874895665806e-06, + "loss": 1.63734455, + "memory(GiB)": 111.15, + "step": 31100, + "train_speed(iter/s)": 0.448539 + }, + { + "epoch": 0.7889396245560629, + "eval_acc": 0.6429355339087103, + "eval_loss": 1.6107193231582642, + "eval_runtime": 114.0963, + "eval_samples_per_second": 55.83, + "eval_steps_per_second": 27.915, + "step": 31100 + }, + { + "acc": 0.65356402, + "epoch": 0.789066463723998, + "grad_norm": 6.34375, + "learning_rate": 7.09092241162967e-06, + "loss": 1.63494625, + "memory(GiB)": 111.15, + "step": 31105, + "train_speed(iter/s)": 0.447811 + }, + { + "acc": 0.65918703, + "epoch": 0.789193302891933, + "grad_norm": 4.90625, + "learning_rate": 7.089969835625432e-06, + "loss": 1.63414268, + "memory(GiB)": 111.15, + "step": 31110, + "train_speed(iter/s)": 0.447848 + }, + { + "acc": 0.63956089, + "epoch": 0.7893201420598681, + "grad_norm": 5.78125, + "learning_rate": 7.089017167694988e-06, + "loss": 1.65694809, + "memory(GiB)": 111.15, + "step": 31115, + "train_speed(iter/s)": 0.447886 + }, + { + "acc": 0.64109316, + "epoch": 0.7894469812278031, + "grad_norm": 5.5625, + "learning_rate": 7.088064407880244e-06, + "loss": 1.68776035, + "memory(GiB)": 111.15, + "step": 31120, + "train_speed(iter/s)": 0.447923 + }, + { + "acc": 0.65231352, + "epoch": 0.7895738203957382, + "grad_norm": 5.25, + "learning_rate": 7.087111556223103e-06, + "loss": 1.64642849, + "memory(GiB)": 111.15, + "step": 31125, + "train_speed(iter/s)": 0.447961 + }, + { + "acc": 0.64406476, + "epoch": 0.7897006595636733, + "grad_norm": 5.90625, + "learning_rate": 7.08615861276548e-06, + "loss": 1.67204704, + "memory(GiB)": 111.15, + "step": 31130, + "train_speed(iter/s)": 0.447999 + }, + { + "acc": 0.63983431, + "epoch": 0.7898274987316083, + "grad_norm": 4.8125, + "learning_rate": 7.085205577549285e-06, + "loss": 1.62904968, + "memory(GiB)": 111.15, + "step": 31135, + "train_speed(iter/s)": 0.448036 + }, + { + "acc": 0.65634842, + "epoch": 0.7899543378995434, + "grad_norm": 4.34375, + "learning_rate": 7.08425245061644e-06, + "loss": 1.61398563, + "memory(GiB)": 111.15, + "step": 31140, + "train_speed(iter/s)": 0.448074 + }, + { + "acc": 0.64045734, + "epoch": 0.7900811770674785, + "grad_norm": 5.125, + "learning_rate": 7.083299232008867e-06, + "loss": 1.66314354, + "memory(GiB)": 111.15, + "step": 31145, + "train_speed(iter/s)": 0.448111 + }, + { + "acc": 0.65300369, + "epoch": 0.7902080162354135, + "grad_norm": 5.59375, + "learning_rate": 7.082345921768492e-06, + "loss": 1.56575985, + "memory(GiB)": 111.15, + "step": 31150, + "train_speed(iter/s)": 0.448149 + }, + { + "acc": 0.65374107, + "epoch": 0.7903348554033486, + "grad_norm": 5.71875, + "learning_rate": 7.0813925199372455e-06, + "loss": 1.61517086, + "memory(GiB)": 111.15, + "step": 31155, + "train_speed(iter/s)": 0.448186 + }, + { + "acc": 0.62918968, + "epoch": 0.7904616945712836, + "grad_norm": 4.6875, + "learning_rate": 7.080439026557065e-06, + "loss": 1.64710522, + "memory(GiB)": 111.15, + "step": 31160, + "train_speed(iter/s)": 0.448224 + }, + { + "acc": 0.66331806, + "epoch": 0.7905885337392187, + "grad_norm": 4.53125, + "learning_rate": 7.079485441669887e-06, + "loss": 1.51792068, + "memory(GiB)": 111.15, + "step": 31165, + "train_speed(iter/s)": 0.448261 + }, + { + "acc": 0.63385506, + "epoch": 0.7907153729071538, + "grad_norm": 5.3125, + "learning_rate": 7.0785317653176534e-06, + "loss": 1.65183105, + "memory(GiB)": 111.15, + "step": 31170, + "train_speed(iter/s)": 0.448299 + }, + { + "acc": 0.63161554, + "epoch": 0.7908422120750888, + "grad_norm": 5.59375, + "learning_rate": 7.077577997542316e-06, + "loss": 1.67356949, + "memory(GiB)": 111.15, + "step": 31175, + "train_speed(iter/s)": 0.448336 + }, + { + "acc": 0.64063845, + "epoch": 0.7909690512430239, + "grad_norm": 5.59375, + "learning_rate": 7.0766241383858195e-06, + "loss": 1.74691181, + "memory(GiB)": 111.15, + "step": 31180, + "train_speed(iter/s)": 0.448374 + }, + { + "acc": 0.6450038, + "epoch": 0.791095890410959, + "grad_norm": 5.21875, + "learning_rate": 7.075670187890123e-06, + "loss": 1.64314842, + "memory(GiB)": 111.15, + "step": 31185, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.65761538, + "epoch": 0.791222729578894, + "grad_norm": 5.25, + "learning_rate": 7.0747161460971845e-06, + "loss": 1.612117, + "memory(GiB)": 111.15, + "step": 31190, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.64225979, + "epoch": 0.791349568746829, + "grad_norm": 6.21875, + "learning_rate": 7.073762013048966e-06, + "loss": 1.71116943, + "memory(GiB)": 111.15, + "step": 31195, + "train_speed(iter/s)": 0.448487 + }, + { + "acc": 0.64990358, + "epoch": 0.791476407914764, + "grad_norm": 5.21875, + "learning_rate": 7.072807788787437e-06, + "loss": 1.59435291, + "memory(GiB)": 111.15, + "step": 31200, + "train_speed(iter/s)": 0.448525 + }, + { + "epoch": 0.791476407914764, + "eval_acc": 0.6428122925524205, + "eval_loss": 1.610496163368225, + "eval_runtime": 113.6278, + "eval_samples_per_second": 56.06, + "eval_steps_per_second": 28.03, + "step": 31200 + }, + { + "acc": 0.6481596, + "epoch": 0.7916032470826991, + "grad_norm": 6.09375, + "learning_rate": 7.071853473354566e-06, + "loss": 1.64403172, + "memory(GiB)": 111.15, + "step": 31205, + "train_speed(iter/s)": 0.447802 + }, + { + "acc": 0.65084019, + "epoch": 0.7917300862506342, + "grad_norm": 4.78125, + "learning_rate": 7.070899066792329e-06, + "loss": 1.6133461, + "memory(GiB)": 111.15, + "step": 31210, + "train_speed(iter/s)": 0.44784 + }, + { + "acc": 0.65622911, + "epoch": 0.7918569254185692, + "grad_norm": 5.3125, + "learning_rate": 7.069944569142706e-06, + "loss": 1.60215988, + "memory(GiB)": 111.15, + "step": 31215, + "train_speed(iter/s)": 0.447878 + }, + { + "acc": 0.64859939, + "epoch": 0.7919837645865043, + "grad_norm": 5.90625, + "learning_rate": 7.068989980447679e-06, + "loss": 1.67581291, + "memory(GiB)": 111.15, + "step": 31220, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.65272231, + "epoch": 0.7921106037544394, + "grad_norm": 5.96875, + "learning_rate": 7.068035300749237e-06, + "loss": 1.6181881, + "memory(GiB)": 111.15, + "step": 31225, + "train_speed(iter/s)": 0.447953 + }, + { + "acc": 0.6427022, + "epoch": 0.7922374429223744, + "grad_norm": 4.71875, + "learning_rate": 7.067080530089366e-06, + "loss": 1.70042667, + "memory(GiB)": 111.15, + "step": 31230, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.64109869, + "epoch": 0.7923642820903095, + "grad_norm": 5.46875, + "learning_rate": 7.066125668510067e-06, + "loss": 1.68296776, + "memory(GiB)": 111.15, + "step": 31235, + "train_speed(iter/s)": 0.448028 + }, + { + "acc": 0.64118376, + "epoch": 0.7924911212582445, + "grad_norm": 5.40625, + "learning_rate": 7.065170716053336e-06, + "loss": 1.65491753, + "memory(GiB)": 111.15, + "step": 31240, + "train_speed(iter/s)": 0.448066 + }, + { + "acc": 0.64315076, + "epoch": 0.7926179604261796, + "grad_norm": 5.9375, + "learning_rate": 7.064215672761175e-06, + "loss": 1.65513268, + "memory(GiB)": 111.15, + "step": 31245, + "train_speed(iter/s)": 0.448104 + }, + { + "acc": 0.66230831, + "epoch": 0.7927447995941147, + "grad_norm": 4.90625, + "learning_rate": 7.063260538675594e-06, + "loss": 1.59357901, + "memory(GiB)": 111.15, + "step": 31250, + "train_speed(iter/s)": 0.448141 + }, + { + "acc": 0.65274773, + "epoch": 0.7928716387620497, + "grad_norm": 5.125, + "learning_rate": 7.062305313838601e-06, + "loss": 1.59322681, + "memory(GiB)": 111.15, + "step": 31255, + "train_speed(iter/s)": 0.448179 + }, + { + "acc": 0.64179802, + "epoch": 0.7929984779299848, + "grad_norm": 5.40625, + "learning_rate": 7.061349998292215e-06, + "loss": 1.69425545, + "memory(GiB)": 111.15, + "step": 31260, + "train_speed(iter/s)": 0.448217 + }, + { + "acc": 0.64792447, + "epoch": 0.7931253170979199, + "grad_norm": 5.3125, + "learning_rate": 7.060394592078452e-06, + "loss": 1.64353371, + "memory(GiB)": 111.15, + "step": 31265, + "train_speed(iter/s)": 0.448254 + }, + { + "acc": 0.64805927, + "epoch": 0.7932521562658549, + "grad_norm": 5.25, + "learning_rate": 7.0594390952393365e-06, + "loss": 1.63469429, + "memory(GiB)": 111.15, + "step": 31270, + "train_speed(iter/s)": 0.448292 + }, + { + "acc": 0.6365171, + "epoch": 0.79337899543379, + "grad_norm": 5.1875, + "learning_rate": 7.058483507816894e-06, + "loss": 1.66427002, + "memory(GiB)": 111.15, + "step": 31275, + "train_speed(iter/s)": 0.44833 + }, + { + "acc": 0.65665178, + "epoch": 0.793505834601725, + "grad_norm": 4.53125, + "learning_rate": 7.057527829853157e-06, + "loss": 1.58186893, + "memory(GiB)": 111.15, + "step": 31280, + "train_speed(iter/s)": 0.448368 + }, + { + "acc": 0.64379082, + "epoch": 0.7936326737696601, + "grad_norm": 6.4375, + "learning_rate": 7.056572061390159e-06, + "loss": 1.65864563, + "memory(GiB)": 111.15, + "step": 31285, + "train_speed(iter/s)": 0.448405 + }, + { + "acc": 0.63721948, + "epoch": 0.7937595129375952, + "grad_norm": 4.78125, + "learning_rate": 7.055616202469939e-06, + "loss": 1.66354313, + "memory(GiB)": 111.15, + "step": 31290, + "train_speed(iter/s)": 0.448443 + }, + { + "acc": 0.65148191, + "epoch": 0.7938863521055302, + "grad_norm": 6.34375, + "learning_rate": 7.054660253134543e-06, + "loss": 1.63543282, + "memory(GiB)": 111.15, + "step": 31295, + "train_speed(iter/s)": 0.448481 + }, + { + "acc": 0.64979401, + "epoch": 0.7940131912734653, + "grad_norm": 5.21875, + "learning_rate": 7.053704213426015e-06, + "loss": 1.70438004, + "memory(GiB)": 111.15, + "step": 31300, + "train_speed(iter/s)": 0.448519 + }, + { + "epoch": 0.7940131912734653, + "eval_acc": 0.6428298387794176, + "eval_loss": 1.6106901168823242, + "eval_runtime": 113.4291, + "eval_samples_per_second": 56.158, + "eval_steps_per_second": 28.079, + "step": 31300 + }, + { + "acc": 0.64462376, + "epoch": 0.7941400304414004, + "grad_norm": 5.34375, + "learning_rate": 7.052748083386406e-06, + "loss": 1.67642441, + "memory(GiB)": 111.15, + "step": 31305, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.66111736, + "epoch": 0.7942668696093353, + "grad_norm": 5.75, + "learning_rate": 7.051791863057772e-06, + "loss": 1.6098505, + "memory(GiB)": 111.15, + "step": 31310, + "train_speed(iter/s)": 0.447836 + }, + { + "acc": 0.64311485, + "epoch": 0.7943937087772704, + "grad_norm": 6.0, + "learning_rate": 7.050835552482171e-06, + "loss": 1.62449951, + "memory(GiB)": 111.15, + "step": 31315, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.65576372, + "epoch": 0.7945205479452054, + "grad_norm": 5.96875, + "learning_rate": 7.049879151701666e-06, + "loss": 1.6281929, + "memory(GiB)": 111.15, + "step": 31320, + "train_speed(iter/s)": 0.447912 + }, + { + "acc": 0.63418741, + "epoch": 0.7946473871131405, + "grad_norm": 5.4375, + "learning_rate": 7.048922660758324e-06, + "loss": 1.63147793, + "memory(GiB)": 111.15, + "step": 31325, + "train_speed(iter/s)": 0.447949 + }, + { + "acc": 0.6562851, + "epoch": 0.7947742262810756, + "grad_norm": 5.28125, + "learning_rate": 7.047966079694215e-06, + "loss": 1.55582857, + "memory(GiB)": 111.15, + "step": 31330, + "train_speed(iter/s)": 0.447986 + }, + { + "acc": 0.65491343, + "epoch": 0.7949010654490106, + "grad_norm": 6.03125, + "learning_rate": 7.047009408551414e-06, + "loss": 1.56434479, + "memory(GiB)": 111.15, + "step": 31335, + "train_speed(iter/s)": 0.448024 + }, + { + "acc": 0.64060459, + "epoch": 0.7950279046169457, + "grad_norm": 5.53125, + "learning_rate": 7.046052647372002e-06, + "loss": 1.64723206, + "memory(GiB)": 111.15, + "step": 31340, + "train_speed(iter/s)": 0.448061 + }, + { + "acc": 0.65123663, + "epoch": 0.7951547437848808, + "grad_norm": 5.78125, + "learning_rate": 7.045095796198057e-06, + "loss": 1.66899834, + "memory(GiB)": 111.15, + "step": 31345, + "train_speed(iter/s)": 0.448098 + }, + { + "acc": 0.63876448, + "epoch": 0.7952815829528158, + "grad_norm": 5.34375, + "learning_rate": 7.044138855071671e-06, + "loss": 1.75702152, + "memory(GiB)": 111.15, + "step": 31350, + "train_speed(iter/s)": 0.448136 + }, + { + "acc": 0.64831133, + "epoch": 0.7954084221207509, + "grad_norm": 5.78125, + "learning_rate": 7.043181824034929e-06, + "loss": 1.65338669, + "memory(GiB)": 111.15, + "step": 31355, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.66197309, + "epoch": 0.7955352612886859, + "grad_norm": 6.21875, + "learning_rate": 7.042224703129929e-06, + "loss": 1.59324617, + "memory(GiB)": 111.15, + "step": 31360, + "train_speed(iter/s)": 0.448212 + }, + { + "acc": 0.64458365, + "epoch": 0.795662100456621, + "grad_norm": 5.6875, + "learning_rate": 7.0412674923987705e-06, + "loss": 1.68621178, + "memory(GiB)": 111.15, + "step": 31365, + "train_speed(iter/s)": 0.44825 + }, + { + "acc": 0.65985699, + "epoch": 0.7957889396245561, + "grad_norm": 6.3125, + "learning_rate": 7.040310191883552e-06, + "loss": 1.55742931, + "memory(GiB)": 111.15, + "step": 31370, + "train_speed(iter/s)": 0.448287 + }, + { + "acc": 0.6507226, + "epoch": 0.7959157787924911, + "grad_norm": 5.25, + "learning_rate": 7.039352801626383e-06, + "loss": 1.60855865, + "memory(GiB)": 111.15, + "step": 31375, + "train_speed(iter/s)": 0.448325 + }, + { + "acc": 0.67202559, + "epoch": 0.7960426179604262, + "grad_norm": 5.21875, + "learning_rate": 7.0383953216693725e-06, + "loss": 1.48202114, + "memory(GiB)": 111.15, + "step": 31380, + "train_speed(iter/s)": 0.448364 + }, + { + "acc": 0.66186886, + "epoch": 0.7961694571283613, + "grad_norm": 5.625, + "learning_rate": 7.037437752054635e-06, + "loss": 1.64839172, + "memory(GiB)": 111.15, + "step": 31385, + "train_speed(iter/s)": 0.448401 + }, + { + "acc": 0.65394578, + "epoch": 0.7962962962962963, + "grad_norm": 5.75, + "learning_rate": 7.036480092824288e-06, + "loss": 1.60997868, + "memory(GiB)": 111.15, + "step": 31390, + "train_speed(iter/s)": 0.448439 + }, + { + "acc": 0.6463728, + "epoch": 0.7964231354642314, + "grad_norm": 5.6875, + "learning_rate": 7.035522344020455e-06, + "loss": 1.62575378, + "memory(GiB)": 111.15, + "step": 31395, + "train_speed(iter/s)": 0.448477 + }, + { + "acc": 0.63021846, + "epoch": 0.7965499746321664, + "grad_norm": 5.59375, + "learning_rate": 7.034564505685262e-06, + "loss": 1.60279312, + "memory(GiB)": 111.15, + "step": 31400, + "train_speed(iter/s)": 0.448514 + }, + { + "epoch": 0.7965499746321664, + "eval_acc": 0.6428628423968648, + "eval_loss": 1.610276699066162, + "eval_runtime": 112.9567, + "eval_samples_per_second": 56.393, + "eval_steps_per_second": 28.197, + "step": 31400 + }, + { + "acc": 0.64464951, + "epoch": 0.7966768138001015, + "grad_norm": 5.96875, + "learning_rate": 7.0336065778608365e-06, + "loss": 1.62779808, + "memory(GiB)": 111.15, + "step": 31405, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.66113567, + "epoch": 0.7968036529680366, + "grad_norm": 4.40625, + "learning_rate": 7.032648560589316e-06, + "loss": 1.57510586, + "memory(GiB)": 111.15, + "step": 31410, + "train_speed(iter/s)": 0.447839 + }, + { + "acc": 0.65662103, + "epoch": 0.7969304921359716, + "grad_norm": 5.75, + "learning_rate": 7.031690453912835e-06, + "loss": 1.57331696, + "memory(GiB)": 111.15, + "step": 31415, + "train_speed(iter/s)": 0.447877 + }, + { + "acc": 0.64789066, + "epoch": 0.7970573313039067, + "grad_norm": 6.53125, + "learning_rate": 7.030732257873539e-06, + "loss": 1.6587677, + "memory(GiB)": 111.15, + "step": 31420, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.65363169, + "epoch": 0.7971841704718418, + "grad_norm": 5.15625, + "learning_rate": 7.02977397251357e-06, + "loss": 1.62958221, + "memory(GiB)": 111.15, + "step": 31425, + "train_speed(iter/s)": 0.447953 + }, + { + "acc": 0.66571565, + "epoch": 0.7973110096397767, + "grad_norm": 4.6875, + "learning_rate": 7.028815597875081e-06, + "loss": 1.56579132, + "memory(GiB)": 111.15, + "step": 31430, + "train_speed(iter/s)": 0.447991 + }, + { + "acc": 0.63231802, + "epoch": 0.7974378488077118, + "grad_norm": 4.9375, + "learning_rate": 7.027857134000223e-06, + "loss": 1.67255478, + "memory(GiB)": 111.15, + "step": 31435, + "train_speed(iter/s)": 0.448029 + }, + { + "acc": 0.65727959, + "epoch": 0.7975646879756468, + "grad_norm": 6.1875, + "learning_rate": 7.026898580931154e-06, + "loss": 1.63699112, + "memory(GiB)": 111.15, + "step": 31440, + "train_speed(iter/s)": 0.448068 + }, + { + "acc": 0.66182232, + "epoch": 0.7976915271435819, + "grad_norm": 5.125, + "learning_rate": 7.025939938710037e-06, + "loss": 1.58103342, + "memory(GiB)": 111.15, + "step": 31445, + "train_speed(iter/s)": 0.448106 + }, + { + "acc": 0.626756, + "epoch": 0.797818366311517, + "grad_norm": 5.375, + "learning_rate": 7.024981207379036e-06, + "loss": 1.70124702, + "memory(GiB)": 111.15, + "step": 31450, + "train_speed(iter/s)": 0.448144 + }, + { + "acc": 0.63196211, + "epoch": 0.797945205479452, + "grad_norm": 5.4375, + "learning_rate": 7.02402238698032e-06, + "loss": 1.68358116, + "memory(GiB)": 111.15, + "step": 31455, + "train_speed(iter/s)": 0.448182 + }, + { + "acc": 0.64372854, + "epoch": 0.7980720446473871, + "grad_norm": 6.25, + "learning_rate": 7.023063477556064e-06, + "loss": 1.67916298, + "memory(GiB)": 111.15, + "step": 31460, + "train_speed(iter/s)": 0.44822 + }, + { + "acc": 0.65356493, + "epoch": 0.7981988838153222, + "grad_norm": 5.1875, + "learning_rate": 7.0221044791484424e-06, + "loss": 1.58596096, + "memory(GiB)": 111.15, + "step": 31465, + "train_speed(iter/s)": 0.448259 + }, + { + "acc": 0.64748917, + "epoch": 0.7983257229832572, + "grad_norm": 5.03125, + "learning_rate": 7.021145391799639e-06, + "loss": 1.60946789, + "memory(GiB)": 111.15, + "step": 31470, + "train_speed(iter/s)": 0.448297 + }, + { + "acc": 0.64902372, + "epoch": 0.7984525621511923, + "grad_norm": 4.9375, + "learning_rate": 7.020186215551837e-06, + "loss": 1.63870678, + "memory(GiB)": 111.15, + "step": 31475, + "train_speed(iter/s)": 0.448335 + }, + { + "acc": 0.65061336, + "epoch": 0.7985794013191273, + "grad_norm": 5.6875, + "learning_rate": 7.019226950447227e-06, + "loss": 1.61500988, + "memory(GiB)": 111.15, + "step": 31480, + "train_speed(iter/s)": 0.448373 + }, + { + "acc": 0.63138103, + "epoch": 0.7987062404870624, + "grad_norm": 5.53125, + "learning_rate": 7.018267596527998e-06, + "loss": 1.6874258, + "memory(GiB)": 111.15, + "step": 31485, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.66947923, + "epoch": 0.7988330796549975, + "grad_norm": 5.3125, + "learning_rate": 7.017308153836352e-06, + "loss": 1.5617588, + "memory(GiB)": 111.15, + "step": 31490, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.64681044, + "epoch": 0.7989599188229325, + "grad_norm": 4.875, + "learning_rate": 7.016348622414484e-06, + "loss": 1.66893044, + "memory(GiB)": 111.15, + "step": 31495, + "train_speed(iter/s)": 0.448488 + }, + { + "acc": 0.65383692, + "epoch": 0.7990867579908676, + "grad_norm": 4.90625, + "learning_rate": 7.015389002304604e-06, + "loss": 1.59984856, + "memory(GiB)": 111.15, + "step": 31500, + "train_speed(iter/s)": 0.448526 + }, + { + "epoch": 0.7990867579908676, + "eval_acc": 0.6429246719586644, + "eval_loss": 1.6103376150131226, + "eval_runtime": 112.6465, + "eval_samples_per_second": 56.549, + "eval_steps_per_second": 28.274, + "step": 31500 + }, + { + "acc": 0.66249008, + "epoch": 0.7992135971588027, + "grad_norm": 5.5, + "learning_rate": 7.014429293548916e-06, + "loss": 1.55225859, + "memory(GiB)": 111.15, + "step": 31505, + "train_speed(iter/s)": 0.447816 + }, + { + "acc": 0.65054154, + "epoch": 0.7993404363267377, + "grad_norm": 5.8125, + "learning_rate": 7.013469496189633e-06, + "loss": 1.64387569, + "memory(GiB)": 111.15, + "step": 31510, + "train_speed(iter/s)": 0.447854 + }, + { + "acc": 0.65694227, + "epoch": 0.7994672754946728, + "grad_norm": 5.46875, + "learning_rate": 7.012509610268974e-06, + "loss": 1.54886665, + "memory(GiB)": 111.15, + "step": 31515, + "train_speed(iter/s)": 0.447891 + }, + { + "acc": 0.6482636, + "epoch": 0.7995941146626078, + "grad_norm": 5.625, + "learning_rate": 7.011549635829156e-06, + "loss": 1.61335354, + "memory(GiB)": 111.15, + "step": 31520, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.63865757, + "epoch": 0.7997209538305429, + "grad_norm": 5.34375, + "learning_rate": 7.010589572912404e-06, + "loss": 1.61058369, + "memory(GiB)": 111.15, + "step": 31525, + "train_speed(iter/s)": 0.447965 + }, + { + "acc": 0.64044333, + "epoch": 0.799847792998478, + "grad_norm": 4.9375, + "learning_rate": 7.009629421560946e-06, + "loss": 1.68246002, + "memory(GiB)": 111.15, + "step": 31530, + "train_speed(iter/s)": 0.448003 + }, + { + "acc": 0.65134134, + "epoch": 0.799974632166413, + "grad_norm": 5.46875, + "learning_rate": 7.008669181817015e-06, + "loss": 1.6632349, + "memory(GiB)": 111.15, + "step": 31535, + "train_speed(iter/s)": 0.44804 + }, + { + "acc": 0.64692097, + "epoch": 0.8001014713343481, + "grad_norm": 5.15625, + "learning_rate": 7.007708853722844e-06, + "loss": 1.60725288, + "memory(GiB)": 111.15, + "step": 31540, + "train_speed(iter/s)": 0.448077 + }, + { + "acc": 0.63877325, + "epoch": 0.8002283105022832, + "grad_norm": 5.78125, + "learning_rate": 7.006748437320674e-06, + "loss": 1.76492195, + "memory(GiB)": 111.15, + "step": 31545, + "train_speed(iter/s)": 0.448114 + }, + { + "acc": 0.63140173, + "epoch": 0.8003551496702181, + "grad_norm": 4.96875, + "learning_rate": 7.005787932652749e-06, + "loss": 1.61773033, + "memory(GiB)": 111.15, + "step": 31550, + "train_speed(iter/s)": 0.448151 + }, + { + "acc": 0.63873987, + "epoch": 0.8004819888381532, + "grad_norm": 4.75, + "learning_rate": 7.0048273397613145e-06, + "loss": 1.69546909, + "memory(GiB)": 111.15, + "step": 31555, + "train_speed(iter/s)": 0.448189 + }, + { + "acc": 0.65602751, + "epoch": 0.8006088280060882, + "grad_norm": 5.3125, + "learning_rate": 7.003866658688624e-06, + "loss": 1.60963116, + "memory(GiB)": 111.15, + "step": 31560, + "train_speed(iter/s)": 0.448226 + }, + { + "acc": 0.64864664, + "epoch": 0.8007356671740233, + "grad_norm": 6.0625, + "learning_rate": 7.0029058894769295e-06, + "loss": 1.68241806, + "memory(GiB)": 111.15, + "step": 31565, + "train_speed(iter/s)": 0.448263 + }, + { + "acc": 0.64374285, + "epoch": 0.8008625063419584, + "grad_norm": 5.28125, + "learning_rate": 7.001945032168493e-06, + "loss": 1.63186874, + "memory(GiB)": 111.15, + "step": 31570, + "train_speed(iter/s)": 0.448301 + }, + { + "acc": 0.65160418, + "epoch": 0.8009893455098934, + "grad_norm": 5.3125, + "learning_rate": 7.000984086805575e-06, + "loss": 1.6174469, + "memory(GiB)": 111.15, + "step": 31575, + "train_speed(iter/s)": 0.448338 + }, + { + "acc": 0.64409556, + "epoch": 0.8011161846778285, + "grad_norm": 5.46875, + "learning_rate": 7.000023053430444e-06, + "loss": 1.6469429, + "memory(GiB)": 111.15, + "step": 31580, + "train_speed(iter/s)": 0.448375 + }, + { + "acc": 0.6415966, + "epoch": 0.8012430238457636, + "grad_norm": 6.3125, + "learning_rate": 6.999061932085369e-06, + "loss": 1.69104767, + "memory(GiB)": 111.15, + "step": 31585, + "train_speed(iter/s)": 0.448413 + }, + { + "acc": 0.65153656, + "epoch": 0.8013698630136986, + "grad_norm": 5.375, + "learning_rate": 6.9981007228126255e-06, + "loss": 1.54551964, + "memory(GiB)": 111.15, + "step": 31590, + "train_speed(iter/s)": 0.44845 + }, + { + "acc": 0.63518896, + "epoch": 0.8014967021816337, + "grad_norm": 5.1875, + "learning_rate": 6.997139425654491e-06, + "loss": 1.66448784, + "memory(GiB)": 111.15, + "step": 31595, + "train_speed(iter/s)": 0.448488 + }, + { + "acc": 0.66454706, + "epoch": 0.8016235413495687, + "grad_norm": 6.15625, + "learning_rate": 6.996178040653248e-06, + "loss": 1.51267223, + "memory(GiB)": 111.15, + "step": 31600, + "train_speed(iter/s)": 0.448525 + }, + { + "epoch": 0.8016235413495687, + "eval_acc": 0.642831092081346, + "eval_loss": 1.6102632284164429, + "eval_runtime": 113.3171, + "eval_samples_per_second": 56.214, + "eval_steps_per_second": 28.107, + "step": 31600 + }, + { + "acc": 0.66257629, + "epoch": 0.8017503805175038, + "grad_norm": 5.78125, + "learning_rate": 6.995216567851183e-06, + "loss": 1.57968321, + "memory(GiB)": 111.15, + "step": 31605, + "train_speed(iter/s)": 0.447814 + }, + { + "acc": 0.63492985, + "epoch": 0.8018772196854389, + "grad_norm": 5.5625, + "learning_rate": 6.994255007290585e-06, + "loss": 1.63558197, + "memory(GiB)": 111.15, + "step": 31610, + "train_speed(iter/s)": 0.447852 + }, + { + "acc": 0.65745273, + "epoch": 0.8020040588533739, + "grad_norm": 4.375, + "learning_rate": 6.993293359013747e-06, + "loss": 1.62971916, + "memory(GiB)": 111.15, + "step": 31615, + "train_speed(iter/s)": 0.447889 + }, + { + "acc": 0.62745018, + "epoch": 0.802130898021309, + "grad_norm": 5.125, + "learning_rate": 6.992331623062969e-06, + "loss": 1.69053936, + "memory(GiB)": 111.15, + "step": 31620, + "train_speed(iter/s)": 0.447927 + }, + { + "acc": 0.64743938, + "epoch": 0.8022577371892441, + "grad_norm": 5.125, + "learning_rate": 6.9913697994805505e-06, + "loss": 1.67083321, + "memory(GiB)": 111.15, + "step": 31625, + "train_speed(iter/s)": 0.447964 + }, + { + "acc": 0.6489357, + "epoch": 0.8023845763571791, + "grad_norm": 5.28125, + "learning_rate": 6.990407888308799e-06, + "loss": 1.5593338, + "memory(GiB)": 111.15, + "step": 31630, + "train_speed(iter/s)": 0.448002 + }, + { + "acc": 0.64411135, + "epoch": 0.8025114155251142, + "grad_norm": 4.59375, + "learning_rate": 6.98944588959002e-06, + "loss": 1.64494915, + "memory(GiB)": 111.15, + "step": 31635, + "train_speed(iter/s)": 0.44804 + }, + { + "acc": 0.64670315, + "epoch": 0.8026382546930492, + "grad_norm": 6.1875, + "learning_rate": 6.9884838033665305e-06, + "loss": 1.62156601, + "memory(GiB)": 111.15, + "step": 31640, + "train_speed(iter/s)": 0.448077 + }, + { + "acc": 0.63672462, + "epoch": 0.8027650938609843, + "grad_norm": 4.84375, + "learning_rate": 6.987521629680643e-06, + "loss": 1.69469833, + "memory(GiB)": 111.15, + "step": 31645, + "train_speed(iter/s)": 0.448114 + }, + { + "acc": 0.64447265, + "epoch": 0.8028919330289194, + "grad_norm": 5.21875, + "learning_rate": 6.9865593685746815e-06, + "loss": 1.67437134, + "memory(GiB)": 111.15, + "step": 31650, + "train_speed(iter/s)": 0.448152 + }, + { + "acc": 0.64424434, + "epoch": 0.8030187721968544, + "grad_norm": 5.1875, + "learning_rate": 6.98559702009097e-06, + "loss": 1.71344376, + "memory(GiB)": 111.15, + "step": 31655, + "train_speed(iter/s)": 0.448189 + }, + { + "acc": 0.65000815, + "epoch": 0.8031456113647895, + "grad_norm": 7.5, + "learning_rate": 6.984634584271836e-06, + "loss": 1.65807037, + "memory(GiB)": 111.15, + "step": 31660, + "train_speed(iter/s)": 0.448227 + }, + { + "acc": 0.64679241, + "epoch": 0.8032724505327246, + "grad_norm": 5.21875, + "learning_rate": 6.983672061159612e-06, + "loss": 1.62801151, + "memory(GiB)": 111.15, + "step": 31665, + "train_speed(iter/s)": 0.448264 + }, + { + "acc": 0.64746494, + "epoch": 0.8033992897006595, + "grad_norm": 5.78125, + "learning_rate": 6.982709450796636e-06, + "loss": 1.61684456, + "memory(GiB)": 111.15, + "step": 31670, + "train_speed(iter/s)": 0.448302 + }, + { + "acc": 0.63285832, + "epoch": 0.8035261288685946, + "grad_norm": 5.40625, + "learning_rate": 6.981746753225245e-06, + "loss": 1.70137234, + "memory(GiB)": 111.15, + "step": 31675, + "train_speed(iter/s)": 0.448339 + }, + { + "acc": 0.65130863, + "epoch": 0.8036529680365296, + "grad_norm": 5.84375, + "learning_rate": 6.980783968487783e-06, + "loss": 1.59011993, + "memory(GiB)": 111.15, + "step": 31680, + "train_speed(iter/s)": 0.448377 + }, + { + "acc": 0.62673101, + "epoch": 0.8037798072044647, + "grad_norm": 6.5, + "learning_rate": 6.9798210966266e-06, + "loss": 1.72415333, + "memory(GiB)": 111.15, + "step": 31685, + "train_speed(iter/s)": 0.448414 + }, + { + "acc": 0.63956404, + "epoch": 0.8039066463723998, + "grad_norm": 5.03125, + "learning_rate": 6.9788581376840455e-06, + "loss": 1.62628403, + "memory(GiB)": 111.15, + "step": 31690, + "train_speed(iter/s)": 0.448452 + }, + { + "acc": 0.65633068, + "epoch": 0.8040334855403348, + "grad_norm": 5.625, + "learning_rate": 6.977895091702474e-06, + "loss": 1.63948917, + "memory(GiB)": 111.15, + "step": 31695, + "train_speed(iter/s)": 0.44849 + }, + { + "acc": 0.6592926, + "epoch": 0.8041603247082699, + "grad_norm": 5.34375, + "learning_rate": 6.976931958724248e-06, + "loss": 1.56827097, + "memory(GiB)": 111.15, + "step": 31700, + "train_speed(iter/s)": 0.448527 + }, + { + "epoch": 0.8041603247082699, + "eval_acc": 0.6428361052890594, + "eval_loss": 1.6104809045791626, + "eval_runtime": 114.0309, + "eval_samples_per_second": 55.862, + "eval_steps_per_second": 27.931, + "step": 31700 + }, + { + "acc": 0.64247723, + "epoch": 0.804287163876205, + "grad_norm": 5.15625, + "learning_rate": 6.975968738791726e-06, + "loss": 1.67576675, + "memory(GiB)": 111.15, + "step": 31705, + "train_speed(iter/s)": 0.447813 + }, + { + "acc": 0.6379591, + "epoch": 0.80441400304414, + "grad_norm": 5.40625, + "learning_rate": 6.9750054319472785e-06, + "loss": 1.70365257, + "memory(GiB)": 111.15, + "step": 31710, + "train_speed(iter/s)": 0.44785 + }, + { + "acc": 0.66468825, + "epoch": 0.8045408422120751, + "grad_norm": 5.09375, + "learning_rate": 6.974042038233272e-06, + "loss": 1.61351738, + "memory(GiB)": 111.15, + "step": 31715, + "train_speed(iter/s)": 0.447887 + }, + { + "acc": 0.62934761, + "epoch": 0.8046676813800101, + "grad_norm": 5.3125, + "learning_rate": 6.9730785576920855e-06, + "loss": 1.75182533, + "memory(GiB)": 111.15, + "step": 31720, + "train_speed(iter/s)": 0.447924 + }, + { + "acc": 0.66422758, + "epoch": 0.8047945205479452, + "grad_norm": 5.28125, + "learning_rate": 6.972114990366094e-06, + "loss": 1.56667309, + "memory(GiB)": 111.15, + "step": 31725, + "train_speed(iter/s)": 0.447961 + }, + { + "acc": 0.62641339, + "epoch": 0.8049213597158803, + "grad_norm": 6.78125, + "learning_rate": 6.97115133629768e-06, + "loss": 1.61781559, + "memory(GiB)": 111.15, + "step": 31730, + "train_speed(iter/s)": 0.447998 + }, + { + "acc": 0.66205974, + "epoch": 0.8050481988838153, + "grad_norm": 4.90625, + "learning_rate": 6.970187595529229e-06, + "loss": 1.57347279, + "memory(GiB)": 111.15, + "step": 31735, + "train_speed(iter/s)": 0.448035 + }, + { + "acc": 0.64540377, + "epoch": 0.8051750380517504, + "grad_norm": 5.71875, + "learning_rate": 6.969223768103133e-06, + "loss": 1.69937973, + "memory(GiB)": 111.15, + "step": 31740, + "train_speed(iter/s)": 0.448072 + }, + { + "acc": 0.64445391, + "epoch": 0.8053018772196855, + "grad_norm": 5.03125, + "learning_rate": 6.968259854061783e-06, + "loss": 1.62817326, + "memory(GiB)": 111.15, + "step": 31745, + "train_speed(iter/s)": 0.448109 + }, + { + "acc": 0.6530035, + "epoch": 0.8054287163876205, + "grad_norm": 5.0, + "learning_rate": 6.967295853447578e-06, + "loss": 1.64273682, + "memory(GiB)": 111.15, + "step": 31750, + "train_speed(iter/s)": 0.448146 + }, + { + "acc": 0.65467186, + "epoch": 0.8055555555555556, + "grad_norm": 5.9375, + "learning_rate": 6.966331766302916e-06, + "loss": 1.63779221, + "memory(GiB)": 111.15, + "step": 31755, + "train_speed(iter/s)": 0.448183 + }, + { + "acc": 0.65035753, + "epoch": 0.8056823947234906, + "grad_norm": 5.125, + "learning_rate": 6.965367592670206e-06, + "loss": 1.61562119, + "memory(GiB)": 111.15, + "step": 31760, + "train_speed(iter/s)": 0.44822 + }, + { + "acc": 0.65400381, + "epoch": 0.8058092338914257, + "grad_norm": 5.78125, + "learning_rate": 6.964403332591854e-06, + "loss": 1.68322487, + "memory(GiB)": 111.15, + "step": 31765, + "train_speed(iter/s)": 0.448257 + }, + { + "acc": 0.64700384, + "epoch": 0.8059360730593608, + "grad_norm": 5.6875, + "learning_rate": 6.963438986110272e-06, + "loss": 1.62388287, + "memory(GiB)": 111.15, + "step": 31770, + "train_speed(iter/s)": 0.448294 + }, + { + "acc": 0.65799122, + "epoch": 0.8060629122272958, + "grad_norm": 5.96875, + "learning_rate": 6.962474553267877e-06, + "loss": 1.60779762, + "memory(GiB)": 111.15, + "step": 31775, + "train_speed(iter/s)": 0.448331 + }, + { + "acc": 0.63359766, + "epoch": 0.8061897513952309, + "grad_norm": 5.21875, + "learning_rate": 6.96151003410709e-06, + "loss": 1.73813095, + "memory(GiB)": 111.15, + "step": 31780, + "train_speed(iter/s)": 0.448368 + }, + { + "acc": 0.65172038, + "epoch": 0.806316590563166, + "grad_norm": 4.78125, + "learning_rate": 6.960545428670333e-06, + "loss": 1.66956406, + "memory(GiB)": 111.15, + "step": 31785, + "train_speed(iter/s)": 0.448405 + }, + { + "acc": 0.66390686, + "epoch": 0.806443429731101, + "grad_norm": 5.9375, + "learning_rate": 6.959580737000038e-06, + "loss": 1.57240562, + "memory(GiB)": 111.15, + "step": 31790, + "train_speed(iter/s)": 0.448442 + }, + { + "acc": 0.65922794, + "epoch": 0.806570268899036, + "grad_norm": 5.0, + "learning_rate": 6.95861595913863e-06, + "loss": 1.57239265, + "memory(GiB)": 111.15, + "step": 31795, + "train_speed(iter/s)": 0.448479 + }, + { + "acc": 0.67181396, + "epoch": 0.806697108066971, + "grad_norm": 4.59375, + "learning_rate": 6.95765109512855e-06, + "loss": 1.51912479, + "memory(GiB)": 111.15, + "step": 31800, + "train_speed(iter/s)": 0.448516 + }, + { + "epoch": 0.806697108066971, + "eval_acc": 0.6429405471164238, + "eval_loss": 1.610428810119629, + "eval_runtime": 114.744, + "eval_samples_per_second": 55.515, + "eval_steps_per_second": 27.757, + "step": 31800 + }, + { + "acc": 0.64618759, + "epoch": 0.8068239472349061, + "grad_norm": 5.6875, + "learning_rate": 6.956686145012233e-06, + "loss": 1.61521854, + "memory(GiB)": 111.15, + "step": 31805, + "train_speed(iter/s)": 0.4478 + }, + { + "acc": 0.64400768, + "epoch": 0.8069507864028412, + "grad_norm": 5.03125, + "learning_rate": 6.955721108832124e-06, + "loss": 1.62332096, + "memory(GiB)": 111.15, + "step": 31810, + "train_speed(iter/s)": 0.447836 + }, + { + "acc": 0.65436106, + "epoch": 0.8070776255707762, + "grad_norm": 5.625, + "learning_rate": 6.9547559866306695e-06, + "loss": 1.58100796, + "memory(GiB)": 111.15, + "step": 31815, + "train_speed(iter/s)": 0.447873 + }, + { + "acc": 0.65407009, + "epoch": 0.8072044647387113, + "grad_norm": 5.9375, + "learning_rate": 6.953790778450318e-06, + "loss": 1.62880211, + "memory(GiB)": 111.15, + "step": 31820, + "train_speed(iter/s)": 0.44791 + }, + { + "acc": 0.6432745, + "epoch": 0.8073313039066464, + "grad_norm": 5.90625, + "learning_rate": 6.9528254843335254e-06, + "loss": 1.63235779, + "memory(GiB)": 111.15, + "step": 31825, + "train_speed(iter/s)": 0.447947 + }, + { + "acc": 0.6581563, + "epoch": 0.8074581430745814, + "grad_norm": 5.46875, + "learning_rate": 6.95186010432275e-06, + "loss": 1.57537041, + "memory(GiB)": 111.15, + "step": 31830, + "train_speed(iter/s)": 0.447984 + }, + { + "acc": 0.64470692, + "epoch": 0.8075849822425165, + "grad_norm": 4.875, + "learning_rate": 6.950894638460452e-06, + "loss": 1.65555878, + "memory(GiB)": 111.15, + "step": 31835, + "train_speed(iter/s)": 0.448022 + }, + { + "acc": 0.64439201, + "epoch": 0.8077118214104515, + "grad_norm": 5.15625, + "learning_rate": 6.949929086789098e-06, + "loss": 1.64161129, + "memory(GiB)": 111.15, + "step": 31840, + "train_speed(iter/s)": 0.448059 + }, + { + "acc": 0.63912249, + "epoch": 0.8078386605783866, + "grad_norm": 4.71875, + "learning_rate": 6.948963449351156e-06, + "loss": 1.69429169, + "memory(GiB)": 111.15, + "step": 31845, + "train_speed(iter/s)": 0.448096 + }, + { + "acc": 0.6373486, + "epoch": 0.8079654997463217, + "grad_norm": 5.875, + "learning_rate": 6.947997726189102e-06, + "loss": 1.67736931, + "memory(GiB)": 111.15, + "step": 31850, + "train_speed(iter/s)": 0.448134 + }, + { + "acc": 0.64799404, + "epoch": 0.8080923389142567, + "grad_norm": 6.96875, + "learning_rate": 6.947031917345409e-06, + "loss": 1.67106133, + "memory(GiB)": 111.15, + "step": 31855, + "train_speed(iter/s)": 0.448171 + }, + { + "acc": 0.63622622, + "epoch": 0.8082191780821918, + "grad_norm": 5.6875, + "learning_rate": 6.946066022862561e-06, + "loss": 1.66762047, + "memory(GiB)": 111.15, + "step": 31860, + "train_speed(iter/s)": 0.448208 + }, + { + "acc": 0.65017257, + "epoch": 0.8083460172501269, + "grad_norm": 5.1875, + "learning_rate": 6.945100042783039e-06, + "loss": 1.68033905, + "memory(GiB)": 111.15, + "step": 31865, + "train_speed(iter/s)": 0.448245 + }, + { + "acc": 0.64822478, + "epoch": 0.8084728564180619, + "grad_norm": 5.375, + "learning_rate": 6.9441339771493345e-06, + "loss": 1.60199432, + "memory(GiB)": 111.15, + "step": 31870, + "train_speed(iter/s)": 0.448282 + }, + { + "acc": 0.65957184, + "epoch": 0.808599695585997, + "grad_norm": 5.3125, + "learning_rate": 6.943167826003937e-06, + "loss": 1.53220997, + "memory(GiB)": 111.15, + "step": 31875, + "train_speed(iter/s)": 0.44832 + }, + { + "acc": 0.65626011, + "epoch": 0.808726534753932, + "grad_norm": 4.8125, + "learning_rate": 6.942201589389344e-06, + "loss": 1.62029152, + "memory(GiB)": 111.15, + "step": 31880, + "train_speed(iter/s)": 0.448357 + }, + { + "acc": 0.66537304, + "epoch": 0.8088533739218671, + "grad_norm": 4.5625, + "learning_rate": 6.9412352673480525e-06, + "loss": 1.5949441, + "memory(GiB)": 111.15, + "step": 31885, + "train_speed(iter/s)": 0.448395 + }, + { + "acc": 0.63273411, + "epoch": 0.8089802130898022, + "grad_norm": 4.625, + "learning_rate": 6.940268859922566e-06, + "loss": 1.67406769, + "memory(GiB)": 111.15, + "step": 31890, + "train_speed(iter/s)": 0.448431 + }, + { + "acc": 0.65248775, + "epoch": 0.8091070522577372, + "grad_norm": 5.25, + "learning_rate": 6.939302367155394e-06, + "loss": 1.6192543, + "memory(GiB)": 111.15, + "step": 31895, + "train_speed(iter/s)": 0.448468 + }, + { + "acc": 0.63136826, + "epoch": 0.8092338914256723, + "grad_norm": 6.40625, + "learning_rate": 6.9383357890890454e-06, + "loss": 1.78237381, + "memory(GiB)": 111.15, + "step": 31900, + "train_speed(iter/s)": 0.448505 + }, + { + "epoch": 0.8092338914256723, + "eval_acc": 0.642909632335524, + "eval_loss": 1.610245943069458, + "eval_runtime": 113.5665, + "eval_samples_per_second": 56.09, + "eval_steps_per_second": 28.045, + "step": 31900 + }, + { + "acc": 0.63244495, + "epoch": 0.8093607305936074, + "grad_norm": 5.78125, + "learning_rate": 6.937369125766033e-06, + "loss": 1.66098251, + "memory(GiB)": 111.15, + "step": 31905, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.6472302, + "epoch": 0.8094875697615423, + "grad_norm": 5.65625, + "learning_rate": 6.936402377228879e-06, + "loss": 1.67172985, + "memory(GiB)": 111.15, + "step": 31910, + "train_speed(iter/s)": 0.447836 + }, + { + "acc": 0.64511914, + "epoch": 0.8096144089294774, + "grad_norm": 4.90625, + "learning_rate": 6.9354355435201015e-06, + "loss": 1.57840033, + "memory(GiB)": 111.15, + "step": 31915, + "train_speed(iter/s)": 0.447872 + }, + { + "acc": 0.64322915, + "epoch": 0.8097412480974124, + "grad_norm": 5.34375, + "learning_rate": 6.934468624682229e-06, + "loss": 1.6797924, + "memory(GiB)": 111.15, + "step": 31920, + "train_speed(iter/s)": 0.447909 + }, + { + "acc": 0.65080938, + "epoch": 0.8098680872653475, + "grad_norm": 5.53125, + "learning_rate": 6.933501620757789e-06, + "loss": 1.57485313, + "memory(GiB)": 111.15, + "step": 31925, + "train_speed(iter/s)": 0.447946 + }, + { + "acc": 0.65878, + "epoch": 0.8099949264332826, + "grad_norm": 4.90625, + "learning_rate": 6.932534531789317e-06, + "loss": 1.5632822, + "memory(GiB)": 111.15, + "step": 31930, + "train_speed(iter/s)": 0.447982 + }, + { + "acc": 0.64428167, + "epoch": 0.8101217656012176, + "grad_norm": 5.15625, + "learning_rate": 6.931567357819344e-06, + "loss": 1.62149239, + "memory(GiB)": 111.15, + "step": 31935, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.63154564, + "epoch": 0.8102486047691527, + "grad_norm": 5.03125, + "learning_rate": 6.930600098890419e-06, + "loss": 1.64114494, + "memory(GiB)": 111.15, + "step": 31940, + "train_speed(iter/s)": 0.448056 + }, + { + "acc": 0.63778973, + "epoch": 0.8103754439370878, + "grad_norm": 5.0625, + "learning_rate": 6.929632755045079e-06, + "loss": 1.64285583, + "memory(GiB)": 111.15, + "step": 31945, + "train_speed(iter/s)": 0.448093 + }, + { + "acc": 0.6517036, + "epoch": 0.8105022831050228, + "grad_norm": 7.53125, + "learning_rate": 6.9286653263258765e-06, + "loss": 1.64077339, + "memory(GiB)": 111.15, + "step": 31950, + "train_speed(iter/s)": 0.44813 + }, + { + "acc": 0.63462081, + "epoch": 0.8106291222729579, + "grad_norm": 5.5625, + "learning_rate": 6.927697812775363e-06, + "loss": 1.70597153, + "memory(GiB)": 111.15, + "step": 31955, + "train_speed(iter/s)": 0.448166 + }, + { + "acc": 0.63532829, + "epoch": 0.8107559614408929, + "grad_norm": 6.53125, + "learning_rate": 6.926730214436091e-06, + "loss": 1.69389038, + "memory(GiB)": 111.15, + "step": 31960, + "train_speed(iter/s)": 0.448203 + }, + { + "acc": 0.65800147, + "epoch": 0.810882800608828, + "grad_norm": 6.09375, + "learning_rate": 6.925762531350624e-06, + "loss": 1.61896172, + "memory(GiB)": 111.15, + "step": 31965, + "train_speed(iter/s)": 0.44824 + }, + { + "acc": 0.64313412, + "epoch": 0.8110096397767631, + "grad_norm": 6.28125, + "learning_rate": 6.924794763561522e-06, + "loss": 1.64976444, + "memory(GiB)": 111.15, + "step": 31970, + "train_speed(iter/s)": 0.448277 + }, + { + "acc": 0.64435272, + "epoch": 0.8111364789446981, + "grad_norm": 5.375, + "learning_rate": 6.923826911111353e-06, + "loss": 1.63263569, + "memory(GiB)": 111.15, + "step": 31975, + "train_speed(iter/s)": 0.448314 + }, + { + "acc": 0.65464978, + "epoch": 0.8112633181126332, + "grad_norm": 4.375, + "learning_rate": 6.922858974042688e-06, + "loss": 1.60467491, + "memory(GiB)": 111.15, + "step": 31980, + "train_speed(iter/s)": 0.448351 + }, + { + "acc": 0.66116729, + "epoch": 0.8113901572805683, + "grad_norm": 6.125, + "learning_rate": 6.921890952398098e-06, + "loss": 1.58285847, + "memory(GiB)": 111.15, + "step": 31985, + "train_speed(iter/s)": 0.448387 + }, + { + "acc": 0.64935994, + "epoch": 0.8115169964485033, + "grad_norm": 5.75, + "learning_rate": 6.920922846220166e-06, + "loss": 1.67587318, + "memory(GiB)": 111.15, + "step": 31990, + "train_speed(iter/s)": 0.448424 + }, + { + "acc": 0.63335123, + "epoch": 0.8116438356164384, + "grad_norm": 5.53125, + "learning_rate": 6.919954655551469e-06, + "loss": 1.63188457, + "memory(GiB)": 111.15, + "step": 31995, + "train_speed(iter/s)": 0.448461 + }, + { + "acc": 0.67098413, + "epoch": 0.8117706747843734, + "grad_norm": 5.0625, + "learning_rate": 6.918986380434594e-06, + "loss": 1.56689825, + "memory(GiB)": 111.15, + "step": 32000, + "train_speed(iter/s)": 0.448498 + }, + { + "epoch": 0.8117706747843734, + "eval_acc": 0.6430057188166992, + "eval_loss": 1.6101093292236328, + "eval_runtime": 114.0623, + "eval_samples_per_second": 55.847, + "eval_steps_per_second": 27.923, + "step": 32000 + }, + { + "acc": 0.65931153, + "epoch": 0.8118975139523085, + "grad_norm": 6.375, + "learning_rate": 6.918018020912132e-06, + "loss": 1.66122608, + "memory(GiB)": 111.15, + "step": 32005, + "train_speed(iter/s)": 0.447791 + }, + { + "acc": 0.66606636, + "epoch": 0.8120243531202436, + "grad_norm": 5.21875, + "learning_rate": 6.917049577026673e-06, + "loss": 1.57549, + "memory(GiB)": 111.15, + "step": 32010, + "train_speed(iter/s)": 0.447828 + }, + { + "acc": 0.65245199, + "epoch": 0.8121511922881786, + "grad_norm": 4.5, + "learning_rate": 6.916081048820815e-06, + "loss": 1.57656918, + "memory(GiB)": 111.15, + "step": 32015, + "train_speed(iter/s)": 0.447865 + }, + { + "acc": 0.65004497, + "epoch": 0.8122780314561137, + "grad_norm": 4.90625, + "learning_rate": 6.915112436337157e-06, + "loss": 1.63452911, + "memory(GiB)": 111.15, + "step": 32020, + "train_speed(iter/s)": 0.447901 + }, + { + "acc": 0.64325523, + "epoch": 0.8124048706240488, + "grad_norm": 5.09375, + "learning_rate": 6.914143739618305e-06, + "loss": 1.67313023, + "memory(GiB)": 111.15, + "step": 32025, + "train_speed(iter/s)": 0.447938 + }, + { + "acc": 0.67870016, + "epoch": 0.8125317097919837, + "grad_norm": 5.875, + "learning_rate": 6.913174958706865e-06, + "loss": 1.55473747, + "memory(GiB)": 111.15, + "step": 32030, + "train_speed(iter/s)": 0.447975 + }, + { + "acc": 0.64887376, + "epoch": 0.8126585489599188, + "grad_norm": 5.21875, + "learning_rate": 6.912206093645448e-06, + "loss": 1.64953308, + "memory(GiB)": 111.15, + "step": 32035, + "train_speed(iter/s)": 0.448012 + }, + { + "acc": 0.66025133, + "epoch": 0.8127853881278538, + "grad_norm": 7.65625, + "learning_rate": 6.91123714447667e-06, + "loss": 1.62360096, + "memory(GiB)": 111.15, + "step": 32040, + "train_speed(iter/s)": 0.448048 + }, + { + "acc": 0.64142513, + "epoch": 0.8129122272957889, + "grad_norm": 4.75, + "learning_rate": 6.910268111243149e-06, + "loss": 1.61002121, + "memory(GiB)": 111.15, + "step": 32045, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.63471193, + "epoch": 0.813039066463724, + "grad_norm": 5.84375, + "learning_rate": 6.909298993987508e-06, + "loss": 1.74754066, + "memory(GiB)": 111.15, + "step": 32050, + "train_speed(iter/s)": 0.448121 + }, + { + "acc": 0.6482111, + "epoch": 0.813165905631659, + "grad_norm": 4.40625, + "learning_rate": 6.908329792752373e-06, + "loss": 1.57019348, + "memory(GiB)": 111.15, + "step": 32055, + "train_speed(iter/s)": 0.448158 + }, + { + "acc": 0.65040941, + "epoch": 0.8132927447995941, + "grad_norm": 4.8125, + "learning_rate": 6.907360507580374e-06, + "loss": 1.64693375, + "memory(GiB)": 111.15, + "step": 32060, + "train_speed(iter/s)": 0.448195 + }, + { + "acc": 0.64002314, + "epoch": 0.8134195839675292, + "grad_norm": 5.03125, + "learning_rate": 6.9063911385141425e-06, + "loss": 1.63660316, + "memory(GiB)": 111.15, + "step": 32065, + "train_speed(iter/s)": 0.448232 + }, + { + "acc": 0.62601032, + "epoch": 0.8135464231354642, + "grad_norm": 5.15625, + "learning_rate": 6.9054216855963194e-06, + "loss": 1.64721909, + "memory(GiB)": 111.15, + "step": 32070, + "train_speed(iter/s)": 0.448269 + }, + { + "acc": 0.64120579, + "epoch": 0.8136732623033993, + "grad_norm": 5.96875, + "learning_rate": 6.904452148869541e-06, + "loss": 1.65252571, + "memory(GiB)": 111.15, + "step": 32075, + "train_speed(iter/s)": 0.448305 + }, + { + "acc": 0.64080791, + "epoch": 0.8138001014713343, + "grad_norm": 4.75, + "learning_rate": 6.903482528376457e-06, + "loss": 1.59783144, + "memory(GiB)": 111.15, + "step": 32080, + "train_speed(iter/s)": 0.448341 + }, + { + "acc": 0.64521198, + "epoch": 0.8139269406392694, + "grad_norm": 5.53125, + "learning_rate": 6.902512824159711e-06, + "loss": 1.68569107, + "memory(GiB)": 111.15, + "step": 32085, + "train_speed(iter/s)": 0.448378 + }, + { + "acc": 0.65083151, + "epoch": 0.8140537798072045, + "grad_norm": 5.28125, + "learning_rate": 6.901543036261957e-06, + "loss": 1.61486492, + "memory(GiB)": 111.15, + "step": 32090, + "train_speed(iter/s)": 0.448415 + }, + { + "acc": 0.63751502, + "epoch": 0.8141806189751395, + "grad_norm": 7.9375, + "learning_rate": 6.900573164725852e-06, + "loss": 1.71959343, + "memory(GiB)": 111.15, + "step": 32095, + "train_speed(iter/s)": 0.448452 + }, + { + "acc": 0.65948038, + "epoch": 0.8143074581430746, + "grad_norm": 5.625, + "learning_rate": 6.899603209594052e-06, + "loss": 1.61185646, + "memory(GiB)": 111.15, + "step": 32100, + "train_speed(iter/s)": 0.448488 + }, + { + "epoch": 0.8143074581430746, + "eval_acc": 0.6430199229052207, + "eval_loss": 1.6102267503738403, + "eval_runtime": 112.2979, + "eval_samples_per_second": 56.724, + "eval_steps_per_second": 28.362, + "step": 32100 + }, + { + "acc": 0.65387573, + "epoch": 0.8144342973110097, + "grad_norm": 3.78125, + "learning_rate": 6.898633170909224e-06, + "loss": 1.60829029, + "memory(GiB)": 111.15, + "step": 32105, + "train_speed(iter/s)": 0.447794 + }, + { + "acc": 0.65083055, + "epoch": 0.8145611364789447, + "grad_norm": 5.71875, + "learning_rate": 6.897663048714031e-06, + "loss": 1.62080517, + "memory(GiB)": 111.15, + "step": 32110, + "train_speed(iter/s)": 0.447832 + }, + { + "acc": 0.65785837, + "epoch": 0.8146879756468798, + "grad_norm": 5.0625, + "learning_rate": 6.896692843051145e-06, + "loss": 1.64479866, + "memory(GiB)": 111.15, + "step": 32115, + "train_speed(iter/s)": 0.447869 + }, + { + "acc": 0.65228848, + "epoch": 0.8148148148148148, + "grad_norm": 5.75, + "learning_rate": 6.895722553963239e-06, + "loss": 1.63371506, + "memory(GiB)": 111.15, + "step": 32120, + "train_speed(iter/s)": 0.447906 + }, + { + "acc": 0.65338068, + "epoch": 0.8149416539827499, + "grad_norm": 5.46875, + "learning_rate": 6.8947521814929915e-06, + "loss": 1.67630653, + "memory(GiB)": 111.15, + "step": 32125, + "train_speed(iter/s)": 0.447943 + }, + { + "acc": 0.63410273, + "epoch": 0.815068493150685, + "grad_norm": 6.90625, + "learning_rate": 6.8937817256830834e-06, + "loss": 1.75306816, + "memory(GiB)": 111.15, + "step": 32130, + "train_speed(iter/s)": 0.447981 + }, + { + "acc": 0.64836688, + "epoch": 0.81519533231862, + "grad_norm": 5.90625, + "learning_rate": 6.892811186576199e-06, + "loss": 1.64142876, + "memory(GiB)": 111.15, + "step": 32135, + "train_speed(iter/s)": 0.448018 + }, + { + "acc": 0.64785552, + "epoch": 0.8153221714865551, + "grad_norm": 5.375, + "learning_rate": 6.8918405642150295e-06, + "loss": 1.6485157, + "memory(GiB)": 111.15, + "step": 32140, + "train_speed(iter/s)": 0.448055 + }, + { + "acc": 0.64854913, + "epoch": 0.8154490106544902, + "grad_norm": 4.5, + "learning_rate": 6.890869858642264e-06, + "loss": 1.6303381, + "memory(GiB)": 111.15, + "step": 32145, + "train_speed(iter/s)": 0.448092 + }, + { + "acc": 0.64658766, + "epoch": 0.8155758498224251, + "grad_norm": 5.1875, + "learning_rate": 6.889899069900603e-06, + "loss": 1.66202736, + "memory(GiB)": 111.15, + "step": 32150, + "train_speed(iter/s)": 0.448129 + }, + { + "acc": 0.64655228, + "epoch": 0.8157026889903602, + "grad_norm": 5.46875, + "learning_rate": 6.888928198032741e-06, + "loss": 1.627285, + "memory(GiB)": 111.15, + "step": 32155, + "train_speed(iter/s)": 0.448166 + }, + { + "acc": 0.65543647, + "epoch": 0.8158295281582952, + "grad_norm": 6.4375, + "learning_rate": 6.887957243081384e-06, + "loss": 1.62440529, + "memory(GiB)": 111.15, + "step": 32160, + "train_speed(iter/s)": 0.448204 + }, + { + "acc": 0.65511284, + "epoch": 0.8159563673262303, + "grad_norm": 5.0, + "learning_rate": 6.886986205089237e-06, + "loss": 1.61765823, + "memory(GiB)": 111.15, + "step": 32165, + "train_speed(iter/s)": 0.448241 + }, + { + "acc": 0.64843044, + "epoch": 0.8160832064941654, + "grad_norm": 5.4375, + "learning_rate": 6.886015084099011e-06, + "loss": 1.6294775, + "memory(GiB)": 111.15, + "step": 32170, + "train_speed(iter/s)": 0.448278 + }, + { + "acc": 0.6519856, + "epoch": 0.8162100456621004, + "grad_norm": 5.4375, + "learning_rate": 6.885043880153424e-06, + "loss": 1.59830055, + "memory(GiB)": 111.15, + "step": 32175, + "train_speed(iter/s)": 0.448315 + }, + { + "acc": 0.66188107, + "epoch": 0.8163368848300355, + "grad_norm": 5.78125, + "learning_rate": 6.88407259329519e-06, + "loss": 1.55038738, + "memory(GiB)": 111.15, + "step": 32180, + "train_speed(iter/s)": 0.448353 + }, + { + "acc": 0.65805902, + "epoch": 0.8164637239979706, + "grad_norm": 5.1875, + "learning_rate": 6.883101223567031e-06, + "loss": 1.63716621, + "memory(GiB)": 111.15, + "step": 32185, + "train_speed(iter/s)": 0.44839 + }, + { + "acc": 0.65508947, + "epoch": 0.8165905631659056, + "grad_norm": 4.625, + "learning_rate": 6.882129771011674e-06, + "loss": 1.56608591, + "memory(GiB)": 111.15, + "step": 32190, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.66098647, + "epoch": 0.8167174023338407, + "grad_norm": 5.4375, + "learning_rate": 6.881158235671845e-06, + "loss": 1.52881451, + "memory(GiB)": 111.15, + "step": 32195, + "train_speed(iter/s)": 0.448464 + }, + { + "acc": 0.6498518, + "epoch": 0.8168442415017757, + "grad_norm": 5.40625, + "learning_rate": 6.8801866175902785e-06, + "loss": 1.60693474, + "memory(GiB)": 111.15, + "step": 32200, + "train_speed(iter/s)": 0.448501 + }, + { + "epoch": 0.8168442415017757, + "eval_acc": 0.6430090609551748, + "eval_loss": 1.6098793745040894, + "eval_runtime": 112.4846, + "eval_samples_per_second": 56.63, + "eval_steps_per_second": 28.315, + "step": 32200 + }, + { + "acc": 0.64329233, + "epoch": 0.8169710806697108, + "grad_norm": 5.09375, + "learning_rate": 6.87921491680971e-06, + "loss": 1.68426514, + "memory(GiB)": 111.15, + "step": 32205, + "train_speed(iter/s)": 0.447808 + }, + { + "acc": 0.65418148, + "epoch": 0.8170979198376459, + "grad_norm": 5.5625, + "learning_rate": 6.878243133372882e-06, + "loss": 1.58572884, + "memory(GiB)": 111.15, + "step": 32210, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.65189028, + "epoch": 0.8172247590055809, + "grad_norm": 5.53125, + "learning_rate": 6.877271267322532e-06, + "loss": 1.65421162, + "memory(GiB)": 111.15, + "step": 32215, + "train_speed(iter/s)": 0.44788 + }, + { + "acc": 0.63424625, + "epoch": 0.817351598173516, + "grad_norm": 4.4375, + "learning_rate": 6.876299318701412e-06, + "loss": 1.68095589, + "memory(GiB)": 111.15, + "step": 32220, + "train_speed(iter/s)": 0.447916 + }, + { + "acc": 0.64811325, + "epoch": 0.8174784373414511, + "grad_norm": 5.375, + "learning_rate": 6.875327287552269e-06, + "loss": 1.62843742, + "memory(GiB)": 111.15, + "step": 32225, + "train_speed(iter/s)": 0.447953 + }, + { + "acc": 0.64201808, + "epoch": 0.8176052765093861, + "grad_norm": 5.65625, + "learning_rate": 6.8743551739178615e-06, + "loss": 1.66659088, + "memory(GiB)": 111.15, + "step": 32230, + "train_speed(iter/s)": 0.447989 + }, + { + "acc": 0.64306602, + "epoch": 0.8177321156773212, + "grad_norm": 5.1875, + "learning_rate": 6.8733829778409425e-06, + "loss": 1.66120968, + "memory(GiB)": 111.15, + "step": 32235, + "train_speed(iter/s)": 0.448026 + }, + { + "acc": 0.66174555, + "epoch": 0.8178589548452562, + "grad_norm": 5.5, + "learning_rate": 6.872410699364278e-06, + "loss": 1.47891312, + "memory(GiB)": 111.15, + "step": 32240, + "train_speed(iter/s)": 0.448063 + }, + { + "acc": 0.64959626, + "epoch": 0.8179857940131913, + "grad_norm": 5.0, + "learning_rate": 6.8714383385306305e-06, + "loss": 1.64989433, + "memory(GiB)": 111.15, + "step": 32245, + "train_speed(iter/s)": 0.448099 + }, + { + "acc": 0.65615549, + "epoch": 0.8181126331811264, + "grad_norm": 4.34375, + "learning_rate": 6.870465895382769e-06, + "loss": 1.57974682, + "memory(GiB)": 111.15, + "step": 32250, + "train_speed(iter/s)": 0.448136 + }, + { + "acc": 0.66252422, + "epoch": 0.8182394723490614, + "grad_norm": 5.5, + "learning_rate": 6.869493369963468e-06, + "loss": 1.58807945, + "memory(GiB)": 111.15, + "step": 32255, + "train_speed(iter/s)": 0.448172 + }, + { + "acc": 0.64822531, + "epoch": 0.8183663115169965, + "grad_norm": 6.34375, + "learning_rate": 6.8685207623155e-06, + "loss": 1.62792683, + "memory(GiB)": 111.15, + "step": 32260, + "train_speed(iter/s)": 0.448209 + }, + { + "acc": 0.64095178, + "epoch": 0.8184931506849316, + "grad_norm": 5.03125, + "learning_rate": 6.867548072481649e-06, + "loss": 1.68384514, + "memory(GiB)": 111.15, + "step": 32265, + "train_speed(iter/s)": 0.448245 + }, + { + "acc": 0.62751408, + "epoch": 0.8186199898528665, + "grad_norm": 5.1875, + "learning_rate": 6.866575300504695e-06, + "loss": 1.63640022, + "memory(GiB)": 111.15, + "step": 32270, + "train_speed(iter/s)": 0.448282 + }, + { + "acc": 0.64976854, + "epoch": 0.8187468290208016, + "grad_norm": 6.6875, + "learning_rate": 6.865602446427424e-06, + "loss": 1.62372932, + "memory(GiB)": 111.15, + "step": 32275, + "train_speed(iter/s)": 0.448319 + }, + { + "acc": 0.64941311, + "epoch": 0.8188736681887366, + "grad_norm": 6.4375, + "learning_rate": 6.864629510292629e-06, + "loss": 1.65745926, + "memory(GiB)": 111.15, + "step": 32280, + "train_speed(iter/s)": 0.448356 + }, + { + "acc": 0.66120553, + "epoch": 0.8190005073566717, + "grad_norm": 4.125, + "learning_rate": 6.863656492143103e-06, + "loss": 1.56767559, + "memory(GiB)": 111.15, + "step": 32285, + "train_speed(iter/s)": 0.448392 + }, + { + "acc": 0.63920107, + "epoch": 0.8191273465246068, + "grad_norm": 6.78125, + "learning_rate": 6.862683392021644e-06, + "loss": 1.63457832, + "memory(GiB)": 111.15, + "step": 32290, + "train_speed(iter/s)": 0.448428 + }, + { + "acc": 0.66457024, + "epoch": 0.8192541856925418, + "grad_norm": 5.8125, + "learning_rate": 6.861710209971052e-06, + "loss": 1.50409756, + "memory(GiB)": 111.15, + "step": 32295, + "train_speed(iter/s)": 0.448465 + }, + { + "acc": 0.63537397, + "epoch": 0.8193810248604769, + "grad_norm": 4.875, + "learning_rate": 6.860736946034136e-06, + "loss": 1.68049278, + "memory(GiB)": 111.15, + "step": 32300, + "train_speed(iter/s)": 0.448501 + }, + { + "epoch": 0.8193810248604769, + "eval_acc": 0.6430128208609599, + "eval_loss": 1.6101309061050415, + "eval_runtime": 113.3191, + "eval_samples_per_second": 56.213, + "eval_steps_per_second": 28.106, + "step": 32300 + }, + { + "acc": 0.63649826, + "epoch": 0.819507864028412, + "grad_norm": 5.3125, + "learning_rate": 6.859763600253698e-06, + "loss": 1.74757595, + "memory(GiB)": 111.15, + "step": 32305, + "train_speed(iter/s)": 0.447805 + }, + { + "acc": 0.64740829, + "epoch": 0.819634703196347, + "grad_norm": 4.75, + "learning_rate": 6.858790172672556e-06, + "loss": 1.64428558, + "memory(GiB)": 111.15, + "step": 32310, + "train_speed(iter/s)": 0.447842 + }, + { + "acc": 0.6554275, + "epoch": 0.8197615423642821, + "grad_norm": 5.34375, + "learning_rate": 6.857816663333523e-06, + "loss": 1.6234087, + "memory(GiB)": 111.15, + "step": 32315, + "train_speed(iter/s)": 0.447878 + }, + { + "acc": 0.66094451, + "epoch": 0.8198883815322171, + "grad_norm": 7.4375, + "learning_rate": 6.856843072279418e-06, + "loss": 1.56929703, + "memory(GiB)": 111.15, + "step": 32320, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.64763412, + "epoch": 0.8200152207001522, + "grad_norm": 5.03125, + "learning_rate": 6.855869399553065e-06, + "loss": 1.68899479, + "memory(GiB)": 111.15, + "step": 32325, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.6500658, + "epoch": 0.8201420598680873, + "grad_norm": 6.71875, + "learning_rate": 6.85489564519729e-06, + "loss": 1.60307236, + "memory(GiB)": 111.15, + "step": 32330, + "train_speed(iter/s)": 0.447988 + }, + { + "acc": 0.66262226, + "epoch": 0.8202688990360223, + "grad_norm": 5.0625, + "learning_rate": 6.853921809254922e-06, + "loss": 1.60607357, + "memory(GiB)": 111.15, + "step": 32335, + "train_speed(iter/s)": 0.448025 + }, + { + "acc": 0.63533254, + "epoch": 0.8203957382039574, + "grad_norm": 5.84375, + "learning_rate": 6.852947891768796e-06, + "loss": 1.6254261, + "memory(GiB)": 111.15, + "step": 32340, + "train_speed(iter/s)": 0.448061 + }, + { + "acc": 0.67298594, + "epoch": 0.8205225773718925, + "grad_norm": 5.71875, + "learning_rate": 6.851973892781749e-06, + "loss": 1.54040718, + "memory(GiB)": 111.15, + "step": 32345, + "train_speed(iter/s)": 0.448098 + }, + { + "acc": 0.64756932, + "epoch": 0.8206494165398275, + "grad_norm": 5.53125, + "learning_rate": 6.850999812336623e-06, + "loss": 1.63710575, + "memory(GiB)": 111.15, + "step": 32350, + "train_speed(iter/s)": 0.448135 + }, + { + "acc": 0.64512329, + "epoch": 0.8207762557077626, + "grad_norm": 5.75, + "learning_rate": 6.850025650476259e-06, + "loss": 1.66190796, + "memory(GiB)": 111.15, + "step": 32355, + "train_speed(iter/s)": 0.448171 + }, + { + "acc": 0.6438077, + "epoch": 0.8209030948756976, + "grad_norm": 4.875, + "learning_rate": 6.849051407243509e-06, + "loss": 1.66104431, + "memory(GiB)": 111.15, + "step": 32360, + "train_speed(iter/s)": 0.448208 + }, + { + "acc": 0.64460835, + "epoch": 0.8210299340436327, + "grad_norm": 5.8125, + "learning_rate": 6.8480770826812205e-06, + "loss": 1.65248108, + "memory(GiB)": 111.15, + "step": 32365, + "train_speed(iter/s)": 0.448245 + }, + { + "acc": 0.64384751, + "epoch": 0.8211567732115678, + "grad_norm": 5.4375, + "learning_rate": 6.847102676832253e-06, + "loss": 1.67847061, + "memory(GiB)": 111.15, + "step": 32370, + "train_speed(iter/s)": 0.448281 + }, + { + "acc": 0.64187651, + "epoch": 0.8212836123795028, + "grad_norm": 5.96875, + "learning_rate": 6.8461281897394615e-06, + "loss": 1.64848404, + "memory(GiB)": 111.15, + "step": 32375, + "train_speed(iter/s)": 0.448318 + }, + { + "acc": 0.63397183, + "epoch": 0.8214104515474379, + "grad_norm": 5.71875, + "learning_rate": 6.845153621445711e-06, + "loss": 1.68159142, + "memory(GiB)": 111.15, + "step": 32380, + "train_speed(iter/s)": 0.448354 + }, + { + "acc": 0.65238295, + "epoch": 0.821537290715373, + "grad_norm": 5.28125, + "learning_rate": 6.844178971993866e-06, + "loss": 1.67321091, + "memory(GiB)": 111.15, + "step": 32385, + "train_speed(iter/s)": 0.448391 + }, + { + "acc": 0.66326227, + "epoch": 0.821664129883308, + "grad_norm": 5.46875, + "learning_rate": 6.843204241426797e-06, + "loss": 1.59882946, + "memory(GiB)": 111.15, + "step": 32390, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.66306458, + "epoch": 0.821790969051243, + "grad_norm": 4.5625, + "learning_rate": 6.842229429787375e-06, + "loss": 1.61632442, + "memory(GiB)": 111.15, + "step": 32395, + "train_speed(iter/s)": 0.448464 + }, + { + "acc": 0.65038009, + "epoch": 0.821917808219178, + "grad_norm": 4.875, + "learning_rate": 6.841254537118477e-06, + "loss": 1.5503458, + "memory(GiB)": 111.15, + "step": 32400, + "train_speed(iter/s)": 0.448501 + }, + { + "epoch": 0.821917808219178, + "eval_acc": 0.6429685375261575, + "eval_loss": 1.6100209951400757, + "eval_runtime": 112.7379, + "eval_samples_per_second": 56.503, + "eval_steps_per_second": 28.251, + "step": 32400 + }, + { + "acc": 0.63776364, + "epoch": 0.8220446473871131, + "grad_norm": 5.34375, + "learning_rate": 6.840279563462985e-06, + "loss": 1.72160969, + "memory(GiB)": 111.15, + "step": 32405, + "train_speed(iter/s)": 0.44781 + }, + { + "acc": 0.64764876, + "epoch": 0.8221714865550482, + "grad_norm": 5.40625, + "learning_rate": 6.839304508863781e-06, + "loss": 1.62405815, + "memory(GiB)": 111.15, + "step": 32410, + "train_speed(iter/s)": 0.447846 + }, + { + "acc": 0.65508089, + "epoch": 0.8222983257229832, + "grad_norm": 4.625, + "learning_rate": 6.838329373363753e-06, + "loss": 1.62292938, + "memory(GiB)": 111.15, + "step": 32415, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.65851521, + "epoch": 0.8224251648909183, + "grad_norm": 4.46875, + "learning_rate": 6.8373541570057924e-06, + "loss": 1.61938, + "memory(GiB)": 111.15, + "step": 32420, + "train_speed(iter/s)": 0.447919 + }, + { + "acc": 0.65963554, + "epoch": 0.8225520040588534, + "grad_norm": 5.75, + "learning_rate": 6.836378859832791e-06, + "loss": 1.64844246, + "memory(GiB)": 111.15, + "step": 32425, + "train_speed(iter/s)": 0.447955 + }, + { + "acc": 0.64661956, + "epoch": 0.8226788432267884, + "grad_norm": 4.6875, + "learning_rate": 6.83540348188765e-06, + "loss": 1.62796707, + "memory(GiB)": 111.15, + "step": 32430, + "train_speed(iter/s)": 0.447991 + }, + { + "acc": 0.65333099, + "epoch": 0.8228056823947235, + "grad_norm": 6.90625, + "learning_rate": 6.834428023213268e-06, + "loss": 1.58288517, + "memory(GiB)": 111.15, + "step": 32435, + "train_speed(iter/s)": 0.448028 + }, + { + "acc": 0.64179029, + "epoch": 0.8229325215626585, + "grad_norm": 4.4375, + "learning_rate": 6.833452483852554e-06, + "loss": 1.60386391, + "memory(GiB)": 111.15, + "step": 32440, + "train_speed(iter/s)": 0.448064 + }, + { + "acc": 0.65531411, + "epoch": 0.8230593607305936, + "grad_norm": 6.46875, + "learning_rate": 6.832476863848411e-06, + "loss": 1.62659531, + "memory(GiB)": 111.15, + "step": 32445, + "train_speed(iter/s)": 0.4481 + }, + { + "acc": 0.63803382, + "epoch": 0.8231861998985287, + "grad_norm": 5.90625, + "learning_rate": 6.831501163243756e-06, + "loss": 1.70130196, + "memory(GiB)": 111.15, + "step": 32450, + "train_speed(iter/s)": 0.448137 + }, + { + "acc": 0.63967266, + "epoch": 0.8233130390664637, + "grad_norm": 6.71875, + "learning_rate": 6.830525382081501e-06, + "loss": 1.69018707, + "memory(GiB)": 111.15, + "step": 32455, + "train_speed(iter/s)": 0.448173 + }, + { + "acc": 0.64958649, + "epoch": 0.8234398782343988, + "grad_norm": 4.59375, + "learning_rate": 6.829549520404568e-06, + "loss": 1.60736275, + "memory(GiB)": 111.15, + "step": 32460, + "train_speed(iter/s)": 0.44821 + }, + { + "acc": 0.66044745, + "epoch": 0.8235667174023339, + "grad_norm": 5.1875, + "learning_rate": 6.828573578255879e-06, + "loss": 1.56229744, + "memory(GiB)": 111.15, + "step": 32465, + "train_speed(iter/s)": 0.448246 + }, + { + "acc": 0.64875484, + "epoch": 0.8236935565702689, + "grad_norm": 4.96875, + "learning_rate": 6.82759755567836e-06, + "loss": 1.62134933, + "memory(GiB)": 111.15, + "step": 32470, + "train_speed(iter/s)": 0.448282 + }, + { + "acc": 0.64224496, + "epoch": 0.823820395738204, + "grad_norm": 5.46875, + "learning_rate": 6.826621452714941e-06, + "loss": 1.70467224, + "memory(GiB)": 111.15, + "step": 32475, + "train_speed(iter/s)": 0.448318 + }, + { + "acc": 0.64410958, + "epoch": 0.823947234906139, + "grad_norm": 4.875, + "learning_rate": 6.825645269408556e-06, + "loss": 1.62777939, + "memory(GiB)": 111.15, + "step": 32480, + "train_speed(iter/s)": 0.448354 + }, + { + "acc": 0.64841433, + "epoch": 0.8240740740740741, + "grad_norm": 5.625, + "learning_rate": 6.82466900580214e-06, + "loss": 1.59784231, + "memory(GiB)": 111.15, + "step": 32485, + "train_speed(iter/s)": 0.44839 + }, + { + "acc": 0.63771162, + "epoch": 0.8242009132420092, + "grad_norm": 5.25, + "learning_rate": 6.823692661938634e-06, + "loss": 1.66963768, + "memory(GiB)": 111.15, + "step": 32490, + "train_speed(iter/s)": 0.448427 + }, + { + "acc": 0.65454922, + "epoch": 0.8243277524099442, + "grad_norm": 5.96875, + "learning_rate": 6.822716237860984e-06, + "loss": 1.64312172, + "memory(GiB)": 111.15, + "step": 32495, + "train_speed(iter/s)": 0.448462 + }, + { + "acc": 0.65863819, + "epoch": 0.8244545915778793, + "grad_norm": 4.65625, + "learning_rate": 6.821739733612135e-06, + "loss": 1.58759604, + "memory(GiB)": 111.15, + "step": 32500, + "train_speed(iter/s)": 0.448498 + }, + { + "epoch": 0.8244545915778793, + "eval_acc": 0.6429163166124753, + "eval_loss": 1.6101748943328857, + "eval_runtime": 114.4078, + "eval_samples_per_second": 55.678, + "eval_steps_per_second": 27.839, + "step": 32500 + }, + { + "acc": 0.65006766, + "epoch": 0.8245814307458144, + "grad_norm": 5.9375, + "learning_rate": 6.820763149235039e-06, + "loss": 1.69699669, + "memory(GiB)": 111.15, + "step": 32505, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.65078034, + "epoch": 0.8247082699137493, + "grad_norm": 5.0, + "learning_rate": 6.819786484772652e-06, + "loss": 1.60453625, + "memory(GiB)": 111.15, + "step": 32510, + "train_speed(iter/s)": 0.447835 + }, + { + "acc": 0.61758137, + "epoch": 0.8248351090816844, + "grad_norm": 5.40625, + "learning_rate": 6.8188097402679275e-06, + "loss": 1.73696442, + "memory(GiB)": 111.15, + "step": 32515, + "train_speed(iter/s)": 0.447871 + }, + { + "acc": 0.65949969, + "epoch": 0.8249619482496194, + "grad_norm": 4.8125, + "learning_rate": 6.817832915763833e-06, + "loss": 1.5610321, + "memory(GiB)": 111.15, + "step": 32520, + "train_speed(iter/s)": 0.447907 + }, + { + "acc": 0.64406414, + "epoch": 0.8250887874175545, + "grad_norm": 4.75, + "learning_rate": 6.81685601130333e-06, + "loss": 1.64460735, + "memory(GiB)": 111.15, + "step": 32525, + "train_speed(iter/s)": 0.447943 + }, + { + "acc": 0.66168027, + "epoch": 0.8252156265854896, + "grad_norm": 6.21875, + "learning_rate": 6.8158790269293885e-06, + "loss": 1.57091646, + "memory(GiB)": 111.15, + "step": 32530, + "train_speed(iter/s)": 0.447979 + }, + { + "acc": 0.66214733, + "epoch": 0.8253424657534246, + "grad_norm": 6.1875, + "learning_rate": 6.8149019626849785e-06, + "loss": 1.53996449, + "memory(GiB)": 111.15, + "step": 32535, + "train_speed(iter/s)": 0.448015 + }, + { + "acc": 0.64911966, + "epoch": 0.8254693049213597, + "grad_norm": 5.03125, + "learning_rate": 6.813924818613079e-06, + "loss": 1.63346252, + "memory(GiB)": 111.15, + "step": 32540, + "train_speed(iter/s)": 0.448051 + }, + { + "acc": 0.66107364, + "epoch": 0.8255961440892948, + "grad_norm": 5.53125, + "learning_rate": 6.812947594756667e-06, + "loss": 1.66344757, + "memory(GiB)": 111.15, + "step": 32545, + "train_speed(iter/s)": 0.448087 + }, + { + "acc": 0.6637958, + "epoch": 0.8257229832572298, + "grad_norm": 5.1875, + "learning_rate": 6.811970291158725e-06, + "loss": 1.60018368, + "memory(GiB)": 111.15, + "step": 32550, + "train_speed(iter/s)": 0.448123 + }, + { + "acc": 0.64954395, + "epoch": 0.8258498224251649, + "grad_norm": 5.09375, + "learning_rate": 6.810992907862239e-06, + "loss": 1.64513016, + "memory(GiB)": 111.15, + "step": 32555, + "train_speed(iter/s)": 0.448159 + }, + { + "acc": 0.63043914, + "epoch": 0.8259766615930999, + "grad_norm": 5.84375, + "learning_rate": 6.810015444910202e-06, + "loss": 1.64187317, + "memory(GiB)": 111.15, + "step": 32560, + "train_speed(iter/s)": 0.448196 + }, + { + "acc": 0.65492306, + "epoch": 0.826103500761035, + "grad_norm": 6.34375, + "learning_rate": 6.809037902345603e-06, + "loss": 1.67356892, + "memory(GiB)": 111.15, + "step": 32565, + "train_speed(iter/s)": 0.448232 + }, + { + "acc": 0.64298658, + "epoch": 0.8262303399289701, + "grad_norm": 6.03125, + "learning_rate": 6.808060280211439e-06, + "loss": 1.5911211, + "memory(GiB)": 111.15, + "step": 32570, + "train_speed(iter/s)": 0.448268 + }, + { + "acc": 0.65569539, + "epoch": 0.8263571790969051, + "grad_norm": 5.5, + "learning_rate": 6.807082578550713e-06, + "loss": 1.59871149, + "memory(GiB)": 111.15, + "step": 32575, + "train_speed(iter/s)": 0.448304 + }, + { + "acc": 0.6435678, + "epoch": 0.8264840182648402, + "grad_norm": 5.125, + "learning_rate": 6.806104797406428e-06, + "loss": 1.62056847, + "memory(GiB)": 111.15, + "step": 32580, + "train_speed(iter/s)": 0.44834 + }, + { + "acc": 0.64942517, + "epoch": 0.8266108574327753, + "grad_norm": 5.3125, + "learning_rate": 6.805126936821588e-06, + "loss": 1.61817665, + "memory(GiB)": 111.15, + "step": 32585, + "train_speed(iter/s)": 0.448376 + }, + { + "acc": 0.64908605, + "epoch": 0.8267376966007103, + "grad_norm": 5.4375, + "learning_rate": 6.804148996839208e-06, + "loss": 1.63608322, + "memory(GiB)": 111.15, + "step": 32590, + "train_speed(iter/s)": 0.448412 + }, + { + "acc": 0.64975762, + "epoch": 0.8268645357686454, + "grad_norm": 5.28125, + "learning_rate": 6.803170977502298e-06, + "loss": 1.66273613, + "memory(GiB)": 111.15, + "step": 32595, + "train_speed(iter/s)": 0.448449 + }, + { + "acc": 0.65923128, + "epoch": 0.8269913749365804, + "grad_norm": 7.78125, + "learning_rate": 6.802192878853879e-06, + "loss": 1.62350521, + "memory(GiB)": 111.15, + "step": 32600, + "train_speed(iter/s)": 0.448485 + }, + { + "epoch": 0.8269913749365804, + "eval_acc": 0.6429121389393807, + "eval_loss": 1.6098910570144653, + "eval_runtime": 114.671, + "eval_samples_per_second": 55.55, + "eval_steps_per_second": 27.775, + "step": 32600 + }, + { + "acc": 0.65719733, + "epoch": 0.8271182141045155, + "grad_norm": 5.71875, + "learning_rate": 6.801214700936972e-06, + "loss": 1.54453564, + "memory(GiB)": 111.15, + "step": 32605, + "train_speed(iter/s)": 0.447786 + }, + { + "acc": 0.65681901, + "epoch": 0.8272450532724506, + "grad_norm": 5.25, + "learning_rate": 6.8002364437946e-06, + "loss": 1.61418667, + "memory(GiB)": 111.15, + "step": 32610, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.64888945, + "epoch": 0.8273718924403856, + "grad_norm": 5.84375, + "learning_rate": 6.799258107469792e-06, + "loss": 1.66314201, + "memory(GiB)": 111.15, + "step": 32615, + "train_speed(iter/s)": 0.447858 + }, + { + "acc": 0.65021448, + "epoch": 0.8274987316083207, + "grad_norm": 4.25, + "learning_rate": 6.798279692005578e-06, + "loss": 1.60341358, + "memory(GiB)": 111.15, + "step": 32620, + "train_speed(iter/s)": 0.447893 + }, + { + "acc": 0.65714846, + "epoch": 0.8276255707762558, + "grad_norm": 5.75, + "learning_rate": 6.7973011974449965e-06, + "loss": 1.5717104, + "memory(GiB)": 111.15, + "step": 32625, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.65262489, + "epoch": 0.8277524099441907, + "grad_norm": 6.0, + "learning_rate": 6.796322623831082e-06, + "loss": 1.56863785, + "memory(GiB)": 111.15, + "step": 32630, + "train_speed(iter/s)": 0.447964 + }, + { + "acc": 0.65096188, + "epoch": 0.8278792491121258, + "grad_norm": 5.625, + "learning_rate": 6.795343971206879e-06, + "loss": 1.58098469, + "memory(GiB)": 111.15, + "step": 32635, + "train_speed(iter/s)": 0.448 + }, + { + "acc": 0.65714645, + "epoch": 0.8280060882800608, + "grad_norm": 6.90625, + "learning_rate": 6.794365239615433e-06, + "loss": 1.60648766, + "memory(GiB)": 111.15, + "step": 32640, + "train_speed(iter/s)": 0.448036 + }, + { + "acc": 0.65380702, + "epoch": 0.8281329274479959, + "grad_norm": 4.4375, + "learning_rate": 6.793386429099792e-06, + "loss": 1.63176231, + "memory(GiB)": 111.15, + "step": 32645, + "train_speed(iter/s)": 0.448073 + }, + { + "acc": 0.65917969, + "epoch": 0.828259766615931, + "grad_norm": 4.6875, + "learning_rate": 6.79240753970301e-06, + "loss": 1.60385113, + "memory(GiB)": 111.15, + "step": 32650, + "train_speed(iter/s)": 0.448108 + }, + { + "acc": 0.64168706, + "epoch": 0.828386605783866, + "grad_norm": 5.03125, + "learning_rate": 6.791428571468139e-06, + "loss": 1.67332458, + "memory(GiB)": 111.15, + "step": 32655, + "train_speed(iter/s)": 0.448145 + }, + { + "acc": 0.63637853, + "epoch": 0.8285134449518011, + "grad_norm": 6.5625, + "learning_rate": 6.7904495244382454e-06, + "loss": 1.68497143, + "memory(GiB)": 111.15, + "step": 32660, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.64454956, + "epoch": 0.8286402841197362, + "grad_norm": 4.625, + "learning_rate": 6.789470398656385e-06, + "loss": 1.6971344, + "memory(GiB)": 111.15, + "step": 32665, + "train_speed(iter/s)": 0.448217 + }, + { + "acc": 0.65143733, + "epoch": 0.8287671232876712, + "grad_norm": 5.40625, + "learning_rate": 6.788491194165629e-06, + "loss": 1.62084446, + "memory(GiB)": 111.15, + "step": 32670, + "train_speed(iter/s)": 0.448253 + }, + { + "acc": 0.67059383, + "epoch": 0.8288939624556063, + "grad_norm": 5.34375, + "learning_rate": 6.787511911009044e-06, + "loss": 1.52378235, + "memory(GiB)": 111.15, + "step": 32675, + "train_speed(iter/s)": 0.448289 + }, + { + "acc": 0.63780518, + "epoch": 0.8290208016235413, + "grad_norm": 5.0625, + "learning_rate": 6.786532549229704e-06, + "loss": 1.6726038, + "memory(GiB)": 111.15, + "step": 32680, + "train_speed(iter/s)": 0.448325 + }, + { + "acc": 0.6582231, + "epoch": 0.8291476407914764, + "grad_norm": 5.15625, + "learning_rate": 6.785553108870686e-06, + "loss": 1.57881584, + "memory(GiB)": 111.15, + "step": 32685, + "train_speed(iter/s)": 0.448361 + }, + { + "acc": 0.63775282, + "epoch": 0.8292744799594115, + "grad_norm": 5.0, + "learning_rate": 6.784573589975072e-06, + "loss": 1.64340076, + "memory(GiB)": 111.15, + "step": 32690, + "train_speed(iter/s)": 0.448398 + }, + { + "acc": 0.63034964, + "epoch": 0.8294013191273465, + "grad_norm": 6.09375, + "learning_rate": 6.783593992585943e-06, + "loss": 1.69854774, + "memory(GiB)": 111.15, + "step": 32695, + "train_speed(iter/s)": 0.448433 + }, + { + "acc": 0.64613237, + "epoch": 0.8295281582952816, + "grad_norm": 5.21875, + "learning_rate": 6.7826143167463876e-06, + "loss": 1.66833515, + "memory(GiB)": 111.15, + "step": 32700, + "train_speed(iter/s)": 0.448469 + }, + { + "epoch": 0.8295281582952816, + "eval_acc": 0.6430086431878653, + "eval_loss": 1.6100075244903564, + "eval_runtime": 113.2712, + "eval_samples_per_second": 56.237, + "eval_steps_per_second": 28.118, + "step": 32700 + }, + { + "acc": 0.65465412, + "epoch": 0.8296549974632167, + "grad_norm": 4.9375, + "learning_rate": 6.781634562499495e-06, + "loss": 1.64357262, + "memory(GiB)": 111.15, + "step": 32705, + "train_speed(iter/s)": 0.447781 + }, + { + "acc": 0.65387259, + "epoch": 0.8297818366311517, + "grad_norm": 5.46875, + "learning_rate": 6.780654729888361e-06, + "loss": 1.6355978, + "memory(GiB)": 111.15, + "step": 32710, + "train_speed(iter/s)": 0.447817 + }, + { + "acc": 0.65253215, + "epoch": 0.8299086757990868, + "grad_norm": 6.875, + "learning_rate": 6.779674818956081e-06, + "loss": 1.63154449, + "memory(GiB)": 111.15, + "step": 32715, + "train_speed(iter/s)": 0.447852 + }, + { + "acc": 0.6314887, + "epoch": 0.8300355149670218, + "grad_norm": 5.75, + "learning_rate": 6.778694829745756e-06, + "loss": 1.64977112, + "memory(GiB)": 111.15, + "step": 32720, + "train_speed(iter/s)": 0.447888 + }, + { + "acc": 0.63668919, + "epoch": 0.8301623541349569, + "grad_norm": 5.34375, + "learning_rate": 6.777714762300492e-06, + "loss": 1.6203289, + "memory(GiB)": 111.15, + "step": 32725, + "train_speed(iter/s)": 0.447924 + }, + { + "acc": 0.66070194, + "epoch": 0.830289193302892, + "grad_norm": 5.15625, + "learning_rate": 6.776734616663397e-06, + "loss": 1.58673019, + "memory(GiB)": 111.15, + "step": 32730, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.66710291, + "epoch": 0.830416032470827, + "grad_norm": 4.96875, + "learning_rate": 6.77575439287758e-06, + "loss": 1.50803852, + "memory(GiB)": 111.15, + "step": 32735, + "train_speed(iter/s)": 0.447995 + }, + { + "acc": 0.64855824, + "epoch": 0.8305428716387621, + "grad_norm": 5.59375, + "learning_rate": 6.774774090986157e-06, + "loss": 1.6296299, + "memory(GiB)": 111.15, + "step": 32740, + "train_speed(iter/s)": 0.448031 + }, + { + "acc": 0.66819649, + "epoch": 0.8306697108066972, + "grad_norm": 5.03125, + "learning_rate": 6.773793711032244e-06, + "loss": 1.61443233, + "memory(GiB)": 111.15, + "step": 32745, + "train_speed(iter/s)": 0.448067 + }, + { + "acc": 0.64040546, + "epoch": 0.8307965499746321, + "grad_norm": 5.59375, + "learning_rate": 6.772813253058965e-06, + "loss": 1.65505238, + "memory(GiB)": 111.15, + "step": 32750, + "train_speed(iter/s)": 0.448102 + }, + { + "acc": 0.6573648, + "epoch": 0.8309233891425672, + "grad_norm": 6.0, + "learning_rate": 6.771832717109444e-06, + "loss": 1.62006741, + "memory(GiB)": 111.15, + "step": 32755, + "train_speed(iter/s)": 0.448137 + }, + { + "acc": 0.6478117, + "epoch": 0.8310502283105022, + "grad_norm": 4.6875, + "learning_rate": 6.77085210322681e-06, + "loss": 1.66119843, + "memory(GiB)": 111.15, + "step": 32760, + "train_speed(iter/s)": 0.448173 + }, + { + "acc": 0.63774009, + "epoch": 0.8311770674784373, + "grad_norm": 5.0625, + "learning_rate": 6.769871411454195e-06, + "loss": 1.65243301, + "memory(GiB)": 111.15, + "step": 32765, + "train_speed(iter/s)": 0.448208 + }, + { + "acc": 0.6276474, + "epoch": 0.8313039066463724, + "grad_norm": 5.3125, + "learning_rate": 6.768890641834732e-06, + "loss": 1.68650894, + "memory(GiB)": 111.15, + "step": 32770, + "train_speed(iter/s)": 0.448244 + }, + { + "acc": 0.64909286, + "epoch": 0.8314307458143074, + "grad_norm": 4.53125, + "learning_rate": 6.767909794411562e-06, + "loss": 1.63493137, + "memory(GiB)": 111.15, + "step": 32775, + "train_speed(iter/s)": 0.44828 + }, + { + "acc": 0.66092644, + "epoch": 0.8315575849822425, + "grad_norm": 7.5, + "learning_rate": 6.7669288692278256e-06, + "loss": 1.62956963, + "memory(GiB)": 111.15, + "step": 32780, + "train_speed(iter/s)": 0.448316 + }, + { + "acc": 0.64859414, + "epoch": 0.8316844241501776, + "grad_norm": 5.0625, + "learning_rate": 6.76594786632667e-06, + "loss": 1.55125647, + "memory(GiB)": 111.15, + "step": 32785, + "train_speed(iter/s)": 0.448351 + }, + { + "acc": 0.64908266, + "epoch": 0.8318112633181126, + "grad_norm": 5.375, + "learning_rate": 6.764966785751242e-06, + "loss": 1.64024887, + "memory(GiB)": 111.15, + "step": 32790, + "train_speed(iter/s)": 0.448386 + }, + { + "acc": 0.6502162, + "epoch": 0.8319381024860477, + "grad_norm": 4.9375, + "learning_rate": 6.763985627544693e-06, + "loss": 1.64023151, + "memory(GiB)": 111.15, + "step": 32795, + "train_speed(iter/s)": 0.448422 + }, + { + "acc": 0.63774815, + "epoch": 0.8320649416539827, + "grad_norm": 5.25, + "learning_rate": 6.763004391750183e-06, + "loss": 1.64442825, + "memory(GiB)": 111.15, + "step": 32800, + "train_speed(iter/s)": 0.448458 + }, + { + "epoch": 0.8320649416539827, + "eval_acc": 0.6430132386282694, + "eval_loss": 1.6097626686096191, + "eval_runtime": 114.6915, + "eval_samples_per_second": 55.54, + "eval_steps_per_second": 27.77, + "step": 32800 + }, + { + "acc": 0.65060616, + "epoch": 0.8321917808219178, + "grad_norm": 7.375, + "learning_rate": 6.762023078410867e-06, + "loss": 1.66409531, + "memory(GiB)": 111.15, + "step": 32805, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.65033741, + "epoch": 0.8323186199898529, + "grad_norm": 4.65625, + "learning_rate": 6.7610416875699095e-06, + "loss": 1.60913353, + "memory(GiB)": 111.15, + "step": 32810, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.66197157, + "epoch": 0.8324454591577879, + "grad_norm": 4.875, + "learning_rate": 6.760060219270476e-06, + "loss": 1.6297905, + "memory(GiB)": 111.15, + "step": 32815, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.66332273, + "epoch": 0.832572298325723, + "grad_norm": 5.1875, + "learning_rate": 6.759078673555736e-06, + "loss": 1.62605858, + "memory(GiB)": 111.15, + "step": 32820, + "train_speed(iter/s)": 0.44787 + }, + { + "acc": 0.64675913, + "epoch": 0.8326991374936581, + "grad_norm": 5.875, + "learning_rate": 6.758097050468862e-06, + "loss": 1.61459904, + "memory(GiB)": 111.15, + "step": 32825, + "train_speed(iter/s)": 0.447905 + }, + { + "acc": 0.65628443, + "epoch": 0.8328259766615931, + "grad_norm": 5.59375, + "learning_rate": 6.757115350053032e-06, + "loss": 1.66839905, + "memory(GiB)": 111.15, + "step": 32830, + "train_speed(iter/s)": 0.447904 + }, + { + "acc": 0.64399061, + "epoch": 0.8329528158295282, + "grad_norm": 4.84375, + "learning_rate": 6.756133572351422e-06, + "loss": 1.63957481, + "memory(GiB)": 111.15, + "step": 32835, + "train_speed(iter/s)": 0.44794 + }, + { + "acc": 0.65229459, + "epoch": 0.8330796549974632, + "grad_norm": 5.375, + "learning_rate": 6.755151717407218e-06, + "loss": 1.58449955, + "memory(GiB)": 111.15, + "step": 32840, + "train_speed(iter/s)": 0.447975 + }, + { + "acc": 0.64108706, + "epoch": 0.8332064941653983, + "grad_norm": 4.53125, + "learning_rate": 6.754169785263605e-06, + "loss": 1.69349861, + "memory(GiB)": 111.15, + "step": 32845, + "train_speed(iter/s)": 0.448011 + }, + { + "acc": 0.65346699, + "epoch": 0.8333333333333334, + "grad_norm": 6.21875, + "learning_rate": 6.753187775963773e-06, + "loss": 1.62150764, + "memory(GiB)": 111.15, + "step": 32850, + "train_speed(iter/s)": 0.448047 + }, + { + "acc": 0.65815678, + "epoch": 0.8334601725012684, + "grad_norm": 5.90625, + "learning_rate": 6.752205689550915e-06, + "loss": 1.56463375, + "memory(GiB)": 111.15, + "step": 32855, + "train_speed(iter/s)": 0.448082 + }, + { + "acc": 0.65201516, + "epoch": 0.8335870116692035, + "grad_norm": 5.0, + "learning_rate": 6.751223526068228e-06, + "loss": 1.62399406, + "memory(GiB)": 111.15, + "step": 32860, + "train_speed(iter/s)": 0.448118 + }, + { + "acc": 0.63955178, + "epoch": 0.8337138508371386, + "grad_norm": 5.0625, + "learning_rate": 6.75024128555891e-06, + "loss": 1.65600052, + "memory(GiB)": 111.15, + "step": 32865, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.64648237, + "epoch": 0.8338406900050735, + "grad_norm": 4.84375, + "learning_rate": 6.7492589680661695e-06, + "loss": 1.63914108, + "memory(GiB)": 111.15, + "step": 32870, + "train_speed(iter/s)": 0.44819 + }, + { + "acc": 0.64333458, + "epoch": 0.8339675291730086, + "grad_norm": 5.21875, + "learning_rate": 6.748276573633207e-06, + "loss": 1.66612473, + "memory(GiB)": 111.15, + "step": 32875, + "train_speed(iter/s)": 0.448225 + }, + { + "acc": 0.65232787, + "epoch": 0.8340943683409436, + "grad_norm": 5.09375, + "learning_rate": 6.747294102303237e-06, + "loss": 1.62393417, + "memory(GiB)": 111.15, + "step": 32880, + "train_speed(iter/s)": 0.448261 + }, + { + "acc": 0.66196537, + "epoch": 0.8342212075088787, + "grad_norm": 5.40625, + "learning_rate": 6.746311554119469e-06, + "loss": 1.58140163, + "memory(GiB)": 111.15, + "step": 32885, + "train_speed(iter/s)": 0.448296 + }, + { + "acc": 0.66298809, + "epoch": 0.8343480466768138, + "grad_norm": 5.875, + "learning_rate": 6.745328929125125e-06, + "loss": 1.58378067, + "memory(GiB)": 111.15, + "step": 32890, + "train_speed(iter/s)": 0.448332 + }, + { + "acc": 0.64894247, + "epoch": 0.8344748858447488, + "grad_norm": 5.375, + "learning_rate": 6.7443462273634195e-06, + "loss": 1.54763718, + "memory(GiB)": 111.15, + "step": 32895, + "train_speed(iter/s)": 0.448368 + }, + { + "acc": 0.64872894, + "epoch": 0.8346017250126839, + "grad_norm": 5.875, + "learning_rate": 6.74336344887758e-06, + "loss": 1.59882526, + "memory(GiB)": 111.15, + "step": 32900, + "train_speed(iter/s)": 0.448404 + }, + { + "epoch": 0.8346017250126839, + "eval_acc": 0.6430529265226679, + "eval_loss": 1.61006760597229, + "eval_runtime": 112.7572, + "eval_samples_per_second": 56.493, + "eval_steps_per_second": 28.247, + "step": 32900 + }, + { + "acc": 0.64996443, + "epoch": 0.834728564180619, + "grad_norm": 5.625, + "learning_rate": 6.742380593710834e-06, + "loss": 1.62278862, + "memory(GiB)": 111.15, + "step": 32905, + "train_speed(iter/s)": 0.447724 + }, + { + "acc": 0.64250236, + "epoch": 0.834855403348554, + "grad_norm": 5.96875, + "learning_rate": 6.7413976619064085e-06, + "loss": 1.68773212, + "memory(GiB)": 111.15, + "step": 32910, + "train_speed(iter/s)": 0.447759 + }, + { + "acc": 0.65227995, + "epoch": 0.8349822425164891, + "grad_norm": 5.40625, + "learning_rate": 6.74041465350754e-06, + "loss": 1.5280241, + "memory(GiB)": 111.15, + "step": 32915, + "train_speed(iter/s)": 0.447794 + }, + { + "acc": 0.66012201, + "epoch": 0.8351090816844241, + "grad_norm": 5.03125, + "learning_rate": 6.739431568557464e-06, + "loss": 1.53043833, + "memory(GiB)": 111.15, + "step": 32920, + "train_speed(iter/s)": 0.44783 + }, + { + "acc": 0.63174486, + "epoch": 0.8352359208523592, + "grad_norm": 5.0, + "learning_rate": 6.738448407099423e-06, + "loss": 1.71644936, + "memory(GiB)": 111.15, + "step": 32925, + "train_speed(iter/s)": 0.447865 + }, + { + "acc": 0.64693007, + "epoch": 0.8353627600202943, + "grad_norm": 5.59375, + "learning_rate": 6.737465169176658e-06, + "loss": 1.57929029, + "memory(GiB)": 111.15, + "step": 32930, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.64836311, + "epoch": 0.8354895991882293, + "grad_norm": 5.375, + "learning_rate": 6.736481854832418e-06, + "loss": 1.63191147, + "memory(GiB)": 111.15, + "step": 32935, + "train_speed(iter/s)": 0.447936 + }, + { + "acc": 0.64984336, + "epoch": 0.8356164383561644, + "grad_norm": 6.3125, + "learning_rate": 6.735498464109953e-06, + "loss": 1.61379204, + "memory(GiB)": 111.15, + "step": 32940, + "train_speed(iter/s)": 0.447971 + }, + { + "acc": 0.65568976, + "epoch": 0.8357432775240995, + "grad_norm": 6.25, + "learning_rate": 6.734514997052517e-06, + "loss": 1.58366652, + "memory(GiB)": 111.15, + "step": 32945, + "train_speed(iter/s)": 0.448007 + }, + { + "acc": 0.66102796, + "epoch": 0.8358701166920345, + "grad_norm": 6.21875, + "learning_rate": 6.733531453703368e-06, + "loss": 1.61800709, + "memory(GiB)": 111.15, + "step": 32950, + "train_speed(iter/s)": 0.448042 + }, + { + "acc": 0.6513133, + "epoch": 0.8359969558599696, + "grad_norm": 6.28125, + "learning_rate": 6.732547834105765e-06, + "loss": 1.55495281, + "memory(GiB)": 111.15, + "step": 32955, + "train_speed(iter/s)": 0.448078 + }, + { + "acc": 0.65970721, + "epoch": 0.8361237950279046, + "grad_norm": 5.59375, + "learning_rate": 6.731564138302975e-06, + "loss": 1.60372982, + "memory(GiB)": 111.15, + "step": 32960, + "train_speed(iter/s)": 0.448114 + }, + { + "acc": 0.65962839, + "epoch": 0.8362506341958397, + "grad_norm": 4.875, + "learning_rate": 6.730580366338261e-06, + "loss": 1.60945663, + "memory(GiB)": 111.15, + "step": 32965, + "train_speed(iter/s)": 0.448149 + }, + { + "acc": 0.66402388, + "epoch": 0.8363774733637748, + "grad_norm": 4.75, + "learning_rate": 6.729596518254897e-06, + "loss": 1.56634359, + "memory(GiB)": 111.15, + "step": 32970, + "train_speed(iter/s)": 0.448184 + }, + { + "acc": 0.64052973, + "epoch": 0.8365043125317098, + "grad_norm": 6.34375, + "learning_rate": 6.728612594096155e-06, + "loss": 1.61817017, + "memory(GiB)": 111.15, + "step": 32975, + "train_speed(iter/s)": 0.44822 + }, + { + "acc": 0.6338439, + "epoch": 0.8366311516996449, + "grad_norm": 4.59375, + "learning_rate": 6.727628593905315e-06, + "loss": 1.66176434, + "memory(GiB)": 111.15, + "step": 32980, + "train_speed(iter/s)": 0.448256 + }, + { + "acc": 0.66152577, + "epoch": 0.83675799086758, + "grad_norm": 5.71875, + "learning_rate": 6.726644517725655e-06, + "loss": 1.63252106, + "memory(GiB)": 111.15, + "step": 32985, + "train_speed(iter/s)": 0.448291 + }, + { + "acc": 0.62931762, + "epoch": 0.836884830035515, + "grad_norm": 6.0, + "learning_rate": 6.725660365600462e-06, + "loss": 1.70414371, + "memory(GiB)": 111.15, + "step": 32990, + "train_speed(iter/s)": 0.448326 + }, + { + "acc": 0.63023186, + "epoch": 0.83701166920345, + "grad_norm": 5.96875, + "learning_rate": 6.724676137573021e-06, + "loss": 1.66071396, + "memory(GiB)": 111.15, + "step": 32995, + "train_speed(iter/s)": 0.448362 + }, + { + "acc": 0.62570782, + "epoch": 0.837138508371385, + "grad_norm": 6.78125, + "learning_rate": 6.723691833686622e-06, + "loss": 1.65840912, + "memory(GiB)": 111.15, + "step": 33000, + "train_speed(iter/s)": 0.448397 + }, + { + "epoch": 0.837138508371385, + "eval_acc": 0.6430942854863041, + "eval_loss": 1.6098781824111938, + "eval_runtime": 113.3463, + "eval_samples_per_second": 56.199, + "eval_steps_per_second": 28.1, + "step": 33000 + }, + { + "acc": 0.64703274, + "epoch": 0.8372653475393201, + "grad_norm": 6.84375, + "learning_rate": 6.722707453984561e-06, + "loss": 1.65039902, + "memory(GiB)": 111.15, + "step": 33005, + "train_speed(iter/s)": 0.447716 + }, + { + "acc": 0.66382532, + "epoch": 0.8373921867072552, + "grad_norm": 5.625, + "learning_rate": 6.721722998510135e-06, + "loss": 1.64579926, + "memory(GiB)": 111.15, + "step": 33010, + "train_speed(iter/s)": 0.44775 + }, + { + "acc": 0.65656996, + "epoch": 0.8375190258751902, + "grad_norm": 5.0, + "learning_rate": 6.720738467306644e-06, + "loss": 1.58595943, + "memory(GiB)": 111.15, + "step": 33015, + "train_speed(iter/s)": 0.447786 + }, + { + "acc": 0.65878968, + "epoch": 0.8376458650431253, + "grad_norm": 5.09375, + "learning_rate": 6.719753860417394e-06, + "loss": 1.65688725, + "memory(GiB)": 111.15, + "step": 33020, + "train_speed(iter/s)": 0.447821 + }, + { + "acc": 0.64140415, + "epoch": 0.8377727042110604, + "grad_norm": 6.0625, + "learning_rate": 6.718769177885689e-06, + "loss": 1.67757874, + "memory(GiB)": 111.15, + "step": 33025, + "train_speed(iter/s)": 0.447856 + }, + { + "acc": 0.67006321, + "epoch": 0.8378995433789954, + "grad_norm": 4.75, + "learning_rate": 6.717784419754845e-06, + "loss": 1.52639618, + "memory(GiB)": 111.15, + "step": 33030, + "train_speed(iter/s)": 0.447892 + }, + { + "acc": 0.65474906, + "epoch": 0.8380263825469305, + "grad_norm": 5.46875, + "learning_rate": 6.71679958606817e-06, + "loss": 1.5921545, + "memory(GiB)": 111.15, + "step": 33035, + "train_speed(iter/s)": 0.447927 + }, + { + "acc": 0.63418345, + "epoch": 0.8381532217148655, + "grad_norm": 5.46875, + "learning_rate": 6.715814676868985e-06, + "loss": 1.68496685, + "memory(GiB)": 111.15, + "step": 33040, + "train_speed(iter/s)": 0.447962 + }, + { + "acc": 0.63637309, + "epoch": 0.8382800608828006, + "grad_norm": 4.8125, + "learning_rate": 6.714829692200611e-06, + "loss": 1.69638748, + "memory(GiB)": 111.15, + "step": 33045, + "train_speed(iter/s)": 0.447997 + }, + { + "acc": 0.64222589, + "epoch": 0.8384069000507357, + "grad_norm": 7.03125, + "learning_rate": 6.71384463210637e-06, + "loss": 1.70058556, + "memory(GiB)": 111.15, + "step": 33050, + "train_speed(iter/s)": 0.448033 + }, + { + "acc": 0.64808073, + "epoch": 0.8385337392186707, + "grad_norm": 5.15625, + "learning_rate": 6.7128594966295904e-06, + "loss": 1.60938702, + "memory(GiB)": 111.15, + "step": 33055, + "train_speed(iter/s)": 0.448068 + }, + { + "acc": 0.65790215, + "epoch": 0.8386605783866058, + "grad_norm": 5.71875, + "learning_rate": 6.711874285813602e-06, + "loss": 1.59583702, + "memory(GiB)": 111.15, + "step": 33060, + "train_speed(iter/s)": 0.448104 + }, + { + "acc": 0.65777206, + "epoch": 0.8387874175545409, + "grad_norm": 4.65625, + "learning_rate": 6.710888999701741e-06, + "loss": 1.63709354, + "memory(GiB)": 111.15, + "step": 33065, + "train_speed(iter/s)": 0.448139 + }, + { + "acc": 0.66115046, + "epoch": 0.8389142567224759, + "grad_norm": 5.40625, + "learning_rate": 6.7099036383373425e-06, + "loss": 1.56979275, + "memory(GiB)": 111.15, + "step": 33070, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.64699354, + "epoch": 0.839041095890411, + "grad_norm": 5.34375, + "learning_rate": 6.708918201763748e-06, + "loss": 1.65200806, + "memory(GiB)": 111.15, + "step": 33075, + "train_speed(iter/s)": 0.448209 + }, + { + "acc": 0.65511065, + "epoch": 0.839167935058346, + "grad_norm": 4.875, + "learning_rate": 6.707932690024302e-06, + "loss": 1.6356308, + "memory(GiB)": 111.15, + "step": 33080, + "train_speed(iter/s)": 0.448244 + }, + { + "acc": 0.6490922, + "epoch": 0.8392947742262811, + "grad_norm": 5.40625, + "learning_rate": 6.706947103162348e-06, + "loss": 1.61149139, + "memory(GiB)": 111.15, + "step": 33085, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.65337124, + "epoch": 0.8394216133942162, + "grad_norm": 5.6875, + "learning_rate": 6.7059614412212425e-06, + "loss": 1.67223358, + "memory(GiB)": 111.15, + "step": 33090, + "train_speed(iter/s)": 0.448315 + }, + { + "acc": 0.65159082, + "epoch": 0.8395484525621512, + "grad_norm": 5.0625, + "learning_rate": 6.704975704244334e-06, + "loss": 1.62722168, + "memory(GiB)": 111.15, + "step": 33095, + "train_speed(iter/s)": 0.44835 + }, + { + "acc": 0.64347687, + "epoch": 0.8396752917300863, + "grad_norm": 5.28125, + "learning_rate": 6.703989892274985e-06, + "loss": 1.68422127, + "memory(GiB)": 111.15, + "step": 33100, + "train_speed(iter/s)": 0.448385 + }, + { + "epoch": 0.8396752917300863, + "eval_acc": 0.6430800813977826, + "eval_loss": 1.6098905801773071, + "eval_runtime": 115.2119, + "eval_samples_per_second": 55.289, + "eval_steps_per_second": 27.645, + "step": 33100 + }, + { + "acc": 0.6633316, + "epoch": 0.8398021308980214, + "grad_norm": 6.25, + "learning_rate": 6.703004005356549e-06, + "loss": 1.58559914, + "memory(GiB)": 111.15, + "step": 33105, + "train_speed(iter/s)": 0.447694 + }, + { + "acc": 0.62422762, + "epoch": 0.8399289700659563, + "grad_norm": 6.15625, + "learning_rate": 6.7020180435323965e-06, + "loss": 1.68098927, + "memory(GiB)": 111.15, + "step": 33110, + "train_speed(iter/s)": 0.447729 + }, + { + "acc": 0.62980299, + "epoch": 0.8400558092338914, + "grad_norm": 5.4375, + "learning_rate": 6.701032006845889e-06, + "loss": 1.66067467, + "memory(GiB)": 111.15, + "step": 33115, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.66773529, + "epoch": 0.8401826484018264, + "grad_norm": 5.78125, + "learning_rate": 6.700045895340401e-06, + "loss": 1.56850796, + "memory(GiB)": 111.15, + "step": 33120, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.64705577, + "epoch": 0.8403094875697615, + "grad_norm": 8.0, + "learning_rate": 6.699059709059304e-06, + "loss": 1.66024208, + "memory(GiB)": 111.15, + "step": 33125, + "train_speed(iter/s)": 0.447835 + }, + { + "acc": 0.66956596, + "epoch": 0.8404363267376966, + "grad_norm": 5.25, + "learning_rate": 6.698073448045975e-06, + "loss": 1.6202116, + "memory(GiB)": 111.15, + "step": 33130, + "train_speed(iter/s)": 0.44787 + }, + { + "acc": 0.66566906, + "epoch": 0.8405631659056316, + "grad_norm": 5.0625, + "learning_rate": 6.697087112343795e-06, + "loss": 1.55232964, + "memory(GiB)": 111.15, + "step": 33135, + "train_speed(iter/s)": 0.447906 + }, + { + "acc": 0.63859696, + "epoch": 0.8406900050735667, + "grad_norm": 5.8125, + "learning_rate": 6.696100701996146e-06, + "loss": 1.70770493, + "memory(GiB)": 111.15, + "step": 33140, + "train_speed(iter/s)": 0.447941 + }, + { + "acc": 0.65811925, + "epoch": 0.8408168442415018, + "grad_norm": 5.8125, + "learning_rate": 6.6951142170464164e-06, + "loss": 1.58993721, + "memory(GiB)": 111.15, + "step": 33145, + "train_speed(iter/s)": 0.447977 + }, + { + "acc": 0.64766684, + "epoch": 0.8409436834094368, + "grad_norm": 5.5625, + "learning_rate": 6.694127657537995e-06, + "loss": 1.64419041, + "memory(GiB)": 111.15, + "step": 33150, + "train_speed(iter/s)": 0.448012 + }, + { + "acc": 0.64497776, + "epoch": 0.8410705225773719, + "grad_norm": 4.9375, + "learning_rate": 6.693141023514276e-06, + "loss": 1.6223114, + "memory(GiB)": 111.15, + "step": 33155, + "train_speed(iter/s)": 0.448047 + }, + { + "acc": 0.64539342, + "epoch": 0.8411973617453069, + "grad_norm": 5.375, + "learning_rate": 6.6921543150186555e-06, + "loss": 1.65548916, + "memory(GiB)": 111.15, + "step": 33160, + "train_speed(iter/s)": 0.448082 + }, + { + "acc": 0.65835824, + "epoch": 0.841324200913242, + "grad_norm": 5.15625, + "learning_rate": 6.691167532094531e-06, + "loss": 1.61784248, + "memory(GiB)": 111.15, + "step": 33165, + "train_speed(iter/s)": 0.448118 + }, + { + "acc": 0.6581017, + "epoch": 0.8414510400811771, + "grad_norm": 7.03125, + "learning_rate": 6.690180674785311e-06, + "loss": 1.70589447, + "memory(GiB)": 111.15, + "step": 33170, + "train_speed(iter/s)": 0.448153 + }, + { + "acc": 0.63799458, + "epoch": 0.8415778792491121, + "grad_norm": 5.1875, + "learning_rate": 6.689193743134397e-06, + "loss": 1.63857613, + "memory(GiB)": 111.15, + "step": 33175, + "train_speed(iter/s)": 0.448188 + }, + { + "acc": 0.64314218, + "epoch": 0.8417047184170472, + "grad_norm": 5.90625, + "learning_rate": 6.688206737185201e-06, + "loss": 1.6137516, + "memory(GiB)": 111.15, + "step": 33180, + "train_speed(iter/s)": 0.448224 + }, + { + "acc": 0.64059944, + "epoch": 0.8418315575849823, + "grad_norm": 5.78125, + "learning_rate": 6.687219656981135e-06, + "loss": 1.61017551, + "memory(GiB)": 111.15, + "step": 33185, + "train_speed(iter/s)": 0.44826 + }, + { + "acc": 0.67158375, + "epoch": 0.8419583967529173, + "grad_norm": 5.625, + "learning_rate": 6.686232502565616e-06, + "loss": 1.59326458, + "memory(GiB)": 111.15, + "step": 33190, + "train_speed(iter/s)": 0.448295 + }, + { + "acc": 0.65127487, + "epoch": 0.8420852359208524, + "grad_norm": 5.34375, + "learning_rate": 6.685245273982063e-06, + "loss": 1.66372623, + "memory(GiB)": 111.15, + "step": 33195, + "train_speed(iter/s)": 0.44833 + }, + { + "acc": 0.63178692, + "epoch": 0.8422120750887874, + "grad_norm": 5.59375, + "learning_rate": 6.684257971273899e-06, + "loss": 1.72483921, + "memory(GiB)": 111.15, + "step": 33200, + "train_speed(iter/s)": 0.448366 + }, + { + "epoch": 0.8422120750887874, + "eval_acc": 0.6430345447610517, + "eval_loss": 1.6096926927566528, + "eval_runtime": 114.2875, + "eval_samples_per_second": 55.737, + "eval_steps_per_second": 27.868, + "step": 33200 + }, + { + "acc": 0.65709572, + "epoch": 0.8423389142567225, + "grad_norm": 4.9375, + "learning_rate": 6.68327059448455e-06, + "loss": 1.59193382, + "memory(GiB)": 111.15, + "step": 33205, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.66269197, + "epoch": 0.8424657534246576, + "grad_norm": 6.28125, + "learning_rate": 6.682283143657444e-06, + "loss": 1.57913322, + "memory(GiB)": 111.15, + "step": 33210, + "train_speed(iter/s)": 0.447717 + }, + { + "acc": 0.64450436, + "epoch": 0.8425925925925926, + "grad_norm": 7.59375, + "learning_rate": 6.681295618836015e-06, + "loss": 1.69801826, + "memory(GiB)": 111.15, + "step": 33215, + "train_speed(iter/s)": 0.447752 + }, + { + "acc": 0.63954353, + "epoch": 0.8427194317605277, + "grad_norm": 4.9375, + "learning_rate": 6.680308020063699e-06, + "loss": 1.61760445, + "memory(GiB)": 111.15, + "step": 33220, + "train_speed(iter/s)": 0.447787 + }, + { + "acc": 0.64750366, + "epoch": 0.8428462709284628, + "grad_norm": 5.71875, + "learning_rate": 6.679320347383933e-06, + "loss": 1.65201035, + "memory(GiB)": 111.15, + "step": 33225, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.65556054, + "epoch": 0.8429731100963977, + "grad_norm": 6.90625, + "learning_rate": 6.678332600840161e-06, + "loss": 1.64039726, + "memory(GiB)": 111.15, + "step": 33230, + "train_speed(iter/s)": 0.447858 + }, + { + "acc": 0.66262279, + "epoch": 0.8430999492643328, + "grad_norm": 5.21875, + "learning_rate": 6.677344780475827e-06, + "loss": 1.57748775, + "memory(GiB)": 111.15, + "step": 33235, + "train_speed(iter/s)": 0.447893 + }, + { + "acc": 0.66933451, + "epoch": 0.8432267884322678, + "grad_norm": 4.96875, + "learning_rate": 6.676356886334383e-06, + "loss": 1.58385143, + "memory(GiB)": 111.15, + "step": 33240, + "train_speed(iter/s)": 0.447929 + }, + { + "acc": 0.63872638, + "epoch": 0.8433536276002029, + "grad_norm": 5.78125, + "learning_rate": 6.675368918459276e-06, + "loss": 1.69033852, + "memory(GiB)": 111.15, + "step": 33245, + "train_speed(iter/s)": 0.447964 + }, + { + "acc": 0.66944065, + "epoch": 0.843480466768138, + "grad_norm": 5.03125, + "learning_rate": 6.674380876893967e-06, + "loss": 1.54368095, + "memory(GiB)": 111.15, + "step": 33250, + "train_speed(iter/s)": 0.447999 + }, + { + "acc": 0.65621281, + "epoch": 0.843607305936073, + "grad_norm": 6.625, + "learning_rate": 6.673392761681908e-06, + "loss": 1.58544693, + "memory(GiB)": 111.15, + "step": 33255, + "train_speed(iter/s)": 0.448034 + }, + { + "acc": 0.66579866, + "epoch": 0.8437341451040081, + "grad_norm": 4.75, + "learning_rate": 6.672404572866566e-06, + "loss": 1.56401043, + "memory(GiB)": 111.15, + "step": 33260, + "train_speed(iter/s)": 0.448069 + }, + { + "acc": 0.65932899, + "epoch": 0.8438609842719432, + "grad_norm": 4.875, + "learning_rate": 6.671416310491406e-06, + "loss": 1.54996433, + "memory(GiB)": 111.15, + "step": 33265, + "train_speed(iter/s)": 0.448104 + }, + { + "acc": 0.63654847, + "epoch": 0.8439878234398782, + "grad_norm": 5.09375, + "learning_rate": 6.670427974599891e-06, + "loss": 1.68755684, + "memory(GiB)": 111.15, + "step": 33270, + "train_speed(iter/s)": 0.448139 + }, + { + "acc": 0.65346222, + "epoch": 0.8441146626078133, + "grad_norm": 6.5625, + "learning_rate": 6.669439565235498e-06, + "loss": 1.66946735, + "memory(GiB)": 111.15, + "step": 33275, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.64862328, + "epoch": 0.8442415017757483, + "grad_norm": 5.625, + "learning_rate": 6.668451082441698e-06, + "loss": 1.65061588, + "memory(GiB)": 111.15, + "step": 33280, + "train_speed(iter/s)": 0.448209 + }, + { + "acc": 0.63458776, + "epoch": 0.8443683409436834, + "grad_norm": 5.21875, + "learning_rate": 6.667462526261972e-06, + "loss": 1.66292915, + "memory(GiB)": 111.15, + "step": 33285, + "train_speed(iter/s)": 0.448244 + }, + { + "acc": 0.65645447, + "epoch": 0.8444951801116185, + "grad_norm": 4.4375, + "learning_rate": 6.666473896739798e-06, + "loss": 1.60450363, + "memory(GiB)": 111.15, + "step": 33290, + "train_speed(iter/s)": 0.448279 + }, + { + "acc": 0.6568069, + "epoch": 0.8446220192795535, + "grad_norm": 5.46875, + "learning_rate": 6.665485193918663e-06, + "loss": 1.62272491, + "memory(GiB)": 111.15, + "step": 33295, + "train_speed(iter/s)": 0.448315 + }, + { + "acc": 0.63424015, + "epoch": 0.8447488584474886, + "grad_norm": 5.9375, + "learning_rate": 6.664496417842053e-06, + "loss": 1.64584999, + "memory(GiB)": 111.15, + "step": 33300, + "train_speed(iter/s)": 0.44835 + }, + { + "epoch": 0.8447488584474886, + "eval_acc": 0.6430863479074245, + "eval_loss": 1.6098783016204834, + "eval_runtime": 114.3274, + "eval_samples_per_second": 55.717, + "eval_steps_per_second": 27.859, + "step": 33300 + }, + { + "acc": 0.65417161, + "epoch": 0.8448756976154237, + "grad_norm": 4.5625, + "learning_rate": 6.6635075685534566e-06, + "loss": 1.59129286, + "memory(GiB)": 111.15, + "step": 33305, + "train_speed(iter/s)": 0.447668 + }, + { + "acc": 0.6417614, + "epoch": 0.8450025367833587, + "grad_norm": 6.09375, + "learning_rate": 6.662518646096374e-06, + "loss": 1.64797993, + "memory(GiB)": 111.15, + "step": 33310, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.64793053, + "epoch": 0.8451293759512938, + "grad_norm": 4.53125, + "learning_rate": 6.661529650514296e-06, + "loss": 1.62770519, + "memory(GiB)": 111.15, + "step": 33315, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.64082332, + "epoch": 0.8452562151192288, + "grad_norm": 4.84375, + "learning_rate": 6.6605405818507274e-06, + "loss": 1.66928177, + "memory(GiB)": 111.15, + "step": 33320, + "train_speed(iter/s)": 0.447773 + }, + { + "acc": 0.63929791, + "epoch": 0.8453830542871639, + "grad_norm": 5.5625, + "learning_rate": 6.659551440149169e-06, + "loss": 1.62811127, + "memory(GiB)": 111.15, + "step": 33325, + "train_speed(iter/s)": 0.447808 + }, + { + "acc": 0.65722704, + "epoch": 0.845509893455099, + "grad_norm": 5.71875, + "learning_rate": 6.65856222545313e-06, + "loss": 1.5611187, + "memory(GiB)": 111.15, + "step": 33330, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.65442138, + "epoch": 0.845636732623034, + "grad_norm": 4.40625, + "learning_rate": 6.657572937806118e-06, + "loss": 1.62152672, + "memory(GiB)": 111.15, + "step": 33335, + "train_speed(iter/s)": 0.44788 + }, + { + "acc": 0.66389108, + "epoch": 0.8457635717909691, + "grad_norm": 4.9375, + "learning_rate": 6.656583577251649e-06, + "loss": 1.52432976, + "memory(GiB)": 111.15, + "step": 33340, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.63178072, + "epoch": 0.8458904109589042, + "grad_norm": 4.75, + "learning_rate": 6.655594143833237e-06, + "loss": 1.70309944, + "memory(GiB)": 111.15, + "step": 33345, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.64744501, + "epoch": 0.8460172501268391, + "grad_norm": 5.21875, + "learning_rate": 6.654604637594404e-06, + "loss": 1.63489761, + "memory(GiB)": 111.15, + "step": 33350, + "train_speed(iter/s)": 0.447986 + }, + { + "acc": 0.64361839, + "epoch": 0.8461440892947742, + "grad_norm": 5.46875, + "learning_rate": 6.653615058578672e-06, + "loss": 1.67503853, + "memory(GiB)": 111.15, + "step": 33355, + "train_speed(iter/s)": 0.448021 + }, + { + "acc": 0.66781349, + "epoch": 0.8462709284627092, + "grad_norm": 5.3125, + "learning_rate": 6.652625406829566e-06, + "loss": 1.57780266, + "memory(GiB)": 111.15, + "step": 33360, + "train_speed(iter/s)": 0.448057 + }, + { + "acc": 0.65137467, + "epoch": 0.8463977676306443, + "grad_norm": 5.84375, + "learning_rate": 6.651635682390616e-06, + "loss": 1.62241669, + "memory(GiB)": 111.15, + "step": 33365, + "train_speed(iter/s)": 0.448093 + }, + { + "acc": 0.65866723, + "epoch": 0.8465246067985794, + "grad_norm": 5.6875, + "learning_rate": 6.650645885305356e-06, + "loss": 1.61662235, + "memory(GiB)": 111.15, + "step": 33370, + "train_speed(iter/s)": 0.448129 + }, + { + "acc": 0.63742476, + "epoch": 0.8466514459665144, + "grad_norm": 6.6875, + "learning_rate": 6.649656015617319e-06, + "loss": 1.70518875, + "memory(GiB)": 111.15, + "step": 33375, + "train_speed(iter/s)": 0.448164 + }, + { + "acc": 0.64194765, + "epoch": 0.8467782851344495, + "grad_norm": 4.375, + "learning_rate": 6.648666073370046e-06, + "loss": 1.59113159, + "memory(GiB)": 111.15, + "step": 33380, + "train_speed(iter/s)": 0.4482 + }, + { + "acc": 0.65195837, + "epoch": 0.8469051243023846, + "grad_norm": 4.84375, + "learning_rate": 6.647676058607076e-06, + "loss": 1.63477859, + "memory(GiB)": 111.15, + "step": 33385, + "train_speed(iter/s)": 0.448235 + }, + { + "acc": 0.64590721, + "epoch": 0.8470319634703196, + "grad_norm": 5.8125, + "learning_rate": 6.64668597137196e-06, + "loss": 1.6787796, + "memory(GiB)": 111.15, + "step": 33390, + "train_speed(iter/s)": 0.448271 + }, + { + "acc": 0.66143999, + "epoch": 0.8471588026382547, + "grad_norm": 6.40625, + "learning_rate": 6.645695811708241e-06, + "loss": 1.61696777, + "memory(GiB)": 111.15, + "step": 33395, + "train_speed(iter/s)": 0.448307 + }, + { + "acc": 0.66040993, + "epoch": 0.8472856418061897, + "grad_norm": 5.09375, + "learning_rate": 6.644705579659474e-06, + "loss": 1.63563118, + "memory(GiB)": 111.15, + "step": 33400, + "train_speed(iter/s)": 0.448343 + }, + { + "epoch": 0.8472856418061897, + "eval_acc": 0.6430157452321261, + "eval_loss": 1.6097849607467651, + "eval_runtime": 113.3956, + "eval_samples_per_second": 56.175, + "eval_steps_per_second": 28.088, + "step": 33400 + }, + { + "acc": 0.65127106, + "epoch": 0.8474124809741248, + "grad_norm": 6.25, + "learning_rate": 6.643715275269212e-06, + "loss": 1.62248077, + "memory(GiB)": 111.15, + "step": 33405, + "train_speed(iter/s)": 0.447669 + }, + { + "acc": 0.6492054, + "epoch": 0.8475393201420599, + "grad_norm": 5.9375, + "learning_rate": 6.642724898581013e-06, + "loss": 1.65614929, + "memory(GiB)": 111.15, + "step": 33410, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.66185493, + "epoch": 0.8476661593099949, + "grad_norm": 5.375, + "learning_rate": 6.6417344496384394e-06, + "loss": 1.63182945, + "memory(GiB)": 111.15, + "step": 33415, + "train_speed(iter/s)": 0.44774 + }, + { + "acc": 0.61915178, + "epoch": 0.84779299847793, + "grad_norm": 4.21875, + "learning_rate": 6.640743928485054e-06, + "loss": 1.69921913, + "memory(GiB)": 111.15, + "step": 33420, + "train_speed(iter/s)": 0.447776 + }, + { + "acc": 0.65246449, + "epoch": 0.8479198376458651, + "grad_norm": 7.71875, + "learning_rate": 6.639753335164426e-06, + "loss": 1.67973499, + "memory(GiB)": 111.15, + "step": 33425, + "train_speed(iter/s)": 0.447811 + }, + { + "acc": 0.6490746, + "epoch": 0.8480466768138001, + "grad_norm": 5.3125, + "learning_rate": 6.638762669720126e-06, + "loss": 1.61516972, + "memory(GiB)": 111.15, + "step": 33430, + "train_speed(iter/s)": 0.447846 + }, + { + "acc": 0.66500316, + "epoch": 0.8481735159817352, + "grad_norm": 5.9375, + "learning_rate": 6.637771932195726e-06, + "loss": 1.56462555, + "memory(GiB)": 111.15, + "step": 33435, + "train_speed(iter/s)": 0.447881 + }, + { + "acc": 0.67313681, + "epoch": 0.8483003551496702, + "grad_norm": 5.125, + "learning_rate": 6.636781122634804e-06, + "loss": 1.60113678, + "memory(GiB)": 111.15, + "step": 33440, + "train_speed(iter/s)": 0.447916 + }, + { + "acc": 0.67301178, + "epoch": 0.8484271943176053, + "grad_norm": 4.625, + "learning_rate": 6.635790241080941e-06, + "loss": 1.47054367, + "memory(GiB)": 111.15, + "step": 33445, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.6498107, + "epoch": 0.8485540334855404, + "grad_norm": 5.34375, + "learning_rate": 6.634799287577721e-06, + "loss": 1.61631012, + "memory(GiB)": 111.15, + "step": 33450, + "train_speed(iter/s)": 0.447986 + }, + { + "acc": 0.66294231, + "epoch": 0.8486808726534754, + "grad_norm": 5.9375, + "learning_rate": 6.6338082621687286e-06, + "loss": 1.56814537, + "memory(GiB)": 111.15, + "step": 33455, + "train_speed(iter/s)": 0.448021 + }, + { + "acc": 0.64329062, + "epoch": 0.8488077118214105, + "grad_norm": 7.4375, + "learning_rate": 6.6328171648975545e-06, + "loss": 1.71611786, + "memory(GiB)": 111.15, + "step": 33460, + "train_speed(iter/s)": 0.448056 + }, + { + "acc": 0.66428714, + "epoch": 0.8489345509893456, + "grad_norm": 5.65625, + "learning_rate": 6.63182599580779e-06, + "loss": 1.579842, + "memory(GiB)": 111.15, + "step": 33465, + "train_speed(iter/s)": 0.448091 + }, + { + "acc": 0.65501623, + "epoch": 0.8490613901572805, + "grad_norm": 5.0, + "learning_rate": 6.630834754943036e-06, + "loss": 1.61677361, + "memory(GiB)": 111.15, + "step": 33470, + "train_speed(iter/s)": 0.448126 + }, + { + "acc": 0.65107756, + "epoch": 0.8491882293252156, + "grad_norm": 6.03125, + "learning_rate": 6.629843442346886e-06, + "loss": 1.62601814, + "memory(GiB)": 111.15, + "step": 33475, + "train_speed(iter/s)": 0.448162 + }, + { + "acc": 0.65384941, + "epoch": 0.8493150684931506, + "grad_norm": 6.125, + "learning_rate": 6.628852058062944e-06, + "loss": 1.64040565, + "memory(GiB)": 111.15, + "step": 33480, + "train_speed(iter/s)": 0.448197 + }, + { + "acc": 0.62846069, + "epoch": 0.8494419076610857, + "grad_norm": 4.84375, + "learning_rate": 6.627860602134818e-06, + "loss": 1.75492535, + "memory(GiB)": 111.15, + "step": 33485, + "train_speed(iter/s)": 0.448232 + }, + { + "acc": 0.63784924, + "epoch": 0.8495687468290208, + "grad_norm": 4.75, + "learning_rate": 6.626869074606113e-06, + "loss": 1.6860321, + "memory(GiB)": 111.15, + "step": 33490, + "train_speed(iter/s)": 0.448268 + }, + { + "acc": 0.6329545, + "epoch": 0.8496955859969558, + "grad_norm": 5.0625, + "learning_rate": 6.625877475520445e-06, + "loss": 1.69934692, + "memory(GiB)": 111.15, + "step": 33495, + "train_speed(iter/s)": 0.448303 + }, + { + "acc": 0.66101437, + "epoch": 0.8498224251648909, + "grad_norm": 5.46875, + "learning_rate": 6.624885804921425e-06, + "loss": 1.53357592, + "memory(GiB)": 111.15, + "step": 33500, + "train_speed(iter/s)": 0.448339 + }, + { + "epoch": 0.8498224251648909, + "eval_acc": 0.6430236828110059, + "eval_loss": 1.6096959114074707, + "eval_runtime": 113.2397, + "eval_samples_per_second": 56.252, + "eval_steps_per_second": 28.126, + "step": 33500 + }, + { + "acc": 0.65558043, + "epoch": 0.849949264332826, + "grad_norm": 6.40625, + "learning_rate": 6.623894062852673e-06, + "loss": 1.58202248, + "memory(GiB)": 111.15, + "step": 33505, + "train_speed(iter/s)": 0.447668 + }, + { + "acc": 0.67567086, + "epoch": 0.850076103500761, + "grad_norm": 4.59375, + "learning_rate": 6.62290224935781e-06, + "loss": 1.55376205, + "memory(GiB)": 111.15, + "step": 33510, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.65079336, + "epoch": 0.8502029426686961, + "grad_norm": 4.375, + "learning_rate": 6.621910364480461e-06, + "loss": 1.61496696, + "memory(GiB)": 111.15, + "step": 33515, + "train_speed(iter/s)": 0.447739 + }, + { + "acc": 0.65058041, + "epoch": 0.8503297818366311, + "grad_norm": 5.5625, + "learning_rate": 6.620918408264252e-06, + "loss": 1.66228218, + "memory(GiB)": 111.15, + "step": 33520, + "train_speed(iter/s)": 0.447774 + }, + { + "acc": 0.63595057, + "epoch": 0.8504566210045662, + "grad_norm": 5.09375, + "learning_rate": 6.6199263807528136e-06, + "loss": 1.70064678, + "memory(GiB)": 111.15, + "step": 33525, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.65302148, + "epoch": 0.8505834601725013, + "grad_norm": 5.53125, + "learning_rate": 6.618934281989783e-06, + "loss": 1.65606384, + "memory(GiB)": 111.15, + "step": 33530, + "train_speed(iter/s)": 0.447845 + }, + { + "acc": 0.64750624, + "epoch": 0.8507102993404363, + "grad_norm": 5.21875, + "learning_rate": 6.6179421120187915e-06, + "loss": 1.58632927, + "memory(GiB)": 111.15, + "step": 33535, + "train_speed(iter/s)": 0.44788 + }, + { + "acc": 0.66727066, + "epoch": 0.8508371385083714, + "grad_norm": 5.46875, + "learning_rate": 6.616949870883486e-06, + "loss": 1.54287758, + "memory(GiB)": 111.15, + "step": 33540, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.64481211, + "epoch": 0.8509639776763065, + "grad_norm": 5.9375, + "learning_rate": 6.615957558627503e-06, + "loss": 1.68597183, + "memory(GiB)": 111.15, + "step": 33545, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.65548801, + "epoch": 0.8510908168442415, + "grad_norm": 5.59375, + "learning_rate": 6.6149651752944945e-06, + "loss": 1.60086823, + "memory(GiB)": 111.15, + "step": 33550, + "train_speed(iter/s)": 0.447986 + }, + { + "acc": 0.64369044, + "epoch": 0.8512176560121766, + "grad_norm": 5.625, + "learning_rate": 6.613972720928105e-06, + "loss": 1.64105549, + "memory(GiB)": 111.15, + "step": 33555, + "train_speed(iter/s)": 0.448022 + }, + { + "acc": 0.65047326, + "epoch": 0.8513444951801116, + "grad_norm": 6.25, + "learning_rate": 6.61298019557199e-06, + "loss": 1.59020205, + "memory(GiB)": 111.15, + "step": 33560, + "train_speed(iter/s)": 0.448057 + }, + { + "acc": 0.67222328, + "epoch": 0.8514713343480467, + "grad_norm": 4.71875, + "learning_rate": 6.6119875992698045e-06, + "loss": 1.54036808, + "memory(GiB)": 111.15, + "step": 33565, + "train_speed(iter/s)": 0.448092 + }, + { + "acc": 0.65799875, + "epoch": 0.8515981735159818, + "grad_norm": 5.65625, + "learning_rate": 6.610994932065207e-06, + "loss": 1.5938921, + "memory(GiB)": 111.15, + "step": 33570, + "train_speed(iter/s)": 0.448128 + }, + { + "acc": 0.6477354, + "epoch": 0.8517250126839168, + "grad_norm": 5.46875, + "learning_rate": 6.610002194001861e-06, + "loss": 1.65589218, + "memory(GiB)": 111.15, + "step": 33575, + "train_speed(iter/s)": 0.448163 + }, + { + "acc": 0.65947609, + "epoch": 0.8518518518518519, + "grad_norm": 5.6875, + "learning_rate": 6.609009385123429e-06, + "loss": 1.54161491, + "memory(GiB)": 111.15, + "step": 33580, + "train_speed(iter/s)": 0.448198 + }, + { + "acc": 0.65867071, + "epoch": 0.851978691019787, + "grad_norm": 5.3125, + "learning_rate": 6.608016505473582e-06, + "loss": 1.628088, + "memory(GiB)": 111.15, + "step": 33585, + "train_speed(iter/s)": 0.448234 + }, + { + "acc": 0.64496975, + "epoch": 0.852105530187722, + "grad_norm": 5.8125, + "learning_rate": 6.60702355509599e-06, + "loss": 1.63381042, + "memory(GiB)": 111.15, + "step": 33590, + "train_speed(iter/s)": 0.448269 + }, + { + "acc": 0.65399361, + "epoch": 0.852232369355657, + "grad_norm": 4.90625, + "learning_rate": 6.606030534034326e-06, + "loss": 1.64799538, + "memory(GiB)": 111.15, + "step": 33595, + "train_speed(iter/s)": 0.448305 + }, + { + "acc": 0.64432917, + "epoch": 0.852359208523592, + "grad_norm": 5.03125, + "learning_rate": 6.6050374423322685e-06, + "loss": 1.68717976, + "memory(GiB)": 111.15, + "step": 33600, + "train_speed(iter/s)": 0.44834 + }, + { + "epoch": 0.852359208523592, + "eval_acc": 0.6430491666168827, + "eval_loss": 1.6097373962402344, + "eval_runtime": 113.468, + "eval_samples_per_second": 56.139, + "eval_steps_per_second": 28.07, + "step": 33600 + }, + { + "acc": 0.63243098, + "epoch": 0.8524860476915271, + "grad_norm": 5.875, + "learning_rate": 6.604044280033498e-06, + "loss": 1.685812, + "memory(GiB)": 111.15, + "step": 33605, + "train_speed(iter/s)": 0.44767 + }, + { + "acc": 0.64197149, + "epoch": 0.8526128868594622, + "grad_norm": 6.25, + "learning_rate": 6.6030510471817e-06, + "loss": 1.6759449, + "memory(GiB)": 111.15, + "step": 33610, + "train_speed(iter/s)": 0.447706 + }, + { + "acc": 0.63028383, + "epoch": 0.8527397260273972, + "grad_norm": 4.84375, + "learning_rate": 6.602057743820558e-06, + "loss": 1.7353054, + "memory(GiB)": 111.15, + "step": 33615, + "train_speed(iter/s)": 0.447741 + }, + { + "acc": 0.66099129, + "epoch": 0.8528665651953323, + "grad_norm": 7.625, + "learning_rate": 6.601064369993766e-06, + "loss": 1.55387506, + "memory(GiB)": 111.15, + "step": 33620, + "train_speed(iter/s)": 0.447776 + }, + { + "acc": 0.64724174, + "epoch": 0.8529934043632674, + "grad_norm": 4.9375, + "learning_rate": 6.600070925745012e-06, + "loss": 1.62212868, + "memory(GiB)": 111.15, + "step": 33625, + "train_speed(iter/s)": 0.447811 + }, + { + "acc": 0.64514408, + "epoch": 0.8531202435312024, + "grad_norm": 5.46875, + "learning_rate": 6.599077411117998e-06, + "loss": 1.67097015, + "memory(GiB)": 111.15, + "step": 33630, + "train_speed(iter/s)": 0.447846 + }, + { + "acc": 0.66402731, + "epoch": 0.8532470826991375, + "grad_norm": 5.1875, + "learning_rate": 6.598083826156418e-06, + "loss": 1.56577082, + "memory(GiB)": 111.15, + "step": 33635, + "train_speed(iter/s)": 0.447882 + }, + { + "acc": 0.65487385, + "epoch": 0.8533739218670725, + "grad_norm": 5.21875, + "learning_rate": 6.597090170903977e-06, + "loss": 1.63060265, + "memory(GiB)": 111.15, + "step": 33640, + "train_speed(iter/s)": 0.447917 + }, + { + "acc": 0.64574356, + "epoch": 0.8535007610350076, + "grad_norm": 7.0, + "learning_rate": 6.596096445404381e-06, + "loss": 1.62216644, + "memory(GiB)": 111.15, + "step": 33645, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.64172659, + "epoch": 0.8536276002029427, + "grad_norm": 5.21875, + "learning_rate": 6.595102649701336e-06, + "loss": 1.65510178, + "memory(GiB)": 111.15, + "step": 33650, + "train_speed(iter/s)": 0.447988 + }, + { + "acc": 0.64771619, + "epoch": 0.8537544393708777, + "grad_norm": 5.1875, + "learning_rate": 6.5941087838385545e-06, + "loss": 1.62130165, + "memory(GiB)": 111.15, + "step": 33655, + "train_speed(iter/s)": 0.448023 + }, + { + "acc": 0.66473866, + "epoch": 0.8538812785388128, + "grad_norm": 5.25, + "learning_rate": 6.593114847859752e-06, + "loss": 1.58470669, + "memory(GiB)": 111.15, + "step": 33660, + "train_speed(iter/s)": 0.448059 + }, + { + "acc": 0.64592905, + "epoch": 0.8540081177067479, + "grad_norm": 6.90625, + "learning_rate": 6.592120841808646e-06, + "loss": 1.67211609, + "memory(GiB)": 111.15, + "step": 33665, + "train_speed(iter/s)": 0.448094 + }, + { + "acc": 0.65438709, + "epoch": 0.8541349568746829, + "grad_norm": 4.90625, + "learning_rate": 6.5911267657289564e-06, + "loss": 1.60621834, + "memory(GiB)": 111.15, + "step": 33670, + "train_speed(iter/s)": 0.448129 + }, + { + "acc": 0.64332075, + "epoch": 0.854261796042618, + "grad_norm": 4.71875, + "learning_rate": 6.590132619664408e-06, + "loss": 1.61449375, + "memory(GiB)": 111.15, + "step": 33675, + "train_speed(iter/s)": 0.448164 + }, + { + "acc": 0.66235428, + "epoch": 0.854388635210553, + "grad_norm": 4.8125, + "learning_rate": 6.589138403658728e-06, + "loss": 1.59146252, + "memory(GiB)": 111.15, + "step": 33680, + "train_speed(iter/s)": 0.448199 + }, + { + "acc": 0.63562546, + "epoch": 0.8545154743784881, + "grad_norm": 5.34375, + "learning_rate": 6.588144117755645e-06, + "loss": 1.69099464, + "memory(GiB)": 111.15, + "step": 33685, + "train_speed(iter/s)": 0.448233 + }, + { + "acc": 0.6595715, + "epoch": 0.8546423135464232, + "grad_norm": 5.875, + "learning_rate": 6.5871497619988945e-06, + "loss": 1.62001686, + "memory(GiB)": 111.15, + "step": 33690, + "train_speed(iter/s)": 0.448269 + }, + { + "acc": 0.63432779, + "epoch": 0.8547691527143582, + "grad_norm": 4.9375, + "learning_rate": 6.586155336432211e-06, + "loss": 1.66417141, + "memory(GiB)": 111.15, + "step": 33695, + "train_speed(iter/s)": 0.448304 + }, + { + "acc": 0.66260719, + "epoch": 0.8548959918822933, + "grad_norm": 4.71875, + "learning_rate": 6.585160841099333e-06, + "loss": 1.53961687, + "memory(GiB)": 111.15, + "step": 33700, + "train_speed(iter/s)": 0.448339 + }, + { + "epoch": 0.8548959918822933, + "eval_acc": 0.6430299493206477, + "eval_loss": 1.6098220348358154, + "eval_runtime": 114.0828, + "eval_samples_per_second": 55.837, + "eval_steps_per_second": 27.918, + "step": 33700 + }, + { + "acc": 0.65671787, + "epoch": 0.8550228310502284, + "grad_norm": 5.46875, + "learning_rate": 6.584166276044005e-06, + "loss": 1.62164822, + "memory(GiB)": 111.15, + "step": 33705, + "train_speed(iter/s)": 0.447666 + }, + { + "acc": 0.63920784, + "epoch": 0.8551496702181633, + "grad_norm": 4.4375, + "learning_rate": 6.583171641309971e-06, + "loss": 1.68052139, + "memory(GiB)": 111.15, + "step": 33710, + "train_speed(iter/s)": 0.447701 + }, + { + "acc": 0.65373049, + "epoch": 0.8552765093860984, + "grad_norm": 6.09375, + "learning_rate": 6.58217693694098e-06, + "loss": 1.58439808, + "memory(GiB)": 111.15, + "step": 33715, + "train_speed(iter/s)": 0.447736 + }, + { + "acc": 0.6487236, + "epoch": 0.8554033485540334, + "grad_norm": 5.84375, + "learning_rate": 6.581182162980784e-06, + "loss": 1.60790405, + "memory(GiB)": 111.15, + "step": 33720, + "train_speed(iter/s)": 0.44777 + }, + { + "acc": 0.63687, + "epoch": 0.8555301877219685, + "grad_norm": 4.84375, + "learning_rate": 6.580187319473137e-06, + "loss": 1.69592552, + "memory(GiB)": 111.15, + "step": 33725, + "train_speed(iter/s)": 0.447805 + }, + { + "acc": 0.63249493, + "epoch": 0.8556570268899036, + "grad_norm": 5.0625, + "learning_rate": 6.579192406461796e-06, + "loss": 1.68351135, + "memory(GiB)": 111.15, + "step": 33730, + "train_speed(iter/s)": 0.44784 + }, + { + "acc": 0.64201794, + "epoch": 0.8557838660578386, + "grad_norm": 6.25, + "learning_rate": 6.5781974239905225e-06, + "loss": 1.69529018, + "memory(GiB)": 111.15, + "step": 33735, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.63101611, + "epoch": 0.8559107052257737, + "grad_norm": 6.03125, + "learning_rate": 6.57720237210308e-06, + "loss": 1.64155407, + "memory(GiB)": 111.15, + "step": 33740, + "train_speed(iter/s)": 0.44791 + }, + { + "acc": 0.6413847, + "epoch": 0.8560375443937088, + "grad_norm": 5.78125, + "learning_rate": 6.576207250843235e-06, + "loss": 1.60753651, + "memory(GiB)": 111.15, + "step": 33745, + "train_speed(iter/s)": 0.447945 + }, + { + "acc": 0.66926622, + "epoch": 0.8561643835616438, + "grad_norm": 5.0625, + "learning_rate": 6.575212060254759e-06, + "loss": 1.5810626, + "memory(GiB)": 111.15, + "step": 33750, + "train_speed(iter/s)": 0.447979 + }, + { + "acc": 0.65662031, + "epoch": 0.8562912227295789, + "grad_norm": 4.96875, + "learning_rate": 6.574216800381424e-06, + "loss": 1.6521246, + "memory(GiB)": 111.15, + "step": 33755, + "train_speed(iter/s)": 0.448014 + }, + { + "acc": 0.65657015, + "epoch": 0.8564180618975139, + "grad_norm": 6.21875, + "learning_rate": 6.573221471267005e-06, + "loss": 1.55959435, + "memory(GiB)": 111.15, + "step": 33760, + "train_speed(iter/s)": 0.448049 + }, + { + "acc": 0.65175214, + "epoch": 0.856544901065449, + "grad_norm": 4.78125, + "learning_rate": 6.572226072955281e-06, + "loss": 1.62210617, + "memory(GiB)": 111.15, + "step": 33765, + "train_speed(iter/s)": 0.448084 + }, + { + "acc": 0.67082224, + "epoch": 0.8566717402333841, + "grad_norm": 6.09375, + "learning_rate": 6.571230605490036e-06, + "loss": 1.56718063, + "memory(GiB)": 111.15, + "step": 33770, + "train_speed(iter/s)": 0.448119 + }, + { + "acc": 0.66851468, + "epoch": 0.8567985794013191, + "grad_norm": 5.5, + "learning_rate": 6.570235068915053e-06, + "loss": 1.55819979, + "memory(GiB)": 111.15, + "step": 33775, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.64724255, + "epoch": 0.8569254185692542, + "grad_norm": 5.21875, + "learning_rate": 6.569239463274122e-06, + "loss": 1.67202682, + "memory(GiB)": 111.15, + "step": 33780, + "train_speed(iter/s)": 0.448189 + }, + { + "acc": 0.63582888, + "epoch": 0.8570522577371893, + "grad_norm": 4.5, + "learning_rate": 6.568243788611033e-06, + "loss": 1.62125702, + "memory(GiB)": 111.15, + "step": 33785, + "train_speed(iter/s)": 0.448223 + }, + { + "acc": 0.65640116, + "epoch": 0.8571790969051243, + "grad_norm": 4.65625, + "learning_rate": 6.56724804496958e-06, + "loss": 1.59647884, + "memory(GiB)": 111.15, + "step": 33790, + "train_speed(iter/s)": 0.448258 + }, + { + "acc": 0.64559865, + "epoch": 0.8573059360730594, + "grad_norm": 4.59375, + "learning_rate": 6.566252232393561e-06, + "loss": 1.63756351, + "memory(GiB)": 111.15, + "step": 33795, + "train_speed(iter/s)": 0.448293 + }, + { + "acc": 0.65920854, + "epoch": 0.8574327752409944, + "grad_norm": 4.84375, + "learning_rate": 6.565256350926777e-06, + "loss": 1.59970531, + "memory(GiB)": 111.15, + "step": 33800, + "train_speed(iter/s)": 0.448328 + }, + { + "epoch": 0.8574327752409944, + "eval_acc": 0.643090525580519, + "eval_loss": 1.6096980571746826, + "eval_runtime": 113.8806, + "eval_samples_per_second": 55.936, + "eval_steps_per_second": 27.968, + "step": 33800 + }, + { + "acc": 0.66458426, + "epoch": 0.8575596144089295, + "grad_norm": 4.59375, + "learning_rate": 6.5642604006130286e-06, + "loss": 1.55669985, + "memory(GiB)": 111.15, + "step": 33805, + "train_speed(iter/s)": 0.44766 + }, + { + "acc": 0.64524298, + "epoch": 0.8576864535768646, + "grad_norm": 4.78125, + "learning_rate": 6.563264381496124e-06, + "loss": 1.7032629, + "memory(GiB)": 111.15, + "step": 33810, + "train_speed(iter/s)": 0.447694 + }, + { + "acc": 0.6447608, + "epoch": 0.8578132927447996, + "grad_norm": 7.46875, + "learning_rate": 6.562268293619872e-06, + "loss": 1.70825539, + "memory(GiB)": 111.15, + "step": 33815, + "train_speed(iter/s)": 0.447729 + }, + { + "acc": 0.6517633, + "epoch": 0.8579401319127347, + "grad_norm": 5.6875, + "learning_rate": 6.561272137028089e-06, + "loss": 1.63892555, + "memory(GiB)": 111.15, + "step": 33820, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.6431633, + "epoch": 0.8580669710806698, + "grad_norm": 5.34375, + "learning_rate": 6.560275911764582e-06, + "loss": 1.59136581, + "memory(GiB)": 111.15, + "step": 33825, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.65595384, + "epoch": 0.8581938102486047, + "grad_norm": 5.8125, + "learning_rate": 6.5592796178731776e-06, + "loss": 1.62723656, + "memory(GiB)": 111.15, + "step": 33830, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.63395481, + "epoch": 0.8583206494165398, + "grad_norm": 4.875, + "learning_rate": 6.5582832553976924e-06, + "loss": 1.62815971, + "memory(GiB)": 111.15, + "step": 33835, + "train_speed(iter/s)": 0.447869 + }, + { + "acc": 0.66568928, + "epoch": 0.8584474885844748, + "grad_norm": 4.5, + "learning_rate": 6.557286824381955e-06, + "loss": 1.5772296, + "memory(GiB)": 111.15, + "step": 33840, + "train_speed(iter/s)": 0.447904 + }, + { + "acc": 0.65601931, + "epoch": 0.8585743277524099, + "grad_norm": 6.71875, + "learning_rate": 6.556290324869786e-06, + "loss": 1.64466362, + "memory(GiB)": 111.15, + "step": 33845, + "train_speed(iter/s)": 0.447938 + }, + { + "acc": 0.65988665, + "epoch": 0.858701166920345, + "grad_norm": 5.21875, + "learning_rate": 6.555293756905024e-06, + "loss": 1.59397125, + "memory(GiB)": 111.15, + "step": 33850, + "train_speed(iter/s)": 0.447973 + }, + { + "acc": 0.65314541, + "epoch": 0.85882800608828, + "grad_norm": 4.90625, + "learning_rate": 6.554297120531497e-06, + "loss": 1.53464174, + "memory(GiB)": 111.15, + "step": 33855, + "train_speed(iter/s)": 0.448008 + }, + { + "acc": 0.64250631, + "epoch": 0.8589548452562151, + "grad_norm": 4.46875, + "learning_rate": 6.553300415793042e-06, + "loss": 1.65261078, + "memory(GiB)": 111.15, + "step": 33860, + "train_speed(iter/s)": 0.448042 + }, + { + "acc": 0.64237485, + "epoch": 0.8590816844241502, + "grad_norm": 4.96875, + "learning_rate": 6.552303642733502e-06, + "loss": 1.64660168, + "memory(GiB)": 111.15, + "step": 33865, + "train_speed(iter/s)": 0.448077 + }, + { + "acc": 0.65425329, + "epoch": 0.8592085235920852, + "grad_norm": 5.0625, + "learning_rate": 6.551306801396715e-06, + "loss": 1.64215279, + "memory(GiB)": 111.15, + "step": 33870, + "train_speed(iter/s)": 0.448112 + }, + { + "acc": 0.64815531, + "epoch": 0.8593353627600203, + "grad_norm": 5.375, + "learning_rate": 6.550309891826531e-06, + "loss": 1.57700758, + "memory(GiB)": 111.15, + "step": 33875, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.64734631, + "epoch": 0.8594622019279553, + "grad_norm": 5.75, + "learning_rate": 6.5493129140667955e-06, + "loss": 1.60265007, + "memory(GiB)": 111.15, + "step": 33880, + "train_speed(iter/s)": 0.448182 + }, + { + "acc": 0.63714714, + "epoch": 0.8595890410958904, + "grad_norm": 6.375, + "learning_rate": 6.54831586816136e-06, + "loss": 1.6706707, + "memory(GiB)": 111.15, + "step": 33885, + "train_speed(iter/s)": 0.448217 + }, + { + "acc": 0.65309372, + "epoch": 0.8597158802638255, + "grad_norm": 5.25, + "learning_rate": 6.54731875415408e-06, + "loss": 1.61687813, + "memory(GiB)": 111.15, + "step": 33890, + "train_speed(iter/s)": 0.448252 + }, + { + "acc": 0.65954657, + "epoch": 0.8598427194317605, + "grad_norm": 4.46875, + "learning_rate": 6.546321572088814e-06, + "loss": 1.57327976, + "memory(GiB)": 111.15, + "step": 33895, + "train_speed(iter/s)": 0.448286 + }, + { + "acc": 0.64983215, + "epoch": 0.8599695585996956, + "grad_norm": 5.03125, + "learning_rate": 6.545324322009421e-06, + "loss": 1.66807709, + "memory(GiB)": 111.15, + "step": 33900, + "train_speed(iter/s)": 0.448321 + }, + { + "epoch": 0.8599695585996956, + "eval_acc": 0.6430696372150462, + "eval_loss": 1.6096121072769165, + "eval_runtime": 114.2269, + "eval_samples_per_second": 55.766, + "eval_steps_per_second": 27.883, + "step": 33900 + }, + { + "acc": 0.66443129, + "epoch": 0.8600963977676307, + "grad_norm": 5.71875, + "learning_rate": 6.544327003959765e-06, + "loss": 1.57760592, + "memory(GiB)": 111.15, + "step": 33905, + "train_speed(iter/s)": 0.447653 + }, + { + "acc": 0.66145468, + "epoch": 0.8602232369355657, + "grad_norm": 5.375, + "learning_rate": 6.543329617983713e-06, + "loss": 1.64106712, + "memory(GiB)": 111.15, + "step": 33910, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.66133275, + "epoch": 0.8603500761035008, + "grad_norm": 5.40625, + "learning_rate": 6.5423321641251316e-06, + "loss": 1.64823666, + "memory(GiB)": 111.15, + "step": 33915, + "train_speed(iter/s)": 0.447723 + }, + { + "acc": 0.63717117, + "epoch": 0.8604769152714358, + "grad_norm": 6.0625, + "learning_rate": 6.541334642427898e-06, + "loss": 1.67163277, + "memory(GiB)": 111.15, + "step": 33920, + "train_speed(iter/s)": 0.447758 + }, + { + "acc": 0.64837456, + "epoch": 0.8606037544393709, + "grad_norm": 4.59375, + "learning_rate": 6.540337052935884e-06, + "loss": 1.58884373, + "memory(GiB)": 111.15, + "step": 33925, + "train_speed(iter/s)": 0.447793 + }, + { + "acc": 0.64843798, + "epoch": 0.860730593607306, + "grad_norm": 5.5625, + "learning_rate": 6.53933939569297e-06, + "loss": 1.64086437, + "memory(GiB)": 111.15, + "step": 33930, + "train_speed(iter/s)": 0.447828 + }, + { + "acc": 0.65010824, + "epoch": 0.860857432775241, + "grad_norm": 5.46875, + "learning_rate": 6.538341670743037e-06, + "loss": 1.65124435, + "memory(GiB)": 111.15, + "step": 33935, + "train_speed(iter/s)": 0.447863 + }, + { + "acc": 0.65677457, + "epoch": 0.8609842719431761, + "grad_norm": 5.625, + "learning_rate": 6.537343878129969e-06, + "loss": 1.62289085, + "memory(GiB)": 111.15, + "step": 33940, + "train_speed(iter/s)": 0.447898 + }, + { + "acc": 0.64079742, + "epoch": 0.8611111111111112, + "grad_norm": 6.1875, + "learning_rate": 6.5363460178976524e-06, + "loss": 1.671912, + "memory(GiB)": 111.15, + "step": 33945, + "train_speed(iter/s)": 0.447933 + }, + { + "acc": 0.64105024, + "epoch": 0.8612379502790461, + "grad_norm": 5.09375, + "learning_rate": 6.53534809008998e-06, + "loss": 1.66428261, + "memory(GiB)": 111.15, + "step": 33950, + "train_speed(iter/s)": 0.447968 + }, + { + "acc": 0.64861894, + "epoch": 0.8613647894469812, + "grad_norm": 4.6875, + "learning_rate": 6.534350094750843e-06, + "loss": 1.64753418, + "memory(GiB)": 111.15, + "step": 33955, + "train_speed(iter/s)": 0.448003 + }, + { + "acc": 0.64230828, + "epoch": 0.8614916286149162, + "grad_norm": 5.375, + "learning_rate": 6.5333520319241385e-06, + "loss": 1.64396248, + "memory(GiB)": 111.15, + "step": 33960, + "train_speed(iter/s)": 0.448038 + }, + { + "acc": 0.63586211, + "epoch": 0.8616184677828513, + "grad_norm": 6.0625, + "learning_rate": 6.532353901653765e-06, + "loss": 1.69052887, + "memory(GiB)": 111.15, + "step": 33965, + "train_speed(iter/s)": 0.448073 + }, + { + "acc": 0.64366751, + "epoch": 0.8617453069507864, + "grad_norm": 5.09375, + "learning_rate": 6.531355703983627e-06, + "loss": 1.68118172, + "memory(GiB)": 111.15, + "step": 33970, + "train_speed(iter/s)": 0.448108 + }, + { + "acc": 0.63940649, + "epoch": 0.8618721461187214, + "grad_norm": 5.5, + "learning_rate": 6.530357438957626e-06, + "loss": 1.66125145, + "memory(GiB)": 111.15, + "step": 33975, + "train_speed(iter/s)": 0.448143 + }, + { + "acc": 0.65495729, + "epoch": 0.8619989852866565, + "grad_norm": 5.84375, + "learning_rate": 6.529359106619675e-06, + "loss": 1.62696686, + "memory(GiB)": 111.15, + "step": 33980, + "train_speed(iter/s)": 0.448178 + }, + { + "acc": 0.64230642, + "epoch": 0.8621258244545916, + "grad_norm": 4.75, + "learning_rate": 6.528360707013681e-06, + "loss": 1.65546036, + "memory(GiB)": 111.15, + "step": 33985, + "train_speed(iter/s)": 0.448213 + }, + { + "acc": 0.66767263, + "epoch": 0.8622526636225266, + "grad_norm": 6.6875, + "learning_rate": 6.52736224018356e-06, + "loss": 1.64611168, + "memory(GiB)": 111.15, + "step": 33990, + "train_speed(iter/s)": 0.448248 + }, + { + "acc": 0.65848632, + "epoch": 0.8623795027904617, + "grad_norm": 4.6875, + "learning_rate": 6.526363706173227e-06, + "loss": 1.66232738, + "memory(GiB)": 111.15, + "step": 33995, + "train_speed(iter/s)": 0.448283 + }, + { + "acc": 0.66594439, + "epoch": 0.8625063419583967, + "grad_norm": 5.375, + "learning_rate": 6.525365105026605e-06, + "loss": 1.5234683, + "memory(GiB)": 111.15, + "step": 34000, + "train_speed(iter/s)": 0.448318 + }, + { + "epoch": 0.8625063419583967, + "eval_acc": 0.6430855123728055, + "eval_loss": 1.6095094680786133, + "eval_runtime": 113.9048, + "eval_samples_per_second": 55.924, + "eval_steps_per_second": 27.962, + "step": 34000 + }, + { + "acc": 0.67361836, + "epoch": 0.8626331811263318, + "grad_norm": 5.5, + "learning_rate": 6.524366436787615e-06, + "loss": 1.51645641, + "memory(GiB)": 111.15, + "step": 34005, + "train_speed(iter/s)": 0.447654 + }, + { + "acc": 0.65412579, + "epoch": 0.8627600202942669, + "grad_norm": 5.03125, + "learning_rate": 6.523367701500183e-06, + "loss": 1.6312912, + "memory(GiB)": 111.15, + "step": 34010, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.6508791, + "epoch": 0.8628868594622019, + "grad_norm": 4.96875, + "learning_rate": 6.5223688992082375e-06, + "loss": 1.60583305, + "memory(GiB)": 111.15, + "step": 34015, + "train_speed(iter/s)": 0.447723 + }, + { + "acc": 0.64036684, + "epoch": 0.863013698630137, + "grad_norm": 6.15625, + "learning_rate": 6.521370029955713e-06, + "loss": 1.66815567, + "memory(GiB)": 111.15, + "step": 34020, + "train_speed(iter/s)": 0.447757 + }, + { + "acc": 0.63880415, + "epoch": 0.8631405377980721, + "grad_norm": 5.0, + "learning_rate": 6.520371093786541e-06, + "loss": 1.70974808, + "memory(GiB)": 111.15, + "step": 34025, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.66748142, + "epoch": 0.8632673769660071, + "grad_norm": 6.53125, + "learning_rate": 6.51937209074466e-06, + "loss": 1.57717495, + "memory(GiB)": 111.15, + "step": 34030, + "train_speed(iter/s)": 0.447827 + }, + { + "acc": 0.64719415, + "epoch": 0.8633942161339422, + "grad_norm": 5.375, + "learning_rate": 6.51837302087401e-06, + "loss": 1.60572033, + "memory(GiB)": 111.15, + "step": 34035, + "train_speed(iter/s)": 0.447861 + }, + { + "acc": 0.64406948, + "epoch": 0.8635210553018772, + "grad_norm": 6.71875, + "learning_rate": 6.517373884218539e-06, + "loss": 1.67631607, + "memory(GiB)": 111.15, + "step": 34040, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.64681015, + "epoch": 0.8636478944698123, + "grad_norm": 5.71875, + "learning_rate": 6.5163746808221865e-06, + "loss": 1.66889915, + "memory(GiB)": 111.15, + "step": 34045, + "train_speed(iter/s)": 0.447931 + }, + { + "acc": 0.6349124, + "epoch": 0.8637747336377474, + "grad_norm": 4.53125, + "learning_rate": 6.515375410728907e-06, + "loss": 1.62266026, + "memory(GiB)": 111.15, + "step": 34050, + "train_speed(iter/s)": 0.447966 + }, + { + "acc": 0.65088034, + "epoch": 0.8639015728056824, + "grad_norm": 4.4375, + "learning_rate": 6.51437607398265e-06, + "loss": 1.62683144, + "memory(GiB)": 111.15, + "step": 34055, + "train_speed(iter/s)": 0.448 + }, + { + "acc": 0.66499357, + "epoch": 0.8640284119736175, + "grad_norm": 6.1875, + "learning_rate": 6.513376670627374e-06, + "loss": 1.61137257, + "memory(GiB)": 111.15, + "step": 34060, + "train_speed(iter/s)": 0.448035 + }, + { + "acc": 0.63947821, + "epoch": 0.8641552511415526, + "grad_norm": 7.40625, + "learning_rate": 6.512377200707033e-06, + "loss": 1.68223114, + "memory(GiB)": 111.15, + "step": 34065, + "train_speed(iter/s)": 0.44807 + }, + { + "acc": 0.65389161, + "epoch": 0.8642820903094875, + "grad_norm": 6.03125, + "learning_rate": 6.511377664265591e-06, + "loss": 1.6292366, + "memory(GiB)": 111.15, + "step": 34070, + "train_speed(iter/s)": 0.448105 + }, + { + "acc": 0.65306759, + "epoch": 0.8644089294774226, + "grad_norm": 4.90625, + "learning_rate": 6.510378061347013e-06, + "loss": 1.62874413, + "memory(GiB)": 111.15, + "step": 34075, + "train_speed(iter/s)": 0.448139 + }, + { + "acc": 0.64595814, + "epoch": 0.8645357686453576, + "grad_norm": 6.21875, + "learning_rate": 6.509378391995264e-06, + "loss": 1.66335678, + "memory(GiB)": 111.15, + "step": 34080, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.64070282, + "epoch": 0.8646626078132927, + "grad_norm": 4.40625, + "learning_rate": 6.508378656254314e-06, + "loss": 1.70875092, + "memory(GiB)": 111.15, + "step": 34085, + "train_speed(iter/s)": 0.448209 + }, + { + "acc": 0.66133313, + "epoch": 0.8647894469812278, + "grad_norm": 5.0625, + "learning_rate": 6.507378854168136e-06, + "loss": 1.57960548, + "memory(GiB)": 111.15, + "step": 34090, + "train_speed(iter/s)": 0.448243 + }, + { + "acc": 0.64189892, + "epoch": 0.8649162861491628, + "grad_norm": 5.21875, + "learning_rate": 6.506378985780707e-06, + "loss": 1.6196991, + "memory(GiB)": 111.15, + "step": 34095, + "train_speed(iter/s)": 0.448278 + }, + { + "acc": 0.66332951, + "epoch": 0.8650431253170979, + "grad_norm": 5.375, + "learning_rate": 6.505379051136004e-06, + "loss": 1.60023766, + "memory(GiB)": 111.15, + "step": 34100, + "train_speed(iter/s)": 0.448313 + }, + { + "epoch": 0.8650431253170979, + "eval_acc": 0.6430529265226679, + "eval_loss": 1.609708547592163, + "eval_runtime": 113.2478, + "eval_samples_per_second": 56.248, + "eval_steps_per_second": 28.124, + "step": 34100 + }, + { + "acc": 0.65595145, + "epoch": 0.865169964485033, + "grad_norm": 5.125, + "learning_rate": 6.504379050278009e-06, + "loss": 1.59937859, + "memory(GiB)": 111.15, + "step": 34105, + "train_speed(iter/s)": 0.447654 + }, + { + "acc": 0.63419781, + "epoch": 0.865296803652968, + "grad_norm": 5.4375, + "learning_rate": 6.503378983250707e-06, + "loss": 1.72474403, + "memory(GiB)": 111.15, + "step": 34110, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.6487473, + "epoch": 0.8654236428209031, + "grad_norm": 7.03125, + "learning_rate": 6.5023788500980855e-06, + "loss": 1.62384415, + "memory(GiB)": 111.15, + "step": 34115, + "train_speed(iter/s)": 0.447723 + }, + { + "acc": 0.64523125, + "epoch": 0.8655504819888381, + "grad_norm": 4.96875, + "learning_rate": 6.501378650864135e-06, + "loss": 1.64504051, + "memory(GiB)": 111.15, + "step": 34120, + "train_speed(iter/s)": 0.447758 + }, + { + "acc": 0.64977465, + "epoch": 0.8656773211567732, + "grad_norm": 5.4375, + "learning_rate": 6.500378385592847e-06, + "loss": 1.6302372, + "memory(GiB)": 111.15, + "step": 34125, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.65709362, + "epoch": 0.8658041603247083, + "grad_norm": 6.5, + "learning_rate": 6.49937805432822e-06, + "loss": 1.60396118, + "memory(GiB)": 111.15, + "step": 34130, + "train_speed(iter/s)": 0.447826 + }, + { + "acc": 0.65094242, + "epoch": 0.8659309994926433, + "grad_norm": 5.75, + "learning_rate": 6.498377657114251e-06, + "loss": 1.59760151, + "memory(GiB)": 111.15, + "step": 34135, + "train_speed(iter/s)": 0.44786 + }, + { + "acc": 0.657862, + "epoch": 0.8660578386605784, + "grad_norm": 5.4375, + "learning_rate": 6.497377193994944e-06, + "loss": 1.60301018, + "memory(GiB)": 111.15, + "step": 34140, + "train_speed(iter/s)": 0.447894 + }, + { + "acc": 0.66481495, + "epoch": 0.8661846778285135, + "grad_norm": 5.3125, + "learning_rate": 6.496376665014301e-06, + "loss": 1.60108242, + "memory(GiB)": 111.15, + "step": 34145, + "train_speed(iter/s)": 0.447929 + }, + { + "acc": 0.64220085, + "epoch": 0.8663115169964485, + "grad_norm": 4.625, + "learning_rate": 6.4953760702163325e-06, + "loss": 1.65890102, + "memory(GiB)": 111.15, + "step": 34150, + "train_speed(iter/s)": 0.447963 + }, + { + "acc": 0.65207853, + "epoch": 0.8664383561643836, + "grad_norm": 4.59375, + "learning_rate": 6.494375409645049e-06, + "loss": 1.61466465, + "memory(GiB)": 111.15, + "step": 34155, + "train_speed(iter/s)": 0.447998 + }, + { + "acc": 0.6472579, + "epoch": 0.8665651953323186, + "grad_norm": 5.15625, + "learning_rate": 6.493374683344462e-06, + "loss": 1.62163239, + "memory(GiB)": 111.15, + "step": 34160, + "train_speed(iter/s)": 0.448032 + }, + { + "acc": 0.65412726, + "epoch": 0.8666920345002537, + "grad_norm": 6.0625, + "learning_rate": 6.492373891358589e-06, + "loss": 1.57932072, + "memory(GiB)": 111.15, + "step": 34165, + "train_speed(iter/s)": 0.448066 + }, + { + "acc": 0.65061989, + "epoch": 0.8668188736681888, + "grad_norm": 6.71875, + "learning_rate": 6.4913730337314495e-06, + "loss": 1.63625336, + "memory(GiB)": 111.15, + "step": 34170, + "train_speed(iter/s)": 0.448101 + }, + { + "acc": 0.64501915, + "epoch": 0.8669457128361238, + "grad_norm": 4.40625, + "learning_rate": 6.490372110507066e-06, + "loss": 1.66398087, + "memory(GiB)": 111.15, + "step": 34175, + "train_speed(iter/s)": 0.448135 + }, + { + "acc": 0.65097809, + "epoch": 0.8670725520040589, + "grad_norm": 4.90625, + "learning_rate": 6.489371121729462e-06, + "loss": 1.5847126, + "memory(GiB)": 111.15, + "step": 34180, + "train_speed(iter/s)": 0.44817 + }, + { + "acc": 0.64408445, + "epoch": 0.867199391171994, + "grad_norm": 5.375, + "learning_rate": 6.4883700674426666e-06, + "loss": 1.58456707, + "memory(GiB)": 111.15, + "step": 34185, + "train_speed(iter/s)": 0.448204 + }, + { + "acc": 0.64847574, + "epoch": 0.867326230339929, + "grad_norm": 5.15625, + "learning_rate": 6.4873689476907105e-06, + "loss": 1.63229408, + "memory(GiB)": 111.15, + "step": 34190, + "train_speed(iter/s)": 0.448238 + }, + { + "acc": 0.65201225, + "epoch": 0.867453069507864, + "grad_norm": 5.8125, + "learning_rate": 6.486367762517628e-06, + "loss": 1.62793159, + "memory(GiB)": 111.15, + "step": 34195, + "train_speed(iter/s)": 0.448272 + }, + { + "acc": 0.64351645, + "epoch": 0.867579908675799, + "grad_norm": 5.03125, + "learning_rate": 6.4853665119674556e-06, + "loss": 1.66754227, + "memory(GiB)": 111.15, + "step": 34200, + "train_speed(iter/s)": 0.448306 + }, + { + "epoch": 0.867579908675799, + "eval_acc": 0.6430341269937423, + "eval_loss": 1.6096562147140503, + "eval_runtime": 113.8922, + "eval_samples_per_second": 55.93, + "eval_steps_per_second": 27.965, + "step": 34200 + }, + { + "acc": 0.63922691, + "epoch": 0.8677067478437341, + "grad_norm": 4.46875, + "learning_rate": 6.484365196084231e-06, + "loss": 1.60705242, + "memory(GiB)": 111.15, + "step": 34205, + "train_speed(iter/s)": 0.447646 + }, + { + "acc": 0.65318489, + "epoch": 0.8678335870116692, + "grad_norm": 4.625, + "learning_rate": 6.4833638149119985e-06, + "loss": 1.62388973, + "memory(GiB)": 111.15, + "step": 34210, + "train_speed(iter/s)": 0.44768 + }, + { + "acc": 0.64865079, + "epoch": 0.8679604261796042, + "grad_norm": 5.59375, + "learning_rate": 6.4823623684948034e-06, + "loss": 1.64490166, + "memory(GiB)": 111.15, + "step": 34215, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.6547688, + "epoch": 0.8680872653475393, + "grad_norm": 5.5, + "learning_rate": 6.4813608568766924e-06, + "loss": 1.67377319, + "memory(GiB)": 111.15, + "step": 34220, + "train_speed(iter/s)": 0.447749 + }, + { + "acc": 0.63875847, + "epoch": 0.8682141045154744, + "grad_norm": 5.6875, + "learning_rate": 6.480359280101717e-06, + "loss": 1.6355896, + "memory(GiB)": 111.15, + "step": 34225, + "train_speed(iter/s)": 0.447783 + }, + { + "acc": 0.65882783, + "epoch": 0.8683409436834094, + "grad_norm": 5.40625, + "learning_rate": 6.479357638213931e-06, + "loss": 1.58050327, + "memory(GiB)": 111.15, + "step": 34230, + "train_speed(iter/s)": 0.447817 + }, + { + "acc": 0.64707541, + "epoch": 0.8684677828513445, + "grad_norm": 5.4375, + "learning_rate": 6.478355931257392e-06, + "loss": 1.62168007, + "memory(GiB)": 111.15, + "step": 34235, + "train_speed(iter/s)": 0.447852 + }, + { + "acc": 0.65711718, + "epoch": 0.8685946220192795, + "grad_norm": 5.09375, + "learning_rate": 6.477354159276158e-06, + "loss": 1.58132343, + "memory(GiB)": 111.15, + "step": 34240, + "train_speed(iter/s)": 0.447886 + }, + { + "acc": 0.641045, + "epoch": 0.8687214611872146, + "grad_norm": 5.1875, + "learning_rate": 6.476352322314292e-06, + "loss": 1.62512379, + "memory(GiB)": 111.15, + "step": 34245, + "train_speed(iter/s)": 0.447921 + }, + { + "acc": 0.64161339, + "epoch": 0.8688483003551497, + "grad_norm": 5.84375, + "learning_rate": 6.47535042041586e-06, + "loss": 1.69849243, + "memory(GiB)": 111.15, + "step": 34250, + "train_speed(iter/s)": 0.447956 + }, + { + "acc": 0.62002602, + "epoch": 0.8689751395230847, + "grad_norm": 4.90625, + "learning_rate": 6.474348453624929e-06, + "loss": 1.72490959, + "memory(GiB)": 111.15, + "step": 34255, + "train_speed(iter/s)": 0.447991 + }, + { + "acc": 0.64783931, + "epoch": 0.8691019786910198, + "grad_norm": 5.8125, + "learning_rate": 6.473346421985571e-06, + "loss": 1.63896446, + "memory(GiB)": 111.15, + "step": 34260, + "train_speed(iter/s)": 0.448025 + }, + { + "acc": 0.64536734, + "epoch": 0.8692288178589549, + "grad_norm": 5.875, + "learning_rate": 6.472344325541859e-06, + "loss": 1.66664047, + "memory(GiB)": 111.15, + "step": 34265, + "train_speed(iter/s)": 0.44806 + }, + { + "acc": 0.64086199, + "epoch": 0.8693556570268899, + "grad_norm": 5.0625, + "learning_rate": 6.4713421643378715e-06, + "loss": 1.68285332, + "memory(GiB)": 111.15, + "step": 34270, + "train_speed(iter/s)": 0.448094 + }, + { + "acc": 0.64286485, + "epoch": 0.869482496194825, + "grad_norm": 5.25, + "learning_rate": 6.470339938417685e-06, + "loss": 1.5917141, + "memory(GiB)": 111.15, + "step": 34275, + "train_speed(iter/s)": 0.448129 + }, + { + "acc": 0.66218743, + "epoch": 0.86960933536276, + "grad_norm": 5.59375, + "learning_rate": 6.469337647825384e-06, + "loss": 1.6064188, + "memory(GiB)": 111.15, + "step": 34280, + "train_speed(iter/s)": 0.448164 + }, + { + "acc": 0.63540525, + "epoch": 0.8697361745306951, + "grad_norm": 7.71875, + "learning_rate": 6.468335292605053e-06, + "loss": 1.63565807, + "memory(GiB)": 111.15, + "step": 34285, + "train_speed(iter/s)": 0.448199 + }, + { + "acc": 0.66602631, + "epoch": 0.8698630136986302, + "grad_norm": 5.5625, + "learning_rate": 6.467332872800779e-06, + "loss": 1.52931166, + "memory(GiB)": 111.15, + "step": 34290, + "train_speed(iter/s)": 0.448233 + }, + { + "acc": 0.65252357, + "epoch": 0.8699898528665652, + "grad_norm": 5.21875, + "learning_rate": 6.466330388456655e-06, + "loss": 1.65180588, + "memory(GiB)": 111.15, + "step": 34295, + "train_speed(iter/s)": 0.448268 + }, + { + "acc": 0.64218988, + "epoch": 0.8701166920345003, + "grad_norm": 6.25, + "learning_rate": 6.465327839616774e-06, + "loss": 1.66812859, + "memory(GiB)": 111.15, + "step": 34300, + "train_speed(iter/s)": 0.448303 + }, + { + "epoch": 0.8701166920345003, + "eval_acc": 0.6431231114306567, + "eval_loss": 1.6095571517944336, + "eval_runtime": 113.8791, + "eval_samples_per_second": 55.936, + "eval_steps_per_second": 27.968, + "step": 34300 + }, + { + "acc": 0.65879116, + "epoch": 0.8702435312024354, + "grad_norm": 5.46875, + "learning_rate": 6.464325226325232e-06, + "loss": 1.59946556, + "memory(GiB)": 111.15, + "step": 34305, + "train_speed(iter/s)": 0.447644 + }, + { + "acc": 0.64151196, + "epoch": 0.8703703703703703, + "grad_norm": 5.9375, + "learning_rate": 6.46332254862613e-06, + "loss": 1.69849758, + "memory(GiB)": 111.15, + "step": 34310, + "train_speed(iter/s)": 0.447678 + }, + { + "acc": 0.652669, + "epoch": 0.8704972095383054, + "grad_norm": 4.4375, + "learning_rate": 6.462319806563568e-06, + "loss": 1.62395058, + "memory(GiB)": 111.15, + "step": 34315, + "train_speed(iter/s)": 0.447712 + }, + { + "acc": 0.6418541, + "epoch": 0.8706240487062404, + "grad_norm": 5.53125, + "learning_rate": 6.461317000181653e-06, + "loss": 1.57451, + "memory(GiB)": 111.15, + "step": 34320, + "train_speed(iter/s)": 0.447746 + }, + { + "acc": 0.64713812, + "epoch": 0.8707508878741755, + "grad_norm": 6.40625, + "learning_rate": 6.460314129524491e-06, + "loss": 1.66305046, + "memory(GiB)": 111.15, + "step": 34325, + "train_speed(iter/s)": 0.44778 + }, + { + "acc": 0.65009332, + "epoch": 0.8708777270421106, + "grad_norm": 4.78125, + "learning_rate": 6.4593111946361945e-06, + "loss": 1.62637062, + "memory(GiB)": 111.15, + "step": 34330, + "train_speed(iter/s)": 0.447814 + }, + { + "acc": 0.6416604, + "epoch": 0.8710045662100456, + "grad_norm": 5.09375, + "learning_rate": 6.458308195560874e-06, + "loss": 1.62182865, + "memory(GiB)": 111.15, + "step": 34335, + "train_speed(iter/s)": 0.447849 + }, + { + "acc": 0.63678265, + "epoch": 0.8711314053779807, + "grad_norm": 6.28125, + "learning_rate": 6.4573051323426515e-06, + "loss": 1.70713615, + "memory(GiB)": 111.15, + "step": 34340, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.65044904, + "epoch": 0.8712582445459158, + "grad_norm": 6.75, + "learning_rate": 6.456302005025641e-06, + "loss": 1.67481651, + "memory(GiB)": 111.15, + "step": 34345, + "train_speed(iter/s)": 0.447917 + }, + { + "acc": 0.63638797, + "epoch": 0.8713850837138508, + "grad_norm": 6.3125, + "learning_rate": 6.4552988136539675e-06, + "loss": 1.70379868, + "memory(GiB)": 111.15, + "step": 34350, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.65994253, + "epoch": 0.8715119228817859, + "grad_norm": 6.0, + "learning_rate": 6.454295558271752e-06, + "loss": 1.60382729, + "memory(GiB)": 111.15, + "step": 34355, + "train_speed(iter/s)": 0.447985 + }, + { + "acc": 0.64621019, + "epoch": 0.8716387620497209, + "grad_norm": 5.09375, + "learning_rate": 6.4532922389231275e-06, + "loss": 1.58820381, + "memory(GiB)": 111.15, + "step": 34360, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.66187363, + "epoch": 0.871765601217656, + "grad_norm": 4.90625, + "learning_rate": 6.452288855652222e-06, + "loss": 1.58590364, + "memory(GiB)": 111.15, + "step": 34365, + "train_speed(iter/s)": 0.448053 + }, + { + "acc": 0.65449305, + "epoch": 0.8718924403855911, + "grad_norm": 4.8125, + "learning_rate": 6.451285408503167e-06, + "loss": 1.60006485, + "memory(GiB)": 111.15, + "step": 34370, + "train_speed(iter/s)": 0.448087 + }, + { + "acc": 0.6422163, + "epoch": 0.8720192795535261, + "grad_norm": 5.21875, + "learning_rate": 6.450281897520102e-06, + "loss": 1.65589714, + "memory(GiB)": 111.15, + "step": 34375, + "train_speed(iter/s)": 0.448121 + }, + { + "acc": 0.66184196, + "epoch": 0.8721461187214612, + "grad_norm": 4.90625, + "learning_rate": 6.449278322747164e-06, + "loss": 1.59704943, + "memory(GiB)": 111.15, + "step": 34380, + "train_speed(iter/s)": 0.448155 + }, + { + "acc": 0.66488466, + "epoch": 0.8722729578893963, + "grad_norm": 5.46875, + "learning_rate": 6.448274684228494e-06, + "loss": 1.61541309, + "memory(GiB)": 111.15, + "step": 34385, + "train_speed(iter/s)": 0.44819 + }, + { + "acc": 0.67182274, + "epoch": 0.8723997970573313, + "grad_norm": 5.6875, + "learning_rate": 6.447270982008237e-06, + "loss": 1.56033363, + "memory(GiB)": 111.15, + "step": 34390, + "train_speed(iter/s)": 0.448224 + }, + { + "acc": 0.64698496, + "epoch": 0.8725266362252664, + "grad_norm": 5.84375, + "learning_rate": 6.446267216130541e-06, + "loss": 1.67104511, + "memory(GiB)": 111.15, + "step": 34395, + "train_speed(iter/s)": 0.448258 + }, + { + "acc": 0.65476718, + "epoch": 0.8726534753932014, + "grad_norm": 5.0, + "learning_rate": 6.4452633866395555e-06, + "loss": 1.63049641, + "memory(GiB)": 111.15, + "step": 34400, + "train_speed(iter/s)": 0.448292 + }, + { + "epoch": 0.8726534753932014, + "eval_acc": 0.643139822123035, + "eval_loss": 1.6097142696380615, + "eval_runtime": 114.0061, + "eval_samples_per_second": 55.874, + "eval_steps_per_second": 27.937, + "step": 34400 + }, + { + "acc": 0.66110506, + "epoch": 0.8727803145611365, + "grad_norm": 4.65625, + "learning_rate": 6.444259493579433e-06, + "loss": 1.6210043, + "memory(GiB)": 111.15, + "step": 34405, + "train_speed(iter/s)": 0.447636 + }, + { + "acc": 0.65170808, + "epoch": 0.8729071537290716, + "grad_norm": 5.625, + "learning_rate": 6.443255536994331e-06, + "loss": 1.61521606, + "memory(GiB)": 111.15, + "step": 34410, + "train_speed(iter/s)": 0.44767 + }, + { + "acc": 0.6474081, + "epoch": 0.8730339928970066, + "grad_norm": 5.5, + "learning_rate": 6.442251516928406e-06, + "loss": 1.62728405, + "memory(GiB)": 111.15, + "step": 34415, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.66347561, + "epoch": 0.8731608320649417, + "grad_norm": 4.46875, + "learning_rate": 6.441247433425821e-06, + "loss": 1.58518572, + "memory(GiB)": 111.15, + "step": 34420, + "train_speed(iter/s)": 0.44774 + }, + { + "acc": 0.64371972, + "epoch": 0.8732876712328768, + "grad_norm": 5.34375, + "learning_rate": 6.4402432865307384e-06, + "loss": 1.67025642, + "memory(GiB)": 111.15, + "step": 34425, + "train_speed(iter/s)": 0.447775 + }, + { + "acc": 0.64176245, + "epoch": 0.8734145104008117, + "grad_norm": 5.1875, + "learning_rate": 6.439239076287327e-06, + "loss": 1.68048458, + "memory(GiB)": 111.15, + "step": 34430, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.65044518, + "epoch": 0.8735413495687468, + "grad_norm": 5.34375, + "learning_rate": 6.438234802739753e-06, + "loss": 1.6400465, + "memory(GiB)": 111.15, + "step": 34435, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.65707197, + "epoch": 0.8736681887366818, + "grad_norm": 6.25, + "learning_rate": 6.4372304659321935e-06, + "loss": 1.62755203, + "memory(GiB)": 111.15, + "step": 34440, + "train_speed(iter/s)": 0.447879 + }, + { + "acc": 0.65013928, + "epoch": 0.8737950279046169, + "grad_norm": 5.0, + "learning_rate": 6.43622606590882e-06, + "loss": 1.65497513, + "memory(GiB)": 111.15, + "step": 34445, + "train_speed(iter/s)": 0.447914 + }, + { + "acc": 0.63802776, + "epoch": 0.873921867072552, + "grad_norm": 5.25, + "learning_rate": 6.4352216027138125e-06, + "loss": 1.61181698, + "memory(GiB)": 111.15, + "step": 34450, + "train_speed(iter/s)": 0.447948 + }, + { + "acc": 0.640662, + "epoch": 0.874048706240487, + "grad_norm": 5.96875, + "learning_rate": 6.434217076391351e-06, + "loss": 1.63862419, + "memory(GiB)": 111.15, + "step": 34455, + "train_speed(iter/s)": 0.447983 + }, + { + "acc": 0.64873023, + "epoch": 0.8741755454084221, + "grad_norm": 5.28125, + "learning_rate": 6.433212486985618e-06, + "loss": 1.63195534, + "memory(GiB)": 111.15, + "step": 34460, + "train_speed(iter/s)": 0.448018 + }, + { + "acc": 0.64791031, + "epoch": 0.8743023845763572, + "grad_norm": 4.53125, + "learning_rate": 6.432207834540802e-06, + "loss": 1.63890209, + "memory(GiB)": 111.15, + "step": 34465, + "train_speed(iter/s)": 0.448053 + }, + { + "acc": 0.6639709, + "epoch": 0.8744292237442922, + "grad_norm": 4.84375, + "learning_rate": 6.431203119101093e-06, + "loss": 1.6085247, + "memory(GiB)": 111.15, + "step": 34470, + "train_speed(iter/s)": 0.448088 + }, + { + "acc": 0.64108644, + "epoch": 0.8745560629122273, + "grad_norm": 5.625, + "learning_rate": 6.430198340710677e-06, + "loss": 1.61292038, + "memory(GiB)": 111.15, + "step": 34475, + "train_speed(iter/s)": 0.448123 + }, + { + "acc": 0.66255517, + "epoch": 0.8746829020801623, + "grad_norm": 5.59375, + "learning_rate": 6.4291934994137566e-06, + "loss": 1.59935665, + "memory(GiB)": 111.15, + "step": 34480, + "train_speed(iter/s)": 0.448158 + }, + { + "acc": 0.6590415, + "epoch": 0.8748097412480974, + "grad_norm": 6.28125, + "learning_rate": 6.428188595254521e-06, + "loss": 1.62850609, + "memory(GiB)": 111.15, + "step": 34485, + "train_speed(iter/s)": 0.448193 + }, + { + "acc": 0.66166492, + "epoch": 0.8749365804160325, + "grad_norm": 6.78125, + "learning_rate": 6.427183628277178e-06, + "loss": 1.65128384, + "memory(GiB)": 111.15, + "step": 34490, + "train_speed(iter/s)": 0.448227 + }, + { + "acc": 0.64320507, + "epoch": 0.8750634195839675, + "grad_norm": 4.8125, + "learning_rate": 6.426178598525925e-06, + "loss": 1.69023361, + "memory(GiB)": 111.15, + "step": 34495, + "train_speed(iter/s)": 0.448262 + }, + { + "acc": 0.64880209, + "epoch": 0.8751902587519026, + "grad_norm": 4.53125, + "learning_rate": 6.4251735060449725e-06, + "loss": 1.67470207, + "memory(GiB)": 111.15, + "step": 34500, + "train_speed(iter/s)": 0.448297 + }, + { + "epoch": 0.8751902587519026, + "eval_acc": 0.6430224295090775, + "eval_loss": 1.609557032585144, + "eval_runtime": 114.3858, + "eval_samples_per_second": 55.689, + "eval_steps_per_second": 27.844, + "step": 34500 + }, + { + "acc": 0.63595996, + "epoch": 0.8753170979198377, + "grad_norm": 5.125, + "learning_rate": 6.424168350878524e-06, + "loss": 1.64846992, + "memory(GiB)": 111.15, + "step": 34505, + "train_speed(iter/s)": 0.447639 + }, + { + "acc": 0.65803919, + "epoch": 0.8754439370877727, + "grad_norm": 6.375, + "learning_rate": 6.423163133070792e-06, + "loss": 1.56049986, + "memory(GiB)": 111.15, + "step": 34510, + "train_speed(iter/s)": 0.447674 + }, + { + "acc": 0.653409, + "epoch": 0.8755707762557078, + "grad_norm": 5.59375, + "learning_rate": 6.422157852665993e-06, + "loss": 1.69092751, + "memory(GiB)": 111.15, + "step": 34515, + "train_speed(iter/s)": 0.447708 + }, + { + "acc": 0.65726824, + "epoch": 0.8756976154236428, + "grad_norm": 5.28125, + "learning_rate": 6.421152509708342e-06, + "loss": 1.57798347, + "memory(GiB)": 111.15, + "step": 34520, + "train_speed(iter/s)": 0.447742 + }, + { + "acc": 0.6533144, + "epoch": 0.8758244545915779, + "grad_norm": 6.15625, + "learning_rate": 6.4201471042420595e-06, + "loss": 1.59954567, + "memory(GiB)": 111.15, + "step": 34525, + "train_speed(iter/s)": 0.447777 + }, + { + "acc": 0.64692063, + "epoch": 0.875951293759513, + "grad_norm": 4.78125, + "learning_rate": 6.419141636311366e-06, + "loss": 1.60424633, + "memory(GiB)": 111.15, + "step": 34530, + "train_speed(iter/s)": 0.447811 + }, + { + "acc": 0.65290022, + "epoch": 0.876078132927448, + "grad_norm": 5.5625, + "learning_rate": 6.4181361059604875e-06, + "loss": 1.59551659, + "memory(GiB)": 111.15, + "step": 34535, + "train_speed(iter/s)": 0.447846 + }, + { + "acc": 0.63786926, + "epoch": 0.8762049720953831, + "grad_norm": 5.21875, + "learning_rate": 6.4171305132336515e-06, + "loss": 1.65199051, + "memory(GiB)": 111.15, + "step": 34540, + "train_speed(iter/s)": 0.44788 + }, + { + "acc": 0.64891663, + "epoch": 0.8763318112633182, + "grad_norm": 4.84375, + "learning_rate": 6.416124858175088e-06, + "loss": 1.60882149, + "memory(GiB)": 111.15, + "step": 34545, + "train_speed(iter/s)": 0.447914 + }, + { + "acc": 0.65133753, + "epoch": 0.8764586504312532, + "grad_norm": 5.40625, + "learning_rate": 6.415119140829031e-06, + "loss": 1.73576355, + "memory(GiB)": 111.15, + "step": 34550, + "train_speed(iter/s)": 0.447948 + }, + { + "acc": 0.65058374, + "epoch": 0.8765854895991883, + "grad_norm": 5.40625, + "learning_rate": 6.414113361239715e-06, + "loss": 1.62504807, + "memory(GiB)": 111.15, + "step": 34555, + "train_speed(iter/s)": 0.447983 + }, + { + "acc": 0.66102953, + "epoch": 0.8767123287671232, + "grad_norm": 5.53125, + "learning_rate": 6.4131075194513825e-06, + "loss": 1.51957378, + "memory(GiB)": 111.15, + "step": 34560, + "train_speed(iter/s)": 0.448017 + }, + { + "acc": 0.66913786, + "epoch": 0.8768391679350583, + "grad_norm": 4.5, + "learning_rate": 6.41210161550827e-06, + "loss": 1.44996338, + "memory(GiB)": 111.15, + "step": 34565, + "train_speed(iter/s)": 0.448051 + }, + { + "acc": 0.64988966, + "epoch": 0.8769660071029934, + "grad_norm": 4.71875, + "learning_rate": 6.411095649454626e-06, + "loss": 1.61549377, + "memory(GiB)": 111.15, + "step": 34570, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.64860592, + "epoch": 0.8770928462709284, + "grad_norm": 5.625, + "learning_rate": 6.410089621334693e-06, + "loss": 1.6418087, + "memory(GiB)": 111.15, + "step": 34575, + "train_speed(iter/s)": 0.44812 + }, + { + "acc": 0.65377812, + "epoch": 0.8772196854388635, + "grad_norm": 5.09375, + "learning_rate": 6.4090835311927236e-06, + "loss": 1.53700314, + "memory(GiB)": 111.15, + "step": 34580, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.67108707, + "epoch": 0.8773465246067986, + "grad_norm": 5.40625, + "learning_rate": 6.40807737907297e-06, + "loss": 1.58395863, + "memory(GiB)": 111.15, + "step": 34585, + "train_speed(iter/s)": 0.448188 + }, + { + "acc": 0.64728599, + "epoch": 0.8774733637747336, + "grad_norm": 6.875, + "learning_rate": 6.407071165019686e-06, + "loss": 1.67352276, + "memory(GiB)": 111.15, + "step": 34590, + "train_speed(iter/s)": 0.448222 + }, + { + "acc": 0.66170816, + "epoch": 0.8776002029426687, + "grad_norm": 4.90625, + "learning_rate": 6.40606488907713e-06, + "loss": 1.61527843, + "memory(GiB)": 111.15, + "step": 34595, + "train_speed(iter/s)": 0.448256 + }, + { + "acc": 0.62888117, + "epoch": 0.8777270421106037, + "grad_norm": 4.9375, + "learning_rate": 6.4050585512895624e-06, + "loss": 1.6425396, + "memory(GiB)": 111.15, + "step": 34600, + "train_speed(iter/s)": 0.448291 + }, + { + "epoch": 0.8777270421106037, + "eval_acc": 0.6430892722785906, + "eval_loss": 1.609403371810913, + "eval_runtime": 114.1131, + "eval_samples_per_second": 55.822, + "eval_steps_per_second": 27.911, + "step": 34600 + }, + { + "acc": 0.65663419, + "epoch": 0.8778538812785388, + "grad_norm": 5.09375, + "learning_rate": 6.4040521517012475e-06, + "loss": 1.57980537, + "memory(GiB)": 111.15, + "step": 34605, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.64003115, + "epoch": 0.8779807204464739, + "grad_norm": 5.6875, + "learning_rate": 6.40304569035645e-06, + "loss": 1.65386276, + "memory(GiB)": 111.15, + "step": 34610, + "train_speed(iter/s)": 0.447671 + }, + { + "acc": 0.65018702, + "epoch": 0.8781075596144089, + "grad_norm": 5.34375, + "learning_rate": 6.402039167299439e-06, + "loss": 1.6275425, + "memory(GiB)": 111.15, + "step": 34615, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.64198742, + "epoch": 0.878234398782344, + "grad_norm": 5.40625, + "learning_rate": 6.401032582574485e-06, + "loss": 1.65648232, + "memory(GiB)": 111.15, + "step": 34620, + "train_speed(iter/s)": 0.447739 + }, + { + "acc": 0.65115027, + "epoch": 0.8783612379502791, + "grad_norm": 6.40625, + "learning_rate": 6.400025936225862e-06, + "loss": 1.63318577, + "memory(GiB)": 111.15, + "step": 34625, + "train_speed(iter/s)": 0.447773 + }, + { + "acc": 0.6400795, + "epoch": 0.8784880771182141, + "grad_norm": 4.90625, + "learning_rate": 6.399019228297851e-06, + "loss": 1.67132969, + "memory(GiB)": 111.15, + "step": 34630, + "train_speed(iter/s)": 0.447807 + }, + { + "acc": 0.65464153, + "epoch": 0.8786149162861492, + "grad_norm": 7.15625, + "learning_rate": 6.398012458834724e-06, + "loss": 1.60673714, + "memory(GiB)": 111.15, + "step": 34635, + "train_speed(iter/s)": 0.447841 + }, + { + "acc": 0.64395008, + "epoch": 0.8787417554540842, + "grad_norm": 4.78125, + "learning_rate": 6.397005627880771e-06, + "loss": 1.6186985, + "memory(GiB)": 111.15, + "step": 34640, + "train_speed(iter/s)": 0.447875 + }, + { + "acc": 0.65225563, + "epoch": 0.8788685946220193, + "grad_norm": 5.84375, + "learning_rate": 6.395998735480271e-06, + "loss": 1.66707726, + "memory(GiB)": 111.15, + "step": 34645, + "train_speed(iter/s)": 0.44791 + }, + { + "acc": 0.65440006, + "epoch": 0.8789954337899544, + "grad_norm": 4.9375, + "learning_rate": 6.394991781677516e-06, + "loss": 1.56823139, + "memory(GiB)": 111.15, + "step": 34650, + "train_speed(iter/s)": 0.447944 + }, + { + "acc": 0.64193106, + "epoch": 0.8791222729578894, + "grad_norm": 5.6875, + "learning_rate": 6.393984766516792e-06, + "loss": 1.68392334, + "memory(GiB)": 111.15, + "step": 34655, + "train_speed(iter/s)": 0.447978 + }, + { + "acc": 0.63486409, + "epoch": 0.8792491121258245, + "grad_norm": 5.375, + "learning_rate": 6.392977690042395e-06, + "loss": 1.75684338, + "memory(GiB)": 111.15, + "step": 34660, + "train_speed(iter/s)": 0.448012 + }, + { + "acc": 0.65631189, + "epoch": 0.8793759512937596, + "grad_norm": 6.3125, + "learning_rate": 6.3919705522986205e-06, + "loss": 1.60642853, + "memory(GiB)": 111.15, + "step": 34665, + "train_speed(iter/s)": 0.448046 + }, + { + "acc": 0.67523003, + "epoch": 0.8795027904616946, + "grad_norm": 5.28125, + "learning_rate": 6.390963353329767e-06, + "loss": 1.50093975, + "memory(GiB)": 111.15, + "step": 34670, + "train_speed(iter/s)": 0.44808 + }, + { + "acc": 0.63496456, + "epoch": 0.8796296296296297, + "grad_norm": 4.90625, + "learning_rate": 6.389956093180134e-06, + "loss": 1.68216171, + "memory(GiB)": 111.15, + "step": 34675, + "train_speed(iter/s)": 0.448114 + }, + { + "acc": 0.66046185, + "epoch": 0.8797564687975646, + "grad_norm": 4.96875, + "learning_rate": 6.388948771894025e-06, + "loss": 1.59488506, + "memory(GiB)": 111.15, + "step": 34680, + "train_speed(iter/s)": 0.448148 + }, + { + "acc": 0.66158018, + "epoch": 0.8798833079654997, + "grad_norm": 4.71875, + "learning_rate": 6.38794138951575e-06, + "loss": 1.59309235, + "memory(GiB)": 111.15, + "step": 34685, + "train_speed(iter/s)": 0.448182 + }, + { + "acc": 0.66133351, + "epoch": 0.8800101471334348, + "grad_norm": 5.46875, + "learning_rate": 6.386933946089615e-06, + "loss": 1.59417095, + "memory(GiB)": 111.15, + "step": 34690, + "train_speed(iter/s)": 0.448216 + }, + { + "acc": 0.65513678, + "epoch": 0.8801369863013698, + "grad_norm": 6.15625, + "learning_rate": 6.385926441659933e-06, + "loss": 1.62426605, + "memory(GiB)": 111.15, + "step": 34695, + "train_speed(iter/s)": 0.44825 + }, + { + "acc": 0.64424953, + "epoch": 0.8802638254693049, + "grad_norm": 6.0625, + "learning_rate": 6.38491887627102e-06, + "loss": 1.65667686, + "memory(GiB)": 111.15, + "step": 34700, + "train_speed(iter/s)": 0.448285 + }, + { + "epoch": 0.8802638254693049, + "eval_acc": 0.6431193515248715, + "eval_loss": 1.609327793121338, + "eval_runtime": 113.5366, + "eval_samples_per_second": 56.105, + "eval_steps_per_second": 28.053, + "step": 34700 + }, + { + "acc": 0.64580173, + "epoch": 0.88039066463724, + "grad_norm": 6.28125, + "learning_rate": 6.383911249967188e-06, + "loss": 1.64650841, + "memory(GiB)": 111.15, + "step": 34705, + "train_speed(iter/s)": 0.447635 + }, + { + "acc": 0.64073262, + "epoch": 0.880517503805175, + "grad_norm": 5.09375, + "learning_rate": 6.382903562792764e-06, + "loss": 1.63277855, + "memory(GiB)": 111.15, + "step": 34710, + "train_speed(iter/s)": 0.447669 + }, + { + "acc": 0.63402076, + "epoch": 0.8806443429731101, + "grad_norm": 6.125, + "learning_rate": 6.381895814792065e-06, + "loss": 1.71601219, + "memory(GiB)": 111.15, + "step": 34715, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.64763756, + "epoch": 0.8807711821410451, + "grad_norm": 5.0625, + "learning_rate": 6.38088800600942e-06, + "loss": 1.63242321, + "memory(GiB)": 111.15, + "step": 34720, + "train_speed(iter/s)": 0.447737 + }, + { + "acc": 0.65155511, + "epoch": 0.8808980213089802, + "grad_norm": 5.0, + "learning_rate": 6.3798801364891535e-06, + "loss": 1.64902935, + "memory(GiB)": 111.15, + "step": 34725, + "train_speed(iter/s)": 0.447771 + }, + { + "acc": 0.64373446, + "epoch": 0.8810248604769153, + "grad_norm": 5.9375, + "learning_rate": 6.378872206275599e-06, + "loss": 1.63523464, + "memory(GiB)": 111.15, + "step": 34730, + "train_speed(iter/s)": 0.447805 + }, + { + "acc": 0.65634322, + "epoch": 0.8811516996448503, + "grad_norm": 5.34375, + "learning_rate": 6.377864215413088e-06, + "loss": 1.57586002, + "memory(GiB)": 111.15, + "step": 34735, + "train_speed(iter/s)": 0.447839 + }, + { + "acc": 0.65515175, + "epoch": 0.8812785388127854, + "grad_norm": 5.9375, + "learning_rate": 6.376856163945957e-06, + "loss": 1.65526428, + "memory(GiB)": 111.15, + "step": 34740, + "train_speed(iter/s)": 0.447873 + }, + { + "acc": 0.65555949, + "epoch": 0.8814053779807205, + "grad_norm": 5.96875, + "learning_rate": 6.375848051918546e-06, + "loss": 1.60933228, + "memory(GiB)": 111.15, + "step": 34745, + "train_speed(iter/s)": 0.447906 + }, + { + "acc": 0.65565929, + "epoch": 0.8815322171486555, + "grad_norm": 5.09375, + "learning_rate": 6.374839879375194e-06, + "loss": 1.58793812, + "memory(GiB)": 111.15, + "step": 34750, + "train_speed(iter/s)": 0.44794 + }, + { + "acc": 0.64354773, + "epoch": 0.8816590563165906, + "grad_norm": 4.84375, + "learning_rate": 6.373831646360245e-06, + "loss": 1.59737854, + "memory(GiB)": 111.15, + "step": 34755, + "train_speed(iter/s)": 0.447974 + }, + { + "acc": 0.65238714, + "epoch": 0.8817858954845256, + "grad_norm": 5.53125, + "learning_rate": 6.372823352918048e-06, + "loss": 1.64025497, + "memory(GiB)": 111.15, + "step": 34760, + "train_speed(iter/s)": 0.448008 + }, + { + "acc": 0.64372177, + "epoch": 0.8819127346524607, + "grad_norm": 5.34375, + "learning_rate": 6.371814999092951e-06, + "loss": 1.60809784, + "memory(GiB)": 111.15, + "step": 34765, + "train_speed(iter/s)": 0.448042 + }, + { + "acc": 0.64996862, + "epoch": 0.8820395738203958, + "grad_norm": 5.0625, + "learning_rate": 6.370806584929305e-06, + "loss": 1.66322861, + "memory(GiB)": 111.15, + "step": 34770, + "train_speed(iter/s)": 0.448076 + }, + { + "acc": 0.65588951, + "epoch": 0.8821664129883308, + "grad_norm": 4.3125, + "learning_rate": 6.369798110471463e-06, + "loss": 1.61911087, + "memory(GiB)": 111.15, + "step": 34775, + "train_speed(iter/s)": 0.44811 + }, + { + "acc": 0.65067415, + "epoch": 0.8822932521562659, + "grad_norm": 5.6875, + "learning_rate": 6.368789575763787e-06, + "loss": 1.60255051, + "memory(GiB)": 111.15, + "step": 34780, + "train_speed(iter/s)": 0.448144 + }, + { + "acc": 0.64651403, + "epoch": 0.882420091324201, + "grad_norm": 6.15625, + "learning_rate": 6.367780980850633e-06, + "loss": 1.60644455, + "memory(GiB)": 111.15, + "step": 34785, + "train_speed(iter/s)": 0.448178 + }, + { + "acc": 0.64960489, + "epoch": 0.882546930492136, + "grad_norm": 5.78125, + "learning_rate": 6.366772325776367e-06, + "loss": 1.64897404, + "memory(GiB)": 111.15, + "step": 34790, + "train_speed(iter/s)": 0.448212 + }, + { + "acc": 0.65357895, + "epoch": 0.882673769660071, + "grad_norm": 5.5, + "learning_rate": 6.365763610585349e-06, + "loss": 1.58374577, + "memory(GiB)": 111.15, + "step": 34795, + "train_speed(iter/s)": 0.448246 + }, + { + "acc": 0.67054491, + "epoch": 0.882800608828006, + "grad_norm": 5.90625, + "learning_rate": 6.3647548353219515e-06, + "loss": 1.59967499, + "memory(GiB)": 111.15, + "step": 34800, + "train_speed(iter/s)": 0.44828 + }, + { + "epoch": 0.882800608828006, + "eval_acc": 0.6430370513649085, + "eval_loss": 1.6092860698699951, + "eval_runtime": 114.1223, + "eval_samples_per_second": 55.817, + "eval_steps_per_second": 27.909, + "step": 34800 + }, + { + "acc": 0.66368871, + "epoch": 0.8829274479959411, + "grad_norm": 6.03125, + "learning_rate": 6.363746000030543e-06, + "loss": 1.646068, + "memory(GiB)": 111.15, + "step": 34805, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.6460813, + "epoch": 0.8830542871638762, + "grad_norm": 5.96875, + "learning_rate": 6.362737104755497e-06, + "loss": 1.65825863, + "memory(GiB)": 111.15, + "step": 34810, + "train_speed(iter/s)": 0.447665 + }, + { + "acc": 0.66513128, + "epoch": 0.8831811263318112, + "grad_norm": 5.15625, + "learning_rate": 6.361728149541188e-06, + "loss": 1.60586357, + "memory(GiB)": 111.15, + "step": 34815, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.65045853, + "epoch": 0.8833079654997463, + "grad_norm": 5.0, + "learning_rate": 6.360719134431995e-06, + "loss": 1.66029167, + "memory(GiB)": 111.15, + "step": 34820, + "train_speed(iter/s)": 0.447733 + }, + { + "acc": 0.66809187, + "epoch": 0.8834348046676814, + "grad_norm": 6.59375, + "learning_rate": 6.359710059472299e-06, + "loss": 1.61884613, + "memory(GiB)": 111.15, + "step": 34825, + "train_speed(iter/s)": 0.447767 + }, + { + "acc": 0.64987679, + "epoch": 0.8835616438356164, + "grad_norm": 5.59375, + "learning_rate": 6.358700924706486e-06, + "loss": 1.6418808, + "memory(GiB)": 111.15, + "step": 34830, + "train_speed(iter/s)": 0.4478 + }, + { + "acc": 0.65272479, + "epoch": 0.8836884830035515, + "grad_norm": 5.5625, + "learning_rate": 6.357691730178939e-06, + "loss": 1.6324152, + "memory(GiB)": 111.15, + "step": 34835, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.63999968, + "epoch": 0.8838153221714865, + "grad_norm": 5.0625, + "learning_rate": 6.356682475934048e-06, + "loss": 1.69148636, + "memory(GiB)": 111.15, + "step": 34840, + "train_speed(iter/s)": 0.447868 + }, + { + "acc": 0.65458541, + "epoch": 0.8839421613394216, + "grad_norm": 6.03125, + "learning_rate": 6.3556731620162036e-06, + "loss": 1.61839123, + "memory(GiB)": 111.15, + "step": 34845, + "train_speed(iter/s)": 0.447903 + }, + { + "acc": 0.6588851, + "epoch": 0.8840690005073567, + "grad_norm": 6.53125, + "learning_rate": 6.354663788469803e-06, + "loss": 1.60629597, + "memory(GiB)": 111.15, + "step": 34850, + "train_speed(iter/s)": 0.447936 + }, + { + "acc": 0.64992585, + "epoch": 0.8841958396752917, + "grad_norm": 5.59375, + "learning_rate": 6.353654355339238e-06, + "loss": 1.68643875, + "memory(GiB)": 111.15, + "step": 34855, + "train_speed(iter/s)": 0.447971 + }, + { + "acc": 0.65827665, + "epoch": 0.8843226788432268, + "grad_norm": 5.71875, + "learning_rate": 6.352644862668914e-06, + "loss": 1.61285172, + "memory(GiB)": 111.15, + "step": 34860, + "train_speed(iter/s)": 0.448004 + }, + { + "acc": 0.66502066, + "epoch": 0.8844495180111619, + "grad_norm": 5.0625, + "learning_rate": 6.351635310503228e-06, + "loss": 1.58902035, + "memory(GiB)": 111.15, + "step": 34865, + "train_speed(iter/s)": 0.448038 + }, + { + "acc": 0.65392323, + "epoch": 0.8845763571790969, + "grad_norm": 6.6875, + "learning_rate": 6.3506256988865865e-06, + "loss": 1.59489975, + "memory(GiB)": 111.15, + "step": 34870, + "train_speed(iter/s)": 0.448073 + }, + { + "acc": 0.65719404, + "epoch": 0.884703196347032, + "grad_norm": 5.0, + "learning_rate": 6.349616027863397e-06, + "loss": 1.520364, + "memory(GiB)": 111.15, + "step": 34875, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.65147948, + "epoch": 0.884830035514967, + "grad_norm": 4.90625, + "learning_rate": 6.34860629747807e-06, + "loss": 1.6194191, + "memory(GiB)": 111.15, + "step": 34880, + "train_speed(iter/s)": 0.448141 + }, + { + "acc": 0.64253235, + "epoch": 0.8849568746829021, + "grad_norm": 6.71875, + "learning_rate": 6.347596507775016e-06, + "loss": 1.63857574, + "memory(GiB)": 111.15, + "step": 34885, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.65490904, + "epoch": 0.8850837138508372, + "grad_norm": 4.90625, + "learning_rate": 6.3465866587986505e-06, + "loss": 1.53005581, + "memory(GiB)": 111.15, + "step": 34890, + "train_speed(iter/s)": 0.448208 + }, + { + "acc": 0.66387405, + "epoch": 0.8852105530187722, + "grad_norm": 4.78125, + "learning_rate": 6.345576750593392e-06, + "loss": 1.58788033, + "memory(GiB)": 111.15, + "step": 34895, + "train_speed(iter/s)": 0.448242 + }, + { + "acc": 0.63290539, + "epoch": 0.8853373921867073, + "grad_norm": 6.1875, + "learning_rate": 6.34456678320366e-06, + "loss": 1.7029932, + "memory(GiB)": 111.15, + "step": 34900, + "train_speed(iter/s)": 0.448276 + }, + { + "epoch": 0.8853373921867073, + "eval_acc": 0.6431013875305649, + "eval_loss": 1.6092891693115234, + "eval_runtime": 113.4019, + "eval_samples_per_second": 56.172, + "eval_steps_per_second": 28.086, + "step": 34900 + }, + { + "acc": 0.64114714, + "epoch": 0.8854642313546424, + "grad_norm": 4.6875, + "learning_rate": 6.343556756673879e-06, + "loss": 1.6703146, + "memory(GiB)": 111.15, + "step": 34905, + "train_speed(iter/s)": 0.447632 + }, + { + "acc": 0.64677267, + "epoch": 0.8855910705225774, + "grad_norm": 4.65625, + "learning_rate": 6.3425466710484726e-06, + "loss": 1.6394825, + "memory(GiB)": 111.15, + "step": 34910, + "train_speed(iter/s)": 0.447665 + }, + { + "acc": 0.64145279, + "epoch": 0.8857179096905125, + "grad_norm": 5.75, + "learning_rate": 6.3415365263718686e-06, + "loss": 1.67763233, + "memory(GiB)": 111.15, + "step": 34915, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.64709215, + "epoch": 0.8858447488584474, + "grad_norm": 5.0, + "learning_rate": 6.340526322688501e-06, + "loss": 1.57831154, + "memory(GiB)": 111.15, + "step": 34920, + "train_speed(iter/s)": 0.447732 + }, + { + "acc": 0.65505872, + "epoch": 0.8859715880263825, + "grad_norm": 5.875, + "learning_rate": 6.339516060042798e-06, + "loss": 1.65564022, + "memory(GiB)": 111.15, + "step": 34925, + "train_speed(iter/s)": 0.447766 + }, + { + "acc": 0.63558779, + "epoch": 0.8860984271943176, + "grad_norm": 5.5625, + "learning_rate": 6.3385057384792e-06, + "loss": 1.65198135, + "memory(GiB)": 111.15, + "step": 34930, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.63833342, + "epoch": 0.8862252663622526, + "grad_norm": 6.375, + "learning_rate": 6.337495358042143e-06, + "loss": 1.65172653, + "memory(GiB)": 111.15, + "step": 34935, + "train_speed(iter/s)": 0.447832 + }, + { + "acc": 0.64745626, + "epoch": 0.8863521055301877, + "grad_norm": 5.25, + "learning_rate": 6.336484918776069e-06, + "loss": 1.66031723, + "memory(GiB)": 111.15, + "step": 34940, + "train_speed(iter/s)": 0.447866 + }, + { + "acc": 0.66223927, + "epoch": 0.8864789446981228, + "grad_norm": 4.84375, + "learning_rate": 6.335474420725421e-06, + "loss": 1.61480904, + "memory(GiB)": 111.15, + "step": 34945, + "train_speed(iter/s)": 0.447899 + }, + { + "acc": 0.63751144, + "epoch": 0.8866057838660578, + "grad_norm": 5.3125, + "learning_rate": 6.334463863934646e-06, + "loss": 1.66160088, + "memory(GiB)": 111.15, + "step": 34950, + "train_speed(iter/s)": 0.447933 + }, + { + "acc": 0.64297414, + "epoch": 0.8867326230339929, + "grad_norm": 4.6875, + "learning_rate": 6.333453248448192e-06, + "loss": 1.60042152, + "memory(GiB)": 111.15, + "step": 34955, + "train_speed(iter/s)": 0.447967 + }, + { + "acc": 0.62930713, + "epoch": 0.8868594622019279, + "grad_norm": 5.40625, + "learning_rate": 6.33244257431051e-06, + "loss": 1.76536121, + "memory(GiB)": 111.15, + "step": 34960, + "train_speed(iter/s)": 0.448001 + }, + { + "acc": 0.6521862, + "epoch": 0.886986301369863, + "grad_norm": 4.90625, + "learning_rate": 6.331431841566056e-06, + "loss": 1.64796829, + "memory(GiB)": 111.15, + "step": 34965, + "train_speed(iter/s)": 0.448034 + }, + { + "acc": 0.65478029, + "epoch": 0.8871131405377981, + "grad_norm": 6.1875, + "learning_rate": 6.330421050259283e-06, + "loss": 1.63373394, + "memory(GiB)": 111.15, + "step": 34970, + "train_speed(iter/s)": 0.448068 + }, + { + "acc": 0.64528475, + "epoch": 0.8872399797057331, + "grad_norm": 6.40625, + "learning_rate": 6.329410200434655e-06, + "loss": 1.56888676, + "memory(GiB)": 111.15, + "step": 34975, + "train_speed(iter/s)": 0.448102 + }, + { + "acc": 0.64811211, + "epoch": 0.8873668188736682, + "grad_norm": 5.03125, + "learning_rate": 6.328399292136629e-06, + "loss": 1.66445427, + "memory(GiB)": 111.15, + "step": 34980, + "train_speed(iter/s)": 0.448136 + }, + { + "acc": 0.64383354, + "epoch": 0.8874936580416033, + "grad_norm": 5.6875, + "learning_rate": 6.327388325409672e-06, + "loss": 1.65385933, + "memory(GiB)": 111.15, + "step": 34985, + "train_speed(iter/s)": 0.44817 + }, + { + "acc": 0.62844038, + "epoch": 0.8876204972095383, + "grad_norm": 4.78125, + "learning_rate": 6.326377300298251e-06, + "loss": 1.65047245, + "memory(GiB)": 111.15, + "step": 34990, + "train_speed(iter/s)": 0.448204 + }, + { + "acc": 0.66008787, + "epoch": 0.8877473363774734, + "grad_norm": 5.3125, + "learning_rate": 6.325366216846832e-06, + "loss": 1.61452103, + "memory(GiB)": 111.15, + "step": 34995, + "train_speed(iter/s)": 0.448237 + }, + { + "acc": 0.65793257, + "epoch": 0.8878741755454084, + "grad_norm": 6.3125, + "learning_rate": 6.324355075099893e-06, + "loss": 1.67592697, + "memory(GiB)": 111.15, + "step": 35000, + "train_speed(iter/s)": 0.448271 + }, + { + "epoch": 0.8878741755454084, + "eval_acc": 0.6430846768381866, + "eval_loss": 1.6092051267623901, + "eval_runtime": 113.7245, + "eval_samples_per_second": 56.013, + "eval_steps_per_second": 28.006, + "step": 35000 + }, + { + "acc": 0.64716396, + "epoch": 0.8880010147133435, + "grad_norm": 4.875, + "learning_rate": 6.3233438751019016e-06, + "loss": 1.65020752, + "memory(GiB)": 111.15, + "step": 35005, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.65882621, + "epoch": 0.8881278538812786, + "grad_norm": 5.90625, + "learning_rate": 6.322332616897341e-06, + "loss": 1.59352236, + "memory(GiB)": 111.15, + "step": 35010, + "train_speed(iter/s)": 0.44766 + }, + { + "acc": 0.66349912, + "epoch": 0.8882546930492136, + "grad_norm": 5.59375, + "learning_rate": 6.321321300530685e-06, + "loss": 1.50159092, + "memory(GiB)": 111.15, + "step": 35015, + "train_speed(iter/s)": 0.447694 + }, + { + "acc": 0.66465912, + "epoch": 0.8883815322171487, + "grad_norm": 6.0, + "learning_rate": 6.320309926046421e-06, + "loss": 1.58760328, + "memory(GiB)": 111.15, + "step": 35020, + "train_speed(iter/s)": 0.447727 + }, + { + "acc": 0.63615427, + "epoch": 0.8885083713850838, + "grad_norm": 5.9375, + "learning_rate": 6.319298493489032e-06, + "loss": 1.67938538, + "memory(GiB)": 111.15, + "step": 35025, + "train_speed(iter/s)": 0.44776 + }, + { + "acc": 0.64092011, + "epoch": 0.8886352105530188, + "grad_norm": 5.9375, + "learning_rate": 6.318287002903004e-06, + "loss": 1.59933052, + "memory(GiB)": 111.15, + "step": 35030, + "train_speed(iter/s)": 0.447794 + }, + { + "acc": 0.62803988, + "epoch": 0.8887620497209539, + "grad_norm": 5.6875, + "learning_rate": 6.317275454332829e-06, + "loss": 1.66207161, + "memory(GiB)": 111.15, + "step": 35035, + "train_speed(iter/s)": 0.447828 + }, + { + "acc": 0.6430089, + "epoch": 0.8888888888888888, + "grad_norm": 6.4375, + "learning_rate": 6.3162638478229965e-06, + "loss": 1.65603638, + "memory(GiB)": 111.15, + "step": 35040, + "train_speed(iter/s)": 0.447862 + }, + { + "acc": 0.64778814, + "epoch": 0.8890157280568239, + "grad_norm": 5.65625, + "learning_rate": 6.315252183418005e-06, + "loss": 1.62399139, + "memory(GiB)": 111.15, + "step": 35045, + "train_speed(iter/s)": 0.447895 + }, + { + "acc": 0.63704357, + "epoch": 0.889142567224759, + "grad_norm": 4.875, + "learning_rate": 6.31424046116235e-06, + "loss": 1.74830837, + "memory(GiB)": 111.15, + "step": 35050, + "train_speed(iter/s)": 0.447929 + }, + { + "acc": 0.64435329, + "epoch": 0.889269406392694, + "grad_norm": 5.09375, + "learning_rate": 6.313228681100532e-06, + "loss": 1.63337917, + "memory(GiB)": 111.15, + "step": 35055, + "train_speed(iter/s)": 0.447962 + }, + { + "acc": 0.66050153, + "epoch": 0.8893962455606291, + "grad_norm": 5.0625, + "learning_rate": 6.312216843277052e-06, + "loss": 1.59275589, + "memory(GiB)": 111.15, + "step": 35060, + "train_speed(iter/s)": 0.447996 + }, + { + "acc": 0.61366501, + "epoch": 0.8895230847285642, + "grad_norm": 5.34375, + "learning_rate": 6.3112049477364165e-06, + "loss": 1.7233511, + "memory(GiB)": 111.15, + "step": 35065, + "train_speed(iter/s)": 0.448029 + }, + { + "acc": 0.65200768, + "epoch": 0.8896499238964992, + "grad_norm": 6.03125, + "learning_rate": 6.310192994523137e-06, + "loss": 1.66232586, + "memory(GiB)": 111.15, + "step": 35070, + "train_speed(iter/s)": 0.448062 + }, + { + "acc": 0.64550676, + "epoch": 0.8897767630644343, + "grad_norm": 5.28125, + "learning_rate": 6.309180983681716e-06, + "loss": 1.63173599, + "memory(GiB)": 111.15, + "step": 35075, + "train_speed(iter/s)": 0.448096 + }, + { + "acc": 0.63843913, + "epoch": 0.8899036022323693, + "grad_norm": 5.0625, + "learning_rate": 6.308168915256671e-06, + "loss": 1.68556232, + "memory(GiB)": 111.15, + "step": 35080, + "train_speed(iter/s)": 0.448129 + }, + { + "acc": 0.63880029, + "epoch": 0.8900304414003044, + "grad_norm": 5.84375, + "learning_rate": 6.307156789292518e-06, + "loss": 1.69357662, + "memory(GiB)": 111.15, + "step": 35085, + "train_speed(iter/s)": 0.448163 + }, + { + "acc": 0.65982151, + "epoch": 0.8901572805682395, + "grad_norm": 5.3125, + "learning_rate": 6.306144605833773e-06, + "loss": 1.61862621, + "memory(GiB)": 111.15, + "step": 35090, + "train_speed(iter/s)": 0.448196 + }, + { + "acc": 0.63373127, + "epoch": 0.8902841197361745, + "grad_norm": 5.59375, + "learning_rate": 6.305132364924955e-06, + "loss": 1.62930679, + "memory(GiB)": 111.15, + "step": 35095, + "train_speed(iter/s)": 0.44823 + }, + { + "acc": 0.64103928, + "epoch": 0.8904109589041096, + "grad_norm": 6.3125, + "learning_rate": 6.3041200666105905e-06, + "loss": 1.63158073, + "memory(GiB)": 111.15, + "step": 35100, + "train_speed(iter/s)": 0.448263 + }, + { + "epoch": 0.8904109589041096, + "eval_acc": 0.6430500021515017, + "eval_loss": 1.6091042757034302, + "eval_runtime": 113.7128, + "eval_samples_per_second": 56.018, + "eval_steps_per_second": 28.009, + "step": 35100 + }, + { + "acc": 0.645959, + "epoch": 0.8905377980720447, + "grad_norm": 4.625, + "learning_rate": 6.303107710935202e-06, + "loss": 1.65226059, + "memory(GiB)": 111.15, + "step": 35105, + "train_speed(iter/s)": 0.447621 + }, + { + "acc": 0.63429556, + "epoch": 0.8906646372399797, + "grad_norm": 5.125, + "learning_rate": 6.302095297943319e-06, + "loss": 1.62112045, + "memory(GiB)": 111.15, + "step": 35110, + "train_speed(iter/s)": 0.447654 + }, + { + "acc": 0.64011493, + "epoch": 0.8907914764079148, + "grad_norm": 4.90625, + "learning_rate": 6.301082827679472e-06, + "loss": 1.63805733, + "memory(GiB)": 111.15, + "step": 35115, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.67653227, + "epoch": 0.8909183155758498, + "grad_norm": 6.0625, + "learning_rate": 6.300070300188192e-06, + "loss": 1.56956062, + "memory(GiB)": 111.15, + "step": 35120, + "train_speed(iter/s)": 0.447721 + }, + { + "acc": 0.65438004, + "epoch": 0.8910451547437849, + "grad_norm": 7.0625, + "learning_rate": 6.2990577155140164e-06, + "loss": 1.64688797, + "memory(GiB)": 111.15, + "step": 35125, + "train_speed(iter/s)": 0.447755 + }, + { + "acc": 0.65322514, + "epoch": 0.89117199391172, + "grad_norm": 5.53125, + "learning_rate": 6.298045073701483e-06, + "loss": 1.58845615, + "memory(GiB)": 111.15, + "step": 35130, + "train_speed(iter/s)": 0.447789 + }, + { + "acc": 0.64744053, + "epoch": 0.891298833079655, + "grad_norm": 5.15625, + "learning_rate": 6.29703237479513e-06, + "loss": 1.61314106, + "memory(GiB)": 111.15, + "step": 35135, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.64623408, + "epoch": 0.8914256722475901, + "grad_norm": 5.0, + "learning_rate": 6.296019618839505e-06, + "loss": 1.67203598, + "memory(GiB)": 111.15, + "step": 35140, + "train_speed(iter/s)": 0.447856 + }, + { + "acc": 0.66794319, + "epoch": 0.8915525114155252, + "grad_norm": 4.78125, + "learning_rate": 6.295006805879149e-06, + "loss": 1.58099632, + "memory(GiB)": 111.15, + "step": 35145, + "train_speed(iter/s)": 0.447889 + }, + { + "acc": 0.64470253, + "epoch": 0.8916793505834602, + "grad_norm": 5.3125, + "learning_rate": 6.293993935958613e-06, + "loss": 1.58829861, + "memory(GiB)": 111.15, + "step": 35150, + "train_speed(iter/s)": 0.447923 + }, + { + "acc": 0.65120764, + "epoch": 0.8918061897513953, + "grad_norm": 4.9375, + "learning_rate": 6.292981009122445e-06, + "loss": 1.6150774, + "memory(GiB)": 111.15, + "step": 35155, + "train_speed(iter/s)": 0.447957 + }, + { + "acc": 0.64268789, + "epoch": 0.8919330289193302, + "grad_norm": 6.0625, + "learning_rate": 6.291968025415202e-06, + "loss": 1.6526083, + "memory(GiB)": 111.15, + "step": 35160, + "train_speed(iter/s)": 0.447991 + }, + { + "acc": 0.65108261, + "epoch": 0.8920598680872653, + "grad_norm": 5.21875, + "learning_rate": 6.290954984881434e-06, + "loss": 1.62632427, + "memory(GiB)": 111.15, + "step": 35165, + "train_speed(iter/s)": 0.448025 + }, + { + "acc": 0.65372324, + "epoch": 0.8921867072552004, + "grad_norm": 5.875, + "learning_rate": 6.289941887565703e-06, + "loss": 1.65655174, + "memory(GiB)": 111.15, + "step": 35170, + "train_speed(iter/s)": 0.448058 + }, + { + "acc": 0.6686132, + "epoch": 0.8923135464231354, + "grad_norm": 5.625, + "learning_rate": 6.288928733512569e-06, + "loss": 1.57822742, + "memory(GiB)": 111.15, + "step": 35175, + "train_speed(iter/s)": 0.448092 + }, + { + "acc": 0.66168518, + "epoch": 0.8924403855910705, + "grad_norm": 5.53125, + "learning_rate": 6.287915522766596e-06, + "loss": 1.57102699, + "memory(GiB)": 111.15, + "step": 35180, + "train_speed(iter/s)": 0.448125 + }, + { + "acc": 0.64310112, + "epoch": 0.8925672247590056, + "grad_norm": 5.65625, + "learning_rate": 6.2869022553723465e-06, + "loss": 1.61847668, + "memory(GiB)": 111.15, + "step": 35185, + "train_speed(iter/s)": 0.448159 + }, + { + "acc": 0.66190262, + "epoch": 0.8926940639269406, + "grad_norm": 5.0, + "learning_rate": 6.285888931374391e-06, + "loss": 1.58692093, + "memory(GiB)": 111.15, + "step": 35190, + "train_speed(iter/s)": 0.448193 + }, + { + "acc": 0.64556694, + "epoch": 0.8928209030948757, + "grad_norm": 5.1875, + "learning_rate": 6.284875550817299e-06, + "loss": 1.59955931, + "memory(GiB)": 111.15, + "step": 35195, + "train_speed(iter/s)": 0.448227 + }, + { + "acc": 0.64750071, + "epoch": 0.8929477422628107, + "grad_norm": 5.03125, + "learning_rate": 6.2838621137456425e-06, + "loss": 1.62667313, + "memory(GiB)": 111.15, + "step": 35200, + "train_speed(iter/s)": 0.448261 + }, + { + "epoch": 0.8929477422628107, + "eval_acc": 0.6431277068710607, + "eval_loss": 1.60905122756958, + "eval_runtime": 113.3999, + "eval_samples_per_second": 56.173, + "eval_steps_per_second": 28.086, + "step": 35200 + }, + { + "acc": 0.64743481, + "epoch": 0.8930745814307458, + "grad_norm": 4.28125, + "learning_rate": 6.282848620203999e-06, + "loss": 1.60228386, + "memory(GiB)": 111.15, + "step": 35205, + "train_speed(iter/s)": 0.447622 + }, + { + "acc": 0.65784287, + "epoch": 0.8932014205986809, + "grad_norm": 5.1875, + "learning_rate": 6.2818350702369466e-06, + "loss": 1.62720032, + "memory(GiB)": 111.15, + "step": 35210, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.65014086, + "epoch": 0.8933282597666159, + "grad_norm": 6.59375, + "learning_rate": 6.280821463889063e-06, + "loss": 1.65376663, + "memory(GiB)": 111.15, + "step": 35215, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.66236334, + "epoch": 0.893455098934551, + "grad_norm": 5.28125, + "learning_rate": 6.279807801204936e-06, + "loss": 1.60311623, + "memory(GiB)": 111.15, + "step": 35220, + "train_speed(iter/s)": 0.447723 + }, + { + "acc": 0.66626415, + "epoch": 0.8935819381024861, + "grad_norm": 5.15625, + "learning_rate": 6.278794082229145e-06, + "loss": 1.61392879, + "memory(GiB)": 111.15, + "step": 35225, + "train_speed(iter/s)": 0.447756 + }, + { + "acc": 0.65733943, + "epoch": 0.8937087772704211, + "grad_norm": 5.625, + "learning_rate": 6.2777803070062825e-06, + "loss": 1.64342575, + "memory(GiB)": 111.15, + "step": 35230, + "train_speed(iter/s)": 0.447789 + }, + { + "acc": 0.65224304, + "epoch": 0.8938356164383562, + "grad_norm": 5.1875, + "learning_rate": 6.276766475580935e-06, + "loss": 1.59731293, + "memory(GiB)": 111.15, + "step": 35235, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.63356943, + "epoch": 0.8939624556062912, + "grad_norm": 5.34375, + "learning_rate": 6.2757525879977e-06, + "loss": 1.71369019, + "memory(GiB)": 111.15, + "step": 35240, + "train_speed(iter/s)": 0.447856 + }, + { + "acc": 0.64796629, + "epoch": 0.8940892947742263, + "grad_norm": 5.9375, + "learning_rate": 6.27473864430117e-06, + "loss": 1.67707577, + "memory(GiB)": 111.15, + "step": 35245, + "train_speed(iter/s)": 0.44789 + }, + { + "acc": 0.63016853, + "epoch": 0.8942161339421614, + "grad_norm": 5.625, + "learning_rate": 6.273724644535942e-06, + "loss": 1.66127205, + "memory(GiB)": 111.15, + "step": 35250, + "train_speed(iter/s)": 0.447924 + }, + { + "acc": 0.64981527, + "epoch": 0.8943429731100964, + "grad_norm": 4.8125, + "learning_rate": 6.272710588746619e-06, + "loss": 1.58839979, + "memory(GiB)": 111.15, + "step": 35255, + "train_speed(iter/s)": 0.447958 + }, + { + "acc": 0.64008727, + "epoch": 0.8944698122780315, + "grad_norm": 5.03125, + "learning_rate": 6.271696476977801e-06, + "loss": 1.67277279, + "memory(GiB)": 111.15, + "step": 35260, + "train_speed(iter/s)": 0.447992 + }, + { + "acc": 0.64659457, + "epoch": 0.8945966514459666, + "grad_norm": 5.09375, + "learning_rate": 6.270682309274094e-06, + "loss": 1.60620575, + "memory(GiB)": 111.15, + "step": 35265, + "train_speed(iter/s)": 0.448026 + }, + { + "acc": 0.65289087, + "epoch": 0.8947234906139016, + "grad_norm": 5.0625, + "learning_rate": 6.269668085680106e-06, + "loss": 1.6108654, + "memory(GiB)": 111.15, + "step": 35270, + "train_speed(iter/s)": 0.44806 + }, + { + "acc": 0.6387578, + "epoch": 0.8948503297818367, + "grad_norm": 6.4375, + "learning_rate": 6.268653806240448e-06, + "loss": 1.67899342, + "memory(GiB)": 111.15, + "step": 35275, + "train_speed(iter/s)": 0.448094 + }, + { + "acc": 0.65749226, + "epoch": 0.8949771689497716, + "grad_norm": 5.09375, + "learning_rate": 6.26763947099973e-06, + "loss": 1.61092415, + "memory(GiB)": 111.15, + "step": 35280, + "train_speed(iter/s)": 0.448128 + }, + { + "acc": 0.63997116, + "epoch": 0.8951040081177067, + "grad_norm": 5.6875, + "learning_rate": 6.266625080002569e-06, + "loss": 1.73264046, + "memory(GiB)": 111.15, + "step": 35285, + "train_speed(iter/s)": 0.448162 + }, + { + "acc": 0.64351435, + "epoch": 0.8952308472856418, + "grad_norm": 4.90625, + "learning_rate": 6.265610633293582e-06, + "loss": 1.63274193, + "memory(GiB)": 111.15, + "step": 35290, + "train_speed(iter/s)": 0.448195 + }, + { + "acc": 0.64347711, + "epoch": 0.8953576864535768, + "grad_norm": 6.0625, + "learning_rate": 6.264596130917389e-06, + "loss": 1.58518639, + "memory(GiB)": 111.15, + "step": 35295, + "train_speed(iter/s)": 0.448229 + }, + { + "acc": 0.66629229, + "epoch": 0.8954845256215119, + "grad_norm": 5.96875, + "learning_rate": 6.2635815729186124e-06, + "loss": 1.51253223, + "memory(GiB)": 111.15, + "step": 35300, + "train_speed(iter/s)": 0.448263 + }, + { + "epoch": 0.8954845256215119, + "eval_acc": 0.6431180982229432, + "eval_loss": 1.609094500541687, + "eval_runtime": 113.8192, + "eval_samples_per_second": 55.966, + "eval_steps_per_second": 27.983, + "step": 35300 + }, + { + "acc": 0.64582, + "epoch": 0.895611364789447, + "grad_norm": 6.0625, + "learning_rate": 6.2625669593418744e-06, + "loss": 1.61476517, + "memory(GiB)": 111.15, + "step": 35305, + "train_speed(iter/s)": 0.447625 + }, + { + "acc": 0.6387918, + "epoch": 0.895738203957382, + "grad_norm": 6.625, + "learning_rate": 6.261552290231807e-06, + "loss": 1.63134499, + "memory(GiB)": 111.15, + "step": 35310, + "train_speed(iter/s)": 0.447658 + }, + { + "acc": 0.65804834, + "epoch": 0.8958650431253171, + "grad_norm": 4.90625, + "learning_rate": 6.260537565633037e-06, + "loss": 1.56066971, + "memory(GiB)": 111.15, + "step": 35315, + "train_speed(iter/s)": 0.447692 + }, + { + "acc": 0.65116673, + "epoch": 0.8959918822932521, + "grad_norm": 5.15625, + "learning_rate": 6.259522785590197e-06, + "loss": 1.62811966, + "memory(GiB)": 111.15, + "step": 35320, + "train_speed(iter/s)": 0.447726 + }, + { + "acc": 0.63588562, + "epoch": 0.8961187214611872, + "grad_norm": 6.0625, + "learning_rate": 6.2585079501479205e-06, + "loss": 1.6877697, + "memory(GiB)": 111.15, + "step": 35325, + "train_speed(iter/s)": 0.447761 + }, + { + "acc": 0.64812555, + "epoch": 0.8962455606291223, + "grad_norm": 5.375, + "learning_rate": 6.257493059350848e-06, + "loss": 1.57535, + "memory(GiB)": 111.15, + "step": 35330, + "train_speed(iter/s)": 0.447795 + }, + { + "acc": 0.65283685, + "epoch": 0.8963723997970573, + "grad_norm": 4.5, + "learning_rate": 6.256478113243613e-06, + "loss": 1.65464935, + "memory(GiB)": 111.15, + "step": 35335, + "train_speed(iter/s)": 0.447829 + }, + { + "acc": 0.64955029, + "epoch": 0.8964992389649924, + "grad_norm": 5.0625, + "learning_rate": 6.255463111870864e-06, + "loss": 1.62873287, + "memory(GiB)": 111.15, + "step": 35340, + "train_speed(iter/s)": 0.447863 + }, + { + "acc": 0.64963188, + "epoch": 0.8966260781329275, + "grad_norm": 5.28125, + "learning_rate": 6.25444805527724e-06, + "loss": 1.66075191, + "memory(GiB)": 111.15, + "step": 35345, + "train_speed(iter/s)": 0.447897 + }, + { + "acc": 0.64682283, + "epoch": 0.8967529173008625, + "grad_norm": 5.90625, + "learning_rate": 6.253432943507391e-06, + "loss": 1.69245644, + "memory(GiB)": 111.15, + "step": 35350, + "train_speed(iter/s)": 0.447931 + }, + { + "acc": 0.65439024, + "epoch": 0.8968797564687976, + "grad_norm": 5.96875, + "learning_rate": 6.252417776605964e-06, + "loss": 1.60527534, + "memory(GiB)": 111.15, + "step": 35355, + "train_speed(iter/s)": 0.447965 + }, + { + "acc": 0.6608459, + "epoch": 0.8970065956367326, + "grad_norm": 5.25, + "learning_rate": 6.251402554617613e-06, + "loss": 1.54255772, + "memory(GiB)": 111.15, + "step": 35360, + "train_speed(iter/s)": 0.448 + }, + { + "acc": 0.62607093, + "epoch": 0.8971334348046677, + "grad_norm": 7.28125, + "learning_rate": 6.2503872775869886e-06, + "loss": 1.7455513, + "memory(GiB)": 111.15, + "step": 35365, + "train_speed(iter/s)": 0.448034 + }, + { + "acc": 0.66820369, + "epoch": 0.8972602739726028, + "grad_norm": 5.3125, + "learning_rate": 6.249371945558751e-06, + "loss": 1.58529482, + "memory(GiB)": 111.15, + "step": 35370, + "train_speed(iter/s)": 0.448068 + }, + { + "acc": 0.65850353, + "epoch": 0.8973871131405378, + "grad_norm": 5.21875, + "learning_rate": 6.248356558577555e-06, + "loss": 1.59365473, + "memory(GiB)": 111.15, + "step": 35375, + "train_speed(iter/s)": 0.448102 + }, + { + "acc": 0.6594625, + "epoch": 0.8975139523084729, + "grad_norm": 7.09375, + "learning_rate": 6.247341116688067e-06, + "loss": 1.64302788, + "memory(GiB)": 111.15, + "step": 35380, + "train_speed(iter/s)": 0.448136 + }, + { + "acc": 0.65186806, + "epoch": 0.897640791476408, + "grad_norm": 5.125, + "learning_rate": 6.246325619934945e-06, + "loss": 1.5920516, + "memory(GiB)": 111.15, + "step": 35385, + "train_speed(iter/s)": 0.448171 + }, + { + "acc": 0.65177813, + "epoch": 0.897767630644343, + "grad_norm": 6.34375, + "learning_rate": 6.245310068362859e-06, + "loss": 1.61536999, + "memory(GiB)": 111.15, + "step": 35390, + "train_speed(iter/s)": 0.448205 + }, + { + "acc": 0.64882827, + "epoch": 0.897894469812278, + "grad_norm": 5.3125, + "learning_rate": 6.244294462016476e-06, + "loss": 1.66434879, + "memory(GiB)": 111.15, + "step": 35395, + "train_speed(iter/s)": 0.448239 + }, + { + "acc": 0.66944933, + "epoch": 0.898021308980213, + "grad_norm": 6.71875, + "learning_rate": 6.243278800940468e-06, + "loss": 1.55462313, + "memory(GiB)": 111.15, + "step": 35400, + "train_speed(iter/s)": 0.448273 + }, + { + "epoch": 0.898021308980213, + "eval_acc": 0.643109742876754, + "eval_loss": 1.609106421470642, + "eval_runtime": 113.9542, + "eval_samples_per_second": 55.9, + "eval_steps_per_second": 27.95, + "step": 35400 + }, + { + "acc": 0.65264606, + "epoch": 0.8981481481481481, + "grad_norm": 6.09375, + "learning_rate": 6.242263085179506e-06, + "loss": 1.6316534, + "memory(GiB)": 111.15, + "step": 35405, + "train_speed(iter/s)": 0.447636 + }, + { + "acc": 0.6496666, + "epoch": 0.8982749873160832, + "grad_norm": 4.5625, + "learning_rate": 6.241247314778269e-06, + "loss": 1.62656364, + "memory(GiB)": 111.15, + "step": 35410, + "train_speed(iter/s)": 0.44767 + }, + { + "acc": 0.64449596, + "epoch": 0.8984018264840182, + "grad_norm": 4.71875, + "learning_rate": 6.240231489781432e-06, + "loss": 1.67344418, + "memory(GiB)": 111.15, + "step": 35415, + "train_speed(iter/s)": 0.447704 + }, + { + "acc": 0.64022503, + "epoch": 0.8985286656519533, + "grad_norm": 5.65625, + "learning_rate": 6.239215610233678e-06, + "loss": 1.62981796, + "memory(GiB)": 111.15, + "step": 35420, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.65906234, + "epoch": 0.8986555048198884, + "grad_norm": 5.90625, + "learning_rate": 6.238199676179688e-06, + "loss": 1.61109581, + "memory(GiB)": 111.15, + "step": 35425, + "train_speed(iter/s)": 0.447772 + }, + { + "acc": 0.6585978, + "epoch": 0.8987823439878234, + "grad_norm": 5.625, + "learning_rate": 6.2371836876641475e-06, + "loss": 1.598283, + "memory(GiB)": 111.15, + "step": 35430, + "train_speed(iter/s)": 0.447807 + }, + { + "acc": 0.64404993, + "epoch": 0.8989091831557585, + "grad_norm": 5.8125, + "learning_rate": 6.236167644731745e-06, + "loss": 1.63840809, + "memory(GiB)": 111.15, + "step": 35435, + "train_speed(iter/s)": 0.447841 + }, + { + "acc": 0.6568943, + "epoch": 0.8990360223236935, + "grad_norm": 5.3125, + "learning_rate": 6.235151547427172e-06, + "loss": 1.60268364, + "memory(GiB)": 111.15, + "step": 35440, + "train_speed(iter/s)": 0.447875 + }, + { + "acc": 0.65012569, + "epoch": 0.8991628614916286, + "grad_norm": 5.0, + "learning_rate": 6.2341353957951165e-06, + "loss": 1.63024292, + "memory(GiB)": 111.15, + "step": 35445, + "train_speed(iter/s)": 0.447909 + }, + { + "acc": 0.63096809, + "epoch": 0.8992897006595637, + "grad_norm": 6.34375, + "learning_rate": 6.233119189880279e-06, + "loss": 1.66348248, + "memory(GiB)": 111.15, + "step": 35450, + "train_speed(iter/s)": 0.447943 + }, + { + "acc": 0.64097195, + "epoch": 0.8994165398274987, + "grad_norm": 5.0625, + "learning_rate": 6.232102929727353e-06, + "loss": 1.65378685, + "memory(GiB)": 111.15, + "step": 35455, + "train_speed(iter/s)": 0.447977 + }, + { + "acc": 0.64256897, + "epoch": 0.8995433789954338, + "grad_norm": 4.65625, + "learning_rate": 6.231086615381039e-06, + "loss": 1.70467968, + "memory(GiB)": 111.15, + "step": 35460, + "train_speed(iter/s)": 0.448011 + }, + { + "acc": 0.65709534, + "epoch": 0.8996702181633689, + "grad_norm": 6.0, + "learning_rate": 6.2300702468860385e-06, + "loss": 1.62775269, + "memory(GiB)": 111.15, + "step": 35465, + "train_speed(iter/s)": 0.448045 + }, + { + "acc": 0.6559289, + "epoch": 0.8997970573313039, + "grad_norm": 5.46875, + "learning_rate": 6.229053824287058e-06, + "loss": 1.56832676, + "memory(GiB)": 111.15, + "step": 35470, + "train_speed(iter/s)": 0.448079 + }, + { + "acc": 0.64319549, + "epoch": 0.899923896499239, + "grad_norm": 5.90625, + "learning_rate": 6.228037347628803e-06, + "loss": 1.66057606, + "memory(GiB)": 111.15, + "step": 35475, + "train_speed(iter/s)": 0.448113 + }, + { + "acc": 0.66973047, + "epoch": 0.900050735667174, + "grad_norm": 4.59375, + "learning_rate": 6.227020816955982e-06, + "loss": 1.56371164, + "memory(GiB)": 111.15, + "step": 35480, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.65730929, + "epoch": 0.9001775748351091, + "grad_norm": 4.875, + "learning_rate": 6.226004232313308e-06, + "loss": 1.6132555, + "memory(GiB)": 111.15, + "step": 35485, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.63706207, + "epoch": 0.9003044140030442, + "grad_norm": 5.09375, + "learning_rate": 6.224987593745493e-06, + "loss": 1.63383255, + "memory(GiB)": 111.15, + "step": 35490, + "train_speed(iter/s)": 0.448215 + }, + { + "acc": 0.63092566, + "epoch": 0.9004312531709792, + "grad_norm": 5.9375, + "learning_rate": 6.223970901297255e-06, + "loss": 1.62231483, + "memory(GiB)": 111.15, + "step": 35495, + "train_speed(iter/s)": 0.448249 + }, + { + "acc": 0.65571547, + "epoch": 0.9005580923389143, + "grad_norm": 6.25, + "learning_rate": 6.222954155013312e-06, + "loss": 1.63363686, + "memory(GiB)": 111.15, + "step": 35500, + "train_speed(iter/s)": 0.448283 + }, + { + "epoch": 0.9005580923389143, + "eval_acc": 0.6431427464942012, + "eval_loss": 1.608956217765808, + "eval_runtime": 114.0513, + "eval_samples_per_second": 55.852, + "eval_steps_per_second": 27.926, + "step": 35500 + }, + { + "acc": 0.66399984, + "epoch": 0.9006849315068494, + "grad_norm": 5.5, + "learning_rate": 6.221937354938386e-06, + "loss": 1.59637842, + "memory(GiB)": 111.15, + "step": 35505, + "train_speed(iter/s)": 0.447646 + }, + { + "acc": 0.65598078, + "epoch": 0.9008117706747844, + "grad_norm": 4.96875, + "learning_rate": 6.2209205011171995e-06, + "loss": 1.64317551, + "memory(GiB)": 111.15, + "step": 35510, + "train_speed(iter/s)": 0.44768 + }, + { + "acc": 0.62706747, + "epoch": 0.9009386098427195, + "grad_norm": 5.3125, + "learning_rate": 6.219903593594476e-06, + "loss": 1.6574379, + "memory(GiB)": 111.15, + "step": 35515, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.6614994, + "epoch": 0.9010654490106544, + "grad_norm": 6.125, + "learning_rate": 6.218886632414949e-06, + "loss": 1.56817207, + "memory(GiB)": 111.15, + "step": 35520, + "train_speed(iter/s)": 0.447747 + }, + { + "acc": 0.6483047, + "epoch": 0.9011922881785895, + "grad_norm": 5.125, + "learning_rate": 6.217869617623343e-06, + "loss": 1.6236393, + "memory(GiB)": 111.15, + "step": 35525, + "train_speed(iter/s)": 0.447781 + }, + { + "acc": 0.65977879, + "epoch": 0.9013191273465246, + "grad_norm": 5.375, + "learning_rate": 6.216852549264396e-06, + "loss": 1.60086594, + "memory(GiB)": 111.15, + "step": 35530, + "train_speed(iter/s)": 0.447815 + }, + { + "acc": 0.6623106, + "epoch": 0.9014459665144596, + "grad_norm": 5.0, + "learning_rate": 6.215835427382842e-06, + "loss": 1.56604214, + "memory(GiB)": 111.15, + "step": 35535, + "train_speed(iter/s)": 0.447849 + }, + { + "acc": 0.65063453, + "epoch": 0.9015728056823947, + "grad_norm": 5.28125, + "learning_rate": 6.214818252023415e-06, + "loss": 1.64742126, + "memory(GiB)": 111.15, + "step": 35540, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.65589414, + "epoch": 0.9016996448503298, + "grad_norm": 4.71875, + "learning_rate": 6.2138010232308585e-06, + "loss": 1.54843388, + "memory(GiB)": 111.15, + "step": 35545, + "train_speed(iter/s)": 0.447917 + }, + { + "acc": 0.66589632, + "epoch": 0.9018264840182648, + "grad_norm": 5.21875, + "learning_rate": 6.212783741049915e-06, + "loss": 1.60103226, + "memory(GiB)": 111.15, + "step": 35550, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.65041027, + "epoch": 0.9019533231861999, + "grad_norm": 4.375, + "learning_rate": 6.211766405525326e-06, + "loss": 1.57349701, + "memory(GiB)": 111.15, + "step": 35555, + "train_speed(iter/s)": 0.447985 + }, + { + "acc": 0.64116735, + "epoch": 0.9020801623541349, + "grad_norm": 6.03125, + "learning_rate": 6.210749016701842e-06, + "loss": 1.69036179, + "memory(GiB)": 111.15, + "step": 35560, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.63586974, + "epoch": 0.90220700152207, + "grad_norm": 6.125, + "learning_rate": 6.2097315746242095e-06, + "loss": 1.66454926, + "memory(GiB)": 111.15, + "step": 35565, + "train_speed(iter/s)": 0.448052 + }, + { + "acc": 0.62810411, + "epoch": 0.9023338406900051, + "grad_norm": 5.1875, + "learning_rate": 6.208714079337181e-06, + "loss": 1.72403717, + "memory(GiB)": 111.15, + "step": 35570, + "train_speed(iter/s)": 0.448086 + }, + { + "acc": 0.64109163, + "epoch": 0.9024606798579401, + "grad_norm": 4.9375, + "learning_rate": 6.207696530885511e-06, + "loss": 1.64453506, + "memory(GiB)": 111.15, + "step": 35575, + "train_speed(iter/s)": 0.44812 + }, + { + "acc": 0.64811096, + "epoch": 0.9025875190258752, + "grad_norm": 6.09375, + "learning_rate": 6.2066789293139565e-06, + "loss": 1.62164021, + "memory(GiB)": 111.15, + "step": 35580, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.62733355, + "epoch": 0.9027143581938103, + "grad_norm": 5.34375, + "learning_rate": 6.2056612746672736e-06, + "loss": 1.69705219, + "memory(GiB)": 111.15, + "step": 35585, + "train_speed(iter/s)": 0.448188 + }, + { + "acc": 0.65397978, + "epoch": 0.9028411973617453, + "grad_norm": 5.125, + "learning_rate": 6.204643566990227e-06, + "loss": 1.6456316, + "memory(GiB)": 111.15, + "step": 35590, + "train_speed(iter/s)": 0.448222 + }, + { + "acc": 0.64921937, + "epoch": 0.9029680365296804, + "grad_norm": 5.28125, + "learning_rate": 6.2036258063275764e-06, + "loss": 1.59275045, + "memory(GiB)": 111.15, + "step": 35595, + "train_speed(iter/s)": 0.448256 + }, + { + "acc": 0.65227051, + "epoch": 0.9030948756976154, + "grad_norm": 6.28125, + "learning_rate": 6.20260799272409e-06, + "loss": 1.63976135, + "memory(GiB)": 111.15, + "step": 35600, + "train_speed(iter/s)": 0.44829 + }, + { + "epoch": 0.9030948756976154, + "eval_acc": 0.6431343911480121, + "eval_loss": 1.6089954376220703, + "eval_runtime": 113.8019, + "eval_samples_per_second": 55.974, + "eval_steps_per_second": 27.987, + "step": 35600 + }, + { + "acc": 0.65952778, + "epoch": 0.9032217148655505, + "grad_norm": 4.65625, + "learning_rate": 6.201590126224534e-06, + "loss": 1.63060684, + "memory(GiB)": 111.15, + "step": 35605, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.66239605, + "epoch": 0.9033485540334856, + "grad_norm": 4.625, + "learning_rate": 6.20057220687368e-06, + "loss": 1.54678383, + "memory(GiB)": 111.15, + "step": 35610, + "train_speed(iter/s)": 0.44769 + }, + { + "acc": 0.65081806, + "epoch": 0.9034753932014206, + "grad_norm": 4.875, + "learning_rate": 6.199554234716301e-06, + "loss": 1.6197876, + "memory(GiB)": 111.15, + "step": 35615, + "train_speed(iter/s)": 0.447724 + }, + { + "acc": 0.64635363, + "epoch": 0.9036022323693557, + "grad_norm": 5.75, + "learning_rate": 6.19853620979717e-06, + "loss": 1.58390064, + "memory(GiB)": 111.15, + "step": 35620, + "train_speed(iter/s)": 0.447758 + }, + { + "acc": 0.65079012, + "epoch": 0.9037290715372908, + "grad_norm": 5.21875, + "learning_rate": 6.1975181321610655e-06, + "loss": 1.6052002, + "memory(GiB)": 111.15, + "step": 35625, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.64330292, + "epoch": 0.9038559107052258, + "grad_norm": 5.8125, + "learning_rate": 6.1965000018527676e-06, + "loss": 1.68066769, + "memory(GiB)": 111.15, + "step": 35630, + "train_speed(iter/s)": 0.447825 + }, + { + "acc": 0.63557425, + "epoch": 0.9039827498731609, + "grad_norm": 5.125, + "learning_rate": 6.195481818917057e-06, + "loss": 1.65652351, + "memory(GiB)": 111.15, + "step": 35635, + "train_speed(iter/s)": 0.447859 + }, + { + "acc": 0.64512863, + "epoch": 0.9041095890410958, + "grad_norm": 5.53125, + "learning_rate": 6.194463583398719e-06, + "loss": 1.65803299, + "memory(GiB)": 111.15, + "step": 35640, + "train_speed(iter/s)": 0.447893 + }, + { + "acc": 0.6425571, + "epoch": 0.9042364282090309, + "grad_norm": 5.625, + "learning_rate": 6.193445295342538e-06, + "loss": 1.61785259, + "memory(GiB)": 111.15, + "step": 35645, + "train_speed(iter/s)": 0.447927 + }, + { + "acc": 0.62941484, + "epoch": 0.904363267376966, + "grad_norm": 4.71875, + "learning_rate": 6.192426954793308e-06, + "loss": 1.65591526, + "memory(GiB)": 111.15, + "step": 35650, + "train_speed(iter/s)": 0.447961 + }, + { + "acc": 0.64823637, + "epoch": 0.904490106544901, + "grad_norm": 5.3125, + "learning_rate": 6.1914085617958135e-06, + "loss": 1.63498383, + "memory(GiB)": 111.15, + "step": 35655, + "train_speed(iter/s)": 0.447995 + }, + { + "acc": 0.65943398, + "epoch": 0.9046169457128361, + "grad_norm": 5.0, + "learning_rate": 6.190390116394853e-06, + "loss": 1.6169138, + "memory(GiB)": 111.15, + "step": 35660, + "train_speed(iter/s)": 0.448028 + }, + { + "acc": 0.63180733, + "epoch": 0.9047437848807712, + "grad_norm": 6.34375, + "learning_rate": 6.189371618635219e-06, + "loss": 1.64459686, + "memory(GiB)": 111.15, + "step": 35665, + "train_speed(iter/s)": 0.448062 + }, + { + "acc": 0.63583593, + "epoch": 0.9048706240487062, + "grad_norm": 4.84375, + "learning_rate": 6.188353068561714e-06, + "loss": 1.6211338, + "memory(GiB)": 111.15, + "step": 35670, + "train_speed(iter/s)": 0.448096 + }, + { + "acc": 0.65103974, + "epoch": 0.9049974632166413, + "grad_norm": 5.71875, + "learning_rate": 6.187334466219133e-06, + "loss": 1.58105707, + "memory(GiB)": 111.15, + "step": 35675, + "train_speed(iter/s)": 0.44813 + }, + { + "acc": 0.64702568, + "epoch": 0.9051243023845763, + "grad_norm": 5.53125, + "learning_rate": 6.18631581165228e-06, + "loss": 1.58143682, + "memory(GiB)": 111.15, + "step": 35680, + "train_speed(iter/s)": 0.448164 + }, + { + "acc": 0.63582287, + "epoch": 0.9052511415525114, + "grad_norm": 5.78125, + "learning_rate": 6.185297104905963e-06, + "loss": 1.65934448, + "memory(GiB)": 111.15, + "step": 35685, + "train_speed(iter/s)": 0.448198 + }, + { + "acc": 0.6403285, + "epoch": 0.9053779807204465, + "grad_norm": 5.625, + "learning_rate": 6.184278346024988e-06, + "loss": 1.65623627, + "memory(GiB)": 111.15, + "step": 35690, + "train_speed(iter/s)": 0.448232 + }, + { + "acc": 0.64617605, + "epoch": 0.9055048198883815, + "grad_norm": 4.75, + "learning_rate": 6.183259535054163e-06, + "loss": 1.58172112, + "memory(GiB)": 111.15, + "step": 35695, + "train_speed(iter/s)": 0.448266 + }, + { + "acc": 0.63220024, + "epoch": 0.9056316590563166, + "grad_norm": 5.0, + "learning_rate": 6.1822406720383e-06, + "loss": 1.643046, + "memory(GiB)": 111.15, + "step": 35700, + "train_speed(iter/s)": 0.448299 + }, + { + "epoch": 0.9056316590563166, + "eval_acc": 0.6431030585998028, + "eval_loss": 1.6090006828308105, + "eval_runtime": 114.6874, + "eval_samples_per_second": 55.542, + "eval_steps_per_second": 27.771, + "step": 35700 + }, + { + "acc": 0.65170927, + "epoch": 0.9057584982242517, + "grad_norm": 4.5625, + "learning_rate": 6.181221757022215e-06, + "loss": 1.64829865, + "memory(GiB)": 111.15, + "step": 35705, + "train_speed(iter/s)": 0.447662 + }, + { + "acc": 0.64411225, + "epoch": 0.9058853373921867, + "grad_norm": 6.0625, + "learning_rate": 6.180202790050724e-06, + "loss": 1.64063587, + "memory(GiB)": 111.15, + "step": 35710, + "train_speed(iter/s)": 0.447695 + }, + { + "acc": 0.65052633, + "epoch": 0.9060121765601218, + "grad_norm": 5.0625, + "learning_rate": 6.179183771168643e-06, + "loss": 1.63856068, + "memory(GiB)": 111.15, + "step": 35715, + "train_speed(iter/s)": 0.447728 + }, + { + "acc": 0.64454508, + "epoch": 0.9061390157280568, + "grad_norm": 5.40625, + "learning_rate": 6.1781647004207965e-06, + "loss": 1.65536461, + "memory(GiB)": 111.15, + "step": 35720, + "train_speed(iter/s)": 0.447761 + }, + { + "acc": 0.63282146, + "epoch": 0.9062658548959919, + "grad_norm": 5.75, + "learning_rate": 6.177145577852005e-06, + "loss": 1.7028265, + "memory(GiB)": 111.15, + "step": 35725, + "train_speed(iter/s)": 0.447794 + }, + { + "acc": 0.63913956, + "epoch": 0.906392694063927, + "grad_norm": 5.71875, + "learning_rate": 6.176126403507097e-06, + "loss": 1.67081757, + "memory(GiB)": 111.15, + "step": 35730, + "train_speed(iter/s)": 0.447827 + }, + { + "acc": 0.65631618, + "epoch": 0.906519533231862, + "grad_norm": 5.21875, + "learning_rate": 6.175107177430897e-06, + "loss": 1.57355309, + "memory(GiB)": 111.15, + "step": 35735, + "train_speed(iter/s)": 0.447861 + }, + { + "acc": 0.64298353, + "epoch": 0.9066463723997971, + "grad_norm": 5.625, + "learning_rate": 6.17408789966824e-06, + "loss": 1.61905289, + "memory(GiB)": 111.15, + "step": 35740, + "train_speed(iter/s)": 0.447894 + }, + { + "acc": 0.64611435, + "epoch": 0.9067732115677322, + "grad_norm": 4.875, + "learning_rate": 6.173068570263951e-06, + "loss": 1.62951183, + "memory(GiB)": 111.15, + "step": 35745, + "train_speed(iter/s)": 0.447926 + }, + { + "acc": 0.64307547, + "epoch": 0.9069000507356672, + "grad_norm": 5.28125, + "learning_rate": 6.172049189262872e-06, + "loss": 1.62421532, + "memory(GiB)": 111.15, + "step": 35750, + "train_speed(iter/s)": 0.447959 + }, + { + "acc": 0.65865946, + "epoch": 0.9070268899036023, + "grad_norm": 4.96875, + "learning_rate": 6.1710297567098354e-06, + "loss": 1.6256216, + "memory(GiB)": 111.15, + "step": 35755, + "train_speed(iter/s)": 0.447992 + }, + { + "acc": 0.64590178, + "epoch": 0.9071537290715372, + "grad_norm": 6.3125, + "learning_rate": 6.170010272649682e-06, + "loss": 1.65918655, + "memory(GiB)": 111.15, + "step": 35760, + "train_speed(iter/s)": 0.448025 + }, + { + "acc": 0.63658719, + "epoch": 0.9072805682394723, + "grad_norm": 7.8125, + "learning_rate": 6.168990737127254e-06, + "loss": 1.66935215, + "memory(GiB)": 111.15, + "step": 35765, + "train_speed(iter/s)": 0.448058 + }, + { + "acc": 0.65297146, + "epoch": 0.9074074074074074, + "grad_norm": 5.34375, + "learning_rate": 6.167971150187394e-06, + "loss": 1.62342091, + "memory(GiB)": 111.15, + "step": 35770, + "train_speed(iter/s)": 0.448091 + }, + { + "acc": 0.63311672, + "epoch": 0.9075342465753424, + "grad_norm": 4.9375, + "learning_rate": 6.166951511874948e-06, + "loss": 1.68222427, + "memory(GiB)": 111.15, + "step": 35775, + "train_speed(iter/s)": 0.448124 + }, + { + "acc": 0.65563755, + "epoch": 0.9076610857432775, + "grad_norm": 6.4375, + "learning_rate": 6.165931822234764e-06, + "loss": 1.5967555, + "memory(GiB)": 111.15, + "step": 35780, + "train_speed(iter/s)": 0.448157 + }, + { + "acc": 0.65375676, + "epoch": 0.9077879249112126, + "grad_norm": 6.53125, + "learning_rate": 6.164912081311694e-06, + "loss": 1.68392525, + "memory(GiB)": 111.15, + "step": 35785, + "train_speed(iter/s)": 0.44819 + }, + { + "acc": 0.65368843, + "epoch": 0.9079147640791476, + "grad_norm": 5.40625, + "learning_rate": 6.163892289150588e-06, + "loss": 1.57026596, + "memory(GiB)": 111.15, + "step": 35790, + "train_speed(iter/s)": 0.448223 + }, + { + "acc": 0.6553782, + "epoch": 0.9080416032470827, + "grad_norm": 5.96875, + "learning_rate": 6.162872445796303e-06, + "loss": 1.59972782, + "memory(GiB)": 111.15, + "step": 35795, + "train_speed(iter/s)": 0.448256 + }, + { + "acc": 0.64188275, + "epoch": 0.9081684424150177, + "grad_norm": 4.75, + "learning_rate": 6.161852551293697e-06, + "loss": 1.67158165, + "memory(GiB)": 111.15, + "step": 35800, + "train_speed(iter/s)": 0.448289 + }, + { + "epoch": 0.9081684424150177, + "eval_acc": 0.6431786744828145, + "eval_loss": 1.6089307069778442, + "eval_runtime": 114.3112, + "eval_samples_per_second": 55.725, + "eval_steps_per_second": 27.863, + "step": 35800 + }, + { + "acc": 0.64902382, + "epoch": 0.9082952815829528, + "grad_norm": 6.0, + "learning_rate": 6.160832605687628e-06, + "loss": 1.65587311, + "memory(GiB)": 111.15, + "step": 35805, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.65198479, + "epoch": 0.9084221207508879, + "grad_norm": 5.4375, + "learning_rate": 6.159812609022961e-06, + "loss": 1.63417072, + "memory(GiB)": 111.15, + "step": 35810, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.64456992, + "epoch": 0.9085489599188229, + "grad_norm": 4.25, + "learning_rate": 6.158792561344553e-06, + "loss": 1.66733017, + "memory(GiB)": 111.15, + "step": 35815, + "train_speed(iter/s)": 0.447722 + }, + { + "acc": 0.64923744, + "epoch": 0.908675799086758, + "grad_norm": 6.5, + "learning_rate": 6.157772462697277e-06, + "loss": 1.63618431, + "memory(GiB)": 111.15, + "step": 35820, + "train_speed(iter/s)": 0.447754 + }, + { + "acc": 0.65504842, + "epoch": 0.9088026382546931, + "grad_norm": 5.3125, + "learning_rate": 6.156752313125998e-06, + "loss": 1.64682083, + "memory(GiB)": 111.15, + "step": 35825, + "train_speed(iter/s)": 0.447787 + }, + { + "acc": 0.65155468, + "epoch": 0.9089294774226281, + "grad_norm": 6.28125, + "learning_rate": 6.155732112675587e-06, + "loss": 1.64214897, + "memory(GiB)": 111.15, + "step": 35830, + "train_speed(iter/s)": 0.44782 + }, + { + "acc": 0.66278849, + "epoch": 0.9090563165905632, + "grad_norm": 5.5, + "learning_rate": 6.154711861390919e-06, + "loss": 1.54290895, + "memory(GiB)": 111.15, + "step": 35835, + "train_speed(iter/s)": 0.447853 + }, + { + "acc": 0.66010962, + "epoch": 0.9091831557584982, + "grad_norm": 4.65625, + "learning_rate": 6.153691559316868e-06, + "loss": 1.63034172, + "memory(GiB)": 111.15, + "step": 35840, + "train_speed(iter/s)": 0.447886 + }, + { + "acc": 0.63150392, + "epoch": 0.9093099949264333, + "grad_norm": 5.625, + "learning_rate": 6.152671206498311e-06, + "loss": 1.71667824, + "memory(GiB)": 111.15, + "step": 35845, + "train_speed(iter/s)": 0.447919 + }, + { + "acc": 0.66426897, + "epoch": 0.9094368340943684, + "grad_norm": 5.375, + "learning_rate": 6.151650802980128e-06, + "loss": 1.55636425, + "memory(GiB)": 111.15, + "step": 35850, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.64026899, + "epoch": 0.9095636732623034, + "grad_norm": 5.53125, + "learning_rate": 6.150630348807201e-06, + "loss": 1.71430054, + "memory(GiB)": 111.15, + "step": 35855, + "train_speed(iter/s)": 0.447985 + }, + { + "acc": 0.63367405, + "epoch": 0.9096905124302385, + "grad_norm": 6.9375, + "learning_rate": 6.149609844024413e-06, + "loss": 1.66832314, + "memory(GiB)": 111.15, + "step": 35860, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.64547706, + "epoch": 0.9098173515981736, + "grad_norm": 4.71875, + "learning_rate": 6.148589288676652e-06, + "loss": 1.71543808, + "memory(GiB)": 111.15, + "step": 35865, + "train_speed(iter/s)": 0.448052 + }, + { + "acc": 0.64994025, + "epoch": 0.9099441907661086, + "grad_norm": 6.46875, + "learning_rate": 6.147568682808808e-06, + "loss": 1.62821236, + "memory(GiB)": 111.15, + "step": 35870, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.65559435, + "epoch": 0.9100710299340437, + "grad_norm": 5.71875, + "learning_rate": 6.146548026465766e-06, + "loss": 1.67522812, + "memory(GiB)": 111.15, + "step": 35875, + "train_speed(iter/s)": 0.448118 + }, + { + "acc": 0.65121307, + "epoch": 0.9101978691019786, + "grad_norm": 5.75, + "learning_rate": 6.145527319692427e-06, + "loss": 1.55567303, + "memory(GiB)": 111.15, + "step": 35880, + "train_speed(iter/s)": 0.448151 + }, + { + "acc": 0.64801378, + "epoch": 0.9103247082699137, + "grad_norm": 5.46875, + "learning_rate": 6.144506562533678e-06, + "loss": 1.57960014, + "memory(GiB)": 111.15, + "step": 35885, + "train_speed(iter/s)": 0.448185 + }, + { + "acc": 0.64389768, + "epoch": 0.9104515474378488, + "grad_norm": 6.28125, + "learning_rate": 6.143485755034425e-06, + "loss": 1.61633358, + "memory(GiB)": 111.15, + "step": 35890, + "train_speed(iter/s)": 0.448218 + }, + { + "acc": 0.64562316, + "epoch": 0.9105783866057838, + "grad_norm": 5.75, + "learning_rate": 6.14246489723956e-06, + "loss": 1.60378799, + "memory(GiB)": 111.15, + "step": 35895, + "train_speed(iter/s)": 0.448251 + }, + { + "acc": 0.649155, + "epoch": 0.9107052257737189, + "grad_norm": 6.21875, + "learning_rate": 6.141443989193988e-06, + "loss": 1.63050461, + "memory(GiB)": 111.15, + "step": 35900, + "train_speed(iter/s)": 0.448284 + }, + { + "epoch": 0.9107052257737189, + "eval_acc": 0.6431377332864877, + "eval_loss": 1.6089552640914917, + "eval_runtime": 114.7915, + "eval_samples_per_second": 55.492, + "eval_steps_per_second": 27.746, + "step": 35900 + }, + { + "acc": 0.63304882, + "epoch": 0.910832064941654, + "grad_norm": 5.34375, + "learning_rate": 6.140423030942615e-06, + "loss": 1.7576128, + "memory(GiB)": 111.15, + "step": 35905, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.65792561, + "epoch": 0.910958904109589, + "grad_norm": 5.03125, + "learning_rate": 6.139402022530344e-06, + "loss": 1.63652077, + "memory(GiB)": 111.15, + "step": 35910, + "train_speed(iter/s)": 0.447682 + }, + { + "acc": 0.63991976, + "epoch": 0.9110857432775241, + "grad_norm": 5.53125, + "learning_rate": 6.138380964002087e-06, + "loss": 1.6403265, + "memory(GiB)": 111.15, + "step": 35915, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.67180948, + "epoch": 0.9112125824454591, + "grad_norm": 6.4375, + "learning_rate": 6.13735985540275e-06, + "loss": 1.55169067, + "memory(GiB)": 111.15, + "step": 35920, + "train_speed(iter/s)": 0.447747 + }, + { + "acc": 0.65612316, + "epoch": 0.9113394216133942, + "grad_norm": 6.09375, + "learning_rate": 6.13633869677725e-06, + "loss": 1.57194166, + "memory(GiB)": 111.15, + "step": 35925, + "train_speed(iter/s)": 0.44778 + }, + { + "acc": 0.64256191, + "epoch": 0.9114662607813293, + "grad_norm": 4.78125, + "learning_rate": 6.1353174881705e-06, + "loss": 1.63215294, + "memory(GiB)": 111.15, + "step": 35930, + "train_speed(iter/s)": 0.447812 + }, + { + "acc": 0.63702741, + "epoch": 0.9115930999492643, + "grad_norm": 5.40625, + "learning_rate": 6.134296229627419e-06, + "loss": 1.70845795, + "memory(GiB)": 111.15, + "step": 35935, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.65311604, + "epoch": 0.9117199391171994, + "grad_norm": 5.1875, + "learning_rate": 6.1332749211929255e-06, + "loss": 1.6214222, + "memory(GiB)": 111.15, + "step": 35940, + "train_speed(iter/s)": 0.447877 + }, + { + "acc": 0.66994019, + "epoch": 0.9118467782851345, + "grad_norm": 5.59375, + "learning_rate": 6.132253562911941e-06, + "loss": 1.57413368, + "memory(GiB)": 111.15, + "step": 35945, + "train_speed(iter/s)": 0.447909 + }, + { + "acc": 0.65686941, + "epoch": 0.9119736174530695, + "grad_norm": 5.5, + "learning_rate": 6.1312321548293895e-06, + "loss": 1.57552853, + "memory(GiB)": 111.15, + "step": 35950, + "train_speed(iter/s)": 0.447942 + }, + { + "acc": 0.64749374, + "epoch": 0.9121004566210046, + "grad_norm": 5.15625, + "learning_rate": 6.130210696990197e-06, + "loss": 1.69909554, + "memory(GiB)": 111.15, + "step": 35955, + "train_speed(iter/s)": 0.447975 + }, + { + "acc": 0.64754601, + "epoch": 0.9122272957889396, + "grad_norm": 6.09375, + "learning_rate": 6.129189189439293e-06, + "loss": 1.67980633, + "memory(GiB)": 111.15, + "step": 35960, + "train_speed(iter/s)": 0.448008 + }, + { + "acc": 0.64943447, + "epoch": 0.9123541349568747, + "grad_norm": 6.03125, + "learning_rate": 6.128167632221605e-06, + "loss": 1.67521133, + "memory(GiB)": 111.15, + "step": 35965, + "train_speed(iter/s)": 0.44804 + }, + { + "acc": 0.65470285, + "epoch": 0.9124809741248098, + "grad_norm": 5.03125, + "learning_rate": 6.127146025382069e-06, + "loss": 1.56109333, + "memory(GiB)": 111.15, + "step": 35970, + "train_speed(iter/s)": 0.448073 + }, + { + "acc": 0.64479866, + "epoch": 0.9126078132927448, + "grad_norm": 5.46875, + "learning_rate": 6.126124368965619e-06, + "loss": 1.67771759, + "memory(GiB)": 111.15, + "step": 35975, + "train_speed(iter/s)": 0.448106 + }, + { + "acc": 0.62697926, + "epoch": 0.9127346524606799, + "grad_norm": 5.3125, + "learning_rate": 6.125102663017191e-06, + "loss": 1.75383434, + "memory(GiB)": 111.15, + "step": 35980, + "train_speed(iter/s)": 0.448139 + }, + { + "acc": 0.6589592, + "epoch": 0.912861491628615, + "grad_norm": 4.9375, + "learning_rate": 6.124080907581724e-06, + "loss": 1.60033054, + "memory(GiB)": 111.15, + "step": 35985, + "train_speed(iter/s)": 0.448171 + }, + { + "acc": 0.65849056, + "epoch": 0.91298833079655, + "grad_norm": 4.96875, + "learning_rate": 6.1230591027041605e-06, + "loss": 1.61527157, + "memory(GiB)": 111.15, + "step": 35990, + "train_speed(iter/s)": 0.448204 + }, + { + "acc": 0.66186309, + "epoch": 0.913115169964485, + "grad_norm": 5.09375, + "learning_rate": 6.1220372484294444e-06, + "loss": 1.57512045, + "memory(GiB)": 111.15, + "step": 35995, + "train_speed(iter/s)": 0.448236 + }, + { + "acc": 0.64300108, + "epoch": 0.91324200913242, + "grad_norm": 5.71875, + "learning_rate": 6.12101534480252e-06, + "loss": 1.63646889, + "memory(GiB)": 111.15, + "step": 36000, + "train_speed(iter/s)": 0.448269 + }, + { + "epoch": 0.91324200913242, + "eval_acc": 0.6431732435077916, + "eval_loss": 1.6090021133422852, + "eval_runtime": 114.2704, + "eval_samples_per_second": 55.745, + "eval_steps_per_second": 27.872, + "step": 36000 + }, + { + "acc": 0.64808969, + "epoch": 0.9133688483003551, + "grad_norm": 5.875, + "learning_rate": 6.119993391868335e-06, + "loss": 1.61564636, + "memory(GiB)": 111.15, + "step": 36005, + "train_speed(iter/s)": 0.44764 + }, + { + "acc": 0.66715131, + "epoch": 0.9134956874682902, + "grad_norm": 5.03125, + "learning_rate": 6.118971389671842e-06, + "loss": 1.61282101, + "memory(GiB)": 111.15, + "step": 36010, + "train_speed(iter/s)": 0.447673 + }, + { + "acc": 0.64225612, + "epoch": 0.9136225266362252, + "grad_norm": 4.625, + "learning_rate": 6.117949338257989e-06, + "loss": 1.65369205, + "memory(GiB)": 111.15, + "step": 36015, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.6604702, + "epoch": 0.9137493658041603, + "grad_norm": 5.40625, + "learning_rate": 6.116927237671735e-06, + "loss": 1.58282585, + "memory(GiB)": 111.15, + "step": 36020, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.64422951, + "epoch": 0.9138762049720954, + "grad_norm": 6.46875, + "learning_rate": 6.115905087958032e-06, + "loss": 1.63165913, + "memory(GiB)": 111.15, + "step": 36025, + "train_speed(iter/s)": 0.447771 + }, + { + "acc": 0.64467354, + "epoch": 0.9140030441400304, + "grad_norm": 5.8125, + "learning_rate": 6.114882889161844e-06, + "loss": 1.66979599, + "memory(GiB)": 111.15, + "step": 36030, + "train_speed(iter/s)": 0.447804 + }, + { + "acc": 0.66051683, + "epoch": 0.9141298833079655, + "grad_norm": 5.4375, + "learning_rate": 6.113860641328127e-06, + "loss": 1.63457184, + "memory(GiB)": 111.15, + "step": 36035, + "train_speed(iter/s)": 0.447837 + }, + { + "acc": 0.65168877, + "epoch": 0.9142567224759005, + "grad_norm": 5.34375, + "learning_rate": 6.112838344501846e-06, + "loss": 1.57108326, + "memory(GiB)": 111.15, + "step": 36040, + "train_speed(iter/s)": 0.447869 + }, + { + "acc": 0.64377279, + "epoch": 0.9143835616438356, + "grad_norm": 5.4375, + "learning_rate": 6.111815998727966e-06, + "loss": 1.67367096, + "memory(GiB)": 111.15, + "step": 36045, + "train_speed(iter/s)": 0.447902 + }, + { + "acc": 0.65810981, + "epoch": 0.9145104008117707, + "grad_norm": 5.25, + "learning_rate": 6.110793604051455e-06, + "loss": 1.58925419, + "memory(GiB)": 111.15, + "step": 36050, + "train_speed(iter/s)": 0.447935 + }, + { + "acc": 0.66094351, + "epoch": 0.9146372399797057, + "grad_norm": 5.9375, + "learning_rate": 6.109771160517283e-06, + "loss": 1.63080444, + "memory(GiB)": 111.15, + "step": 36055, + "train_speed(iter/s)": 0.447968 + }, + { + "acc": 0.64185266, + "epoch": 0.9147640791476408, + "grad_norm": 6.1875, + "learning_rate": 6.108748668170419e-06, + "loss": 1.60242157, + "memory(GiB)": 111.15, + "step": 36060, + "train_speed(iter/s)": 0.448001 + }, + { + "acc": 0.65731583, + "epoch": 0.9148909183155759, + "grad_norm": 4.8125, + "learning_rate": 6.1077261270558385e-06, + "loss": 1.61961937, + "memory(GiB)": 111.15, + "step": 36065, + "train_speed(iter/s)": 0.448034 + }, + { + "acc": 0.65015144, + "epoch": 0.9150177574835109, + "grad_norm": 4.53125, + "learning_rate": 6.106703537218518e-06, + "loss": 1.6514782, + "memory(GiB)": 111.15, + "step": 36070, + "train_speed(iter/s)": 0.448067 + }, + { + "acc": 0.64364433, + "epoch": 0.915144596651446, + "grad_norm": 4.78125, + "learning_rate": 6.105680898703434e-06, + "loss": 1.66553078, + "memory(GiB)": 111.15, + "step": 36075, + "train_speed(iter/s)": 0.4481 + }, + { + "acc": 0.63928485, + "epoch": 0.915271435819381, + "grad_norm": 4.5625, + "learning_rate": 6.104658211555568e-06, + "loss": 1.70847988, + "memory(GiB)": 111.15, + "step": 36080, + "train_speed(iter/s)": 0.448132 + }, + { + "acc": 0.65058866, + "epoch": 0.9153982749873161, + "grad_norm": 6.09375, + "learning_rate": 6.103635475819902e-06, + "loss": 1.65806656, + "memory(GiB)": 111.15, + "step": 36085, + "train_speed(iter/s)": 0.448165 + }, + { + "acc": 0.64787178, + "epoch": 0.9155251141552512, + "grad_norm": 4.8125, + "learning_rate": 6.102612691541422e-06, + "loss": 1.62349339, + "memory(GiB)": 111.15, + "step": 36090, + "train_speed(iter/s)": 0.448199 + }, + { + "acc": 0.62040234, + "epoch": 0.9156519533231862, + "grad_norm": 5.21875, + "learning_rate": 6.10158985876511e-06, + "loss": 1.63794346, + "memory(GiB)": 111.15, + "step": 36095, + "train_speed(iter/s)": 0.448232 + }, + { + "acc": 0.65285196, + "epoch": 0.9157787924911213, + "grad_norm": 5.5, + "learning_rate": 6.10056697753596e-06, + "loss": 1.60144787, + "memory(GiB)": 111.15, + "step": 36100, + "train_speed(iter/s)": 0.448264 + }, + { + "epoch": 0.9157787924911213, + "eval_acc": 0.6432200334464508, + "eval_loss": 1.6089006662368774, + "eval_runtime": 114.7676, + "eval_samples_per_second": 55.503, + "eval_steps_per_second": 27.752, + "step": 36100 + }, + { + "acc": 0.64461823, + "epoch": 0.9159056316590564, + "grad_norm": 4.8125, + "learning_rate": 6.0995440478989595e-06, + "loss": 1.62612782, + "memory(GiB)": 111.15, + "step": 36105, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.67233086, + "epoch": 0.9160324708269914, + "grad_norm": 6.84375, + "learning_rate": 6.098521069899104e-06, + "loss": 1.50297976, + "memory(GiB)": 111.15, + "step": 36110, + "train_speed(iter/s)": 0.447667 + }, + { + "acc": 0.64715214, + "epoch": 0.9161593099949265, + "grad_norm": 6.8125, + "learning_rate": 6.097498043581385e-06, + "loss": 1.70990868, + "memory(GiB)": 111.15, + "step": 36115, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.65904198, + "epoch": 0.9162861491628614, + "grad_norm": 6.5625, + "learning_rate": 6.096474968990804e-06, + "loss": 1.60258942, + "memory(GiB)": 111.15, + "step": 36120, + "train_speed(iter/s)": 0.447732 + }, + { + "acc": 0.6569458, + "epoch": 0.9164129883307965, + "grad_norm": 4.6875, + "learning_rate": 6.095451846172358e-06, + "loss": 1.6251421, + "memory(GiB)": 111.15, + "step": 36125, + "train_speed(iter/s)": 0.447765 + }, + { + "acc": 0.6414144, + "epoch": 0.9165398274987316, + "grad_norm": 5.78125, + "learning_rate": 6.094428675171049e-06, + "loss": 1.65946999, + "memory(GiB)": 111.15, + "step": 36130, + "train_speed(iter/s)": 0.447797 + }, + { + "acc": 0.64585562, + "epoch": 0.9166666666666666, + "grad_norm": 6.09375, + "learning_rate": 6.09340545603188e-06, + "loss": 1.65260944, + "memory(GiB)": 111.15, + "step": 36135, + "train_speed(iter/s)": 0.44783 + }, + { + "acc": 0.66742196, + "epoch": 0.9167935058346017, + "grad_norm": 5.40625, + "learning_rate": 6.092382188799858e-06, + "loss": 1.62193642, + "memory(GiB)": 111.15, + "step": 36140, + "train_speed(iter/s)": 0.447863 + }, + { + "acc": 0.65737352, + "epoch": 0.9169203450025368, + "grad_norm": 5.28125, + "learning_rate": 6.09135887351999e-06, + "loss": 1.61712189, + "memory(GiB)": 111.15, + "step": 36145, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.6460206, + "epoch": 0.9170471841704718, + "grad_norm": 7.59375, + "learning_rate": 6.090335510237286e-06, + "loss": 1.63158703, + "memory(GiB)": 111.15, + "step": 36150, + "train_speed(iter/s)": 0.447929 + }, + { + "acc": 0.65749063, + "epoch": 0.9171740233384069, + "grad_norm": 5.34375, + "learning_rate": 6.089312098996758e-06, + "loss": 1.54442081, + "memory(GiB)": 111.15, + "step": 36155, + "train_speed(iter/s)": 0.447962 + }, + { + "acc": 0.65311604, + "epoch": 0.9173008625063419, + "grad_norm": 5.84375, + "learning_rate": 6.088288639843422e-06, + "loss": 1.64581108, + "memory(GiB)": 111.15, + "step": 36160, + "train_speed(iter/s)": 0.447994 + }, + { + "acc": 0.65726194, + "epoch": 0.917427701674277, + "grad_norm": 5.8125, + "learning_rate": 6.08726513282229e-06, + "loss": 1.61313229, + "memory(GiB)": 111.15, + "step": 36165, + "train_speed(iter/s)": 0.448027 + }, + { + "acc": 0.65243516, + "epoch": 0.9175545408422121, + "grad_norm": 5.09375, + "learning_rate": 6.0862415779783855e-06, + "loss": 1.61155643, + "memory(GiB)": 111.15, + "step": 36170, + "train_speed(iter/s)": 0.448059 + }, + { + "acc": 0.64384398, + "epoch": 0.9176813800101471, + "grad_norm": 6.09375, + "learning_rate": 6.085217975356726e-06, + "loss": 1.67191277, + "memory(GiB)": 111.15, + "step": 36175, + "train_speed(iter/s)": 0.448092 + }, + { + "acc": 0.6422183, + "epoch": 0.9178082191780822, + "grad_norm": 4.59375, + "learning_rate": 6.084194325002335e-06, + "loss": 1.68194408, + "memory(GiB)": 111.15, + "step": 36180, + "train_speed(iter/s)": 0.448124 + }, + { + "acc": 0.66529198, + "epoch": 0.9179350583460173, + "grad_norm": 5.46875, + "learning_rate": 6.083170626960237e-06, + "loss": 1.60145645, + "memory(GiB)": 111.15, + "step": 36185, + "train_speed(iter/s)": 0.448157 + }, + { + "acc": 0.65623322, + "epoch": 0.9180618975139523, + "grad_norm": 4.5625, + "learning_rate": 6.082146881275458e-06, + "loss": 1.614151, + "memory(GiB)": 111.15, + "step": 36190, + "train_speed(iter/s)": 0.44819 + }, + { + "acc": 0.64873705, + "epoch": 0.9181887366818874, + "grad_norm": 5.21875, + "learning_rate": 6.081123087993028e-06, + "loss": 1.65113754, + "memory(GiB)": 111.15, + "step": 36195, + "train_speed(iter/s)": 0.448223 + }, + { + "acc": 0.63634701, + "epoch": 0.9183155758498224, + "grad_norm": 6.03125, + "learning_rate": 6.0800992471579775e-06, + "loss": 1.6678257, + "memory(GiB)": 111.15, + "step": 36200, + "train_speed(iter/s)": 0.448256 + }, + { + "epoch": 0.9183155758498224, + "eval_acc": 0.6431949674078834, + "eval_loss": 1.6087076663970947, + "eval_runtime": 115.5745, + "eval_samples_per_second": 55.116, + "eval_steps_per_second": 27.558, + "step": 36200 + }, + { + "acc": 0.65501719, + "epoch": 0.9184424150177575, + "grad_norm": 5.84375, + "learning_rate": 6.079075358815341e-06, + "loss": 1.58085175, + "memory(GiB)": 111.15, + "step": 36205, + "train_speed(iter/s)": 0.447622 + }, + { + "acc": 0.64535847, + "epoch": 0.9185692541856926, + "grad_norm": 5.21875, + "learning_rate": 6.078051423010152e-06, + "loss": 1.62130623, + "memory(GiB)": 111.15, + "step": 36210, + "train_speed(iter/s)": 0.447655 + }, + { + "acc": 0.64341249, + "epoch": 0.9186960933536276, + "grad_norm": 5.53125, + "learning_rate": 6.077027439787448e-06, + "loss": 1.69657555, + "memory(GiB)": 111.15, + "step": 36215, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.6530139, + "epoch": 0.9188229325215627, + "grad_norm": 5.625, + "learning_rate": 6.076003409192268e-06, + "loss": 1.64521179, + "memory(GiB)": 111.15, + "step": 36220, + "train_speed(iter/s)": 0.44772 + }, + { + "acc": 0.65316257, + "epoch": 0.9189497716894978, + "grad_norm": 5.4375, + "learning_rate": 6.074979331269656e-06, + "loss": 1.59895287, + "memory(GiB)": 111.15, + "step": 36225, + "train_speed(iter/s)": 0.447753 + }, + { + "acc": 0.64304476, + "epoch": 0.9190766108574328, + "grad_norm": 4.8125, + "learning_rate": 6.0739552060646525e-06, + "loss": 1.68178329, + "memory(GiB)": 111.15, + "step": 36230, + "train_speed(iter/s)": 0.447786 + }, + { + "acc": 0.6580514, + "epoch": 0.9192034500253679, + "grad_norm": 6.0625, + "learning_rate": 6.0729310336223025e-06, + "loss": 1.59110003, + "memory(GiB)": 111.15, + "step": 36235, + "train_speed(iter/s)": 0.447819 + }, + { + "acc": 0.66088467, + "epoch": 0.9193302891933028, + "grad_norm": 4.875, + "learning_rate": 6.071906813987658e-06, + "loss": 1.6205452, + "memory(GiB)": 111.15, + "step": 36240, + "train_speed(iter/s)": 0.447852 + }, + { + "acc": 0.65448775, + "epoch": 0.9194571283612379, + "grad_norm": 5.65625, + "learning_rate": 6.070882547205764e-06, + "loss": 1.62742882, + "memory(GiB)": 111.15, + "step": 36245, + "train_speed(iter/s)": 0.447884 + }, + { + "acc": 0.66130304, + "epoch": 0.919583967529173, + "grad_norm": 5.28125, + "learning_rate": 6.069858233321677e-06, + "loss": 1.5564559, + "memory(GiB)": 111.15, + "step": 36250, + "train_speed(iter/s)": 0.447916 + }, + { + "acc": 0.66485438, + "epoch": 0.919710806697108, + "grad_norm": 5.59375, + "learning_rate": 6.068833872380445e-06, + "loss": 1.55910187, + "memory(GiB)": 111.15, + "step": 36255, + "train_speed(iter/s)": 0.447949 + }, + { + "acc": 0.63518887, + "epoch": 0.9198376458650431, + "grad_norm": 6.125, + "learning_rate": 6.067809464427129e-06, + "loss": 1.68682251, + "memory(GiB)": 111.15, + "step": 36260, + "train_speed(iter/s)": 0.447982 + }, + { + "acc": 0.65564861, + "epoch": 0.9199644850329782, + "grad_norm": 5.53125, + "learning_rate": 6.066785009506786e-06, + "loss": 1.59522572, + "memory(GiB)": 111.15, + "step": 36265, + "train_speed(iter/s)": 0.448015 + }, + { + "acc": 0.65231628, + "epoch": 0.9200913242009132, + "grad_norm": 5.5625, + "learning_rate": 6.065760507664474e-06, + "loss": 1.60623112, + "memory(GiB)": 111.15, + "step": 36270, + "train_speed(iter/s)": 0.448048 + }, + { + "acc": 0.65791125, + "epoch": 0.9202181633688483, + "grad_norm": 6.78125, + "learning_rate": 6.064735958945258e-06, + "loss": 1.62483826, + "memory(GiB)": 111.15, + "step": 36275, + "train_speed(iter/s)": 0.448081 + }, + { + "acc": 0.65667534, + "epoch": 0.9203450025367833, + "grad_norm": 6.5, + "learning_rate": 6.0637113633942006e-06, + "loss": 1.64503937, + "memory(GiB)": 111.15, + "step": 36280, + "train_speed(iter/s)": 0.448113 + }, + { + "acc": 0.63414297, + "epoch": 0.9204718417047184, + "grad_norm": 4.90625, + "learning_rate": 6.0626867210563675e-06, + "loss": 1.71982651, + "memory(GiB)": 111.15, + "step": 36285, + "train_speed(iter/s)": 0.448146 + }, + { + "acc": 0.66492729, + "epoch": 0.9205986808726535, + "grad_norm": 4.78125, + "learning_rate": 6.061662031976828e-06, + "loss": 1.58647194, + "memory(GiB)": 111.15, + "step": 36290, + "train_speed(iter/s)": 0.448179 + }, + { + "acc": 0.65190973, + "epoch": 0.9207255200405885, + "grad_norm": 5.125, + "learning_rate": 6.0606372962006534e-06, + "loss": 1.5588953, + "memory(GiB)": 111.15, + "step": 36295, + "train_speed(iter/s)": 0.448212 + }, + { + "acc": 0.64651089, + "epoch": 0.9208523592085236, + "grad_norm": 5.03125, + "learning_rate": 6.0596125137729145e-06, + "loss": 1.63594685, + "memory(GiB)": 111.15, + "step": 36300, + "train_speed(iter/s)": 0.448244 + }, + { + "epoch": 0.9208523592085236, + "eval_acc": 0.6432104247983332, + "eval_loss": 1.608576774597168, + "eval_runtime": 113.0577, + "eval_samples_per_second": 56.343, + "eval_steps_per_second": 28.171, + "step": 36300 + }, + { + "acc": 0.6424367, + "epoch": 0.9209791983764587, + "grad_norm": 5.78125, + "learning_rate": 6.058587684738685e-06, + "loss": 1.68772202, + "memory(GiB)": 111.15, + "step": 36305, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.66101141, + "epoch": 0.9211060375443937, + "grad_norm": 6.3125, + "learning_rate": 6.057562809143045e-06, + "loss": 1.59679413, + "memory(GiB)": 111.15, + "step": 36310, + "train_speed(iter/s)": 0.447659 + }, + { + "acc": 0.64731216, + "epoch": 0.9212328767123288, + "grad_norm": 5.03125, + "learning_rate": 6.056537887031069e-06, + "loss": 1.6431097, + "memory(GiB)": 111.15, + "step": 36315, + "train_speed(iter/s)": 0.447692 + }, + { + "acc": 0.6571404, + "epoch": 0.9213597158802638, + "grad_norm": 4.84375, + "learning_rate": 6.055512918447841e-06, + "loss": 1.63394375, + "memory(GiB)": 111.15, + "step": 36320, + "train_speed(iter/s)": 0.447724 + }, + { + "acc": 0.64703393, + "epoch": 0.9214865550481989, + "grad_norm": 4.875, + "learning_rate": 6.054487903438442e-06, + "loss": 1.62453079, + "memory(GiB)": 111.15, + "step": 36325, + "train_speed(iter/s)": 0.447756 + }, + { + "acc": 0.64894915, + "epoch": 0.921613394216134, + "grad_norm": 5.53125, + "learning_rate": 6.0534628420479576e-06, + "loss": 1.65224075, + "memory(GiB)": 111.15, + "step": 36330, + "train_speed(iter/s)": 0.447789 + }, + { + "acc": 0.66040955, + "epoch": 0.921740233384069, + "grad_norm": 5.3125, + "learning_rate": 6.0524377343214724e-06, + "loss": 1.66840706, + "memory(GiB)": 111.15, + "step": 36335, + "train_speed(iter/s)": 0.447821 + }, + { + "acc": 0.64087572, + "epoch": 0.9218670725520041, + "grad_norm": 5.65625, + "learning_rate": 6.051412580304079e-06, + "loss": 1.6649456, + "memory(GiB)": 111.15, + "step": 36340, + "train_speed(iter/s)": 0.447853 + }, + { + "acc": 0.64455271, + "epoch": 0.9219939117199392, + "grad_norm": 4.5, + "learning_rate": 6.050387380040864e-06, + "loss": 1.66817341, + "memory(GiB)": 111.15, + "step": 36345, + "train_speed(iter/s)": 0.447886 + }, + { + "acc": 0.63157768, + "epoch": 0.9221207508878742, + "grad_norm": 5.09375, + "learning_rate": 6.049362133576924e-06, + "loss": 1.67944489, + "memory(GiB)": 111.15, + "step": 36350, + "train_speed(iter/s)": 0.447917 + }, + { + "acc": 0.67116184, + "epoch": 0.9222475900558093, + "grad_norm": 5.4375, + "learning_rate": 6.048336840957351e-06, + "loss": 1.57984333, + "memory(GiB)": 111.15, + "step": 36355, + "train_speed(iter/s)": 0.447949 + }, + { + "acc": 0.65210872, + "epoch": 0.9223744292237442, + "grad_norm": 4.84375, + "learning_rate": 6.047311502227245e-06, + "loss": 1.62568436, + "memory(GiB)": 111.15, + "step": 36360, + "train_speed(iter/s)": 0.447982 + }, + { + "acc": 0.64931793, + "epoch": 0.9225012683916793, + "grad_norm": 5.53125, + "learning_rate": 6.046286117431703e-06, + "loss": 1.63582954, + "memory(GiB)": 111.15, + "step": 36365, + "train_speed(iter/s)": 0.448014 + }, + { + "acc": 0.6534153, + "epoch": 0.9226281075596144, + "grad_norm": 4.3125, + "learning_rate": 6.0452606866158246e-06, + "loss": 1.65509644, + "memory(GiB)": 111.15, + "step": 36370, + "train_speed(iter/s)": 0.448045 + }, + { + "acc": 0.64870634, + "epoch": 0.9227549467275494, + "grad_norm": 5.53125, + "learning_rate": 6.044235209824716e-06, + "loss": 1.58035374, + "memory(GiB)": 111.15, + "step": 36375, + "train_speed(iter/s)": 0.448078 + }, + { + "acc": 0.64314871, + "epoch": 0.9228817858954845, + "grad_norm": 5.5625, + "learning_rate": 6.04320968710348e-06, + "loss": 1.70649872, + "memory(GiB)": 111.15, + "step": 36380, + "train_speed(iter/s)": 0.448109 + }, + { + "acc": 0.6299408, + "epoch": 0.9230086250634196, + "grad_norm": 6.1875, + "learning_rate": 6.042184118497223e-06, + "loss": 1.70052242, + "memory(GiB)": 111.15, + "step": 36385, + "train_speed(iter/s)": 0.448142 + }, + { + "acc": 0.66827822, + "epoch": 0.9231354642313546, + "grad_norm": 6.0625, + "learning_rate": 6.0411585040510576e-06, + "loss": 1.59164295, + "memory(GiB)": 111.15, + "step": 36390, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.66311469, + "epoch": 0.9232623033992897, + "grad_norm": 5.03125, + "learning_rate": 6.040132843810091e-06, + "loss": 1.60450554, + "memory(GiB)": 111.15, + "step": 36395, + "train_speed(iter/s)": 0.448206 + }, + { + "acc": 0.67393913, + "epoch": 0.9233891425672247, + "grad_norm": 6.0, + "learning_rate": 6.03910713781944e-06, + "loss": 1.51075735, + "memory(GiB)": 111.15, + "step": 36400, + "train_speed(iter/s)": 0.448238 + }, + { + "epoch": 0.9233891425672247, + "eval_acc": 0.6431903719674793, + "eval_loss": 1.6087710857391357, + "eval_runtime": 114.0808, + "eval_samples_per_second": 55.838, + "eval_steps_per_second": 27.919, + "step": 36400 + }, + { + "acc": 0.64326468, + "epoch": 0.9235159817351598, + "grad_norm": 6.46875, + "learning_rate": 6.038081386124216e-06, + "loss": 1.62376366, + "memory(GiB)": 111.15, + "step": 36405, + "train_speed(iter/s)": 0.447617 + }, + { + "acc": 0.64780474, + "epoch": 0.9236428209030949, + "grad_norm": 5.59375, + "learning_rate": 6.037055588769539e-06, + "loss": 1.60299091, + "memory(GiB)": 111.15, + "step": 36410, + "train_speed(iter/s)": 0.44765 + }, + { + "acc": 0.65084329, + "epoch": 0.9237696600710299, + "grad_norm": 4.6875, + "learning_rate": 6.036029745800527e-06, + "loss": 1.57017765, + "memory(GiB)": 111.15, + "step": 36415, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.65075684, + "epoch": 0.923896499238965, + "grad_norm": 5.09375, + "learning_rate": 6.0350038572623e-06, + "loss": 1.67223949, + "memory(GiB)": 111.15, + "step": 36420, + "train_speed(iter/s)": 0.447715 + }, + { + "acc": 0.63734651, + "epoch": 0.9240233384069001, + "grad_norm": 5.09375, + "learning_rate": 6.033977923199984e-06, + "loss": 1.6192379, + "memory(GiB)": 111.15, + "step": 36425, + "train_speed(iter/s)": 0.447748 + }, + { + "acc": 0.66568174, + "epoch": 0.9241501775748351, + "grad_norm": 6.0, + "learning_rate": 6.032951943658702e-06, + "loss": 1.53037319, + "memory(GiB)": 111.15, + "step": 36430, + "train_speed(iter/s)": 0.447781 + }, + { + "acc": 0.6478334, + "epoch": 0.9242770167427702, + "grad_norm": 5.15625, + "learning_rate": 6.031925918683582e-06, + "loss": 1.62719917, + "memory(GiB)": 111.15, + "step": 36435, + "train_speed(iter/s)": 0.447814 + }, + { + "acc": 0.63664618, + "epoch": 0.9244038559107052, + "grad_norm": 6.03125, + "learning_rate": 6.030899848319754e-06, + "loss": 1.67872505, + "memory(GiB)": 111.15, + "step": 36440, + "train_speed(iter/s)": 0.447846 + }, + { + "acc": 0.6377902, + "epoch": 0.9245306950786403, + "grad_norm": 4.9375, + "learning_rate": 6.029873732612346e-06, + "loss": 1.69607506, + "memory(GiB)": 111.15, + "step": 36445, + "train_speed(iter/s)": 0.447879 + }, + { + "acc": 0.64654922, + "epoch": 0.9246575342465754, + "grad_norm": 5.75, + "learning_rate": 6.028847571606493e-06, + "loss": 1.6563446, + "memory(GiB)": 111.15, + "step": 36450, + "train_speed(iter/s)": 0.447912 + }, + { + "acc": 0.64020529, + "epoch": 0.9247843734145104, + "grad_norm": 4.96875, + "learning_rate": 6.0278213653473305e-06, + "loss": 1.63829269, + "memory(GiB)": 111.15, + "step": 36455, + "train_speed(iter/s)": 0.447945 + }, + { + "acc": 0.64891663, + "epoch": 0.9249112125824455, + "grad_norm": 4.90625, + "learning_rate": 6.026795113879998e-06, + "loss": 1.64488754, + "memory(GiB)": 111.15, + "step": 36460, + "train_speed(iter/s)": 0.447978 + }, + { + "acc": 0.64766383, + "epoch": 0.9250380517503806, + "grad_norm": 5.0, + "learning_rate": 6.025768817249629e-06, + "loss": 1.65279617, + "memory(GiB)": 111.15, + "step": 36465, + "train_speed(iter/s)": 0.44801 + }, + { + "acc": 0.62383528, + "epoch": 0.9251648909183156, + "grad_norm": 5.65625, + "learning_rate": 6.024742475501369e-06, + "loss": 1.72132759, + "memory(GiB)": 111.15, + "step": 36470, + "train_speed(iter/s)": 0.448043 + }, + { + "acc": 0.64978466, + "epoch": 0.9252917300862507, + "grad_norm": 5.34375, + "learning_rate": 6.023716088680359e-06, + "loss": 1.62371216, + "memory(GiB)": 111.15, + "step": 36475, + "train_speed(iter/s)": 0.448076 + }, + { + "acc": 0.64134035, + "epoch": 0.9254185692541856, + "grad_norm": 5.65625, + "learning_rate": 6.022689656831746e-06, + "loss": 1.56668301, + "memory(GiB)": 111.15, + "step": 36480, + "train_speed(iter/s)": 0.448109 + }, + { + "acc": 0.64621105, + "epoch": 0.9255454084221207, + "grad_norm": 7.59375, + "learning_rate": 6.021663180000675e-06, + "loss": 1.60215034, + "memory(GiB)": 111.15, + "step": 36485, + "train_speed(iter/s)": 0.448142 + }, + { + "acc": 0.64770026, + "epoch": 0.9256722475900558, + "grad_norm": 5.28125, + "learning_rate": 6.020636658232297e-06, + "loss": 1.63039856, + "memory(GiB)": 111.15, + "step": 36490, + "train_speed(iter/s)": 0.448174 + }, + { + "acc": 0.64439898, + "epoch": 0.9257990867579908, + "grad_norm": 5.65625, + "learning_rate": 6.019610091571762e-06, + "loss": 1.65722694, + "memory(GiB)": 111.15, + "step": 36495, + "train_speed(iter/s)": 0.448207 + }, + { + "acc": 0.64405375, + "epoch": 0.9259259259259259, + "grad_norm": 6.40625, + "learning_rate": 6.018583480064222e-06, + "loss": 1.71776733, + "memory(GiB)": 111.15, + "step": 36500, + "train_speed(iter/s)": 0.44824 + }, + { + "epoch": 0.9259259259259259, + "eval_acc": 0.643144417563439, + "eval_loss": 1.6087801456451416, + "eval_runtime": 113.5712, + "eval_samples_per_second": 56.088, + "eval_steps_per_second": 28.044, + "step": 36500 + }, + { + "acc": 0.64949284, + "epoch": 0.926052765093861, + "grad_norm": 5.0, + "learning_rate": 6.017556823754833e-06, + "loss": 1.53541946, + "memory(GiB)": 111.15, + "step": 36505, + "train_speed(iter/s)": 0.447623 + }, + { + "acc": 0.6513011, + "epoch": 0.926179604261796, + "grad_norm": 7.1875, + "learning_rate": 6.016530122688753e-06, + "loss": 1.61536713, + "memory(GiB)": 111.15, + "step": 36510, + "train_speed(iter/s)": 0.447655 + }, + { + "acc": 0.66707501, + "epoch": 0.9263064434297311, + "grad_norm": 5.71875, + "learning_rate": 6.015503376911138e-06, + "loss": 1.62300262, + "memory(GiB)": 111.15, + "step": 36515, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.63881717, + "epoch": 0.9264332825976661, + "grad_norm": 4.71875, + "learning_rate": 6.0144765864671515e-06, + "loss": 1.64056244, + "memory(GiB)": 111.15, + "step": 36520, + "train_speed(iter/s)": 0.447721 + }, + { + "acc": 0.63129015, + "epoch": 0.9265601217656012, + "grad_norm": 5.0, + "learning_rate": 6.013449751401954e-06, + "loss": 1.68068886, + "memory(GiB)": 111.15, + "step": 36525, + "train_speed(iter/s)": 0.447753 + }, + { + "acc": 0.64978123, + "epoch": 0.9266869609335363, + "grad_norm": 5.875, + "learning_rate": 6.012422871760715e-06, + "loss": 1.63124695, + "memory(GiB)": 111.15, + "step": 36530, + "train_speed(iter/s)": 0.447786 + }, + { + "acc": 0.63995261, + "epoch": 0.9268138001014713, + "grad_norm": 5.09375, + "learning_rate": 6.011395947588594e-06, + "loss": 1.66824284, + "memory(GiB)": 111.15, + "step": 36535, + "train_speed(iter/s)": 0.447818 + }, + { + "acc": 0.64427938, + "epoch": 0.9269406392694064, + "grad_norm": 5.09375, + "learning_rate": 6.010368978930767e-06, + "loss": 1.63021641, + "memory(GiB)": 111.15, + "step": 36540, + "train_speed(iter/s)": 0.447851 + }, + { + "acc": 0.65215015, + "epoch": 0.9270674784373415, + "grad_norm": 5.28125, + "learning_rate": 6.0093419658323995e-06, + "loss": 1.67144966, + "memory(GiB)": 111.15, + "step": 36545, + "train_speed(iter/s)": 0.447884 + }, + { + "acc": 0.63101187, + "epoch": 0.9271943176052765, + "grad_norm": 5.46875, + "learning_rate": 6.0083149083386675e-06, + "loss": 1.68657837, + "memory(GiB)": 111.15, + "step": 36550, + "train_speed(iter/s)": 0.447916 + }, + { + "acc": 0.64131417, + "epoch": 0.9273211567732116, + "grad_norm": 5.53125, + "learning_rate": 6.007287806494742e-06, + "loss": 1.6572464, + "memory(GiB)": 111.15, + "step": 36555, + "train_speed(iter/s)": 0.447949 + }, + { + "acc": 0.65773125, + "epoch": 0.9274479959411466, + "grad_norm": 4.90625, + "learning_rate": 6.006260660345802e-06, + "loss": 1.61939945, + "memory(GiB)": 111.15, + "step": 36560, + "train_speed(iter/s)": 0.447982 + }, + { + "acc": 0.6508194, + "epoch": 0.9275748351090817, + "grad_norm": 5.0625, + "learning_rate": 6.005233469937027e-06, + "loss": 1.65617733, + "memory(GiB)": 111.15, + "step": 36565, + "train_speed(iter/s)": 0.448015 + }, + { + "acc": 0.66327219, + "epoch": 0.9277016742770168, + "grad_norm": 6.0, + "learning_rate": 6.004206235313594e-06, + "loss": 1.5682991, + "memory(GiB)": 111.15, + "step": 36570, + "train_speed(iter/s)": 0.448048 + }, + { + "acc": 0.65131907, + "epoch": 0.9278285134449518, + "grad_norm": 4.96875, + "learning_rate": 6.003178956520688e-06, + "loss": 1.61189213, + "memory(GiB)": 111.15, + "step": 36575, + "train_speed(iter/s)": 0.44808 + }, + { + "acc": 0.64789929, + "epoch": 0.9279553526128869, + "grad_norm": 4.3125, + "learning_rate": 6.002151633603493e-06, + "loss": 1.64973335, + "memory(GiB)": 111.15, + "step": 36580, + "train_speed(iter/s)": 0.448113 + }, + { + "acc": 0.64127016, + "epoch": 0.928082191780822, + "grad_norm": 4.625, + "learning_rate": 6.0011242666071945e-06, + "loss": 1.61005478, + "memory(GiB)": 111.15, + "step": 36585, + "train_speed(iter/s)": 0.448146 + }, + { + "acc": 0.64447584, + "epoch": 0.928209030948757, + "grad_norm": 5.0625, + "learning_rate": 6.000096855576982e-06, + "loss": 1.60900726, + "memory(GiB)": 111.15, + "step": 36590, + "train_speed(iter/s)": 0.448179 + }, + { + "acc": 0.65335054, + "epoch": 0.928335870116692, + "grad_norm": 5.03125, + "learning_rate": 5.999069400558044e-06, + "loss": 1.63926392, + "memory(GiB)": 111.15, + "step": 36595, + "train_speed(iter/s)": 0.448211 + }, + { + "acc": 0.65481186, + "epoch": 0.928462709284627, + "grad_norm": 5.8125, + "learning_rate": 5.998041901595573e-06, + "loss": 1.60935268, + "memory(GiB)": 111.15, + "step": 36600, + "train_speed(iter/s)": 0.448243 + }, + { + "epoch": 0.928462709284627, + "eval_acc": 0.6431602927211983, + "eval_loss": 1.6087822914123535, + "eval_runtime": 114.013, + "eval_samples_per_second": 55.871, + "eval_steps_per_second": 27.935, + "step": 36600 + }, + { + "acc": 0.63276386, + "epoch": 0.9285895484525621, + "grad_norm": 5.59375, + "learning_rate": 5.997014358734763e-06, + "loss": 1.64770737, + "memory(GiB)": 111.15, + "step": 36605, + "train_speed(iter/s)": 0.447626 + }, + { + "acc": 0.65176911, + "epoch": 0.9287163876204972, + "grad_norm": 5.1875, + "learning_rate": 5.995986772020811e-06, + "loss": 1.57933636, + "memory(GiB)": 111.15, + "step": 36610, + "train_speed(iter/s)": 0.447658 + }, + { + "acc": 0.64311409, + "epoch": 0.9288432267884322, + "grad_norm": 4.96875, + "learning_rate": 5.994959141498913e-06, + "loss": 1.64045372, + "memory(GiB)": 111.15, + "step": 36615, + "train_speed(iter/s)": 0.447691 + }, + { + "acc": 0.65488167, + "epoch": 0.9289700659563673, + "grad_norm": 4.3125, + "learning_rate": 5.993931467214272e-06, + "loss": 1.63277283, + "memory(GiB)": 111.15, + "step": 36620, + "train_speed(iter/s)": 0.447724 + }, + { + "acc": 0.64676361, + "epoch": 0.9290969051243024, + "grad_norm": 5.28125, + "learning_rate": 5.992903749212084e-06, + "loss": 1.6516655, + "memory(GiB)": 111.15, + "step": 36625, + "train_speed(iter/s)": 0.447756 + }, + { + "acc": 0.6456274, + "epoch": 0.9292237442922374, + "grad_norm": 5.9375, + "learning_rate": 5.991875987537559e-06, + "loss": 1.66086006, + "memory(GiB)": 111.15, + "step": 36630, + "train_speed(iter/s)": 0.447789 + }, + { + "acc": 0.65170717, + "epoch": 0.9293505834601725, + "grad_norm": 6.09375, + "learning_rate": 5.990848182235898e-06, + "loss": 1.65292854, + "memory(GiB)": 111.15, + "step": 36635, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.63868465, + "epoch": 0.9294774226281075, + "grad_norm": 6.5625, + "learning_rate": 5.98982033335231e-06, + "loss": 1.69883995, + "memory(GiB)": 111.15, + "step": 36640, + "train_speed(iter/s)": 0.447854 + }, + { + "acc": 0.65549889, + "epoch": 0.9296042617960426, + "grad_norm": 5.15625, + "learning_rate": 5.988792440932006e-06, + "loss": 1.51775379, + "memory(GiB)": 111.15, + "step": 36645, + "train_speed(iter/s)": 0.447887 + }, + { + "acc": 0.65993223, + "epoch": 0.9297311009639777, + "grad_norm": 4.875, + "learning_rate": 5.987764505020195e-06, + "loss": 1.52925968, + "memory(GiB)": 111.15, + "step": 36650, + "train_speed(iter/s)": 0.44792 + }, + { + "acc": 0.63512383, + "epoch": 0.9298579401319127, + "grad_norm": 5.53125, + "learning_rate": 5.986736525662091e-06, + "loss": 1.65203342, + "memory(GiB)": 111.15, + "step": 36655, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.65197134, + "epoch": 0.9299847792998478, + "grad_norm": 4.46875, + "learning_rate": 5.985708502902909e-06, + "loss": 1.58356514, + "memory(GiB)": 111.15, + "step": 36660, + "train_speed(iter/s)": 0.447985 + }, + { + "acc": 0.64990916, + "epoch": 0.9301116184677829, + "grad_norm": 5.34375, + "learning_rate": 5.984680436787867e-06, + "loss": 1.6359314, + "memory(GiB)": 111.15, + "step": 36665, + "train_speed(iter/s)": 0.448017 + }, + { + "acc": 0.64861102, + "epoch": 0.9302384576357179, + "grad_norm": 5.0, + "learning_rate": 5.983652327362182e-06, + "loss": 1.63021126, + "memory(GiB)": 111.15, + "step": 36670, + "train_speed(iter/s)": 0.44805 + }, + { + "acc": 0.65491462, + "epoch": 0.930365296803653, + "grad_norm": 4.8125, + "learning_rate": 5.982624174671077e-06, + "loss": 1.65926971, + "memory(GiB)": 111.15, + "step": 36675, + "train_speed(iter/s)": 0.448082 + }, + { + "acc": 0.65743828, + "epoch": 0.930492135971588, + "grad_norm": 5.59375, + "learning_rate": 5.981595978759773e-06, + "loss": 1.61877575, + "memory(GiB)": 111.15, + "step": 36680, + "train_speed(iter/s)": 0.448115 + }, + { + "acc": 0.64636078, + "epoch": 0.9306189751395231, + "grad_norm": 4.53125, + "learning_rate": 5.980567739673495e-06, + "loss": 1.66098442, + "memory(GiB)": 111.15, + "step": 36685, + "train_speed(iter/s)": 0.448147 + }, + { + "acc": 0.65286617, + "epoch": 0.9307458143074582, + "grad_norm": 5.25, + "learning_rate": 5.979539457457472e-06, + "loss": 1.62726936, + "memory(GiB)": 111.15, + "step": 36690, + "train_speed(iter/s)": 0.44818 + }, + { + "acc": 0.65126867, + "epoch": 0.9308726534753932, + "grad_norm": 6.03125, + "learning_rate": 5.978511132156928e-06, + "loss": 1.63660202, + "memory(GiB)": 111.15, + "step": 36695, + "train_speed(iter/s)": 0.448213 + }, + { + "acc": 0.64327288, + "epoch": 0.9309994926433283, + "grad_norm": 5.96875, + "learning_rate": 5.9774827638170965e-06, + "loss": 1.67627754, + "memory(GiB)": 111.15, + "step": 36700, + "train_speed(iter/s)": 0.448245 + }, + { + "epoch": 0.9309994926433283, + "eval_acc": 0.643065877309261, + "eval_loss": 1.6087065935134888, + "eval_runtime": 114.1016, + "eval_samples_per_second": 55.827, + "eval_steps_per_second": 27.914, + "step": 36700 + }, + { + "acc": 0.65392179, + "epoch": 0.9311263318112634, + "grad_norm": 4.90625, + "learning_rate": 5.9764543524832085e-06, + "loss": 1.56462126, + "memory(GiB)": 111.15, + "step": 36705, + "train_speed(iter/s)": 0.447629 + }, + { + "acc": 0.66371756, + "epoch": 0.9312531709791984, + "grad_norm": 5.90625, + "learning_rate": 5.975425898200499e-06, + "loss": 1.60209465, + "memory(GiB)": 111.15, + "step": 36710, + "train_speed(iter/s)": 0.447662 + }, + { + "acc": 0.64541454, + "epoch": 0.9313800101471335, + "grad_norm": 5.84375, + "learning_rate": 5.974397401014202e-06, + "loss": 1.69137001, + "memory(GiB)": 111.15, + "step": 36715, + "train_speed(iter/s)": 0.447694 + }, + { + "acc": 0.64094782, + "epoch": 0.9315068493150684, + "grad_norm": 4.75, + "learning_rate": 5.973368860969559e-06, + "loss": 1.65377998, + "memory(GiB)": 111.15, + "step": 36720, + "train_speed(iter/s)": 0.447727 + }, + { + "acc": 0.64917321, + "epoch": 0.9316336884830035, + "grad_norm": 6.5625, + "learning_rate": 5.972340278111808e-06, + "loss": 1.64718227, + "memory(GiB)": 111.15, + "step": 36725, + "train_speed(iter/s)": 0.44776 + }, + { + "acc": 0.65446157, + "epoch": 0.9317605276509386, + "grad_norm": 4.5625, + "learning_rate": 5.9713116524861895e-06, + "loss": 1.50676193, + "memory(GiB)": 111.15, + "step": 36730, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.65375485, + "epoch": 0.9318873668188736, + "grad_norm": 6.0, + "learning_rate": 5.970282984137947e-06, + "loss": 1.68175316, + "memory(GiB)": 111.15, + "step": 36735, + "train_speed(iter/s)": 0.447824 + }, + { + "acc": 0.67130117, + "epoch": 0.9320142059868087, + "grad_norm": 4.65625, + "learning_rate": 5.969254273112328e-06, + "loss": 1.56214314, + "memory(GiB)": 111.15, + "step": 36740, + "train_speed(iter/s)": 0.447857 + }, + { + "acc": 0.65184336, + "epoch": 0.9321410451547438, + "grad_norm": 5.5, + "learning_rate": 5.968225519454577e-06, + "loss": 1.65125656, + "memory(GiB)": 111.15, + "step": 36745, + "train_speed(iter/s)": 0.44789 + }, + { + "acc": 0.66311512, + "epoch": 0.9322678843226788, + "grad_norm": 5.625, + "learning_rate": 5.967196723209947e-06, + "loss": 1.59511204, + "memory(GiB)": 111.15, + "step": 36750, + "train_speed(iter/s)": 0.447922 + }, + { + "acc": 0.64025402, + "epoch": 0.9323947234906139, + "grad_norm": 6.125, + "learning_rate": 5.966167884423686e-06, + "loss": 1.70963058, + "memory(GiB)": 111.15, + "step": 36755, + "train_speed(iter/s)": 0.447955 + }, + { + "acc": 0.67873564, + "epoch": 0.9325215626585489, + "grad_norm": 5.5, + "learning_rate": 5.965139003141048e-06, + "loss": 1.5078661, + "memory(GiB)": 111.15, + "step": 36760, + "train_speed(iter/s)": 0.447987 + }, + { + "acc": 0.64550447, + "epoch": 0.932648401826484, + "grad_norm": 5.1875, + "learning_rate": 5.964110079407287e-06, + "loss": 1.63001671, + "memory(GiB)": 111.15, + "step": 36765, + "train_speed(iter/s)": 0.44802 + }, + { + "acc": 0.64718099, + "epoch": 0.9327752409944191, + "grad_norm": 5.1875, + "learning_rate": 5.9630811132676625e-06, + "loss": 1.66544781, + "memory(GiB)": 111.15, + "step": 36770, + "train_speed(iter/s)": 0.448053 + }, + { + "acc": 0.65200386, + "epoch": 0.9329020801623541, + "grad_norm": 5.375, + "learning_rate": 5.962052104767427e-06, + "loss": 1.66214523, + "memory(GiB)": 111.15, + "step": 36775, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.65890656, + "epoch": 0.9330289193302892, + "grad_norm": 5.46875, + "learning_rate": 5.961023053951848e-06, + "loss": 1.62643089, + "memory(GiB)": 111.15, + "step": 36780, + "train_speed(iter/s)": 0.448117 + }, + { + "acc": 0.65522518, + "epoch": 0.9331557584982243, + "grad_norm": 5.625, + "learning_rate": 5.9599939608661825e-06, + "loss": 1.66874218, + "memory(GiB)": 111.15, + "step": 36785, + "train_speed(iter/s)": 0.448149 + }, + { + "acc": 0.64050875, + "epoch": 0.9332825976661593, + "grad_norm": 4.875, + "learning_rate": 5.9589648255556975e-06, + "loss": 1.65283127, + "memory(GiB)": 111.15, + "step": 36790, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.64744935, + "epoch": 0.9334094368340944, + "grad_norm": 5.0, + "learning_rate": 5.957935648065658e-06, + "loss": 1.62253113, + "memory(GiB)": 111.15, + "step": 36795, + "train_speed(iter/s)": 0.448213 + }, + { + "acc": 0.65219469, + "epoch": 0.9335362760020294, + "grad_norm": 6.34375, + "learning_rate": 5.956906428441331e-06, + "loss": 1.61010303, + "memory(GiB)": 111.15, + "step": 36800, + "train_speed(iter/s)": 0.448245 + }, + { + "epoch": 0.9335362760020294, + "eval_acc": 0.6432008161502157, + "eval_loss": 1.6087191104888916, + "eval_runtime": 112.8859, + "eval_samples_per_second": 56.429, + "eval_steps_per_second": 28.214, + "step": 36800 + }, + { + "acc": 0.64840775, + "epoch": 0.9336631151699645, + "grad_norm": 4.46875, + "learning_rate": 5.955877166727988e-06, + "loss": 1.61947365, + "memory(GiB)": 111.15, + "step": 36805, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.65272045, + "epoch": 0.9337899543378996, + "grad_norm": 6.15625, + "learning_rate": 5.954847862970898e-06, + "loss": 1.64417877, + "memory(GiB)": 111.15, + "step": 36810, + "train_speed(iter/s)": 0.44767 + }, + { + "acc": 0.64016085, + "epoch": 0.9339167935058346, + "grad_norm": 4.53125, + "learning_rate": 5.953818517215338e-06, + "loss": 1.6786602, + "memory(GiB)": 111.15, + "step": 36815, + "train_speed(iter/s)": 0.447702 + }, + { + "acc": 0.65304708, + "epoch": 0.9340436326737697, + "grad_norm": 5.1875, + "learning_rate": 5.95278912950658e-06, + "loss": 1.64396648, + "memory(GiB)": 111.15, + "step": 36820, + "train_speed(iter/s)": 0.447734 + }, + { + "acc": 0.67129741, + "epoch": 0.9341704718417048, + "grad_norm": 4.875, + "learning_rate": 5.9517596998899e-06, + "loss": 1.53759356, + "memory(GiB)": 111.15, + "step": 36825, + "train_speed(iter/s)": 0.447767 + }, + { + "acc": 0.66465149, + "epoch": 0.9342973110096398, + "grad_norm": 5.65625, + "learning_rate": 5.9507302284105836e-06, + "loss": 1.59951611, + "memory(GiB)": 111.15, + "step": 36830, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.6611104, + "epoch": 0.9344241501775749, + "grad_norm": 5.03125, + "learning_rate": 5.949700715113904e-06, + "loss": 1.61405067, + "memory(GiB)": 111.15, + "step": 36835, + "train_speed(iter/s)": 0.447831 + }, + { + "acc": 0.64961329, + "epoch": 0.9345509893455098, + "grad_norm": 5.96875, + "learning_rate": 5.9486711600451484e-06, + "loss": 1.62284985, + "memory(GiB)": 111.15, + "step": 36840, + "train_speed(iter/s)": 0.447864 + }, + { + "acc": 0.66234055, + "epoch": 0.9346778285134449, + "grad_norm": 5.1875, + "learning_rate": 5.9476415632495974e-06, + "loss": 1.63580761, + "memory(GiB)": 111.15, + "step": 36845, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.64065399, + "epoch": 0.93480466768138, + "grad_norm": 5.09375, + "learning_rate": 5.946611924772542e-06, + "loss": 1.62230911, + "memory(GiB)": 111.15, + "step": 36850, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.66177635, + "epoch": 0.934931506849315, + "grad_norm": 4.78125, + "learning_rate": 5.945582244659267e-06, + "loss": 1.5749506, + "memory(GiB)": 111.15, + "step": 36855, + "train_speed(iter/s)": 0.447961 + }, + { + "acc": 0.65094151, + "epoch": 0.9350583460172501, + "grad_norm": 5.59375, + "learning_rate": 5.944552522955063e-06, + "loss": 1.6492569, + "memory(GiB)": 111.15, + "step": 36860, + "train_speed(iter/s)": 0.447993 + }, + { + "acc": 0.6713872, + "epoch": 0.9351851851851852, + "grad_norm": 4.25, + "learning_rate": 5.943522759705221e-06, + "loss": 1.60539436, + "memory(GiB)": 111.15, + "step": 36865, + "train_speed(iter/s)": 0.448025 + }, + { + "acc": 0.63201346, + "epoch": 0.9353120243531202, + "grad_norm": 4.75, + "learning_rate": 5.942492954955037e-06, + "loss": 1.71706161, + "memory(GiB)": 111.15, + "step": 36870, + "train_speed(iter/s)": 0.448058 + }, + { + "acc": 0.63267298, + "epoch": 0.9354388635210553, + "grad_norm": 5.40625, + "learning_rate": 5.941463108749804e-06, + "loss": 1.70435524, + "memory(GiB)": 111.15, + "step": 36875, + "train_speed(iter/s)": 0.44809 + }, + { + "acc": 0.65122232, + "epoch": 0.9355657026889903, + "grad_norm": 5.9375, + "learning_rate": 5.940433221134821e-06, + "loss": 1.62943726, + "memory(GiB)": 111.15, + "step": 36880, + "train_speed(iter/s)": 0.448123 + }, + { + "acc": 0.66188183, + "epoch": 0.9356925418569254, + "grad_norm": 5.59375, + "learning_rate": 5.9394032921553856e-06, + "loss": 1.62202454, + "memory(GiB)": 111.15, + "step": 36885, + "train_speed(iter/s)": 0.448155 + }, + { + "acc": 0.64588065, + "epoch": 0.9358193810248605, + "grad_norm": 6.1875, + "learning_rate": 5.9383733218568e-06, + "loss": 1.67549191, + "memory(GiB)": 111.15, + "step": 36890, + "train_speed(iter/s)": 0.448187 + }, + { + "acc": 0.6477931, + "epoch": 0.9359462201927955, + "grad_norm": 5.0, + "learning_rate": 5.937343310284365e-06, + "loss": 1.6310688, + "memory(GiB)": 111.15, + "step": 36895, + "train_speed(iter/s)": 0.44822 + }, + { + "acc": 0.64817858, + "epoch": 0.9360730593607306, + "grad_norm": 4.96875, + "learning_rate": 5.936313257483387e-06, + "loss": 1.64522705, + "memory(GiB)": 111.15, + "step": 36900, + "train_speed(iter/s)": 0.448252 + }, + { + "epoch": 0.9360730593607306, + "eval_acc": 0.6431627993250552, + "eval_loss": 1.6087678670883179, + "eval_runtime": 113.3204, + "eval_samples_per_second": 56.212, + "eval_steps_per_second": 28.106, + "step": 36900 + }, + { + "acc": 0.64475031, + "epoch": 0.9361998985286657, + "grad_norm": 4.75, + "learning_rate": 5.935283163499171e-06, + "loss": 1.6564991, + "memory(GiB)": 111.15, + "step": 36905, + "train_speed(iter/s)": 0.447643 + }, + { + "acc": 0.63907037, + "epoch": 0.9363267376966007, + "grad_norm": 5.0, + "learning_rate": 5.9342530283770274e-06, + "loss": 1.70316582, + "memory(GiB)": 111.15, + "step": 36910, + "train_speed(iter/s)": 0.447675 + }, + { + "acc": 0.65809822, + "epoch": 0.9364535768645358, + "grad_norm": 4.3125, + "learning_rate": 5.9332228521622615e-06, + "loss": 1.54398994, + "memory(GiB)": 111.15, + "step": 36915, + "train_speed(iter/s)": 0.447707 + }, + { + "acc": 0.6620399, + "epoch": 0.9365804160324708, + "grad_norm": 5.1875, + "learning_rate": 5.93219263490019e-06, + "loss": 1.62343616, + "memory(GiB)": 111.15, + "step": 36920, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.66314096, + "epoch": 0.9367072552004059, + "grad_norm": 5.40625, + "learning_rate": 5.931162376636123e-06, + "loss": 1.6736824, + "memory(GiB)": 111.15, + "step": 36925, + "train_speed(iter/s)": 0.44777 + }, + { + "acc": 0.65223899, + "epoch": 0.936834094368341, + "grad_norm": 6.0, + "learning_rate": 5.93013207741538e-06, + "loss": 1.59007149, + "memory(GiB)": 111.15, + "step": 36930, + "train_speed(iter/s)": 0.447802 + }, + { + "acc": 0.64796052, + "epoch": 0.936960933536276, + "grad_norm": 4.9375, + "learning_rate": 5.929101737283274e-06, + "loss": 1.59672184, + "memory(GiB)": 111.15, + "step": 36935, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.63852029, + "epoch": 0.9370877727042111, + "grad_norm": 5.28125, + "learning_rate": 5.928071356285126e-06, + "loss": 1.62632904, + "memory(GiB)": 111.15, + "step": 36940, + "train_speed(iter/s)": 0.447866 + }, + { + "acc": 0.66527581, + "epoch": 0.9372146118721462, + "grad_norm": 5.34375, + "learning_rate": 5.927040934466255e-06, + "loss": 1.58500309, + "memory(GiB)": 111.15, + "step": 36945, + "train_speed(iter/s)": 0.447898 + }, + { + "acc": 0.66687412, + "epoch": 0.9373414510400812, + "grad_norm": 4.9375, + "learning_rate": 5.926010471871986e-06, + "loss": 1.54485474, + "memory(GiB)": 111.15, + "step": 36950, + "train_speed(iter/s)": 0.44793 + }, + { + "acc": 0.65134335, + "epoch": 0.9374682902080163, + "grad_norm": 4.59375, + "learning_rate": 5.924979968547642e-06, + "loss": 1.57324781, + "memory(GiB)": 111.15, + "step": 36955, + "train_speed(iter/s)": 0.447962 + }, + { + "acc": 0.6303236, + "epoch": 0.9375951293759512, + "grad_norm": 6.125, + "learning_rate": 5.9239494245385485e-06, + "loss": 1.74390106, + "memory(GiB)": 111.15, + "step": 36960, + "train_speed(iter/s)": 0.447994 + }, + { + "acc": 0.65009909, + "epoch": 0.9377219685438863, + "grad_norm": 5.4375, + "learning_rate": 5.9229188398900325e-06, + "loss": 1.64183292, + "memory(GiB)": 111.15, + "step": 36965, + "train_speed(iter/s)": 0.448026 + }, + { + "acc": 0.63904181, + "epoch": 0.9378488077118214, + "grad_norm": 4.9375, + "learning_rate": 5.921888214647429e-06, + "loss": 1.67749443, + "memory(GiB)": 111.15, + "step": 36970, + "train_speed(iter/s)": 0.448058 + }, + { + "acc": 0.6445075, + "epoch": 0.9379756468797564, + "grad_norm": 5.34375, + "learning_rate": 5.920857548856064e-06, + "loss": 1.57803679, + "memory(GiB)": 111.15, + "step": 36975, + "train_speed(iter/s)": 0.448089 + }, + { + "acc": 0.6463017, + "epoch": 0.9381024860476915, + "grad_norm": 6.0, + "learning_rate": 5.919826842561274e-06, + "loss": 1.61653957, + "memory(GiB)": 111.15, + "step": 36980, + "train_speed(iter/s)": 0.448121 + }, + { + "acc": 0.64693284, + "epoch": 0.9382293252156266, + "grad_norm": 5.4375, + "learning_rate": 5.91879609580839e-06, + "loss": 1.67656441, + "memory(GiB)": 111.15, + "step": 36985, + "train_speed(iter/s)": 0.448153 + }, + { + "acc": 0.64013429, + "epoch": 0.9383561643835616, + "grad_norm": 6.34375, + "learning_rate": 5.917765308642754e-06, + "loss": 1.7001339, + "memory(GiB)": 111.15, + "step": 36990, + "train_speed(iter/s)": 0.448184 + }, + { + "acc": 0.64782805, + "epoch": 0.9384830035514967, + "grad_norm": 5.0625, + "learning_rate": 5.9167344811097014e-06, + "loss": 1.68568745, + "memory(GiB)": 111.15, + "step": 36995, + "train_speed(iter/s)": 0.448216 + }, + { + "acc": 0.66303568, + "epoch": 0.9386098427194317, + "grad_norm": 5.65625, + "learning_rate": 5.9157036132545735e-06, + "loss": 1.57605095, + "memory(GiB)": 111.15, + "step": 37000, + "train_speed(iter/s)": 0.448248 + }, + { + "epoch": 0.9386098427194317, + "eval_acc": 0.643208335961786, + "eval_loss": 1.6087180376052856, + "eval_runtime": 113.0879, + "eval_samples_per_second": 56.328, + "eval_steps_per_second": 28.164, + "step": 37000 + }, + { + "acc": 0.66180563, + "epoch": 0.9387366818873668, + "grad_norm": 6.71875, + "learning_rate": 5.914672705122713e-06, + "loss": 1.64287071, + "memory(GiB)": 111.15, + "step": 37005, + "train_speed(iter/s)": 0.447641 + }, + { + "acc": 0.64640808, + "epoch": 0.9388635210553019, + "grad_norm": 6.1875, + "learning_rate": 5.9136417567594615e-06, + "loss": 1.63961067, + "memory(GiB)": 111.15, + "step": 37010, + "train_speed(iter/s)": 0.447674 + }, + { + "acc": 0.65565939, + "epoch": 0.9389903602232369, + "grad_norm": 5.40625, + "learning_rate": 5.9126107682101675e-06, + "loss": 1.67383423, + "memory(GiB)": 111.15, + "step": 37015, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.63465281, + "epoch": 0.939117199391172, + "grad_norm": 4.09375, + "learning_rate": 5.911579739520178e-06, + "loss": 1.63523598, + "memory(GiB)": 111.15, + "step": 37020, + "train_speed(iter/s)": 0.447737 + }, + { + "acc": 0.63649426, + "epoch": 0.9392440385591071, + "grad_norm": 5.65625, + "learning_rate": 5.91054867073484e-06, + "loss": 1.68743896, + "memory(GiB)": 111.15, + "step": 37025, + "train_speed(iter/s)": 0.447769 + }, + { + "acc": 0.64090424, + "epoch": 0.9393708777270421, + "grad_norm": 5.78125, + "learning_rate": 5.909517561899508e-06, + "loss": 1.67788544, + "memory(GiB)": 111.15, + "step": 37030, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.61976786, + "epoch": 0.9394977168949772, + "grad_norm": 5.25, + "learning_rate": 5.908486413059532e-06, + "loss": 1.67228699, + "memory(GiB)": 111.15, + "step": 37035, + "train_speed(iter/s)": 0.447833 + }, + { + "acc": 0.64659748, + "epoch": 0.9396245560629122, + "grad_norm": 4.3125, + "learning_rate": 5.907455224260268e-06, + "loss": 1.62812958, + "memory(GiB)": 111.15, + "step": 37040, + "train_speed(iter/s)": 0.447865 + }, + { + "acc": 0.64173508, + "epoch": 0.9397513952308473, + "grad_norm": 6.21875, + "learning_rate": 5.9064239955470704e-06, + "loss": 1.65689964, + "memory(GiB)": 111.15, + "step": 37045, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.62297859, + "epoch": 0.9398782343987824, + "grad_norm": 5.6875, + "learning_rate": 5.9053927269653e-06, + "loss": 1.74262123, + "memory(GiB)": 111.15, + "step": 37050, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.66084027, + "epoch": 0.9400050735667174, + "grad_norm": 5.0625, + "learning_rate": 5.904361418560314e-06, + "loss": 1.6167263, + "memory(GiB)": 111.15, + "step": 37055, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.6510972, + "epoch": 0.9401319127346525, + "grad_norm": 5.96875, + "learning_rate": 5.903330070377477e-06, + "loss": 1.65200901, + "memory(GiB)": 111.15, + "step": 37060, + "train_speed(iter/s)": 0.447992 + }, + { + "acc": 0.66018867, + "epoch": 0.9402587519025876, + "grad_norm": 6.0625, + "learning_rate": 5.902298682462147e-06, + "loss": 1.57438011, + "memory(GiB)": 111.15, + "step": 37065, + "train_speed(iter/s)": 0.448024 + }, + { + "acc": 0.65587759, + "epoch": 0.9403855910705226, + "grad_norm": 5.96875, + "learning_rate": 5.901267254859695e-06, + "loss": 1.63363934, + "memory(GiB)": 111.15, + "step": 37070, + "train_speed(iter/s)": 0.448056 + }, + { + "acc": 0.64581499, + "epoch": 0.9405124302384577, + "grad_norm": 4.5625, + "learning_rate": 5.900235787615485e-06, + "loss": 1.69133701, + "memory(GiB)": 111.15, + "step": 37075, + "train_speed(iter/s)": 0.448087 + }, + { + "acc": 0.65443301, + "epoch": 0.9406392694063926, + "grad_norm": 6.0, + "learning_rate": 5.8992042807748866e-06, + "loss": 1.60109081, + "memory(GiB)": 111.15, + "step": 37080, + "train_speed(iter/s)": 0.448119 + }, + { + "acc": 0.6487915, + "epoch": 0.9407661085743277, + "grad_norm": 5.5, + "learning_rate": 5.898172734383267e-06, + "loss": 1.66089115, + "memory(GiB)": 111.15, + "step": 37085, + "train_speed(iter/s)": 0.44815 + }, + { + "acc": 0.64047756, + "epoch": 0.9408929477422628, + "grad_norm": 5.8125, + "learning_rate": 5.897141148486003e-06, + "loss": 1.69680634, + "memory(GiB)": 111.15, + "step": 37090, + "train_speed(iter/s)": 0.448182 + }, + { + "acc": 0.65752144, + "epoch": 0.9410197869101978, + "grad_norm": 5.71875, + "learning_rate": 5.8961095231284645e-06, + "loss": 1.59454041, + "memory(GiB)": 111.15, + "step": 37095, + "train_speed(iter/s)": 0.448214 + }, + { + "acc": 0.63917894, + "epoch": 0.9411466260781329, + "grad_norm": 5.0, + "learning_rate": 5.895077858356029e-06, + "loss": 1.64568748, + "memory(GiB)": 111.15, + "step": 37100, + "train_speed(iter/s)": 0.448246 + }, + { + "epoch": 0.9411466260781329, + "eval_acc": 0.6432208689810697, + "eval_loss": 1.608633041381836, + "eval_runtime": 113.4219, + "eval_samples_per_second": 56.162, + "eval_steps_per_second": 28.081, + "step": 37100 + }, + { + "acc": 0.64294691, + "epoch": 0.941273465246068, + "grad_norm": 5.4375, + "learning_rate": 5.8940461542140725e-06, + "loss": 1.62280102, + "memory(GiB)": 111.15, + "step": 37105, + "train_speed(iter/s)": 0.44764 + }, + { + "acc": 0.66310158, + "epoch": 0.941400304414003, + "grad_norm": 6.09375, + "learning_rate": 5.893014410747975e-06, + "loss": 1.54547243, + "memory(GiB)": 111.15, + "step": 37110, + "train_speed(iter/s)": 0.447672 + }, + { + "acc": 0.66172028, + "epoch": 0.9415271435819381, + "grad_norm": 5.375, + "learning_rate": 5.891982628003114e-06, + "loss": 1.62750931, + "memory(GiB)": 111.15, + "step": 37115, + "train_speed(iter/s)": 0.447704 + }, + { + "acc": 0.64027243, + "epoch": 0.9416539827498731, + "grad_norm": 6.4375, + "learning_rate": 5.890950806024879e-06, + "loss": 1.62078838, + "memory(GiB)": 111.15, + "step": 37120, + "train_speed(iter/s)": 0.447736 + }, + { + "acc": 0.65207281, + "epoch": 0.9417808219178082, + "grad_norm": 5.6875, + "learning_rate": 5.889918944858647e-06, + "loss": 1.63681679, + "memory(GiB)": 111.15, + "step": 37125, + "train_speed(iter/s)": 0.447769 + }, + { + "acc": 0.66141176, + "epoch": 0.9419076610857433, + "grad_norm": 5.8125, + "learning_rate": 5.888887044549808e-06, + "loss": 1.58517666, + "memory(GiB)": 111.15, + "step": 37130, + "train_speed(iter/s)": 0.4478 + }, + { + "acc": 0.64079466, + "epoch": 0.9420345002536783, + "grad_norm": 5.40625, + "learning_rate": 5.887855105143746e-06, + "loss": 1.63401833, + "memory(GiB)": 111.15, + "step": 37135, + "train_speed(iter/s)": 0.447832 + }, + { + "acc": 0.65540824, + "epoch": 0.9421613394216134, + "grad_norm": 4.625, + "learning_rate": 5.886823126685855e-06, + "loss": 1.58768015, + "memory(GiB)": 111.15, + "step": 37140, + "train_speed(iter/s)": 0.447864 + }, + { + "acc": 0.65241961, + "epoch": 0.9422881785895485, + "grad_norm": 5.59375, + "learning_rate": 5.8857911092215214e-06, + "loss": 1.67613602, + "memory(GiB)": 111.15, + "step": 37145, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.65101447, + "epoch": 0.9424150177574835, + "grad_norm": 5.03125, + "learning_rate": 5.884759052796142e-06, + "loss": 1.68357468, + "memory(GiB)": 111.15, + "step": 37150, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.6553473, + "epoch": 0.9425418569254186, + "grad_norm": 5.03125, + "learning_rate": 5.883726957455108e-06, + "loss": 1.63154354, + "memory(GiB)": 111.15, + "step": 37155, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.63755636, + "epoch": 0.9426686960933536, + "grad_norm": 4.9375, + "learning_rate": 5.8826948232438176e-06, + "loss": 1.67211895, + "memory(GiB)": 111.15, + "step": 37160, + "train_speed(iter/s)": 0.447992 + }, + { + "acc": 0.65601282, + "epoch": 0.9427955352612887, + "grad_norm": 4.84375, + "learning_rate": 5.881662650207667e-06, + "loss": 1.56463623, + "memory(GiB)": 111.15, + "step": 37165, + "train_speed(iter/s)": 0.448024 + }, + { + "acc": 0.66890516, + "epoch": 0.9429223744292238, + "grad_norm": 6.1875, + "learning_rate": 5.880630438392057e-06, + "loss": 1.59788685, + "memory(GiB)": 111.15, + "step": 37170, + "train_speed(iter/s)": 0.448056 + }, + { + "acc": 0.6549171, + "epoch": 0.9430492135971588, + "grad_norm": 4.84375, + "learning_rate": 5.879598187842389e-06, + "loss": 1.57209244, + "memory(GiB)": 111.15, + "step": 37175, + "train_speed(iter/s)": 0.448088 + }, + { + "acc": 0.66600657, + "epoch": 0.9431760527650939, + "grad_norm": 5.71875, + "learning_rate": 5.878565898604066e-06, + "loss": 1.64994011, + "memory(GiB)": 111.15, + "step": 37180, + "train_speed(iter/s)": 0.44812 + }, + { + "acc": 0.65453196, + "epoch": 0.943302891933029, + "grad_norm": 4.84375, + "learning_rate": 5.87753357072249e-06, + "loss": 1.57580814, + "memory(GiB)": 111.15, + "step": 37185, + "train_speed(iter/s)": 0.448151 + }, + { + "acc": 0.65838079, + "epoch": 0.943429731100964, + "grad_norm": 5.84375, + "learning_rate": 5.876501204243072e-06, + "loss": 1.61551971, + "memory(GiB)": 111.15, + "step": 37190, + "train_speed(iter/s)": 0.448184 + }, + { + "acc": 0.65803652, + "epoch": 0.943556570268899, + "grad_norm": 6.6875, + "learning_rate": 5.875468799211217e-06, + "loss": 1.56397667, + "memory(GiB)": 111.15, + "step": 37195, + "train_speed(iter/s)": 0.448216 + }, + { + "acc": 0.64595647, + "epoch": 0.943683409436834, + "grad_norm": 4.59375, + "learning_rate": 5.874436355672337e-06, + "loss": 1.58461304, + "memory(GiB)": 111.15, + "step": 37200, + "train_speed(iter/s)": 0.448216 + }, + { + "epoch": 0.943683409436834, + "eval_acc": 0.6432551259004452, + "eval_loss": 1.6084177494049072, + "eval_runtime": 114.0291, + "eval_samples_per_second": 55.863, + "eval_steps_per_second": 27.931, + "step": 37200 + }, + { + "acc": 0.64850054, + "epoch": 0.9438102486047691, + "grad_norm": 5.375, + "learning_rate": 5.873403873671839e-06, + "loss": 1.60039082, + "memory(GiB)": 111.15, + "step": 37205, + "train_speed(iter/s)": 0.447608 + }, + { + "acc": 0.65184994, + "epoch": 0.9439370877727042, + "grad_norm": 5.9375, + "learning_rate": 5.872371353255142e-06, + "loss": 1.63956337, + "memory(GiB)": 111.15, + "step": 37210, + "train_speed(iter/s)": 0.44764 + }, + { + "acc": 0.66383591, + "epoch": 0.9440639269406392, + "grad_norm": 5.875, + "learning_rate": 5.871338794467656e-06, + "loss": 1.62747421, + "memory(GiB)": 111.15, + "step": 37215, + "train_speed(iter/s)": 0.447672 + }, + { + "acc": 0.65046492, + "epoch": 0.9441907661085743, + "grad_norm": 4.6875, + "learning_rate": 5.8703061973548e-06, + "loss": 1.5742487, + "memory(GiB)": 111.15, + "step": 37220, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.65264244, + "epoch": 0.9443176052765094, + "grad_norm": 4.125, + "learning_rate": 5.869273561961992e-06, + "loss": 1.57169285, + "memory(GiB)": 111.15, + "step": 37225, + "train_speed(iter/s)": 0.447735 + }, + { + "acc": 0.66504784, + "epoch": 0.9444444444444444, + "grad_norm": 5.46875, + "learning_rate": 5.8682408883346535e-06, + "loss": 1.61468124, + "memory(GiB)": 111.15, + "step": 37230, + "train_speed(iter/s)": 0.447767 + }, + { + "acc": 0.65034723, + "epoch": 0.9445712836123795, + "grad_norm": 5.53125, + "learning_rate": 5.867208176518202e-06, + "loss": 1.60625248, + "memory(GiB)": 111.15, + "step": 37235, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.66023254, + "epoch": 0.9446981227803145, + "grad_norm": 7.125, + "learning_rate": 5.866175426558064e-06, + "loss": 1.62073631, + "memory(GiB)": 111.15, + "step": 37240, + "train_speed(iter/s)": 0.44783 + }, + { + "acc": 0.64335985, + "epoch": 0.9448249619482496, + "grad_norm": 5.21875, + "learning_rate": 5.865142638499664e-06, + "loss": 1.73200417, + "memory(GiB)": 111.15, + "step": 37245, + "train_speed(iter/s)": 0.447863 + }, + { + "acc": 0.62721634, + "epoch": 0.9449518011161847, + "grad_norm": 5.1875, + "learning_rate": 5.864109812388426e-06, + "loss": 1.73281021, + "memory(GiB)": 111.15, + "step": 37250, + "train_speed(iter/s)": 0.447895 + }, + { + "acc": 0.64263673, + "epoch": 0.9450786402841197, + "grad_norm": 4.875, + "learning_rate": 5.863076948269782e-06, + "loss": 1.6032671, + "memory(GiB)": 111.15, + "step": 37255, + "train_speed(iter/s)": 0.447927 + }, + { + "acc": 0.65331535, + "epoch": 0.9452054794520548, + "grad_norm": 4.6875, + "learning_rate": 5.862044046189162e-06, + "loss": 1.64866333, + "memory(GiB)": 111.15, + "step": 37260, + "train_speed(iter/s)": 0.447958 + }, + { + "acc": 0.66519642, + "epoch": 0.9453323186199899, + "grad_norm": 5.28125, + "learning_rate": 5.8610111061919924e-06, + "loss": 1.61403656, + "memory(GiB)": 111.15, + "step": 37265, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.64394484, + "epoch": 0.9454591577879249, + "grad_norm": 5.90625, + "learning_rate": 5.859978128323713e-06, + "loss": 1.67553635, + "memory(GiB)": 111.15, + "step": 37270, + "train_speed(iter/s)": 0.448022 + }, + { + "acc": 0.65383644, + "epoch": 0.94558599695586, + "grad_norm": 5.375, + "learning_rate": 5.858945112629755e-06, + "loss": 1.58597021, + "memory(GiB)": 111.15, + "step": 37275, + "train_speed(iter/s)": 0.448054 + }, + { + "acc": 0.66297998, + "epoch": 0.945712836123795, + "grad_norm": 4.9375, + "learning_rate": 5.857912059155557e-06, + "loss": 1.60815315, + "memory(GiB)": 111.15, + "step": 37280, + "train_speed(iter/s)": 0.448085 + }, + { + "acc": 0.66370888, + "epoch": 0.9458396752917301, + "grad_norm": 4.875, + "learning_rate": 5.856878967946555e-06, + "loss": 1.58115072, + "memory(GiB)": 111.15, + "step": 37285, + "train_speed(iter/s)": 0.448117 + }, + { + "acc": 0.64213486, + "epoch": 0.9459665144596652, + "grad_norm": 6.03125, + "learning_rate": 5.855845839048191e-06, + "loss": 1.66096001, + "memory(GiB)": 111.15, + "step": 37290, + "train_speed(iter/s)": 0.448149 + }, + { + "acc": 0.65389915, + "epoch": 0.9460933536276002, + "grad_norm": 5.125, + "learning_rate": 5.854812672505906e-06, + "loss": 1.687006, + "memory(GiB)": 111.15, + "step": 37295, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.65251904, + "epoch": 0.9462201927955353, + "grad_norm": 5.59375, + "learning_rate": 5.853779468365144e-06, + "loss": 1.57423725, + "memory(GiB)": 111.15, + "step": 37300, + "train_speed(iter/s)": 0.448213 + }, + { + "epoch": 0.9462201927955353, + "eval_acc": 0.6432359086042102, + "eval_loss": 1.608532428741455, + "eval_runtime": 112.8205, + "eval_samples_per_second": 56.461, + "eval_steps_per_second": 28.231, + "step": 37300 + }, + { + "acc": 0.64296694, + "epoch": 0.9463470319634704, + "grad_norm": 4.9375, + "learning_rate": 5.852746226671348e-06, + "loss": 1.70132923, + "memory(GiB)": 111.15, + "step": 37305, + "train_speed(iter/s)": 0.447613 + }, + { + "acc": 0.64594803, + "epoch": 0.9464738711314054, + "grad_norm": 5.03125, + "learning_rate": 5.851712947469966e-06, + "loss": 1.61736641, + "memory(GiB)": 111.15, + "step": 37310, + "train_speed(iter/s)": 0.447645 + }, + { + "acc": 0.63651905, + "epoch": 0.9466007102993405, + "grad_norm": 5.25, + "learning_rate": 5.850679630806446e-06, + "loss": 1.66887665, + "memory(GiB)": 111.15, + "step": 37315, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.6551609, + "epoch": 0.9467275494672754, + "grad_norm": 4.8125, + "learning_rate": 5.849646276726237e-06, + "loss": 1.63950577, + "memory(GiB)": 111.15, + "step": 37320, + "train_speed(iter/s)": 0.447708 + }, + { + "acc": 0.64945221, + "epoch": 0.9468543886352105, + "grad_norm": 6.125, + "learning_rate": 5.848612885274792e-06, + "loss": 1.66113663, + "memory(GiB)": 111.15, + "step": 37325, + "train_speed(iter/s)": 0.447739 + }, + { + "acc": 0.63907881, + "epoch": 0.9469812278031456, + "grad_norm": 5.125, + "learning_rate": 5.847579456497564e-06, + "loss": 1.69568958, + "memory(GiB)": 111.15, + "step": 37330, + "train_speed(iter/s)": 0.447771 + }, + { + "acc": 0.64542961, + "epoch": 0.9471080669710806, + "grad_norm": 5.34375, + "learning_rate": 5.8465459904400065e-06, + "loss": 1.58881559, + "memory(GiB)": 111.15, + "step": 37335, + "train_speed(iter/s)": 0.447802 + }, + { + "acc": 0.66437607, + "epoch": 0.9472349061390157, + "grad_norm": 5.125, + "learning_rate": 5.845512487147579e-06, + "loss": 1.62925186, + "memory(GiB)": 111.15, + "step": 37340, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.64942589, + "epoch": 0.9473617453069508, + "grad_norm": 5.90625, + "learning_rate": 5.844478946665733e-06, + "loss": 1.64681473, + "memory(GiB)": 111.15, + "step": 37345, + "train_speed(iter/s)": 0.447865 + }, + { + "acc": 0.65807352, + "epoch": 0.9474885844748858, + "grad_norm": 4.84375, + "learning_rate": 5.843445369039937e-06, + "loss": 1.6812294, + "memory(GiB)": 111.15, + "step": 37350, + "train_speed(iter/s)": 0.447897 + }, + { + "acc": 0.66834626, + "epoch": 0.9476154236428209, + "grad_norm": 5.875, + "learning_rate": 5.842411754315645e-06, + "loss": 1.50647354, + "memory(GiB)": 111.15, + "step": 37355, + "train_speed(iter/s)": 0.447929 + }, + { + "acc": 0.64796972, + "epoch": 0.9477422628107559, + "grad_norm": 4.875, + "learning_rate": 5.841378102538324e-06, + "loss": 1.65653152, + "memory(GiB)": 111.15, + "step": 37360, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.65826821, + "epoch": 0.947869101978691, + "grad_norm": 6.0, + "learning_rate": 5.840344413753438e-06, + "loss": 1.57466497, + "memory(GiB)": 111.15, + "step": 37365, + "train_speed(iter/s)": 0.447992 + }, + { + "acc": 0.64652634, + "epoch": 0.9479959411466261, + "grad_norm": 4.6875, + "learning_rate": 5.8393106880064535e-06, + "loss": 1.6179718, + "memory(GiB)": 111.15, + "step": 37370, + "train_speed(iter/s)": 0.448024 + }, + { + "acc": 0.6470643, + "epoch": 0.9481227803145611, + "grad_norm": 5.3125, + "learning_rate": 5.838276925342836e-06, + "loss": 1.6014143, + "memory(GiB)": 111.15, + "step": 37375, + "train_speed(iter/s)": 0.448056 + }, + { + "acc": 0.65447021, + "epoch": 0.9482496194824962, + "grad_norm": 5.53125, + "learning_rate": 5.837243125808058e-06, + "loss": 1.60551376, + "memory(GiB)": 111.15, + "step": 37380, + "train_speed(iter/s)": 0.448088 + }, + { + "acc": 0.63523512, + "epoch": 0.9483764586504313, + "grad_norm": 6.1875, + "learning_rate": 5.8362092894475886e-06, + "loss": 1.72053928, + "memory(GiB)": 111.15, + "step": 37385, + "train_speed(iter/s)": 0.448119 + }, + { + "acc": 0.64642544, + "epoch": 0.9485032978183663, + "grad_norm": 4.9375, + "learning_rate": 5.835175416306901e-06, + "loss": 1.63617458, + "memory(GiB)": 111.15, + "step": 37390, + "train_speed(iter/s)": 0.448151 + }, + { + "acc": 0.65392818, + "epoch": 0.9486301369863014, + "grad_norm": 6.09375, + "learning_rate": 5.83414150643147e-06, + "loss": 1.58535118, + "memory(GiB)": 111.15, + "step": 37395, + "train_speed(iter/s)": 0.448183 + }, + { + "acc": 0.63319101, + "epoch": 0.9487569761542364, + "grad_norm": 5.5, + "learning_rate": 5.833107559866772e-06, + "loss": 1.52308331, + "memory(GiB)": 111.15, + "step": 37400, + "train_speed(iter/s)": 0.448214 + }, + { + "epoch": 0.9487569761542364, + "eval_acc": 0.6432563792023736, + "eval_loss": 1.6085642576217651, + "eval_runtime": 113.6443, + "eval_samples_per_second": 56.052, + "eval_steps_per_second": 28.026, + "step": 37400 + }, + { + "acc": 0.65363026, + "epoch": 0.9488838153221715, + "grad_norm": 6.28125, + "learning_rate": 5.832073576658282e-06, + "loss": 1.63073673, + "memory(GiB)": 111.15, + "step": 37405, + "train_speed(iter/s)": 0.447612 + }, + { + "acc": 0.65721283, + "epoch": 0.9490106544901066, + "grad_norm": 5.875, + "learning_rate": 5.831039556851485e-06, + "loss": 1.58220577, + "memory(GiB)": 111.15, + "step": 37410, + "train_speed(iter/s)": 0.447643 + }, + { + "acc": 0.64072185, + "epoch": 0.9491374936580416, + "grad_norm": 4.5625, + "learning_rate": 5.8300055004918535e-06, + "loss": 1.6418335, + "memory(GiB)": 111.15, + "step": 37415, + "train_speed(iter/s)": 0.447675 + }, + { + "acc": 0.64099765, + "epoch": 0.9492643328259767, + "grad_norm": 5.46875, + "learning_rate": 5.828971407624877e-06, + "loss": 1.68861732, + "memory(GiB)": 111.15, + "step": 37420, + "train_speed(iter/s)": 0.447707 + }, + { + "acc": 0.65273571, + "epoch": 0.9493911719939118, + "grad_norm": 5.03125, + "learning_rate": 5.827937278296037e-06, + "loss": 1.61013088, + "memory(GiB)": 111.15, + "step": 37425, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.66474538, + "epoch": 0.9495180111618468, + "grad_norm": 4.5625, + "learning_rate": 5.826903112550819e-06, + "loss": 1.61405697, + "memory(GiB)": 111.15, + "step": 37430, + "train_speed(iter/s)": 0.44777 + }, + { + "acc": 0.64697742, + "epoch": 0.9496448503297819, + "grad_norm": 5.6875, + "learning_rate": 5.825868910434708e-06, + "loss": 1.60982857, + "memory(GiB)": 111.15, + "step": 37435, + "train_speed(iter/s)": 0.447802 + }, + { + "acc": 0.67022667, + "epoch": 0.9497716894977168, + "grad_norm": 4.875, + "learning_rate": 5.824834671993197e-06, + "loss": 1.55824318, + "memory(GiB)": 111.15, + "step": 37440, + "train_speed(iter/s)": 0.447833 + }, + { + "acc": 0.64745646, + "epoch": 0.9498985286656519, + "grad_norm": 5.9375, + "learning_rate": 5.823800397271774e-06, + "loss": 1.6343441, + "memory(GiB)": 111.15, + "step": 37445, + "train_speed(iter/s)": 0.447865 + }, + { + "acc": 0.645643, + "epoch": 0.950025367833587, + "grad_norm": 6.5, + "learning_rate": 5.822766086315932e-06, + "loss": 1.6365612, + "memory(GiB)": 111.15, + "step": 37450, + "train_speed(iter/s)": 0.447897 + }, + { + "acc": 0.64971976, + "epoch": 0.950152207001522, + "grad_norm": 5.09375, + "learning_rate": 5.821731739171164e-06, + "loss": 1.63438225, + "memory(GiB)": 111.15, + "step": 37455, + "train_speed(iter/s)": 0.447929 + }, + { + "acc": 0.65813961, + "epoch": 0.9502790461694571, + "grad_norm": 5.4375, + "learning_rate": 5.820697355882965e-06, + "loss": 1.67674179, + "memory(GiB)": 111.15, + "step": 37460, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.6489347, + "epoch": 0.9504058853373922, + "grad_norm": 6.0625, + "learning_rate": 5.819662936496833e-06, + "loss": 1.64227295, + "memory(GiB)": 111.15, + "step": 37465, + "train_speed(iter/s)": 0.447992 + }, + { + "acc": 0.64165096, + "epoch": 0.9505327245053272, + "grad_norm": 4.46875, + "learning_rate": 5.818628481058265e-06, + "loss": 1.54955616, + "memory(GiB)": 111.15, + "step": 37470, + "train_speed(iter/s)": 0.448023 + }, + { + "acc": 0.63421674, + "epoch": 0.9506595636732623, + "grad_norm": 4.8125, + "learning_rate": 5.81759398961276e-06, + "loss": 1.64595718, + "memory(GiB)": 111.15, + "step": 37475, + "train_speed(iter/s)": 0.448055 + }, + { + "acc": 0.64826784, + "epoch": 0.9507864028411973, + "grad_norm": 4.6875, + "learning_rate": 5.816559462205824e-06, + "loss": 1.60083237, + "memory(GiB)": 111.15, + "step": 37480, + "train_speed(iter/s)": 0.448087 + }, + { + "acc": 0.65104685, + "epoch": 0.9509132420091324, + "grad_norm": 5.4375, + "learning_rate": 5.815524898882954e-06, + "loss": 1.58138018, + "memory(GiB)": 111.15, + "step": 37485, + "train_speed(iter/s)": 0.448119 + }, + { + "acc": 0.66581125, + "epoch": 0.9510400811770675, + "grad_norm": 6.0, + "learning_rate": 5.8144902996896615e-06, + "loss": 1.60206966, + "memory(GiB)": 111.15, + "step": 37490, + "train_speed(iter/s)": 0.44815 + }, + { + "acc": 0.64139528, + "epoch": 0.9511669203450025, + "grad_norm": 4.875, + "learning_rate": 5.813455664671446e-06, + "loss": 1.67632141, + "memory(GiB)": 111.15, + "step": 37495, + "train_speed(iter/s)": 0.448182 + }, + { + "acc": 0.66485844, + "epoch": 0.9512937595129376, + "grad_norm": 5.625, + "learning_rate": 5.812420993873819e-06, + "loss": 1.50900459, + "memory(GiB)": 111.15, + "step": 37500, + "train_speed(iter/s)": 0.448214 + }, + { + "epoch": 0.9512937595129376, + "eval_acc": 0.6432551259004452, + "eval_loss": 1.6083678007125854, + "eval_runtime": 114.0297, + "eval_samples_per_second": 55.863, + "eval_steps_per_second": 27.931, + "step": 37500 + }, + { + "acc": 0.65191164, + "epoch": 0.9514205986808727, + "grad_norm": 5.5, + "learning_rate": 5.81138628734229e-06, + "loss": 1.60928574, + "memory(GiB)": 111.15, + "step": 37505, + "train_speed(iter/s)": 0.447611 + }, + { + "acc": 0.65303807, + "epoch": 0.9515474378488077, + "grad_norm": 5.65625, + "learning_rate": 5.81035154512237e-06, + "loss": 1.58780251, + "memory(GiB)": 111.15, + "step": 37510, + "train_speed(iter/s)": 0.447643 + }, + { + "acc": 0.63737488, + "epoch": 0.9516742770167428, + "grad_norm": 5.5, + "learning_rate": 5.809316767259571e-06, + "loss": 1.6849823, + "memory(GiB)": 111.15, + "step": 37515, + "train_speed(iter/s)": 0.447674 + }, + { + "acc": 0.64170074, + "epoch": 0.9518011161846778, + "grad_norm": 5.21875, + "learning_rate": 5.808281953799408e-06, + "loss": 1.64985809, + "memory(GiB)": 111.15, + "step": 37520, + "train_speed(iter/s)": 0.447706 + }, + { + "acc": 0.63101082, + "epoch": 0.9519279553526129, + "grad_norm": 5.78125, + "learning_rate": 5.807247104787395e-06, + "loss": 1.65439949, + "memory(GiB)": 111.15, + "step": 37525, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.66345987, + "epoch": 0.952054794520548, + "grad_norm": 5.28125, + "learning_rate": 5.806212220269049e-06, + "loss": 1.63025227, + "memory(GiB)": 111.15, + "step": 37530, + "train_speed(iter/s)": 0.447769 + }, + { + "acc": 0.63604379, + "epoch": 0.952181633688483, + "grad_norm": 5.375, + "learning_rate": 5.805177300289891e-06, + "loss": 1.68179798, + "memory(GiB)": 111.15, + "step": 37535, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.63945336, + "epoch": 0.9523084728564181, + "grad_norm": 5.375, + "learning_rate": 5.804142344895441e-06, + "loss": 1.63861923, + "memory(GiB)": 111.15, + "step": 37540, + "train_speed(iter/s)": 0.447832 + }, + { + "acc": 0.65863495, + "epoch": 0.9524353120243532, + "grad_norm": 5.65625, + "learning_rate": 5.803107354131221e-06, + "loss": 1.60021076, + "memory(GiB)": 111.15, + "step": 37545, + "train_speed(iter/s)": 0.447864 + }, + { + "acc": 0.64391603, + "epoch": 0.9525621511922882, + "grad_norm": 5.15625, + "learning_rate": 5.802072328042753e-06, + "loss": 1.66389675, + "memory(GiB)": 111.15, + "step": 37550, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.64460244, + "epoch": 0.9526889903602233, + "grad_norm": 4.875, + "learning_rate": 5.8010372666755625e-06, + "loss": 1.66317406, + "memory(GiB)": 111.15, + "step": 37555, + "train_speed(iter/s)": 0.447927 + }, + { + "acc": 0.6444612, + "epoch": 0.9528158295281582, + "grad_norm": 5.875, + "learning_rate": 5.800002170075179e-06, + "loss": 1.67170563, + "memory(GiB)": 111.15, + "step": 37560, + "train_speed(iter/s)": 0.447959 + }, + { + "acc": 0.6701086, + "epoch": 0.9529426686960933, + "grad_norm": 5.03125, + "learning_rate": 5.798967038287125e-06, + "loss": 1.56292057, + "memory(GiB)": 111.15, + "step": 37565, + "train_speed(iter/s)": 0.447991 + }, + { + "acc": 0.65586348, + "epoch": 0.9530695078640284, + "grad_norm": 6.1875, + "learning_rate": 5.797931871356936e-06, + "loss": 1.57731657, + "memory(GiB)": 111.15, + "step": 37570, + "train_speed(iter/s)": 0.448023 + }, + { + "acc": 0.65106702, + "epoch": 0.9531963470319634, + "grad_norm": 5.28125, + "learning_rate": 5.796896669330139e-06, + "loss": 1.63664284, + "memory(GiB)": 111.15, + "step": 37575, + "train_speed(iter/s)": 0.448054 + }, + { + "acc": 0.63874702, + "epoch": 0.9533231861998985, + "grad_norm": 5.3125, + "learning_rate": 5.79586143225227e-06, + "loss": 1.71230946, + "memory(GiB)": 111.15, + "step": 37580, + "train_speed(iter/s)": 0.448086 + }, + { + "acc": 0.64593048, + "epoch": 0.9534500253678336, + "grad_norm": 5.9375, + "learning_rate": 5.79482616016886e-06, + "loss": 1.69622154, + "memory(GiB)": 111.15, + "step": 37585, + "train_speed(iter/s)": 0.448118 + }, + { + "acc": 0.64620123, + "epoch": 0.9535768645357686, + "grad_norm": 6.28125, + "learning_rate": 5.793790853125449e-06, + "loss": 1.6487114, + "memory(GiB)": 111.15, + "step": 37590, + "train_speed(iter/s)": 0.448149 + }, + { + "acc": 0.65041313, + "epoch": 0.9537037037037037, + "grad_norm": 5.71875, + "learning_rate": 5.792755511167572e-06, + "loss": 1.6690464, + "memory(GiB)": 111.15, + "step": 37595, + "train_speed(iter/s)": 0.448181 + }, + { + "acc": 0.64211826, + "epoch": 0.9538305428716387, + "grad_norm": 5.375, + "learning_rate": 5.7917201343407685e-06, + "loss": 1.63287888, + "memory(GiB)": 111.15, + "step": 37600, + "train_speed(iter/s)": 0.448213 + }, + { + "epoch": 0.9538305428716387, + "eval_acc": 0.6431707369039348, + "eval_loss": 1.6083132028579712, + "eval_runtime": 112.4992, + "eval_samples_per_second": 56.623, + "eval_steps_per_second": 28.311, + "step": 37600 + }, + { + "acc": 0.65244603, + "epoch": 0.9539573820395738, + "grad_norm": 5.71875, + "learning_rate": 5.790684722690577e-06, + "loss": 1.68373508, + "memory(GiB)": 111.15, + "step": 37605, + "train_speed(iter/s)": 0.44762 + }, + { + "acc": 0.65057707, + "epoch": 0.9540842212075089, + "grad_norm": 4.84375, + "learning_rate": 5.789649276262542e-06, + "loss": 1.69392872, + "memory(GiB)": 111.15, + "step": 37610, + "train_speed(iter/s)": 0.447651 + }, + { + "acc": 0.65457945, + "epoch": 0.9542110603754439, + "grad_norm": 6.03125, + "learning_rate": 5.788613795102207e-06, + "loss": 1.61412888, + "memory(GiB)": 111.15, + "step": 37615, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.65108728, + "epoch": 0.954337899543379, + "grad_norm": 5.0625, + "learning_rate": 5.787578279255116e-06, + "loss": 1.57943745, + "memory(GiB)": 111.15, + "step": 37620, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.6551774, + "epoch": 0.9544647387113141, + "grad_norm": 4.96875, + "learning_rate": 5.786542728766815e-06, + "loss": 1.61650352, + "memory(GiB)": 111.15, + "step": 37625, + "train_speed(iter/s)": 0.447746 + }, + { + "acc": 0.64639063, + "epoch": 0.9545915778792491, + "grad_norm": 8.0625, + "learning_rate": 5.785507143682856e-06, + "loss": 1.60042992, + "memory(GiB)": 111.15, + "step": 37630, + "train_speed(iter/s)": 0.447777 + }, + { + "acc": 0.62880182, + "epoch": 0.9547184170471842, + "grad_norm": 5.125, + "learning_rate": 5.784471524048782e-06, + "loss": 1.6733139, + "memory(GiB)": 111.15, + "step": 37635, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.65209284, + "epoch": 0.9548452562151192, + "grad_norm": 6.21875, + "learning_rate": 5.783435869910151e-06, + "loss": 1.63154678, + "memory(GiB)": 111.15, + "step": 37640, + "train_speed(iter/s)": 0.44784 + }, + { + "acc": 0.65268869, + "epoch": 0.9549720953830543, + "grad_norm": 4.5625, + "learning_rate": 5.782400181312511e-06, + "loss": 1.65504227, + "memory(GiB)": 111.15, + "step": 37645, + "train_speed(iter/s)": 0.447871 + }, + { + "acc": 0.65120478, + "epoch": 0.9550989345509894, + "grad_norm": 5.0625, + "learning_rate": 5.781364458301419e-06, + "loss": 1.58749828, + "memory(GiB)": 111.15, + "step": 37650, + "train_speed(iter/s)": 0.447902 + }, + { + "acc": 0.6569912, + "epoch": 0.9552257737189244, + "grad_norm": 4.84375, + "learning_rate": 5.780328700922427e-06, + "loss": 1.65730877, + "memory(GiB)": 111.15, + "step": 37655, + "train_speed(iter/s)": 0.447934 + }, + { + "acc": 0.66978798, + "epoch": 0.9553526128868595, + "grad_norm": 5.125, + "learning_rate": 5.779292909221097e-06, + "loss": 1.62211533, + "memory(GiB)": 111.15, + "step": 37660, + "train_speed(iter/s)": 0.447965 + }, + { + "acc": 0.65640888, + "epoch": 0.9554794520547946, + "grad_norm": 5.3125, + "learning_rate": 5.778257083242986e-06, + "loss": 1.6011692, + "memory(GiB)": 111.15, + "step": 37665, + "train_speed(iter/s)": 0.447997 + }, + { + "acc": 0.6435955, + "epoch": 0.9556062912227296, + "grad_norm": 4.65625, + "learning_rate": 5.777221223033653e-06, + "loss": 1.65108833, + "memory(GiB)": 111.15, + "step": 37670, + "train_speed(iter/s)": 0.448028 + }, + { + "acc": 0.64209127, + "epoch": 0.9557331303906647, + "grad_norm": 4.3125, + "learning_rate": 5.77618532863866e-06, + "loss": 1.61971474, + "memory(GiB)": 111.15, + "step": 37675, + "train_speed(iter/s)": 0.448059 + }, + { + "acc": 0.66960335, + "epoch": 0.9558599695585996, + "grad_norm": 6.1875, + "learning_rate": 5.775149400103572e-06, + "loss": 1.50451078, + "memory(GiB)": 111.15, + "step": 37680, + "train_speed(iter/s)": 0.44809 + }, + { + "acc": 0.65117331, + "epoch": 0.9559868087265347, + "grad_norm": 5.3125, + "learning_rate": 5.774113437473953e-06, + "loss": 1.60331707, + "memory(GiB)": 111.15, + "step": 37685, + "train_speed(iter/s)": 0.448122 + }, + { + "acc": 0.64285212, + "epoch": 0.9561136478944698, + "grad_norm": 4.5625, + "learning_rate": 5.7730774407953675e-06, + "loss": 1.58822603, + "memory(GiB)": 111.15, + "step": 37690, + "train_speed(iter/s)": 0.448153 + }, + { + "acc": 0.65596104, + "epoch": 0.9562404870624048, + "grad_norm": 6.15625, + "learning_rate": 5.772041410113384e-06, + "loss": 1.59798965, + "memory(GiB)": 111.15, + "step": 37695, + "train_speed(iter/s)": 0.448185 + }, + { + "acc": 0.64053297, + "epoch": 0.9563673262303399, + "grad_norm": 5.0625, + "learning_rate": 5.771005345473575e-06, + "loss": 1.61455936, + "memory(GiB)": 111.15, + "step": 37700, + "train_speed(iter/s)": 0.448217 + }, + { + "epoch": 0.9563673262303399, + "eval_acc": 0.6431866120616943, + "eval_loss": 1.6084281206130981, + "eval_runtime": 112.9423, + "eval_samples_per_second": 56.401, + "eval_steps_per_second": 28.2, + "step": 37700 + }, + { + "acc": 0.65484352, + "epoch": 0.956494165398275, + "grad_norm": 4.8125, + "learning_rate": 5.769969246921505e-06, + "loss": 1.57739544, + "memory(GiB)": 111.15, + "step": 37705, + "train_speed(iter/s)": 0.447623 + }, + { + "acc": 0.64098244, + "epoch": 0.95662100456621, + "grad_norm": 5.09375, + "learning_rate": 5.768933114502753e-06, + "loss": 1.65980415, + "memory(GiB)": 111.15, + "step": 37710, + "train_speed(iter/s)": 0.447654 + }, + { + "acc": 0.6573844, + "epoch": 0.9567478437341451, + "grad_norm": 5.875, + "learning_rate": 5.7678969482628875e-06, + "loss": 1.61390705, + "memory(GiB)": 111.15, + "step": 37715, + "train_speed(iter/s)": 0.447686 + }, + { + "acc": 0.65145321, + "epoch": 0.9568746829020801, + "grad_norm": 6.4375, + "learning_rate": 5.766860748247488e-06, + "loss": 1.57390394, + "memory(GiB)": 111.15, + "step": 37720, + "train_speed(iter/s)": 0.447717 + }, + { + "acc": 0.63683453, + "epoch": 0.9570015220700152, + "grad_norm": 4.84375, + "learning_rate": 5.765824514502126e-06, + "loss": 1.65293827, + "memory(GiB)": 111.15, + "step": 37725, + "train_speed(iter/s)": 0.447748 + }, + { + "acc": 0.67228537, + "epoch": 0.9571283612379503, + "grad_norm": 8.75, + "learning_rate": 5.7647882470723846e-06, + "loss": 1.59528179, + "memory(GiB)": 111.15, + "step": 37730, + "train_speed(iter/s)": 0.447779 + }, + { + "acc": 0.67130551, + "epoch": 0.9572552004058853, + "grad_norm": 6.5, + "learning_rate": 5.763751946003842e-06, + "loss": 1.57221489, + "memory(GiB)": 111.15, + "step": 37735, + "train_speed(iter/s)": 0.447811 + }, + { + "acc": 0.64345775, + "epoch": 0.9573820395738204, + "grad_norm": 5.03125, + "learning_rate": 5.7627156113420775e-06, + "loss": 1.58576717, + "memory(GiB)": 111.15, + "step": 37740, + "train_speed(iter/s)": 0.447842 + }, + { + "acc": 0.64523726, + "epoch": 0.9575088787417555, + "grad_norm": 5.5625, + "learning_rate": 5.761679243132677e-06, + "loss": 1.61721535, + "memory(GiB)": 111.15, + "step": 37745, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.63670321, + "epoch": 0.9576357179096905, + "grad_norm": 5.3125, + "learning_rate": 5.760642841421222e-06, + "loss": 1.64926376, + "memory(GiB)": 111.15, + "step": 37750, + "train_speed(iter/s)": 0.447905 + }, + { + "acc": 0.66271596, + "epoch": 0.9577625570776256, + "grad_norm": 5.8125, + "learning_rate": 5.759606406253299e-06, + "loss": 1.56552401, + "memory(GiB)": 111.15, + "step": 37755, + "train_speed(iter/s)": 0.447936 + }, + { + "acc": 0.65387115, + "epoch": 0.9578893962455606, + "grad_norm": 4.84375, + "learning_rate": 5.758569937674494e-06, + "loss": 1.56112843, + "memory(GiB)": 111.15, + "step": 37760, + "train_speed(iter/s)": 0.447968 + }, + { + "acc": 0.65759916, + "epoch": 0.9580162354134957, + "grad_norm": 5.75, + "learning_rate": 5.7575334357303954e-06, + "loss": 1.5678112, + "memory(GiB)": 111.15, + "step": 37765, + "train_speed(iter/s)": 0.447999 + }, + { + "acc": 0.65374923, + "epoch": 0.9581430745814308, + "grad_norm": 5.125, + "learning_rate": 5.756496900466596e-06, + "loss": 1.60888119, + "memory(GiB)": 111.15, + "step": 37770, + "train_speed(iter/s)": 0.44803 + }, + { + "acc": 0.65898232, + "epoch": 0.9582699137493658, + "grad_norm": 5.09375, + "learning_rate": 5.755460331928684e-06, + "loss": 1.60830364, + "memory(GiB)": 111.15, + "step": 37775, + "train_speed(iter/s)": 0.448062 + }, + { + "acc": 0.64398746, + "epoch": 0.9583967529173009, + "grad_norm": 5.625, + "learning_rate": 5.754423730162257e-06, + "loss": 1.66227531, + "memory(GiB)": 111.15, + "step": 37780, + "train_speed(iter/s)": 0.448093 + }, + { + "acc": 0.64547644, + "epoch": 0.958523592085236, + "grad_norm": 4.71875, + "learning_rate": 5.753387095212901e-06, + "loss": 1.61079979, + "memory(GiB)": 111.15, + "step": 37785, + "train_speed(iter/s)": 0.448124 + }, + { + "acc": 0.64972725, + "epoch": 0.958650431253171, + "grad_norm": 6.5, + "learning_rate": 5.752350427126221e-06, + "loss": 1.6558136, + "memory(GiB)": 111.15, + "step": 37790, + "train_speed(iter/s)": 0.448155 + }, + { + "acc": 0.65403852, + "epoch": 0.958777270421106, + "grad_norm": 4.84375, + "learning_rate": 5.751313725947808e-06, + "loss": 1.56705313, + "memory(GiB)": 111.15, + "step": 37795, + "train_speed(iter/s)": 0.448186 + }, + { + "acc": 0.66735349, + "epoch": 0.958904109589041, + "grad_norm": 5.59375, + "learning_rate": 5.7502769917232635e-06, + "loss": 1.57006321, + "memory(GiB)": 111.15, + "step": 37800, + "train_speed(iter/s)": 0.448218 + }, + { + "epoch": 0.958904109589041, + "eval_acc": 0.6431456708653673, + "eval_loss": 1.6084568500518799, + "eval_runtime": 113.2669, + "eval_samples_per_second": 56.239, + "eval_steps_per_second": 28.119, + "step": 37800 + }, + { + "acc": 0.64457612, + "epoch": 0.9590309487569761, + "grad_norm": 4.75, + "learning_rate": 5.7492402244981885e-06, + "loss": 1.66820545, + "memory(GiB)": 111.15, + "step": 37805, + "train_speed(iter/s)": 0.447624 + }, + { + "acc": 0.64308238, + "epoch": 0.9591577879249112, + "grad_norm": 5.75, + "learning_rate": 5.748203424318182e-06, + "loss": 1.63351116, + "memory(GiB)": 111.15, + "step": 37810, + "train_speed(iter/s)": 0.447655 + }, + { + "acc": 0.67181911, + "epoch": 0.9592846270928462, + "grad_norm": 5.4375, + "learning_rate": 5.747166591228849e-06, + "loss": 1.55307083, + "memory(GiB)": 111.15, + "step": 37815, + "train_speed(iter/s)": 0.447686 + }, + { + "acc": 0.65797386, + "epoch": 0.9594114662607813, + "grad_norm": 5.1875, + "learning_rate": 5.746129725275793e-06, + "loss": 1.62620716, + "memory(GiB)": 111.15, + "step": 37820, + "train_speed(iter/s)": 0.447717 + }, + { + "acc": 0.64643707, + "epoch": 0.9595383054287164, + "grad_norm": 5.40625, + "learning_rate": 5.74509282650462e-06, + "loss": 1.67104721, + "memory(GiB)": 111.15, + "step": 37825, + "train_speed(iter/s)": 0.447748 + }, + { + "acc": 0.6356842, + "epoch": 0.9596651445966514, + "grad_norm": 5.15625, + "learning_rate": 5.744055894960938e-06, + "loss": 1.63086128, + "memory(GiB)": 111.15, + "step": 37830, + "train_speed(iter/s)": 0.447779 + }, + { + "acc": 0.63393965, + "epoch": 0.9597919837645865, + "grad_norm": 5.84375, + "learning_rate": 5.743018930690357e-06, + "loss": 1.71520615, + "memory(GiB)": 111.15, + "step": 37835, + "train_speed(iter/s)": 0.447811 + }, + { + "acc": 0.64578772, + "epoch": 0.9599188229325215, + "grad_norm": 5.3125, + "learning_rate": 5.7419819337384855e-06, + "loss": 1.64941654, + "memory(GiB)": 111.15, + "step": 37840, + "train_speed(iter/s)": 0.447842 + }, + { + "acc": 0.6364687, + "epoch": 0.9600456621004566, + "grad_norm": 5.5, + "learning_rate": 5.740944904150934e-06, + "loss": 1.62561798, + "memory(GiB)": 111.15, + "step": 37845, + "train_speed(iter/s)": 0.447873 + }, + { + "acc": 0.6527379, + "epoch": 0.9601725012683917, + "grad_norm": 5.0625, + "learning_rate": 5.739907841973321e-06, + "loss": 1.59018822, + "memory(GiB)": 111.15, + "step": 37850, + "train_speed(iter/s)": 0.447904 + }, + { + "acc": 0.64697666, + "epoch": 0.9602993404363267, + "grad_norm": 5.9375, + "learning_rate": 5.738870747251255e-06, + "loss": 1.6493721, + "memory(GiB)": 111.15, + "step": 37855, + "train_speed(iter/s)": 0.447935 + }, + { + "acc": 0.63894358, + "epoch": 0.9604261796042618, + "grad_norm": 5.46875, + "learning_rate": 5.737833620030357e-06, + "loss": 1.63304424, + "memory(GiB)": 111.15, + "step": 37860, + "train_speed(iter/s)": 0.447967 + }, + { + "acc": 0.67704525, + "epoch": 0.9605530187721969, + "grad_norm": 4.84375, + "learning_rate": 5.7367964603562385e-06, + "loss": 1.51600857, + "memory(GiB)": 111.15, + "step": 37865, + "train_speed(iter/s)": 0.447998 + }, + { + "acc": 0.6659936, + "epoch": 0.9606798579401319, + "grad_norm": 4.96875, + "learning_rate": 5.7357592682745245e-06, + "loss": 1.62924862, + "memory(GiB)": 111.15, + "step": 37870, + "train_speed(iter/s)": 0.448029 + }, + { + "acc": 0.65637217, + "epoch": 0.960806697108067, + "grad_norm": 6.125, + "learning_rate": 5.734722043830833e-06, + "loss": 1.52634678, + "memory(GiB)": 111.15, + "step": 37875, + "train_speed(iter/s)": 0.44806 + }, + { + "acc": 0.64979467, + "epoch": 0.960933536276002, + "grad_norm": 5.875, + "learning_rate": 5.7336847870707855e-06, + "loss": 1.66275921, + "memory(GiB)": 111.15, + "step": 37880, + "train_speed(iter/s)": 0.448092 + }, + { + "acc": 0.644804, + "epoch": 0.9610603754439371, + "grad_norm": 5.1875, + "learning_rate": 5.732647498040006e-06, + "loss": 1.62733688, + "memory(GiB)": 111.15, + "step": 37885, + "train_speed(iter/s)": 0.448123 + }, + { + "acc": 0.64404159, + "epoch": 0.9611872146118722, + "grad_norm": 5.59375, + "learning_rate": 5.731610176784118e-06, + "loss": 1.63733921, + "memory(GiB)": 111.15, + "step": 37890, + "train_speed(iter/s)": 0.448154 + }, + { + "acc": 0.64561753, + "epoch": 0.9613140537798072, + "grad_norm": 5.03125, + "learning_rate": 5.730572823348748e-06, + "loss": 1.59093256, + "memory(GiB)": 111.15, + "step": 37895, + "train_speed(iter/s)": 0.448186 + }, + { + "acc": 0.64796133, + "epoch": 0.9614408929477423, + "grad_norm": 5.78125, + "learning_rate": 5.729535437779523e-06, + "loss": 1.64088974, + "memory(GiB)": 111.15, + "step": 37900, + "train_speed(iter/s)": 0.448216 + }, + { + "epoch": 0.9614408929477423, + "eval_acc": 0.6432710010582046, + "eval_loss": 1.608338475227356, + "eval_runtime": 113.7654, + "eval_samples_per_second": 55.992, + "eval_steps_per_second": 27.996, + "step": 37900 + }, + { + "acc": 0.63411975, + "epoch": 0.9615677321156774, + "grad_norm": 4.96875, + "learning_rate": 5.728498020122073e-06, + "loss": 1.66930275, + "memory(GiB)": 111.15, + "step": 37905, + "train_speed(iter/s)": 0.447621 + }, + { + "acc": 0.65075417, + "epoch": 0.9616945712836124, + "grad_norm": 5.65625, + "learning_rate": 5.727460570422028e-06, + "loss": 1.59226904, + "memory(GiB)": 111.15, + "step": 37910, + "train_speed(iter/s)": 0.447652 + }, + { + "acc": 0.66885576, + "epoch": 0.9618214104515475, + "grad_norm": 5.78125, + "learning_rate": 5.726423088725017e-06, + "loss": 1.59313049, + "memory(GiB)": 111.15, + "step": 37915, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.64210749, + "epoch": 0.9619482496194824, + "grad_norm": 4.53125, + "learning_rate": 5.725385575076677e-06, + "loss": 1.61188774, + "memory(GiB)": 111.15, + "step": 37920, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.63773499, + "epoch": 0.9620750887874175, + "grad_norm": 5.5, + "learning_rate": 5.7243480295226405e-06, + "loss": 1.64668884, + "memory(GiB)": 111.15, + "step": 37925, + "train_speed(iter/s)": 0.447745 + }, + { + "acc": 0.64736824, + "epoch": 0.9622019279553526, + "grad_norm": 5.40625, + "learning_rate": 5.723310452108545e-06, + "loss": 1.63032036, + "memory(GiB)": 111.15, + "step": 37930, + "train_speed(iter/s)": 0.447777 + }, + { + "acc": 0.63594503, + "epoch": 0.9623287671232876, + "grad_norm": 4.8125, + "learning_rate": 5.722272842880023e-06, + "loss": 1.69734325, + "memory(GiB)": 111.15, + "step": 37935, + "train_speed(iter/s)": 0.447807 + }, + { + "acc": 0.65085282, + "epoch": 0.9624556062912227, + "grad_norm": 5.09375, + "learning_rate": 5.7212352018827215e-06, + "loss": 1.6286375, + "memory(GiB)": 111.15, + "step": 37940, + "train_speed(iter/s)": 0.447838 + }, + { + "acc": 0.65648508, + "epoch": 0.9625824454591578, + "grad_norm": 5.25, + "learning_rate": 5.720197529162272e-06, + "loss": 1.6239233, + "memory(GiB)": 111.15, + "step": 37945, + "train_speed(iter/s)": 0.44787 + }, + { + "acc": 0.65311384, + "epoch": 0.9627092846270928, + "grad_norm": 4.34375, + "learning_rate": 5.719159824764321e-06, + "loss": 1.5598772, + "memory(GiB)": 111.15, + "step": 37950, + "train_speed(iter/s)": 0.447901 + }, + { + "acc": 0.65618963, + "epoch": 0.9628361237950279, + "grad_norm": 5.0, + "learning_rate": 5.71812208873451e-06, + "loss": 1.64736748, + "memory(GiB)": 111.15, + "step": 37955, + "train_speed(iter/s)": 0.447932 + }, + { + "acc": 0.65138025, + "epoch": 0.9629629629629629, + "grad_norm": 5.21875, + "learning_rate": 5.717084321118482e-06, + "loss": 1.57500248, + "memory(GiB)": 111.15, + "step": 37960, + "train_speed(iter/s)": 0.447963 + }, + { + "acc": 0.65213313, + "epoch": 0.963089802130898, + "grad_norm": 6.0, + "learning_rate": 5.716046521961887e-06, + "loss": 1.67080421, + "memory(GiB)": 111.15, + "step": 37965, + "train_speed(iter/s)": 0.447994 + }, + { + "acc": 0.65621161, + "epoch": 0.9632166412988331, + "grad_norm": 6.5625, + "learning_rate": 5.715008691310366e-06, + "loss": 1.64005356, + "memory(GiB)": 111.15, + "step": 37970, + "train_speed(iter/s)": 0.448025 + }, + { + "acc": 0.63909087, + "epoch": 0.9633434804667681, + "grad_norm": 6.15625, + "learning_rate": 5.713970829209573e-06, + "loss": 1.71783047, + "memory(GiB)": 111.15, + "step": 37975, + "train_speed(iter/s)": 0.448056 + }, + { + "acc": 0.64700327, + "epoch": 0.9634703196347032, + "grad_norm": 5.25, + "learning_rate": 5.712932935705153e-06, + "loss": 1.66772385, + "memory(GiB)": 111.15, + "step": 37980, + "train_speed(iter/s)": 0.448087 + }, + { + "acc": 0.63302727, + "epoch": 0.9635971588026383, + "grad_norm": 5.375, + "learning_rate": 5.711895010842762e-06, + "loss": 1.70298977, + "memory(GiB)": 111.15, + "step": 37985, + "train_speed(iter/s)": 0.448119 + }, + { + "acc": 0.65632386, + "epoch": 0.9637239979705733, + "grad_norm": 5.34375, + "learning_rate": 5.710857054668048e-06, + "loss": 1.66115227, + "memory(GiB)": 111.15, + "step": 37990, + "train_speed(iter/s)": 0.44815 + }, + { + "acc": 0.65235395, + "epoch": 0.9638508371385084, + "grad_norm": 5.65625, + "learning_rate": 5.7098190672266675e-06, + "loss": 1.62149544, + "memory(GiB)": 111.15, + "step": 37995, + "train_speed(iter/s)": 0.448182 + }, + { + "acc": 0.64726439, + "epoch": 0.9639776763064434, + "grad_norm": 5.5, + "learning_rate": 5.708781048564276e-06, + "loss": 1.6140337, + "memory(GiB)": 111.15, + "step": 38000, + "train_speed(iter/s)": 0.448213 + }, + { + "epoch": 0.9639776763064434, + "eval_acc": 0.6433211331353395, + "eval_loss": 1.6082912683486938, + "eval_runtime": 112.9303, + "eval_samples_per_second": 56.406, + "eval_steps_per_second": 28.203, + "step": 38000 + }, + { + "acc": 0.65357647, + "epoch": 0.9641045154743785, + "grad_norm": 7.59375, + "learning_rate": 5.707742998726527e-06, + "loss": 1.60278187, + "memory(GiB)": 111.15, + "step": 38005, + "train_speed(iter/s)": 0.447624 + }, + { + "acc": 0.64809303, + "epoch": 0.9642313546423136, + "grad_norm": 6.3125, + "learning_rate": 5.706704917759085e-06, + "loss": 1.56668577, + "memory(GiB)": 111.15, + "step": 38010, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.66595564, + "epoch": 0.9643581938102486, + "grad_norm": 4.9375, + "learning_rate": 5.705666805707603e-06, + "loss": 1.56591253, + "memory(GiB)": 111.15, + "step": 38015, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.63984847, + "epoch": 0.9644850329781837, + "grad_norm": 4.6875, + "learning_rate": 5.704628662617744e-06, + "loss": 1.5984396, + "memory(GiB)": 111.15, + "step": 38020, + "train_speed(iter/s)": 0.447718 + }, + { + "acc": 0.66344433, + "epoch": 0.9646118721461188, + "grad_norm": 5.1875, + "learning_rate": 5.703590488535171e-06, + "loss": 1.61009331, + "memory(GiB)": 111.15, + "step": 38025, + "train_speed(iter/s)": 0.44775 + }, + { + "acc": 0.65580888, + "epoch": 0.9647387113140538, + "grad_norm": 5.75, + "learning_rate": 5.702552283505548e-06, + "loss": 1.62099152, + "memory(GiB)": 111.15, + "step": 38030, + "train_speed(iter/s)": 0.447781 + }, + { + "acc": 0.66327682, + "epoch": 0.9648655504819889, + "grad_norm": 4.78125, + "learning_rate": 5.7015140475745376e-06, + "loss": 1.57617931, + "memory(GiB)": 111.15, + "step": 38035, + "train_speed(iter/s)": 0.447812 + }, + { + "acc": 0.65850315, + "epoch": 0.9649923896499238, + "grad_norm": 5.53125, + "learning_rate": 5.700475780787809e-06, + "loss": 1.5859251, + "memory(GiB)": 111.15, + "step": 38040, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.64547172, + "epoch": 0.9651192288178589, + "grad_norm": 5.09375, + "learning_rate": 5.699437483191027e-06, + "loss": 1.5923214, + "memory(GiB)": 111.15, + "step": 38045, + "train_speed(iter/s)": 0.447875 + }, + { + "acc": 0.64711199, + "epoch": 0.965246067985794, + "grad_norm": 5.34375, + "learning_rate": 5.6983991548298615e-06, + "loss": 1.61994801, + "memory(GiB)": 111.15, + "step": 38050, + "train_speed(iter/s)": 0.447906 + }, + { + "acc": 0.65806456, + "epoch": 0.965372907153729, + "grad_norm": 6.59375, + "learning_rate": 5.697360795749983e-06, + "loss": 1.61230316, + "memory(GiB)": 111.15, + "step": 38055, + "train_speed(iter/s)": 0.447938 + }, + { + "acc": 0.64435802, + "epoch": 0.9654997463216641, + "grad_norm": 4.96875, + "learning_rate": 5.696322405997064e-06, + "loss": 1.65254631, + "memory(GiB)": 111.15, + "step": 38060, + "train_speed(iter/s)": 0.447969 + }, + { + "acc": 0.65069556, + "epoch": 0.9656265854895992, + "grad_norm": 6.15625, + "learning_rate": 5.695283985616775e-06, + "loss": 1.64032631, + "memory(GiB)": 111.15, + "step": 38065, + "train_speed(iter/s)": 0.448 + }, + { + "acc": 0.6406249, + "epoch": 0.9657534246575342, + "grad_norm": 4.8125, + "learning_rate": 5.694245534654795e-06, + "loss": 1.66803761, + "memory(GiB)": 111.15, + "step": 38070, + "train_speed(iter/s)": 0.448032 + }, + { + "acc": 0.65066829, + "epoch": 0.9658802638254693, + "grad_norm": 6.28125, + "learning_rate": 5.693207053156794e-06, + "loss": 1.48808556, + "memory(GiB)": 111.15, + "step": 38075, + "train_speed(iter/s)": 0.448063 + }, + { + "acc": 0.65130844, + "epoch": 0.9660071029934043, + "grad_norm": 4.28125, + "learning_rate": 5.692168541168455e-06, + "loss": 1.61089764, + "memory(GiB)": 111.15, + "step": 38080, + "train_speed(iter/s)": 0.448095 + }, + { + "acc": 0.66354756, + "epoch": 0.9661339421613394, + "grad_norm": 4.6875, + "learning_rate": 5.691129998735449e-06, + "loss": 1.52607975, + "memory(GiB)": 111.15, + "step": 38085, + "train_speed(iter/s)": 0.448126 + }, + { + "acc": 0.64947119, + "epoch": 0.9662607813292745, + "grad_norm": 5.4375, + "learning_rate": 5.690091425903464e-06, + "loss": 1.6870163, + "memory(GiB)": 111.15, + "step": 38090, + "train_speed(iter/s)": 0.448157 + }, + { + "acc": 0.64694004, + "epoch": 0.9663876204972095, + "grad_norm": 5.8125, + "learning_rate": 5.689052822718175e-06, + "loss": 1.65096626, + "memory(GiB)": 111.15, + "step": 38095, + "train_speed(iter/s)": 0.448189 + }, + { + "acc": 0.66464472, + "epoch": 0.9665144596651446, + "grad_norm": 6.09375, + "learning_rate": 5.688014189225266e-06, + "loss": 1.54420137, + "memory(GiB)": 111.15, + "step": 38100, + "train_speed(iter/s)": 0.44822 + }, + { + "epoch": 0.9665144596651446, + "eval_acc": 0.6432354908369007, + "eval_loss": 1.6082361936569214, + "eval_runtime": 113.7129, + "eval_samples_per_second": 56.018, + "eval_steps_per_second": 28.009, + "step": 38100 + }, + { + "acc": 0.64822855, + "epoch": 0.9666412988330797, + "grad_norm": 5.875, + "learning_rate": 5.686975525470423e-06, + "loss": 1.67433529, + "memory(GiB)": 111.15, + "step": 38105, + "train_speed(iter/s)": 0.447628 + }, + { + "acc": 0.65658216, + "epoch": 0.9667681380010147, + "grad_norm": 5.53125, + "learning_rate": 5.685936831499328e-06, + "loss": 1.64664288, + "memory(GiB)": 111.15, + "step": 38110, + "train_speed(iter/s)": 0.447658 + }, + { + "acc": 0.64668617, + "epoch": 0.9668949771689498, + "grad_norm": 5.4375, + "learning_rate": 5.684898107357669e-06, + "loss": 1.64589958, + "memory(GiB)": 111.15, + "step": 38115, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.6414104, + "epoch": 0.9670218163368848, + "grad_norm": 5.34375, + "learning_rate": 5.683859353091133e-06, + "loss": 1.71410179, + "memory(GiB)": 111.15, + "step": 38120, + "train_speed(iter/s)": 0.44772 + }, + { + "acc": 0.63821044, + "epoch": 0.9671486555048199, + "grad_norm": 5.0625, + "learning_rate": 5.6828205687454094e-06, + "loss": 1.6794075, + "memory(GiB)": 111.15, + "step": 38125, + "train_speed(iter/s)": 0.447751 + }, + { + "acc": 0.63549957, + "epoch": 0.967275494672755, + "grad_norm": 5.09375, + "learning_rate": 5.68178175436619e-06, + "loss": 1.70381966, + "memory(GiB)": 111.15, + "step": 38130, + "train_speed(iter/s)": 0.447781 + }, + { + "acc": 0.65297294, + "epoch": 0.96740233384069, + "grad_norm": 5.21875, + "learning_rate": 5.680742909999163e-06, + "loss": 1.65964355, + "memory(GiB)": 111.15, + "step": 38135, + "train_speed(iter/s)": 0.447812 + }, + { + "acc": 0.62622771, + "epoch": 0.9675291730086251, + "grad_norm": 6.53125, + "learning_rate": 5.679704035690026e-06, + "loss": 1.74048843, + "memory(GiB)": 111.15, + "step": 38140, + "train_speed(iter/s)": 0.447843 + }, + { + "acc": 0.65116901, + "epoch": 0.9676560121765602, + "grad_norm": 4.96875, + "learning_rate": 5.6786651314844675e-06, + "loss": 1.56747389, + "memory(GiB)": 111.15, + "step": 38145, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.64638405, + "epoch": 0.9677828513444952, + "grad_norm": 5.71875, + "learning_rate": 5.67762619742819e-06, + "loss": 1.631563, + "memory(GiB)": 111.15, + "step": 38150, + "train_speed(iter/s)": 0.447905 + }, + { + "acc": 0.63896847, + "epoch": 0.9679096905124303, + "grad_norm": 5.53125, + "learning_rate": 5.676587233566885e-06, + "loss": 1.64555531, + "memory(GiB)": 111.15, + "step": 38155, + "train_speed(iter/s)": 0.447936 + }, + { + "acc": 0.64798641, + "epoch": 0.9680365296803652, + "grad_norm": 5.0, + "learning_rate": 5.675548239946254e-06, + "loss": 1.6007452, + "memory(GiB)": 111.15, + "step": 38160, + "train_speed(iter/s)": 0.447967 + }, + { + "acc": 0.63559103, + "epoch": 0.9681633688483003, + "grad_norm": 5.84375, + "learning_rate": 5.674509216611993e-06, + "loss": 1.69154396, + "memory(GiB)": 111.15, + "step": 38165, + "train_speed(iter/s)": 0.447998 + }, + { + "acc": 0.65323105, + "epoch": 0.9682902080162354, + "grad_norm": 5.3125, + "learning_rate": 5.673470163609806e-06, + "loss": 1.58584919, + "memory(GiB)": 111.15, + "step": 38170, + "train_speed(iter/s)": 0.448028 + }, + { + "acc": 0.66741929, + "epoch": 0.9684170471841704, + "grad_norm": 4.59375, + "learning_rate": 5.672431080985395e-06, + "loss": 1.52258396, + "memory(GiB)": 111.15, + "step": 38175, + "train_speed(iter/s)": 0.448059 + }, + { + "acc": 0.64816027, + "epoch": 0.9685438863521055, + "grad_norm": 5.75, + "learning_rate": 5.671391968784464e-06, + "loss": 1.58730564, + "memory(GiB)": 111.15, + "step": 38180, + "train_speed(iter/s)": 0.44809 + }, + { + "acc": 0.65758715, + "epoch": 0.9686707255200406, + "grad_norm": 5.90625, + "learning_rate": 5.670352827052715e-06, + "loss": 1.60149441, + "memory(GiB)": 111.15, + "step": 38185, + "train_speed(iter/s)": 0.448121 + }, + { + "acc": 0.6614171, + "epoch": 0.9687975646879756, + "grad_norm": 5.3125, + "learning_rate": 5.6693136558358565e-06, + "loss": 1.63580875, + "memory(GiB)": 111.15, + "step": 38190, + "train_speed(iter/s)": 0.448152 + }, + { + "acc": 0.64279342, + "epoch": 0.9689244038559107, + "grad_norm": 5.21875, + "learning_rate": 5.668274455179595e-06, + "loss": 1.64372406, + "memory(GiB)": 111.15, + "step": 38195, + "train_speed(iter/s)": 0.448183 + }, + { + "acc": 0.62958226, + "epoch": 0.9690512430238457, + "grad_norm": 5.59375, + "learning_rate": 5.667235225129639e-06, + "loss": 1.68812637, + "memory(GiB)": 111.15, + "step": 38200, + "train_speed(iter/s)": 0.448214 + }, + { + "epoch": 0.9690512430238457, + "eval_acc": 0.6431744968097199, + "eval_loss": 1.6083592176437378, + "eval_runtime": 113.0849, + "eval_samples_per_second": 56.329, + "eval_steps_per_second": 28.165, + "step": 38200 + }, + { + "acc": 0.63212872, + "epoch": 0.9691780821917808, + "grad_norm": 6.78125, + "learning_rate": 5.6661959657317e-06, + "loss": 1.64503784, + "memory(GiB)": 111.15, + "step": 38205, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.66840734, + "epoch": 0.9693049213597159, + "grad_norm": 6.53125, + "learning_rate": 5.665156677031487e-06, + "loss": 1.57848606, + "memory(GiB)": 111.15, + "step": 38210, + "train_speed(iter/s)": 0.447657 + }, + { + "acc": 0.63696737, + "epoch": 0.9694317605276509, + "grad_norm": 5.9375, + "learning_rate": 5.664117359074712e-06, + "loss": 1.69384956, + "memory(GiB)": 111.15, + "step": 38215, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.64122763, + "epoch": 0.969558599695586, + "grad_norm": 5.875, + "learning_rate": 5.6630780119070935e-06, + "loss": 1.67860794, + "memory(GiB)": 111.15, + "step": 38220, + "train_speed(iter/s)": 0.447719 + }, + { + "acc": 0.64971094, + "epoch": 0.9696854388635211, + "grad_norm": 5.34375, + "learning_rate": 5.6620386355743415e-06, + "loss": 1.65264244, + "memory(GiB)": 111.15, + "step": 38225, + "train_speed(iter/s)": 0.44775 + }, + { + "acc": 0.64493017, + "epoch": 0.9698122780314561, + "grad_norm": 5.625, + "learning_rate": 5.660999230122177e-06, + "loss": 1.59779367, + "memory(GiB)": 111.15, + "step": 38230, + "train_speed(iter/s)": 0.447781 + }, + { + "acc": 0.64599018, + "epoch": 0.9699391171993912, + "grad_norm": 5.59375, + "learning_rate": 5.659959795596313e-06, + "loss": 1.64453773, + "memory(GiB)": 111.15, + "step": 38235, + "train_speed(iter/s)": 0.447812 + }, + { + "acc": 0.63827357, + "epoch": 0.9700659563673262, + "grad_norm": 5.4375, + "learning_rate": 5.65892033204247e-06, + "loss": 1.66556206, + "memory(GiB)": 111.15, + "step": 38240, + "train_speed(iter/s)": 0.447843 + }, + { + "acc": 0.66577673, + "epoch": 0.9701927955352613, + "grad_norm": 4.71875, + "learning_rate": 5.657880839506371e-06, + "loss": 1.56589594, + "memory(GiB)": 111.15, + "step": 38245, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.6513279, + "epoch": 0.9703196347031964, + "grad_norm": 7.4375, + "learning_rate": 5.656841318033735e-06, + "loss": 1.61522331, + "memory(GiB)": 111.15, + "step": 38250, + "train_speed(iter/s)": 0.447905 + }, + { + "acc": 0.64937515, + "epoch": 0.9704464738711314, + "grad_norm": 6.125, + "learning_rate": 5.6558017676702846e-06, + "loss": 1.60584507, + "memory(GiB)": 111.15, + "step": 38255, + "train_speed(iter/s)": 0.447935 + }, + { + "acc": 0.64983883, + "epoch": 0.9705733130390665, + "grad_norm": 5.375, + "learning_rate": 5.654762188461744e-06, + "loss": 1.6544239, + "memory(GiB)": 111.15, + "step": 38260, + "train_speed(iter/s)": 0.447966 + }, + { + "acc": 0.64256611, + "epoch": 0.9707001522070016, + "grad_norm": 5.25, + "learning_rate": 5.653722580453841e-06, + "loss": 1.65569916, + "memory(GiB)": 111.15, + "step": 38265, + "train_speed(iter/s)": 0.447998 + }, + { + "acc": 0.63728442, + "epoch": 0.9708269913749366, + "grad_norm": 5.28125, + "learning_rate": 5.652682943692299e-06, + "loss": 1.68643074, + "memory(GiB)": 111.15, + "step": 38270, + "train_speed(iter/s)": 0.448028 + }, + { + "acc": 0.65716839, + "epoch": 0.9709538305428717, + "grad_norm": 6.875, + "learning_rate": 5.651643278222847e-06, + "loss": 1.63960342, + "memory(GiB)": 111.15, + "step": 38275, + "train_speed(iter/s)": 0.44806 + }, + { + "acc": 0.66863909, + "epoch": 0.9710806697108066, + "grad_norm": 5.40625, + "learning_rate": 5.6506035840912145e-06, + "loss": 1.59280872, + "memory(GiB)": 111.15, + "step": 38280, + "train_speed(iter/s)": 0.448091 + }, + { + "acc": 0.66119976, + "epoch": 0.9712075088787417, + "grad_norm": 5.15625, + "learning_rate": 5.649563861343131e-06, + "loss": 1.61689701, + "memory(GiB)": 111.15, + "step": 38285, + "train_speed(iter/s)": 0.448121 + }, + { + "acc": 0.63051305, + "epoch": 0.9713343480466768, + "grad_norm": 5.78125, + "learning_rate": 5.648524110024331e-06, + "loss": 1.63550186, + "memory(GiB)": 111.15, + "step": 38290, + "train_speed(iter/s)": 0.448152 + }, + { + "acc": 0.65839138, + "epoch": 0.9714611872146118, + "grad_norm": 5.0625, + "learning_rate": 5.647484330180542e-06, + "loss": 1.6121748, + "memory(GiB)": 111.15, + "step": 38295, + "train_speed(iter/s)": 0.448183 + }, + { + "acc": 0.65086751, + "epoch": 0.9715880263825469, + "grad_norm": 6.4375, + "learning_rate": 5.646444521857504e-06, + "loss": 1.65338478, + "memory(GiB)": 111.15, + "step": 38300, + "train_speed(iter/s)": 0.448214 + }, + { + "epoch": 0.9715880263825469, + "eval_acc": 0.6432593035735398, + "eval_loss": 1.6082494258880615, + "eval_runtime": 114.0532, + "eval_samples_per_second": 55.851, + "eval_steps_per_second": 27.926, + "step": 38300 + }, + { + "acc": 0.66711283, + "epoch": 0.971714865550482, + "grad_norm": 5.125, + "learning_rate": 5.645404685100948e-06, + "loss": 1.55881481, + "memory(GiB)": 111.15, + "step": 38305, + "train_speed(iter/s)": 0.447624 + }, + { + "acc": 0.65604935, + "epoch": 0.971841704718417, + "grad_norm": 5.5625, + "learning_rate": 5.644364819956613e-06, + "loss": 1.6389492, + "memory(GiB)": 111.15, + "step": 38310, + "train_speed(iter/s)": 0.447655 + }, + { + "acc": 0.64508533, + "epoch": 0.9719685438863521, + "grad_norm": 5.84375, + "learning_rate": 5.643324926470236e-06, + "loss": 1.64507999, + "memory(GiB)": 111.15, + "step": 38315, + "train_speed(iter/s)": 0.447686 + }, + { + "acc": 0.65343857, + "epoch": 0.9720953830542871, + "grad_norm": 6.1875, + "learning_rate": 5.642285004687557e-06, + "loss": 1.62866573, + "memory(GiB)": 111.15, + "step": 38320, + "train_speed(iter/s)": 0.447716 + }, + { + "acc": 0.66259871, + "epoch": 0.9722222222222222, + "grad_norm": 5.53125, + "learning_rate": 5.6412450546543165e-06, + "loss": 1.58839121, + "memory(GiB)": 111.15, + "step": 38325, + "train_speed(iter/s)": 0.447747 + }, + { + "acc": 0.64490886, + "epoch": 0.9723490613901573, + "grad_norm": 5.53125, + "learning_rate": 5.640205076416254e-06, + "loss": 1.74140224, + "memory(GiB)": 111.15, + "step": 38330, + "train_speed(iter/s)": 0.447778 + }, + { + "acc": 0.64434161, + "epoch": 0.9724759005580923, + "grad_norm": 4.84375, + "learning_rate": 5.639165070019116e-06, + "loss": 1.63730507, + "memory(GiB)": 111.15, + "step": 38335, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.64621811, + "epoch": 0.9726027397260274, + "grad_norm": 6.09375, + "learning_rate": 5.638125035508642e-06, + "loss": 1.6301239, + "memory(GiB)": 111.15, + "step": 38340, + "train_speed(iter/s)": 0.447839 + }, + { + "acc": 0.65778227, + "epoch": 0.9727295788939625, + "grad_norm": 4.25, + "learning_rate": 5.6370849729305825e-06, + "loss": 1.59651327, + "memory(GiB)": 111.15, + "step": 38345, + "train_speed(iter/s)": 0.44787 + }, + { + "acc": 0.62804413, + "epoch": 0.9728564180618975, + "grad_norm": 5.15625, + "learning_rate": 5.63604488233068e-06, + "loss": 1.7055727, + "memory(GiB)": 111.15, + "step": 38350, + "train_speed(iter/s)": 0.447901 + }, + { + "acc": 0.65539923, + "epoch": 0.9729832572298326, + "grad_norm": 5.90625, + "learning_rate": 5.635004763754683e-06, + "loss": 1.58288097, + "memory(GiB)": 111.15, + "step": 38355, + "train_speed(iter/s)": 0.447931 + }, + { + "acc": 0.64381843, + "epoch": 0.9731100963977676, + "grad_norm": 6.09375, + "learning_rate": 5.633964617248345e-06, + "loss": 1.65393906, + "memory(GiB)": 111.15, + "step": 38360, + "train_speed(iter/s)": 0.447962 + }, + { + "acc": 0.66399236, + "epoch": 0.9732369355657027, + "grad_norm": 4.75, + "learning_rate": 5.6329244428574085e-06, + "loss": 1.56526489, + "memory(GiB)": 111.15, + "step": 38365, + "train_speed(iter/s)": 0.447992 + }, + { + "acc": 0.64809623, + "epoch": 0.9733637747336378, + "grad_norm": 5.09375, + "learning_rate": 5.631884240627632e-06, + "loss": 1.62166939, + "memory(GiB)": 111.15, + "step": 38370, + "train_speed(iter/s)": 0.448023 + }, + { + "acc": 0.64708586, + "epoch": 0.9734906139015728, + "grad_norm": 4.59375, + "learning_rate": 5.6308440106047634e-06, + "loss": 1.68124371, + "memory(GiB)": 111.15, + "step": 38375, + "train_speed(iter/s)": 0.448054 + }, + { + "acc": 0.65787563, + "epoch": 0.9736174530695079, + "grad_norm": 5.25, + "learning_rate": 5.62980375283456e-06, + "loss": 1.59150095, + "memory(GiB)": 111.15, + "step": 38380, + "train_speed(iter/s)": 0.448084 + }, + { + "acc": 0.65531349, + "epoch": 0.973744292237443, + "grad_norm": 5.71875, + "learning_rate": 5.628763467362775e-06, + "loss": 1.60065918, + "memory(GiB)": 111.15, + "step": 38385, + "train_speed(iter/s)": 0.448115 + }, + { + "acc": 0.63937249, + "epoch": 0.973871131405378, + "grad_norm": 5.28125, + "learning_rate": 5.627723154235165e-06, + "loss": 1.64038506, + "memory(GiB)": 111.15, + "step": 38390, + "train_speed(iter/s)": 0.448146 + }, + { + "acc": 0.64645486, + "epoch": 0.973997970573313, + "grad_norm": 5.28125, + "learning_rate": 5.62668281349749e-06, + "loss": 1.68599854, + "memory(GiB)": 111.15, + "step": 38395, + "train_speed(iter/s)": 0.448177 + }, + { + "acc": 0.63866363, + "epoch": 0.974124809741248, + "grad_norm": 5.6875, + "learning_rate": 5.625642445195505e-06, + "loss": 1.71497192, + "memory(GiB)": 111.15, + "step": 38400, + "train_speed(iter/s)": 0.448208 + }, + { + "epoch": 0.974124809741248, + "eval_acc": 0.6432158557733563, + "eval_loss": 1.6083251237869263, + "eval_runtime": 114.1445, + "eval_samples_per_second": 55.806, + "eval_steps_per_second": 27.903, + "step": 38400 + }, + { + "acc": 0.66302452, + "epoch": 0.9742516489091831, + "grad_norm": 5.78125, + "learning_rate": 5.6246020493749735e-06, + "loss": 1.6190033, + "memory(GiB)": 111.15, + "step": 38405, + "train_speed(iter/s)": 0.447619 + }, + { + "acc": 0.65698891, + "epoch": 0.9743784880771182, + "grad_norm": 5.90625, + "learning_rate": 5.623561626081654e-06, + "loss": 1.63814468, + "memory(GiB)": 111.15, + "step": 38410, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.65272455, + "epoch": 0.9745053272450532, + "grad_norm": 5.4375, + "learning_rate": 5.622521175361311e-06, + "loss": 1.56358852, + "memory(GiB)": 111.15, + "step": 38415, + "train_speed(iter/s)": 0.44768 + }, + { + "acc": 0.66045895, + "epoch": 0.9746321664129883, + "grad_norm": 5.21875, + "learning_rate": 5.621480697259707e-06, + "loss": 1.69206181, + "memory(GiB)": 111.15, + "step": 38420, + "train_speed(iter/s)": 0.447711 + }, + { + "acc": 0.63472724, + "epoch": 0.9747590055809234, + "grad_norm": 5.09375, + "learning_rate": 5.620440191822607e-06, + "loss": 1.64831696, + "memory(GiB)": 111.15, + "step": 38425, + "train_speed(iter/s)": 0.447742 + }, + { + "acc": 0.63210855, + "epoch": 0.9748858447488584, + "grad_norm": 4.875, + "learning_rate": 5.619399659095778e-06, + "loss": 1.65324116, + "memory(GiB)": 111.15, + "step": 38430, + "train_speed(iter/s)": 0.447773 + }, + { + "acc": 0.6514327, + "epoch": 0.9750126839167935, + "grad_norm": 6.25, + "learning_rate": 5.618359099124985e-06, + "loss": 1.66416512, + "memory(GiB)": 111.15, + "step": 38435, + "train_speed(iter/s)": 0.447804 + }, + { + "acc": 0.64643173, + "epoch": 0.9751395230847285, + "grad_norm": 4.90625, + "learning_rate": 5.617318511956001e-06, + "loss": 1.68979759, + "memory(GiB)": 111.15, + "step": 38440, + "train_speed(iter/s)": 0.447835 + }, + { + "acc": 0.64708109, + "epoch": 0.9752663622526636, + "grad_norm": 6.09375, + "learning_rate": 5.61627789763459e-06, + "loss": 1.64605503, + "memory(GiB)": 111.15, + "step": 38445, + "train_speed(iter/s)": 0.447866 + }, + { + "acc": 0.65224552, + "epoch": 0.9753932014205987, + "grad_norm": 5.21875, + "learning_rate": 5.6152372562065275e-06, + "loss": 1.59103374, + "memory(GiB)": 111.15, + "step": 38450, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.64316478, + "epoch": 0.9755200405885337, + "grad_norm": 4.90625, + "learning_rate": 5.614196587717581e-06, + "loss": 1.65290623, + "memory(GiB)": 111.15, + "step": 38455, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.66292963, + "epoch": 0.9756468797564688, + "grad_norm": 4.25, + "learning_rate": 5.613155892213529e-06, + "loss": 1.59761868, + "memory(GiB)": 111.15, + "step": 38460, + "train_speed(iter/s)": 0.447959 + }, + { + "acc": 0.66566935, + "epoch": 0.9757737189244039, + "grad_norm": 4.9375, + "learning_rate": 5.612115169740142e-06, + "loss": 1.57923679, + "memory(GiB)": 111.15, + "step": 38465, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.66416254, + "epoch": 0.9759005580923389, + "grad_norm": 5.59375, + "learning_rate": 5.611074420343197e-06, + "loss": 1.60197506, + "memory(GiB)": 111.15, + "step": 38470, + "train_speed(iter/s)": 0.448021 + }, + { + "acc": 0.66842446, + "epoch": 0.976027397260274, + "grad_norm": 5.21875, + "learning_rate": 5.610033644068471e-06, + "loss": 1.62266502, + "memory(GiB)": 111.15, + "step": 38475, + "train_speed(iter/s)": 0.448052 + }, + { + "acc": 0.66403122, + "epoch": 0.976154236428209, + "grad_norm": 4.40625, + "learning_rate": 5.608992840961742e-06, + "loss": 1.55293427, + "memory(GiB)": 111.15, + "step": 38480, + "train_speed(iter/s)": 0.448083 + }, + { + "acc": 0.64567842, + "epoch": 0.9762810755961441, + "grad_norm": 5.3125, + "learning_rate": 5.6079520110687876e-06, + "loss": 1.6122797, + "memory(GiB)": 111.15, + "step": 38485, + "train_speed(iter/s)": 0.448114 + }, + { + "acc": 0.65812778, + "epoch": 0.9764079147640792, + "grad_norm": 5.1875, + "learning_rate": 5.606911154435392e-06, + "loss": 1.60118866, + "memory(GiB)": 111.15, + "step": 38490, + "train_speed(iter/s)": 0.448146 + }, + { + "acc": 0.65274453, + "epoch": 0.9765347539320142, + "grad_norm": 6.84375, + "learning_rate": 5.605870271107332e-06, + "loss": 1.62341156, + "memory(GiB)": 111.15, + "step": 38495, + "train_speed(iter/s)": 0.448177 + }, + { + "acc": 0.64583621, + "epoch": 0.9766615930999493, + "grad_norm": 5.375, + "learning_rate": 5.6048293611303925e-06, + "loss": 1.64106369, + "memory(GiB)": 111.15, + "step": 38500, + "train_speed(iter/s)": 0.448207 + }, + { + "epoch": 0.9766615930999493, + "eval_acc": 0.6432367441388291, + "eval_loss": 1.6081515550613403, + "eval_runtime": 115.3055, + "eval_samples_per_second": 55.245, + "eval_steps_per_second": 27.622, + "step": 38500 + }, + { + "acc": 0.67334781, + "epoch": 0.9767884322678844, + "grad_norm": 4.5, + "learning_rate": 5.603788424550357e-06, + "loss": 1.61199169, + "memory(GiB)": 111.15, + "step": 38505, + "train_speed(iter/s)": 0.447613 + }, + { + "acc": 0.65221825, + "epoch": 0.9769152714358194, + "grad_norm": 4.5625, + "learning_rate": 5.602747461413014e-06, + "loss": 1.57742996, + "memory(GiB)": 111.15, + "step": 38510, + "train_speed(iter/s)": 0.447644 + }, + { + "acc": 0.65716662, + "epoch": 0.9770421106037545, + "grad_norm": 6.59375, + "learning_rate": 5.6017064717641435e-06, + "loss": 1.52217093, + "memory(GiB)": 111.15, + "step": 38515, + "train_speed(iter/s)": 0.447674 + }, + { + "acc": 0.65146527, + "epoch": 0.9771689497716894, + "grad_norm": 4.96875, + "learning_rate": 5.600665455649538e-06, + "loss": 1.63876972, + "memory(GiB)": 111.15, + "step": 38520, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.64918518, + "epoch": 0.9772957889396245, + "grad_norm": 5.28125, + "learning_rate": 5.599624413114981e-06, + "loss": 1.64157982, + "memory(GiB)": 111.15, + "step": 38525, + "train_speed(iter/s)": 0.447736 + }, + { + "acc": 0.65078211, + "epoch": 0.9774226281075596, + "grad_norm": 4.96875, + "learning_rate": 5.5985833442062676e-06, + "loss": 1.63556976, + "memory(GiB)": 111.15, + "step": 38530, + "train_speed(iter/s)": 0.447766 + }, + { + "acc": 0.66401796, + "epoch": 0.9775494672754946, + "grad_norm": 5.15625, + "learning_rate": 5.597542248969185e-06, + "loss": 1.52866001, + "memory(GiB)": 111.15, + "step": 38535, + "train_speed(iter/s)": 0.447797 + }, + { + "acc": 0.64037709, + "epoch": 0.9776763064434297, + "grad_norm": 5.78125, + "learning_rate": 5.596501127449527e-06, + "loss": 1.64272041, + "memory(GiB)": 111.15, + "step": 38540, + "train_speed(iter/s)": 0.447828 + }, + { + "acc": 0.65261059, + "epoch": 0.9778031456113648, + "grad_norm": 5.34375, + "learning_rate": 5.595459979693086e-06, + "loss": 1.61309071, + "memory(GiB)": 111.15, + "step": 38545, + "train_speed(iter/s)": 0.447859 + }, + { + "acc": 0.65917225, + "epoch": 0.9779299847792998, + "grad_norm": 5.5, + "learning_rate": 5.594418805745657e-06, + "loss": 1.60150414, + "memory(GiB)": 111.15, + "step": 38550, + "train_speed(iter/s)": 0.44789 + }, + { + "acc": 0.64531221, + "epoch": 0.9780568239472349, + "grad_norm": 5.625, + "learning_rate": 5.593377605653035e-06, + "loss": 1.60327721, + "memory(GiB)": 111.15, + "step": 38555, + "train_speed(iter/s)": 0.447921 + }, + { + "acc": 0.65925307, + "epoch": 0.9781836631151699, + "grad_norm": 5.59375, + "learning_rate": 5.592336379461018e-06, + "loss": 1.59987087, + "memory(GiB)": 111.15, + "step": 38560, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.64645205, + "epoch": 0.978310502283105, + "grad_norm": 5.15625, + "learning_rate": 5.5912951272154004e-06, + "loss": 1.58303204, + "memory(GiB)": 111.15, + "step": 38565, + "train_speed(iter/s)": 0.447983 + }, + { + "acc": 0.64635758, + "epoch": 0.9784373414510401, + "grad_norm": 4.8125, + "learning_rate": 5.590253848961984e-06, + "loss": 1.68574295, + "memory(GiB)": 111.15, + "step": 38570, + "train_speed(iter/s)": 0.448014 + }, + { + "acc": 0.63941536, + "epoch": 0.9785641806189751, + "grad_norm": 4.75, + "learning_rate": 5.589212544746566e-06, + "loss": 1.66617508, + "memory(GiB)": 111.15, + "step": 38575, + "train_speed(iter/s)": 0.448045 + }, + { + "acc": 0.65980906, + "epoch": 0.9786910197869102, + "grad_norm": 6.0625, + "learning_rate": 5.588171214614953e-06, + "loss": 1.62895775, + "memory(GiB)": 111.15, + "step": 38580, + "train_speed(iter/s)": 0.448076 + }, + { + "acc": 0.64566207, + "epoch": 0.9788178589548453, + "grad_norm": 4.59375, + "learning_rate": 5.587129858612941e-06, + "loss": 1.63394909, + "memory(GiB)": 111.15, + "step": 38585, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.64351454, + "epoch": 0.9789446981227803, + "grad_norm": 4.4375, + "learning_rate": 5.586088476786339e-06, + "loss": 1.63288803, + "memory(GiB)": 111.15, + "step": 38590, + "train_speed(iter/s)": 0.448138 + }, + { + "acc": 0.63664474, + "epoch": 0.9790715372907154, + "grad_norm": 5.375, + "learning_rate": 5.585047069180947e-06, + "loss": 1.69305496, + "memory(GiB)": 111.15, + "step": 38595, + "train_speed(iter/s)": 0.448169 + }, + { + "acc": 0.65332947, + "epoch": 0.9791983764586504, + "grad_norm": 6.0, + "learning_rate": 5.5840056358425755e-06, + "loss": 1.6180687, + "memory(GiB)": 111.15, + "step": 38600, + "train_speed(iter/s)": 0.4482 + }, + { + "epoch": 0.9791983764586504, + "eval_acc": 0.6432609746427776, + "eval_loss": 1.6082512140274048, + "eval_runtime": 113.7472, + "eval_samples_per_second": 56.001, + "eval_steps_per_second": 28.001, + "step": 38600 + }, + { + "acc": 0.63092561, + "epoch": 0.9793252156265855, + "grad_norm": 5.59375, + "learning_rate": 5.582964176817025e-06, + "loss": 1.70142708, + "memory(GiB)": 111.15, + "step": 38605, + "train_speed(iter/s)": 0.447616 + }, + { + "acc": 0.6661396, + "epoch": 0.9794520547945206, + "grad_norm": 5.90625, + "learning_rate": 5.58192269215011e-06, + "loss": 1.59254217, + "memory(GiB)": 111.15, + "step": 38610, + "train_speed(iter/s)": 0.447646 + }, + { + "acc": 0.66099234, + "epoch": 0.9795788939624556, + "grad_norm": 5.28125, + "learning_rate": 5.580881181887636e-06, + "loss": 1.60646076, + "memory(GiB)": 111.15, + "step": 38615, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.64767895, + "epoch": 0.9797057331303907, + "grad_norm": 4.6875, + "learning_rate": 5.579839646075414e-06, + "loss": 1.61990166, + "memory(GiB)": 111.15, + "step": 38620, + "train_speed(iter/s)": 0.447708 + }, + { + "acc": 0.64067574, + "epoch": 0.9798325722983258, + "grad_norm": 4.28125, + "learning_rate": 5.578798084759257e-06, + "loss": 1.65990906, + "memory(GiB)": 111.15, + "step": 38625, + "train_speed(iter/s)": 0.447739 + }, + { + "acc": 0.65499086, + "epoch": 0.9799594114662608, + "grad_norm": 5.90625, + "learning_rate": 5.577756497984975e-06, + "loss": 1.56445961, + "memory(GiB)": 111.15, + "step": 38630, + "train_speed(iter/s)": 0.447769 + }, + { + "acc": 0.6522963, + "epoch": 0.9800862506341959, + "grad_norm": 5.0, + "learning_rate": 5.576714885798382e-06, + "loss": 1.60522461, + "memory(GiB)": 111.15, + "step": 38635, + "train_speed(iter/s)": 0.4478 + }, + { + "acc": 0.65742149, + "epoch": 0.9802130898021308, + "grad_norm": 6.03125, + "learning_rate": 5.575673248245295e-06, + "loss": 1.59437904, + "memory(GiB)": 111.15, + "step": 38640, + "train_speed(iter/s)": 0.447831 + }, + { + "acc": 0.63906622, + "epoch": 0.9803399289700659, + "grad_norm": 7.78125, + "learning_rate": 5.574631585371527e-06, + "loss": 1.72446709, + "memory(GiB)": 111.15, + "step": 38645, + "train_speed(iter/s)": 0.447861 + }, + { + "acc": 0.64251556, + "epoch": 0.980466768138001, + "grad_norm": 5.8125, + "learning_rate": 5.573589897222897e-06, + "loss": 1.60679359, + "memory(GiB)": 111.15, + "step": 38650, + "train_speed(iter/s)": 0.447892 + }, + { + "acc": 0.65387106, + "epoch": 0.980593607305936, + "grad_norm": 4.75, + "learning_rate": 5.572548183845222e-06, + "loss": 1.57363567, + "memory(GiB)": 111.15, + "step": 38655, + "train_speed(iter/s)": 0.447923 + }, + { + "acc": 0.64696064, + "epoch": 0.9807204464738711, + "grad_norm": 4.6875, + "learning_rate": 5.571506445284322e-06, + "loss": 1.64330444, + "memory(GiB)": 111.15, + "step": 38660, + "train_speed(iter/s)": 0.447954 + }, + { + "acc": 0.66184907, + "epoch": 0.9808472856418062, + "grad_norm": 4.34375, + "learning_rate": 5.570464681586017e-06, + "loss": 1.61074657, + "memory(GiB)": 111.15, + "step": 38665, + "train_speed(iter/s)": 0.447984 + }, + { + "acc": 0.64535456, + "epoch": 0.9809741248097412, + "grad_norm": 5.25, + "learning_rate": 5.569422892796129e-06, + "loss": 1.6140192, + "memory(GiB)": 111.15, + "step": 38670, + "train_speed(iter/s)": 0.448014 + }, + { + "acc": 0.63526869, + "epoch": 0.9811009639776763, + "grad_norm": 5.4375, + "learning_rate": 5.568381078960479e-06, + "loss": 1.76182137, + "memory(GiB)": 111.15, + "step": 38675, + "train_speed(iter/s)": 0.448045 + }, + { + "acc": 0.65909166, + "epoch": 0.9812278031456113, + "grad_norm": 6.5, + "learning_rate": 5.567339240124892e-06, + "loss": 1.59001465, + "memory(GiB)": 111.15, + "step": 38680, + "train_speed(iter/s)": 0.448076 + }, + { + "acc": 0.65199137, + "epoch": 0.9813546423135464, + "grad_norm": 5.09375, + "learning_rate": 5.5662973763351915e-06, + "loss": 1.67126465, + "memory(GiB)": 111.15, + "step": 38685, + "train_speed(iter/s)": 0.448106 + }, + { + "acc": 0.64316497, + "epoch": 0.9814814814814815, + "grad_norm": 6.3125, + "learning_rate": 5.565255487637204e-06, + "loss": 1.65081558, + "memory(GiB)": 111.15, + "step": 38690, + "train_speed(iter/s)": 0.448137 + }, + { + "acc": 0.6480607, + "epoch": 0.9816083206494165, + "grad_norm": 5.8125, + "learning_rate": 5.564213574076757e-06, + "loss": 1.62241325, + "memory(GiB)": 111.15, + "step": 38695, + "train_speed(iter/s)": 0.448168 + }, + { + "acc": 0.63676071, + "epoch": 0.9817351598173516, + "grad_norm": 6.09375, + "learning_rate": 5.563171635699678e-06, + "loss": 1.68695068, + "memory(GiB)": 111.15, + "step": 38700, + "train_speed(iter/s)": 0.448199 + }, + { + "epoch": 0.9817351598173516, + "eval_acc": 0.6432396685099953, + "eval_loss": 1.60833740234375, + "eval_runtime": 115.6096, + "eval_samples_per_second": 55.099, + "eval_steps_per_second": 27.55, + "step": 38700 + }, + { + "acc": 0.66670437, + "epoch": 0.9818619989852867, + "grad_norm": 5.53125, + "learning_rate": 5.562129672551796e-06, + "loss": 1.54412069, + "memory(GiB)": 111.15, + "step": 38705, + "train_speed(iter/s)": 0.447606 + }, + { + "acc": 0.6430994, + "epoch": 0.9819888381532217, + "grad_norm": 5.46875, + "learning_rate": 5.561087684678941e-06, + "loss": 1.65596848, + "memory(GiB)": 111.15, + "step": 38710, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.6431109, + "epoch": 0.9821156773211568, + "grad_norm": 5.15625, + "learning_rate": 5.560045672126945e-06, + "loss": 1.63770256, + "memory(GiB)": 111.15, + "step": 38715, + "train_speed(iter/s)": 0.447668 + }, + { + "acc": 0.65113745, + "epoch": 0.9822425164890918, + "grad_norm": 5.75, + "learning_rate": 5.55900363494164e-06, + "loss": 1.62744102, + "memory(GiB)": 111.15, + "step": 38720, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.65693569, + "epoch": 0.9823693556570269, + "grad_norm": 5.15625, + "learning_rate": 5.557961573168857e-06, + "loss": 1.64446316, + "memory(GiB)": 111.15, + "step": 38725, + "train_speed(iter/s)": 0.447729 + }, + { + "acc": 0.65299716, + "epoch": 0.982496194824962, + "grad_norm": 5.96875, + "learning_rate": 5.5569194868544376e-06, + "loss": 1.65036774, + "memory(GiB)": 111.15, + "step": 38730, + "train_speed(iter/s)": 0.44776 + }, + { + "acc": 0.64541769, + "epoch": 0.982623033992897, + "grad_norm": 5.75, + "learning_rate": 5.555877376044209e-06, + "loss": 1.60355587, + "memory(GiB)": 111.15, + "step": 38735, + "train_speed(iter/s)": 0.447791 + }, + { + "acc": 0.66685667, + "epoch": 0.9827498731608321, + "grad_norm": 5.34375, + "learning_rate": 5.554835240784013e-06, + "loss": 1.60752335, + "memory(GiB)": 111.15, + "step": 38740, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.65326772, + "epoch": 0.9828767123287672, + "grad_norm": 5.46875, + "learning_rate": 5.553793081119685e-06, + "loss": 1.64165192, + "memory(GiB)": 111.15, + "step": 38745, + "train_speed(iter/s)": 0.447853 + }, + { + "acc": 0.64744077, + "epoch": 0.9830035514967022, + "grad_norm": 4.5625, + "learning_rate": 5.552750897097065e-06, + "loss": 1.64683914, + "memory(GiB)": 111.15, + "step": 38750, + "train_speed(iter/s)": 0.447884 + }, + { + "acc": 0.64144907, + "epoch": 0.9831303906646373, + "grad_norm": 6.40625, + "learning_rate": 5.551708688761993e-06, + "loss": 1.64226093, + "memory(GiB)": 111.15, + "step": 38755, + "train_speed(iter/s)": 0.447914 + }, + { + "acc": 0.65029831, + "epoch": 0.9832572298325722, + "grad_norm": 5.09375, + "learning_rate": 5.550666456160311e-06, + "loss": 1.62593269, + "memory(GiB)": 111.15, + "step": 38760, + "train_speed(iter/s)": 0.447945 + }, + { + "acc": 0.65252028, + "epoch": 0.9833840690005073, + "grad_norm": 5.0, + "learning_rate": 5.549624199337857e-06, + "loss": 1.63261166, + "memory(GiB)": 111.15, + "step": 38765, + "train_speed(iter/s)": 0.447976 + }, + { + "acc": 0.63910356, + "epoch": 0.9835109081684424, + "grad_norm": 6.15625, + "learning_rate": 5.548581918340479e-06, + "loss": 1.6522934, + "memory(GiB)": 111.15, + "step": 38770, + "train_speed(iter/s)": 0.448007 + }, + { + "acc": 0.66128902, + "epoch": 0.9836377473363774, + "grad_norm": 5.71875, + "learning_rate": 5.547539613214019e-06, + "loss": 1.5682478, + "memory(GiB)": 111.15, + "step": 38775, + "train_speed(iter/s)": 0.448037 + }, + { + "acc": 0.65054455, + "epoch": 0.9837645865043125, + "grad_norm": 6.75, + "learning_rate": 5.546497284004321e-06, + "loss": 1.64680138, + "memory(GiB)": 111.15, + "step": 38780, + "train_speed(iter/s)": 0.448068 + }, + { + "acc": 0.6533535, + "epoch": 0.9838914256722476, + "grad_norm": 5.875, + "learning_rate": 5.545454930757233e-06, + "loss": 1.61486511, + "memory(GiB)": 111.15, + "step": 38785, + "train_speed(iter/s)": 0.448099 + }, + { + "acc": 0.66209016, + "epoch": 0.9840182648401826, + "grad_norm": 4.875, + "learning_rate": 5.544412553518602e-06, + "loss": 1.54976997, + "memory(GiB)": 111.15, + "step": 38790, + "train_speed(iter/s)": 0.44813 + }, + { + "acc": 0.65023875, + "epoch": 0.9841451040081177, + "grad_norm": 4.625, + "learning_rate": 5.543370152334275e-06, + "loss": 1.59951954, + "memory(GiB)": 111.15, + "step": 38795, + "train_speed(iter/s)": 0.448161 + }, + { + "acc": 0.65026531, + "epoch": 0.9842719431760527, + "grad_norm": 4.625, + "learning_rate": 5.542327727250105e-06, + "loss": 1.61255875, + "memory(GiB)": 111.15, + "step": 38800, + "train_speed(iter/s)": 0.448192 + }, + { + "epoch": 0.9842719431760527, + "eval_acc": 0.6431866120616943, + "eval_loss": 1.6082284450531006, + "eval_runtime": 114.2023, + "eval_samples_per_second": 55.778, + "eval_steps_per_second": 27.889, + "step": 38800 + }, + { + "acc": 0.63830562, + "epoch": 0.9843987823439878, + "grad_norm": 5.78125, + "learning_rate": 5.5412852783119385e-06, + "loss": 1.64790916, + "memory(GiB)": 111.15, + "step": 38805, + "train_speed(iter/s)": 0.447608 + }, + { + "acc": 0.64818711, + "epoch": 0.9845256215119229, + "grad_norm": 4.34375, + "learning_rate": 5.54024280556563e-06, + "loss": 1.5984726, + "memory(GiB)": 111.15, + "step": 38810, + "train_speed(iter/s)": 0.447639 + }, + { + "acc": 0.64142036, + "epoch": 0.9846524606798579, + "grad_norm": 6.6875, + "learning_rate": 5.53920030905703e-06, + "loss": 1.68215446, + "memory(GiB)": 111.15, + "step": 38815, + "train_speed(iter/s)": 0.447669 + }, + { + "acc": 0.64614081, + "epoch": 0.984779299847793, + "grad_norm": 4.84375, + "learning_rate": 5.538157788831993e-06, + "loss": 1.59583282, + "memory(GiB)": 111.15, + "step": 38820, + "train_speed(iter/s)": 0.4477 + }, + { + "acc": 0.64830456, + "epoch": 0.9849061390157281, + "grad_norm": 6.8125, + "learning_rate": 5.537115244936374e-06, + "loss": 1.6144434, + "memory(GiB)": 111.15, + "step": 38825, + "train_speed(iter/s)": 0.447731 + }, + { + "acc": 0.63462982, + "epoch": 0.9850329781836631, + "grad_norm": 6.1875, + "learning_rate": 5.536072677416029e-06, + "loss": 1.68707829, + "memory(GiB)": 111.15, + "step": 38830, + "train_speed(iter/s)": 0.447761 + }, + { + "acc": 0.65889883, + "epoch": 0.9851598173515982, + "grad_norm": 5.15625, + "learning_rate": 5.535030086316814e-06, + "loss": 1.57779512, + "memory(GiB)": 111.15, + "step": 38835, + "train_speed(iter/s)": 0.447791 + }, + { + "acc": 0.66687841, + "epoch": 0.9852866565195332, + "grad_norm": 4.625, + "learning_rate": 5.533987471684586e-06, + "loss": 1.55581455, + "memory(GiB)": 111.15, + "step": 38840, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.64741631, + "epoch": 0.9854134956874683, + "grad_norm": 4.84375, + "learning_rate": 5.532944833565207e-06, + "loss": 1.60340519, + "memory(GiB)": 111.15, + "step": 38845, + "train_speed(iter/s)": 0.447852 + }, + { + "acc": 0.6448493, + "epoch": 0.9855403348554034, + "grad_norm": 4.90625, + "learning_rate": 5.531902172004533e-06, + "loss": 1.64807587, + "memory(GiB)": 111.15, + "step": 38850, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.64658899, + "epoch": 0.9856671740233384, + "grad_norm": 5.40625, + "learning_rate": 5.530859487048427e-06, + "loss": 1.64748135, + "memory(GiB)": 111.15, + "step": 38855, + "train_speed(iter/s)": 0.447914 + }, + { + "acc": 0.63290567, + "epoch": 0.9857940131912735, + "grad_norm": 4.5, + "learning_rate": 5.529816778742752e-06, + "loss": 1.72681808, + "memory(GiB)": 111.15, + "step": 38860, + "train_speed(iter/s)": 0.447944 + }, + { + "acc": 0.67373419, + "epoch": 0.9859208523592086, + "grad_norm": 6.09375, + "learning_rate": 5.528774047133369e-06, + "loss": 1.54914532, + "memory(GiB)": 111.15, + "step": 38865, + "train_speed(iter/s)": 0.447975 + }, + { + "acc": 0.6464582, + "epoch": 0.9860476915271436, + "grad_norm": 4.6875, + "learning_rate": 5.527731292266142e-06, + "loss": 1.62197685, + "memory(GiB)": 111.15, + "step": 38870, + "train_speed(iter/s)": 0.448005 + }, + { + "acc": 0.65546007, + "epoch": 0.9861745306950787, + "grad_norm": 5.03125, + "learning_rate": 5.5266885141869355e-06, + "loss": 1.61491966, + "memory(GiB)": 111.15, + "step": 38875, + "train_speed(iter/s)": 0.448036 + }, + { + "acc": 0.6459547, + "epoch": 0.9863013698630136, + "grad_norm": 6.3125, + "learning_rate": 5.5256457129416185e-06, + "loss": 1.67743912, + "memory(GiB)": 111.15, + "step": 38880, + "train_speed(iter/s)": 0.448067 + }, + { + "acc": 0.66930852, + "epoch": 0.9864282090309487, + "grad_norm": 5.03125, + "learning_rate": 5.524602888576055e-06, + "loss": 1.58891582, + "memory(GiB)": 111.15, + "step": 38885, + "train_speed(iter/s)": 0.448098 + }, + { + "acc": 0.63604403, + "epoch": 0.9865550481988838, + "grad_norm": 5.125, + "learning_rate": 5.523560041136116e-06, + "loss": 1.67907009, + "memory(GiB)": 111.15, + "step": 38890, + "train_speed(iter/s)": 0.448128 + }, + { + "acc": 0.62922287, + "epoch": 0.9866818873668188, + "grad_norm": 5.09375, + "learning_rate": 5.522517170667667e-06, + "loss": 1.71436481, + "memory(GiB)": 111.15, + "step": 38895, + "train_speed(iter/s)": 0.448159 + }, + { + "acc": 0.65854759, + "epoch": 0.9868087265347539, + "grad_norm": 6.03125, + "learning_rate": 5.5214742772165806e-06, + "loss": 1.58648834, + "memory(GiB)": 111.15, + "step": 38900, + "train_speed(iter/s)": 0.44819 + }, + { + "epoch": 0.9868087265347539, + "eval_acc": 0.6432405040446142, + "eval_loss": 1.6082016229629517, + "eval_runtime": 114.7106, + "eval_samples_per_second": 55.531, + "eval_steps_per_second": 27.766, + "step": 38900 + }, + { + "acc": 0.632798, + "epoch": 0.986935565702689, + "grad_norm": 6.0, + "learning_rate": 5.520431360828728e-06, + "loss": 1.69280396, + "memory(GiB)": 111.15, + "step": 38905, + "train_speed(iter/s)": 0.447606 + }, + { + "acc": 0.63993406, + "epoch": 0.987062404870624, + "grad_norm": 5.8125, + "learning_rate": 5.51938842154998e-06, + "loss": 1.66391487, + "memory(GiB)": 111.15, + "step": 38910, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.65736704, + "epoch": 0.9871892440385591, + "grad_norm": 6.25, + "learning_rate": 5.51834545942621e-06, + "loss": 1.58965015, + "memory(GiB)": 111.15, + "step": 38915, + "train_speed(iter/s)": 0.447667 + }, + { + "acc": 0.64566917, + "epoch": 0.9873160832064941, + "grad_norm": 5.65625, + "learning_rate": 5.5173024745032925e-06, + "loss": 1.5968111, + "memory(GiB)": 111.15, + "step": 38920, + "train_speed(iter/s)": 0.447698 + }, + { + "acc": 0.6363894, + "epoch": 0.9874429223744292, + "grad_norm": 5.90625, + "learning_rate": 5.516259466827103e-06, + "loss": 1.67563858, + "memory(GiB)": 111.15, + "step": 38925, + "train_speed(iter/s)": 0.447729 + }, + { + "acc": 0.65155311, + "epoch": 0.9875697615423643, + "grad_norm": 5.375, + "learning_rate": 5.515216436443517e-06, + "loss": 1.59047585, + "memory(GiB)": 111.15, + "step": 38930, + "train_speed(iter/s)": 0.44776 + }, + { + "acc": 0.655405, + "epoch": 0.9876966007102993, + "grad_norm": 6.75, + "learning_rate": 5.514173383398412e-06, + "loss": 1.69908943, + "memory(GiB)": 111.15, + "step": 38935, + "train_speed(iter/s)": 0.447791 + }, + { + "acc": 0.64985099, + "epoch": 0.9878234398782344, + "grad_norm": 5.03125, + "learning_rate": 5.513130307737666e-06, + "loss": 1.61802521, + "memory(GiB)": 111.15, + "step": 38940, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.6479486, + "epoch": 0.9879502790461695, + "grad_norm": 4.71875, + "learning_rate": 5.512087209507157e-06, + "loss": 1.66994839, + "memory(GiB)": 111.15, + "step": 38945, + "train_speed(iter/s)": 0.447853 + }, + { + "acc": 0.65758371, + "epoch": 0.9880771182141045, + "grad_norm": 4.84375, + "learning_rate": 5.5110440887527684e-06, + "loss": 1.53663855, + "memory(GiB)": 111.15, + "step": 38950, + "train_speed(iter/s)": 0.447884 + }, + { + "acc": 0.65314589, + "epoch": 0.9882039573820396, + "grad_norm": 4.4375, + "learning_rate": 5.510000945520377e-06, + "loss": 1.59754486, + "memory(GiB)": 111.15, + "step": 38955, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.65256996, + "epoch": 0.9883307965499746, + "grad_norm": 4.40625, + "learning_rate": 5.508957779855869e-06, + "loss": 1.60757389, + "memory(GiB)": 111.15, + "step": 38960, + "train_speed(iter/s)": 0.447946 + }, + { + "acc": 0.64738321, + "epoch": 0.9884576357179097, + "grad_norm": 6.09375, + "learning_rate": 5.507914591805124e-06, + "loss": 1.62718792, + "memory(GiB)": 111.15, + "step": 38965, + "train_speed(iter/s)": 0.447977 + }, + { + "acc": 0.65215149, + "epoch": 0.9885844748858448, + "grad_norm": 5.0625, + "learning_rate": 5.506871381414027e-06, + "loss": 1.60638371, + "memory(GiB)": 111.15, + "step": 38970, + "train_speed(iter/s)": 0.448008 + }, + { + "acc": 0.64806027, + "epoch": 0.9887113140537798, + "grad_norm": 6.25, + "learning_rate": 5.505828148728465e-06, + "loss": 1.65014534, + "memory(GiB)": 111.15, + "step": 38975, + "train_speed(iter/s)": 0.448039 + }, + { + "acc": 0.65896282, + "epoch": 0.9888381532217149, + "grad_norm": 5.5625, + "learning_rate": 5.5047848937943225e-06, + "loss": 1.61885929, + "memory(GiB)": 111.15, + "step": 38980, + "train_speed(iter/s)": 0.448069 + }, + { + "acc": 0.65670977, + "epoch": 0.98896499238965, + "grad_norm": 5.46875, + "learning_rate": 5.503741616657486e-06, + "loss": 1.5974617, + "memory(GiB)": 111.15, + "step": 38985, + "train_speed(iter/s)": 0.4481 + }, + { + "acc": 0.64853272, + "epoch": 0.989091831557585, + "grad_norm": 6.4375, + "learning_rate": 5.502698317363846e-06, + "loss": 1.66269341, + "memory(GiB)": 111.15, + "step": 38990, + "train_speed(iter/s)": 0.448131 + }, + { + "acc": 0.64214725, + "epoch": 0.98921867072552, + "grad_norm": 5.75, + "learning_rate": 5.501654995959288e-06, + "loss": 1.68369217, + "memory(GiB)": 111.15, + "step": 38995, + "train_speed(iter/s)": 0.448162 + }, + { + "acc": 0.6372438, + "epoch": 0.989345509893455, + "grad_norm": 5.3125, + "learning_rate": 5.5006116524897034e-06, + "loss": 1.63972969, + "memory(GiB)": 111.15, + "step": 39000, + "train_speed(iter/s)": 0.448193 + }, + { + "epoch": 0.989345509893455, + "eval_acc": 0.6433123600218409, + "eval_loss": 1.6082676649093628, + "eval_runtime": 114.0101, + "eval_samples_per_second": 55.872, + "eval_steps_per_second": 27.936, + "step": 39000 + }, + { + "acc": 0.65106945, + "epoch": 0.9894723490613901, + "grad_norm": 6.15625, + "learning_rate": 5.499568287000984e-06, + "loss": 1.66175575, + "memory(GiB)": 111.15, + "step": 39005, + "train_speed(iter/s)": 0.447613 + }, + { + "acc": 0.6523756, + "epoch": 0.9895991882293252, + "grad_norm": 5.5625, + "learning_rate": 5.49852489953902e-06, + "loss": 1.63799248, + "memory(GiB)": 111.15, + "step": 39010, + "train_speed(iter/s)": 0.447643 + }, + { + "acc": 0.65483618, + "epoch": 0.9897260273972602, + "grad_norm": 5.75, + "learning_rate": 5.497481490149705e-06, + "loss": 1.58277607, + "memory(GiB)": 111.15, + "step": 39015, + "train_speed(iter/s)": 0.447674 + }, + { + "acc": 0.64470253, + "epoch": 0.9898528665651953, + "grad_norm": 5.21875, + "learning_rate": 5.496438058878936e-06, + "loss": 1.62632484, + "memory(GiB)": 111.15, + "step": 39020, + "train_speed(iter/s)": 0.447704 + }, + { + "acc": 0.64750156, + "epoch": 0.9899797057331304, + "grad_norm": 5.21875, + "learning_rate": 5.4953946057726005e-06, + "loss": 1.65874405, + "memory(GiB)": 111.15, + "step": 39025, + "train_speed(iter/s)": 0.447734 + }, + { + "acc": 0.66314154, + "epoch": 0.9901065449010654, + "grad_norm": 5.0, + "learning_rate": 5.494351130876602e-06, + "loss": 1.58561316, + "memory(GiB)": 111.15, + "step": 39030, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.66476164, + "epoch": 0.9902333840690005, + "grad_norm": 4.875, + "learning_rate": 5.493307634236831e-06, + "loss": 1.5994173, + "memory(GiB)": 111.15, + "step": 39035, + "train_speed(iter/s)": 0.447794 + }, + { + "acc": 0.64356575, + "epoch": 0.9903602232369355, + "grad_norm": 5.21875, + "learning_rate": 5.492264115899189e-06, + "loss": 1.6328083, + "memory(GiB)": 111.15, + "step": 39040, + "train_speed(iter/s)": 0.447824 + }, + { + "acc": 0.64319687, + "epoch": 0.9904870624048706, + "grad_norm": 7.09375, + "learning_rate": 5.491220575909573e-06, + "loss": 1.67500038, + "memory(GiB)": 111.15, + "step": 39045, + "train_speed(iter/s)": 0.447855 + }, + { + "acc": 0.6514545, + "epoch": 0.9906139015728057, + "grad_norm": 4.65625, + "learning_rate": 5.4901770143138835e-06, + "loss": 1.66373558, + "memory(GiB)": 111.15, + "step": 39050, + "train_speed(iter/s)": 0.447885 + }, + { + "acc": 0.65245466, + "epoch": 0.9907407407407407, + "grad_norm": 4.40625, + "learning_rate": 5.48913343115802e-06, + "loss": 1.59431324, + "memory(GiB)": 111.15, + "step": 39055, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.64514451, + "epoch": 0.9908675799086758, + "grad_norm": 4.46875, + "learning_rate": 5.488089826487884e-06, + "loss": 1.59268484, + "memory(GiB)": 111.15, + "step": 39060, + "train_speed(iter/s)": 0.447945 + }, + { + "acc": 0.65673885, + "epoch": 0.9909944190766109, + "grad_norm": 4.625, + "learning_rate": 5.48704620034938e-06, + "loss": 1.61783028, + "memory(GiB)": 111.15, + "step": 39065, + "train_speed(iter/s)": 0.447976 + }, + { + "acc": 0.66663427, + "epoch": 0.9911212582445459, + "grad_norm": 5.5, + "learning_rate": 5.486002552788408e-06, + "loss": 1.61757336, + "memory(GiB)": 111.15, + "step": 39070, + "train_speed(iter/s)": 0.448006 + }, + { + "acc": 0.64552255, + "epoch": 0.991248097412481, + "grad_norm": 5.71875, + "learning_rate": 5.4849588838508734e-06, + "loss": 1.61322823, + "memory(GiB)": 111.15, + "step": 39075, + "train_speed(iter/s)": 0.448036 + }, + { + "acc": 0.66338196, + "epoch": 0.991374936580416, + "grad_norm": 5.625, + "learning_rate": 5.483915193582684e-06, + "loss": 1.61497135, + "memory(GiB)": 111.15, + "step": 39080, + "train_speed(iter/s)": 0.448066 + }, + { + "acc": 0.63486028, + "epoch": 0.9915017757483511, + "grad_norm": 4.875, + "learning_rate": 5.482871482029742e-06, + "loss": 1.59242516, + "memory(GiB)": 111.15, + "step": 39085, + "train_speed(iter/s)": 0.448096 + }, + { + "acc": 0.65460591, + "epoch": 0.9916286149162862, + "grad_norm": 5.0625, + "learning_rate": 5.4818277492379565e-06, + "loss": 1.64608116, + "memory(GiB)": 111.15, + "step": 39090, + "train_speed(iter/s)": 0.448126 + }, + { + "acc": 0.66008201, + "epoch": 0.9917554540842212, + "grad_norm": 5.3125, + "learning_rate": 5.480783995253236e-06, + "loss": 1.62929153, + "memory(GiB)": 111.15, + "step": 39095, + "train_speed(iter/s)": 0.448157 + }, + { + "acc": 0.66490121, + "epoch": 0.9918822932521563, + "grad_norm": 4.75, + "learning_rate": 5.47974022012149e-06, + "loss": 1.61535759, + "memory(GiB)": 111.15, + "step": 39100, + "train_speed(iter/s)": 0.448187 + }, + { + "epoch": 0.9918822932521563, + "eval_acc": 0.6433382615950273, + "eval_loss": 1.6082067489624023, + "eval_runtime": 114.9472, + "eval_samples_per_second": 55.417, + "eval_steps_per_second": 27.708, + "step": 39100 + }, + { + "acc": 0.66118412, + "epoch": 0.9920091324200914, + "grad_norm": 6.03125, + "learning_rate": 5.478696423888624e-06, + "loss": 1.58910322, + "memory(GiB)": 111.15, + "step": 39105, + "train_speed(iter/s)": 0.447604 + }, + { + "acc": 0.65348701, + "epoch": 0.9921359715880264, + "grad_norm": 5.1875, + "learning_rate": 5.477652606600555e-06, + "loss": 1.61331863, + "memory(GiB)": 111.15, + "step": 39110, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.66350994, + "epoch": 0.9922628107559615, + "grad_norm": 5.75, + "learning_rate": 5.47660876830319e-06, + "loss": 1.56650829, + "memory(GiB)": 111.15, + "step": 39115, + "train_speed(iter/s)": 0.447664 + }, + { + "acc": 0.66485724, + "epoch": 0.9923896499238964, + "grad_norm": 5.40625, + "learning_rate": 5.475564909042444e-06, + "loss": 1.5440834, + "memory(GiB)": 111.15, + "step": 39120, + "train_speed(iter/s)": 0.447694 + }, + { + "acc": 0.65781565, + "epoch": 0.9925164890918315, + "grad_norm": 5.1875, + "learning_rate": 5.4745210288642306e-06, + "loss": 1.59838352, + "memory(GiB)": 111.15, + "step": 39125, + "train_speed(iter/s)": 0.447724 + }, + { + "acc": 0.6324935, + "epoch": 0.9926433282597666, + "grad_norm": 4.65625, + "learning_rate": 5.473477127814464e-06, + "loss": 1.6156538, + "memory(GiB)": 111.15, + "step": 39130, + "train_speed(iter/s)": 0.447754 + }, + { + "acc": 0.63243775, + "epoch": 0.9927701674277016, + "grad_norm": 5.8125, + "learning_rate": 5.472433205939058e-06, + "loss": 1.71217213, + "memory(GiB)": 111.15, + "step": 39135, + "train_speed(iter/s)": 0.447784 + }, + { + "acc": 0.66811447, + "epoch": 0.9928970065956367, + "grad_norm": 6.125, + "learning_rate": 5.471389263283932e-06, + "loss": 1.56518116, + "memory(GiB)": 111.15, + "step": 39140, + "train_speed(iter/s)": 0.447814 + }, + { + "acc": 0.66972437, + "epoch": 0.9930238457635718, + "grad_norm": 6.15625, + "learning_rate": 5.4703452998950005e-06, + "loss": 1.55787668, + "memory(GiB)": 111.15, + "step": 39145, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.64865832, + "epoch": 0.9931506849315068, + "grad_norm": 5.25, + "learning_rate": 5.469301315818183e-06, + "loss": 1.6869381, + "memory(GiB)": 111.15, + "step": 39150, + "train_speed(iter/s)": 0.447875 + }, + { + "acc": 0.6501698, + "epoch": 0.9932775240994419, + "grad_norm": 5.0625, + "learning_rate": 5.468257311099399e-06, + "loss": 1.53545971, + "memory(GiB)": 111.15, + "step": 39155, + "train_speed(iter/s)": 0.447905 + }, + { + "acc": 0.67240143, + "epoch": 0.9934043632673769, + "grad_norm": 5.3125, + "learning_rate": 5.467213285784567e-06, + "loss": 1.5861352, + "memory(GiB)": 111.15, + "step": 39160, + "train_speed(iter/s)": 0.447935 + }, + { + "acc": 0.65195913, + "epoch": 0.993531202435312, + "grad_norm": 5.03125, + "learning_rate": 5.466169239919608e-06, + "loss": 1.66908607, + "memory(GiB)": 111.15, + "step": 39165, + "train_speed(iter/s)": 0.447964 + }, + { + "acc": 0.64898911, + "epoch": 0.9936580416032471, + "grad_norm": 5.8125, + "learning_rate": 5.465125173550446e-06, + "loss": 1.67499466, + "memory(GiB)": 111.15, + "step": 39170, + "train_speed(iter/s)": 0.447994 + }, + { + "acc": 0.64629135, + "epoch": 0.9937848807711821, + "grad_norm": 5.84375, + "learning_rate": 5.464081086723001e-06, + "loss": 1.68281803, + "memory(GiB)": 111.15, + "step": 39175, + "train_speed(iter/s)": 0.448025 + }, + { + "acc": 0.63684025, + "epoch": 0.9939117199391172, + "grad_norm": 5.5625, + "learning_rate": 5.4630369794832006e-06, + "loss": 1.69737682, + "memory(GiB)": 111.15, + "step": 39180, + "train_speed(iter/s)": 0.448055 + }, + { + "acc": 0.65934916, + "epoch": 0.9940385591070523, + "grad_norm": 5.40625, + "learning_rate": 5.461992851876963e-06, + "loss": 1.64266796, + "memory(GiB)": 111.15, + "step": 39185, + "train_speed(iter/s)": 0.448084 + }, + { + "acc": 0.64949117, + "epoch": 0.9941653982749873, + "grad_norm": 6.46875, + "learning_rate": 5.460948703950218e-06, + "loss": 1.71684361, + "memory(GiB)": 111.15, + "step": 39190, + "train_speed(iter/s)": 0.448114 + }, + { + "acc": 0.64473372, + "epoch": 0.9942922374429224, + "grad_norm": 6.15625, + "learning_rate": 5.459904535748892e-06, + "loss": 1.66077843, + "memory(GiB)": 111.15, + "step": 39195, + "train_speed(iter/s)": 0.448145 + }, + { + "acc": 0.67691469, + "epoch": 0.9944190766108574, + "grad_norm": 5.09375, + "learning_rate": 5.458860347318912e-06, + "loss": 1.50590191, + "memory(GiB)": 111.15, + "step": 39200, + "train_speed(iter/s)": 0.448175 + }, + { + "epoch": 0.9944190766108574, + "eval_acc": 0.6432020694521441, + "eval_loss": 1.6082583665847778, + "eval_runtime": 113.4651, + "eval_samples_per_second": 56.141, + "eval_steps_per_second": 28.07, + "step": 39200 + }, + { + "acc": 0.65131121, + "epoch": 0.9945459157787925, + "grad_norm": 5.875, + "learning_rate": 5.457816138706203e-06, + "loss": 1.67283859, + "memory(GiB)": 111.15, + "step": 39205, + "train_speed(iter/s)": 0.447601 + }, + { + "acc": 0.65959349, + "epoch": 0.9946727549467276, + "grad_norm": 6.6875, + "learning_rate": 5.456771909956697e-06, + "loss": 1.62966232, + "memory(GiB)": 111.15, + "step": 39210, + "train_speed(iter/s)": 0.447631 + }, + { + "acc": 0.6658596, + "epoch": 0.9947995941146626, + "grad_norm": 5.34375, + "learning_rate": 5.455727661116324e-06, + "loss": 1.58908377, + "memory(GiB)": 111.15, + "step": 39215, + "train_speed(iter/s)": 0.447661 + }, + { + "acc": 0.65662622, + "epoch": 0.9949264332825977, + "grad_norm": 7.34375, + "learning_rate": 5.454683392231014e-06, + "loss": 1.64433212, + "memory(GiB)": 111.15, + "step": 39220, + "train_speed(iter/s)": 0.447691 + }, + { + "acc": 0.66279936, + "epoch": 0.9950532724505328, + "grad_norm": 4.84375, + "learning_rate": 5.453639103346697e-06, + "loss": 1.56457024, + "memory(GiB)": 111.15, + "step": 39225, + "train_speed(iter/s)": 0.447721 + }, + { + "acc": 0.65143805, + "epoch": 0.9951801116184678, + "grad_norm": 6.625, + "learning_rate": 5.452594794509307e-06, + "loss": 1.59279537, + "memory(GiB)": 111.15, + "step": 39230, + "train_speed(iter/s)": 0.447751 + }, + { + "acc": 0.65677204, + "epoch": 0.9953069507864029, + "grad_norm": 5.3125, + "learning_rate": 5.4515504657647765e-06, + "loss": 1.63405762, + "memory(GiB)": 111.15, + "step": 39235, + "train_speed(iter/s)": 0.44778 + }, + { + "acc": 0.64984322, + "epoch": 0.9954337899543378, + "grad_norm": 5.25, + "learning_rate": 5.450506117159044e-06, + "loss": 1.7271637, + "memory(GiB)": 111.15, + "step": 39240, + "train_speed(iter/s)": 0.44781 + }, + { + "acc": 0.65072136, + "epoch": 0.9955606291222729, + "grad_norm": 6.65625, + "learning_rate": 5.449461748738037e-06, + "loss": 1.65196648, + "memory(GiB)": 111.15, + "step": 39245, + "train_speed(iter/s)": 0.44784 + }, + { + "acc": 0.65900936, + "epoch": 0.995687468290208, + "grad_norm": 5.5625, + "learning_rate": 5.448417360547699e-06, + "loss": 1.58362904, + "memory(GiB)": 111.15, + "step": 39250, + "train_speed(iter/s)": 0.44787 + }, + { + "acc": 0.65894766, + "epoch": 0.995814307458143, + "grad_norm": 4.90625, + "learning_rate": 5.44737295263396e-06, + "loss": 1.55037003, + "memory(GiB)": 111.15, + "step": 39255, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.6400938, + "epoch": 0.9959411466260781, + "grad_norm": 5.6875, + "learning_rate": 5.446328525042764e-06, + "loss": 1.70235977, + "memory(GiB)": 111.15, + "step": 39260, + "train_speed(iter/s)": 0.44793 + }, + { + "acc": 0.64759274, + "epoch": 0.9960679857940132, + "grad_norm": 5.90625, + "learning_rate": 5.4452840778200456e-06, + "loss": 1.65262241, + "memory(GiB)": 111.15, + "step": 39265, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.65061407, + "epoch": 0.9961948249619482, + "grad_norm": 6.0, + "learning_rate": 5.444239611011746e-06, + "loss": 1.64820518, + "memory(GiB)": 111.15, + "step": 39270, + "train_speed(iter/s)": 0.447989 + }, + { + "acc": 0.64439659, + "epoch": 0.9963216641298833, + "grad_norm": 4.84375, + "learning_rate": 5.443195124663804e-06, + "loss": 1.66527863, + "memory(GiB)": 111.15, + "step": 39275, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.64506311, + "epoch": 0.9964485032978183, + "grad_norm": 5.59375, + "learning_rate": 5.442150618822162e-06, + "loss": 1.6430666, + "memory(GiB)": 111.15, + "step": 39280, + "train_speed(iter/s)": 0.44805 + }, + { + "acc": 0.65723066, + "epoch": 0.9965753424657534, + "grad_norm": 5.125, + "learning_rate": 5.441106093532762e-06, + "loss": 1.69065323, + "memory(GiB)": 111.15, + "step": 39285, + "train_speed(iter/s)": 0.44808 + }, + { + "acc": 0.65834045, + "epoch": 0.9967021816336885, + "grad_norm": 4.46875, + "learning_rate": 5.440061548841546e-06, + "loss": 1.58060293, + "memory(GiB)": 111.15, + "step": 39290, + "train_speed(iter/s)": 0.44811 + }, + { + "acc": 0.65455742, + "epoch": 0.9968290208016235, + "grad_norm": 5.75, + "learning_rate": 5.43901698479446e-06, + "loss": 1.62158279, + "memory(GiB)": 111.15, + "step": 39295, + "train_speed(iter/s)": 0.44814 + }, + { + "acc": 0.63209276, + "epoch": 0.9969558599695586, + "grad_norm": 5.40625, + "learning_rate": 5.4379724014374455e-06, + "loss": 1.67374191, + "memory(GiB)": 111.15, + "step": 39300, + "train_speed(iter/s)": 0.448171 + }, + { + "epoch": 0.9969558599695586, + "eval_acc": 0.6432701655235857, + "eval_loss": 1.6082273721694946, + "eval_runtime": 113.6998, + "eval_samples_per_second": 56.025, + "eval_steps_per_second": 28.012, + "step": 39300 + }, + { + "acc": 0.63998156, + "epoch": 0.9970826991374937, + "grad_norm": 4.6875, + "learning_rate": 5.436927798816448e-06, + "loss": 1.6513649, + "memory(GiB)": 111.15, + "step": 39305, + "train_speed(iter/s)": 0.447597 + }, + { + "acc": 0.66049671, + "epoch": 0.9972095383054287, + "grad_norm": 6.53125, + "learning_rate": 5.4358831769774174e-06, + "loss": 1.57565384, + "memory(GiB)": 111.15, + "step": 39310, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.65350571, + "epoch": 0.9973363774733638, + "grad_norm": 4.8125, + "learning_rate": 5.434838535966298e-06, + "loss": 1.67350407, + "memory(GiB)": 111.15, + "step": 39315, + "train_speed(iter/s)": 0.447658 + }, + { + "acc": 0.63826437, + "epoch": 0.9974632166412988, + "grad_norm": 5.75, + "learning_rate": 5.43379387582904e-06, + "loss": 1.64764843, + "memory(GiB)": 111.15, + "step": 39320, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.65622091, + "epoch": 0.9975900558092339, + "grad_norm": 4.90625, + "learning_rate": 5.432749196611587e-06, + "loss": 1.59636812, + "memory(GiB)": 111.15, + "step": 39325, + "train_speed(iter/s)": 0.447718 + }, + { + "acc": 0.64829016, + "epoch": 0.997716894977169, + "grad_norm": 5.53125, + "learning_rate": 5.431704498359896e-06, + "loss": 1.60851097, + "memory(GiB)": 111.15, + "step": 39330, + "train_speed(iter/s)": 0.447748 + }, + { + "acc": 0.64982653, + "epoch": 0.997843734145104, + "grad_norm": 6.53125, + "learning_rate": 5.43065978111991e-06, + "loss": 1.61846027, + "memory(GiB)": 111.15, + "step": 39335, + "train_speed(iter/s)": 0.447779 + }, + { + "acc": 0.647859, + "epoch": 0.9979705733130391, + "grad_norm": 5.75, + "learning_rate": 5.429615044937586e-06, + "loss": 1.58494644, + "memory(GiB)": 111.15, + "step": 39340, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.64695005, + "epoch": 0.9980974124809742, + "grad_norm": 5.125, + "learning_rate": 5.4285702898588754e-06, + "loss": 1.67978935, + "memory(GiB)": 111.15, + "step": 39345, + "train_speed(iter/s)": 0.447839 + }, + { + "acc": 0.6530756, + "epoch": 0.9982242516489092, + "grad_norm": 4.8125, + "learning_rate": 5.427525515929729e-06, + "loss": 1.61268673, + "memory(GiB)": 111.15, + "step": 39350, + "train_speed(iter/s)": 0.447869 + }, + { + "acc": 0.66860709, + "epoch": 0.9983510908168443, + "grad_norm": 6.1875, + "learning_rate": 5.426480723196102e-06, + "loss": 1.52746496, + "memory(GiB)": 111.15, + "step": 39355, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.65226078, + "epoch": 0.9984779299847792, + "grad_norm": 5.15625, + "learning_rate": 5.425435911703948e-06, + "loss": 1.62220879, + "memory(GiB)": 111.15, + "step": 39360, + "train_speed(iter/s)": 0.44793 + }, + { + "acc": 0.64381142, + "epoch": 0.9986047691527143, + "grad_norm": 5.59375, + "learning_rate": 5.424391081499223e-06, + "loss": 1.63242054, + "memory(GiB)": 111.15, + "step": 39365, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.65289841, + "epoch": 0.9987316083206494, + "grad_norm": 4.75, + "learning_rate": 5.423346232627884e-06, + "loss": 1.61484833, + "memory(GiB)": 111.15, + "step": 39370, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.636304, + "epoch": 0.9988584474885844, + "grad_norm": 6.1875, + "learning_rate": 5.422301365135887e-06, + "loss": 1.67468128, + "memory(GiB)": 111.15, + "step": 39375, + "train_speed(iter/s)": 0.44802 + }, + { + "acc": 0.63682871, + "epoch": 0.9989852866565195, + "grad_norm": 6.15625, + "learning_rate": 5.421256479069191e-06, + "loss": 1.64049225, + "memory(GiB)": 111.15, + "step": 39380, + "train_speed(iter/s)": 0.44805 + }, + { + "acc": 0.66197672, + "epoch": 0.9991121258244546, + "grad_norm": 5.84375, + "learning_rate": 5.420211574473754e-06, + "loss": 1.5611393, + "memory(GiB)": 111.15, + "step": 39385, + "train_speed(iter/s)": 0.448081 + }, + { + "acc": 0.65536337, + "epoch": 0.9992389649923896, + "grad_norm": 5.1875, + "learning_rate": 5.419166651395536e-06, + "loss": 1.60675144, + "memory(GiB)": 111.15, + "step": 39390, + "train_speed(iter/s)": 0.448111 + }, + { + "acc": 0.65984936, + "epoch": 0.9993658041603247, + "grad_norm": 5.15625, + "learning_rate": 5.418121709880497e-06, + "loss": 1.56654949, + "memory(GiB)": 111.15, + "step": 39395, + "train_speed(iter/s)": 0.448141 + }, + { + "acc": 0.6483614, + "epoch": 0.9994926433282597, + "grad_norm": 5.90625, + "learning_rate": 5.4170767499746e-06, + "loss": 1.64665527, + "memory(GiB)": 111.15, + "step": 39400, + "train_speed(iter/s)": 0.448171 + }, + { + "epoch": 0.9994926433282597, + "eval_acc": 0.6432179446099036, + "eval_loss": 1.608314871788025, + "eval_runtime": 113.6741, + "eval_samples_per_second": 56.037, + "eval_steps_per_second": 28.019, + "step": 39400 + }, + { + "acc": 0.64883614, + "epoch": 0.9996194824961948, + "grad_norm": 5.75, + "learning_rate": 5.416031771723803e-06, + "loss": 1.66084232, + "memory(GiB)": 111.15, + "step": 39405, + "train_speed(iter/s)": 0.4476 + }, + { + "acc": 0.65329466, + "epoch": 0.9997463216641299, + "grad_norm": 6.5, + "learning_rate": 5.414986775174073e-06, + "loss": 1.59881992, + "memory(GiB)": 111.15, + "step": 39410, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.66442161, + "epoch": 0.9998731608320649, + "grad_norm": 4.375, + "learning_rate": 5.41394176037137e-06, + "loss": 1.57998381, + "memory(GiB)": 111.15, + "step": 39415, + "train_speed(iter/s)": 0.44766 + }, + { + "acc": 0.64412012, + "epoch": 1.0, + "grad_norm": 5.46875, + "learning_rate": 5.412896727361663e-06, + "loss": 1.66640778, + "memory(GiB)": 111.15, + "step": 39420, + "train_speed(iter/s)": 0.447685 + }, + { + "acc": 0.66178389, + "epoch": 1.000126839167935, + "grad_norm": 5.71875, + "learning_rate": 5.411851676190912e-06, + "loss": 1.5887125, + "memory(GiB)": 111.15, + "step": 39425, + "train_speed(iter/s)": 0.447713 + }, + { + "acc": 0.64433618, + "epoch": 1.0002536783358702, + "grad_norm": 5.09375, + "learning_rate": 5.4108066069050864e-06, + "loss": 1.6462265, + "memory(GiB)": 111.15, + "step": 39430, + "train_speed(iter/s)": 0.447743 + }, + { + "acc": 0.62163587, + "epoch": 1.0003805175038052, + "grad_norm": 4.96875, + "learning_rate": 5.409761519550153e-06, + "loss": 1.68288383, + "memory(GiB)": 111.15, + "step": 39435, + "train_speed(iter/s)": 0.447773 + }, + { + "acc": 0.66523914, + "epoch": 1.0005073566717402, + "grad_norm": 5.65625, + "learning_rate": 5.408716414172077e-06, + "loss": 1.571632, + "memory(GiB)": 111.15, + "step": 39440, + "train_speed(iter/s)": 0.447803 + }, + { + "acc": 0.66488414, + "epoch": 1.0006341958396754, + "grad_norm": 6.75, + "learning_rate": 5.407671290816829e-06, + "loss": 1.62046261, + "memory(GiB)": 111.15, + "step": 39445, + "train_speed(iter/s)": 0.447833 + }, + { + "acc": 0.62985287, + "epoch": 1.0007610350076104, + "grad_norm": 5.5, + "learning_rate": 5.406626149530378e-06, + "loss": 1.74389839, + "memory(GiB)": 111.15, + "step": 39450, + "train_speed(iter/s)": 0.447863 + }, + { + "acc": 0.66560316, + "epoch": 1.0008878741755454, + "grad_norm": 4.625, + "learning_rate": 5.405580990358692e-06, + "loss": 1.53729324, + "memory(GiB)": 111.15, + "step": 39455, + "train_speed(iter/s)": 0.447894 + }, + { + "acc": 0.65725031, + "epoch": 1.0010147133434804, + "grad_norm": 5.84375, + "learning_rate": 5.404535813347746e-06, + "loss": 1.62159004, + "memory(GiB)": 111.15, + "step": 39460, + "train_speed(iter/s)": 0.447924 + }, + { + "acc": 0.64566045, + "epoch": 1.0011415525114156, + "grad_norm": 6.40625, + "learning_rate": 5.403490618543505e-06, + "loss": 1.65722046, + "memory(GiB)": 111.15, + "step": 39465, + "train_speed(iter/s)": 0.447954 + }, + { + "acc": 0.66430845, + "epoch": 1.0012683916793506, + "grad_norm": 5.53125, + "learning_rate": 5.40244540599195e-06, + "loss": 1.58184166, + "memory(GiB)": 111.15, + "step": 39470, + "train_speed(iter/s)": 0.447984 + }, + { + "acc": 0.65437417, + "epoch": 1.0013952308472855, + "grad_norm": 4.40625, + "learning_rate": 5.401400175739045e-06, + "loss": 1.6633709, + "memory(GiB)": 111.15, + "step": 39475, + "train_speed(iter/s)": 0.448014 + }, + { + "acc": 0.64184475, + "epoch": 1.0015220700152208, + "grad_norm": 5.65625, + "learning_rate": 5.400354927830769e-06, + "loss": 1.70178795, + "memory(GiB)": 111.15, + "step": 39480, + "train_speed(iter/s)": 0.448044 + }, + { + "acc": 0.66010714, + "epoch": 1.0016489091831557, + "grad_norm": 5.46875, + "learning_rate": 5.399309662313097e-06, + "loss": 1.5991272, + "memory(GiB)": 111.15, + "step": 39485, + "train_speed(iter/s)": 0.448075 + }, + { + "acc": 0.64233284, + "epoch": 1.0017757483510907, + "grad_norm": 5.46875, + "learning_rate": 5.3982643792320024e-06, + "loss": 1.64679756, + "memory(GiB)": 111.15, + "step": 39490, + "train_speed(iter/s)": 0.448105 + }, + { + "acc": 0.64842343, + "epoch": 1.001902587519026, + "grad_norm": 4.875, + "learning_rate": 5.397219078633462e-06, + "loss": 1.60900593, + "memory(GiB)": 111.15, + "step": 39495, + "train_speed(iter/s)": 0.448135 + }, + { + "acc": 0.65039477, + "epoch": 1.002029426686961, + "grad_norm": 6.40625, + "learning_rate": 5.3961737605634546e-06, + "loss": 1.63652115, + "memory(GiB)": 111.15, + "step": 39500, + "train_speed(iter/s)": 0.448165 + }, + { + "epoch": 1.002029426686961, + "eval_acc": 0.6432141847041184, + "eval_loss": 1.60808265209198, + "eval_runtime": 113.7677, + "eval_samples_per_second": 55.991, + "eval_steps_per_second": 27.996, + "step": 39500 + }, + { + "acc": 0.66362972, + "epoch": 1.002156265854896, + "grad_norm": 5.59375, + "learning_rate": 5.395128425067954e-06, + "loss": 1.62422028, + "memory(GiB)": 111.15, + "step": 39505, + "train_speed(iter/s)": 0.447592 + }, + { + "acc": 0.67544489, + "epoch": 1.0022831050228311, + "grad_norm": 4.90625, + "learning_rate": 5.394083072192944e-06, + "loss": 1.53909721, + "memory(GiB)": 111.15, + "step": 39510, + "train_speed(iter/s)": 0.447622 + }, + { + "acc": 0.66685333, + "epoch": 1.0024099441907661, + "grad_norm": 5.9375, + "learning_rate": 5.393037701984399e-06, + "loss": 1.51949902, + "memory(GiB)": 111.15, + "step": 39515, + "train_speed(iter/s)": 0.447652 + }, + { + "acc": 0.64012194, + "epoch": 1.0025367833587011, + "grad_norm": 4.90625, + "learning_rate": 5.391992314488303e-06, + "loss": 1.61038723, + "memory(GiB)": 111.15, + "step": 39520, + "train_speed(iter/s)": 0.447682 + }, + { + "acc": 0.65102844, + "epoch": 1.0026636225266363, + "grad_norm": 4.65625, + "learning_rate": 5.3909469097506314e-06, + "loss": 1.62399578, + "memory(GiB)": 111.15, + "step": 39525, + "train_speed(iter/s)": 0.447711 + }, + { + "acc": 0.63440504, + "epoch": 1.0027904616945713, + "grad_norm": 5.65625, + "learning_rate": 5.389901487817373e-06, + "loss": 1.63620453, + "memory(GiB)": 111.15, + "step": 39530, + "train_speed(iter/s)": 0.447741 + }, + { + "acc": 0.6556098, + "epoch": 1.0029173008625063, + "grad_norm": 5.8125, + "learning_rate": 5.388856048734505e-06, + "loss": 1.60866566, + "memory(GiB)": 111.15, + "step": 39535, + "train_speed(iter/s)": 0.447771 + }, + { + "acc": 0.6431735, + "epoch": 1.0030441400304415, + "grad_norm": 5.25, + "learning_rate": 5.3878105925480115e-06, + "loss": 1.73538666, + "memory(GiB)": 111.15, + "step": 39540, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.65401006, + "epoch": 1.0031709791983765, + "grad_norm": 5.0, + "learning_rate": 5.3867651193038765e-06, + "loss": 1.6365387, + "memory(GiB)": 111.15, + "step": 39545, + "train_speed(iter/s)": 0.447831 + }, + { + "acc": 0.65336471, + "epoch": 1.0032978183663115, + "grad_norm": 5.1875, + "learning_rate": 5.385719629048086e-06, + "loss": 1.59634533, + "memory(GiB)": 111.15, + "step": 39550, + "train_speed(iter/s)": 0.447861 + }, + { + "acc": 0.63970785, + "epoch": 1.0034246575342465, + "grad_norm": 4.5625, + "learning_rate": 5.384674121826622e-06, + "loss": 1.68354187, + "memory(GiB)": 111.15, + "step": 39555, + "train_speed(iter/s)": 0.44789 + }, + { + "acc": 0.65445366, + "epoch": 1.0035514967021817, + "grad_norm": 6.0, + "learning_rate": 5.383628597685474e-06, + "loss": 1.60566216, + "memory(GiB)": 111.15, + "step": 39560, + "train_speed(iter/s)": 0.44792 + }, + { + "acc": 0.67544084, + "epoch": 1.0036783358701167, + "grad_norm": 5.28125, + "learning_rate": 5.382583056670627e-06, + "loss": 1.52926083, + "memory(GiB)": 111.15, + "step": 39565, + "train_speed(iter/s)": 0.44795 + }, + { + "acc": 0.64569597, + "epoch": 1.0038051750380517, + "grad_norm": 5.9375, + "learning_rate": 5.38153749882807e-06, + "loss": 1.67619057, + "memory(GiB)": 111.15, + "step": 39570, + "train_speed(iter/s)": 0.44798 + }, + { + "acc": 0.650702, + "epoch": 1.0039320142059869, + "grad_norm": 5.875, + "learning_rate": 5.38049192420379e-06, + "loss": 1.6190979, + "memory(GiB)": 111.15, + "step": 39575, + "train_speed(iter/s)": 0.44801 + }, + { + "acc": 0.64530878, + "epoch": 1.0040588533739219, + "grad_norm": 4.375, + "learning_rate": 5.3794463328437766e-06, + "loss": 1.61990814, + "memory(GiB)": 111.15, + "step": 39580, + "train_speed(iter/s)": 0.44804 + }, + { + "acc": 0.6591095, + "epoch": 1.0041856925418569, + "grad_norm": 5.5625, + "learning_rate": 5.3784007247940185e-06, + "loss": 1.56346836, + "memory(GiB)": 111.15, + "step": 39585, + "train_speed(iter/s)": 0.44807 + }, + { + "acc": 0.64829483, + "epoch": 1.004312531709792, + "grad_norm": 6.625, + "learning_rate": 5.377355100100508e-06, + "loss": 1.52508392, + "memory(GiB)": 111.15, + "step": 39590, + "train_speed(iter/s)": 0.448099 + }, + { + "acc": 0.66062636, + "epoch": 1.004439370877727, + "grad_norm": 6.25, + "learning_rate": 5.376309458809235e-06, + "loss": 1.58911142, + "memory(GiB)": 111.15, + "step": 39595, + "train_speed(iter/s)": 0.448129 + }, + { + "acc": 0.63790379, + "epoch": 1.004566210045662, + "grad_norm": 5.46875, + "learning_rate": 5.375263800966192e-06, + "loss": 1.633465, + "memory(GiB)": 111.15, + "step": 39600, + "train_speed(iter/s)": 0.448159 + }, + { + "epoch": 1.004566210045662, + "eval_acc": 0.6431932963386455, + "eval_loss": 1.6082106828689575, + "eval_runtime": 114.3675, + "eval_samples_per_second": 55.698, + "eval_steps_per_second": 27.849, + "step": 39600 + }, + { + "acc": 0.64589062, + "epoch": 1.0046930492135973, + "grad_norm": 6.53125, + "learning_rate": 5.374218126617371e-06, + "loss": 1.60336304, + "memory(GiB)": 111.15, + "step": 39605, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.64720907, + "epoch": 1.0048198883815322, + "grad_norm": 4.96875, + "learning_rate": 5.373172435808768e-06, + "loss": 1.66600628, + "memory(GiB)": 111.15, + "step": 39610, + "train_speed(iter/s)": 0.447615 + }, + { + "acc": 0.64777117, + "epoch": 1.0049467275494672, + "grad_norm": 5.71875, + "learning_rate": 5.372126728586372e-06, + "loss": 1.67611465, + "memory(GiB)": 111.15, + "step": 39615, + "train_speed(iter/s)": 0.447645 + }, + { + "acc": 0.64936905, + "epoch": 1.0050735667174022, + "grad_norm": 5.28125, + "learning_rate": 5.371081004996184e-06, + "loss": 1.59677639, + "memory(GiB)": 111.15, + "step": 39620, + "train_speed(iter/s)": 0.447675 + }, + { + "acc": 0.64281044, + "epoch": 1.0052004058853374, + "grad_norm": 4.8125, + "learning_rate": 5.370035265084195e-06, + "loss": 1.61831551, + "memory(GiB)": 111.15, + "step": 39625, + "train_speed(iter/s)": 0.447704 + }, + { + "acc": 0.64756889, + "epoch": 1.0053272450532724, + "grad_norm": 6.0, + "learning_rate": 5.3689895088964025e-06, + "loss": 1.63750305, + "memory(GiB)": 111.15, + "step": 39630, + "train_speed(iter/s)": 0.447734 + }, + { + "acc": 0.64257288, + "epoch": 1.0054540842212074, + "grad_norm": 5.5625, + "learning_rate": 5.367943736478806e-06, + "loss": 1.62911949, + "memory(GiB)": 111.15, + "step": 39635, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.64446812, + "epoch": 1.0055809233891426, + "grad_norm": 5.59375, + "learning_rate": 5.3668979478774e-06, + "loss": 1.69302483, + "memory(GiB)": 111.15, + "step": 39640, + "train_speed(iter/s)": 0.447793 + }, + { + "acc": 0.66250248, + "epoch": 1.0057077625570776, + "grad_norm": 5.6875, + "learning_rate": 5.3658521431381836e-06, + "loss": 1.58643856, + "memory(GiB)": 111.15, + "step": 39645, + "train_speed(iter/s)": 0.447823 + }, + { + "acc": 0.64807558, + "epoch": 1.0058346017250126, + "grad_norm": 4.8125, + "learning_rate": 5.364806322307158e-06, + "loss": 1.59241257, + "memory(GiB)": 111.15, + "step": 39650, + "train_speed(iter/s)": 0.447853 + }, + { + "acc": 0.6422327, + "epoch": 1.0059614408929478, + "grad_norm": 5.65625, + "learning_rate": 5.363760485430321e-06, + "loss": 1.6734705, + "memory(GiB)": 111.15, + "step": 39655, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.65738487, + "epoch": 1.0060882800608828, + "grad_norm": 5.5, + "learning_rate": 5.3627146325536725e-06, + "loss": 1.677174, + "memory(GiB)": 111.15, + "step": 39660, + "train_speed(iter/s)": 0.447913 + }, + { + "acc": 0.66984663, + "epoch": 1.0062151192288178, + "grad_norm": 5.6875, + "learning_rate": 5.361668763723216e-06, + "loss": 1.50745487, + "memory(GiB)": 111.15, + "step": 39665, + "train_speed(iter/s)": 0.447942 + }, + { + "acc": 0.65139236, + "epoch": 1.006341958396753, + "grad_norm": 5.1875, + "learning_rate": 5.360622878984954e-06, + "loss": 1.61013851, + "memory(GiB)": 111.15, + "step": 39670, + "train_speed(iter/s)": 0.447973 + }, + { + "acc": 0.64651446, + "epoch": 1.006468797564688, + "grad_norm": 5.15625, + "learning_rate": 5.359576978384885e-06, + "loss": 1.63361683, + "memory(GiB)": 111.15, + "step": 39675, + "train_speed(iter/s)": 0.448002 + }, + { + "acc": 0.65711145, + "epoch": 1.006595636732623, + "grad_norm": 5.0625, + "learning_rate": 5.358531061969018e-06, + "loss": 1.57543831, + "memory(GiB)": 111.15, + "step": 39680, + "train_speed(iter/s)": 0.448032 + }, + { + "acc": 0.65204124, + "epoch": 1.0067224759005582, + "grad_norm": 5.40625, + "learning_rate": 5.357485129783351e-06, + "loss": 1.70686646, + "memory(GiB)": 111.15, + "step": 39685, + "train_speed(iter/s)": 0.448061 + }, + { + "acc": 0.64723225, + "epoch": 1.0068493150684932, + "grad_norm": 5.65625, + "learning_rate": 5.356439181873895e-06, + "loss": 1.64063587, + "memory(GiB)": 111.15, + "step": 39690, + "train_speed(iter/s)": 0.448091 + }, + { + "acc": 0.66252131, + "epoch": 1.0069761542364282, + "grad_norm": 5.34375, + "learning_rate": 5.35539321828665e-06, + "loss": 1.56069984, + "memory(GiB)": 111.15, + "step": 39695, + "train_speed(iter/s)": 0.44812 + }, + { + "acc": 0.65507216, + "epoch": 1.0071029934043634, + "grad_norm": 5.03125, + "learning_rate": 5.354347239067625e-06, + "loss": 1.61719398, + "memory(GiB)": 111.15, + "step": 39700, + "train_speed(iter/s)": 0.44815 + }, + { + "epoch": 1.0071029934043634, + "eval_acc": 0.6432488593908033, + "eval_loss": 1.6081593036651611, + "eval_runtime": 114.113, + "eval_samples_per_second": 55.822, + "eval_steps_per_second": 27.911, + "step": 39700 + }, + { + "acc": 0.65907316, + "epoch": 1.0072298325722984, + "grad_norm": 5.90625, + "learning_rate": 5.3533012442628275e-06, + "loss": 1.55026798, + "memory(GiB)": 111.15, + "step": 39705, + "train_speed(iter/s)": 0.44758 + }, + { + "acc": 0.66580553, + "epoch": 1.0073566717402334, + "grad_norm": 5.96875, + "learning_rate": 5.3522552339182635e-06, + "loss": 1.55303135, + "memory(GiB)": 111.15, + "step": 39710, + "train_speed(iter/s)": 0.447609 + }, + { + "acc": 0.65620108, + "epoch": 1.0074835109081683, + "grad_norm": 5.59375, + "learning_rate": 5.351209208079941e-06, + "loss": 1.66382751, + "memory(GiB)": 111.15, + "step": 39715, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.65113745, + "epoch": 1.0076103500761036, + "grad_norm": 6.0, + "learning_rate": 5.35016316679387e-06, + "loss": 1.62946949, + "memory(GiB)": 111.15, + "step": 39720, + "train_speed(iter/s)": 0.447668 + }, + { + "acc": 0.66325445, + "epoch": 1.0077371892440385, + "grad_norm": 6.21875, + "learning_rate": 5.349117110106059e-06, + "loss": 1.50890064, + "memory(GiB)": 111.15, + "step": 39725, + "train_speed(iter/s)": 0.447697 + }, + { + "acc": 0.65340486, + "epoch": 1.0078640284119735, + "grad_norm": 5.1875, + "learning_rate": 5.34807103806252e-06, + "loss": 1.59762974, + "memory(GiB)": 111.15, + "step": 39730, + "train_speed(iter/s)": 0.447726 + }, + { + "acc": 0.64442921, + "epoch": 1.0079908675799087, + "grad_norm": 5.65625, + "learning_rate": 5.347024950709262e-06, + "loss": 1.63154182, + "memory(GiB)": 111.15, + "step": 39735, + "train_speed(iter/s)": 0.447756 + }, + { + "acc": 0.63947983, + "epoch": 1.0081177067478437, + "grad_norm": 4.90625, + "learning_rate": 5.345978848092297e-06, + "loss": 1.68277149, + "memory(GiB)": 111.15, + "step": 39740, + "train_speed(iter/s)": 0.447785 + }, + { + "acc": 0.64948325, + "epoch": 1.0082445459157787, + "grad_norm": 7.15625, + "learning_rate": 5.344932730257637e-06, + "loss": 1.60249767, + "memory(GiB)": 111.15, + "step": 39745, + "train_speed(iter/s)": 0.447815 + }, + { + "acc": 0.64680572, + "epoch": 1.008371385083714, + "grad_norm": 7.0625, + "learning_rate": 5.343886597251298e-06, + "loss": 1.65255222, + "memory(GiB)": 111.15, + "step": 39750, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.65496683, + "epoch": 1.008498224251649, + "grad_norm": 5.09375, + "learning_rate": 5.342840449119287e-06, + "loss": 1.64953575, + "memory(GiB)": 111.15, + "step": 39755, + "train_speed(iter/s)": 0.447873 + }, + { + "acc": 0.65032244, + "epoch": 1.008625063419584, + "grad_norm": 5.8125, + "learning_rate": 5.341794285907627e-06, + "loss": 1.62490177, + "memory(GiB)": 111.15, + "step": 39760, + "train_speed(iter/s)": 0.447903 + }, + { + "acc": 0.64107056, + "epoch": 1.0087519025875191, + "grad_norm": 4.875, + "learning_rate": 5.340748107662324e-06, + "loss": 1.67632484, + "memory(GiB)": 111.15, + "step": 39765, + "train_speed(iter/s)": 0.447932 + }, + { + "acc": 0.64493189, + "epoch": 1.0088787417554541, + "grad_norm": 4.90625, + "learning_rate": 5.339701914429402e-06, + "loss": 1.64155045, + "memory(GiB)": 111.15, + "step": 39770, + "train_speed(iter/s)": 0.447961 + }, + { + "acc": 0.64131594, + "epoch": 1.009005580923389, + "grad_norm": 5.78125, + "learning_rate": 5.338655706254871e-06, + "loss": 1.72737808, + "memory(GiB)": 111.15, + "step": 39775, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.64466929, + "epoch": 1.009132420091324, + "grad_norm": 4.3125, + "learning_rate": 5.33760948318475e-06, + "loss": 1.63873825, + "memory(GiB)": 111.15, + "step": 39780, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.66228123, + "epoch": 1.0092592592592593, + "grad_norm": 6.03125, + "learning_rate": 5.336563245265056e-06, + "loss": 1.63846722, + "memory(GiB)": 111.15, + "step": 39785, + "train_speed(iter/s)": 0.448048 + }, + { + "acc": 0.6400176, + "epoch": 1.0093860984271943, + "grad_norm": 5.75, + "learning_rate": 5.3355169925418095e-06, + "loss": 1.70039463, + "memory(GiB)": 111.15, + "step": 39790, + "train_speed(iter/s)": 0.448078 + }, + { + "acc": 0.64856739, + "epoch": 1.0095129375951293, + "grad_norm": 4.59375, + "learning_rate": 5.334470725061027e-06, + "loss": 1.6694088, + "memory(GiB)": 111.15, + "step": 39795, + "train_speed(iter/s)": 0.448107 + }, + { + "acc": 0.65386062, + "epoch": 1.0096397767630645, + "grad_norm": 5.46875, + "learning_rate": 5.333424442868729e-06, + "loss": 1.6667141, + "memory(GiB)": 111.15, + "step": 39800, + "train_speed(iter/s)": 0.448136 + }, + { + "epoch": 1.0096397767630645, + "eval_acc": 0.6433144488583882, + "eval_loss": 1.6081514358520508, + "eval_runtime": 114.0822, + "eval_samples_per_second": 55.837, + "eval_steps_per_second": 27.918, + "step": 39800 + }, + { + "acc": 0.64495335, + "epoch": 1.0097666159309995, + "grad_norm": 5.9375, + "learning_rate": 5.3323781460109345e-06, + "loss": 1.67381058, + "memory(GiB)": 111.15, + "step": 39805, + "train_speed(iter/s)": 0.447568 + }, + { + "acc": 0.65276031, + "epoch": 1.0098934550989345, + "grad_norm": 5.6875, + "learning_rate": 5.3313318345336665e-06, + "loss": 1.62089081, + "memory(GiB)": 111.15, + "step": 39810, + "train_speed(iter/s)": 0.447597 + }, + { + "acc": 0.66422386, + "epoch": 1.0100202942668697, + "grad_norm": 5.84375, + "learning_rate": 5.330285508482944e-06, + "loss": 1.61745186, + "memory(GiB)": 111.15, + "step": 39815, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.65481091, + "epoch": 1.0101471334348047, + "grad_norm": 7.03125, + "learning_rate": 5.3292391679047905e-06, + "loss": 1.59996452, + "memory(GiB)": 111.15, + "step": 39820, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.64787683, + "epoch": 1.0102739726027397, + "grad_norm": 4.78125, + "learning_rate": 5.328192812845228e-06, + "loss": 1.66148891, + "memory(GiB)": 111.15, + "step": 39825, + "train_speed(iter/s)": 0.447686 + }, + { + "acc": 0.65874586, + "epoch": 1.0104008117706749, + "grad_norm": 5.4375, + "learning_rate": 5.3271464433502805e-06, + "loss": 1.57473946, + "memory(GiB)": 111.15, + "step": 39830, + "train_speed(iter/s)": 0.447716 + }, + { + "acc": 0.64933901, + "epoch": 1.0105276509386099, + "grad_norm": 5.25, + "learning_rate": 5.3261000594659715e-06, + "loss": 1.65782299, + "memory(GiB)": 111.15, + "step": 39835, + "train_speed(iter/s)": 0.447745 + }, + { + "acc": 0.62788258, + "epoch": 1.0106544901065448, + "grad_norm": 5.40625, + "learning_rate": 5.3250536612383275e-06, + "loss": 1.71120052, + "memory(GiB)": 111.15, + "step": 39840, + "train_speed(iter/s)": 0.447775 + }, + { + "acc": 0.66780062, + "epoch": 1.01078132927448, + "grad_norm": 4.375, + "learning_rate": 5.32400724871337e-06, + "loss": 1.53020706, + "memory(GiB)": 111.15, + "step": 39845, + "train_speed(iter/s)": 0.447804 + }, + { + "acc": 0.65816269, + "epoch": 1.010908168442415, + "grad_norm": 5.09375, + "learning_rate": 5.322960821937129e-06, + "loss": 1.60437107, + "memory(GiB)": 111.15, + "step": 39850, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.6356698, + "epoch": 1.01103500761035, + "grad_norm": 6.53125, + "learning_rate": 5.321914380955628e-06, + "loss": 1.67131119, + "memory(GiB)": 111.15, + "step": 39855, + "train_speed(iter/s)": 0.447864 + }, + { + "acc": 0.65512314, + "epoch": 1.0111618467782852, + "grad_norm": 5.46875, + "learning_rate": 5.320867925814896e-06, + "loss": 1.62351112, + "memory(GiB)": 111.15, + "step": 39860, + "train_speed(iter/s)": 0.447893 + }, + { + "acc": 0.6532279, + "epoch": 1.0112886859462202, + "grad_norm": 5.5625, + "learning_rate": 5.31982145656096e-06, + "loss": 1.63477726, + "memory(GiB)": 111.15, + "step": 39865, + "train_speed(iter/s)": 0.447923 + }, + { + "acc": 0.65063858, + "epoch": 1.0114155251141552, + "grad_norm": 4.53125, + "learning_rate": 5.318774973239849e-06, + "loss": 1.57653027, + "memory(GiB)": 111.15, + "step": 39870, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.64717431, + "epoch": 1.0115423642820902, + "grad_norm": 5.03125, + "learning_rate": 5.31772847589759e-06, + "loss": 1.64279976, + "memory(GiB)": 111.15, + "step": 39875, + "train_speed(iter/s)": 0.447982 + }, + { + "acc": 0.66402674, + "epoch": 1.0116692034500254, + "grad_norm": 6.125, + "learning_rate": 5.316681964580215e-06, + "loss": 1.58774252, + "memory(GiB)": 111.15, + "step": 39880, + "train_speed(iter/s)": 0.448011 + }, + { + "acc": 0.66032367, + "epoch": 1.0117960426179604, + "grad_norm": 5.625, + "learning_rate": 5.315635439333753e-06, + "loss": 1.60470581, + "memory(GiB)": 111.15, + "step": 39885, + "train_speed(iter/s)": 0.44804 + }, + { + "acc": 0.64424925, + "epoch": 1.0119228817858954, + "grad_norm": 5.5, + "learning_rate": 5.314588900204235e-06, + "loss": 1.62823982, + "memory(GiB)": 111.15, + "step": 39890, + "train_speed(iter/s)": 0.44807 + }, + { + "acc": 0.65040388, + "epoch": 1.0120497209538306, + "grad_norm": 6.0625, + "learning_rate": 5.313542347237692e-06, + "loss": 1.68264332, + "memory(GiB)": 111.15, + "step": 39895, + "train_speed(iter/s)": 0.448099 + }, + { + "acc": 0.6289144, + "epoch": 1.0121765601217656, + "grad_norm": 6.4375, + "learning_rate": 5.312495780480159e-06, + "loss": 1.67910194, + "memory(GiB)": 111.15, + "step": 39900, + "train_speed(iter/s)": 0.448129 + }, + { + "epoch": 1.0121765601217656, + "eval_acc": 0.6433186265314828, + "eval_loss": 1.6081836223602295, + "eval_runtime": 115.0039, + "eval_samples_per_second": 55.389, + "eval_steps_per_second": 27.695, + "step": 39900 + }, + { + "acc": 0.65393057, + "epoch": 1.0123033992897006, + "grad_norm": 4.96875, + "learning_rate": 5.311449199977664e-06, + "loss": 1.61747627, + "memory(GiB)": 111.15, + "step": 39905, + "train_speed(iter/s)": 0.447557 + }, + { + "acc": 0.66160479, + "epoch": 1.0124302384576358, + "grad_norm": 5.875, + "learning_rate": 5.310402605776245e-06, + "loss": 1.59220953, + "memory(GiB)": 111.15, + "step": 39910, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.65786276, + "epoch": 1.0125570776255708, + "grad_norm": 5.0, + "learning_rate": 5.309355997921931e-06, + "loss": 1.59512062, + "memory(GiB)": 111.15, + "step": 39915, + "train_speed(iter/s)": 0.447615 + }, + { + "acc": 0.67027893, + "epoch": 1.0126839167935058, + "grad_norm": 5.125, + "learning_rate": 5.308309376460761e-06, + "loss": 1.61905785, + "memory(GiB)": 111.15, + "step": 39920, + "train_speed(iter/s)": 0.447644 + }, + { + "acc": 0.66347847, + "epoch": 1.012810755961441, + "grad_norm": 4.5625, + "learning_rate": 5.307262741438767e-06, + "loss": 1.5459487, + "memory(GiB)": 111.15, + "step": 39925, + "train_speed(iter/s)": 0.447674 + }, + { + "acc": 0.65626087, + "epoch": 1.012937595129376, + "grad_norm": 5.25, + "learning_rate": 5.3062160929019855e-06, + "loss": 1.63151112, + "memory(GiB)": 111.15, + "step": 39930, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.66607256, + "epoch": 1.013064434297311, + "grad_norm": 4.9375, + "learning_rate": 5.305169430896454e-06, + "loss": 1.57903204, + "memory(GiB)": 111.15, + "step": 39935, + "train_speed(iter/s)": 0.447733 + }, + { + "acc": 0.65347519, + "epoch": 1.013191273465246, + "grad_norm": 4.75, + "learning_rate": 5.304122755468209e-06, + "loss": 1.590135, + "memory(GiB)": 111.15, + "step": 39940, + "train_speed(iter/s)": 0.447762 + }, + { + "acc": 0.65314207, + "epoch": 1.0133181126331812, + "grad_norm": 5.875, + "learning_rate": 5.303076066663286e-06, + "loss": 1.5560667, + "memory(GiB)": 111.15, + "step": 39945, + "train_speed(iter/s)": 0.447791 + }, + { + "acc": 0.64245901, + "epoch": 1.0134449518011162, + "grad_norm": 5.28125, + "learning_rate": 5.302029364527726e-06, + "loss": 1.67264977, + "memory(GiB)": 111.15, + "step": 39950, + "train_speed(iter/s)": 0.447821 + }, + { + "acc": 0.64368553, + "epoch": 1.0135717909690511, + "grad_norm": 6.0, + "learning_rate": 5.3009826491075645e-06, + "loss": 1.68944817, + "memory(GiB)": 111.15, + "step": 39955, + "train_speed(iter/s)": 0.447851 + }, + { + "acc": 0.65242434, + "epoch": 1.0136986301369864, + "grad_norm": 5.53125, + "learning_rate": 5.299935920448843e-06, + "loss": 1.5475214, + "memory(GiB)": 111.15, + "step": 39960, + "train_speed(iter/s)": 0.44788 + }, + { + "acc": 0.65826397, + "epoch": 1.0138254693049213, + "grad_norm": 5.1875, + "learning_rate": 5.298889178597599e-06, + "loss": 1.58585119, + "memory(GiB)": 111.15, + "step": 39965, + "train_speed(iter/s)": 0.447909 + }, + { + "acc": 0.64279633, + "epoch": 1.0139523084728563, + "grad_norm": 6.25, + "learning_rate": 5.297842423599877e-06, + "loss": 1.62738323, + "memory(GiB)": 111.15, + "step": 39970, + "train_speed(iter/s)": 0.447939 + }, + { + "acc": 0.6587266, + "epoch": 1.0140791476407915, + "grad_norm": 5.75, + "learning_rate": 5.296795655501714e-06, + "loss": 1.64810314, + "memory(GiB)": 111.15, + "step": 39975, + "train_speed(iter/s)": 0.447969 + }, + { + "acc": 0.64716606, + "epoch": 1.0142059868087265, + "grad_norm": 5.59375, + "learning_rate": 5.295748874349155e-06, + "loss": 1.64768143, + "memory(GiB)": 111.15, + "step": 39980, + "train_speed(iter/s)": 0.447998 + }, + { + "acc": 0.64591742, + "epoch": 1.0143328259766615, + "grad_norm": 5.125, + "learning_rate": 5.294702080188236e-06, + "loss": 1.56741829, + "memory(GiB)": 111.15, + "step": 39985, + "train_speed(iter/s)": 0.448027 + }, + { + "acc": 0.64270277, + "epoch": 1.0144596651445967, + "grad_norm": 5.375, + "learning_rate": 5.293655273065008e-06, + "loss": 1.62689133, + "memory(GiB)": 111.15, + "step": 39990, + "train_speed(iter/s)": 0.448057 + }, + { + "acc": 0.65954552, + "epoch": 1.0145865043125317, + "grad_norm": 4.8125, + "learning_rate": 5.2926084530255076e-06, + "loss": 1.61682243, + "memory(GiB)": 111.15, + "step": 39995, + "train_speed(iter/s)": 0.448086 + }, + { + "acc": 0.65227718, + "epoch": 1.0147133434804667, + "grad_norm": 5.53125, + "learning_rate": 5.291561620115781e-06, + "loss": 1.60877037, + "memory(GiB)": 111.15, + "step": 40000, + "train_speed(iter/s)": 0.448116 + }, + { + "epoch": 1.0147133434804667, + "eval_acc": 0.6433315773180759, + "eval_loss": 1.6080617904663086, + "eval_runtime": 114.2967, + "eval_samples_per_second": 55.732, + "eval_steps_per_second": 27.866, + "step": 40000 + }, + { + "acc": 0.64977036, + "epoch": 1.014840182648402, + "grad_norm": 5.65625, + "learning_rate": 5.290514774381874e-06, + "loss": 1.65115051, + "memory(GiB)": 111.15, + "step": 40005, + "train_speed(iter/s)": 0.447549 + }, + { + "acc": 0.65103402, + "epoch": 1.014967021816337, + "grad_norm": 6.15625, + "learning_rate": 5.289467915869829e-06, + "loss": 1.63566742, + "memory(GiB)": 111.15, + "step": 40010, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.65834255, + "epoch": 1.015093860984272, + "grad_norm": 6.1875, + "learning_rate": 5.288421044625694e-06, + "loss": 1.61507778, + "memory(GiB)": 111.15, + "step": 40015, + "train_speed(iter/s)": 0.447608 + }, + { + "acc": 0.64797544, + "epoch": 1.0152207001522071, + "grad_norm": 5.09375, + "learning_rate": 5.287374160695513e-06, + "loss": 1.64077225, + "memory(GiB)": 111.15, + "step": 40020, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.64045777, + "epoch": 1.015347539320142, + "grad_norm": 4.84375, + "learning_rate": 5.286327264125332e-06, + "loss": 1.69212914, + "memory(GiB)": 111.15, + "step": 40025, + "train_speed(iter/s)": 0.447667 + }, + { + "acc": 0.6399971, + "epoch": 1.015474378488077, + "grad_norm": 5.5625, + "learning_rate": 5.285280354961202e-06, + "loss": 1.70978355, + "memory(GiB)": 111.15, + "step": 40030, + "train_speed(iter/s)": 0.447697 + }, + { + "acc": 0.65272961, + "epoch": 1.015601217656012, + "grad_norm": 5.3125, + "learning_rate": 5.284233433249167e-06, + "loss": 1.53898296, + "memory(GiB)": 111.15, + "step": 40035, + "train_speed(iter/s)": 0.447726 + }, + { + "acc": 0.64337845, + "epoch": 1.0157280568239473, + "grad_norm": 6.46875, + "learning_rate": 5.283186499035276e-06, + "loss": 1.5846879, + "memory(GiB)": 111.15, + "step": 40040, + "train_speed(iter/s)": 0.447755 + }, + { + "acc": 0.66363692, + "epoch": 1.0158548959918823, + "grad_norm": 5.1875, + "learning_rate": 5.2821395523655795e-06, + "loss": 1.55863438, + "memory(GiB)": 111.15, + "step": 40045, + "train_speed(iter/s)": 0.447785 + }, + { + "acc": 0.66166706, + "epoch": 1.0159817351598173, + "grad_norm": 5.4375, + "learning_rate": 5.281092593286127e-06, + "loss": 1.56614332, + "memory(GiB)": 111.15, + "step": 40050, + "train_speed(iter/s)": 0.447814 + }, + { + "acc": 0.64470692, + "epoch": 1.0161085743277525, + "grad_norm": 5.5625, + "learning_rate": 5.280045621842964e-06, + "loss": 1.60553131, + "memory(GiB)": 111.15, + "step": 40055, + "train_speed(iter/s)": 0.447843 + }, + { + "acc": 0.63212118, + "epoch": 1.0162354134956875, + "grad_norm": 4.65625, + "learning_rate": 5.278998638082148e-06, + "loss": 1.651824, + "memory(GiB)": 111.15, + "step": 40060, + "train_speed(iter/s)": 0.447872 + }, + { + "acc": 0.65166306, + "epoch": 1.0163622526636225, + "grad_norm": 4.75, + "learning_rate": 5.277951642049722e-06, + "loss": 1.62924156, + "memory(GiB)": 111.15, + "step": 40065, + "train_speed(iter/s)": 0.447902 + }, + { + "acc": 0.66398573, + "epoch": 1.0164890918315577, + "grad_norm": 4.65625, + "learning_rate": 5.276904633791745e-06, + "loss": 1.58433189, + "memory(GiB)": 111.15, + "step": 40070, + "train_speed(iter/s)": 0.447931 + }, + { + "acc": 0.65241776, + "epoch": 1.0166159309994927, + "grad_norm": 5.5625, + "learning_rate": 5.275857613354265e-06, + "loss": 1.58089972, + "memory(GiB)": 111.15, + "step": 40075, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.66117334, + "epoch": 1.0167427701674276, + "grad_norm": 4.625, + "learning_rate": 5.274810580783335e-06, + "loss": 1.58382664, + "memory(GiB)": 111.15, + "step": 40080, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.67078323, + "epoch": 1.0168696093353629, + "grad_norm": 5.21875, + "learning_rate": 5.2737635361250094e-06, + "loss": 1.57147865, + "memory(GiB)": 111.15, + "step": 40085, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.64773293, + "epoch": 1.0169964485032978, + "grad_norm": 5.59375, + "learning_rate": 5.2727164794253415e-06, + "loss": 1.62201405, + "memory(GiB)": 111.15, + "step": 40090, + "train_speed(iter/s)": 0.448049 + }, + { + "acc": 0.64658189, + "epoch": 1.0171232876712328, + "grad_norm": 5.4375, + "learning_rate": 5.271669410730384e-06, + "loss": 1.61984863, + "memory(GiB)": 111.15, + "step": 40095, + "train_speed(iter/s)": 0.448078 + }, + { + "acc": 0.66722693, + "epoch": 1.0172501268391678, + "grad_norm": 5.40625, + "learning_rate": 5.270622330086194e-06, + "loss": 1.60455856, + "memory(GiB)": 111.15, + "step": 40100, + "train_speed(iter/s)": 0.448107 + }, + { + "epoch": 1.0172501268391678, + "eval_acc": 0.6433441103373596, + "eval_loss": 1.6080188751220703, + "eval_runtime": 116.0743, + "eval_samples_per_second": 54.879, + "eval_steps_per_second": 27.439, + "step": 40100 + }, + { + "acc": 0.63481288, + "epoch": 1.017376966007103, + "grad_norm": 4.78125, + "learning_rate": 5.269575237538827e-06, + "loss": 1.62236443, + "memory(GiB)": 111.15, + "step": 40105, + "train_speed(iter/s)": 0.447532 + }, + { + "acc": 0.63913541, + "epoch": 1.017503805175038, + "grad_norm": 5.78125, + "learning_rate": 5.268528133134335e-06, + "loss": 1.67866859, + "memory(GiB)": 111.15, + "step": 40110, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.65746269, + "epoch": 1.017630644342973, + "grad_norm": 5.4375, + "learning_rate": 5.267481016918776e-06, + "loss": 1.57248344, + "memory(GiB)": 111.15, + "step": 40115, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.65432539, + "epoch": 1.0177574835109082, + "grad_norm": 5.65625, + "learning_rate": 5.266433888938212e-06, + "loss": 1.6101614, + "memory(GiB)": 111.15, + "step": 40120, + "train_speed(iter/s)": 0.44762 + }, + { + "acc": 0.66521573, + "epoch": 1.0178843226788432, + "grad_norm": 5.1875, + "learning_rate": 5.265386749238691e-06, + "loss": 1.57280483, + "memory(GiB)": 111.15, + "step": 40125, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.64449773, + "epoch": 1.0180111618467782, + "grad_norm": 5.5, + "learning_rate": 5.26433959786628e-06, + "loss": 1.61938801, + "memory(GiB)": 111.15, + "step": 40130, + "train_speed(iter/s)": 0.447678 + }, + { + "acc": 0.64395514, + "epoch": 1.0181380010147134, + "grad_norm": 5.875, + "learning_rate": 5.263292434867031e-06, + "loss": 1.60944977, + "memory(GiB)": 111.15, + "step": 40135, + "train_speed(iter/s)": 0.447708 + }, + { + "acc": 0.67059765, + "epoch": 1.0182648401826484, + "grad_norm": 4.5, + "learning_rate": 5.262245260287006e-06, + "loss": 1.52574415, + "memory(GiB)": 111.15, + "step": 40140, + "train_speed(iter/s)": 0.447737 + }, + { + "acc": 0.63782482, + "epoch": 1.0183916793505834, + "grad_norm": 7.4375, + "learning_rate": 5.261198074172262e-06, + "loss": 1.70457802, + "memory(GiB)": 111.15, + "step": 40145, + "train_speed(iter/s)": 0.447766 + }, + { + "acc": 0.65610008, + "epoch": 1.0185185185185186, + "grad_norm": 5.59375, + "learning_rate": 5.260150876568862e-06, + "loss": 1.64210548, + "memory(GiB)": 111.15, + "step": 40150, + "train_speed(iter/s)": 0.447796 + }, + { + "acc": 0.65463495, + "epoch": 1.0186453576864536, + "grad_norm": 4.53125, + "learning_rate": 5.259103667522866e-06, + "loss": 1.62486877, + "memory(GiB)": 111.15, + "step": 40155, + "train_speed(iter/s)": 0.447825 + }, + { + "acc": 0.65004196, + "epoch": 1.0187721968543886, + "grad_norm": 4.78125, + "learning_rate": 5.258056447080333e-06, + "loss": 1.68350677, + "memory(GiB)": 111.15, + "step": 40160, + "train_speed(iter/s)": 0.447854 + }, + { + "acc": 0.6483242, + "epoch": 1.0188990360223238, + "grad_norm": 5.0625, + "learning_rate": 5.257009215287325e-06, + "loss": 1.68151913, + "memory(GiB)": 111.15, + "step": 40165, + "train_speed(iter/s)": 0.447884 + }, + { + "acc": 0.65709085, + "epoch": 1.0190258751902588, + "grad_norm": 5.15625, + "learning_rate": 5.255961972189905e-06, + "loss": 1.62931232, + "memory(GiB)": 111.15, + "step": 40170, + "train_speed(iter/s)": 0.447913 + }, + { + "acc": 0.64324722, + "epoch": 1.0191527143581938, + "grad_norm": 5.53125, + "learning_rate": 5.254914717834133e-06, + "loss": 1.65946484, + "memory(GiB)": 111.15, + "step": 40175, + "train_speed(iter/s)": 0.447942 + }, + { + "acc": 0.64323854, + "epoch": 1.019279553526129, + "grad_norm": 4.84375, + "learning_rate": 5.253867452266075e-06, + "loss": 1.61372318, + "memory(GiB)": 111.15, + "step": 40180, + "train_speed(iter/s)": 0.447971 + }, + { + "acc": 0.66695662, + "epoch": 1.019406392694064, + "grad_norm": 6.5625, + "learning_rate": 5.252820175531792e-06, + "loss": 1.61986389, + "memory(GiB)": 111.15, + "step": 40185, + "train_speed(iter/s)": 0.448 + }, + { + "acc": 0.63874846, + "epoch": 1.019533231861999, + "grad_norm": 6.9375, + "learning_rate": 5.25177288767735e-06, + "loss": 1.65603523, + "memory(GiB)": 111.15, + "step": 40190, + "train_speed(iter/s)": 0.44803 + }, + { + "acc": 0.65913134, + "epoch": 1.019660071029934, + "grad_norm": 5.5625, + "learning_rate": 5.250725588748811e-06, + "loss": 1.59358406, + "memory(GiB)": 111.15, + "step": 40195, + "train_speed(iter/s)": 0.448059 + }, + { + "acc": 0.6444993, + "epoch": 1.0197869101978692, + "grad_norm": 5.15625, + "learning_rate": 5.249678278792243e-06, + "loss": 1.70611305, + "memory(GiB)": 111.15, + "step": 40200, + "train_speed(iter/s)": 0.448087 + }, + { + "epoch": 1.0197869101978692, + "eval_acc": 0.6433311595507665, + "eval_loss": 1.608157753944397, + "eval_runtime": 114.0998, + "eval_samples_per_second": 55.828, + "eval_steps_per_second": 27.914, + "step": 40200 + }, + { + "acc": 0.66980801, + "epoch": 1.0199137493658041, + "grad_norm": 5.375, + "learning_rate": 5.248630957853708e-06, + "loss": 1.57858696, + "memory(GiB)": 111.15, + "step": 40205, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.64949827, + "epoch": 1.0200405885337391, + "grad_norm": 5.28125, + "learning_rate": 5.247583625979276e-06, + "loss": 1.56204052, + "memory(GiB)": 111.15, + "step": 40210, + "train_speed(iter/s)": 0.447553 + }, + { + "acc": 0.63844352, + "epoch": 1.0201674277016743, + "grad_norm": 6.375, + "learning_rate": 5.246536283215007e-06, + "loss": 1.62905426, + "memory(GiB)": 111.15, + "step": 40215, + "train_speed(iter/s)": 0.447583 + }, + { + "acc": 0.64625292, + "epoch": 1.0202942668696093, + "grad_norm": 5.28125, + "learning_rate": 5.245488929606974e-06, + "loss": 1.64053345, + "memory(GiB)": 111.15, + "step": 40220, + "train_speed(iter/s)": 0.447612 + }, + { + "acc": 0.64884081, + "epoch": 1.0204211060375443, + "grad_norm": 4.34375, + "learning_rate": 5.244441565201241e-06, + "loss": 1.6384655, + "memory(GiB)": 111.15, + "step": 40225, + "train_speed(iter/s)": 0.447641 + }, + { + "acc": 0.64363327, + "epoch": 1.0205479452054795, + "grad_norm": 5.375, + "learning_rate": 5.243394190043877e-06, + "loss": 1.56483517, + "memory(GiB)": 111.15, + "step": 40230, + "train_speed(iter/s)": 0.44767 + }, + { + "acc": 0.65625615, + "epoch": 1.0206747843734145, + "grad_norm": 5.28125, + "learning_rate": 5.242346804180949e-06, + "loss": 1.60521927, + "memory(GiB)": 111.15, + "step": 40235, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.65169392, + "epoch": 1.0208016235413495, + "grad_norm": 5.21875, + "learning_rate": 5.241299407658528e-06, + "loss": 1.62840633, + "memory(GiB)": 111.15, + "step": 40240, + "train_speed(iter/s)": 0.447728 + }, + { + "acc": 0.65289021, + "epoch": 1.0209284627092847, + "grad_norm": 5.125, + "learning_rate": 5.240252000522681e-06, + "loss": 1.57140865, + "memory(GiB)": 111.15, + "step": 40245, + "train_speed(iter/s)": 0.447757 + }, + { + "acc": 0.64046588, + "epoch": 1.0210553018772197, + "grad_norm": 5.96875, + "learning_rate": 5.239204582819479e-06, + "loss": 1.66789856, + "memory(GiB)": 111.15, + "step": 40250, + "train_speed(iter/s)": 0.447786 + }, + { + "acc": 0.64328222, + "epoch": 1.0211821410451547, + "grad_norm": 6.40625, + "learning_rate": 5.238157154594989e-06, + "loss": 1.62314739, + "memory(GiB)": 111.15, + "step": 40255, + "train_speed(iter/s)": 0.447816 + }, + { + "acc": 0.66804323, + "epoch": 1.0213089802130897, + "grad_norm": 5.28125, + "learning_rate": 5.237109715895287e-06, + "loss": 1.50406475, + "memory(GiB)": 111.15, + "step": 40260, + "train_speed(iter/s)": 0.447845 + }, + { + "acc": 0.65789499, + "epoch": 1.021435819381025, + "grad_norm": 5.34375, + "learning_rate": 5.2360622667664385e-06, + "loss": 1.59499855, + "memory(GiB)": 111.15, + "step": 40265, + "train_speed(iter/s)": 0.447874 + }, + { + "acc": 0.65987034, + "epoch": 1.02156265854896, + "grad_norm": 5.8125, + "learning_rate": 5.235014807254521e-06, + "loss": 1.56635761, + "memory(GiB)": 111.15, + "step": 40270, + "train_speed(iter/s)": 0.447903 + }, + { + "acc": 0.63681631, + "epoch": 1.0216894977168949, + "grad_norm": 5.9375, + "learning_rate": 5.233967337405599e-06, + "loss": 1.6512434, + "memory(GiB)": 111.15, + "step": 40275, + "train_speed(iter/s)": 0.447932 + }, + { + "acc": 0.65742435, + "epoch": 1.02181633688483, + "grad_norm": 4.9375, + "learning_rate": 5.232919857265752e-06, + "loss": 1.6080864, + "memory(GiB)": 111.15, + "step": 40280, + "train_speed(iter/s)": 0.447961 + }, + { + "acc": 0.64107962, + "epoch": 1.021943176052765, + "grad_norm": 5.125, + "learning_rate": 5.231872366881048e-06, + "loss": 1.63518219, + "memory(GiB)": 111.15, + "step": 40285, + "train_speed(iter/s)": 0.44799 + }, + { + "acc": 0.65424418, + "epoch": 1.0220700152207, + "grad_norm": 4.84375, + "learning_rate": 5.230824866297563e-06, + "loss": 1.61906376, + "memory(GiB)": 111.15, + "step": 40290, + "train_speed(iter/s)": 0.448019 + }, + { + "acc": 0.65756526, + "epoch": 1.0221968543886353, + "grad_norm": 5.5625, + "learning_rate": 5.229777355561368e-06, + "loss": 1.60591393, + "memory(GiB)": 111.15, + "step": 40295, + "train_speed(iter/s)": 0.448048 + }, + { + "acc": 0.65153389, + "epoch": 1.0223236935565703, + "grad_norm": 6.25, + "learning_rate": 5.2287298347185415e-06, + "loss": 1.60807457, + "memory(GiB)": 111.15, + "step": 40300, + "train_speed(iter/s)": 0.448078 + }, + { + "epoch": 1.0223236935565703, + "eval_acc": 0.6433608210297379, + "eval_loss": 1.608086347579956, + "eval_runtime": 113.6276, + "eval_samples_per_second": 56.06, + "eval_steps_per_second": 28.03, + "step": 40300 + }, + { + "acc": 0.63398457, + "epoch": 1.0224505327245053, + "grad_norm": 5.5, + "learning_rate": 5.227682303815155e-06, + "loss": 1.6592289, + "memory(GiB)": 111.15, + "step": 40305, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.64890108, + "epoch": 1.0225773718924405, + "grad_norm": 4.78125, + "learning_rate": 5.226634762897284e-06, + "loss": 1.66181583, + "memory(GiB)": 111.15, + "step": 40310, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.64910316, + "epoch": 1.0227042110603755, + "grad_norm": 5.125, + "learning_rate": 5.225587212011004e-06, + "loss": 1.57215853, + "memory(GiB)": 111.15, + "step": 40315, + "train_speed(iter/s)": 0.447576 + }, + { + "acc": 0.65539274, + "epoch": 1.0228310502283104, + "grad_norm": 5.25, + "learning_rate": 5.224539651202391e-06, + "loss": 1.59055653, + "memory(GiB)": 111.15, + "step": 40320, + "train_speed(iter/s)": 0.447605 + }, + { + "acc": 0.64403586, + "epoch": 1.0229578893962457, + "grad_norm": 4.9375, + "learning_rate": 5.223492080517523e-06, + "loss": 1.6279438, + "memory(GiB)": 111.15, + "step": 40325, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.64567447, + "epoch": 1.0230847285641806, + "grad_norm": 6.5625, + "learning_rate": 5.2224445000024744e-06, + "loss": 1.69446888, + "memory(GiB)": 111.15, + "step": 40330, + "train_speed(iter/s)": 0.447663 + }, + { + "acc": 0.6426548, + "epoch": 1.0232115677321156, + "grad_norm": 5.78125, + "learning_rate": 5.221396909703322e-06, + "loss": 1.60868073, + "memory(GiB)": 111.15, + "step": 40335, + "train_speed(iter/s)": 0.447692 + }, + { + "acc": 0.65774355, + "epoch": 1.0233384069000508, + "grad_norm": 5.4375, + "learning_rate": 5.220349309666148e-06, + "loss": 1.55166941, + "memory(GiB)": 111.15, + "step": 40340, + "train_speed(iter/s)": 0.447721 + }, + { + "acc": 0.64283972, + "epoch": 1.0234652460679858, + "grad_norm": 6.3125, + "learning_rate": 5.2193016999370265e-06, + "loss": 1.64165573, + "memory(GiB)": 111.15, + "step": 40345, + "train_speed(iter/s)": 0.447749 + }, + { + "acc": 0.64905815, + "epoch": 1.0235920852359208, + "grad_norm": 5.03125, + "learning_rate": 5.218254080562038e-06, + "loss": 1.58729601, + "memory(GiB)": 111.15, + "step": 40350, + "train_speed(iter/s)": 0.447778 + }, + { + "acc": 0.67201099, + "epoch": 1.0237189244038558, + "grad_norm": 6.3125, + "learning_rate": 5.2172064515872585e-06, + "loss": 1.55817757, + "memory(GiB)": 111.15, + "step": 40355, + "train_speed(iter/s)": 0.447808 + }, + { + "acc": 0.6558238, + "epoch": 1.023845763571791, + "grad_norm": 6.125, + "learning_rate": 5.21615881305877e-06, + "loss": 1.62978954, + "memory(GiB)": 111.15, + "step": 40360, + "train_speed(iter/s)": 0.447837 + }, + { + "acc": 0.64109697, + "epoch": 1.023972602739726, + "grad_norm": 5.25, + "learning_rate": 5.215111165022653e-06, + "loss": 1.60632095, + "memory(GiB)": 111.15, + "step": 40365, + "train_speed(iter/s)": 0.447865 + }, + { + "acc": 0.66899137, + "epoch": 1.024099441907661, + "grad_norm": 5.0, + "learning_rate": 5.2140635075249856e-06, + "loss": 1.53117485, + "memory(GiB)": 111.15, + "step": 40370, + "train_speed(iter/s)": 0.447894 + }, + { + "acc": 0.6543644, + "epoch": 1.0242262810755962, + "grad_norm": 5.5, + "learning_rate": 5.213015840611851e-06, + "loss": 1.60895348, + "memory(GiB)": 111.15, + "step": 40375, + "train_speed(iter/s)": 0.447923 + }, + { + "acc": 0.6459518, + "epoch": 1.0243531202435312, + "grad_norm": 6.25, + "learning_rate": 5.211968164329328e-06, + "loss": 1.67447128, + "memory(GiB)": 111.15, + "step": 40380, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.65584316, + "epoch": 1.0244799594114662, + "grad_norm": 5.3125, + "learning_rate": 5.210920478723497e-06, + "loss": 1.63530731, + "memory(GiB)": 111.15, + "step": 40385, + "train_speed(iter/s)": 0.447981 + }, + { + "acc": 0.63666363, + "epoch": 1.0246067985794014, + "grad_norm": 5.3125, + "learning_rate": 5.209872783840443e-06, + "loss": 1.68106613, + "memory(GiB)": 111.15, + "step": 40390, + "train_speed(iter/s)": 0.448009 + }, + { + "acc": 0.65629983, + "epoch": 1.0247336377473364, + "grad_norm": 4.5625, + "learning_rate": 5.208825079726248e-06, + "loss": 1.6246357, + "memory(GiB)": 111.15, + "step": 40395, + "train_speed(iter/s)": 0.448039 + }, + { + "acc": 0.65562038, + "epoch": 1.0248604769152714, + "grad_norm": 5.125, + "learning_rate": 5.207777366426992e-06, + "loss": 1.61266136, + "memory(GiB)": 111.15, + "step": 40400, + "train_speed(iter/s)": 0.448067 + }, + { + "epoch": 1.0248604769152714, + "eval_acc": 0.6432822807755599, + "eval_loss": 1.6082926988601685, + "eval_runtime": 114.8666, + "eval_samples_per_second": 55.456, + "eval_steps_per_second": 27.728, + "step": 40400 + }, + { + "acc": 0.65114346, + "epoch": 1.0249873160832066, + "grad_norm": 5.3125, + "learning_rate": 5.206729643988759e-06, + "loss": 1.6144207, + "memory(GiB)": 111.15, + "step": 40405, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.65620499, + "epoch": 1.0251141552511416, + "grad_norm": 5.21875, + "learning_rate": 5.205681912457635e-06, + "loss": 1.5876009, + "memory(GiB)": 111.15, + "step": 40410, + "train_speed(iter/s)": 0.447533 + }, + { + "acc": 0.63470745, + "epoch": 1.0252409944190766, + "grad_norm": 6.6875, + "learning_rate": 5.204634171879701e-06, + "loss": 1.71323051, + "memory(GiB)": 111.15, + "step": 40415, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.64604206, + "epoch": 1.0253678335870116, + "grad_norm": 4.59375, + "learning_rate": 5.2035864223010445e-06, + "loss": 1.52141266, + "memory(GiB)": 111.15, + "step": 40420, + "train_speed(iter/s)": 0.447559 + }, + { + "acc": 0.65233707, + "epoch": 1.0254946727549468, + "grad_norm": 5.59375, + "learning_rate": 5.202538663767746e-06, + "loss": 1.62205582, + "memory(GiB)": 111.15, + "step": 40425, + "train_speed(iter/s)": 0.447589 + }, + { + "acc": 0.64416108, + "epoch": 1.0256215119228818, + "grad_norm": 5.53125, + "learning_rate": 5.201490896325895e-06, + "loss": 1.64863739, + "memory(GiB)": 111.15, + "step": 40430, + "train_speed(iter/s)": 0.447618 + }, + { + "acc": 0.66002836, + "epoch": 1.0257483510908167, + "grad_norm": 5.125, + "learning_rate": 5.200443120021572e-06, + "loss": 1.58110418, + "memory(GiB)": 111.15, + "step": 40435, + "train_speed(iter/s)": 0.447648 + }, + { + "acc": 0.66408958, + "epoch": 1.025875190258752, + "grad_norm": 5.3125, + "learning_rate": 5.199395334900868e-06, + "loss": 1.59011822, + "memory(GiB)": 111.15, + "step": 40440, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.65090561, + "epoch": 1.026002029426687, + "grad_norm": 6.1875, + "learning_rate": 5.198347541009866e-06, + "loss": 1.59583721, + "memory(GiB)": 111.15, + "step": 40445, + "train_speed(iter/s)": 0.447707 + }, + { + "acc": 0.648947, + "epoch": 1.026128868594622, + "grad_norm": 5.875, + "learning_rate": 5.197299738394654e-06, + "loss": 1.66350498, + "memory(GiB)": 111.15, + "step": 40450, + "train_speed(iter/s)": 0.447736 + }, + { + "acc": 0.64902525, + "epoch": 1.0262557077625571, + "grad_norm": 5.03125, + "learning_rate": 5.196251927101318e-06, + "loss": 1.63024521, + "memory(GiB)": 111.15, + "step": 40455, + "train_speed(iter/s)": 0.447765 + }, + { + "acc": 0.65747595, + "epoch": 1.0263825469304921, + "grad_norm": 5.90625, + "learning_rate": 5.195204107175946e-06, + "loss": 1.51528168, + "memory(GiB)": 111.15, + "step": 40460, + "train_speed(iter/s)": 0.447795 + }, + { + "acc": 0.66066685, + "epoch": 1.0265093860984271, + "grad_norm": 5.8125, + "learning_rate": 5.194156278664627e-06, + "loss": 1.63173485, + "memory(GiB)": 111.15, + "step": 40465, + "train_speed(iter/s)": 0.447824 + }, + { + "acc": 0.65442181, + "epoch": 1.0266362252663623, + "grad_norm": 4.75, + "learning_rate": 5.1931084416134466e-06, + "loss": 1.60343056, + "memory(GiB)": 111.15, + "step": 40470, + "train_speed(iter/s)": 0.447854 + }, + { + "acc": 0.64815531, + "epoch": 1.0267630644342973, + "grad_norm": 4.40625, + "learning_rate": 5.192060596068496e-06, + "loss": 1.63886967, + "memory(GiB)": 111.15, + "step": 40475, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.65427256, + "epoch": 1.0268899036022323, + "grad_norm": 5.21875, + "learning_rate": 5.191012742075863e-06, + "loss": 1.61426926, + "memory(GiB)": 111.15, + "step": 40480, + "train_speed(iter/s)": 0.447913 + }, + { + "acc": 0.66056013, + "epoch": 1.0270167427701675, + "grad_norm": 7.3125, + "learning_rate": 5.189964879681635e-06, + "loss": 1.65126152, + "memory(GiB)": 111.15, + "step": 40485, + "train_speed(iter/s)": 0.447942 + }, + { + "acc": 0.64422908, + "epoch": 1.0271435819381025, + "grad_norm": 5.53125, + "learning_rate": 5.188917008931905e-06, + "loss": 1.62142105, + "memory(GiB)": 111.15, + "step": 40490, + "train_speed(iter/s)": 0.447972 + }, + { + "acc": 0.64890475, + "epoch": 1.0272704211060375, + "grad_norm": 6.46875, + "learning_rate": 5.18786912987276e-06, + "loss": 1.59351673, + "memory(GiB)": 111.15, + "step": 40495, + "train_speed(iter/s)": 0.448001 + }, + { + "acc": 0.65617952, + "epoch": 1.0273972602739727, + "grad_norm": 5.28125, + "learning_rate": 5.186821242550294e-06, + "loss": 1.57704592, + "memory(GiB)": 111.15, + "step": 40500, + "train_speed(iter/s)": 0.448031 + }, + { + "epoch": 1.0273972602739727, + "eval_acc": 0.643355390054715, + "eval_loss": 1.60795259475708, + "eval_runtime": 112.9602, + "eval_samples_per_second": 56.392, + "eval_steps_per_second": 28.196, + "step": 40500 + }, + { + "acc": 0.63562841, + "epoch": 1.0275240994419077, + "grad_norm": 5.375, + "learning_rate": 5.185773347010594e-06, + "loss": 1.65379295, + "memory(GiB)": 111.15, + "step": 40505, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.65012007, + "epoch": 1.0276509386098427, + "grad_norm": 5.03125, + "learning_rate": 5.184725443299753e-06, + "loss": 1.63796329, + "memory(GiB)": 111.15, + "step": 40510, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.65802174, + "epoch": 1.0277777777777777, + "grad_norm": 5.59375, + "learning_rate": 5.183677531463863e-06, + "loss": 1.60880356, + "memory(GiB)": 111.15, + "step": 40515, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.63450532, + "epoch": 1.027904616945713, + "grad_norm": 5.15625, + "learning_rate": 5.182629611549015e-06, + "loss": 1.66586418, + "memory(GiB)": 111.15, + "step": 40520, + "train_speed(iter/s)": 0.447566 + }, + { + "acc": 0.64967351, + "epoch": 1.0280314561136479, + "grad_norm": 5.59375, + "learning_rate": 5.181581683601301e-06, + "loss": 1.57608366, + "memory(GiB)": 111.15, + "step": 40525, + "train_speed(iter/s)": 0.447596 + }, + { + "acc": 0.65348783, + "epoch": 1.0281582952815829, + "grad_norm": 6.125, + "learning_rate": 5.1805337476668135e-06, + "loss": 1.71041298, + "memory(GiB)": 111.15, + "step": 40530, + "train_speed(iter/s)": 0.447625 + }, + { + "acc": 0.65535736, + "epoch": 1.028285134449518, + "grad_norm": 3.953125, + "learning_rate": 5.179485803791646e-06, + "loss": 1.63384514, + "memory(GiB)": 111.15, + "step": 40535, + "train_speed(iter/s)": 0.447654 + }, + { + "acc": 0.64854808, + "epoch": 1.028411973617453, + "grad_norm": 5.34375, + "learning_rate": 5.178437852021892e-06, + "loss": 1.69926491, + "memory(GiB)": 111.15, + "step": 40540, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.65768862, + "epoch": 1.028538812785388, + "grad_norm": 5.09375, + "learning_rate": 5.177389892403645e-06, + "loss": 1.59175692, + "memory(GiB)": 111.15, + "step": 40545, + "train_speed(iter/s)": 0.447713 + }, + { + "acc": 0.6422122, + "epoch": 1.0286656519533233, + "grad_norm": 5.21875, + "learning_rate": 5.176341924982997e-06, + "loss": 1.60292702, + "memory(GiB)": 111.15, + "step": 40550, + "train_speed(iter/s)": 0.447742 + }, + { + "acc": 0.6605608, + "epoch": 1.0287924911212583, + "grad_norm": 5.65625, + "learning_rate": 5.1752939498060435e-06, + "loss": 1.58780041, + "memory(GiB)": 111.15, + "step": 40555, + "train_speed(iter/s)": 0.447771 + }, + { + "acc": 0.64195137, + "epoch": 1.0289193302891932, + "grad_norm": 5.28125, + "learning_rate": 5.174245966918883e-06, + "loss": 1.70906525, + "memory(GiB)": 111.15, + "step": 40560, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.63318062, + "epoch": 1.0290461694571285, + "grad_norm": 5.15625, + "learning_rate": 5.173197976367603e-06, + "loss": 1.62271614, + "memory(GiB)": 111.15, + "step": 40565, + "train_speed(iter/s)": 0.44783 + }, + { + "acc": 0.65527067, + "epoch": 1.0291730086250634, + "grad_norm": 5.28125, + "learning_rate": 5.1721499781983055e-06, + "loss": 1.5782361, + "memory(GiB)": 111.15, + "step": 40570, + "train_speed(iter/s)": 0.44786 + }, + { + "acc": 0.64043713, + "epoch": 1.0292998477929984, + "grad_norm": 4.6875, + "learning_rate": 5.171101972457081e-06, + "loss": 1.62701206, + "memory(GiB)": 111.15, + "step": 40575, + "train_speed(iter/s)": 0.447889 + }, + { + "acc": 0.64376483, + "epoch": 1.0294266869609334, + "grad_norm": 5.09375, + "learning_rate": 5.170053959190029e-06, + "loss": 1.63317795, + "memory(GiB)": 111.15, + "step": 40580, + "train_speed(iter/s)": 0.447918 + }, + { + "acc": 0.64261417, + "epoch": 1.0295535261288686, + "grad_norm": 6.0, + "learning_rate": 5.169005938443245e-06, + "loss": 1.64981098, + "memory(GiB)": 111.15, + "step": 40585, + "train_speed(iter/s)": 0.447948 + }, + { + "acc": 0.65794153, + "epoch": 1.0296803652968036, + "grad_norm": 5.25, + "learning_rate": 5.1679579102628245e-06, + "loss": 1.62129059, + "memory(GiB)": 111.15, + "step": 40590, + "train_speed(iter/s)": 0.447977 + }, + { + "acc": 0.64069643, + "epoch": 1.0298072044647386, + "grad_norm": 5.1875, + "learning_rate": 5.166909874694866e-06, + "loss": 1.63066673, + "memory(GiB)": 111.15, + "step": 40595, + "train_speed(iter/s)": 0.448007 + }, + { + "acc": 0.6375, + "epoch": 1.0299340436326738, + "grad_norm": 7.0, + "learning_rate": 5.165861831785465e-06, + "loss": 1.74127274, + "memory(GiB)": 111.15, + "step": 40600, + "train_speed(iter/s)": 0.448036 + }, + { + "epoch": 1.0299340436326738, + "eval_acc": 0.6433345016892421, + "eval_loss": 1.608051061630249, + "eval_runtime": 112.9986, + "eval_samples_per_second": 56.372, + "eval_steps_per_second": 28.186, + "step": 40600 + }, + { + "acc": 0.63999095, + "epoch": 1.0300608828006088, + "grad_norm": 5.5, + "learning_rate": 5.164813781580721e-06, + "loss": 1.64957886, + "memory(GiB)": 111.15, + "step": 40605, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64852219, + "epoch": 1.0301877219685438, + "grad_norm": 4.96875, + "learning_rate": 5.16376572412673e-06, + "loss": 1.60484295, + "memory(GiB)": 111.15, + "step": 40610, + "train_speed(iter/s)": 0.447513 + }, + { + "acc": 0.66338205, + "epoch": 1.030314561136479, + "grad_norm": 5.375, + "learning_rate": 5.162717659469593e-06, + "loss": 1.64841385, + "memory(GiB)": 111.15, + "step": 40615, + "train_speed(iter/s)": 0.447542 + }, + { + "acc": 0.66090255, + "epoch": 1.030441400304414, + "grad_norm": 5.3125, + "learning_rate": 5.161669587655406e-06, + "loss": 1.55668163, + "memory(GiB)": 111.15, + "step": 40620, + "train_speed(iter/s)": 0.447571 + }, + { + "acc": 0.65121937, + "epoch": 1.030568239472349, + "grad_norm": 6.625, + "learning_rate": 5.160621508730267e-06, + "loss": 1.6102356, + "memory(GiB)": 111.15, + "step": 40625, + "train_speed(iter/s)": 0.4476 + }, + { + "acc": 0.64548273, + "epoch": 1.0306950786402842, + "grad_norm": 5.375, + "learning_rate": 5.15957342274028e-06, + "loss": 1.65638199, + "memory(GiB)": 111.15, + "step": 40630, + "train_speed(iter/s)": 0.447629 + }, + { + "acc": 0.64255619, + "epoch": 1.0308219178082192, + "grad_norm": 6.78125, + "learning_rate": 5.158525329731539e-06, + "loss": 1.63898296, + "memory(GiB)": 111.15, + "step": 40635, + "train_speed(iter/s)": 0.447659 + }, + { + "acc": 0.64803109, + "epoch": 1.0309487569761542, + "grad_norm": 6.125, + "learning_rate": 5.157477229750149e-06, + "loss": 1.70335655, + "memory(GiB)": 111.15, + "step": 40640, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.6595911, + "epoch": 1.0310755961440894, + "grad_norm": 5.15625, + "learning_rate": 5.156429122842204e-06, + "loss": 1.64078808, + "memory(GiB)": 111.15, + "step": 40645, + "train_speed(iter/s)": 0.447718 + }, + { + "acc": 0.65450773, + "epoch": 1.0312024353120244, + "grad_norm": 5.875, + "learning_rate": 5.15538100905381e-06, + "loss": 1.58373766, + "memory(GiB)": 111.15, + "step": 40650, + "train_speed(iter/s)": 0.447747 + }, + { + "acc": 0.65133629, + "epoch": 1.0313292744799594, + "grad_norm": 4.4375, + "learning_rate": 5.154332888431064e-06, + "loss": 1.61011696, + "memory(GiB)": 111.15, + "step": 40655, + "train_speed(iter/s)": 0.447776 + }, + { + "acc": 0.64573202, + "epoch": 1.0314561136478946, + "grad_norm": 5.71875, + "learning_rate": 5.15328476102007e-06, + "loss": 1.65489025, + "memory(GiB)": 111.15, + "step": 40660, + "train_speed(iter/s)": 0.447805 + }, + { + "acc": 0.64169312, + "epoch": 1.0315829528158296, + "grad_norm": 4.75, + "learning_rate": 5.1522366268669264e-06, + "loss": 1.6431839, + "memory(GiB)": 111.15, + "step": 40665, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.64894066, + "epoch": 1.0317097919837646, + "grad_norm": 5.34375, + "learning_rate": 5.1511884860177376e-06, + "loss": 1.60927906, + "memory(GiB)": 111.15, + "step": 40670, + "train_speed(iter/s)": 0.447864 + }, + { + "acc": 0.65176334, + "epoch": 1.0318366311516995, + "grad_norm": 5.75, + "learning_rate": 5.150140338518603e-06, + "loss": 1.63491745, + "memory(GiB)": 111.15, + "step": 40675, + "train_speed(iter/s)": 0.447893 + }, + { + "acc": 0.65949211, + "epoch": 1.0319634703196348, + "grad_norm": 4.5625, + "learning_rate": 5.149092184415627e-06, + "loss": 1.52849216, + "memory(GiB)": 111.15, + "step": 40680, + "train_speed(iter/s)": 0.447922 + }, + { + "acc": 0.67078447, + "epoch": 1.0320903094875697, + "grad_norm": 5.0, + "learning_rate": 5.148044023754911e-06, + "loss": 1.58773632, + "memory(GiB)": 111.15, + "step": 40685, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.63752394, + "epoch": 1.0322171486555047, + "grad_norm": 4.53125, + "learning_rate": 5.146995856582557e-06, + "loss": 1.65227089, + "memory(GiB)": 111.15, + "step": 40690, + "train_speed(iter/s)": 0.44798 + }, + { + "acc": 0.64474263, + "epoch": 1.03234398782344, + "grad_norm": 5.21875, + "learning_rate": 5.14594768294467e-06, + "loss": 1.68951874, + "memory(GiB)": 111.15, + "step": 40695, + "train_speed(iter/s)": 0.44801 + }, + { + "acc": 0.65284805, + "epoch": 1.032470826991375, + "grad_norm": 5.71875, + "learning_rate": 5.1448995028873515e-06, + "loss": 1.61398735, + "memory(GiB)": 111.15, + "step": 40700, + "train_speed(iter/s)": 0.448039 + }, + { + "epoch": 1.032470826991375, + "eval_acc": 0.643354554520096, + "eval_loss": 1.6081308126449585, + "eval_runtime": 114.0068, + "eval_samples_per_second": 55.874, + "eval_steps_per_second": 27.937, + "step": 40700 + }, + { + "acc": 0.64201126, + "epoch": 1.03259766615931, + "grad_norm": 5.3125, + "learning_rate": 5.143851316456706e-06, + "loss": 1.5807785, + "memory(GiB)": 111.15, + "step": 40705, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64368792, + "epoch": 1.0327245053272451, + "grad_norm": 6.0625, + "learning_rate": 5.142803123698838e-06, + "loss": 1.67756577, + "memory(GiB)": 111.15, + "step": 40710, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.65963807, + "epoch": 1.0328513444951801, + "grad_norm": 5.4375, + "learning_rate": 5.14175492465985e-06, + "loss": 1.57015285, + "memory(GiB)": 111.15, + "step": 40715, + "train_speed(iter/s)": 0.447541 + }, + { + "acc": 0.6568099, + "epoch": 1.0329781836631151, + "grad_norm": 5.375, + "learning_rate": 5.14070671938585e-06, + "loss": 1.61709709, + "memory(GiB)": 111.15, + "step": 40720, + "train_speed(iter/s)": 0.44757 + }, + { + "acc": 0.65632639, + "epoch": 1.0331050228310503, + "grad_norm": 4.96875, + "learning_rate": 5.139658507922937e-06, + "loss": 1.6002697, + "memory(GiB)": 111.15, + "step": 40725, + "train_speed(iter/s)": 0.447598 + }, + { + "acc": 0.64592943, + "epoch": 1.0332318619989853, + "grad_norm": 6.03125, + "learning_rate": 5.138610290317221e-06, + "loss": 1.67325859, + "memory(GiB)": 111.15, + "step": 40730, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.64382057, + "epoch": 1.0333587011669203, + "grad_norm": 5.40625, + "learning_rate": 5.137562066614805e-06, + "loss": 1.64746399, + "memory(GiB)": 111.15, + "step": 40735, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.65734453, + "epoch": 1.0334855403348553, + "grad_norm": 7.0625, + "learning_rate": 5.136513836861795e-06, + "loss": 1.51711617, + "memory(GiB)": 111.15, + "step": 40740, + "train_speed(iter/s)": 0.447685 + }, + { + "acc": 0.65307245, + "epoch": 1.0336123795027905, + "grad_norm": 5.0, + "learning_rate": 5.135465601104298e-06, + "loss": 1.66664238, + "memory(GiB)": 111.15, + "step": 40745, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.68397794, + "epoch": 1.0337392186707255, + "grad_norm": 4.96875, + "learning_rate": 5.134417359388418e-06, + "loss": 1.56502991, + "memory(GiB)": 111.15, + "step": 40750, + "train_speed(iter/s)": 0.447743 + }, + { + "acc": 0.64395456, + "epoch": 1.0338660578386605, + "grad_norm": 5.5, + "learning_rate": 5.133369111760264e-06, + "loss": 1.6396513, + "memory(GiB)": 111.15, + "step": 40755, + "train_speed(iter/s)": 0.447772 + }, + { + "acc": 0.66405067, + "epoch": 1.0339928970065957, + "grad_norm": 5.46875, + "learning_rate": 5.132320858265939e-06, + "loss": 1.59878254, + "memory(GiB)": 111.15, + "step": 40760, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.66319175, + "epoch": 1.0341197361745307, + "grad_norm": 6.75, + "learning_rate": 5.131272598951554e-06, + "loss": 1.61093979, + "memory(GiB)": 111.15, + "step": 40765, + "train_speed(iter/s)": 0.44783 + }, + { + "acc": 0.65984297, + "epoch": 1.0342465753424657, + "grad_norm": 4.5, + "learning_rate": 5.130224333863212e-06, + "loss": 1.52692137, + "memory(GiB)": 111.15, + "step": 40770, + "train_speed(iter/s)": 0.447858 + }, + { + "acc": 0.64868245, + "epoch": 1.0343734145104009, + "grad_norm": 5.5, + "learning_rate": 5.129176063047022e-06, + "loss": 1.63619518, + "memory(GiB)": 111.15, + "step": 40775, + "train_speed(iter/s)": 0.447887 + }, + { + "acc": 0.6407342, + "epoch": 1.0345002536783359, + "grad_norm": 5.78125, + "learning_rate": 5.128127786549094e-06, + "loss": 1.70756321, + "memory(GiB)": 111.15, + "step": 40780, + "train_speed(iter/s)": 0.447916 + }, + { + "acc": 0.64937325, + "epoch": 1.0346270928462709, + "grad_norm": 5.21875, + "learning_rate": 5.127079504415532e-06, + "loss": 1.63127022, + "memory(GiB)": 111.15, + "step": 40785, + "train_speed(iter/s)": 0.447945 + }, + { + "acc": 0.6689785, + "epoch": 1.034753932014206, + "grad_norm": 6.28125, + "learning_rate": 5.126031216692449e-06, + "loss": 1.52826786, + "memory(GiB)": 111.15, + "step": 40790, + "train_speed(iter/s)": 0.447974 + }, + { + "acc": 0.65140305, + "epoch": 1.034880771182141, + "grad_norm": 4.5625, + "learning_rate": 5.124982923425947e-06, + "loss": 1.60285378, + "memory(GiB)": 111.15, + "step": 40795, + "train_speed(iter/s)": 0.448003 + }, + { + "acc": 0.64701653, + "epoch": 1.035007610350076, + "grad_norm": 4.78125, + "learning_rate": 5.123934624662139e-06, + "loss": 1.63468208, + "memory(GiB)": 111.15, + "step": 40800, + "train_speed(iter/s)": 0.448031 + }, + { + "epoch": 1.035007610350076, + "eval_acc": 0.6432973203987004, + "eval_loss": 1.6079540252685547, + "eval_runtime": 114.3305, + "eval_samples_per_second": 55.716, + "eval_steps_per_second": 27.858, + "step": 40800 + }, + { + "acc": 0.63488569, + "epoch": 1.0351344495180113, + "grad_norm": 6.53125, + "learning_rate": 5.1228863204471335e-06, + "loss": 1.64909077, + "memory(GiB)": 111.15, + "step": 40805, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.65733476, + "epoch": 1.0352612886859462, + "grad_norm": 4.375, + "learning_rate": 5.121838010827039e-06, + "loss": 1.57177639, + "memory(GiB)": 111.15, + "step": 40810, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.64333277, + "epoch": 1.0353881278538812, + "grad_norm": 5.21875, + "learning_rate": 5.120789695847965e-06, + "loss": 1.65285835, + "memory(GiB)": 111.15, + "step": 40815, + "train_speed(iter/s)": 0.447533 + }, + { + "acc": 0.64745102, + "epoch": 1.0355149670218164, + "grad_norm": 5.6875, + "learning_rate": 5.119741375556021e-06, + "loss": 1.65078621, + "memory(GiB)": 111.15, + "step": 40820, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.67034769, + "epoch": 1.0356418061897514, + "grad_norm": 5.0625, + "learning_rate": 5.118693049997316e-06, + "loss": 1.54314833, + "memory(GiB)": 111.15, + "step": 40825, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.64423437, + "epoch": 1.0357686453576864, + "grad_norm": 5.875, + "learning_rate": 5.117644719217961e-06, + "loss": 1.64635735, + "memory(GiB)": 111.15, + "step": 40830, + "train_speed(iter/s)": 0.447619 + }, + { + "acc": 0.65076056, + "epoch": 1.0358954845256214, + "grad_norm": 5.59375, + "learning_rate": 5.116596383264066e-06, + "loss": 1.61809425, + "memory(GiB)": 111.15, + "step": 40835, + "train_speed(iter/s)": 0.447648 + }, + { + "acc": 0.6380867, + "epoch": 1.0360223236935566, + "grad_norm": 9.0625, + "learning_rate": 5.115548042181742e-06, + "loss": 1.67739067, + "memory(GiB)": 111.15, + "step": 40840, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.65940332, + "epoch": 1.0361491628614916, + "grad_norm": 4.4375, + "learning_rate": 5.114499696017098e-06, + "loss": 1.51522026, + "memory(GiB)": 111.15, + "step": 40845, + "train_speed(iter/s)": 0.447706 + }, + { + "acc": 0.64435668, + "epoch": 1.0362760020294266, + "grad_norm": 5.0, + "learning_rate": 5.1134513448162475e-06, + "loss": 1.6797226, + "memory(GiB)": 111.15, + "step": 40850, + "train_speed(iter/s)": 0.447735 + }, + { + "acc": 0.63998327, + "epoch": 1.0364028411973618, + "grad_norm": 5.125, + "learning_rate": 5.112402988625299e-06, + "loss": 1.64177666, + "memory(GiB)": 111.15, + "step": 40855, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.65968466, + "epoch": 1.0365296803652968, + "grad_norm": 5.75, + "learning_rate": 5.111354627490367e-06, + "loss": 1.57887211, + "memory(GiB)": 111.15, + "step": 40860, + "train_speed(iter/s)": 0.447793 + }, + { + "acc": 0.63749566, + "epoch": 1.0366565195332318, + "grad_norm": 5.21875, + "learning_rate": 5.110306261457559e-06, + "loss": 1.695998, + "memory(GiB)": 111.15, + "step": 40865, + "train_speed(iter/s)": 0.447821 + }, + { + "acc": 0.64302111, + "epoch": 1.036783358701167, + "grad_norm": 4.65625, + "learning_rate": 5.109257890572991e-06, + "loss": 1.70650558, + "memory(GiB)": 111.15, + "step": 40870, + "train_speed(iter/s)": 0.44785 + }, + { + "acc": 0.65584531, + "epoch": 1.036910197869102, + "grad_norm": 6.0, + "learning_rate": 5.108209514882772e-06, + "loss": 1.62428818, + "memory(GiB)": 111.15, + "step": 40875, + "train_speed(iter/s)": 0.447879 + }, + { + "acc": 0.63860765, + "epoch": 1.037037037037037, + "grad_norm": 5.96875, + "learning_rate": 5.107161134433017e-06, + "loss": 1.64309654, + "memory(GiB)": 111.15, + "step": 40880, + "train_speed(iter/s)": 0.447908 + }, + { + "acc": 0.6521987, + "epoch": 1.0371638762049722, + "grad_norm": 6.09375, + "learning_rate": 5.106112749269835e-06, + "loss": 1.58687725, + "memory(GiB)": 111.15, + "step": 40885, + "train_speed(iter/s)": 0.447937 + }, + { + "acc": 0.64667993, + "epoch": 1.0372907153729072, + "grad_norm": 5.65625, + "learning_rate": 5.105064359439341e-06, + "loss": 1.66483879, + "memory(GiB)": 111.15, + "step": 40890, + "train_speed(iter/s)": 0.447966 + }, + { + "acc": 0.66261978, + "epoch": 1.0374175545408422, + "grad_norm": 5.375, + "learning_rate": 5.1040159649876485e-06, + "loss": 1.54089003, + "memory(GiB)": 111.15, + "step": 40895, + "train_speed(iter/s)": 0.447995 + }, + { + "acc": 0.63992376, + "epoch": 1.0375443937087772, + "grad_norm": 5.875, + "learning_rate": 5.102967565960868e-06, + "loss": 1.67910347, + "memory(GiB)": 111.15, + "step": 40900, + "train_speed(iter/s)": 0.448024 + }, + { + "epoch": 1.0375443937087772, + "eval_acc": 0.6432864584486545, + "eval_loss": 1.6080029010772705, + "eval_runtime": 114.4644, + "eval_samples_per_second": 55.651, + "eval_steps_per_second": 27.825, + "step": 40900 + }, + { + "acc": 0.65248785, + "epoch": 1.0376712328767124, + "grad_norm": 5.21875, + "learning_rate": 5.101919162405116e-06, + "loss": 1.62851849, + "memory(GiB)": 111.15, + "step": 40905, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.65560546, + "epoch": 1.0377980720446474, + "grad_norm": 5.59375, + "learning_rate": 5.100870754366503e-06, + "loss": 1.59730148, + "memory(GiB)": 111.15, + "step": 40910, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.66916599, + "epoch": 1.0379249112125823, + "grad_norm": 5.75, + "learning_rate": 5.099822341891144e-06, + "loss": 1.53786001, + "memory(GiB)": 111.15, + "step": 40915, + "train_speed(iter/s)": 0.447526 + }, + { + "acc": 0.64343257, + "epoch": 1.0380517503805176, + "grad_norm": 5.5, + "learning_rate": 5.098773925025152e-06, + "loss": 1.6442028, + "memory(GiB)": 111.15, + "step": 40920, + "train_speed(iter/s)": 0.447555 + }, + { + "acc": 0.63281708, + "epoch": 1.0381785895484525, + "grad_norm": 5.59375, + "learning_rate": 5.097725503814643e-06, + "loss": 1.69906387, + "memory(GiB)": 111.15, + "step": 40925, + "train_speed(iter/s)": 0.447584 + }, + { + "acc": 0.6451139, + "epoch": 1.0383054287163875, + "grad_norm": 4.71875, + "learning_rate": 5.09667707830573e-06, + "loss": 1.58250751, + "memory(GiB)": 111.15, + "step": 40930, + "train_speed(iter/s)": 0.447613 + }, + { + "acc": 0.67246876, + "epoch": 1.0384322678843227, + "grad_norm": 5.6875, + "learning_rate": 5.095628648544526e-06, + "loss": 1.49694986, + "memory(GiB)": 111.15, + "step": 40935, + "train_speed(iter/s)": 0.447642 + }, + { + "acc": 0.65144882, + "epoch": 1.0385591070522577, + "grad_norm": 5.5, + "learning_rate": 5.0945802145771495e-06, + "loss": 1.56341877, + "memory(GiB)": 111.15, + "step": 40940, + "train_speed(iter/s)": 0.44767 + }, + { + "acc": 0.65438194, + "epoch": 1.0386859462201927, + "grad_norm": 5.3125, + "learning_rate": 5.093531776449711e-06, + "loss": 1.57775364, + "memory(GiB)": 111.15, + "step": 40945, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.6567029, + "epoch": 1.038812785388128, + "grad_norm": 5.65625, + "learning_rate": 5.092483334208327e-06, + "loss": 1.64072227, + "memory(GiB)": 111.15, + "step": 40950, + "train_speed(iter/s)": 0.447728 + }, + { + "acc": 0.63843875, + "epoch": 1.038939624556063, + "grad_norm": 5.3125, + "learning_rate": 5.091434887899114e-06, + "loss": 1.65412827, + "memory(GiB)": 111.15, + "step": 40955, + "train_speed(iter/s)": 0.447757 + }, + { + "acc": 0.65137196, + "epoch": 1.039066463723998, + "grad_norm": 5.875, + "learning_rate": 5.0903864375681866e-06, + "loss": 1.63410091, + "memory(GiB)": 111.15, + "step": 40960, + "train_speed(iter/s)": 0.447786 + }, + { + "acc": 0.65558853, + "epoch": 1.0391933028919331, + "grad_norm": 5.09375, + "learning_rate": 5.0893379832616594e-06, + "loss": 1.60920525, + "memory(GiB)": 111.15, + "step": 40965, + "train_speed(iter/s)": 0.447814 + }, + { + "acc": 0.65729518, + "epoch": 1.0393201420598681, + "grad_norm": 5.75, + "learning_rate": 5.08828952502565e-06, + "loss": 1.58674374, + "memory(GiB)": 111.15, + "step": 40970, + "train_speed(iter/s)": 0.447843 + }, + { + "acc": 0.65411425, + "epoch": 1.039446981227803, + "grad_norm": 4.8125, + "learning_rate": 5.087241062906272e-06, + "loss": 1.6174469, + "memory(GiB)": 111.15, + "step": 40975, + "train_speed(iter/s)": 0.447872 + }, + { + "acc": 0.63433638, + "epoch": 1.0395738203957383, + "grad_norm": 4.78125, + "learning_rate": 5.086192596949643e-06, + "loss": 1.65459747, + "memory(GiB)": 111.15, + "step": 40980, + "train_speed(iter/s)": 0.447901 + }, + { + "acc": 0.65891299, + "epoch": 1.0397006595636733, + "grad_norm": 5.21875, + "learning_rate": 5.085144127201879e-06, + "loss": 1.57019787, + "memory(GiB)": 111.15, + "step": 40985, + "train_speed(iter/s)": 0.44793 + }, + { + "acc": 0.6594985, + "epoch": 1.0398274987316083, + "grad_norm": 5.875, + "learning_rate": 5.084095653709096e-06, + "loss": 1.52025614, + "memory(GiB)": 111.15, + "step": 40990, + "train_speed(iter/s)": 0.447959 + }, + { + "acc": 0.66843004, + "epoch": 1.0399543378995433, + "grad_norm": 5.4375, + "learning_rate": 5.0830471765174096e-06, + "loss": 1.61086578, + "memory(GiB)": 111.15, + "step": 40995, + "train_speed(iter/s)": 0.447988 + }, + { + "acc": 0.64343576, + "epoch": 1.0400811770674785, + "grad_norm": 5.59375, + "learning_rate": 5.0819986956729395e-06, + "loss": 1.64727268, + "memory(GiB)": 111.15, + "step": 41000, + "train_speed(iter/s)": 0.448017 + }, + { + "epoch": 1.0400811770674785, + "eval_acc": 0.6433441103373596, + "eval_loss": 1.6080927848815918, + "eval_runtime": 113.876, + "eval_samples_per_second": 55.938, + "eval_steps_per_second": 27.969, + "step": 41000 + }, + { + "acc": 0.66668983, + "epoch": 1.0402080162354135, + "grad_norm": 6.75, + "learning_rate": 5.080950211221799e-06, + "loss": 1.56072845, + "memory(GiB)": 111.15, + "step": 41005, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.6455523, + "epoch": 1.0403348554033485, + "grad_norm": 5.625, + "learning_rate": 5.079901723210109e-06, + "loss": 1.62132072, + "memory(GiB)": 111.15, + "step": 41010, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.65115528, + "epoch": 1.0404616945712837, + "grad_norm": 5.5625, + "learning_rate": 5.078853231683981e-06, + "loss": 1.601297, + "memory(GiB)": 111.15, + "step": 41015, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.65062113, + "epoch": 1.0405885337392187, + "grad_norm": 5.28125, + "learning_rate": 5.077804736689539e-06, + "loss": 1.67061729, + "memory(GiB)": 111.15, + "step": 41020, + "train_speed(iter/s)": 0.447552 + }, + { + "acc": 0.65944533, + "epoch": 1.0407153729071537, + "grad_norm": 5.96875, + "learning_rate": 5.0767562382728955e-06, + "loss": 1.57506714, + "memory(GiB)": 111.15, + "step": 41025, + "train_speed(iter/s)": 0.447581 + }, + { + "acc": 0.63766208, + "epoch": 1.0408422120750889, + "grad_norm": 5.9375, + "learning_rate": 5.075707736480171e-06, + "loss": 1.70299683, + "memory(GiB)": 111.15, + "step": 41030, + "train_speed(iter/s)": 0.44761 + }, + { + "acc": 0.64936695, + "epoch": 1.0409690512430239, + "grad_norm": 5.46875, + "learning_rate": 5.074659231357482e-06, + "loss": 1.63858891, + "memory(GiB)": 111.15, + "step": 41035, + "train_speed(iter/s)": 0.447639 + }, + { + "acc": 0.65061126, + "epoch": 1.0410958904109588, + "grad_norm": 4.4375, + "learning_rate": 5.073610722950947e-06, + "loss": 1.63703995, + "memory(GiB)": 111.15, + "step": 41040, + "train_speed(iter/s)": 0.447667 + }, + { + "acc": 0.64435196, + "epoch": 1.041222729578894, + "grad_norm": 5.9375, + "learning_rate": 5.072562211306683e-06, + "loss": 1.60127449, + "memory(GiB)": 111.15, + "step": 41045, + "train_speed(iter/s)": 0.447696 + }, + { + "acc": 0.64154825, + "epoch": 1.041349568746829, + "grad_norm": 5.8125, + "learning_rate": 5.071513696470809e-06, + "loss": 1.67070961, + "memory(GiB)": 111.15, + "step": 41050, + "train_speed(iter/s)": 0.447725 + }, + { + "acc": 0.65889044, + "epoch": 1.041476407914764, + "grad_norm": 6.6875, + "learning_rate": 5.070465178489443e-06, + "loss": 1.65068092, + "memory(GiB)": 111.15, + "step": 41055, + "train_speed(iter/s)": 0.447755 + }, + { + "acc": 0.65784783, + "epoch": 1.041603247082699, + "grad_norm": 4.90625, + "learning_rate": 5.069416657408704e-06, + "loss": 1.58803673, + "memory(GiB)": 111.15, + "step": 41060, + "train_speed(iter/s)": 0.447784 + }, + { + "acc": 0.65429068, + "epoch": 1.0417300862506342, + "grad_norm": 4.53125, + "learning_rate": 5.0683681332747105e-06, + "loss": 1.67527542, + "memory(GiB)": 111.15, + "step": 41065, + "train_speed(iter/s)": 0.447813 + }, + { + "acc": 0.65455174, + "epoch": 1.0418569254185692, + "grad_norm": 4.625, + "learning_rate": 5.067319606133583e-06, + "loss": 1.65811615, + "memory(GiB)": 111.15, + "step": 41070, + "train_speed(iter/s)": 0.447842 + }, + { + "acc": 0.66336813, + "epoch": 1.0419837645865042, + "grad_norm": 6.09375, + "learning_rate": 5.066271076031436e-06, + "loss": 1.5706151, + "memory(GiB)": 111.15, + "step": 41075, + "train_speed(iter/s)": 0.447871 + }, + { + "acc": 0.6484437, + "epoch": 1.0421106037544394, + "grad_norm": 5.25, + "learning_rate": 5.065222543014394e-06, + "loss": 1.64866486, + "memory(GiB)": 111.15, + "step": 41080, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.64865565, + "epoch": 1.0422374429223744, + "grad_norm": 6.34375, + "learning_rate": 5.06417400712857e-06, + "loss": 1.64748135, + "memory(GiB)": 111.15, + "step": 41085, + "train_speed(iter/s)": 0.447929 + }, + { + "acc": 0.6580224, + "epoch": 1.0423642820903094, + "grad_norm": 5.3125, + "learning_rate": 5.0631254684200906e-06, + "loss": 1.56194439, + "memory(GiB)": 111.15, + "step": 41090, + "train_speed(iter/s)": 0.447958 + }, + { + "acc": 0.66102161, + "epoch": 1.0424911212582446, + "grad_norm": 5.4375, + "learning_rate": 5.062076926935068e-06, + "loss": 1.61562958, + "memory(GiB)": 111.15, + "step": 41095, + "train_speed(iter/s)": 0.447988 + }, + { + "acc": 0.65637355, + "epoch": 1.0426179604261796, + "grad_norm": 5.53125, + "learning_rate": 5.061028382719626e-06, + "loss": 1.56043463, + "memory(GiB)": 111.15, + "step": 41100, + "train_speed(iter/s)": 0.448017 + }, + { + "epoch": 1.0426179604261796, + "eval_acc": 0.6432789386370843, + "eval_loss": 1.6078935861587524, + "eval_runtime": 114.3941, + "eval_samples_per_second": 55.685, + "eval_steps_per_second": 27.842, + "step": 41100 + }, + { + "acc": 0.65185738, + "epoch": 1.0427447995941146, + "grad_norm": 5.4375, + "learning_rate": 5.0599798358198835e-06, + "loss": 1.58938799, + "memory(GiB)": 111.15, + "step": 41105, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.64666548, + "epoch": 1.0428716387620498, + "grad_norm": 5.34375, + "learning_rate": 5.0589312862819605e-06, + "loss": 1.61957588, + "memory(GiB)": 111.15, + "step": 41110, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.66259336, + "epoch": 1.0429984779299848, + "grad_norm": 5.15625, + "learning_rate": 5.057882734151977e-06, + "loss": 1.54000654, + "memory(GiB)": 111.15, + "step": 41115, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64513125, + "epoch": 1.0431253170979198, + "grad_norm": 5.3125, + "learning_rate": 5.05683417947605e-06, + "loss": 1.63933792, + "memory(GiB)": 111.15, + "step": 41120, + "train_speed(iter/s)": 0.447552 + }, + { + "acc": 0.64459567, + "epoch": 1.043252156265855, + "grad_norm": 5.8125, + "learning_rate": 5.055785622300303e-06, + "loss": 1.63526134, + "memory(GiB)": 111.15, + "step": 41125, + "train_speed(iter/s)": 0.447581 + }, + { + "acc": 0.65142193, + "epoch": 1.04337899543379, + "grad_norm": 4.59375, + "learning_rate": 5.054737062670857e-06, + "loss": 1.63840141, + "memory(GiB)": 111.15, + "step": 41130, + "train_speed(iter/s)": 0.447609 + }, + { + "acc": 0.64377842, + "epoch": 1.043505834601725, + "grad_norm": 5.21875, + "learning_rate": 5.053688500633828e-06, + "loss": 1.59433794, + "memory(GiB)": 111.15, + "step": 41135, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.65733337, + "epoch": 1.0436326737696602, + "grad_norm": 6.0, + "learning_rate": 5.052639936235341e-06, + "loss": 1.57682009, + "memory(GiB)": 111.15, + "step": 41140, + "train_speed(iter/s)": 0.447667 + }, + { + "acc": 0.65228024, + "epoch": 1.0437595129375952, + "grad_norm": 7.0625, + "learning_rate": 5.051591369521513e-06, + "loss": 1.62955685, + "memory(GiB)": 111.15, + "step": 41145, + "train_speed(iter/s)": 0.447696 + }, + { + "acc": 0.64480276, + "epoch": 1.0438863521055302, + "grad_norm": 5.75, + "learning_rate": 5.050542800538469e-06, + "loss": 1.60388947, + "memory(GiB)": 111.15, + "step": 41150, + "train_speed(iter/s)": 0.447725 + }, + { + "acc": 0.65521536, + "epoch": 1.0440131912734651, + "grad_norm": 5.125, + "learning_rate": 5.049494229332324e-06, + "loss": 1.66966076, + "memory(GiB)": 111.15, + "step": 41155, + "train_speed(iter/s)": 0.447754 + }, + { + "acc": 0.65751781, + "epoch": 1.0441400304414004, + "grad_norm": 5.0, + "learning_rate": 5.048445655949204e-06, + "loss": 1.61418495, + "memory(GiB)": 111.15, + "step": 41160, + "train_speed(iter/s)": 0.447782 + }, + { + "acc": 0.63896503, + "epoch": 1.0442668696093353, + "grad_norm": 5.1875, + "learning_rate": 5.047397080435225e-06, + "loss": 1.70956783, + "memory(GiB)": 111.15, + "step": 41165, + "train_speed(iter/s)": 0.447811 + }, + { + "acc": 0.65428047, + "epoch": 1.0443937087772703, + "grad_norm": 5.28125, + "learning_rate": 5.046348502836512e-06, + "loss": 1.68791084, + "memory(GiB)": 111.15, + "step": 41170, + "train_speed(iter/s)": 0.447841 + }, + { + "acc": 0.6444273, + "epoch": 1.0445205479452055, + "grad_norm": 4.90625, + "learning_rate": 5.045299923199186e-06, + "loss": 1.62266731, + "memory(GiB)": 111.15, + "step": 41175, + "train_speed(iter/s)": 0.447869 + }, + { + "acc": 0.65184245, + "epoch": 1.0446473871131405, + "grad_norm": 4.5, + "learning_rate": 5.044251341569366e-06, + "loss": 1.57923622, + "memory(GiB)": 111.15, + "step": 41180, + "train_speed(iter/s)": 0.447898 + }, + { + "acc": 0.63445482, + "epoch": 1.0447742262810755, + "grad_norm": 6.8125, + "learning_rate": 5.043202757993175e-06, + "loss": 1.72848911, + "memory(GiB)": 111.15, + "step": 41185, + "train_speed(iter/s)": 0.447927 + }, + { + "acc": 0.62730656, + "epoch": 1.0449010654490107, + "grad_norm": 7.78125, + "learning_rate": 5.042154172516734e-06, + "loss": 1.65335712, + "memory(GiB)": 111.15, + "step": 41190, + "train_speed(iter/s)": 0.447955 + }, + { + "acc": 0.6533968, + "epoch": 1.0450279046169457, + "grad_norm": 5.4375, + "learning_rate": 5.041105585186164e-06, + "loss": 1.61289768, + "memory(GiB)": 111.15, + "step": 41195, + "train_speed(iter/s)": 0.447984 + }, + { + "acc": 0.66363182, + "epoch": 1.0451547437848807, + "grad_norm": 4.6875, + "learning_rate": 5.040056996047587e-06, + "loss": 1.63038387, + "memory(GiB)": 111.15, + "step": 41200, + "train_speed(iter/s)": 0.448012 + }, + { + "epoch": 1.0451547437848807, + "eval_acc": 0.6433278174122908, + "eval_loss": 1.607922911643982, + "eval_runtime": 113.7005, + "eval_samples_per_second": 56.024, + "eval_steps_per_second": 28.012, + "step": 41200 + }, + { + "acc": 0.65747271, + "epoch": 1.045281582952816, + "grad_norm": 5.21875, + "learning_rate": 5.039008405147125e-06, + "loss": 1.55639734, + "memory(GiB)": 111.15, + "step": 41205, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.64198084, + "epoch": 1.045408422120751, + "grad_norm": 6.0625, + "learning_rate": 5.0379598125308984e-06, + "loss": 1.66575546, + "memory(GiB)": 111.15, + "step": 41210, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.63691568, + "epoch": 1.045535261288686, + "grad_norm": 5.5625, + "learning_rate": 5.036911218245029e-06, + "loss": 1.70568542, + "memory(GiB)": 111.15, + "step": 41215, + "train_speed(iter/s)": 0.447522 + }, + { + "acc": 0.65117254, + "epoch": 1.045662100456621, + "grad_norm": 4.75, + "learning_rate": 5.035862622335641e-06, + "loss": 1.6058918, + "memory(GiB)": 111.15, + "step": 41220, + "train_speed(iter/s)": 0.447551 + }, + { + "acc": 0.66283898, + "epoch": 1.045788939624556, + "grad_norm": 5.15625, + "learning_rate": 5.034814024848853e-06, + "loss": 1.59637547, + "memory(GiB)": 111.15, + "step": 41225, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.66006317, + "epoch": 1.045915778792491, + "grad_norm": 5.09375, + "learning_rate": 5.033765425830791e-06, + "loss": 1.59055967, + "memory(GiB)": 111.15, + "step": 41230, + "train_speed(iter/s)": 0.447608 + }, + { + "acc": 0.65510993, + "epoch": 1.046042617960426, + "grad_norm": 6.875, + "learning_rate": 5.032716825327573e-06, + "loss": 1.61936722, + "memory(GiB)": 111.15, + "step": 41235, + "train_speed(iter/s)": 0.447636 + }, + { + "acc": 0.65027905, + "epoch": 1.0461694571283613, + "grad_norm": 5.75, + "learning_rate": 5.031668223385323e-06, + "loss": 1.59913139, + "memory(GiB)": 111.15, + "step": 41240, + "train_speed(iter/s)": 0.447665 + }, + { + "acc": 0.64304953, + "epoch": 1.0462962962962963, + "grad_norm": 5.875, + "learning_rate": 5.030619620050163e-06, + "loss": 1.64939537, + "memory(GiB)": 111.15, + "step": 41245, + "train_speed(iter/s)": 0.447693 + }, + { + "acc": 0.67448773, + "epoch": 1.0464231354642313, + "grad_norm": 5.71875, + "learning_rate": 5.029571015368217e-06, + "loss": 1.50006161, + "memory(GiB)": 111.15, + "step": 41250, + "train_speed(iter/s)": 0.447722 + }, + { + "acc": 0.64383898, + "epoch": 1.0465499746321665, + "grad_norm": 5.15625, + "learning_rate": 5.028522409385605e-06, + "loss": 1.61359253, + "memory(GiB)": 111.15, + "step": 41255, + "train_speed(iter/s)": 0.447751 + }, + { + "acc": 0.65242805, + "epoch": 1.0466768138001015, + "grad_norm": 4.71875, + "learning_rate": 5.0274738021484495e-06, + "loss": 1.55669518, + "memory(GiB)": 111.15, + "step": 41260, + "train_speed(iter/s)": 0.447779 + }, + { + "acc": 0.63750105, + "epoch": 1.0468036529680365, + "grad_norm": 5.78125, + "learning_rate": 5.026425193702874e-06, + "loss": 1.65287399, + "memory(GiB)": 111.15, + "step": 41265, + "train_speed(iter/s)": 0.447808 + }, + { + "acc": 0.6729043, + "epoch": 1.0469304921359717, + "grad_norm": 5.09375, + "learning_rate": 5.025376584095001e-06, + "loss": 1.49361486, + "memory(GiB)": 111.15, + "step": 41270, + "train_speed(iter/s)": 0.447836 + }, + { + "acc": 0.66202035, + "epoch": 1.0470573313039067, + "grad_norm": 6.03125, + "learning_rate": 5.024327973370951e-06, + "loss": 1.5671957, + "memory(GiB)": 111.15, + "step": 41275, + "train_speed(iter/s)": 0.447865 + }, + { + "acc": 0.63682976, + "epoch": 1.0471841704718416, + "grad_norm": 5.96875, + "learning_rate": 5.02327936157685e-06, + "loss": 1.67426186, + "memory(GiB)": 111.15, + "step": 41280, + "train_speed(iter/s)": 0.447894 + }, + { + "acc": 0.6498147, + "epoch": 1.0473110096397769, + "grad_norm": 5.25, + "learning_rate": 5.022230748758816e-06, + "loss": 1.59235554, + "memory(GiB)": 111.15, + "step": 41285, + "train_speed(iter/s)": 0.447922 + }, + { + "acc": 0.66580114, + "epoch": 1.0474378488077118, + "grad_norm": 5.875, + "learning_rate": 5.021182134962978e-06, + "loss": 1.62084732, + "memory(GiB)": 111.15, + "step": 41290, + "train_speed(iter/s)": 0.447951 + }, + { + "acc": 0.65972061, + "epoch": 1.0475646879756468, + "grad_norm": 4.375, + "learning_rate": 5.020133520235453e-06, + "loss": 1.61086044, + "memory(GiB)": 111.15, + "step": 41295, + "train_speed(iter/s)": 0.447979 + }, + { + "acc": 0.64280195, + "epoch": 1.047691527143582, + "grad_norm": 5.0, + "learning_rate": 5.019084904622367e-06, + "loss": 1.66787224, + "memory(GiB)": 111.15, + "step": 41300, + "train_speed(iter/s)": 0.448008 + }, + { + "epoch": 1.047691527143582, + "eval_acc": 0.6433332483873138, + "eval_loss": 1.6079574823379517, + "eval_runtime": 113.607, + "eval_samples_per_second": 56.07, + "eval_steps_per_second": 28.035, + "step": 41300 + }, + { + "acc": 0.64389014, + "epoch": 1.047818366311517, + "grad_norm": 4.375, + "learning_rate": 5.01803628816984e-06, + "loss": 1.61635475, + "memory(GiB)": 111.15, + "step": 41305, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.64811792, + "epoch": 1.047945205479452, + "grad_norm": 5.96875, + "learning_rate": 5.016987670923998e-06, + "loss": 1.58635883, + "memory(GiB)": 111.15, + "step": 41310, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.66456432, + "epoch": 1.048072044647387, + "grad_norm": 5.65625, + "learning_rate": 5.0159390529309615e-06, + "loss": 1.51359539, + "memory(GiB)": 111.15, + "step": 41315, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.6507175, + "epoch": 1.0481988838153222, + "grad_norm": 5.5, + "learning_rate": 5.014890434236854e-06, + "loss": 1.589674, + "memory(GiB)": 111.15, + "step": 41320, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.63273487, + "epoch": 1.0483257229832572, + "grad_norm": 5.84375, + "learning_rate": 5.0138418148878e-06, + "loss": 1.69515762, + "memory(GiB)": 111.15, + "step": 41325, + "train_speed(iter/s)": 0.447576 + }, + { + "acc": 0.66087823, + "epoch": 1.0484525621511922, + "grad_norm": 6.09375, + "learning_rate": 5.01279319492992e-06, + "loss": 1.61725693, + "memory(GiB)": 111.15, + "step": 41330, + "train_speed(iter/s)": 0.447604 + }, + { + "acc": 0.63669882, + "epoch": 1.0485794013191274, + "grad_norm": 4.84375, + "learning_rate": 5.01174457440934e-06, + "loss": 1.61847954, + "memory(GiB)": 111.15, + "step": 41335, + "train_speed(iter/s)": 0.447632 + }, + { + "acc": 0.66644945, + "epoch": 1.0487062404870624, + "grad_norm": 4.875, + "learning_rate": 5.010695953372179e-06, + "loss": 1.59719458, + "memory(GiB)": 111.15, + "step": 41340, + "train_speed(iter/s)": 0.447661 + }, + { + "acc": 0.65537276, + "epoch": 1.0488330796549974, + "grad_norm": 5.6875, + "learning_rate": 5.009647331864563e-06, + "loss": 1.59391222, + "memory(GiB)": 111.15, + "step": 41345, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.65975494, + "epoch": 1.0489599188229326, + "grad_norm": 5.5625, + "learning_rate": 5.008598709932615e-06, + "loss": 1.62572289, + "memory(GiB)": 111.15, + "step": 41350, + "train_speed(iter/s)": 0.447718 + }, + { + "acc": 0.6454555, + "epoch": 1.0490867579908676, + "grad_norm": 5.4375, + "learning_rate": 5.007550087622456e-06, + "loss": 1.68592033, + "memory(GiB)": 111.15, + "step": 41355, + "train_speed(iter/s)": 0.447746 + }, + { + "acc": 0.66192613, + "epoch": 1.0492135971588026, + "grad_norm": 5.96875, + "learning_rate": 5.0065014649802124e-06, + "loss": 1.5834548, + "memory(GiB)": 111.15, + "step": 41360, + "train_speed(iter/s)": 0.447775 + }, + { + "acc": 0.64973631, + "epoch": 1.0493404363267378, + "grad_norm": 4.84375, + "learning_rate": 5.005452842052003e-06, + "loss": 1.65350857, + "memory(GiB)": 111.15, + "step": 41365, + "train_speed(iter/s)": 0.447803 + }, + { + "acc": 0.65237398, + "epoch": 1.0494672754946728, + "grad_norm": 5.34375, + "learning_rate": 5.004404218883955e-06, + "loss": 1.63232288, + "memory(GiB)": 111.15, + "step": 41370, + "train_speed(iter/s)": 0.447831 + }, + { + "acc": 0.66964989, + "epoch": 1.0495941146626078, + "grad_norm": 6.46875, + "learning_rate": 5.0033555955221875e-06, + "loss": 1.60175323, + "memory(GiB)": 111.15, + "step": 41375, + "train_speed(iter/s)": 0.44786 + }, + { + "acc": 0.6546792, + "epoch": 1.0497209538305428, + "grad_norm": 5.125, + "learning_rate": 5.002306972012829e-06, + "loss": 1.59816303, + "memory(GiB)": 111.15, + "step": 41380, + "train_speed(iter/s)": 0.447889 + }, + { + "acc": 0.65405612, + "epoch": 1.049847792998478, + "grad_norm": 4.625, + "learning_rate": 5.001258348401998e-06, + "loss": 1.60086021, + "memory(GiB)": 111.15, + "step": 41385, + "train_speed(iter/s)": 0.447917 + }, + { + "acc": 0.65473151, + "epoch": 1.049974632166413, + "grad_norm": 5.21875, + "learning_rate": 5.000209724735819e-06, + "loss": 1.63437271, + "memory(GiB)": 111.15, + "step": 41390, + "train_speed(iter/s)": 0.447946 + }, + { + "acc": 0.64984846, + "epoch": 1.050101471334348, + "grad_norm": 5.59375, + "learning_rate": 4.999161101060416e-06, + "loss": 1.66689453, + "memory(GiB)": 111.15, + "step": 41395, + "train_speed(iter/s)": 0.447974 + }, + { + "acc": 0.63587079, + "epoch": 1.0502283105022832, + "grad_norm": 5.78125, + "learning_rate": 4.99811247742191e-06, + "loss": 1.68938675, + "memory(GiB)": 111.15, + "step": 41400, + "train_speed(iter/s)": 0.448003 + }, + { + "epoch": 1.0502283105022832, + "eval_acc": 0.6433867226029243, + "eval_loss": 1.6079118251800537, + "eval_runtime": 113.3231, + "eval_samples_per_second": 56.211, + "eval_steps_per_second": 28.105, + "step": 41400 + }, + { + "acc": 0.64716449, + "epoch": 1.0503551496702181, + "grad_norm": 5.375, + "learning_rate": 4.9970638538664275e-06, + "loss": 1.63772545, + "memory(GiB)": 111.15, + "step": 41405, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.64550753, + "epoch": 1.0504819888381531, + "grad_norm": 4.375, + "learning_rate": 4.996015230440091e-06, + "loss": 1.5745451, + "memory(GiB)": 111.15, + "step": 41410, + "train_speed(iter/s)": 0.447488 + }, + { + "acc": 0.65790439, + "epoch": 1.0506088280060883, + "grad_norm": 5.53125, + "learning_rate": 4.99496660718902e-06, + "loss": 1.52142334, + "memory(GiB)": 111.15, + "step": 41415, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.65228167, + "epoch": 1.0507356671740233, + "grad_norm": 5.0, + "learning_rate": 4.99391798415934e-06, + "loss": 1.5975812, + "memory(GiB)": 111.15, + "step": 41420, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.65375223, + "epoch": 1.0508625063419583, + "grad_norm": 5.1875, + "learning_rate": 4.992869361397175e-06, + "loss": 1.61226254, + "memory(GiB)": 111.15, + "step": 41425, + "train_speed(iter/s)": 0.447573 + }, + { + "acc": 0.64917965, + "epoch": 1.0509893455098935, + "grad_norm": 4.96875, + "learning_rate": 4.991820738948649e-06, + "loss": 1.57846651, + "memory(GiB)": 111.15, + "step": 41430, + "train_speed(iter/s)": 0.447602 + }, + { + "acc": 0.64134197, + "epoch": 1.0511161846778285, + "grad_norm": 6.40625, + "learning_rate": 4.9907721168598805e-06, + "loss": 1.64574299, + "memory(GiB)": 111.15, + "step": 41435, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.66372609, + "epoch": 1.0512430238457635, + "grad_norm": 4.875, + "learning_rate": 4.989723495176997e-06, + "loss": 1.58076591, + "memory(GiB)": 111.15, + "step": 41440, + "train_speed(iter/s)": 0.447659 + }, + { + "acc": 0.66658134, + "epoch": 1.0513698630136987, + "grad_norm": 5.1875, + "learning_rate": 4.988674873946118e-06, + "loss": 1.54508991, + "memory(GiB)": 111.15, + "step": 41445, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.66063328, + "epoch": 1.0514967021816337, + "grad_norm": 6.1875, + "learning_rate": 4.987626253213373e-06, + "loss": 1.60391693, + "memory(GiB)": 111.15, + "step": 41450, + "train_speed(iter/s)": 0.447716 + }, + { + "acc": 0.6602973, + "epoch": 1.0516235413495687, + "grad_norm": 5.65625, + "learning_rate": 4.986577633024877e-06, + "loss": 1.63489685, + "memory(GiB)": 111.15, + "step": 41455, + "train_speed(iter/s)": 0.447744 + }, + { + "acc": 0.64594297, + "epoch": 1.051750380517504, + "grad_norm": 5.71875, + "learning_rate": 4.985529013426758e-06, + "loss": 1.64574318, + "memory(GiB)": 111.15, + "step": 41460, + "train_speed(iter/s)": 0.447773 + }, + { + "acc": 0.65249085, + "epoch": 1.051877219685439, + "grad_norm": 5.34375, + "learning_rate": 4.984480394465136e-06, + "loss": 1.55723352, + "memory(GiB)": 111.15, + "step": 41465, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.64868441, + "epoch": 1.052004058853374, + "grad_norm": 4.34375, + "learning_rate": 4.9834317761861385e-06, + "loss": 1.61537743, + "memory(GiB)": 111.15, + "step": 41470, + "train_speed(iter/s)": 0.44783 + }, + { + "acc": 0.64188175, + "epoch": 1.0521308980213089, + "grad_norm": 5.71875, + "learning_rate": 4.982383158635884e-06, + "loss": 1.6617466, + "memory(GiB)": 111.15, + "step": 41475, + "train_speed(iter/s)": 0.447858 + }, + { + "acc": 0.64264688, + "epoch": 1.052257737189244, + "grad_norm": 5.59375, + "learning_rate": 4.981334541860496e-06, + "loss": 1.64584541, + "memory(GiB)": 111.15, + "step": 41480, + "train_speed(iter/s)": 0.447887 + }, + { + "acc": 0.64641542, + "epoch": 1.052384576357179, + "grad_norm": 5.59375, + "learning_rate": 4.980285925906098e-06, + "loss": 1.6408493, + "memory(GiB)": 111.15, + "step": 41485, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.64465728, + "epoch": 1.052511415525114, + "grad_norm": 5.09375, + "learning_rate": 4.9792373108188155e-06, + "loss": 1.63844452, + "memory(GiB)": 111.15, + "step": 41490, + "train_speed(iter/s)": 0.447943 + }, + { + "acc": 0.65630188, + "epoch": 1.0526382546930493, + "grad_norm": 6.53125, + "learning_rate": 4.978188696644767e-06, + "loss": 1.57658367, + "memory(GiB)": 111.15, + "step": 41495, + "train_speed(iter/s)": 0.447972 + }, + { + "acc": 0.66106791, + "epoch": 1.0527650938609843, + "grad_norm": 4.8125, + "learning_rate": 4.977140083430075e-06, + "loss": 1.56866264, + "memory(GiB)": 111.15, + "step": 41500, + "train_speed(iter/s)": 0.448 + }, + { + "epoch": 1.0527650938609843, + "eval_acc": 0.6433771139548068, + "eval_loss": 1.6079896688461304, + "eval_runtime": 114.4588, + "eval_samples_per_second": 55.653, + "eval_steps_per_second": 27.827, + "step": 41500 + }, + { + "acc": 0.66230946, + "epoch": 1.0528919330289193, + "grad_norm": 4.3125, + "learning_rate": 4.976091471220867e-06, + "loss": 1.60003586, + "memory(GiB)": 111.15, + "step": 41505, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.64725595, + "epoch": 1.0530187721968545, + "grad_norm": 4.6875, + "learning_rate": 4.975042860063263e-06, + "loss": 1.6337513, + "memory(GiB)": 111.15, + "step": 41510, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.64228354, + "epoch": 1.0531456113647895, + "grad_norm": 5.375, + "learning_rate": 4.973994250003384e-06, + "loss": 1.6573494, + "memory(GiB)": 111.15, + "step": 41515, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.63467693, + "epoch": 1.0532724505327244, + "grad_norm": 5.1875, + "learning_rate": 4.972945641087355e-06, + "loss": 1.70228691, + "memory(GiB)": 111.15, + "step": 41520, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.65360041, + "epoch": 1.0533992897006597, + "grad_norm": 4.9375, + "learning_rate": 4.9718970333612955e-06, + "loss": 1.58373804, + "memory(GiB)": 111.15, + "step": 41525, + "train_speed(iter/s)": 0.447565 + }, + { + "acc": 0.64937391, + "epoch": 1.0535261288685946, + "grad_norm": 5.71875, + "learning_rate": 4.970848426871333e-06, + "loss": 1.63750954, + "memory(GiB)": 111.15, + "step": 41530, + "train_speed(iter/s)": 0.447594 + }, + { + "acc": 0.64278803, + "epoch": 1.0536529680365296, + "grad_norm": 5.3125, + "learning_rate": 4.9697998216635854e-06, + "loss": 1.66214066, + "memory(GiB)": 111.15, + "step": 41535, + "train_speed(iter/s)": 0.447622 + }, + { + "acc": 0.65317836, + "epoch": 1.0537798072044646, + "grad_norm": 4.78125, + "learning_rate": 4.9687512177841765e-06, + "loss": 1.57719746, + "memory(GiB)": 111.15, + "step": 41540, + "train_speed(iter/s)": 0.44765 + }, + { + "acc": 0.64128928, + "epoch": 1.0539066463723998, + "grad_norm": 5.0, + "learning_rate": 4.967702615279227e-06, + "loss": 1.64484367, + "memory(GiB)": 111.15, + "step": 41545, + "train_speed(iter/s)": 0.447678 + }, + { + "acc": 0.64409766, + "epoch": 1.0540334855403348, + "grad_norm": 5.375, + "learning_rate": 4.966654014194863e-06, + "loss": 1.67312775, + "memory(GiB)": 111.15, + "step": 41550, + "train_speed(iter/s)": 0.447706 + }, + { + "acc": 0.65694685, + "epoch": 1.0541603247082698, + "grad_norm": 4.65625, + "learning_rate": 4.965605414577204e-06, + "loss": 1.55383034, + "memory(GiB)": 111.15, + "step": 41555, + "train_speed(iter/s)": 0.447735 + }, + { + "acc": 0.66345983, + "epoch": 1.054287163876205, + "grad_norm": 5.59375, + "learning_rate": 4.964556816472371e-06, + "loss": 1.59527283, + "memory(GiB)": 111.15, + "step": 41560, + "train_speed(iter/s)": 0.447763 + }, + { + "acc": 0.65390444, + "epoch": 1.05441400304414, + "grad_norm": 4.5625, + "learning_rate": 4.9635082199264874e-06, + "loss": 1.61077957, + "memory(GiB)": 111.15, + "step": 41565, + "train_speed(iter/s)": 0.447791 + }, + { + "acc": 0.66283593, + "epoch": 1.054540842212075, + "grad_norm": 5.0625, + "learning_rate": 4.962459624985677e-06, + "loss": 1.60091324, + "memory(GiB)": 111.15, + "step": 41570, + "train_speed(iter/s)": 0.447819 + }, + { + "acc": 0.64899426, + "epoch": 1.0546676813800102, + "grad_norm": 5.25, + "learning_rate": 4.961411031696059e-06, + "loss": 1.64444885, + "memory(GiB)": 111.15, + "step": 41575, + "train_speed(iter/s)": 0.447847 + }, + { + "acc": 0.66245971, + "epoch": 1.0547945205479452, + "grad_norm": 5.875, + "learning_rate": 4.960362440103756e-06, + "loss": 1.61947212, + "memory(GiB)": 111.15, + "step": 41580, + "train_speed(iter/s)": 0.447875 + }, + { + "acc": 0.66432285, + "epoch": 1.0549213597158802, + "grad_norm": 6.78125, + "learning_rate": 4.95931385025489e-06, + "loss": 1.55193405, + "memory(GiB)": 111.15, + "step": 41585, + "train_speed(iter/s)": 0.447903 + }, + { + "acc": 0.65418472, + "epoch": 1.0550481988838154, + "grad_norm": 4.625, + "learning_rate": 4.958265262195584e-06, + "loss": 1.6047451, + "memory(GiB)": 111.15, + "step": 41590, + "train_speed(iter/s)": 0.447931 + }, + { + "acc": 0.67184582, + "epoch": 1.0551750380517504, + "grad_norm": 5.625, + "learning_rate": 4.957216675971955e-06, + "loss": 1.60239735, + "memory(GiB)": 111.15, + "step": 41595, + "train_speed(iter/s)": 0.44796 + }, + { + "acc": 0.65619879, + "epoch": 1.0553018772196854, + "grad_norm": 5.03125, + "learning_rate": 4.9561680916301295e-06, + "loss": 1.66729145, + "memory(GiB)": 111.15, + "step": 41600, + "train_speed(iter/s)": 0.447988 + }, + { + "epoch": 1.0553018772196854, + "eval_acc": 0.6433319950853854, + "eval_loss": 1.6081465482711792, + "eval_runtime": 113.3846, + "eval_samples_per_second": 56.18, + "eval_steps_per_second": 28.09, + "step": 41600 + }, + { + "acc": 0.65336027, + "epoch": 1.0554287163876206, + "grad_norm": 5.5625, + "learning_rate": 4.955119509216226e-06, + "loss": 1.60565681, + "memory(GiB)": 111.15, + "step": 41605, + "train_speed(iter/s)": 0.447448 + }, + { + "acc": 0.65778151, + "epoch": 1.0555555555555556, + "grad_norm": 4.6875, + "learning_rate": 4.9540709287763685e-06, + "loss": 1.63132935, + "memory(GiB)": 111.15, + "step": 41610, + "train_speed(iter/s)": 0.447477 + }, + { + "acc": 0.64583302, + "epoch": 1.0556823947234906, + "grad_norm": 6.53125, + "learning_rate": 4.953022350356676e-06, + "loss": 1.63953476, + "memory(GiB)": 111.15, + "step": 41615, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.65336466, + "epoch": 1.0558092338914258, + "grad_norm": 6.125, + "learning_rate": 4.951973774003269e-06, + "loss": 1.64433517, + "memory(GiB)": 111.15, + "step": 41620, + "train_speed(iter/s)": 0.447534 + }, + { + "acc": 0.65398035, + "epoch": 1.0559360730593608, + "grad_norm": 5.53125, + "learning_rate": 4.950925199762271e-06, + "loss": 1.69238739, + "memory(GiB)": 111.15, + "step": 41625, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.64929085, + "epoch": 1.0560629122272958, + "grad_norm": 5.875, + "learning_rate": 4.949876627679803e-06, + "loss": 1.65933266, + "memory(GiB)": 111.15, + "step": 41630, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.66287713, + "epoch": 1.0561897513952307, + "grad_norm": 5.5625, + "learning_rate": 4.948828057801983e-06, + "loss": 1.61840954, + "memory(GiB)": 111.15, + "step": 41635, + "train_speed(iter/s)": 0.44762 + }, + { + "acc": 0.63991489, + "epoch": 1.056316590563166, + "grad_norm": 5.4375, + "learning_rate": 4.947779490174933e-06, + "loss": 1.65793953, + "memory(GiB)": 111.15, + "step": 41640, + "train_speed(iter/s)": 0.447648 + }, + { + "acc": 0.64888453, + "epoch": 1.056443429731101, + "grad_norm": 5.9375, + "learning_rate": 4.946730924844775e-06, + "loss": 1.65340195, + "memory(GiB)": 111.15, + "step": 41645, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.64696922, + "epoch": 1.056570268899036, + "grad_norm": 5.21875, + "learning_rate": 4.945682361857631e-06, + "loss": 1.63181572, + "memory(GiB)": 111.15, + "step": 41650, + "train_speed(iter/s)": 0.447706 + }, + { + "acc": 0.65906734, + "epoch": 1.0566971080669711, + "grad_norm": 5.34375, + "learning_rate": 4.944633801259615e-06, + "loss": 1.60643387, + "memory(GiB)": 111.15, + "step": 41655, + "train_speed(iter/s)": 0.447735 + }, + { + "acc": 0.62939491, + "epoch": 1.0568239472349061, + "grad_norm": 6.09375, + "learning_rate": 4.943585243096854e-06, + "loss": 1.69056721, + "memory(GiB)": 111.15, + "step": 41660, + "train_speed(iter/s)": 0.447763 + }, + { + "acc": 0.66375179, + "epoch": 1.0569507864028411, + "grad_norm": 4.9375, + "learning_rate": 4.942536687415465e-06, + "loss": 1.55037909, + "memory(GiB)": 111.15, + "step": 41665, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.63826828, + "epoch": 1.0570776255707763, + "grad_norm": 6.03125, + "learning_rate": 4.941488134261571e-06, + "loss": 1.71210632, + "memory(GiB)": 111.15, + "step": 41670, + "train_speed(iter/s)": 0.447821 + }, + { + "acc": 0.66714935, + "epoch": 1.0572044647387113, + "grad_norm": 5.75, + "learning_rate": 4.940439583681288e-06, + "loss": 1.48552799, + "memory(GiB)": 111.15, + "step": 41675, + "train_speed(iter/s)": 0.447849 + }, + { + "acc": 0.65472527, + "epoch": 1.0573313039066463, + "grad_norm": 5.59375, + "learning_rate": 4.939391035720739e-06, + "loss": 1.59369564, + "memory(GiB)": 111.15, + "step": 41680, + "train_speed(iter/s)": 0.447878 + }, + { + "acc": 0.66626196, + "epoch": 1.0574581430745815, + "grad_norm": 5.125, + "learning_rate": 4.938342490426041e-06, + "loss": 1.54022617, + "memory(GiB)": 111.15, + "step": 41685, + "train_speed(iter/s)": 0.447907 + }, + { + "acc": 0.64749217, + "epoch": 1.0575849822425165, + "grad_norm": 5.96875, + "learning_rate": 4.937293947843318e-06, + "loss": 1.62703934, + "memory(GiB)": 111.15, + "step": 41690, + "train_speed(iter/s)": 0.447936 + }, + { + "acc": 0.66188059, + "epoch": 1.0577118214104515, + "grad_norm": 4.4375, + "learning_rate": 4.936245408018687e-06, + "loss": 1.58038101, + "memory(GiB)": 111.15, + "step": 41695, + "train_speed(iter/s)": 0.447964 + }, + { + "acc": 0.64972219, + "epoch": 1.0578386605783865, + "grad_norm": 5.46875, + "learning_rate": 4.935196870998265e-06, + "loss": 1.66318321, + "memory(GiB)": 111.15, + "step": 41700, + "train_speed(iter/s)": 0.447993 + }, + { + "epoch": 1.0578386605783865, + "eval_acc": 0.643321550902649, + "eval_loss": 1.6078789234161377, + "eval_runtime": 113.1564, + "eval_samples_per_second": 56.294, + "eval_steps_per_second": 28.147, + "step": 41700 + }, + { + "acc": 0.64919252, + "epoch": 1.0579654997463217, + "grad_norm": 7.4375, + "learning_rate": 4.934148336828176e-06, + "loss": 1.67132511, + "memory(GiB)": 111.15, + "step": 41705, + "train_speed(iter/s)": 0.447455 + }, + { + "acc": 0.63010468, + "epoch": 1.0580923389142567, + "grad_norm": 6.15625, + "learning_rate": 4.933099805554538e-06, + "loss": 1.74319992, + "memory(GiB)": 111.15, + "step": 41710, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64316578, + "epoch": 1.0582191780821917, + "grad_norm": 5.1875, + "learning_rate": 4.932051277223468e-06, + "loss": 1.6727066, + "memory(GiB)": 111.15, + "step": 41715, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.65144887, + "epoch": 1.058346017250127, + "grad_norm": 5.8125, + "learning_rate": 4.931002751881086e-06, + "loss": 1.62946434, + "memory(GiB)": 111.15, + "step": 41720, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.64521546, + "epoch": 1.0584728564180619, + "grad_norm": 6.5625, + "learning_rate": 4.929954229573512e-06, + "loss": 1.63819504, + "memory(GiB)": 111.15, + "step": 41725, + "train_speed(iter/s)": 0.447568 + }, + { + "acc": 0.64819469, + "epoch": 1.0585996955859969, + "grad_norm": 5.0625, + "learning_rate": 4.9289057103468635e-06, + "loss": 1.58806763, + "memory(GiB)": 111.15, + "step": 41730, + "train_speed(iter/s)": 0.447597 + }, + { + "acc": 0.65480847, + "epoch": 1.058726534753932, + "grad_norm": 4.8125, + "learning_rate": 4.927857194247258e-06, + "loss": 1.59844437, + "memory(GiB)": 111.15, + "step": 41735, + "train_speed(iter/s)": 0.447625 + }, + { + "acc": 0.63103924, + "epoch": 1.058853373921867, + "grad_norm": 4.90625, + "learning_rate": 4.926808681320816e-06, + "loss": 1.65533524, + "memory(GiB)": 111.15, + "step": 41740, + "train_speed(iter/s)": 0.447654 + }, + { + "acc": 0.63293095, + "epoch": 1.058980213089802, + "grad_norm": 6.78125, + "learning_rate": 4.925760171613654e-06, + "loss": 1.72081032, + "memory(GiB)": 111.15, + "step": 41745, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.63749876, + "epoch": 1.0591070522577373, + "grad_norm": 5.125, + "learning_rate": 4.9247116651718925e-06, + "loss": 1.67388039, + "memory(GiB)": 111.15, + "step": 41750, + "train_speed(iter/s)": 0.447711 + }, + { + "acc": 0.65185189, + "epoch": 1.0592338914256723, + "grad_norm": 6.3125, + "learning_rate": 4.9236631620416486e-06, + "loss": 1.69824562, + "memory(GiB)": 111.15, + "step": 41755, + "train_speed(iter/s)": 0.44774 + }, + { + "acc": 0.65148354, + "epoch": 1.0593607305936072, + "grad_norm": 4.78125, + "learning_rate": 4.922614662269038e-06, + "loss": 1.64696178, + "memory(GiB)": 111.15, + "step": 41760, + "train_speed(iter/s)": 0.447768 + }, + { + "acc": 0.66139345, + "epoch": 1.0594875697615425, + "grad_norm": 4.8125, + "learning_rate": 4.9215661659001805e-06, + "loss": 1.58040371, + "memory(GiB)": 111.15, + "step": 41765, + "train_speed(iter/s)": 0.447797 + }, + { + "acc": 0.66876173, + "epoch": 1.0596144089294774, + "grad_norm": 5.6875, + "learning_rate": 4.920517672981195e-06, + "loss": 1.54324284, + "memory(GiB)": 111.15, + "step": 41770, + "train_speed(iter/s)": 0.447826 + }, + { + "acc": 0.65028963, + "epoch": 1.0597412480974124, + "grad_norm": 5.34375, + "learning_rate": 4.919469183558195e-06, + "loss": 1.61663628, + "memory(GiB)": 111.15, + "step": 41775, + "train_speed(iter/s)": 0.447854 + }, + { + "acc": 0.65981278, + "epoch": 1.0598680872653476, + "grad_norm": 4.9375, + "learning_rate": 4.9184206976773e-06, + "loss": 1.57383986, + "memory(GiB)": 111.15, + "step": 41780, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.64819117, + "epoch": 1.0599949264332826, + "grad_norm": 5.75, + "learning_rate": 4.917372215384627e-06, + "loss": 1.69450836, + "memory(GiB)": 111.15, + "step": 41785, + "train_speed(iter/s)": 0.447911 + }, + { + "acc": 0.6503056, + "epoch": 1.0601217656012176, + "grad_norm": 5.75, + "learning_rate": 4.916323736726295e-06, + "loss": 1.60568657, + "memory(GiB)": 111.15, + "step": 41790, + "train_speed(iter/s)": 0.44794 + }, + { + "acc": 0.64542637, + "epoch": 1.0602486047691526, + "grad_norm": 5.125, + "learning_rate": 4.9152752617484156e-06, + "loss": 1.60724602, + "memory(GiB)": 111.15, + "step": 41795, + "train_speed(iter/s)": 0.447968 + }, + { + "acc": 0.64659729, + "epoch": 1.0603754439370878, + "grad_norm": 5.125, + "learning_rate": 4.91422679049711e-06, + "loss": 1.65084209, + "memory(GiB)": 111.15, + "step": 41800, + "train_speed(iter/s)": 0.447997 + }, + { + "epoch": 1.0603754439370878, + "eval_acc": 0.6433374260604083, + "eval_loss": 1.6078965663909912, + "eval_runtime": 113.0609, + "eval_samples_per_second": 56.341, + "eval_steps_per_second": 28.171, + "step": 41800 + }, + { + "acc": 0.66245699, + "epoch": 1.0605022831050228, + "grad_norm": 5.40625, + "learning_rate": 4.913178323018493e-06, + "loss": 1.57084637, + "memory(GiB)": 111.15, + "step": 41805, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.65230112, + "epoch": 1.0606291222729578, + "grad_norm": 6.5625, + "learning_rate": 4.912129859358682e-06, + "loss": 1.63472214, + "memory(GiB)": 111.15, + "step": 41810, + "train_speed(iter/s)": 0.447489 + }, + { + "acc": 0.66080217, + "epoch": 1.060755961440893, + "grad_norm": 5.78125, + "learning_rate": 4.9110813995637905e-06, + "loss": 1.56788025, + "memory(GiB)": 111.15, + "step": 41815, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.65907879, + "epoch": 1.060882800608828, + "grad_norm": 4.59375, + "learning_rate": 4.910032943679936e-06, + "loss": 1.59788551, + "memory(GiB)": 111.15, + "step": 41820, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.66098442, + "epoch": 1.061009639776763, + "grad_norm": 5.5, + "learning_rate": 4.908984491753234e-06, + "loss": 1.58864365, + "memory(GiB)": 111.15, + "step": 41825, + "train_speed(iter/s)": 0.447573 + }, + { + "acc": 0.64891291, + "epoch": 1.0611364789446982, + "grad_norm": 5.78125, + "learning_rate": 4.907936043829802e-06, + "loss": 1.61914139, + "memory(GiB)": 111.15, + "step": 41830, + "train_speed(iter/s)": 0.447602 + }, + { + "acc": 0.65188279, + "epoch": 1.0612633181126332, + "grad_norm": 4.75, + "learning_rate": 4.906887599955754e-06, + "loss": 1.63676891, + "memory(GiB)": 111.15, + "step": 41835, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.66076803, + "epoch": 1.0613901572805682, + "grad_norm": 5.625, + "learning_rate": 4.905839160177203e-06, + "loss": 1.56510029, + "memory(GiB)": 111.15, + "step": 41840, + "train_speed(iter/s)": 0.447658 + }, + { + "acc": 0.64615254, + "epoch": 1.0615169964485034, + "grad_norm": 5.90625, + "learning_rate": 4.904790724540267e-06, + "loss": 1.60787983, + "memory(GiB)": 111.15, + "step": 41845, + "train_speed(iter/s)": 0.447686 + }, + { + "acc": 0.65833616, + "epoch": 1.0616438356164384, + "grad_norm": 5.40625, + "learning_rate": 4.903742293091061e-06, + "loss": 1.59069233, + "memory(GiB)": 111.15, + "step": 41850, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.64784856, + "epoch": 1.0617706747843734, + "grad_norm": 5.0, + "learning_rate": 4.902693865875698e-06, + "loss": 1.65351543, + "memory(GiB)": 111.15, + "step": 41855, + "train_speed(iter/s)": 0.447742 + }, + { + "acc": 0.65048389, + "epoch": 1.0618975139523084, + "grad_norm": 6.125, + "learning_rate": 4.901645442940293e-06, + "loss": 1.62082329, + "memory(GiB)": 111.15, + "step": 41860, + "train_speed(iter/s)": 0.44777 + }, + { + "acc": 0.65620852, + "epoch": 1.0620243531202436, + "grad_norm": 5.125, + "learning_rate": 4.900597024330961e-06, + "loss": 1.60522633, + "memory(GiB)": 111.15, + "step": 41865, + "train_speed(iter/s)": 0.447798 + }, + { + "acc": 0.65147662, + "epoch": 1.0621511922881786, + "grad_norm": 5.84375, + "learning_rate": 4.899548610093816e-06, + "loss": 1.66997337, + "memory(GiB)": 111.15, + "step": 41870, + "train_speed(iter/s)": 0.447826 + }, + { + "acc": 0.66023006, + "epoch": 1.0622780314561135, + "grad_norm": 7.1875, + "learning_rate": 4.89850020027497e-06, + "loss": 1.59744205, + "memory(GiB)": 111.15, + "step": 41875, + "train_speed(iter/s)": 0.447854 + }, + { + "acc": 0.66146517, + "epoch": 1.0624048706240488, + "grad_norm": 5.21875, + "learning_rate": 4.89745179492054e-06, + "loss": 1.57576199, + "memory(GiB)": 111.15, + "step": 41880, + "train_speed(iter/s)": 0.447881 + }, + { + "acc": 0.65449839, + "epoch": 1.0625317097919837, + "grad_norm": 5.5625, + "learning_rate": 4.896403394076636e-06, + "loss": 1.6449482, + "memory(GiB)": 111.15, + "step": 41885, + "train_speed(iter/s)": 0.44791 + }, + { + "acc": 0.65123515, + "epoch": 1.0626585489599187, + "grad_norm": 6.59375, + "learning_rate": 4.895354997789377e-06, + "loss": 1.66186676, + "memory(GiB)": 111.15, + "step": 41890, + "train_speed(iter/s)": 0.447938 + }, + { + "acc": 0.65771194, + "epoch": 1.062785388127854, + "grad_norm": 5.09375, + "learning_rate": 4.894306606104869e-06, + "loss": 1.60809669, + "memory(GiB)": 111.15, + "step": 41895, + "train_speed(iter/s)": 0.447966 + }, + { + "acc": 0.66397886, + "epoch": 1.062912227295789, + "grad_norm": 4.71875, + "learning_rate": 4.893258219069229e-06, + "loss": 1.59437504, + "memory(GiB)": 111.15, + "step": 41900, + "train_speed(iter/s)": 0.447993 + }, + { + "epoch": 1.062912227295789, + "eval_acc": 0.6433177909968638, + "eval_loss": 1.6079145669937134, + "eval_runtime": 113.4914, + "eval_samples_per_second": 56.128, + "eval_steps_per_second": 28.064, + "step": 41900 + }, + { + "acc": 0.66107016, + "epoch": 1.063039066463724, + "grad_norm": 5.15625, + "learning_rate": 4.892209836728569e-06, + "loss": 1.57260399, + "memory(GiB)": 111.15, + "step": 41905, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.65187168, + "epoch": 1.0631659056316591, + "grad_norm": 5.03125, + "learning_rate": 4.891161459129003e-06, + "loss": 1.62901077, + "memory(GiB)": 111.15, + "step": 41910, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.66040144, + "epoch": 1.0632927447995941, + "grad_norm": 5.6875, + "learning_rate": 4.890113086316641e-06, + "loss": 1.65080624, + "memory(GiB)": 111.15, + "step": 41915, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.65543137, + "epoch": 1.0634195839675291, + "grad_norm": 4.84375, + "learning_rate": 4.889064718337595e-06, + "loss": 1.61014786, + "memory(GiB)": 111.15, + "step": 41920, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.65282369, + "epoch": 1.0635464231354643, + "grad_norm": 4.59375, + "learning_rate": 4.888016355237979e-06, + "loss": 1.6042841, + "memory(GiB)": 111.15, + "step": 41925, + "train_speed(iter/s)": 0.447568 + }, + { + "acc": 0.64654007, + "epoch": 1.0636732623033993, + "grad_norm": 9.8125, + "learning_rate": 4.886967997063905e-06, + "loss": 1.60976868, + "memory(GiB)": 111.15, + "step": 41930, + "train_speed(iter/s)": 0.447595 + }, + { + "acc": 0.65265007, + "epoch": 1.0638001014713343, + "grad_norm": 6.0, + "learning_rate": 4.885919643861482e-06, + "loss": 1.53563662, + "memory(GiB)": 111.15, + "step": 41935, + "train_speed(iter/s)": 0.447623 + }, + { + "acc": 0.6696969, + "epoch": 1.0639269406392695, + "grad_norm": 6.25, + "learning_rate": 4.884871295676821e-06, + "loss": 1.53086767, + "memory(GiB)": 111.15, + "step": 41940, + "train_speed(iter/s)": 0.447652 + }, + { + "acc": 0.65311594, + "epoch": 1.0640537798072045, + "grad_norm": 5.3125, + "learning_rate": 4.883822952556036e-06, + "loss": 1.60818233, + "memory(GiB)": 111.15, + "step": 41945, + "train_speed(iter/s)": 0.44768 + }, + { + "acc": 0.64462595, + "epoch": 1.0641806189751395, + "grad_norm": 5.75, + "learning_rate": 4.882774614545237e-06, + "loss": 1.60149612, + "memory(GiB)": 111.15, + "step": 41950, + "train_speed(iter/s)": 0.447708 + }, + { + "acc": 0.65336847, + "epoch": 1.0643074581430745, + "grad_norm": 5.6875, + "learning_rate": 4.881726281690531e-06, + "loss": 1.63061409, + "memory(GiB)": 111.15, + "step": 41955, + "train_speed(iter/s)": 0.447736 + }, + { + "acc": 0.64028168, + "epoch": 1.0644342973110097, + "grad_norm": 6.21875, + "learning_rate": 4.8806779540380335e-06, + "loss": 1.67095222, + "memory(GiB)": 111.15, + "step": 41960, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.66788568, + "epoch": 1.0645611364789447, + "grad_norm": 4.6875, + "learning_rate": 4.879629631633851e-06, + "loss": 1.61166553, + "memory(GiB)": 111.15, + "step": 41965, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.63985977, + "epoch": 1.0646879756468797, + "grad_norm": 6.375, + "learning_rate": 4.8785813145240965e-06, + "loss": 1.63767319, + "memory(GiB)": 111.15, + "step": 41970, + "train_speed(iter/s)": 0.44782 + }, + { + "acc": 0.65098057, + "epoch": 1.0648148148148149, + "grad_norm": 5.3125, + "learning_rate": 4.877533002754877e-06, + "loss": 1.60758495, + "memory(GiB)": 111.15, + "step": 41975, + "train_speed(iter/s)": 0.447849 + }, + { + "acc": 0.63141074, + "epoch": 1.0649416539827499, + "grad_norm": 4.8125, + "learning_rate": 4.8764846963723025e-06, + "loss": 1.69148827, + "memory(GiB)": 111.15, + "step": 41980, + "train_speed(iter/s)": 0.447877 + }, + { + "acc": 0.65166168, + "epoch": 1.0650684931506849, + "grad_norm": 6.09375, + "learning_rate": 4.875436395422481e-06, + "loss": 1.62675076, + "memory(GiB)": 111.15, + "step": 41985, + "train_speed(iter/s)": 0.447905 + }, + { + "acc": 0.64788809, + "epoch": 1.06519533231862, + "grad_norm": 5.15625, + "learning_rate": 4.874388099951527e-06, + "loss": 1.59000778, + "memory(GiB)": 111.15, + "step": 41990, + "train_speed(iter/s)": 0.447933 + }, + { + "acc": 0.67918553, + "epoch": 1.065322171486555, + "grad_norm": 4.8125, + "learning_rate": 4.873339810005543e-06, + "loss": 1.55519466, + "memory(GiB)": 111.15, + "step": 41995, + "train_speed(iter/s)": 0.447961 + }, + { + "acc": 0.63840542, + "epoch": 1.06544901065449, + "grad_norm": 5.3125, + "learning_rate": 4.872291525630638e-06, + "loss": 1.68950081, + "memory(GiB)": 111.15, + "step": 42000, + "train_speed(iter/s)": 0.447989 + }, + { + "epoch": 1.06544901065449, + "eval_acc": 0.6433340839219327, + "eval_loss": 1.6078428030014038, + "eval_runtime": 113.3725, + "eval_samples_per_second": 56.186, + "eval_steps_per_second": 28.093, + "step": 42000 + }, + { + "acc": 0.65729804, + "epoch": 1.0655758498224253, + "grad_norm": 5.46875, + "learning_rate": 4.871243246872923e-06, + "loss": 1.59312334, + "memory(GiB)": 111.15, + "step": 42005, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.65402613, + "epoch": 1.0657026889903602, + "grad_norm": 5.34375, + "learning_rate": 4.870194973778506e-06, + "loss": 1.6272253, + "memory(GiB)": 111.15, + "step": 42010, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.64179764, + "epoch": 1.0658295281582952, + "grad_norm": 4.875, + "learning_rate": 4.869146706393493e-06, + "loss": 1.65246601, + "memory(GiB)": 111.15, + "step": 42015, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.65462732, + "epoch": 1.0659563673262302, + "grad_norm": 5.9375, + "learning_rate": 4.868098444763991e-06, + "loss": 1.5815218, + "memory(GiB)": 111.15, + "step": 42020, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.64720526, + "epoch": 1.0660832064941654, + "grad_norm": 6.75, + "learning_rate": 4.86705018893611e-06, + "loss": 1.61470165, + "memory(GiB)": 111.15, + "step": 42025, + "train_speed(iter/s)": 0.447565 + }, + { + "acc": 0.65128374, + "epoch": 1.0662100456621004, + "grad_norm": 4.75, + "learning_rate": 4.866001938955955e-06, + "loss": 1.60142593, + "memory(GiB)": 111.15, + "step": 42030, + "train_speed(iter/s)": 0.447593 + }, + { + "acc": 0.65026035, + "epoch": 1.0663368848300354, + "grad_norm": 4.5, + "learning_rate": 4.864953694869632e-06, + "loss": 1.59540253, + "memory(GiB)": 111.15, + "step": 42035, + "train_speed(iter/s)": 0.447621 + }, + { + "acc": 0.65704904, + "epoch": 1.0664637239979706, + "grad_norm": 7.25, + "learning_rate": 4.863905456723249e-06, + "loss": 1.63357162, + "memory(GiB)": 111.15, + "step": 42040, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.63413544, + "epoch": 1.0665905631659056, + "grad_norm": 5.0625, + "learning_rate": 4.8628572245629105e-06, + "loss": 1.7015274, + "memory(GiB)": 111.15, + "step": 42045, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.64236126, + "epoch": 1.0667174023338406, + "grad_norm": 5.84375, + "learning_rate": 4.861808998434726e-06, + "loss": 1.61057396, + "memory(GiB)": 111.15, + "step": 42050, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.65882177, + "epoch": 1.0668442415017758, + "grad_norm": 6.0625, + "learning_rate": 4.860760778384797e-06, + "loss": 1.56174812, + "memory(GiB)": 111.15, + "step": 42055, + "train_speed(iter/s)": 0.447733 + }, + { + "acc": 0.64902668, + "epoch": 1.0669710806697108, + "grad_norm": 4.78125, + "learning_rate": 4.85971256445923e-06, + "loss": 1.58841963, + "memory(GiB)": 111.15, + "step": 42060, + "train_speed(iter/s)": 0.44776 + }, + { + "acc": 0.64837356, + "epoch": 1.0670979198376458, + "grad_norm": 6.375, + "learning_rate": 4.858664356704131e-06, + "loss": 1.63287716, + "memory(GiB)": 111.15, + "step": 42065, + "train_speed(iter/s)": 0.447788 + }, + { + "acc": 0.6820281, + "epoch": 1.067224759005581, + "grad_norm": 6.25, + "learning_rate": 4.857616155165606e-06, + "loss": 1.49113922, + "memory(GiB)": 111.15, + "step": 42070, + "train_speed(iter/s)": 0.447816 + }, + { + "acc": 0.62892075, + "epoch": 1.067351598173516, + "grad_norm": 4.96875, + "learning_rate": 4.856567959889758e-06, + "loss": 1.71526108, + "memory(GiB)": 111.15, + "step": 42075, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.6530241, + "epoch": 1.067478437341451, + "grad_norm": 5.0625, + "learning_rate": 4.855519770922691e-06, + "loss": 1.65960541, + "memory(GiB)": 111.15, + "step": 42080, + "train_speed(iter/s)": 0.447872 + }, + { + "acc": 0.64618864, + "epoch": 1.0676052765093862, + "grad_norm": 6.25, + "learning_rate": 4.8544715883105084e-06, + "loss": 1.61837482, + "memory(GiB)": 111.15, + "step": 42085, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.6499299, + "epoch": 1.0677321156773212, + "grad_norm": 5.5, + "learning_rate": 4.853423412099318e-06, + "loss": 1.63688354, + "memory(GiB)": 111.15, + "step": 42090, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.64437094, + "epoch": 1.0678589548452562, + "grad_norm": 4.90625, + "learning_rate": 4.852375242335217e-06, + "loss": 1.64836521, + "memory(GiB)": 111.15, + "step": 42095, + "train_speed(iter/s)": 0.447956 + }, + { + "acc": 0.64936285, + "epoch": 1.0679857940131914, + "grad_norm": 5.125, + "learning_rate": 4.851327079064314e-06, + "loss": 1.62126427, + "memory(GiB)": 111.15, + "step": 42100, + "train_speed(iter/s)": 0.447984 + }, + { + "epoch": 1.0679857940131914, + "eval_acc": 0.6432985737006288, + "eval_loss": 1.6079578399658203, + "eval_runtime": 112.8729, + "eval_samples_per_second": 56.435, + "eval_steps_per_second": 28.218, + "step": 42100 + }, + { + "acc": 0.64184604, + "epoch": 1.0681126331811264, + "grad_norm": 5.59375, + "learning_rate": 4.850278922332708e-06, + "loss": 1.71156197, + "memory(GiB)": 111.15, + "step": 42105, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.65807724, + "epoch": 1.0682394723490614, + "grad_norm": 6.09375, + "learning_rate": 4.849230772186508e-06, + "loss": 1.62731743, + "memory(GiB)": 111.15, + "step": 42110, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.65833473, + "epoch": 1.0683663115169963, + "grad_norm": 4.75, + "learning_rate": 4.848182628671806e-06, + "loss": 1.56651812, + "memory(GiB)": 111.15, + "step": 42115, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.66958933, + "epoch": 1.0684931506849316, + "grad_norm": 4.71875, + "learning_rate": 4.847134491834713e-06, + "loss": 1.54585619, + "memory(GiB)": 111.15, + "step": 42120, + "train_speed(iter/s)": 0.447534 + }, + { + "acc": 0.6515511, + "epoch": 1.0686199898528665, + "grad_norm": 4.90625, + "learning_rate": 4.846086361721326e-06, + "loss": 1.57193289, + "memory(GiB)": 111.15, + "step": 42125, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.6369154, + "epoch": 1.0687468290208015, + "grad_norm": 5.71875, + "learning_rate": 4.84503823837775e-06, + "loss": 1.63303871, + "memory(GiB)": 111.15, + "step": 42130, + "train_speed(iter/s)": 0.44759 + }, + { + "acc": 0.65595679, + "epoch": 1.0688736681887367, + "grad_norm": 4.75, + "learning_rate": 4.843990121850083e-06, + "loss": 1.59799109, + "memory(GiB)": 111.15, + "step": 42135, + "train_speed(iter/s)": 0.447618 + }, + { + "acc": 0.63519468, + "epoch": 1.0690005073566717, + "grad_norm": 5.625, + "learning_rate": 4.842942012184426e-06, + "loss": 1.63287354, + "memory(GiB)": 111.15, + "step": 42140, + "train_speed(iter/s)": 0.447646 + }, + { + "acc": 0.63613129, + "epoch": 1.0691273465246067, + "grad_norm": 4.65625, + "learning_rate": 4.841893909426881e-06, + "loss": 1.66413803, + "memory(GiB)": 111.15, + "step": 42145, + "train_speed(iter/s)": 0.447673 + }, + { + "acc": 0.64686556, + "epoch": 1.069254185692542, + "grad_norm": 6.46875, + "learning_rate": 4.84084581362355e-06, + "loss": 1.6707222, + "memory(GiB)": 111.15, + "step": 42150, + "train_speed(iter/s)": 0.447701 + }, + { + "acc": 0.65531702, + "epoch": 1.069381024860477, + "grad_norm": 5.125, + "learning_rate": 4.839797724820529e-06, + "loss": 1.63433762, + "memory(GiB)": 111.15, + "step": 42155, + "train_speed(iter/s)": 0.447729 + }, + { + "acc": 0.64087443, + "epoch": 1.069507864028412, + "grad_norm": 6.375, + "learning_rate": 4.838749643063918e-06, + "loss": 1.64347954, + "memory(GiB)": 111.15, + "step": 42160, + "train_speed(iter/s)": 0.447757 + }, + { + "acc": 0.64798174, + "epoch": 1.0696347031963471, + "grad_norm": 5.15625, + "learning_rate": 4.837701568399819e-06, + "loss": 1.65603561, + "memory(GiB)": 111.15, + "step": 42165, + "train_speed(iter/s)": 0.447784 + }, + { + "acc": 0.63539333, + "epoch": 1.0697615423642821, + "grad_norm": 4.5, + "learning_rate": 4.836653500874331e-06, + "loss": 1.66145554, + "memory(GiB)": 111.15, + "step": 42170, + "train_speed(iter/s)": 0.447812 + }, + { + "acc": 0.64523268, + "epoch": 1.069888381532217, + "grad_norm": 5.25, + "learning_rate": 4.835605440533549e-06, + "loss": 1.66880531, + "memory(GiB)": 111.15, + "step": 42175, + "train_speed(iter/s)": 0.44784 + }, + { + "acc": 0.66297474, + "epoch": 1.070015220700152, + "grad_norm": 5.78125, + "learning_rate": 4.834557387423575e-06, + "loss": 1.51575947, + "memory(GiB)": 111.15, + "step": 42180, + "train_speed(iter/s)": 0.447868 + }, + { + "acc": 0.65415649, + "epoch": 1.0701420598680873, + "grad_norm": 4.5625, + "learning_rate": 4.833509341590503e-06, + "loss": 1.59952297, + "memory(GiB)": 111.15, + "step": 42185, + "train_speed(iter/s)": 0.447896 + }, + { + "acc": 0.65597782, + "epoch": 1.0702688990360223, + "grad_norm": 5.5, + "learning_rate": 4.8324613030804374e-06, + "loss": 1.56670532, + "memory(GiB)": 111.15, + "step": 42190, + "train_speed(iter/s)": 0.447924 + }, + { + "acc": 0.64246426, + "epoch": 1.0703957382039573, + "grad_norm": 5.6875, + "learning_rate": 4.83141327193947e-06, + "loss": 1.66084213, + "memory(GiB)": 111.15, + "step": 42195, + "train_speed(iter/s)": 0.447952 + }, + { + "acc": 0.66203041, + "epoch": 1.0705225773718925, + "grad_norm": 5.4375, + "learning_rate": 4.8303652482137e-06, + "loss": 1.5848753, + "memory(GiB)": 111.15, + "step": 42200, + "train_speed(iter/s)": 0.447979 + }, + { + "epoch": 1.0705225773718925, + "eval_acc": 0.6433090178833653, + "eval_loss": 1.6080025434494019, + "eval_runtime": 114.7536, + "eval_samples_per_second": 55.51, + "eval_steps_per_second": 27.755, + "step": 42200 + }, + { + "acc": 0.63279877, + "epoch": 1.0706494165398275, + "grad_norm": 5.5625, + "learning_rate": 4.829317231949222e-06, + "loss": 1.6698967, + "memory(GiB)": 111.15, + "step": 42205, + "train_speed(iter/s)": 0.44744 + }, + { + "acc": 0.66080055, + "epoch": 1.0707762557077625, + "grad_norm": 5.6875, + "learning_rate": 4.828269223192137e-06, + "loss": 1.61632996, + "memory(GiB)": 111.15, + "step": 42210, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.65124178, + "epoch": 1.0709030948756977, + "grad_norm": 5.40625, + "learning_rate": 4.827221221988537e-06, + "loss": 1.60223389, + "memory(GiB)": 111.15, + "step": 42215, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.65630131, + "epoch": 1.0710299340436327, + "grad_norm": 4.90625, + "learning_rate": 4.826173228384518e-06, + "loss": 1.57442532, + "memory(GiB)": 111.15, + "step": 42220, + "train_speed(iter/s)": 0.447522 + }, + { + "acc": 0.65567865, + "epoch": 1.0711567732115677, + "grad_norm": 4.9375, + "learning_rate": 4.8251252424261775e-06, + "loss": 1.62923698, + "memory(GiB)": 111.15, + "step": 42225, + "train_speed(iter/s)": 0.44755 + }, + { + "acc": 0.67063842, + "epoch": 1.0712836123795029, + "grad_norm": 4.6875, + "learning_rate": 4.8240772641596105e-06, + "loss": 1.50680742, + "memory(GiB)": 111.15, + "step": 42230, + "train_speed(iter/s)": 0.447577 + }, + { + "acc": 0.65935564, + "epoch": 1.0714104515474379, + "grad_norm": 5.28125, + "learning_rate": 4.82302929363091e-06, + "loss": 1.63713493, + "memory(GiB)": 111.15, + "step": 42235, + "train_speed(iter/s)": 0.447605 + }, + { + "acc": 0.64822226, + "epoch": 1.0715372907153728, + "grad_norm": 5.3125, + "learning_rate": 4.8219813308861705e-06, + "loss": 1.64099064, + "memory(GiB)": 111.15, + "step": 42240, + "train_speed(iter/s)": 0.447633 + }, + { + "acc": 0.66384859, + "epoch": 1.071664129883308, + "grad_norm": 5.15625, + "learning_rate": 4.820933375971487e-06, + "loss": 1.56793594, + "memory(GiB)": 111.15, + "step": 42245, + "train_speed(iter/s)": 0.447661 + }, + { + "acc": 0.63692331, + "epoch": 1.071790969051243, + "grad_norm": 5.4375, + "learning_rate": 4.819885428932955e-06, + "loss": 1.64721413, + "memory(GiB)": 111.15, + "step": 42250, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.65747166, + "epoch": 1.071917808219178, + "grad_norm": 5.28125, + "learning_rate": 4.818837489816664e-06, + "loss": 1.57155237, + "memory(GiB)": 111.15, + "step": 42255, + "train_speed(iter/s)": 0.447716 + }, + { + "acc": 0.64331398, + "epoch": 1.0720446473871132, + "grad_norm": 4.65625, + "learning_rate": 4.81778955866871e-06, + "loss": 1.71358528, + "memory(GiB)": 111.15, + "step": 42260, + "train_speed(iter/s)": 0.447744 + }, + { + "acc": 0.65925589, + "epoch": 1.0721714865550482, + "grad_norm": 5.09375, + "learning_rate": 4.816741635535183e-06, + "loss": 1.58301716, + "memory(GiB)": 111.15, + "step": 42265, + "train_speed(iter/s)": 0.447772 + }, + { + "acc": 0.63877544, + "epoch": 1.0722983257229832, + "grad_norm": 5.09375, + "learning_rate": 4.81569372046218e-06, + "loss": 1.68303795, + "memory(GiB)": 111.15, + "step": 42270, + "train_speed(iter/s)": 0.4478 + }, + { + "acc": 0.63884192, + "epoch": 1.0724251648909182, + "grad_norm": 6.0625, + "learning_rate": 4.814645813495788e-06, + "loss": 1.68935814, + "memory(GiB)": 111.15, + "step": 42275, + "train_speed(iter/s)": 0.447828 + }, + { + "acc": 0.65086803, + "epoch": 1.0725520040588534, + "grad_norm": 4.5, + "learning_rate": 4.8135979146821e-06, + "loss": 1.6348484, + "memory(GiB)": 111.15, + "step": 42280, + "train_speed(iter/s)": 0.447855 + }, + { + "acc": 0.63860865, + "epoch": 1.0726788432267884, + "grad_norm": 5.53125, + "learning_rate": 4.81255002406721e-06, + "loss": 1.62826576, + "memory(GiB)": 111.15, + "step": 42285, + "train_speed(iter/s)": 0.447883 + }, + { + "acc": 0.66557293, + "epoch": 1.0728056823947234, + "grad_norm": 4.71875, + "learning_rate": 4.811502141697206e-06, + "loss": 1.53216391, + "memory(GiB)": 111.15, + "step": 42290, + "train_speed(iter/s)": 0.44791 + }, + { + "acc": 0.6535913, + "epoch": 1.0729325215626586, + "grad_norm": 4.53125, + "learning_rate": 4.81045426761818e-06, + "loss": 1.60770531, + "memory(GiB)": 111.15, + "step": 42295, + "train_speed(iter/s)": 0.447938 + }, + { + "acc": 0.65668974, + "epoch": 1.0730593607305936, + "grad_norm": 4.71875, + "learning_rate": 4.80940640187622e-06, + "loss": 1.63087654, + "memory(GiB)": 111.15, + "step": 42300, + "train_speed(iter/s)": 0.447966 + }, + { + "epoch": 1.0730593607305936, + "eval_acc": 0.6433975845529701, + "eval_loss": 1.6080869436264038, + "eval_runtime": 114.6427, + "eval_samples_per_second": 55.564, + "eval_steps_per_second": 27.782, + "step": 42300 + }, + { + "acc": 0.65788221, + "epoch": 1.0731861998985286, + "grad_norm": 5.6875, + "learning_rate": 4.808358544517418e-06, + "loss": 1.54398746, + "memory(GiB)": 111.15, + "step": 42305, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.65854883, + "epoch": 1.0733130390664638, + "grad_norm": 5.59375, + "learning_rate": 4.807310695587865e-06, + "loss": 1.56340799, + "memory(GiB)": 111.15, + "step": 42310, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64213314, + "epoch": 1.0734398782343988, + "grad_norm": 5.15625, + "learning_rate": 4.8062628551336445e-06, + "loss": 1.69329185, + "memory(GiB)": 111.15, + "step": 42315, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.65046864, + "epoch": 1.0735667174023338, + "grad_norm": 5.25, + "learning_rate": 4.80521502320085e-06, + "loss": 1.65091934, + "memory(GiB)": 111.15, + "step": 42320, + "train_speed(iter/s)": 0.447511 + }, + { + "acc": 0.64869566, + "epoch": 1.073693556570269, + "grad_norm": 5.96875, + "learning_rate": 4.804167199835567e-06, + "loss": 1.60231094, + "memory(GiB)": 111.15, + "step": 42325, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.65662122, + "epoch": 1.073820395738204, + "grad_norm": 5.25, + "learning_rate": 4.8031193850838894e-06, + "loss": 1.56717491, + "memory(GiB)": 111.15, + "step": 42330, + "train_speed(iter/s)": 0.447567 + }, + { + "acc": 0.64752436, + "epoch": 1.073947234906139, + "grad_norm": 5.0, + "learning_rate": 4.802071578991896e-06, + "loss": 1.64022694, + "memory(GiB)": 111.15, + "step": 42335, + "train_speed(iter/s)": 0.447595 + }, + { + "acc": 0.64729605, + "epoch": 1.074074074074074, + "grad_norm": 4.8125, + "learning_rate": 4.801023781605679e-06, + "loss": 1.63269806, + "memory(GiB)": 111.15, + "step": 42340, + "train_speed(iter/s)": 0.447623 + }, + { + "acc": 0.65244746, + "epoch": 1.0742009132420092, + "grad_norm": 4.46875, + "learning_rate": 4.799975992971325e-06, + "loss": 1.62109928, + "memory(GiB)": 111.15, + "step": 42345, + "train_speed(iter/s)": 0.447651 + }, + { + "acc": 0.65786071, + "epoch": 1.0743277524099442, + "grad_norm": 5.125, + "learning_rate": 4.798928213134921e-06, + "loss": 1.57612333, + "memory(GiB)": 111.15, + "step": 42350, + "train_speed(iter/s)": 0.447679 + }, + { + "acc": 0.65906668, + "epoch": 1.0744545915778791, + "grad_norm": 5.6875, + "learning_rate": 4.797880442142551e-06, + "loss": 1.62653961, + "memory(GiB)": 111.15, + "step": 42355, + "train_speed(iter/s)": 0.447707 + }, + { + "acc": 0.65131454, + "epoch": 1.0745814307458144, + "grad_norm": 5.53125, + "learning_rate": 4.7968326800403e-06, + "loss": 1.58531933, + "memory(GiB)": 111.15, + "step": 42360, + "train_speed(iter/s)": 0.447735 + }, + { + "acc": 0.64075627, + "epoch": 1.0747082699137493, + "grad_norm": 6.40625, + "learning_rate": 4.795784926874255e-06, + "loss": 1.65971165, + "memory(GiB)": 111.15, + "step": 42365, + "train_speed(iter/s)": 0.447764 + }, + { + "acc": 0.6582942, + "epoch": 1.0748351090816843, + "grad_norm": 5.9375, + "learning_rate": 4.794737182690503e-06, + "loss": 1.61465836, + "memory(GiB)": 111.15, + "step": 42370, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.6313972, + "epoch": 1.0749619482496195, + "grad_norm": 4.65625, + "learning_rate": 4.793689447535126e-06, + "loss": 1.63567257, + "memory(GiB)": 111.15, + "step": 42375, + "train_speed(iter/s)": 0.44782 + }, + { + "acc": 0.66560392, + "epoch": 1.0750887874175545, + "grad_norm": 5.65625, + "learning_rate": 4.792641721454206e-06, + "loss": 1.59289656, + "memory(GiB)": 111.15, + "step": 42380, + "train_speed(iter/s)": 0.447848 + }, + { + "acc": 0.65964684, + "epoch": 1.0752156265854895, + "grad_norm": 5.375, + "learning_rate": 4.79159400449383e-06, + "loss": 1.57748537, + "memory(GiB)": 111.15, + "step": 42385, + "train_speed(iter/s)": 0.447876 + }, + { + "acc": 0.65489941, + "epoch": 1.0753424657534247, + "grad_norm": 4.84375, + "learning_rate": 4.7905462967000816e-06, + "loss": 1.60706482, + "memory(GiB)": 111.15, + "step": 42390, + "train_speed(iter/s)": 0.447904 + }, + { + "acc": 0.64812918, + "epoch": 1.0754693049213597, + "grad_norm": 5.84375, + "learning_rate": 4.789498598119039e-06, + "loss": 1.65929451, + "memory(GiB)": 111.15, + "step": 42395, + "train_speed(iter/s)": 0.447932 + }, + { + "acc": 0.64938974, + "epoch": 1.0755961440892947, + "grad_norm": 4.90625, + "learning_rate": 4.78845090879679e-06, + "loss": 1.60679092, + "memory(GiB)": 111.15, + "step": 42400, + "train_speed(iter/s)": 0.447961 + }, + { + "epoch": 1.0755961440892947, + "eval_acc": 0.6433750251182595, + "eval_loss": 1.6079102754592896, + "eval_runtime": 114.4677, + "eval_samples_per_second": 55.649, + "eval_steps_per_second": 27.824, + "step": 42400 + }, + { + "acc": 0.64990029, + "epoch": 1.07572298325723, + "grad_norm": 6.125, + "learning_rate": 4.787403228779413e-06, + "loss": 1.61475754, + "memory(GiB)": 111.15, + "step": 42405, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.64791603, + "epoch": 1.075849822425165, + "grad_norm": 6.75, + "learning_rate": 4.786355558112994e-06, + "loss": 1.68084602, + "memory(GiB)": 111.15, + "step": 42410, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.66396799, + "epoch": 1.0759766615931, + "grad_norm": 4.8125, + "learning_rate": 4.78530789684361e-06, + "loss": 1.57055225, + "memory(GiB)": 111.15, + "step": 42415, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.63911309, + "epoch": 1.0761035007610351, + "grad_norm": 4.125, + "learning_rate": 4.784260245017343e-06, + "loss": 1.67114429, + "memory(GiB)": 111.15, + "step": 42420, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.6534699, + "epoch": 1.07623033992897, + "grad_norm": 5.3125, + "learning_rate": 4.7832126026802725e-06, + "loss": 1.63753948, + "memory(GiB)": 111.15, + "step": 42425, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.65466781, + "epoch": 1.076357179096905, + "grad_norm": 5.4375, + "learning_rate": 4.782164969878482e-06, + "loss": 1.57898817, + "memory(GiB)": 111.15, + "step": 42430, + "train_speed(iter/s)": 0.447565 + }, + { + "acc": 0.62913675, + "epoch": 1.07648401826484, + "grad_norm": 4.75, + "learning_rate": 4.781117346658047e-06, + "loss": 1.6569315, + "memory(GiB)": 111.15, + "step": 42435, + "train_speed(iter/s)": 0.447593 + }, + { + "acc": 0.64169998, + "epoch": 1.0766108574327753, + "grad_norm": 5.0625, + "learning_rate": 4.780069733065048e-06, + "loss": 1.62741814, + "memory(GiB)": 111.15, + "step": 42440, + "train_speed(iter/s)": 0.447621 + }, + { + "acc": 0.65704737, + "epoch": 1.0767376966007103, + "grad_norm": 7.4375, + "learning_rate": 4.779022129145566e-06, + "loss": 1.57679758, + "memory(GiB)": 111.15, + "step": 42445, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.64747906, + "epoch": 1.0768645357686453, + "grad_norm": 4.8125, + "learning_rate": 4.777974534945677e-06, + "loss": 1.65724602, + "memory(GiB)": 111.15, + "step": 42450, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.64793148, + "epoch": 1.0769913749365805, + "grad_norm": 4.4375, + "learning_rate": 4.776926950511457e-06, + "loss": 1.68300209, + "memory(GiB)": 111.15, + "step": 42455, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.6358758, + "epoch": 1.0771182141045155, + "grad_norm": 4.71875, + "learning_rate": 4.775879375888986e-06, + "loss": 1.69937038, + "memory(GiB)": 111.15, + "step": 42460, + "train_speed(iter/s)": 0.447733 + }, + { + "acc": 0.66269622, + "epoch": 1.0772450532724505, + "grad_norm": 5.21875, + "learning_rate": 4.774831811124343e-06, + "loss": 1.58237514, + "memory(GiB)": 111.15, + "step": 42465, + "train_speed(iter/s)": 0.447761 + }, + { + "acc": 0.65211596, + "epoch": 1.0773718924403857, + "grad_norm": 5.03125, + "learning_rate": 4.773784256263601e-06, + "loss": 1.6161253, + "memory(GiB)": 111.15, + "step": 42470, + "train_speed(iter/s)": 0.447789 + }, + { + "acc": 0.64606943, + "epoch": 1.0774987316083207, + "grad_norm": 5.15625, + "learning_rate": 4.7727367113528374e-06, + "loss": 1.63733406, + "memory(GiB)": 111.15, + "step": 42475, + "train_speed(iter/s)": 0.447816 + }, + { + "acc": 0.64288712, + "epoch": 1.0776255707762556, + "grad_norm": 4.84375, + "learning_rate": 4.771689176438128e-06, + "loss": 1.65903435, + "memory(GiB)": 111.15, + "step": 42480, + "train_speed(iter/s)": 0.447844 + }, + { + "acc": 0.65304546, + "epoch": 1.0777524099441909, + "grad_norm": 7.09375, + "learning_rate": 4.770641651565546e-06, + "loss": 1.54605789, + "memory(GiB)": 111.15, + "step": 42485, + "train_speed(iter/s)": 0.447872 + }, + { + "acc": 0.6501338, + "epoch": 1.0778792491121258, + "grad_norm": 5.375, + "learning_rate": 4.769594136781172e-06, + "loss": 1.61108551, + "memory(GiB)": 111.15, + "step": 42490, + "train_speed(iter/s)": 0.4479 + }, + { + "acc": 0.66053123, + "epoch": 1.0780060882800608, + "grad_norm": 6.28125, + "learning_rate": 4.768546632131074e-06, + "loss": 1.62741489, + "memory(GiB)": 111.15, + "step": 42495, + "train_speed(iter/s)": 0.447928 + }, + { + "acc": 0.6540844, + "epoch": 1.0781329274479958, + "grad_norm": 6.6875, + "learning_rate": 4.767499137661328e-06, + "loss": 1.66894722, + "memory(GiB)": 111.15, + "step": 42500, + "train_speed(iter/s)": 0.447956 + }, + { + "epoch": 1.0781329274479958, + "eval_acc": 0.6433140310910787, + "eval_loss": 1.608110785484314, + "eval_runtime": 112.9595, + "eval_samples_per_second": 56.392, + "eval_steps_per_second": 28.196, + "step": 42500 + }, + { + "acc": 0.66638837, + "epoch": 1.078259766615931, + "grad_norm": 4.375, + "learning_rate": 4.76645165341801e-06, + "loss": 1.51792583, + "memory(GiB)": 111.15, + "step": 42505, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.63633523, + "epoch": 1.078386605783866, + "grad_norm": 5.09375, + "learning_rate": 4.76540417944719e-06, + "loss": 1.62671757, + "memory(GiB)": 111.15, + "step": 42510, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.66191425, + "epoch": 1.078513444951801, + "grad_norm": 4.53125, + "learning_rate": 4.764356715794942e-06, + "loss": 1.62717953, + "memory(GiB)": 111.15, + "step": 42515, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.65805836, + "epoch": 1.0786402841197362, + "grad_norm": 6.9375, + "learning_rate": 4.763309262507336e-06, + "loss": 1.57801075, + "memory(GiB)": 111.15, + "step": 42520, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.64669514, + "epoch": 1.0787671232876712, + "grad_norm": 5.8125, + "learning_rate": 4.762261819630447e-06, + "loss": 1.63321724, + "memory(GiB)": 111.15, + "step": 42525, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.66242962, + "epoch": 1.0788939624556062, + "grad_norm": 5.5, + "learning_rate": 4.761214387210345e-06, + "loss": 1.5374156, + "memory(GiB)": 111.15, + "step": 42530, + "train_speed(iter/s)": 0.447567 + }, + { + "acc": 0.66498322, + "epoch": 1.0790208016235414, + "grad_norm": 5.125, + "learning_rate": 4.760166965293099e-06, + "loss": 1.61086388, + "memory(GiB)": 111.15, + "step": 42535, + "train_speed(iter/s)": 0.447595 + }, + { + "acc": 0.65004244, + "epoch": 1.0791476407914764, + "grad_norm": 5.34375, + "learning_rate": 4.759119553924781e-06, + "loss": 1.62843895, + "memory(GiB)": 111.15, + "step": 42540, + "train_speed(iter/s)": 0.447622 + }, + { + "acc": 0.64620929, + "epoch": 1.0792744799594114, + "grad_norm": 5.46875, + "learning_rate": 4.758072153151461e-06, + "loss": 1.6055687, + "memory(GiB)": 111.15, + "step": 42545, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.64984074, + "epoch": 1.0794013191273466, + "grad_norm": 5.0625, + "learning_rate": 4.757024763019209e-06, + "loss": 1.55049896, + "memory(GiB)": 111.15, + "step": 42550, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.65108652, + "epoch": 1.0795281582952816, + "grad_norm": 4.78125, + "learning_rate": 4.755977383574091e-06, + "loss": 1.61213303, + "memory(GiB)": 111.15, + "step": 42555, + "train_speed(iter/s)": 0.447705 + }, + { + "acc": 0.65702071, + "epoch": 1.0796549974632166, + "grad_norm": 5.5, + "learning_rate": 4.754930014862177e-06, + "loss": 1.63071251, + "memory(GiB)": 111.15, + "step": 42560, + "train_speed(iter/s)": 0.447733 + }, + { + "acc": 0.6611783, + "epoch": 1.0797818366311518, + "grad_norm": 4.90625, + "learning_rate": 4.753882656929535e-06, + "loss": 1.65588474, + "memory(GiB)": 111.15, + "step": 42565, + "train_speed(iter/s)": 0.44776 + }, + { + "acc": 0.64996538, + "epoch": 1.0799086757990868, + "grad_norm": 5.09375, + "learning_rate": 4.752835309822234e-06, + "loss": 1.62630806, + "memory(GiB)": 111.15, + "step": 42570, + "train_speed(iter/s)": 0.447788 + }, + { + "acc": 0.65728741, + "epoch": 1.0800355149670218, + "grad_norm": 5.6875, + "learning_rate": 4.7517879735863385e-06, + "loss": 1.59335899, + "memory(GiB)": 111.15, + "step": 42575, + "train_speed(iter/s)": 0.447815 + }, + { + "acc": 0.63495207, + "epoch": 1.080162354134957, + "grad_norm": 4.5625, + "learning_rate": 4.750740648267916e-06, + "loss": 1.64548187, + "memory(GiB)": 111.15, + "step": 42580, + "train_speed(iter/s)": 0.447843 + }, + { + "acc": 0.648312, + "epoch": 1.080289193302892, + "grad_norm": 4.5, + "learning_rate": 4.749693333913033e-06, + "loss": 1.62279701, + "memory(GiB)": 111.15, + "step": 42585, + "train_speed(iter/s)": 0.447871 + }, + { + "acc": 0.64716854, + "epoch": 1.080416032470827, + "grad_norm": 5.3125, + "learning_rate": 4.748646030567755e-06, + "loss": 1.6170126, + "memory(GiB)": 111.15, + "step": 42590, + "train_speed(iter/s)": 0.447898 + }, + { + "acc": 0.65540886, + "epoch": 1.080542871638762, + "grad_norm": 5.75, + "learning_rate": 4.747598738278147e-06, + "loss": 1.65511131, + "memory(GiB)": 111.15, + "step": 42595, + "train_speed(iter/s)": 0.447926 + }, + { + "acc": 0.65587034, + "epoch": 1.0806697108066972, + "grad_norm": 4.5, + "learning_rate": 4.746551457090272e-06, + "loss": 1.6335247, + "memory(GiB)": 111.15, + "step": 42600, + "train_speed(iter/s)": 0.447953 + }, + { + "epoch": 1.0806697108066972, + "eval_acc": 0.643359985495119, + "eval_loss": 1.608039379119873, + "eval_runtime": 114.8983, + "eval_samples_per_second": 55.44, + "eval_steps_per_second": 27.72, + "step": 42600 + }, + { + "acc": 0.6475275, + "epoch": 1.0807965499746321, + "grad_norm": 5.25, + "learning_rate": 4.745504187050197e-06, + "loss": 1.66056252, + "memory(GiB)": 111.15, + "step": 42605, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.63938088, + "epoch": 1.0809233891425671, + "grad_norm": 5.0, + "learning_rate": 4.744456928203985e-06, + "loss": 1.64244041, + "memory(GiB)": 111.15, + "step": 42610, + "train_speed(iter/s)": 0.447446 + }, + { + "acc": 0.6555337, + "epoch": 1.0810502283105023, + "grad_norm": 6.53125, + "learning_rate": 4.743409680597695e-06, + "loss": 1.60825748, + "memory(GiB)": 111.15, + "step": 42615, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.63922367, + "epoch": 1.0811770674784373, + "grad_norm": 5.65625, + "learning_rate": 4.742362444277394e-06, + "loss": 1.66628761, + "memory(GiB)": 111.15, + "step": 42620, + "train_speed(iter/s)": 0.4475 + }, + { + "acc": 0.6513371, + "epoch": 1.0813039066463723, + "grad_norm": 5.03125, + "learning_rate": 4.741315219289142e-06, + "loss": 1.6523737, + "memory(GiB)": 111.15, + "step": 42625, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.65962896, + "epoch": 1.0814307458143075, + "grad_norm": 5.40625, + "learning_rate": 4.740268005679005e-06, + "loss": 1.57260571, + "memory(GiB)": 111.15, + "step": 42630, + "train_speed(iter/s)": 0.447555 + }, + { + "acc": 0.65712008, + "epoch": 1.0815575849822425, + "grad_norm": 6.75, + "learning_rate": 4.739220803493039e-06, + "loss": 1.56484509, + "memory(GiB)": 111.15, + "step": 42635, + "train_speed(iter/s)": 0.447582 + }, + { + "acc": 0.66033769, + "epoch": 1.0816844241501775, + "grad_norm": 6.40625, + "learning_rate": 4.738173612777306e-06, + "loss": 1.67986526, + "memory(GiB)": 111.15, + "step": 42640, + "train_speed(iter/s)": 0.44761 + }, + { + "acc": 0.63509431, + "epoch": 1.0818112633181127, + "grad_norm": 5.125, + "learning_rate": 4.737126433577866e-06, + "loss": 1.73638515, + "memory(GiB)": 111.15, + "step": 42645, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.6361464, + "epoch": 1.0819381024860477, + "grad_norm": 5.46875, + "learning_rate": 4.736079265940781e-06, + "loss": 1.64229202, + "memory(GiB)": 111.15, + "step": 42650, + "train_speed(iter/s)": 0.447666 + }, + { + "acc": 0.63813267, + "epoch": 1.0820649416539827, + "grad_norm": 5.8125, + "learning_rate": 4.735032109912107e-06, + "loss": 1.63184185, + "memory(GiB)": 111.15, + "step": 42655, + "train_speed(iter/s)": 0.447693 + }, + { + "acc": 0.64233227, + "epoch": 1.0821917808219177, + "grad_norm": 5.0625, + "learning_rate": 4.733984965537903e-06, + "loss": 1.61498661, + "memory(GiB)": 111.15, + "step": 42660, + "train_speed(iter/s)": 0.447721 + }, + { + "acc": 0.66911764, + "epoch": 1.082318619989853, + "grad_norm": 5.34375, + "learning_rate": 4.732937832864229e-06, + "loss": 1.53687544, + "memory(GiB)": 111.15, + "step": 42665, + "train_speed(iter/s)": 0.447748 + }, + { + "acc": 0.63870249, + "epoch": 1.082445459157788, + "grad_norm": 5.40625, + "learning_rate": 4.731890711937141e-06, + "loss": 1.66874638, + "memory(GiB)": 111.15, + "step": 42670, + "train_speed(iter/s)": 0.447776 + }, + { + "acc": 0.65349216, + "epoch": 1.0825722983257229, + "grad_norm": 6.09375, + "learning_rate": 4.730843602802696e-06, + "loss": 1.56692314, + "memory(GiB)": 111.15, + "step": 42675, + "train_speed(iter/s)": 0.447803 + }, + { + "acc": 0.65690975, + "epoch": 1.082699137493658, + "grad_norm": 6.53125, + "learning_rate": 4.729796505506951e-06, + "loss": 1.58195248, + "memory(GiB)": 111.15, + "step": 42680, + "train_speed(iter/s)": 0.447831 + }, + { + "acc": 0.63565168, + "epoch": 1.082825976661593, + "grad_norm": 4.78125, + "learning_rate": 4.728749420095964e-06, + "loss": 1.68743172, + "memory(GiB)": 111.15, + "step": 42685, + "train_speed(iter/s)": 0.447859 + }, + { + "acc": 0.65260105, + "epoch": 1.082952815829528, + "grad_norm": 5.46875, + "learning_rate": 4.727702346615788e-06, + "loss": 1.67320499, + "memory(GiB)": 111.15, + "step": 42690, + "train_speed(iter/s)": 0.447886 + }, + { + "acc": 0.6545435, + "epoch": 1.0830796549974633, + "grad_norm": 4.5, + "learning_rate": 4.726655285112477e-06, + "loss": 1.58356848, + "memory(GiB)": 111.15, + "step": 42695, + "train_speed(iter/s)": 0.447914 + }, + { + "acc": 0.64574442, + "epoch": 1.0832064941653983, + "grad_norm": 5.65625, + "learning_rate": 4.725608235632088e-06, + "loss": 1.64110832, + "memory(GiB)": 111.15, + "step": 42700, + "train_speed(iter/s)": 0.447942 + }, + { + "epoch": 1.0832064941653983, + "eval_acc": 0.6433048402102707, + "eval_loss": 1.6080348491668701, + "eval_runtime": 114.3889, + "eval_samples_per_second": 55.687, + "eval_steps_per_second": 27.844, + "step": 42700 + }, + { + "acc": 0.6555387, + "epoch": 1.0833333333333333, + "grad_norm": 5.96875, + "learning_rate": 4.724561198220672e-06, + "loss": 1.61674061, + "memory(GiB)": 111.15, + "step": 42705, + "train_speed(iter/s)": 0.44741 + }, + { + "acc": 0.64328675, + "epoch": 1.0834601725012685, + "grad_norm": 5.125, + "learning_rate": 4.723514172924287e-06, + "loss": 1.63138638, + "memory(GiB)": 111.15, + "step": 42710, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.64533577, + "epoch": 1.0835870116692035, + "grad_norm": 4.875, + "learning_rate": 4.7224671597889825e-06, + "loss": 1.61416206, + "memory(GiB)": 111.15, + "step": 42715, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.63563566, + "epoch": 1.0837138508371384, + "grad_norm": 6.15625, + "learning_rate": 4.72142015886081e-06, + "loss": 1.69171658, + "memory(GiB)": 111.15, + "step": 42720, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.66792364, + "epoch": 1.0838406900050737, + "grad_norm": 5.25, + "learning_rate": 4.720373170185823e-06, + "loss": 1.5581192, + "memory(GiB)": 111.15, + "step": 42725, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.65652103, + "epoch": 1.0839675291730086, + "grad_norm": 5.5625, + "learning_rate": 4.719326193810075e-06, + "loss": 1.646418, + "memory(GiB)": 111.15, + "step": 42730, + "train_speed(iter/s)": 0.447548 + }, + { + "acc": 0.64024019, + "epoch": 1.0840943683409436, + "grad_norm": 6.21875, + "learning_rate": 4.718279229779612e-06, + "loss": 1.64975052, + "memory(GiB)": 111.15, + "step": 42735, + "train_speed(iter/s)": 0.447576 + }, + { + "acc": 0.64308114, + "epoch": 1.0842212075088788, + "grad_norm": 5.28125, + "learning_rate": 4.717232278140485e-06, + "loss": 1.72063713, + "memory(GiB)": 111.15, + "step": 42740, + "train_speed(iter/s)": 0.447603 + }, + { + "acc": 0.6637732, + "epoch": 1.0843480466768138, + "grad_norm": 4.6875, + "learning_rate": 4.716185338938746e-06, + "loss": 1.57291613, + "memory(GiB)": 111.15, + "step": 42745, + "train_speed(iter/s)": 0.447631 + }, + { + "acc": 0.6394186, + "epoch": 1.0844748858447488, + "grad_norm": 6.53125, + "learning_rate": 4.7151384122204445e-06, + "loss": 1.60274467, + "memory(GiB)": 111.15, + "step": 42750, + "train_speed(iter/s)": 0.447659 + }, + { + "acc": 0.64304848, + "epoch": 1.0846017250126838, + "grad_norm": 5.6875, + "learning_rate": 4.7140914980316254e-06, + "loss": 1.60801468, + "memory(GiB)": 111.15, + "step": 42755, + "train_speed(iter/s)": 0.447686 + }, + { + "acc": 0.64666204, + "epoch": 1.084728564180619, + "grad_norm": 5.90625, + "learning_rate": 4.713044596418339e-06, + "loss": 1.6410614, + "memory(GiB)": 111.15, + "step": 42760, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.65263901, + "epoch": 1.084855403348554, + "grad_norm": 5.15625, + "learning_rate": 4.711997707426632e-06, + "loss": 1.61113319, + "memory(GiB)": 111.15, + "step": 42765, + "train_speed(iter/s)": 0.447741 + }, + { + "acc": 0.64947672, + "epoch": 1.084982242516489, + "grad_norm": 5.4375, + "learning_rate": 4.710950831102555e-06, + "loss": 1.65112762, + "memory(GiB)": 111.15, + "step": 42770, + "train_speed(iter/s)": 0.447769 + }, + { + "acc": 0.65196028, + "epoch": 1.0851090816844242, + "grad_norm": 5.9375, + "learning_rate": 4.709903967492147e-06, + "loss": 1.59210529, + "memory(GiB)": 111.15, + "step": 42775, + "train_speed(iter/s)": 0.447797 + }, + { + "acc": 0.64059129, + "epoch": 1.0852359208523592, + "grad_norm": 5.375, + "learning_rate": 4.7088571166414595e-06, + "loss": 1.60884666, + "memory(GiB)": 111.15, + "step": 42780, + "train_speed(iter/s)": 0.447825 + }, + { + "acc": 0.65471478, + "epoch": 1.0853627600202942, + "grad_norm": 5.09375, + "learning_rate": 4.707810278596534e-06, + "loss": 1.63755646, + "memory(GiB)": 111.15, + "step": 42785, + "train_speed(iter/s)": 0.447853 + }, + { + "acc": 0.65636454, + "epoch": 1.0854895991882294, + "grad_norm": 6.65625, + "learning_rate": 4.7067634534034205e-06, + "loss": 1.60719414, + "memory(GiB)": 111.15, + "step": 42790, + "train_speed(iter/s)": 0.44788 + }, + { + "acc": 0.64232798, + "epoch": 1.0856164383561644, + "grad_norm": 5.875, + "learning_rate": 4.705716641108157e-06, + "loss": 1.6291769, + "memory(GiB)": 111.15, + "step": 42795, + "train_speed(iter/s)": 0.447908 + }, + { + "acc": 0.65043526, + "epoch": 1.0857432775240994, + "grad_norm": 5.9375, + "learning_rate": 4.7046698417567894e-06, + "loss": 1.65495186, + "memory(GiB)": 111.15, + "step": 42800, + "train_speed(iter/s)": 0.447936 + }, + { + "epoch": 1.0857432775240994, + "eval_acc": 0.6433257285757435, + "eval_loss": 1.6080865859985352, + "eval_runtime": 114.082, + "eval_samples_per_second": 55.837, + "eval_steps_per_second": 27.919, + "step": 42800 + }, + { + "acc": 0.65143194, + "epoch": 1.0858701166920346, + "grad_norm": 5.8125, + "learning_rate": 4.7036230553953616e-06, + "loss": 1.62103233, + "memory(GiB)": 111.15, + "step": 42805, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.66299734, + "epoch": 1.0859969558599696, + "grad_norm": 6.0625, + "learning_rate": 4.702576282069916e-06, + "loss": 1.54169788, + "memory(GiB)": 111.15, + "step": 42810, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.64317827, + "epoch": 1.0861237950279046, + "grad_norm": 5.875, + "learning_rate": 4.701529521826492e-06, + "loss": 1.68515339, + "memory(GiB)": 111.15, + "step": 42815, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.66326375, + "epoch": 1.0862506341958396, + "grad_norm": 5.125, + "learning_rate": 4.700482774711131e-06, + "loss": 1.62345314, + "memory(GiB)": 111.15, + "step": 42820, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.66727629, + "epoch": 1.0863774733637748, + "grad_norm": 5.65625, + "learning_rate": 4.699436040769877e-06, + "loss": 1.51841269, + "memory(GiB)": 111.15, + "step": 42825, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.6602231, + "epoch": 1.0865043125317098, + "grad_norm": 5.21875, + "learning_rate": 4.698389320048768e-06, + "loss": 1.59528408, + "memory(GiB)": 111.15, + "step": 42830, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.63144603, + "epoch": 1.0866311516996447, + "grad_norm": 5.3125, + "learning_rate": 4.697342612593841e-06, + "loss": 1.67240829, + "memory(GiB)": 111.15, + "step": 42835, + "train_speed(iter/s)": 0.447572 + }, + { + "acc": 0.64681754, + "epoch": 1.08675799086758, + "grad_norm": 4.71875, + "learning_rate": 4.696295918451139e-06, + "loss": 1.59435062, + "memory(GiB)": 111.15, + "step": 42840, + "train_speed(iter/s)": 0.4476 + }, + { + "acc": 0.65802159, + "epoch": 1.086884830035515, + "grad_norm": 4.9375, + "learning_rate": 4.695249237666697e-06, + "loss": 1.59759254, + "memory(GiB)": 111.15, + "step": 42845, + "train_speed(iter/s)": 0.447628 + }, + { + "acc": 0.66968393, + "epoch": 1.08701166920345, + "grad_norm": 5.78125, + "learning_rate": 4.694202570286556e-06, + "loss": 1.56533194, + "memory(GiB)": 111.15, + "step": 42850, + "train_speed(iter/s)": 0.447655 + }, + { + "acc": 0.64820061, + "epoch": 1.0871385083713851, + "grad_norm": 5.0625, + "learning_rate": 4.693155916356751e-06, + "loss": 1.64918976, + "memory(GiB)": 111.15, + "step": 42855, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.63204174, + "epoch": 1.0872653475393201, + "grad_norm": 5.53125, + "learning_rate": 4.692109275923318e-06, + "loss": 1.72071457, + "memory(GiB)": 111.15, + "step": 42860, + "train_speed(iter/s)": 0.44771 + }, + { + "acc": 0.65164766, + "epoch": 1.0873921867072551, + "grad_norm": 6.09375, + "learning_rate": 4.6910626490322925e-06, + "loss": 1.67460384, + "memory(GiB)": 111.15, + "step": 42865, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.64980364, + "epoch": 1.0875190258751903, + "grad_norm": 6.375, + "learning_rate": 4.690016035729714e-06, + "loss": 1.62289696, + "memory(GiB)": 111.15, + "step": 42870, + "train_speed(iter/s)": 0.447766 + }, + { + "acc": 0.64433002, + "epoch": 1.0876458650431253, + "grad_norm": 6.25, + "learning_rate": 4.688969436061612e-06, + "loss": 1.6158556, + "memory(GiB)": 111.15, + "step": 42875, + "train_speed(iter/s)": 0.447793 + }, + { + "acc": 0.65760841, + "epoch": 1.0877727042110603, + "grad_norm": 5.9375, + "learning_rate": 4.687922850074022e-06, + "loss": 1.59640808, + "memory(GiB)": 111.15, + "step": 42880, + "train_speed(iter/s)": 0.447821 + }, + { + "acc": 0.65554414, + "epoch": 1.0878995433789955, + "grad_norm": 5.09375, + "learning_rate": 4.686876277812981e-06, + "loss": 1.66610069, + "memory(GiB)": 111.15, + "step": 42885, + "train_speed(iter/s)": 0.447849 + }, + { + "acc": 0.63980007, + "epoch": 1.0880263825469305, + "grad_norm": 5.5, + "learning_rate": 4.685829719324519e-06, + "loss": 1.69300232, + "memory(GiB)": 111.15, + "step": 42890, + "train_speed(iter/s)": 0.447877 + }, + { + "acc": 0.63820162, + "epoch": 1.0881532217148655, + "grad_norm": 5.5625, + "learning_rate": 4.6847831746546664e-06, + "loss": 1.69423447, + "memory(GiB)": 111.15, + "step": 42895, + "train_speed(iter/s)": 0.447904 + }, + { + "acc": 0.65422244, + "epoch": 1.0882800608828007, + "grad_norm": 7.1875, + "learning_rate": 4.683736643849459e-06, + "loss": 1.64154873, + "memory(GiB)": 111.15, + "step": 42900, + "train_speed(iter/s)": 0.447932 + }, + { + "epoch": 1.0882800608828007, + "eval_acc": 0.6432868762159639, + "eval_loss": 1.6078615188598633, + "eval_runtime": 113.4697, + "eval_samples_per_second": 56.138, + "eval_steps_per_second": 28.069, + "step": 42900 + }, + { + "acc": 0.65394912, + "epoch": 1.0884069000507357, + "grad_norm": 5.65625, + "learning_rate": 4.6826901269549255e-06, + "loss": 1.58369274, + "memory(GiB)": 111.15, + "step": 42905, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.64701324, + "epoch": 1.0885337392186707, + "grad_norm": 6.15625, + "learning_rate": 4.681643624017097e-06, + "loss": 1.64469929, + "memory(GiB)": 111.15, + "step": 42910, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.66808262, + "epoch": 1.0886605783866057, + "grad_norm": 5.78125, + "learning_rate": 4.680597135082002e-06, + "loss": 1.5405324, + "memory(GiB)": 111.15, + "step": 42915, + "train_speed(iter/s)": 0.447463 + }, + { + "acc": 0.67221107, + "epoch": 1.088787417554541, + "grad_norm": 5.9375, + "learning_rate": 4.679550660195673e-06, + "loss": 1.58627586, + "memory(GiB)": 111.15, + "step": 42920, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.64403071, + "epoch": 1.0889142567224759, + "grad_norm": 6.1875, + "learning_rate": 4.6785041994041345e-06, + "loss": 1.64165344, + "memory(GiB)": 111.15, + "step": 42925, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.63942022, + "epoch": 1.0890410958904109, + "grad_norm": 5.59375, + "learning_rate": 4.6774577527534195e-06, + "loss": 1.68262138, + "memory(GiB)": 111.15, + "step": 42930, + "train_speed(iter/s)": 0.447546 + }, + { + "acc": 0.65656362, + "epoch": 1.089167935058346, + "grad_norm": 4.78125, + "learning_rate": 4.676411320289551e-06, + "loss": 1.65069294, + "memory(GiB)": 111.15, + "step": 42935, + "train_speed(iter/s)": 0.447573 + }, + { + "acc": 0.65362797, + "epoch": 1.089294774226281, + "grad_norm": 5.78125, + "learning_rate": 4.675364902058556e-06, + "loss": 1.62113438, + "memory(GiB)": 111.15, + "step": 42940, + "train_speed(iter/s)": 0.447601 + }, + { + "acc": 0.6762516, + "epoch": 1.089421613394216, + "grad_norm": 5.75, + "learning_rate": 4.674318498106464e-06, + "loss": 1.58532314, + "memory(GiB)": 111.15, + "step": 42945, + "train_speed(iter/s)": 0.447629 + }, + { + "acc": 0.66250973, + "epoch": 1.0895484525621513, + "grad_norm": 6.28125, + "learning_rate": 4.6732721084792985e-06, + "loss": 1.61214523, + "memory(GiB)": 111.15, + "step": 42950, + "train_speed(iter/s)": 0.447657 + }, + { + "acc": 0.63973207, + "epoch": 1.0896752917300863, + "grad_norm": 5.625, + "learning_rate": 4.672225733223084e-06, + "loss": 1.68354359, + "memory(GiB)": 111.15, + "step": 42955, + "train_speed(iter/s)": 0.447685 + }, + { + "acc": 0.66553125, + "epoch": 1.0898021308980212, + "grad_norm": 5.90625, + "learning_rate": 4.671179372383844e-06, + "loss": 1.59900837, + "memory(GiB)": 111.15, + "step": 42960, + "train_speed(iter/s)": 0.447712 + }, + { + "acc": 0.64338713, + "epoch": 1.0899289700659565, + "grad_norm": 5.15625, + "learning_rate": 4.670133026007604e-06, + "loss": 1.67145157, + "memory(GiB)": 111.15, + "step": 42965, + "train_speed(iter/s)": 0.44774 + }, + { + "acc": 0.65400829, + "epoch": 1.0900558092338914, + "grad_norm": 5.9375, + "learning_rate": 4.669086694140388e-06, + "loss": 1.68493576, + "memory(GiB)": 111.15, + "step": 42970, + "train_speed(iter/s)": 0.447767 + }, + { + "acc": 0.64420414, + "epoch": 1.0901826484018264, + "grad_norm": 4.5, + "learning_rate": 4.668040376828214e-06, + "loss": 1.58679953, + "memory(GiB)": 111.15, + "step": 42975, + "train_speed(iter/s)": 0.447795 + }, + { + "acc": 0.66479387, + "epoch": 1.0903094875697614, + "grad_norm": 5.5, + "learning_rate": 4.666994074117108e-06, + "loss": 1.61896133, + "memory(GiB)": 111.15, + "step": 42980, + "train_speed(iter/s)": 0.447822 + }, + { + "acc": 0.64758492, + "epoch": 1.0904363267376966, + "grad_norm": 5.25, + "learning_rate": 4.665947786053088e-06, + "loss": 1.67446289, + "memory(GiB)": 111.15, + "step": 42985, + "train_speed(iter/s)": 0.44785 + }, + { + "acc": 0.64877167, + "epoch": 1.0905631659056316, + "grad_norm": 5.3125, + "learning_rate": 4.664901512682179e-06, + "loss": 1.61705761, + "memory(GiB)": 111.15, + "step": 42990, + "train_speed(iter/s)": 0.447878 + }, + { + "acc": 0.65659275, + "epoch": 1.0906900050735666, + "grad_norm": 6.59375, + "learning_rate": 4.663855254050394e-06, + "loss": 1.5930562, + "memory(GiB)": 111.15, + "step": 42995, + "train_speed(iter/s)": 0.447906 + }, + { + "acc": 0.6341464, + "epoch": 1.0908168442415018, + "grad_norm": 5.5, + "learning_rate": 4.662809010203757e-06, + "loss": 1.6670826, + "memory(GiB)": 111.15, + "step": 43000, + "train_speed(iter/s)": 0.447934 + }, + { + "epoch": 1.0908168442415018, + "eval_acc": 0.6433073468141274, + "eval_loss": 1.6077696084976196, + "eval_runtime": 114.1001, + "eval_samples_per_second": 55.828, + "eval_steps_per_second": 27.914, + "step": 43000 + }, + { + "acc": 0.64584084, + "epoch": 1.0909436834094368, + "grad_norm": 5.1875, + "learning_rate": 4.661762781188284e-06, + "loss": 1.63387032, + "memory(GiB)": 111.15, + "step": 43005, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.64543142, + "epoch": 1.0910705225773718, + "grad_norm": 4.6875, + "learning_rate": 4.660716567049997e-06, + "loss": 1.62504539, + "memory(GiB)": 111.15, + "step": 43010, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.6500843, + "epoch": 1.091197361745307, + "grad_norm": 5.5625, + "learning_rate": 4.659670367834908e-06, + "loss": 1.62744827, + "memory(GiB)": 111.15, + "step": 43015, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.64741364, + "epoch": 1.091324200913242, + "grad_norm": 6.3125, + "learning_rate": 4.658624183589035e-06, + "loss": 1.64770241, + "memory(GiB)": 111.15, + "step": 43020, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.65534477, + "epoch": 1.091451040081177, + "grad_norm": 4.625, + "learning_rate": 4.657578014358395e-06, + "loss": 1.63118877, + "memory(GiB)": 111.15, + "step": 43025, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.64144487, + "epoch": 1.0915778792491122, + "grad_norm": 5.15625, + "learning_rate": 4.656531860189005e-06, + "loss": 1.63590393, + "memory(GiB)": 111.15, + "step": 43030, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.6399765, + "epoch": 1.0917047184170472, + "grad_norm": 5.15625, + "learning_rate": 4.655485721126875e-06, + "loss": 1.62361069, + "memory(GiB)": 111.15, + "step": 43035, + "train_speed(iter/s)": 0.447574 + }, + { + "acc": 0.63829618, + "epoch": 1.0918315575849822, + "grad_norm": 6.5, + "learning_rate": 4.6544395972180214e-06, + "loss": 1.65863419, + "memory(GiB)": 111.15, + "step": 43040, + "train_speed(iter/s)": 0.447602 + }, + { + "acc": 0.63826284, + "epoch": 1.0919583967529174, + "grad_norm": 5.15625, + "learning_rate": 4.653393488508457e-06, + "loss": 1.65007706, + "memory(GiB)": 111.15, + "step": 43045, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.64083128, + "epoch": 1.0920852359208524, + "grad_norm": 5.9375, + "learning_rate": 4.652347395044197e-06, + "loss": 1.60255299, + "memory(GiB)": 111.15, + "step": 43050, + "train_speed(iter/s)": 0.447657 + }, + { + "acc": 0.65764699, + "epoch": 1.0922120750887874, + "grad_norm": 4.65625, + "learning_rate": 4.651301316871247e-06, + "loss": 1.63141975, + "memory(GiB)": 111.15, + "step": 43055, + "train_speed(iter/s)": 0.447685 + }, + { + "acc": 0.64725056, + "epoch": 1.0923389142567226, + "grad_norm": 5.28125, + "learning_rate": 4.6502552540356235e-06, + "loss": 1.62078667, + "memory(GiB)": 111.15, + "step": 43060, + "train_speed(iter/s)": 0.447713 + }, + { + "acc": 0.65702257, + "epoch": 1.0924657534246576, + "grad_norm": 5.4375, + "learning_rate": 4.649209206583335e-06, + "loss": 1.66143494, + "memory(GiB)": 111.15, + "step": 43065, + "train_speed(iter/s)": 0.447741 + }, + { + "acc": 0.67002554, + "epoch": 1.0925925925925926, + "grad_norm": 4.96875, + "learning_rate": 4.648163174560393e-06, + "loss": 1.60738335, + "memory(GiB)": 111.15, + "step": 43070, + "train_speed(iter/s)": 0.447769 + }, + { + "acc": 0.65469995, + "epoch": 1.0927194317605275, + "grad_norm": 5.15625, + "learning_rate": 4.647117158012804e-06, + "loss": 1.62020607, + "memory(GiB)": 111.15, + "step": 43075, + "train_speed(iter/s)": 0.447796 + }, + { + "acc": 0.64555955, + "epoch": 1.0928462709284628, + "grad_norm": 5.25, + "learning_rate": 4.646071156986579e-06, + "loss": 1.58331337, + "memory(GiB)": 111.15, + "step": 43080, + "train_speed(iter/s)": 0.447824 + }, + { + "acc": 0.64111156, + "epoch": 1.0929731100963977, + "grad_norm": 5.78125, + "learning_rate": 4.645025171527723e-06, + "loss": 1.67377243, + "memory(GiB)": 111.15, + "step": 43085, + "train_speed(iter/s)": 0.447852 + }, + { + "acc": 0.64014711, + "epoch": 1.0930999492643327, + "grad_norm": 5.0, + "learning_rate": 4.643979201682247e-06, + "loss": 1.65312767, + "memory(GiB)": 111.15, + "step": 43090, + "train_speed(iter/s)": 0.44788 + }, + { + "acc": 0.64275312, + "epoch": 1.093226788432268, + "grad_norm": 5.6875, + "learning_rate": 4.642933247496155e-06, + "loss": 1.65204659, + "memory(GiB)": 111.15, + "step": 43095, + "train_speed(iter/s)": 0.447908 + }, + { + "acc": 0.64876466, + "epoch": 1.093353627600203, + "grad_norm": 6.34375, + "learning_rate": 4.641887309015451e-06, + "loss": 1.64270668, + "memory(GiB)": 111.15, + "step": 43100, + "train_speed(iter/s)": 0.447936 + }, + { + "epoch": 1.093353627600203, + "eval_acc": 0.6433900647413999, + "eval_loss": 1.6079121828079224, + "eval_runtime": 113.8484, + "eval_samples_per_second": 55.952, + "eval_steps_per_second": 27.976, + "step": 43100 + }, + { + "acc": 0.65540686, + "epoch": 1.093480466768138, + "grad_norm": 4.5625, + "learning_rate": 4.640841386286143e-06, + "loss": 1.5706913, + "memory(GiB)": 111.15, + "step": 43105, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.64994993, + "epoch": 1.0936073059360731, + "grad_norm": 5.5, + "learning_rate": 4.639795479354236e-06, + "loss": 1.69318714, + "memory(GiB)": 111.15, + "step": 43110, + "train_speed(iter/s)": 0.44744 + }, + { + "acc": 0.66305676, + "epoch": 1.0937341451040081, + "grad_norm": 7.9375, + "learning_rate": 4.6387495882657295e-06, + "loss": 1.64373035, + "memory(GiB)": 111.15, + "step": 43115, + "train_speed(iter/s)": 0.447468 + }, + { + "acc": 0.65417633, + "epoch": 1.0938609842719431, + "grad_norm": 5.125, + "learning_rate": 4.63770371306663e-06, + "loss": 1.6305233, + "memory(GiB)": 111.15, + "step": 43120, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.6518446, + "epoch": 1.0939878234398783, + "grad_norm": 5.59375, + "learning_rate": 4.636657853802939e-06, + "loss": 1.55965977, + "memory(GiB)": 111.15, + "step": 43125, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.66722093, + "epoch": 1.0941146626078133, + "grad_norm": 5.78125, + "learning_rate": 4.635612010520659e-06, + "loss": 1.6997118, + "memory(GiB)": 111.15, + "step": 43130, + "train_speed(iter/s)": 0.447551 + }, + { + "acc": 0.64425454, + "epoch": 1.0942415017757483, + "grad_norm": 5.875, + "learning_rate": 4.6345661832657866e-06, + "loss": 1.64130669, + "memory(GiB)": 111.15, + "step": 43135, + "train_speed(iter/s)": 0.447578 + }, + { + "acc": 0.65306797, + "epoch": 1.0943683409436833, + "grad_norm": 5.09375, + "learning_rate": 4.633520372084327e-06, + "loss": 1.60207863, + "memory(GiB)": 111.15, + "step": 43140, + "train_speed(iter/s)": 0.447606 + }, + { + "acc": 0.64471245, + "epoch": 1.0944951801116185, + "grad_norm": 5.3125, + "learning_rate": 4.632474577022276e-06, + "loss": 1.68639107, + "memory(GiB)": 111.15, + "step": 43145, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.66564145, + "epoch": 1.0946220192795535, + "grad_norm": 5.21875, + "learning_rate": 4.631428798125637e-06, + "loss": 1.60233879, + "memory(GiB)": 111.15, + "step": 43150, + "train_speed(iter/s)": 0.447661 + }, + { + "acc": 0.65127916, + "epoch": 1.0947488584474885, + "grad_norm": 4.46875, + "learning_rate": 4.630383035440403e-06, + "loss": 1.61197567, + "memory(GiB)": 111.15, + "step": 43155, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.66479182, + "epoch": 1.0948756976154237, + "grad_norm": 5.0, + "learning_rate": 4.6293372890125724e-06, + "loss": 1.61481304, + "memory(GiB)": 111.15, + "step": 43160, + "train_speed(iter/s)": 0.447717 + }, + { + "acc": 0.65257854, + "epoch": 1.0950025367833587, + "grad_norm": 4.4375, + "learning_rate": 4.628291558888144e-06, + "loss": 1.65578022, + "memory(GiB)": 111.15, + "step": 43165, + "train_speed(iter/s)": 0.447744 + }, + { + "acc": 0.65413942, + "epoch": 1.0951293759512937, + "grad_norm": 5.4375, + "learning_rate": 4.627245845113113e-06, + "loss": 1.63474026, + "memory(GiB)": 111.15, + "step": 43170, + "train_speed(iter/s)": 0.447772 + }, + { + "acc": 0.64557409, + "epoch": 1.0952562151192289, + "grad_norm": 5.71875, + "learning_rate": 4.626200147733474e-06, + "loss": 1.60908928, + "memory(GiB)": 111.15, + "step": 43175, + "train_speed(iter/s)": 0.447799 + }, + { + "acc": 0.64561458, + "epoch": 1.0953830542871639, + "grad_norm": 5.15625, + "learning_rate": 4.62515446679522e-06, + "loss": 1.69374199, + "memory(GiB)": 111.15, + "step": 43180, + "train_speed(iter/s)": 0.447827 + }, + { + "acc": 0.67279644, + "epoch": 1.0955098934550989, + "grad_norm": 4.84375, + "learning_rate": 4.624108802344347e-06, + "loss": 1.52150116, + "memory(GiB)": 111.15, + "step": 43185, + "train_speed(iter/s)": 0.447855 + }, + { + "acc": 0.66117582, + "epoch": 1.095636732623034, + "grad_norm": 6.5625, + "learning_rate": 4.623063154426848e-06, + "loss": 1.60155468, + "memory(GiB)": 111.15, + "step": 43190, + "train_speed(iter/s)": 0.447882 + }, + { + "acc": 0.64417887, + "epoch": 1.095763571790969, + "grad_norm": 5.3125, + "learning_rate": 4.622017523088712e-06, + "loss": 1.62947731, + "memory(GiB)": 111.15, + "step": 43195, + "train_speed(iter/s)": 0.44791 + }, + { + "acc": 0.64856844, + "epoch": 1.095890410958904, + "grad_norm": 5.59375, + "learning_rate": 4.620971908375934e-06, + "loss": 1.64301128, + "memory(GiB)": 111.15, + "step": 43200, + "train_speed(iter/s)": 0.447937 + }, + { + "epoch": 1.095890410958904, + "eval_acc": 0.6432672411524195, + "eval_loss": 1.6079301834106445, + "eval_runtime": 114.4255, + "eval_samples_per_second": 55.669, + "eval_steps_per_second": 27.835, + "step": 43200 + }, + { + "acc": 0.66859722, + "epoch": 1.0960172501268393, + "grad_norm": 6.46875, + "learning_rate": 4.619926310334503e-06, + "loss": 1.55285549, + "memory(GiB)": 111.15, + "step": 43205, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.64241614, + "epoch": 1.0961440892947742, + "grad_norm": 5.53125, + "learning_rate": 4.618880729010413e-06, + "loss": 1.63571796, + "memory(GiB)": 111.15, + "step": 43210, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.66556234, + "epoch": 1.0962709284627092, + "grad_norm": 5.28125, + "learning_rate": 4.617835164449647e-06, + "loss": 1.54211273, + "memory(GiB)": 111.15, + "step": 43215, + "train_speed(iter/s)": 0.447468 + }, + { + "acc": 0.65664206, + "epoch": 1.0963977676306444, + "grad_norm": 6.28125, + "learning_rate": 4.616789616698197e-06, + "loss": 1.61631546, + "memory(GiB)": 111.15, + "step": 43220, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.6360239, + "epoch": 1.0965246067985794, + "grad_norm": 6.25, + "learning_rate": 4.61574408580205e-06, + "loss": 1.68419266, + "memory(GiB)": 111.15, + "step": 43225, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.66399536, + "epoch": 1.0966514459665144, + "grad_norm": 5.65625, + "learning_rate": 4.614698571807196e-06, + "loss": 1.57540979, + "memory(GiB)": 111.15, + "step": 43230, + "train_speed(iter/s)": 0.447552 + }, + { + "acc": 0.65063257, + "epoch": 1.0967782851344494, + "grad_norm": 5.4375, + "learning_rate": 4.6136530747596185e-06, + "loss": 1.63200531, + "memory(GiB)": 111.15, + "step": 43235, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.64680958, + "epoch": 1.0969051243023846, + "grad_norm": 5.125, + "learning_rate": 4.612607594705301e-06, + "loss": 1.67614746, + "memory(GiB)": 111.15, + "step": 43240, + "train_speed(iter/s)": 0.447607 + }, + { + "acc": 0.66331425, + "epoch": 1.0970319634703196, + "grad_norm": 5.25, + "learning_rate": 4.611562131690234e-06, + "loss": 1.6006815, + "memory(GiB)": 111.15, + "step": 43245, + "train_speed(iter/s)": 0.447635 + }, + { + "acc": 0.6593277, + "epoch": 1.0971588026382546, + "grad_norm": 5.4375, + "learning_rate": 4.610516685760399e-06, + "loss": 1.59404202, + "memory(GiB)": 111.15, + "step": 43250, + "train_speed(iter/s)": 0.447662 + }, + { + "acc": 0.6653554, + "epoch": 1.0972856418061898, + "grad_norm": 5.1875, + "learning_rate": 4.6094712569617775e-06, + "loss": 1.54944744, + "memory(GiB)": 111.15, + "step": 43255, + "train_speed(iter/s)": 0.44769 + }, + { + "acc": 0.64194574, + "epoch": 1.0974124809741248, + "grad_norm": 5.09375, + "learning_rate": 4.608425845340353e-06, + "loss": 1.66923714, + "memory(GiB)": 111.15, + "step": 43260, + "train_speed(iter/s)": 0.447718 + }, + { + "acc": 0.63083611, + "epoch": 1.0975393201420598, + "grad_norm": 5.8125, + "learning_rate": 4.607380450942109e-06, + "loss": 1.67406273, + "memory(GiB)": 111.15, + "step": 43265, + "train_speed(iter/s)": 0.447746 + }, + { + "acc": 0.64443703, + "epoch": 1.097666159309995, + "grad_norm": 5.78125, + "learning_rate": 4.606335073813028e-06, + "loss": 1.67276096, + "memory(GiB)": 111.15, + "step": 43270, + "train_speed(iter/s)": 0.447774 + }, + { + "acc": 0.63445864, + "epoch": 1.09779299847793, + "grad_norm": 4.84375, + "learning_rate": 4.605289713999085e-06, + "loss": 1.68042507, + "memory(GiB)": 111.15, + "step": 43275, + "train_speed(iter/s)": 0.447801 + }, + { + "acc": 0.64469566, + "epoch": 1.097919837645865, + "grad_norm": 4.375, + "learning_rate": 4.604244371546263e-06, + "loss": 1.62685776, + "memory(GiB)": 111.15, + "step": 43280, + "train_speed(iter/s)": 0.447829 + }, + { + "acc": 0.63526926, + "epoch": 1.0980466768138002, + "grad_norm": 4.9375, + "learning_rate": 4.603199046500539e-06, + "loss": 1.68029537, + "memory(GiB)": 111.15, + "step": 43285, + "train_speed(iter/s)": 0.447857 + }, + { + "acc": 0.6506074, + "epoch": 1.0981735159817352, + "grad_norm": 5.46875, + "learning_rate": 4.602153738907896e-06, + "loss": 1.64666214, + "memory(GiB)": 111.15, + "step": 43290, + "train_speed(iter/s)": 0.447884 + }, + { + "acc": 0.65821037, + "epoch": 1.0983003551496702, + "grad_norm": 5.40625, + "learning_rate": 4.601108448814306e-06, + "loss": 1.60583076, + "memory(GiB)": 111.15, + "step": 43295, + "train_speed(iter/s)": 0.447912 + }, + { + "acc": 0.63974838, + "epoch": 1.0984271943176052, + "grad_norm": 4.59375, + "learning_rate": 4.600063176265749e-06, + "loss": 1.55254173, + "memory(GiB)": 111.15, + "step": 43300, + "train_speed(iter/s)": 0.44794 + }, + { + "epoch": 1.0984271943176052, + "eval_acc": 0.6433152843930071, + "eval_loss": 1.6079305410385132, + "eval_runtime": 113.0512, + "eval_samples_per_second": 56.346, + "eval_steps_per_second": 28.173, + "step": 43300 + }, + { + "acc": 0.65727563, + "epoch": 1.0985540334855404, + "grad_norm": 5.53125, + "learning_rate": 4.599017921308196e-06, + "loss": 1.60708199, + "memory(GiB)": 111.15, + "step": 43305, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.63844452, + "epoch": 1.0986808726534754, + "grad_norm": 5.3125, + "learning_rate": 4.5979726839876285e-06, + "loss": 1.6336071, + "memory(GiB)": 111.15, + "step": 43310, + "train_speed(iter/s)": 0.44745 + }, + { + "acc": 0.63467751, + "epoch": 1.0988077118214103, + "grad_norm": 4.96875, + "learning_rate": 4.596927464350015e-06, + "loss": 1.62585945, + "memory(GiB)": 111.15, + "step": 43315, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.63524923, + "epoch": 1.0989345509893456, + "grad_norm": 5.0625, + "learning_rate": 4.595882262441331e-06, + "loss": 1.63681736, + "memory(GiB)": 111.15, + "step": 43320, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.66428719, + "epoch": 1.0990613901572805, + "grad_norm": 4.25, + "learning_rate": 4.5948370783075505e-06, + "loss": 1.56777344, + "memory(GiB)": 111.15, + "step": 43325, + "train_speed(iter/s)": 0.447533 + }, + { + "acc": 0.65139337, + "epoch": 1.0991882293252155, + "grad_norm": 5.9375, + "learning_rate": 4.5937919119946445e-06, + "loss": 1.64537086, + "memory(GiB)": 111.15, + "step": 43330, + "train_speed(iter/s)": 0.447561 + }, + { + "acc": 0.66201086, + "epoch": 1.0993150684931507, + "grad_norm": 6.09375, + "learning_rate": 4.592746763548582e-06, + "loss": 1.58151798, + "memory(GiB)": 111.15, + "step": 43335, + "train_speed(iter/s)": 0.447588 + }, + { + "acc": 0.6571106, + "epoch": 1.0994419076610857, + "grad_norm": 5.0625, + "learning_rate": 4.591701633015336e-06, + "loss": 1.63487015, + "memory(GiB)": 111.15, + "step": 43340, + "train_speed(iter/s)": 0.447616 + }, + { + "acc": 0.66150961, + "epoch": 1.0995687468290207, + "grad_norm": 4.8125, + "learning_rate": 4.590656520440876e-06, + "loss": 1.63231621, + "memory(GiB)": 111.15, + "step": 43345, + "train_speed(iter/s)": 0.447644 + }, + { + "acc": 0.67355814, + "epoch": 1.099695585996956, + "grad_norm": 5.28125, + "learning_rate": 4.58961142587117e-06, + "loss": 1.54215012, + "memory(GiB)": 111.15, + "step": 43350, + "train_speed(iter/s)": 0.447671 + }, + { + "acc": 0.64136171, + "epoch": 1.099822425164891, + "grad_norm": 5.375, + "learning_rate": 4.588566349352185e-06, + "loss": 1.57358217, + "memory(GiB)": 111.15, + "step": 43355, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.65167885, + "epoch": 1.099949264332826, + "grad_norm": 5.90625, + "learning_rate": 4.5875212909298885e-06, + "loss": 1.59133663, + "memory(GiB)": 111.15, + "step": 43360, + "train_speed(iter/s)": 0.447726 + }, + { + "acc": 0.66197891, + "epoch": 1.1000761035007611, + "grad_norm": 4.90625, + "learning_rate": 4.586476250650246e-06, + "loss": 1.56490755, + "memory(GiB)": 111.15, + "step": 43365, + "train_speed(iter/s)": 0.447754 + }, + { + "acc": 0.65178537, + "epoch": 1.1002029426686961, + "grad_norm": 4.75, + "learning_rate": 4.585431228559228e-06, + "loss": 1.5937314, + "memory(GiB)": 111.15, + "step": 43370, + "train_speed(iter/s)": 0.447782 + }, + { + "acc": 0.6509562, + "epoch": 1.100329781836631, + "grad_norm": 5.0, + "learning_rate": 4.584386224702792e-06, + "loss": 1.57431412, + "memory(GiB)": 111.15, + "step": 43375, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.65648108, + "epoch": 1.1004566210045663, + "grad_norm": 6.0, + "learning_rate": 4.583341239126906e-06, + "loss": 1.629426, + "memory(GiB)": 111.15, + "step": 43380, + "train_speed(iter/s)": 0.447837 + }, + { + "acc": 0.6466567, + "epoch": 1.1005834601725013, + "grad_norm": 5.25, + "learning_rate": 4.582296271877534e-06, + "loss": 1.68363113, + "memory(GiB)": 111.15, + "step": 43385, + "train_speed(iter/s)": 0.447864 + }, + { + "acc": 0.64667435, + "epoch": 1.1007102993404363, + "grad_norm": 5.75, + "learning_rate": 4.581251323000636e-06, + "loss": 1.63045006, + "memory(GiB)": 111.15, + "step": 43390, + "train_speed(iter/s)": 0.447892 + }, + { + "acc": 0.63172913, + "epoch": 1.1008371385083713, + "grad_norm": 4.9375, + "learning_rate": 4.580206392542175e-06, + "loss": 1.66979218, + "memory(GiB)": 111.15, + "step": 43395, + "train_speed(iter/s)": 0.447919 + }, + { + "acc": 0.66061268, + "epoch": 1.1009639776763065, + "grad_norm": 5.15625, + "learning_rate": 4.579161480548109e-06, + "loss": 1.64188118, + "memory(GiB)": 111.15, + "step": 43400, + "train_speed(iter/s)": 0.447947 + }, + { + "epoch": 1.1009639776763065, + "eval_acc": 0.6432902183544397, + "eval_loss": 1.6081310510635376, + "eval_runtime": 114.5032, + "eval_samples_per_second": 55.632, + "eval_steps_per_second": 27.816, + "step": 43400 + }, + { + "acc": 0.65021391, + "epoch": 1.1010908168442415, + "grad_norm": 4.71875, + "learning_rate": 4.578116587064402e-06, + "loss": 1.65231171, + "memory(GiB)": 111.15, + "step": 43405, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.65788941, + "epoch": 1.1012176560121765, + "grad_norm": 8.0, + "learning_rate": 4.577071712137012e-06, + "loss": 1.61938438, + "memory(GiB)": 111.15, + "step": 43410, + "train_speed(iter/s)": 0.447451 + }, + { + "acc": 0.65242443, + "epoch": 1.1013444951801117, + "grad_norm": 5.1875, + "learning_rate": 4.576026855811893e-06, + "loss": 1.63090172, + "memory(GiB)": 111.15, + "step": 43415, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.66317925, + "epoch": 1.1014713343480467, + "grad_norm": 8.8125, + "learning_rate": 4.5749820181350095e-06, + "loss": 1.60119286, + "memory(GiB)": 111.15, + "step": 43420, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.64037914, + "epoch": 1.1015981735159817, + "grad_norm": 5.125, + "learning_rate": 4.57393719915231e-06, + "loss": 1.64320908, + "memory(GiB)": 111.15, + "step": 43425, + "train_speed(iter/s)": 0.447533 + }, + { + "acc": 0.64961319, + "epoch": 1.1017250126839169, + "grad_norm": 5.28125, + "learning_rate": 4.5728923989097604e-06, + "loss": 1.59660254, + "memory(GiB)": 111.15, + "step": 43430, + "train_speed(iter/s)": 0.44756 + }, + { + "acc": 0.64875803, + "epoch": 1.1018518518518519, + "grad_norm": 5.46875, + "learning_rate": 4.571847617453306e-06, + "loss": 1.56322613, + "memory(GiB)": 111.15, + "step": 43435, + "train_speed(iter/s)": 0.447587 + }, + { + "acc": 0.65476894, + "epoch": 1.1019786910197868, + "grad_norm": 5.375, + "learning_rate": 4.570802854828906e-06, + "loss": 1.6484024, + "memory(GiB)": 111.15, + "step": 43440, + "train_speed(iter/s)": 0.447615 + }, + { + "acc": 0.65546503, + "epoch": 1.102105530187722, + "grad_norm": 5.4375, + "learning_rate": 4.569758111082512e-06, + "loss": 1.56123734, + "memory(GiB)": 111.15, + "step": 43445, + "train_speed(iter/s)": 0.447642 + }, + { + "acc": 0.66491127, + "epoch": 1.102232369355657, + "grad_norm": 6.125, + "learning_rate": 4.568713386260078e-06, + "loss": 1.54644613, + "memory(GiB)": 111.15, + "step": 43450, + "train_speed(iter/s)": 0.447669 + }, + { + "acc": 0.6680563, + "epoch": 1.102359208523592, + "grad_norm": 6.0625, + "learning_rate": 4.567668680407555e-06, + "loss": 1.59939766, + "memory(GiB)": 111.15, + "step": 43455, + "train_speed(iter/s)": 0.447697 + }, + { + "acc": 0.64586029, + "epoch": 1.102486047691527, + "grad_norm": 6.1875, + "learning_rate": 4.566623993570893e-06, + "loss": 1.66532745, + "memory(GiB)": 111.15, + "step": 43460, + "train_speed(iter/s)": 0.447724 + }, + { + "acc": 0.63109884, + "epoch": 1.1026128868594622, + "grad_norm": 5.46875, + "learning_rate": 4.565579325796043e-06, + "loss": 1.73518524, + "memory(GiB)": 111.15, + "step": 43465, + "train_speed(iter/s)": 0.447752 + }, + { + "acc": 0.65850573, + "epoch": 1.1027397260273972, + "grad_norm": 5.09375, + "learning_rate": 4.564534677128954e-06, + "loss": 1.63165188, + "memory(GiB)": 111.15, + "step": 43470, + "train_speed(iter/s)": 0.447779 + }, + { + "acc": 0.64889488, + "epoch": 1.1028665651953322, + "grad_norm": 4.59375, + "learning_rate": 4.563490047615574e-06, + "loss": 1.62906837, + "memory(GiB)": 111.15, + "step": 43475, + "train_speed(iter/s)": 0.447806 + }, + { + "acc": 0.65114002, + "epoch": 1.1029934043632674, + "grad_norm": 5.75, + "learning_rate": 4.56244543730185e-06, + "loss": 1.61588478, + "memory(GiB)": 111.15, + "step": 43480, + "train_speed(iter/s)": 0.447834 + }, + { + "acc": 0.65293856, + "epoch": 1.1031202435312024, + "grad_norm": 4.90625, + "learning_rate": 4.561400846233729e-06, + "loss": 1.57419949, + "memory(GiB)": 111.15, + "step": 43485, + "train_speed(iter/s)": 0.447861 + }, + { + "acc": 0.64807096, + "epoch": 1.1032470826991374, + "grad_norm": 4.75, + "learning_rate": 4.56035627445716e-06, + "loss": 1.57269135, + "memory(GiB)": 111.15, + "step": 43490, + "train_speed(iter/s)": 0.447888 + }, + { + "acc": 0.63613081, + "epoch": 1.1033739218670726, + "grad_norm": 6.0, + "learning_rate": 4.55931172201808e-06, + "loss": 1.7466217, + "memory(GiB)": 111.15, + "step": 43495, + "train_speed(iter/s)": 0.447915 + }, + { + "acc": 0.64539809, + "epoch": 1.1035007610350076, + "grad_norm": 5.09375, + "learning_rate": 4.558267188962441e-06, + "loss": 1.59908066, + "memory(GiB)": 111.15, + "step": 43500, + "train_speed(iter/s)": 0.447943 + }, + { + "epoch": 1.1035007610350076, + "eval_acc": 0.6432735076620614, + "eval_loss": 1.6079615354537964, + "eval_runtime": 115.645, + "eval_samples_per_second": 55.082, + "eval_steps_per_second": 27.541, + "step": 43500 + }, + { + "acc": 0.65048227, + "epoch": 1.1036276002029426, + "grad_norm": 4.5625, + "learning_rate": 4.557222675336182e-06, + "loss": 1.66699333, + "memory(GiB)": 111.15, + "step": 43505, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.64676151, + "epoch": 1.1037544393708778, + "grad_norm": 5.4375, + "learning_rate": 4.556178181185249e-06, + "loss": 1.5879631, + "memory(GiB)": 111.15, + "step": 43510, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.64554763, + "epoch": 1.1038812785388128, + "grad_norm": 4.90625, + "learning_rate": 4.555133706555579e-06, + "loss": 1.68190308, + "memory(GiB)": 111.15, + "step": 43515, + "train_speed(iter/s)": 0.44747 + }, + { + "acc": 0.64101658, + "epoch": 1.1040081177067478, + "grad_norm": 5.875, + "learning_rate": 4.554089251493115e-06, + "loss": 1.68351669, + "memory(GiB)": 111.15, + "step": 43520, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.66073627, + "epoch": 1.104134956874683, + "grad_norm": 5.71875, + "learning_rate": 4.553044816043796e-06, + "loss": 1.5887537, + "memory(GiB)": 111.15, + "step": 43525, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.65184112, + "epoch": 1.104261796042618, + "grad_norm": 6.125, + "learning_rate": 4.552000400253563e-06, + "loss": 1.61050129, + "memory(GiB)": 111.15, + "step": 43530, + "train_speed(iter/s)": 0.447551 + }, + { + "acc": 0.65982385, + "epoch": 1.104388635210553, + "grad_norm": 5.9375, + "learning_rate": 4.550956004168352e-06, + "loss": 1.6211298, + "memory(GiB)": 111.15, + "step": 43535, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.64694424, + "epoch": 1.1045154743784882, + "grad_norm": 4.5625, + "learning_rate": 4.5499116278341e-06, + "loss": 1.6568264, + "memory(GiB)": 111.15, + "step": 43540, + "train_speed(iter/s)": 0.447606 + }, + { + "acc": 0.64717407, + "epoch": 1.1046423135464232, + "grad_norm": 5.5625, + "learning_rate": 4.548867271296745e-06, + "loss": 1.63925514, + "memory(GiB)": 111.15, + "step": 43545, + "train_speed(iter/s)": 0.447633 + }, + { + "acc": 0.668786, + "epoch": 1.1047691527143582, + "grad_norm": 5.34375, + "learning_rate": 4.547822934602222e-06, + "loss": 1.56878967, + "memory(GiB)": 111.15, + "step": 43550, + "train_speed(iter/s)": 0.44766 + }, + { + "acc": 0.65407186, + "epoch": 1.1048959918822931, + "grad_norm": 6.03125, + "learning_rate": 4.5467786177964635e-06, + "loss": 1.61240654, + "memory(GiB)": 111.15, + "step": 43555, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.64877152, + "epoch": 1.1050228310502284, + "grad_norm": 5.28125, + "learning_rate": 4.545734320925406e-06, + "loss": 1.64874153, + "memory(GiB)": 111.15, + "step": 43560, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.66243849, + "epoch": 1.1051496702181633, + "grad_norm": 6.21875, + "learning_rate": 4.544690044034981e-06, + "loss": 1.63725128, + "memory(GiB)": 111.15, + "step": 43565, + "train_speed(iter/s)": 0.447742 + }, + { + "acc": 0.65426559, + "epoch": 1.1052765093860983, + "grad_norm": 5.15625, + "learning_rate": 4.543645787171122e-06, + "loss": 1.61616249, + "memory(GiB)": 111.15, + "step": 43570, + "train_speed(iter/s)": 0.447769 + }, + { + "acc": 0.63866663, + "epoch": 1.1054033485540335, + "grad_norm": 5.0625, + "learning_rate": 4.5426015503797565e-06, + "loss": 1.67352505, + "memory(GiB)": 111.15, + "step": 43575, + "train_speed(iter/s)": 0.447796 + }, + { + "acc": 0.64731407, + "epoch": 1.1055301877219685, + "grad_norm": 5.65625, + "learning_rate": 4.5415573337068185e-06, + "loss": 1.65331612, + "memory(GiB)": 111.15, + "step": 43580, + "train_speed(iter/s)": 0.447823 + }, + { + "acc": 0.64814091, + "epoch": 1.1056570268899035, + "grad_norm": 6.0625, + "learning_rate": 4.540513137198233e-06, + "loss": 1.65565243, + "memory(GiB)": 111.15, + "step": 43585, + "train_speed(iter/s)": 0.44785 + }, + { + "acc": 0.66227474, + "epoch": 1.1057838660578387, + "grad_norm": 5.375, + "learning_rate": 4.539468960899936e-06, + "loss": 1.52074642, + "memory(GiB)": 111.15, + "step": 43590, + "train_speed(iter/s)": 0.447877 + }, + { + "acc": 0.63618751, + "epoch": 1.1059107052257737, + "grad_norm": 6.96875, + "learning_rate": 4.538424804857847e-06, + "loss": 1.68571815, + "memory(GiB)": 111.15, + "step": 43595, + "train_speed(iter/s)": 0.447904 + }, + { + "acc": 0.65476508, + "epoch": 1.1060375443937087, + "grad_norm": 4.78125, + "learning_rate": 4.537380669117896e-06, + "loss": 1.64121552, + "memory(GiB)": 111.15, + "step": 43600, + "train_speed(iter/s)": 0.447931 + }, + { + "epoch": 1.1060375443937087, + "eval_acc": 0.6433604032624285, + "eval_loss": 1.6079680919647217, + "eval_runtime": 114.0327, + "eval_samples_per_second": 55.861, + "eval_steps_per_second": 27.931, + "step": 43600 + }, + { + "acc": 0.65443902, + "epoch": 1.106164383561644, + "grad_norm": 6.375, + "learning_rate": 4.536336553726008e-06, + "loss": 1.58929148, + "memory(GiB)": 111.15, + "step": 43605, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.65535836, + "epoch": 1.106291222729579, + "grad_norm": 4.8125, + "learning_rate": 4.535292458728112e-06, + "loss": 1.61774578, + "memory(GiB)": 111.15, + "step": 43610, + "train_speed(iter/s)": 0.44744 + }, + { + "acc": 0.6451571, + "epoch": 1.106418061897514, + "grad_norm": 5.84375, + "learning_rate": 4.534248384170126e-06, + "loss": 1.63786201, + "memory(GiB)": 111.15, + "step": 43615, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.64870329, + "epoch": 1.106544901065449, + "grad_norm": 5.4375, + "learning_rate": 4.533204330097974e-06, + "loss": 1.62143669, + "memory(GiB)": 111.15, + "step": 43620, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.66259298, + "epoch": 1.106671740233384, + "grad_norm": 4.4375, + "learning_rate": 4.532160296557581e-06, + "loss": 1.62355938, + "memory(GiB)": 111.15, + "step": 43625, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.66635122, + "epoch": 1.106798579401319, + "grad_norm": 7.53125, + "learning_rate": 4.531116283594868e-06, + "loss": 1.55716248, + "memory(GiB)": 111.15, + "step": 43630, + "train_speed(iter/s)": 0.447548 + }, + { + "acc": 0.67196503, + "epoch": 1.106925418569254, + "grad_norm": 5.75, + "learning_rate": 4.530072291255753e-06, + "loss": 1.57118921, + "memory(GiB)": 111.15, + "step": 43635, + "train_speed(iter/s)": 0.447575 + }, + { + "acc": 0.65551481, + "epoch": 1.1070522577371893, + "grad_norm": 5.34375, + "learning_rate": 4.529028319586157e-06, + "loss": 1.61909847, + "memory(GiB)": 111.15, + "step": 43640, + "train_speed(iter/s)": 0.447603 + }, + { + "acc": 0.64416184, + "epoch": 1.1071790969051243, + "grad_norm": 6.03125, + "learning_rate": 4.527984368631997e-06, + "loss": 1.71984138, + "memory(GiB)": 111.15, + "step": 43645, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.64371533, + "epoch": 1.1073059360730593, + "grad_norm": 5.4375, + "learning_rate": 4.526940438439196e-06, + "loss": 1.53390865, + "memory(GiB)": 111.15, + "step": 43650, + "train_speed(iter/s)": 0.447657 + }, + { + "acc": 0.65462222, + "epoch": 1.1074327752409945, + "grad_norm": 6.84375, + "learning_rate": 4.525896529053662e-06, + "loss": 1.63867989, + "memory(GiB)": 111.15, + "step": 43655, + "train_speed(iter/s)": 0.447684 + }, + { + "acc": 0.6484807, + "epoch": 1.1075596144089295, + "grad_norm": 4.9375, + "learning_rate": 4.524852640521318e-06, + "loss": 1.63207855, + "memory(GiB)": 111.15, + "step": 43660, + "train_speed(iter/s)": 0.447712 + }, + { + "acc": 0.65583491, + "epoch": 1.1076864535768645, + "grad_norm": 5.25, + "learning_rate": 4.523808772888073e-06, + "loss": 1.57315559, + "memory(GiB)": 111.15, + "step": 43665, + "train_speed(iter/s)": 0.447739 + }, + { + "acc": 0.65190787, + "epoch": 1.1078132927447997, + "grad_norm": 4.6875, + "learning_rate": 4.522764926199848e-06, + "loss": 1.63750877, + "memory(GiB)": 111.15, + "step": 43670, + "train_speed(iter/s)": 0.447766 + }, + { + "acc": 0.64582796, + "epoch": 1.1079401319127347, + "grad_norm": 5.625, + "learning_rate": 4.5217211005025516e-06, + "loss": 1.57468548, + "memory(GiB)": 111.15, + "step": 43675, + "train_speed(iter/s)": 0.447793 + }, + { + "acc": 0.64481063, + "epoch": 1.1080669710806696, + "grad_norm": 5.9375, + "learning_rate": 4.520677295842095e-06, + "loss": 1.59120722, + "memory(GiB)": 111.15, + "step": 43680, + "train_speed(iter/s)": 0.44782 + }, + { + "acc": 0.6792923, + "epoch": 1.1081938102486049, + "grad_norm": 6.375, + "learning_rate": 4.5196335122643915e-06, + "loss": 1.50496445, + "memory(GiB)": 111.15, + "step": 43685, + "train_speed(iter/s)": 0.447848 + }, + { + "acc": 0.65321269, + "epoch": 1.1083206494165398, + "grad_norm": 5.65625, + "learning_rate": 4.518589749815352e-06, + "loss": 1.57540808, + "memory(GiB)": 111.15, + "step": 43690, + "train_speed(iter/s)": 0.447875 + }, + { + "acc": 0.64867992, + "epoch": 1.1084474885844748, + "grad_norm": 5.6875, + "learning_rate": 4.517546008540884e-06, + "loss": 1.62391605, + "memory(GiB)": 111.15, + "step": 43695, + "train_speed(iter/s)": 0.447902 + }, + { + "acc": 0.64859099, + "epoch": 1.10857432775241, + "grad_norm": 5.21875, + "learning_rate": 4.5165022884868946e-06, + "loss": 1.61436348, + "memory(GiB)": 111.15, + "step": 43700, + "train_speed(iter/s)": 0.447929 + }, + { + "epoch": 1.10857432775241, + "eval_acc": 0.6433299062488381, + "eval_loss": 1.607836365699768, + "eval_runtime": 113.4399, + "eval_samples_per_second": 56.153, + "eval_steps_per_second": 28.077, + "step": 43700 + }, + { + "acc": 0.63205709, + "epoch": 1.108701166920345, + "grad_norm": 5.78125, + "learning_rate": 4.515458589699295e-06, + "loss": 1.70671482, + "memory(GiB)": 111.15, + "step": 43705, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.65764709, + "epoch": 1.10882800608828, + "grad_norm": 4.65625, + "learning_rate": 4.514414912223991e-06, + "loss": 1.59906254, + "memory(GiB)": 111.15, + "step": 43710, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.66517401, + "epoch": 1.108954845256215, + "grad_norm": 5.625, + "learning_rate": 4.513371256106885e-06, + "loss": 1.57330589, + "memory(GiB)": 111.15, + "step": 43715, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.63972378, + "epoch": 1.1090816844241502, + "grad_norm": 5.9375, + "learning_rate": 4.512327621393885e-06, + "loss": 1.63225956, + "memory(GiB)": 111.15, + "step": 43720, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.67330337, + "epoch": 1.1092085235920852, + "grad_norm": 5.28125, + "learning_rate": 4.511284008130892e-06, + "loss": 1.55555668, + "memory(GiB)": 111.15, + "step": 43725, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64832211, + "epoch": 1.1093353627600202, + "grad_norm": 5.25, + "learning_rate": 4.510240416363813e-06, + "loss": 1.60936871, + "memory(GiB)": 111.15, + "step": 43730, + "train_speed(iter/s)": 0.44755 + }, + { + "acc": 0.63588448, + "epoch": 1.1094622019279554, + "grad_norm": 4.90625, + "learning_rate": 4.5091968461385455e-06, + "loss": 1.67983971, + "memory(GiB)": 111.15, + "step": 43735, + "train_speed(iter/s)": 0.447577 + }, + { + "acc": 0.6545136, + "epoch": 1.1095890410958904, + "grad_norm": 5.09375, + "learning_rate": 4.508153297500993e-06, + "loss": 1.60611191, + "memory(GiB)": 111.15, + "step": 43740, + "train_speed(iter/s)": 0.447604 + }, + { + "acc": 0.65158997, + "epoch": 1.1097158802638254, + "grad_norm": 5.3125, + "learning_rate": 4.507109770497052e-06, + "loss": 1.60791626, + "memory(GiB)": 111.15, + "step": 43745, + "train_speed(iter/s)": 0.447631 + }, + { + "acc": 0.6608737, + "epoch": 1.1098427194317606, + "grad_norm": 5.15625, + "learning_rate": 4.506066265172626e-06, + "loss": 1.57341871, + "memory(GiB)": 111.15, + "step": 43750, + "train_speed(iter/s)": 0.447658 + }, + { + "acc": 0.62668147, + "epoch": 1.1099695585996956, + "grad_norm": 5.4375, + "learning_rate": 4.505022781573611e-06, + "loss": 1.70450211, + "memory(GiB)": 111.15, + "step": 43755, + "train_speed(iter/s)": 0.447685 + }, + { + "acc": 0.65914383, + "epoch": 1.1100963977676306, + "grad_norm": 4.84375, + "learning_rate": 4.503979319745902e-06, + "loss": 1.65097389, + "memory(GiB)": 111.15, + "step": 43760, + "train_speed(iter/s)": 0.447712 + }, + { + "acc": 0.64734373, + "epoch": 1.1102232369355658, + "grad_norm": 5.15625, + "learning_rate": 4.502935879735398e-06, + "loss": 1.6246088, + "memory(GiB)": 111.15, + "step": 43765, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.65419312, + "epoch": 1.1103500761035008, + "grad_norm": 4.9375, + "learning_rate": 4.5018924615879956e-06, + "loss": 1.61998558, + "memory(GiB)": 111.15, + "step": 43770, + "train_speed(iter/s)": 0.447765 + }, + { + "acc": 0.65592775, + "epoch": 1.1104769152714358, + "grad_norm": 4.78125, + "learning_rate": 4.500849065349584e-06, + "loss": 1.6009388, + "memory(GiB)": 111.15, + "step": 43775, + "train_speed(iter/s)": 0.447792 + }, + { + "acc": 0.64072948, + "epoch": 1.1106037544393708, + "grad_norm": 6.34375, + "learning_rate": 4.499805691066059e-06, + "loss": 1.59774771, + "memory(GiB)": 111.15, + "step": 43780, + "train_speed(iter/s)": 0.447819 + }, + { + "acc": 0.64454565, + "epoch": 1.110730593607306, + "grad_norm": 5.8125, + "learning_rate": 4.498762338783314e-06, + "loss": 1.63502274, + "memory(GiB)": 111.15, + "step": 43785, + "train_speed(iter/s)": 0.447845 + }, + { + "acc": 0.65958037, + "epoch": 1.110857432775241, + "grad_norm": 5.28125, + "learning_rate": 4.49771900854724e-06, + "loss": 1.58114986, + "memory(GiB)": 111.15, + "step": 43790, + "train_speed(iter/s)": 0.447872 + }, + { + "acc": 0.64242678, + "epoch": 1.110984271943176, + "grad_norm": 5.8125, + "learning_rate": 4.496675700403724e-06, + "loss": 1.69098415, + "memory(GiB)": 111.15, + "step": 43795, + "train_speed(iter/s)": 0.447899 + }, + { + "acc": 0.64786968, + "epoch": 1.1111111111111112, + "grad_norm": 6.0, + "learning_rate": 4.495632414398659e-06, + "loss": 1.69711342, + "memory(GiB)": 111.15, + "step": 43800, + "train_speed(iter/s)": 0.447926 + }, + { + "epoch": 1.1111111111111112, + "eval_acc": 0.6432442639503994, + "eval_loss": 1.607789158821106, + "eval_runtime": 114.488, + "eval_samples_per_second": 55.639, + "eval_steps_per_second": 27.82, + "step": 43800 + }, + { + "acc": 0.63777924, + "epoch": 1.1112379502790461, + "grad_norm": 5.40625, + "learning_rate": 4.494589150577932e-06, + "loss": 1.71778812, + "memory(GiB)": 111.15, + "step": 43805, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.66468453, + "epoch": 1.1113647894469811, + "grad_norm": 5.96875, + "learning_rate": 4.493545908987432e-06, + "loss": 1.63095341, + "memory(GiB)": 111.15, + "step": 43810, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.63431921, + "epoch": 1.1114916286149163, + "grad_norm": 4.90625, + "learning_rate": 4.492502689673044e-06, + "loss": 1.71169815, + "memory(GiB)": 111.15, + "step": 43815, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.66547384, + "epoch": 1.1116184677828513, + "grad_norm": 5.53125, + "learning_rate": 4.491459492680651e-06, + "loss": 1.60673866, + "memory(GiB)": 111.15, + "step": 43820, + "train_speed(iter/s)": 0.447488 + }, + { + "acc": 0.65815578, + "epoch": 1.1117453069507863, + "grad_norm": 5.59375, + "learning_rate": 4.4904163180561425e-06, + "loss": 1.53903627, + "memory(GiB)": 111.15, + "step": 43825, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.63364606, + "epoch": 1.1118721461187215, + "grad_norm": 4.9375, + "learning_rate": 4.4893731658453996e-06, + "loss": 1.62379837, + "memory(GiB)": 111.15, + "step": 43830, + "train_speed(iter/s)": 0.447542 + }, + { + "acc": 0.65906, + "epoch": 1.1119989852866565, + "grad_norm": 4.96875, + "learning_rate": 4.4883300360943035e-06, + "loss": 1.60301685, + "memory(GiB)": 111.15, + "step": 43835, + "train_speed(iter/s)": 0.447569 + }, + { + "acc": 0.64513493, + "epoch": 1.1121258244545915, + "grad_norm": 5.5625, + "learning_rate": 4.4872869288487366e-06, + "loss": 1.61025333, + "memory(GiB)": 111.15, + "step": 43840, + "train_speed(iter/s)": 0.447595 + }, + { + "acc": 0.64752474, + "epoch": 1.1122526636225267, + "grad_norm": 4.71875, + "learning_rate": 4.48624384415458e-06, + "loss": 1.56753292, + "memory(GiB)": 111.15, + "step": 43845, + "train_speed(iter/s)": 0.447622 + }, + { + "acc": 0.65127726, + "epoch": 1.1123795027904617, + "grad_norm": 5.34375, + "learning_rate": 4.485200782057715e-06, + "loss": 1.67358112, + "memory(GiB)": 111.15, + "step": 43850, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.64066677, + "epoch": 1.1125063419583967, + "grad_norm": 5.84375, + "learning_rate": 4.4841577426040145e-06, + "loss": 1.64540443, + "memory(GiB)": 111.15, + "step": 43855, + "train_speed(iter/s)": 0.447676 + }, + { + "acc": 0.65520682, + "epoch": 1.112633181126332, + "grad_norm": 5.1875, + "learning_rate": 4.483114725839361e-06, + "loss": 1.61334229, + "memory(GiB)": 111.15, + "step": 43860, + "train_speed(iter/s)": 0.447702 + }, + { + "acc": 0.64900894, + "epoch": 1.112760020294267, + "grad_norm": 5.875, + "learning_rate": 4.482071731809629e-06, + "loss": 1.65554008, + "memory(GiB)": 111.15, + "step": 43865, + "train_speed(iter/s)": 0.447729 + }, + { + "acc": 0.63979573, + "epoch": 1.112886859462202, + "grad_norm": 5.5, + "learning_rate": 4.481028760560697e-06, + "loss": 1.65602264, + "memory(GiB)": 111.15, + "step": 43870, + "train_speed(iter/s)": 0.447755 + }, + { + "acc": 0.65832863, + "epoch": 1.1130136986301369, + "grad_norm": 5.1875, + "learning_rate": 4.479985812138435e-06, + "loss": 1.61183891, + "memory(GiB)": 111.15, + "step": 43875, + "train_speed(iter/s)": 0.447782 + }, + { + "acc": 0.64696579, + "epoch": 1.113140537798072, + "grad_norm": 5.40625, + "learning_rate": 4.478942886588719e-06, + "loss": 1.58558369, + "memory(GiB)": 111.15, + "step": 43880, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.64988232, + "epoch": 1.113267376966007, + "grad_norm": 4.875, + "learning_rate": 4.47789998395742e-06, + "loss": 1.614501, + "memory(GiB)": 111.15, + "step": 43885, + "train_speed(iter/s)": 0.447836 + }, + { + "acc": 0.65577488, + "epoch": 1.113394216133942, + "grad_norm": 6.40625, + "learning_rate": 4.476857104290413e-06, + "loss": 1.65937443, + "memory(GiB)": 111.15, + "step": 43890, + "train_speed(iter/s)": 0.447863 + }, + { + "acc": 0.66699953, + "epoch": 1.1135210553018773, + "grad_norm": 6.5, + "learning_rate": 4.4758142476335655e-06, + "loss": 1.61105213, + "memory(GiB)": 111.15, + "step": 43895, + "train_speed(iter/s)": 0.44789 + }, + { + "acc": 0.66692457, + "epoch": 1.1136478944698123, + "grad_norm": 4.46875, + "learning_rate": 4.474771414032747e-06, + "loss": 1.68279419, + "memory(GiB)": 111.15, + "step": 43900, + "train_speed(iter/s)": 0.447916 + }, + { + "epoch": 1.1136478944698123, + "eval_acc": 0.6432785208697748, + "eval_loss": 1.607942819595337, + "eval_runtime": 115.4914, + "eval_samples_per_second": 55.156, + "eval_steps_per_second": 27.578, + "step": 43900 + }, + { + "acc": 0.65618267, + "epoch": 1.1137747336377473, + "grad_norm": 5.5625, + "learning_rate": 4.473728603533827e-06, + "loss": 1.60887527, + "memory(GiB)": 111.15, + "step": 43905, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.64473619, + "epoch": 1.1139015728056825, + "grad_norm": 6.09375, + "learning_rate": 4.472685816182674e-06, + "loss": 1.65027294, + "memory(GiB)": 111.15, + "step": 43910, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.64944496, + "epoch": 1.1140284119736175, + "grad_norm": 5.90625, + "learning_rate": 4.471643052025152e-06, + "loss": 1.6497776, + "memory(GiB)": 111.15, + "step": 43915, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.63943596, + "epoch": 1.1141552511415524, + "grad_norm": 5.53125, + "learning_rate": 4.470600311107127e-06, + "loss": 1.69407959, + "memory(GiB)": 111.15, + "step": 43920, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.64081774, + "epoch": 1.1142820903094877, + "grad_norm": 5.3125, + "learning_rate": 4.469557593474464e-06, + "loss": 1.61218548, + "memory(GiB)": 111.15, + "step": 43925, + "train_speed(iter/s)": 0.447502 + }, + { + "acc": 0.64238997, + "epoch": 1.1144089294774226, + "grad_norm": 5.59375, + "learning_rate": 4.468514899173027e-06, + "loss": 1.63466797, + "memory(GiB)": 111.15, + "step": 43930, + "train_speed(iter/s)": 0.447529 + }, + { + "acc": 0.66194081, + "epoch": 1.1145357686453576, + "grad_norm": 5.46875, + "learning_rate": 4.4674722282486775e-06, + "loss": 1.57595196, + "memory(GiB)": 111.15, + "step": 43935, + "train_speed(iter/s)": 0.447556 + }, + { + "acc": 0.65851707, + "epoch": 1.1146626078132926, + "grad_norm": 5.59375, + "learning_rate": 4.4664295807472765e-06, + "loss": 1.60494995, + "memory(GiB)": 111.15, + "step": 43940, + "train_speed(iter/s)": 0.447583 + }, + { + "acc": 0.64260588, + "epoch": 1.1147894469812278, + "grad_norm": 6.3125, + "learning_rate": 4.465386956714684e-06, + "loss": 1.61171074, + "memory(GiB)": 111.15, + "step": 43945, + "train_speed(iter/s)": 0.44761 + }, + { + "acc": 0.64786301, + "epoch": 1.1149162861491628, + "grad_norm": 5.59375, + "learning_rate": 4.4643443561967625e-06, + "loss": 1.62274284, + "memory(GiB)": 111.15, + "step": 43950, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.65653734, + "epoch": 1.1150431253170978, + "grad_norm": 5.4375, + "learning_rate": 4.463301779239366e-06, + "loss": 1.52905922, + "memory(GiB)": 111.15, + "step": 43955, + "train_speed(iter/s)": 0.447663 + }, + { + "acc": 0.67558603, + "epoch": 1.115169964485033, + "grad_norm": 5.25, + "learning_rate": 4.462259225888354e-06, + "loss": 1.56166439, + "memory(GiB)": 111.15, + "step": 43960, + "train_speed(iter/s)": 0.44769 + }, + { + "acc": 0.64364748, + "epoch": 1.115296803652968, + "grad_norm": 5.0625, + "learning_rate": 4.4612166961895805e-06, + "loss": 1.68547668, + "memory(GiB)": 111.15, + "step": 43965, + "train_speed(iter/s)": 0.447717 + }, + { + "acc": 0.65125332, + "epoch": 1.115423642820903, + "grad_norm": 4.8125, + "learning_rate": 4.460174190188905e-06, + "loss": 1.59385376, + "memory(GiB)": 111.15, + "step": 43970, + "train_speed(iter/s)": 0.447743 + }, + { + "acc": 0.64289832, + "epoch": 1.1155504819888382, + "grad_norm": 5.75, + "learning_rate": 4.459131707932177e-06, + "loss": 1.69408722, + "memory(GiB)": 111.15, + "step": 43975, + "train_speed(iter/s)": 0.44777 + }, + { + "acc": 0.6463275, + "epoch": 1.1156773211567732, + "grad_norm": 5.28125, + "learning_rate": 4.458089249465251e-06, + "loss": 1.66346512, + "memory(GiB)": 111.15, + "step": 43980, + "train_speed(iter/s)": 0.447797 + }, + { + "acc": 0.64140282, + "epoch": 1.1158041603247082, + "grad_norm": 4.625, + "learning_rate": 4.45704681483398e-06, + "loss": 1.66770821, + "memory(GiB)": 111.15, + "step": 43985, + "train_speed(iter/s)": 0.447823 + }, + { + "acc": 0.6409235, + "epoch": 1.1159309994926434, + "grad_norm": 5.09375, + "learning_rate": 4.456004404084215e-06, + "loss": 1.63699207, + "memory(GiB)": 111.15, + "step": 43990, + "train_speed(iter/s)": 0.447851 + }, + { + "acc": 0.63896856, + "epoch": 1.1160578386605784, + "grad_norm": 6.8125, + "learning_rate": 4.454962017261803e-06, + "loss": 1.69916878, + "memory(GiB)": 111.15, + "step": 43995, + "train_speed(iter/s)": 0.447877 + }, + { + "acc": 0.64159613, + "epoch": 1.1161846778285134, + "grad_norm": 5.3125, + "learning_rate": 4.453919654412596e-06, + "loss": 1.64852219, + "memory(GiB)": 111.15, + "step": 44000, + "train_speed(iter/s)": 0.447904 + }, + { + "epoch": 1.1161846778285134, + "eval_acc": 0.6433432748027407, + "eval_loss": 1.607741355895996, + "eval_runtime": 115.5934, + "eval_samples_per_second": 55.107, + "eval_steps_per_second": 27.553, + "step": 44000 + }, + { + "acc": 0.65112696, + "epoch": 1.1163115169964486, + "grad_norm": 5.25, + "learning_rate": 4.45287731558244e-06, + "loss": 1.57612791, + "memory(GiB)": 111.15, + "step": 44005, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.65021548, + "epoch": 1.1164383561643836, + "grad_norm": 4.4375, + "learning_rate": 4.451835000817185e-06, + "loss": 1.60348492, + "memory(GiB)": 111.15, + "step": 44010, + "train_speed(iter/s)": 0.44741 + }, + { + "acc": 0.65743942, + "epoch": 1.1165651953323186, + "grad_norm": 5.46875, + "learning_rate": 4.450792710162672e-06, + "loss": 1.63968811, + "memory(GiB)": 111.15, + "step": 44015, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.64624481, + "epoch": 1.1166920345002538, + "grad_norm": 5.46875, + "learning_rate": 4.449750443664747e-06, + "loss": 1.61443634, + "memory(GiB)": 111.15, + "step": 44020, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.65451622, + "epoch": 1.1168188736681888, + "grad_norm": 4.5625, + "learning_rate": 4.448708201369254e-06, + "loss": 1.60247784, + "memory(GiB)": 111.15, + "step": 44025, + "train_speed(iter/s)": 0.447491 + }, + { + "acc": 0.64684162, + "epoch": 1.1169457128361238, + "grad_norm": 5.125, + "learning_rate": 4.4476659833220374e-06, + "loss": 1.63735123, + "memory(GiB)": 111.15, + "step": 44030, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.66011486, + "epoch": 1.1170725520040587, + "grad_norm": 5.53125, + "learning_rate": 4.4466237895689365e-06, + "loss": 1.57280703, + "memory(GiB)": 111.15, + "step": 44035, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.65363779, + "epoch": 1.117199391171994, + "grad_norm": 6.25, + "learning_rate": 4.44558162015579e-06, + "loss": 1.61035843, + "memory(GiB)": 111.15, + "step": 44040, + "train_speed(iter/s)": 0.447572 + }, + { + "acc": 0.64679914, + "epoch": 1.117326230339929, + "grad_norm": 5.90625, + "learning_rate": 4.444539475128441e-06, + "loss": 1.63318138, + "memory(GiB)": 111.15, + "step": 44045, + "train_speed(iter/s)": 0.447599 + }, + { + "acc": 0.64766855, + "epoch": 1.117453069507864, + "grad_norm": 5.03125, + "learning_rate": 4.443497354532726e-06, + "loss": 1.65816517, + "memory(GiB)": 111.15, + "step": 44050, + "train_speed(iter/s)": 0.447626 + }, + { + "acc": 0.65258675, + "epoch": 1.1175799086757991, + "grad_norm": 5.5625, + "learning_rate": 4.442455258414482e-06, + "loss": 1.56705914, + "memory(GiB)": 111.15, + "step": 44055, + "train_speed(iter/s)": 0.447653 + }, + { + "acc": 0.63883114, + "epoch": 1.1177067478437341, + "grad_norm": 5.34375, + "learning_rate": 4.441413186819543e-06, + "loss": 1.6178257, + "memory(GiB)": 111.15, + "step": 44060, + "train_speed(iter/s)": 0.44768 + }, + { + "acc": 0.65384569, + "epoch": 1.1178335870116691, + "grad_norm": 6.625, + "learning_rate": 4.440371139793747e-06, + "loss": 1.5860672, + "memory(GiB)": 111.15, + "step": 44065, + "train_speed(iter/s)": 0.447707 + }, + { + "acc": 0.65505285, + "epoch": 1.1179604261796043, + "grad_norm": 5.03125, + "learning_rate": 4.43932911738293e-06, + "loss": 1.64198551, + "memory(GiB)": 111.15, + "step": 44070, + "train_speed(iter/s)": 0.447734 + }, + { + "acc": 0.64321499, + "epoch": 1.1180872653475393, + "grad_norm": 5.96875, + "learning_rate": 4.438287119632917e-06, + "loss": 1.6269474, + "memory(GiB)": 111.15, + "step": 44075, + "train_speed(iter/s)": 0.447761 + }, + { + "acc": 0.64389811, + "epoch": 1.1182141045154743, + "grad_norm": 9.125, + "learning_rate": 4.4372451465895465e-06, + "loss": 1.61616402, + "memory(GiB)": 111.15, + "step": 44080, + "train_speed(iter/s)": 0.447788 + }, + { + "acc": 0.64391565, + "epoch": 1.1183409436834095, + "grad_norm": 5.4375, + "learning_rate": 4.436203198298645e-06, + "loss": 1.62471066, + "memory(GiB)": 111.15, + "step": 44085, + "train_speed(iter/s)": 0.447815 + }, + { + "acc": 0.64482422, + "epoch": 1.1184677828513445, + "grad_norm": 4.75, + "learning_rate": 4.435161274806049e-06, + "loss": 1.61286068, + "memory(GiB)": 111.15, + "step": 44090, + "train_speed(iter/s)": 0.447842 + }, + { + "acc": 0.63948755, + "epoch": 1.1185946220192795, + "grad_norm": 4.71875, + "learning_rate": 4.4341193761575765e-06, + "loss": 1.67480049, + "memory(GiB)": 111.15, + "step": 44095, + "train_speed(iter/s)": 0.447869 + }, + { + "acc": 0.66657171, + "epoch": 1.1187214611872145, + "grad_norm": 5.9375, + "learning_rate": 4.433077502399063e-06, + "loss": 1.61271782, + "memory(GiB)": 111.15, + "step": 44100, + "train_speed(iter/s)": 0.447896 + }, + { + "epoch": 1.1187214611872145, + "eval_acc": 0.6433202976007205, + "eval_loss": 1.6078840494155884, + "eval_runtime": 113.5781, + "eval_samples_per_second": 56.085, + "eval_steps_per_second": 28.042, + "step": 44100 + }, + { + "acc": 0.65443869, + "epoch": 1.1188483003551497, + "grad_norm": 4.90625, + "learning_rate": 4.43203565357633e-06, + "loss": 1.54661493, + "memory(GiB)": 111.15, + "step": 44105, + "train_speed(iter/s)": 0.447385 + }, + { + "acc": 0.650489, + "epoch": 1.1189751395230847, + "grad_norm": 5.09375, + "learning_rate": 4.430993829735208e-06, + "loss": 1.6056778, + "memory(GiB)": 111.15, + "step": 44110, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.65703754, + "epoch": 1.1191019786910197, + "grad_norm": 5.1875, + "learning_rate": 4.429952030921516e-06, + "loss": 1.52913475, + "memory(GiB)": 111.15, + "step": 44115, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.6673522, + "epoch": 1.119228817858955, + "grad_norm": 6.65625, + "learning_rate": 4.428910257181077e-06, + "loss": 1.61710567, + "memory(GiB)": 111.15, + "step": 44120, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.65801573, + "epoch": 1.1193556570268899, + "grad_norm": 7.6875, + "learning_rate": 4.427868508559717e-06, + "loss": 1.6121624, + "memory(GiB)": 111.15, + "step": 44125, + "train_speed(iter/s)": 0.447491 + }, + { + "acc": 0.67140007, + "epoch": 1.1194824961948249, + "grad_norm": 6.3125, + "learning_rate": 4.426826785103256e-06, + "loss": 1.53208256, + "memory(GiB)": 111.15, + "step": 44130, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.6505619, + "epoch": 1.11960933536276, + "grad_norm": 6.5625, + "learning_rate": 4.425785086857509e-06, + "loss": 1.64225731, + "memory(GiB)": 111.15, + "step": 44135, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.64814262, + "epoch": 1.119736174530695, + "grad_norm": 5.3125, + "learning_rate": 4.424743413868298e-06, + "loss": 1.64004974, + "memory(GiB)": 111.15, + "step": 44140, + "train_speed(iter/s)": 0.44757 + }, + { + "acc": 0.65566292, + "epoch": 1.11986301369863, + "grad_norm": 6.125, + "learning_rate": 4.42370176618144e-06, + "loss": 1.59869289, + "memory(GiB)": 111.15, + "step": 44145, + "train_speed(iter/s)": 0.447597 + }, + { + "acc": 0.63815413, + "epoch": 1.1199898528665653, + "grad_norm": 5.03125, + "learning_rate": 4.422660143842753e-06, + "loss": 1.61365929, + "memory(GiB)": 111.15, + "step": 44150, + "train_speed(iter/s)": 0.447623 + }, + { + "acc": 0.64841948, + "epoch": 1.1201166920345003, + "grad_norm": 6.15625, + "learning_rate": 4.421618546898048e-06, + "loss": 1.668256, + "memory(GiB)": 111.15, + "step": 44155, + "train_speed(iter/s)": 0.44765 + }, + { + "acc": 0.6525094, + "epoch": 1.1202435312024352, + "grad_norm": 5.0, + "learning_rate": 4.420576975393143e-06, + "loss": 1.53784256, + "memory(GiB)": 111.15, + "step": 44160, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.66498985, + "epoch": 1.1203703703703705, + "grad_norm": 5.125, + "learning_rate": 4.4195354293738484e-06, + "loss": 1.55815372, + "memory(GiB)": 111.15, + "step": 44165, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.64175463, + "epoch": 1.1204972095383054, + "grad_norm": 5.34375, + "learning_rate": 4.418493908885979e-06, + "loss": 1.57835846, + "memory(GiB)": 111.15, + "step": 44170, + "train_speed(iter/s)": 0.44773 + }, + { + "acc": 0.65107002, + "epoch": 1.1206240487062404, + "grad_norm": 4.71875, + "learning_rate": 4.417452413975343e-06, + "loss": 1.61611595, + "memory(GiB)": 111.15, + "step": 44175, + "train_speed(iter/s)": 0.447756 + }, + { + "acc": 0.65821552, + "epoch": 1.1207508878741756, + "grad_norm": 4.78125, + "learning_rate": 4.4164109446877514e-06, + "loss": 1.61503448, + "memory(GiB)": 111.15, + "step": 44180, + "train_speed(iter/s)": 0.447782 + }, + { + "acc": 0.65994129, + "epoch": 1.1208777270421106, + "grad_norm": 4.5625, + "learning_rate": 4.41536950106901e-06, + "loss": 1.58942909, + "memory(GiB)": 111.15, + "step": 44185, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.66450405, + "epoch": 1.1210045662100456, + "grad_norm": 6.0625, + "learning_rate": 4.414328083164931e-06, + "loss": 1.5735651, + "memory(GiB)": 111.15, + "step": 44190, + "train_speed(iter/s)": 0.447835 + }, + { + "acc": 0.6425241, + "epoch": 1.1211314053779806, + "grad_norm": 5.1875, + "learning_rate": 4.4132866910213154e-06, + "loss": 1.68819065, + "memory(GiB)": 111.15, + "step": 44195, + "train_speed(iter/s)": 0.447861 + }, + { + "acc": 0.65582895, + "epoch": 1.1212582445459158, + "grad_norm": 5.28125, + "learning_rate": 4.41224532468397e-06, + "loss": 1.52001848, + "memory(GiB)": 111.15, + "step": 44200, + "train_speed(iter/s)": 0.447888 + }, + { + "epoch": 1.1212582445459158, + "eval_acc": 0.6433754428855689, + "eval_loss": 1.6078749895095825, + "eval_runtime": 115.018, + "eval_samples_per_second": 55.383, + "eval_steps_per_second": 27.691, + "step": 44200 + }, + { + "acc": 0.65578642, + "epoch": 1.1213850837138508, + "grad_norm": 5.65625, + "learning_rate": 4.411203984198701e-06, + "loss": 1.64828663, + "memory(GiB)": 111.15, + "step": 44205, + "train_speed(iter/s)": 0.447372 + }, + { + "acc": 0.63984108, + "epoch": 1.1215119228817858, + "grad_norm": 4.28125, + "learning_rate": 4.41016266961131e-06, + "loss": 1.61994762, + "memory(GiB)": 111.15, + "step": 44210, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.64766026, + "epoch": 1.121638762049721, + "grad_norm": 7.59375, + "learning_rate": 4.409121380967597e-06, + "loss": 1.66457253, + "memory(GiB)": 111.15, + "step": 44215, + "train_speed(iter/s)": 0.447425 + }, + { + "acc": 0.66569662, + "epoch": 1.121765601217656, + "grad_norm": 7.625, + "learning_rate": 4.408080118313364e-06, + "loss": 1.60522537, + "memory(GiB)": 111.15, + "step": 44220, + "train_speed(iter/s)": 0.447451 + }, + { + "acc": 0.64866714, + "epoch": 1.121892440385591, + "grad_norm": 6.125, + "learning_rate": 4.40703888169441e-06, + "loss": 1.65662689, + "memory(GiB)": 111.15, + "step": 44225, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.65370255, + "epoch": 1.1220192795535262, + "grad_norm": 5.40625, + "learning_rate": 4.4059976711565355e-06, + "loss": 1.55274525, + "memory(GiB)": 111.15, + "step": 44230, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.64949899, + "epoch": 1.1221461187214612, + "grad_norm": 5.5625, + "learning_rate": 4.404956486745532e-06, + "loss": 1.6032217, + "memory(GiB)": 111.15, + "step": 44235, + "train_speed(iter/s)": 0.447531 + }, + { + "acc": 0.64297504, + "epoch": 1.1222729578893962, + "grad_norm": 6.625, + "learning_rate": 4.403915328507201e-06, + "loss": 1.66941166, + "memory(GiB)": 111.15, + "step": 44240, + "train_speed(iter/s)": 0.447558 + }, + { + "acc": 0.64786777, + "epoch": 1.1223997970573314, + "grad_norm": 5.125, + "learning_rate": 4.4028741964873334e-06, + "loss": 1.60844574, + "memory(GiB)": 111.15, + "step": 44245, + "train_speed(iter/s)": 0.447585 + }, + { + "acc": 0.63982377, + "epoch": 1.1225266362252664, + "grad_norm": 6.6875, + "learning_rate": 4.4018330907317275e-06, + "loss": 1.61652489, + "memory(GiB)": 111.15, + "step": 44250, + "train_speed(iter/s)": 0.447611 + }, + { + "acc": 0.63406534, + "epoch": 1.1226534753932014, + "grad_norm": 5.375, + "learning_rate": 4.400792011286171e-06, + "loss": 1.74306545, + "memory(GiB)": 111.15, + "step": 44255, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.66363192, + "epoch": 1.1227803145611364, + "grad_norm": 5.21875, + "learning_rate": 4.3997509581964566e-06, + "loss": 1.59445553, + "memory(GiB)": 111.15, + "step": 44260, + "train_speed(iter/s)": 0.447664 + }, + { + "acc": 0.6495513, + "epoch": 1.1229071537290716, + "grad_norm": 5.5, + "learning_rate": 4.398709931508376e-06, + "loss": 1.56180058, + "memory(GiB)": 111.15, + "step": 44265, + "train_speed(iter/s)": 0.447691 + }, + { + "acc": 0.65189104, + "epoch": 1.1230339928970066, + "grad_norm": 6.375, + "learning_rate": 4.397668931267718e-06, + "loss": 1.54950695, + "memory(GiB)": 111.15, + "step": 44270, + "train_speed(iter/s)": 0.447717 + }, + { + "acc": 0.65029039, + "epoch": 1.1231608320649416, + "grad_norm": 5.125, + "learning_rate": 4.396627957520269e-06, + "loss": 1.65279121, + "memory(GiB)": 111.15, + "step": 44275, + "train_speed(iter/s)": 0.447743 + }, + { + "acc": 0.634725, + "epoch": 1.1232876712328768, + "grad_norm": 5.5, + "learning_rate": 4.395587010311815e-06, + "loss": 1.65929642, + "memory(GiB)": 111.15, + "step": 44280, + "train_speed(iter/s)": 0.44777 + }, + { + "acc": 0.65903997, + "epoch": 1.1234145104008117, + "grad_norm": 4.875, + "learning_rate": 4.394546089688143e-06, + "loss": 1.60289326, + "memory(GiB)": 111.15, + "step": 44285, + "train_speed(iter/s)": 0.447796 + }, + { + "acc": 0.65372772, + "epoch": 1.1235413495687467, + "grad_norm": 6.46875, + "learning_rate": 4.3935051956950395e-06, + "loss": 1.63487148, + "memory(GiB)": 111.15, + "step": 44290, + "train_speed(iter/s)": 0.447823 + }, + { + "acc": 0.64541187, + "epoch": 1.123668188736682, + "grad_norm": 5.21875, + "learning_rate": 4.3924643283782824e-06, + "loss": 1.63864079, + "memory(GiB)": 111.15, + "step": 44295, + "train_speed(iter/s)": 0.447849 + }, + { + "acc": 0.65878868, + "epoch": 1.123795027904617, + "grad_norm": 6.15625, + "learning_rate": 4.391423487783657e-06, + "loss": 1.60517616, + "memory(GiB)": 111.15, + "step": 44300, + "train_speed(iter/s)": 0.447875 + }, + { + "epoch": 1.123795027904617, + "eval_acc": 0.6433595677278096, + "eval_loss": 1.6079118251800537, + "eval_runtime": 115.0935, + "eval_samples_per_second": 55.346, + "eval_steps_per_second": 27.673, + "step": 44300 + }, + { + "acc": 0.65777192, + "epoch": 1.123921867072552, + "grad_norm": 5.71875, + "learning_rate": 4.3903826739569444e-06, + "loss": 1.60267563, + "memory(GiB)": 111.15, + "step": 44305, + "train_speed(iter/s)": 0.44736 + }, + { + "acc": 0.66067238, + "epoch": 1.1240487062404871, + "grad_norm": 5.875, + "learning_rate": 4.389341886943926e-06, + "loss": 1.59889183, + "memory(GiB)": 111.15, + "step": 44310, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.64479022, + "epoch": 1.1241755454084221, + "grad_norm": 5.46875, + "learning_rate": 4.388301126790374e-06, + "loss": 1.63187466, + "memory(GiB)": 111.15, + "step": 44315, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.6545929, + "epoch": 1.1243023845763571, + "grad_norm": 5.5, + "learning_rate": 4.387260393542071e-06, + "loss": 1.64414196, + "memory(GiB)": 111.15, + "step": 44320, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.64408798, + "epoch": 1.1244292237442923, + "grad_norm": 5.28125, + "learning_rate": 4.38621968724479e-06, + "loss": 1.70518379, + "memory(GiB)": 111.15, + "step": 44325, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.64886532, + "epoch": 1.1245560629122273, + "grad_norm": 5.25, + "learning_rate": 4.385179007944311e-06, + "loss": 1.58664742, + "memory(GiB)": 111.15, + "step": 44330, + "train_speed(iter/s)": 0.447491 + }, + { + "acc": 0.64567008, + "epoch": 1.1246829020801623, + "grad_norm": 7.34375, + "learning_rate": 4.384138355686402e-06, + "loss": 1.67950497, + "memory(GiB)": 111.15, + "step": 44335, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.63414545, + "epoch": 1.1248097412480975, + "grad_norm": 5.21875, + "learning_rate": 4.383097730516837e-06, + "loss": 1.66204834, + "memory(GiB)": 111.15, + "step": 44340, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.66399784, + "epoch": 1.1249365804160325, + "grad_norm": 5.5, + "learning_rate": 4.382057132481389e-06, + "loss": 1.58795309, + "memory(GiB)": 111.15, + "step": 44345, + "train_speed(iter/s)": 0.447571 + }, + { + "acc": 0.64943981, + "epoch": 1.1250634195839675, + "grad_norm": 6.3125, + "learning_rate": 4.381016561625829e-06, + "loss": 1.55676098, + "memory(GiB)": 111.15, + "step": 44350, + "train_speed(iter/s)": 0.447597 + }, + { + "acc": 0.64191642, + "epoch": 1.1251902587519025, + "grad_norm": 5.84375, + "learning_rate": 4.379976017995922e-06, + "loss": 1.70260162, + "memory(GiB)": 111.15, + "step": 44355, + "train_speed(iter/s)": 0.447623 + }, + { + "acc": 0.6549799, + "epoch": 1.1253170979198377, + "grad_norm": 6.15625, + "learning_rate": 4.378935501637438e-06, + "loss": 1.65214729, + "memory(GiB)": 111.15, + "step": 44360, + "train_speed(iter/s)": 0.44765 + }, + { + "acc": 0.65513363, + "epoch": 1.1254439370877727, + "grad_norm": 5.5625, + "learning_rate": 4.377895012596144e-06, + "loss": 1.59331493, + "memory(GiB)": 111.15, + "step": 44365, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.64305301, + "epoch": 1.1255707762557077, + "grad_norm": 5.40625, + "learning_rate": 4.376854550917805e-06, + "loss": 1.67815609, + "memory(GiB)": 111.15, + "step": 44370, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.64244843, + "epoch": 1.1256976154236429, + "grad_norm": 5.40625, + "learning_rate": 4.375814116648184e-06, + "loss": 1.67063408, + "memory(GiB)": 111.15, + "step": 44375, + "train_speed(iter/s)": 0.44773 + }, + { + "acc": 0.65648556, + "epoch": 1.1258244545915779, + "grad_norm": 4.6875, + "learning_rate": 4.374773709833045e-06, + "loss": 1.6102253, + "memory(GiB)": 111.15, + "step": 44380, + "train_speed(iter/s)": 0.447756 + }, + { + "acc": 0.66167498, + "epoch": 1.1259512937595129, + "grad_norm": 7.03125, + "learning_rate": 4.37373333051815e-06, + "loss": 1.6173193, + "memory(GiB)": 111.15, + "step": 44385, + "train_speed(iter/s)": 0.447782 + }, + { + "acc": 0.63655729, + "epoch": 1.126078132927448, + "grad_norm": 5.34375, + "learning_rate": 4.37269297874926e-06, + "loss": 1.70184288, + "memory(GiB)": 111.15, + "step": 44390, + "train_speed(iter/s)": 0.447809 + }, + { + "acc": 0.64288807, + "epoch": 1.126204972095383, + "grad_norm": 5.28125, + "learning_rate": 4.371652654572134e-06, + "loss": 1.70077591, + "memory(GiB)": 111.15, + "step": 44395, + "train_speed(iter/s)": 0.447835 + }, + { + "acc": 0.65500755, + "epoch": 1.126331811263318, + "grad_norm": 5.40625, + "learning_rate": 4.370612358032529e-06, + "loss": 1.57173691, + "memory(GiB)": 111.15, + "step": 44400, + "train_speed(iter/s)": 0.447861 + }, + { + "epoch": 1.126331811263318, + "eval_acc": 0.6433712652124743, + "eval_loss": 1.6079192161560059, + "eval_runtime": 114.13, + "eval_samples_per_second": 55.814, + "eval_steps_per_second": 27.907, + "step": 44400 + }, + { + "acc": 0.65262108, + "epoch": 1.1264586504312533, + "grad_norm": 6.5625, + "learning_rate": 4.369572089176201e-06, + "loss": 1.56066046, + "memory(GiB)": 111.15, + "step": 44405, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.64441929, + "epoch": 1.1265854895991883, + "grad_norm": 6.03125, + "learning_rate": 4.3685318480489095e-06, + "loss": 1.62346668, + "memory(GiB)": 111.15, + "step": 44410, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.66127205, + "epoch": 1.1267123287671232, + "grad_norm": 5.71875, + "learning_rate": 4.367491634696405e-06, + "loss": 1.56290569, + "memory(GiB)": 111.15, + "step": 44415, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.6432724, + "epoch": 1.1268391679350582, + "grad_norm": 6.28125, + "learning_rate": 4.366451449164442e-06, + "loss": 1.6180481, + "memory(GiB)": 111.15, + "step": 44420, + "train_speed(iter/s)": 0.447431 + }, + { + "acc": 0.65104713, + "epoch": 1.1269660071029934, + "grad_norm": 4.84375, + "learning_rate": 4.365411291498774e-06, + "loss": 1.63149815, + "memory(GiB)": 111.15, + "step": 44425, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.64610243, + "epoch": 1.1270928462709284, + "grad_norm": 5.6875, + "learning_rate": 4.364371161745151e-06, + "loss": 1.68140335, + "memory(GiB)": 111.15, + "step": 44430, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.6367178, + "epoch": 1.1272196854388636, + "grad_norm": 5.1875, + "learning_rate": 4.363331059949321e-06, + "loss": 1.64778919, + "memory(GiB)": 111.15, + "step": 44435, + "train_speed(iter/s)": 0.447511 + }, + { + "acc": 0.65939946, + "epoch": 1.1273465246067986, + "grad_norm": 4.9375, + "learning_rate": 4.362290986157034e-06, + "loss": 1.61843014, + "memory(GiB)": 111.15, + "step": 44440, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.64637117, + "epoch": 1.1274733637747336, + "grad_norm": 5.40625, + "learning_rate": 4.361250940414036e-06, + "loss": 1.63354836, + "memory(GiB)": 111.15, + "step": 44445, + "train_speed(iter/s)": 0.447564 + }, + { + "acc": 0.64943676, + "epoch": 1.1276002029426686, + "grad_norm": 5.96875, + "learning_rate": 4.360210922766076e-06, + "loss": 1.5989336, + "memory(GiB)": 111.15, + "step": 44450, + "train_speed(iter/s)": 0.44759 + }, + { + "acc": 0.62589984, + "epoch": 1.1277270421106038, + "grad_norm": 5.375, + "learning_rate": 4.359170933258893e-06, + "loss": 1.70158501, + "memory(GiB)": 111.15, + "step": 44455, + "train_speed(iter/s)": 0.447617 + }, + { + "acc": 0.66030955, + "epoch": 1.1278538812785388, + "grad_norm": 5.15625, + "learning_rate": 4.358130971938235e-06, + "loss": 1.59031773, + "memory(GiB)": 111.15, + "step": 44460, + "train_speed(iter/s)": 0.447643 + }, + { + "acc": 0.64051161, + "epoch": 1.1279807204464738, + "grad_norm": 5.125, + "learning_rate": 4.357091038849841e-06, + "loss": 1.66350937, + "memory(GiB)": 111.15, + "step": 44465, + "train_speed(iter/s)": 0.44767 + }, + { + "acc": 0.65867968, + "epoch": 1.128107559614409, + "grad_norm": 5.46875, + "learning_rate": 4.356051134039455e-06, + "loss": 1.6016777, + "memory(GiB)": 111.15, + "step": 44470, + "train_speed(iter/s)": 0.447696 + }, + { + "acc": 0.66057205, + "epoch": 1.128234398782344, + "grad_norm": 4.875, + "learning_rate": 4.3550112575528155e-06, + "loss": 1.59703064, + "memory(GiB)": 111.15, + "step": 44475, + "train_speed(iter/s)": 0.447723 + }, + { + "acc": 0.66386662, + "epoch": 1.128361237950279, + "grad_norm": 4.9375, + "learning_rate": 4.353971409435659e-06, + "loss": 1.52899818, + "memory(GiB)": 111.15, + "step": 44480, + "train_speed(iter/s)": 0.447749 + }, + { + "acc": 0.65949125, + "epoch": 1.1284880771182142, + "grad_norm": 6.21875, + "learning_rate": 4.352931589733725e-06, + "loss": 1.57208672, + "memory(GiB)": 111.15, + "step": 44485, + "train_speed(iter/s)": 0.447776 + }, + { + "acc": 0.6396142, + "epoch": 1.1286149162861492, + "grad_norm": 7.25, + "learning_rate": 4.35189179849275e-06, + "loss": 1.66934242, + "memory(GiB)": 111.15, + "step": 44490, + "train_speed(iter/s)": 0.447803 + }, + { + "acc": 0.65546818, + "epoch": 1.1287417554540842, + "grad_norm": 5.09375, + "learning_rate": 4.350852035758466e-06, + "loss": 1.70654545, + "memory(GiB)": 111.15, + "step": 44495, + "train_speed(iter/s)": 0.447829 + }, + { + "acc": 0.63755651, + "epoch": 1.1288685946220194, + "grad_norm": 4.4375, + "learning_rate": 4.3498123015766066e-06, + "loss": 1.65399857, + "memory(GiB)": 111.15, + "step": 44500, + "train_speed(iter/s)": 0.447856 + }, + { + "epoch": 1.1288685946220194, + "eval_acc": 0.64337460735095, + "eval_loss": 1.6078941822052002, + "eval_runtime": 114.7377, + "eval_samples_per_second": 55.518, + "eval_steps_per_second": 27.759, + "step": 44500 + }, + { + "acc": 0.66181526, + "epoch": 1.1289954337899544, + "grad_norm": 6.71875, + "learning_rate": 4.348772595992906e-06, + "loss": 1.59675512, + "memory(GiB)": 111.15, + "step": 44505, + "train_speed(iter/s)": 0.447345 + }, + { + "acc": 0.63864489, + "epoch": 1.1291222729578894, + "grad_norm": 5.46875, + "learning_rate": 4.347732919053096e-06, + "loss": 1.60583763, + "memory(GiB)": 111.15, + "step": 44510, + "train_speed(iter/s)": 0.447371 + }, + { + "acc": 0.64036913, + "epoch": 1.1292491121258244, + "grad_norm": 5.1875, + "learning_rate": 4.346693270802902e-06, + "loss": 1.66709995, + "memory(GiB)": 111.15, + "step": 44515, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.65434275, + "epoch": 1.1293759512937596, + "grad_norm": 5.0625, + "learning_rate": 4.345653651288055e-06, + "loss": 1.57456503, + "memory(GiB)": 111.15, + "step": 44520, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.64219542, + "epoch": 1.1295027904616946, + "grad_norm": 5.84375, + "learning_rate": 4.344614060554281e-06, + "loss": 1.59333324, + "memory(GiB)": 111.15, + "step": 44525, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.65165234, + "epoch": 1.1296296296296295, + "grad_norm": 5.3125, + "learning_rate": 4.343574498647311e-06, + "loss": 1.62638092, + "memory(GiB)": 111.15, + "step": 44530, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.65469542, + "epoch": 1.1297564687975648, + "grad_norm": 5.28125, + "learning_rate": 4.342534965612861e-06, + "loss": 1.57522259, + "memory(GiB)": 111.15, + "step": 44535, + "train_speed(iter/s)": 0.447502 + }, + { + "acc": 0.64744778, + "epoch": 1.1298833079654997, + "grad_norm": 5.65625, + "learning_rate": 4.34149546149666e-06, + "loss": 1.59292679, + "memory(GiB)": 111.15, + "step": 44540, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.65541029, + "epoch": 1.1300101471334347, + "grad_norm": 5.1875, + "learning_rate": 4.340455986344428e-06, + "loss": 1.67056522, + "memory(GiB)": 111.15, + "step": 44545, + "train_speed(iter/s)": 0.447554 + }, + { + "acc": 0.65752726, + "epoch": 1.13013698630137, + "grad_norm": 5.4375, + "learning_rate": 4.3394165402018875e-06, + "loss": 1.61908607, + "memory(GiB)": 111.15, + "step": 44550, + "train_speed(iter/s)": 0.447581 + }, + { + "acc": 0.64977808, + "epoch": 1.130263825469305, + "grad_norm": 5.40625, + "learning_rate": 4.338377123114757e-06, + "loss": 1.6142868, + "memory(GiB)": 111.15, + "step": 44555, + "train_speed(iter/s)": 0.447607 + }, + { + "acc": 0.6419672, + "epoch": 1.13039066463724, + "grad_norm": 4.875, + "learning_rate": 4.337337735128752e-06, + "loss": 1.67025986, + "memory(GiB)": 111.15, + "step": 44560, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.65293622, + "epoch": 1.1305175038051751, + "grad_norm": 4.15625, + "learning_rate": 4.336298376289594e-06, + "loss": 1.6309948, + "memory(GiB)": 111.15, + "step": 44565, + "train_speed(iter/s)": 0.44766 + }, + { + "acc": 0.63311815, + "epoch": 1.1306443429731101, + "grad_norm": 5.96875, + "learning_rate": 4.335259046642998e-06, + "loss": 1.68553944, + "memory(GiB)": 111.15, + "step": 44570, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.6459197, + "epoch": 1.130771182141045, + "grad_norm": 4.59375, + "learning_rate": 4.334219746234675e-06, + "loss": 1.63104992, + "memory(GiB)": 111.15, + "step": 44575, + "train_speed(iter/s)": 0.447713 + }, + { + "acc": 0.64463377, + "epoch": 1.13089802130898, + "grad_norm": 5.21875, + "learning_rate": 4.3331804751103395e-06, + "loss": 1.61752815, + "memory(GiB)": 111.15, + "step": 44580, + "train_speed(iter/s)": 0.44774 + }, + { + "acc": 0.6461257, + "epoch": 1.1310248604769153, + "grad_norm": 6.40625, + "learning_rate": 4.332141233315705e-06, + "loss": 1.62877007, + "memory(GiB)": 111.15, + "step": 44585, + "train_speed(iter/s)": 0.447766 + }, + { + "acc": 0.65750513, + "epoch": 1.1311516996448503, + "grad_norm": 5.0, + "learning_rate": 4.331102020896482e-06, + "loss": 1.64890251, + "memory(GiB)": 111.15, + "step": 44590, + "train_speed(iter/s)": 0.447793 + }, + { + "acc": 0.66294122, + "epoch": 1.1312785388127855, + "grad_norm": 6.28125, + "learning_rate": 4.330062837898376e-06, + "loss": 1.55516186, + "memory(GiB)": 111.15, + "step": 44595, + "train_speed(iter/s)": 0.44782 + }, + { + "acc": 0.64195776, + "epoch": 1.1314053779807205, + "grad_norm": 5.3125, + "learning_rate": 4.3290236843670985e-06, + "loss": 1.65180855, + "memory(GiB)": 111.15, + "step": 44600, + "train_speed(iter/s)": 0.447847 + }, + { + "epoch": 1.1314053779807205, + "eval_acc": 0.6433232219718867, + "eval_loss": 1.607893705368042, + "eval_runtime": 113.6202, + "eval_samples_per_second": 56.064, + "eval_steps_per_second": 28.032, + "step": 44600 + }, + { + "acc": 0.65265818, + "epoch": 1.1315322171486555, + "grad_norm": 5.40625, + "learning_rate": 4.327984560348354e-06, + "loss": 1.64216805, + "memory(GiB)": 111.15, + "step": 44605, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.66440697, + "epoch": 1.1316590563165905, + "grad_norm": 5.0, + "learning_rate": 4.3269454658878516e-06, + "loss": 1.60079765, + "memory(GiB)": 111.15, + "step": 44610, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.66019144, + "epoch": 1.1317858954845257, + "grad_norm": 4.84375, + "learning_rate": 4.325906401031291e-06, + "loss": 1.63402786, + "memory(GiB)": 111.15, + "step": 44615, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.65743327, + "epoch": 1.1319127346524607, + "grad_norm": 5.65625, + "learning_rate": 4.324867365824376e-06, + "loss": 1.58513508, + "memory(GiB)": 111.15, + "step": 44620, + "train_speed(iter/s)": 0.447421 + }, + { + "acc": 0.65438323, + "epoch": 1.1320395738203957, + "grad_norm": 5.40625, + "learning_rate": 4.323828360312809e-06, + "loss": 1.63268127, + "memory(GiB)": 111.15, + "step": 44625, + "train_speed(iter/s)": 0.447448 + }, + { + "acc": 0.65129538, + "epoch": 1.1321664129883309, + "grad_norm": 5.53125, + "learning_rate": 4.32278938454229e-06, + "loss": 1.58695202, + "memory(GiB)": 111.15, + "step": 44630, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.656599, + "epoch": 1.1322932521562659, + "grad_norm": 6.21875, + "learning_rate": 4.321750438558517e-06, + "loss": 1.54436321, + "memory(GiB)": 111.15, + "step": 44635, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.64657326, + "epoch": 1.1324200913242009, + "grad_norm": 5.03125, + "learning_rate": 4.3207115224071874e-06, + "loss": 1.54409609, + "memory(GiB)": 111.15, + "step": 44640, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.63263578, + "epoch": 1.132546930492136, + "grad_norm": 5.25, + "learning_rate": 4.319672636133998e-06, + "loss": 1.68333588, + "memory(GiB)": 111.15, + "step": 44645, + "train_speed(iter/s)": 0.447553 + }, + { + "acc": 0.6551465, + "epoch": 1.132673769660071, + "grad_norm": 5.6875, + "learning_rate": 4.318633779784646e-06, + "loss": 1.65211239, + "memory(GiB)": 111.15, + "step": 44650, + "train_speed(iter/s)": 0.44758 + }, + { + "acc": 0.64118099, + "epoch": 1.132800608828006, + "grad_norm": 5.34375, + "learning_rate": 4.317594953404818e-06, + "loss": 1.66570492, + "memory(GiB)": 111.15, + "step": 44655, + "train_speed(iter/s)": 0.447606 + }, + { + "acc": 0.65399265, + "epoch": 1.1329274479959413, + "grad_norm": 4.46875, + "learning_rate": 4.316556157040213e-06, + "loss": 1.59479761, + "memory(GiB)": 111.15, + "step": 44660, + "train_speed(iter/s)": 0.447633 + }, + { + "acc": 0.63089466, + "epoch": 1.1330542871638762, + "grad_norm": 5.28125, + "learning_rate": 4.315517390736519e-06, + "loss": 1.65857143, + "memory(GiB)": 111.15, + "step": 44665, + "train_speed(iter/s)": 0.44766 + }, + { + "acc": 0.64124131, + "epoch": 1.1331811263318112, + "grad_norm": 4.34375, + "learning_rate": 4.314478654539429e-06, + "loss": 1.60869389, + "memory(GiB)": 111.15, + "step": 44670, + "train_speed(iter/s)": 0.447661 + }, + { + "acc": 0.64692812, + "epoch": 1.1333079654997462, + "grad_norm": 4.875, + "learning_rate": 4.313439948494625e-06, + "loss": 1.64736538, + "memory(GiB)": 111.15, + "step": 44675, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.65326467, + "epoch": 1.1334348046676814, + "grad_norm": 6.53125, + "learning_rate": 4.312401272647799e-06, + "loss": 1.59976978, + "memory(GiB)": 111.15, + "step": 44680, + "train_speed(iter/s)": 0.447714 + }, + { + "acc": 0.64872317, + "epoch": 1.1335616438356164, + "grad_norm": 5.125, + "learning_rate": 4.311362627044633e-06, + "loss": 1.60678692, + "memory(GiB)": 111.15, + "step": 44685, + "train_speed(iter/s)": 0.44774 + }, + { + "acc": 0.65031595, + "epoch": 1.1336884830035514, + "grad_norm": 5.71875, + "learning_rate": 4.310324011730816e-06, + "loss": 1.62481575, + "memory(GiB)": 111.15, + "step": 44690, + "train_speed(iter/s)": 0.447766 + }, + { + "acc": 0.64237304, + "epoch": 1.1338153221714866, + "grad_norm": 5.5, + "learning_rate": 4.309285426752027e-06, + "loss": 1.66444435, + "memory(GiB)": 111.15, + "step": 44695, + "train_speed(iter/s)": 0.447793 + }, + { + "acc": 0.65581474, + "epoch": 1.1339421613394216, + "grad_norm": 5.375, + "learning_rate": 4.308246872153947e-06, + "loss": 1.6261097, + "memory(GiB)": 111.15, + "step": 44700, + "train_speed(iter/s)": 0.447819 + }, + { + "epoch": 1.1339421613394216, + "eval_acc": 0.6433420215008123, + "eval_loss": 1.6078213453292847, + "eval_runtime": 115.0178, + "eval_samples_per_second": 55.383, + "eval_steps_per_second": 27.691, + "step": 44700 + }, + { + "acc": 0.65990391, + "epoch": 1.1340690005073566, + "grad_norm": 4.90625, + "learning_rate": 4.307208347982259e-06, + "loss": 1.52997751, + "memory(GiB)": 111.15, + "step": 44705, + "train_speed(iter/s)": 0.447309 + }, + { + "acc": 0.66295257, + "epoch": 1.1341958396752918, + "grad_norm": 5.5625, + "learning_rate": 4.306169854282643e-06, + "loss": 1.55549612, + "memory(GiB)": 111.15, + "step": 44710, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.65863123, + "epoch": 1.1343226788432268, + "grad_norm": 6.1875, + "learning_rate": 4.305131391100773e-06, + "loss": 1.6091053, + "memory(GiB)": 111.15, + "step": 44715, + "train_speed(iter/s)": 0.447362 + }, + { + "acc": 0.66823378, + "epoch": 1.1344495180111618, + "grad_norm": 4.90625, + "learning_rate": 4.304092958482325e-06, + "loss": 1.57553587, + "memory(GiB)": 111.15, + "step": 44720, + "train_speed(iter/s)": 0.447388 + }, + { + "acc": 0.64841604, + "epoch": 1.134576357179097, + "grad_norm": 7.78125, + "learning_rate": 4.303054556472978e-06, + "loss": 1.60884628, + "memory(GiB)": 111.15, + "step": 44725, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.65681896, + "epoch": 1.134703196347032, + "grad_norm": 5.15625, + "learning_rate": 4.3020161851184036e-06, + "loss": 1.63113213, + "memory(GiB)": 111.15, + "step": 44730, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.64134669, + "epoch": 1.134830035514967, + "grad_norm": 5.84375, + "learning_rate": 4.300977844464273e-06, + "loss": 1.61652412, + "memory(GiB)": 111.15, + "step": 44735, + "train_speed(iter/s)": 0.447468 + }, + { + "acc": 0.65704737, + "epoch": 1.134956874682902, + "grad_norm": 5.65625, + "learning_rate": 4.2999395345562564e-06, + "loss": 1.57149048, + "memory(GiB)": 111.15, + "step": 44740, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.63782835, + "epoch": 1.1350837138508372, + "grad_norm": 7.21875, + "learning_rate": 4.298901255440025e-06, + "loss": 1.64432793, + "memory(GiB)": 111.15, + "step": 44745, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.64547672, + "epoch": 1.1352105530187722, + "grad_norm": 6.90625, + "learning_rate": 4.297863007161249e-06, + "loss": 1.64089737, + "memory(GiB)": 111.15, + "step": 44750, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.64888906, + "epoch": 1.1353373921867074, + "grad_norm": 4.625, + "learning_rate": 4.29682478976559e-06, + "loss": 1.56311178, + "memory(GiB)": 111.15, + "step": 44755, + "train_speed(iter/s)": 0.447574 + }, + { + "acc": 0.65365257, + "epoch": 1.1354642313546424, + "grad_norm": 5.71875, + "learning_rate": 4.295786603298717e-06, + "loss": 1.58876095, + "memory(GiB)": 111.15, + "step": 44760, + "train_speed(iter/s)": 0.447601 + }, + { + "acc": 0.65111709, + "epoch": 1.1355910705225774, + "grad_norm": 4.75, + "learning_rate": 4.294748447806293e-06, + "loss": 1.60851212, + "memory(GiB)": 111.15, + "step": 44765, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.65492334, + "epoch": 1.1357179096905123, + "grad_norm": 5.90625, + "learning_rate": 4.293710323333983e-06, + "loss": 1.60140438, + "memory(GiB)": 111.15, + "step": 44770, + "train_speed(iter/s)": 0.447654 + }, + { + "acc": 0.65428615, + "epoch": 1.1358447488584476, + "grad_norm": 5.4375, + "learning_rate": 4.292672229927445e-06, + "loss": 1.60452995, + "memory(GiB)": 111.15, + "step": 44775, + "train_speed(iter/s)": 0.44768 + }, + { + "acc": 0.66485419, + "epoch": 1.1359715880263825, + "grad_norm": 5.21875, + "learning_rate": 4.2916341676323386e-06, + "loss": 1.59235992, + "memory(GiB)": 111.15, + "step": 44780, + "train_speed(iter/s)": 0.447707 + }, + { + "acc": 0.66126938, + "epoch": 1.1360984271943175, + "grad_norm": 5.6875, + "learning_rate": 4.290596136494326e-06, + "loss": 1.5681448, + "memory(GiB)": 111.15, + "step": 44785, + "train_speed(iter/s)": 0.447733 + }, + { + "acc": 0.64210567, + "epoch": 1.1362252663622527, + "grad_norm": 6.8125, + "learning_rate": 4.289558136559063e-06, + "loss": 1.69922714, + "memory(GiB)": 111.15, + "step": 44790, + "train_speed(iter/s)": 0.44776 + }, + { + "acc": 0.64447508, + "epoch": 1.1363521055301877, + "grad_norm": 5.125, + "learning_rate": 4.288520167872203e-06, + "loss": 1.58402319, + "memory(GiB)": 111.15, + "step": 44795, + "train_speed(iter/s)": 0.447787 + }, + { + "acc": 0.63325844, + "epoch": 1.1364789446981227, + "grad_norm": 5.09375, + "learning_rate": 4.287482230479404e-06, + "loss": 1.70664062, + "memory(GiB)": 111.15, + "step": 44800, + "train_speed(iter/s)": 0.447813 + }, + { + "epoch": 1.1364789446981227, + "eval_acc": 0.6433457814065975, + "eval_loss": 1.6078972816467285, + "eval_runtime": 113.862, + "eval_samples_per_second": 55.945, + "eval_steps_per_second": 27.972, + "step": 44800 + }, + { + "acc": 0.63784957, + "epoch": 1.136605783866058, + "grad_norm": 6.03125, + "learning_rate": 4.286444324426318e-06, + "loss": 1.66635513, + "memory(GiB)": 111.15, + "step": 44805, + "train_speed(iter/s)": 0.44731 + }, + { + "acc": 0.65475955, + "epoch": 1.136732623033993, + "grad_norm": 4.9375, + "learning_rate": 4.2854064497585964e-06, + "loss": 1.60457191, + "memory(GiB)": 111.15, + "step": 44810, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.64360304, + "epoch": 1.136859462201928, + "grad_norm": 5.1875, + "learning_rate": 4.284368606521888e-06, + "loss": 1.66360283, + "memory(GiB)": 111.15, + "step": 44815, + "train_speed(iter/s)": 0.447362 + }, + { + "acc": 0.67575889, + "epoch": 1.1369863013698631, + "grad_norm": 4.84375, + "learning_rate": 4.283330794761845e-06, + "loss": 1.60619316, + "memory(GiB)": 111.15, + "step": 44820, + "train_speed(iter/s)": 0.447388 + }, + { + "acc": 0.65168228, + "epoch": 1.137113140537798, + "grad_norm": 4.84375, + "learning_rate": 4.282293014524112e-06, + "loss": 1.59728832, + "memory(GiB)": 111.15, + "step": 44825, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.64944725, + "epoch": 1.137239979705733, + "grad_norm": 4.6875, + "learning_rate": 4.281255265854338e-06, + "loss": 1.61133175, + "memory(GiB)": 111.15, + "step": 44830, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.65651388, + "epoch": 1.137366818873668, + "grad_norm": 4.71875, + "learning_rate": 4.280217548798166e-06, + "loss": 1.57524509, + "memory(GiB)": 111.15, + "step": 44835, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.64412088, + "epoch": 1.1374936580416033, + "grad_norm": 6.125, + "learning_rate": 4.279179863401239e-06, + "loss": 1.64911118, + "memory(GiB)": 111.15, + "step": 44840, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.66259184, + "epoch": 1.1376204972095383, + "grad_norm": 6.34375, + "learning_rate": 4.278142209709199e-06, + "loss": 1.56671677, + "memory(GiB)": 111.15, + "step": 44845, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.65551834, + "epoch": 1.1377473363774733, + "grad_norm": 5.84375, + "learning_rate": 4.277104587767691e-06, + "loss": 1.65050697, + "memory(GiB)": 111.15, + "step": 44850, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.64185448, + "epoch": 1.1378741755454085, + "grad_norm": 7.71875, + "learning_rate": 4.276066997622348e-06, + "loss": 1.64568253, + "memory(GiB)": 111.15, + "step": 44855, + "train_speed(iter/s)": 0.447573 + }, + { + "acc": 0.66512671, + "epoch": 1.1380010147133435, + "grad_norm": 6.15625, + "learning_rate": 4.27502943931881e-06, + "loss": 1.56770077, + "memory(GiB)": 111.15, + "step": 44860, + "train_speed(iter/s)": 0.4476 + }, + { + "acc": 0.64297218, + "epoch": 1.1381278538812785, + "grad_norm": 4.875, + "learning_rate": 4.273991912902716e-06, + "loss": 1.64431419, + "memory(GiB)": 111.15, + "step": 44865, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.64143085, + "epoch": 1.1382546930492137, + "grad_norm": 5.3125, + "learning_rate": 4.272954418419699e-06, + "loss": 1.63828869, + "memory(GiB)": 111.15, + "step": 44870, + "train_speed(iter/s)": 0.447653 + }, + { + "acc": 0.65571337, + "epoch": 1.1383815322171487, + "grad_norm": 5.40625, + "learning_rate": 4.2719169559153905e-06, + "loss": 1.57127457, + "memory(GiB)": 111.15, + "step": 44875, + "train_speed(iter/s)": 0.44768 + }, + { + "acc": 0.66970882, + "epoch": 1.1385083713850837, + "grad_norm": 4.90625, + "learning_rate": 4.270879525435426e-06, + "loss": 1.57865715, + "memory(GiB)": 111.15, + "step": 44880, + "train_speed(iter/s)": 0.447706 + }, + { + "acc": 0.65452795, + "epoch": 1.1386352105530189, + "grad_norm": 4.9375, + "learning_rate": 4.269842127025435e-06, + "loss": 1.62263565, + "memory(GiB)": 111.15, + "step": 44885, + "train_speed(iter/s)": 0.447732 + }, + { + "acc": 0.64971747, + "epoch": 1.1387620497209539, + "grad_norm": 5.34375, + "learning_rate": 4.2688047607310504e-06, + "loss": 1.60528603, + "memory(GiB)": 111.15, + "step": 44890, + "train_speed(iter/s)": 0.447759 + }, + { + "acc": 0.63943243, + "epoch": 1.1388888888888888, + "grad_norm": 5.5625, + "learning_rate": 4.267767426597893e-06, + "loss": 1.67741432, + "memory(GiB)": 111.15, + "step": 44895, + "train_speed(iter/s)": 0.447785 + }, + { + "acc": 0.65345116, + "epoch": 1.1390157280568238, + "grad_norm": 4.03125, + "learning_rate": 4.266730124671594e-06, + "loss": 1.58672047, + "memory(GiB)": 111.15, + "step": 44900, + "train_speed(iter/s)": 0.447812 + }, + { + "epoch": 1.1390157280568238, + "eval_acc": 0.6433984200875891, + "eval_loss": 1.6078051328659058, + "eval_runtime": 114.6062, + "eval_samples_per_second": 55.582, + "eval_steps_per_second": 27.791, + "step": 44900 + }, + { + "acc": 0.65645156, + "epoch": 1.139142567224759, + "grad_norm": 5.34375, + "learning_rate": 4.265692854997778e-06, + "loss": 1.60282154, + "memory(GiB)": 111.15, + "step": 44905, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.66472616, + "epoch": 1.139269406392694, + "grad_norm": 5.0625, + "learning_rate": 4.2646556176220714e-06, + "loss": 1.58359585, + "memory(GiB)": 111.15, + "step": 44910, + "train_speed(iter/s)": 0.447333 + }, + { + "acc": 0.65980868, + "epoch": 1.1393962455606292, + "grad_norm": 5.03125, + "learning_rate": 4.263618412590092e-06, + "loss": 1.56915112, + "memory(GiB)": 111.15, + "step": 44915, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.63603125, + "epoch": 1.1395230847285642, + "grad_norm": 5.21875, + "learning_rate": 4.2625812399474604e-06, + "loss": 1.63994274, + "memory(GiB)": 111.15, + "step": 44920, + "train_speed(iter/s)": 0.447385 + }, + { + "acc": 0.6582839, + "epoch": 1.1396499238964992, + "grad_norm": 4.9375, + "learning_rate": 4.2615440997398e-06, + "loss": 1.64318962, + "memory(GiB)": 111.15, + "step": 44925, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.65905972, + "epoch": 1.1397767630644342, + "grad_norm": 5.90625, + "learning_rate": 4.2605069920127284e-06, + "loss": 1.56430721, + "memory(GiB)": 111.15, + "step": 44930, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.66130757, + "epoch": 1.1399036022323694, + "grad_norm": 5.65625, + "learning_rate": 4.25946991681186e-06, + "loss": 1.60133877, + "memory(GiB)": 111.15, + "step": 44935, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.64481907, + "epoch": 1.1400304414003044, + "grad_norm": 5.3125, + "learning_rate": 4.258432874182809e-06, + "loss": 1.65277824, + "memory(GiB)": 111.15, + "step": 44940, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.64446163, + "epoch": 1.1401572805682394, + "grad_norm": 5.125, + "learning_rate": 4.2573958641711925e-06, + "loss": 1.65079155, + "memory(GiB)": 111.15, + "step": 44945, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.65792527, + "epoch": 1.1402841197361746, + "grad_norm": 4.625, + "learning_rate": 4.256358886822622e-06, + "loss": 1.55748215, + "memory(GiB)": 111.15, + "step": 44950, + "train_speed(iter/s)": 0.447542 + }, + { + "acc": 0.6582077, + "epoch": 1.1404109589041096, + "grad_norm": 6.40625, + "learning_rate": 4.255321942182707e-06, + "loss": 1.57377377, + "memory(GiB)": 111.15, + "step": 44955, + "train_speed(iter/s)": 0.447569 + }, + { + "acc": 0.64529648, + "epoch": 1.1405377980720446, + "grad_norm": 5.15625, + "learning_rate": 4.254285030297058e-06, + "loss": 1.64692459, + "memory(GiB)": 111.15, + "step": 44960, + "train_speed(iter/s)": 0.447594 + }, + { + "acc": 0.65990734, + "epoch": 1.1406646372399798, + "grad_norm": 4.8125, + "learning_rate": 4.2532481512112814e-06, + "loss": 1.64500065, + "memory(GiB)": 111.15, + "step": 44965, + "train_speed(iter/s)": 0.447621 + }, + { + "acc": 0.64051323, + "epoch": 1.1407914764079148, + "grad_norm": 6.15625, + "learning_rate": 4.252211304970988e-06, + "loss": 1.6603941, + "memory(GiB)": 111.15, + "step": 44970, + "train_speed(iter/s)": 0.447647 + }, + { + "acc": 0.64884377, + "epoch": 1.1409183155758498, + "grad_norm": 6.3125, + "learning_rate": 4.251174491621778e-06, + "loss": 1.60946922, + "memory(GiB)": 111.15, + "step": 44975, + "train_speed(iter/s)": 0.447673 + }, + { + "acc": 0.6506639, + "epoch": 1.141045154743785, + "grad_norm": 4.90625, + "learning_rate": 4.250137711209258e-06, + "loss": 1.67017517, + "memory(GiB)": 111.15, + "step": 44980, + "train_speed(iter/s)": 0.447699 + }, + { + "acc": 0.64830098, + "epoch": 1.14117199391172, + "grad_norm": 5.59375, + "learning_rate": 4.249100963779028e-06, + "loss": 1.7100935, + "memory(GiB)": 111.15, + "step": 44985, + "train_speed(iter/s)": 0.447725 + }, + { + "acc": 0.65711031, + "epoch": 1.141298833079655, + "grad_norm": 6.1875, + "learning_rate": 4.248064249376692e-06, + "loss": 1.61508427, + "memory(GiB)": 111.15, + "step": 44990, + "train_speed(iter/s)": 0.447751 + }, + { + "acc": 0.65331855, + "epoch": 1.14142567224759, + "grad_norm": 5.09375, + "learning_rate": 4.2470275680478466e-06, + "loss": 1.59913597, + "memory(GiB)": 111.15, + "step": 44995, + "train_speed(iter/s)": 0.447778 + }, + { + "acc": 0.65907564, + "epoch": 1.1415525114155252, + "grad_norm": 5.9375, + "learning_rate": 4.2459909198380886e-06, + "loss": 1.58414555, + "memory(GiB)": 111.15, + "step": 45000, + "train_speed(iter/s)": 0.447803 + }, + { + "epoch": 1.1415525114155252, + "eval_acc": 0.6433624920989758, + "eval_loss": 1.607709288597107, + "eval_runtime": 114.3564, + "eval_samples_per_second": 55.703, + "eval_steps_per_second": 27.852, + "step": 45000 + }, + { + "acc": 0.66488857, + "epoch": 1.1416793505834602, + "grad_norm": 5.34375, + "learning_rate": 4.244954304793019e-06, + "loss": 1.56073761, + "memory(GiB)": 111.15, + "step": 45005, + "train_speed(iter/s)": 0.4473 + }, + { + "acc": 0.6437233, + "epoch": 1.1418061897513951, + "grad_norm": 5.125, + "learning_rate": 4.2439177229582304e-06, + "loss": 1.63757057, + "memory(GiB)": 111.15, + "step": 45010, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.65384283, + "epoch": 1.1419330289193304, + "grad_norm": 5.1875, + "learning_rate": 4.242881174379313e-06, + "loss": 1.60467892, + "memory(GiB)": 111.15, + "step": 45015, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.64354811, + "epoch": 1.1420598680872653, + "grad_norm": 5.1875, + "learning_rate": 4.241844659101865e-06, + "loss": 1.64326077, + "memory(GiB)": 111.15, + "step": 45020, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.63499041, + "epoch": 1.1421867072552003, + "grad_norm": 4.59375, + "learning_rate": 4.240808177171472e-06, + "loss": 1.71998444, + "memory(GiB)": 111.15, + "step": 45025, + "train_speed(iter/s)": 0.447404 + }, + { + "acc": 0.64933777, + "epoch": 1.1423135464231355, + "grad_norm": 5.53125, + "learning_rate": 4.239771728633727e-06, + "loss": 1.59443846, + "memory(GiB)": 111.15, + "step": 45030, + "train_speed(iter/s)": 0.447431 + }, + { + "acc": 0.61932917, + "epoch": 1.1424403855910705, + "grad_norm": 4.78125, + "learning_rate": 4.238735313534213e-06, + "loss": 1.7109457, + "memory(GiB)": 111.15, + "step": 45035, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.6501318, + "epoch": 1.1425672247590055, + "grad_norm": 5.875, + "learning_rate": 4.23769893191852e-06, + "loss": 1.67295132, + "memory(GiB)": 111.15, + "step": 45040, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.6526823, + "epoch": 1.1426940639269407, + "grad_norm": 4.8125, + "learning_rate": 4.236662583832229e-06, + "loss": 1.5519186, + "memory(GiB)": 111.15, + "step": 45045, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.63970671, + "epoch": 1.1428209030948757, + "grad_norm": 5.03125, + "learning_rate": 4.23562626932093e-06, + "loss": 1.73732872, + "memory(GiB)": 111.15, + "step": 45050, + "train_speed(iter/s)": 0.447536 + }, + { + "acc": 0.63720961, + "epoch": 1.1429477422628107, + "grad_norm": 5.28125, + "learning_rate": 4.234589988430198e-06, + "loss": 1.64321747, + "memory(GiB)": 111.15, + "step": 45055, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.64486284, + "epoch": 1.1430745814307457, + "grad_norm": 5.5625, + "learning_rate": 4.233553741205615e-06, + "loss": 1.67430878, + "memory(GiB)": 111.15, + "step": 45060, + "train_speed(iter/s)": 0.447588 + }, + { + "acc": 0.63723578, + "epoch": 1.143201420598681, + "grad_norm": 5.75, + "learning_rate": 4.2325175276927614e-06, + "loss": 1.61616135, + "memory(GiB)": 111.15, + "step": 45065, + "train_speed(iter/s)": 0.447615 + }, + { + "acc": 0.64830894, + "epoch": 1.143328259766616, + "grad_norm": 5.0625, + "learning_rate": 4.231481347937214e-06, + "loss": 1.60404282, + "memory(GiB)": 111.15, + "step": 45070, + "train_speed(iter/s)": 0.447641 + }, + { + "acc": 0.65453186, + "epoch": 1.143455098934551, + "grad_norm": 5.25, + "learning_rate": 4.230445201984547e-06, + "loss": 1.61684265, + "memory(GiB)": 111.15, + "step": 45075, + "train_speed(iter/s)": 0.447667 + }, + { + "acc": 0.64846401, + "epoch": 1.143581938102486, + "grad_norm": 4.78125, + "learning_rate": 4.229409089880336e-06, + "loss": 1.64810181, + "memory(GiB)": 111.15, + "step": 45080, + "train_speed(iter/s)": 0.447693 + }, + { + "acc": 0.64508514, + "epoch": 1.143708777270421, + "grad_norm": 5.78125, + "learning_rate": 4.2283730116701535e-06, + "loss": 1.65871544, + "memory(GiB)": 111.15, + "step": 45085, + "train_speed(iter/s)": 0.44772 + }, + { + "acc": 0.65791316, + "epoch": 1.143835616438356, + "grad_norm": 5.65625, + "learning_rate": 4.227336967399573e-06, + "loss": 1.62084579, + "memory(GiB)": 111.15, + "step": 45090, + "train_speed(iter/s)": 0.447746 + }, + { + "acc": 0.65844827, + "epoch": 1.1439624556062913, + "grad_norm": 5.3125, + "learning_rate": 4.2263009571141585e-06, + "loss": 1.59415226, + "memory(GiB)": 111.15, + "step": 45095, + "train_speed(iter/s)": 0.447772 + }, + { + "acc": 0.65585384, + "epoch": 1.1440892947742263, + "grad_norm": 5.03125, + "learning_rate": 4.225264980859485e-06, + "loss": 1.62433815, + "memory(GiB)": 111.15, + "step": 45100, + "train_speed(iter/s)": 0.447799 + }, + { + "epoch": 1.1440892947742263, + "eval_acc": 0.6433290707142192, + "eval_loss": 1.6076991558074951, + "eval_runtime": 113.7255, + "eval_samples_per_second": 56.012, + "eval_steps_per_second": 28.006, + "step": 45100 + }, + { + "acc": 0.64157009, + "epoch": 1.1442161339421613, + "grad_norm": 5.46875, + "learning_rate": 4.224229038681115e-06, + "loss": 1.66799831, + "memory(GiB)": 111.15, + "step": 45105, + "train_speed(iter/s)": 0.4473 + }, + { + "acc": 0.64234676, + "epoch": 1.1443429731100965, + "grad_norm": 5.1875, + "learning_rate": 4.223193130624619e-06, + "loss": 1.66464672, + "memory(GiB)": 111.15, + "step": 45110, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.6652236, + "epoch": 1.1444698122780315, + "grad_norm": 5.25, + "learning_rate": 4.222157256735553e-06, + "loss": 1.53412838, + "memory(GiB)": 111.15, + "step": 45115, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.66031313, + "epoch": 1.1445966514459665, + "grad_norm": 5.59375, + "learning_rate": 4.2211214170594865e-06, + "loss": 1.60161018, + "memory(GiB)": 111.15, + "step": 45120, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.6410696, + "epoch": 1.1447234906139017, + "grad_norm": 6.0, + "learning_rate": 4.220085611641976e-06, + "loss": 1.64588242, + "memory(GiB)": 111.15, + "step": 45125, + "train_speed(iter/s)": 0.447404 + }, + { + "acc": 0.65715427, + "epoch": 1.1448503297818367, + "grad_norm": 4.75, + "learning_rate": 4.2190498405285826e-06, + "loss": 1.68535824, + "memory(GiB)": 111.15, + "step": 45130, + "train_speed(iter/s)": 0.44743 + }, + { + "acc": 0.65598812, + "epoch": 1.1449771689497716, + "grad_norm": 5.40625, + "learning_rate": 4.218014103764865e-06, + "loss": 1.61414146, + "memory(GiB)": 111.15, + "step": 45135, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.63089905, + "epoch": 1.1451040081177069, + "grad_norm": 5.0625, + "learning_rate": 4.216978401396376e-06, + "loss": 1.68057976, + "memory(GiB)": 111.15, + "step": 45140, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64114971, + "epoch": 1.1452308472856418, + "grad_norm": 6.09375, + "learning_rate": 4.215942733468675e-06, + "loss": 1.66314812, + "memory(GiB)": 111.15, + "step": 45145, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.67007871, + "epoch": 1.1453576864535768, + "grad_norm": 5.375, + "learning_rate": 4.2149071000273134e-06, + "loss": 1.55185585, + "memory(GiB)": 111.15, + "step": 45150, + "train_speed(iter/s)": 0.447535 + }, + { + "acc": 0.65442743, + "epoch": 1.1454845256215118, + "grad_norm": 6.5625, + "learning_rate": 4.213871501117842e-06, + "loss": 1.55826378, + "memory(GiB)": 111.15, + "step": 45155, + "train_speed(iter/s)": 0.447561 + }, + { + "acc": 0.64640036, + "epoch": 1.145611364789447, + "grad_norm": 5.3125, + "learning_rate": 4.212835936785811e-06, + "loss": 1.74158249, + "memory(GiB)": 111.15, + "step": 45160, + "train_speed(iter/s)": 0.447587 + }, + { + "acc": 0.6636385, + "epoch": 1.145738203957382, + "grad_norm": 5.40625, + "learning_rate": 4.21180040707677e-06, + "loss": 1.54877739, + "memory(GiB)": 111.15, + "step": 45165, + "train_speed(iter/s)": 0.447614 + }, + { + "acc": 0.65347195, + "epoch": 1.145865043125317, + "grad_norm": 4.96875, + "learning_rate": 4.2107649120362684e-06, + "loss": 1.66960812, + "memory(GiB)": 111.15, + "step": 45170, + "train_speed(iter/s)": 0.44764 + }, + { + "acc": 0.63969984, + "epoch": 1.1459918822932522, + "grad_norm": 4.46875, + "learning_rate": 4.2097294517098465e-06, + "loss": 1.6722599, + "memory(GiB)": 111.15, + "step": 45175, + "train_speed(iter/s)": 0.447666 + }, + { + "acc": 0.65433836, + "epoch": 1.1461187214611872, + "grad_norm": 5.15625, + "learning_rate": 4.208694026143054e-06, + "loss": 1.68519917, + "memory(GiB)": 111.15, + "step": 45180, + "train_speed(iter/s)": 0.447692 + }, + { + "acc": 0.65259008, + "epoch": 1.1462455606291222, + "grad_norm": 6.09375, + "learning_rate": 4.2076586353814295e-06, + "loss": 1.71392841, + "memory(GiB)": 111.15, + "step": 45185, + "train_speed(iter/s)": 0.447718 + }, + { + "acc": 0.65216694, + "epoch": 1.1463723997970574, + "grad_norm": 5.03125, + "learning_rate": 4.2066232794705174e-06, + "loss": 1.60369568, + "memory(GiB)": 111.15, + "step": 45190, + "train_speed(iter/s)": 0.447744 + }, + { + "acc": 0.64312439, + "epoch": 1.1464992389649924, + "grad_norm": 5.125, + "learning_rate": 4.205587958455854e-06, + "loss": 1.61999855, + "memory(GiB)": 111.15, + "step": 45195, + "train_speed(iter/s)": 0.44777 + }, + { + "acc": 0.64466305, + "epoch": 1.1466260781329274, + "grad_norm": 4.96875, + "learning_rate": 4.204552672382981e-06, + "loss": 1.66653423, + "memory(GiB)": 111.15, + "step": 45200, + "train_speed(iter/s)": 0.447796 + }, + { + "epoch": 1.1466260781329274, + "eval_acc": 0.6433487057777637, + "eval_loss": 1.6078672409057617, + "eval_runtime": 113.5799, + "eval_samples_per_second": 56.084, + "eval_steps_per_second": 28.042, + "step": 45200 + }, + { + "acc": 0.64580183, + "epoch": 1.1467529173008626, + "grad_norm": 4.8125, + "learning_rate": 4.203517421297431e-06, + "loss": 1.59700603, + "memory(GiB)": 111.15, + "step": 45205, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.63825407, + "epoch": 1.1468797564687976, + "grad_norm": 5.3125, + "learning_rate": 4.202482205244742e-06, + "loss": 1.60356331, + "memory(GiB)": 111.15, + "step": 45210, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.66770859, + "epoch": 1.1470065956367326, + "grad_norm": 5.5, + "learning_rate": 4.201447024270446e-06, + "loss": 1.5577117, + "memory(GiB)": 111.15, + "step": 45215, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.64789791, + "epoch": 1.1471334348046676, + "grad_norm": 5.125, + "learning_rate": 4.200411878420074e-06, + "loss": 1.67158813, + "memory(GiB)": 111.15, + "step": 45220, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.6459281, + "epoch": 1.1472602739726028, + "grad_norm": 5.65625, + "learning_rate": 4.199376767739158e-06, + "loss": 1.61804886, + "memory(GiB)": 111.15, + "step": 45225, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.66516528, + "epoch": 1.1473871131405378, + "grad_norm": 5.90625, + "learning_rate": 4.1983416922732276e-06, + "loss": 1.61684914, + "memory(GiB)": 111.15, + "step": 45230, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.64896107, + "epoch": 1.147513952308473, + "grad_norm": 4.875, + "learning_rate": 4.197306652067807e-06, + "loss": 1.58107491, + "memory(GiB)": 111.15, + "step": 45235, + "train_speed(iter/s)": 0.447455 + }, + { + "acc": 0.66222744, + "epoch": 1.147640791476408, + "grad_norm": 5.8125, + "learning_rate": 4.196271647168425e-06, + "loss": 1.57762699, + "memory(GiB)": 111.15, + "step": 45240, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.63739042, + "epoch": 1.147767630644343, + "grad_norm": 5.5625, + "learning_rate": 4.195236677620604e-06, + "loss": 1.68059273, + "memory(GiB)": 111.15, + "step": 45245, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.66593103, + "epoch": 1.147894469812278, + "grad_norm": 6.46875, + "learning_rate": 4.1942017434698675e-06, + "loss": 1.49003239, + "memory(GiB)": 111.15, + "step": 45250, + "train_speed(iter/s)": 0.447533 + }, + { + "acc": 0.65071239, + "epoch": 1.1480213089802132, + "grad_norm": 4.8125, + "learning_rate": 4.1931668447617346e-06, + "loss": 1.60624237, + "memory(GiB)": 111.15, + "step": 45255, + "train_speed(iter/s)": 0.447559 + }, + { + "acc": 0.66678309, + "epoch": 1.1481481481481481, + "grad_norm": 5.78125, + "learning_rate": 4.192131981541727e-06, + "loss": 1.6613369, + "memory(GiB)": 111.15, + "step": 45260, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.65607386, + "epoch": 1.1482749873160831, + "grad_norm": 5.21875, + "learning_rate": 4.19109715385536e-06, + "loss": 1.5743391, + "memory(GiB)": 111.15, + "step": 45265, + "train_speed(iter/s)": 0.447612 + }, + { + "acc": 0.63558378, + "epoch": 1.1484018264840183, + "grad_norm": 5.03125, + "learning_rate": 4.190062361748154e-06, + "loss": 1.62784996, + "memory(GiB)": 111.15, + "step": 45270, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.64226389, + "epoch": 1.1485286656519533, + "grad_norm": 6.125, + "learning_rate": 4.189027605265621e-06, + "loss": 1.62233315, + "memory(GiB)": 111.15, + "step": 45275, + "train_speed(iter/s)": 0.447664 + }, + { + "acc": 0.64902964, + "epoch": 1.1486555048198883, + "grad_norm": 6.25, + "learning_rate": 4.187992884453273e-06, + "loss": 1.63907604, + "memory(GiB)": 111.15, + "step": 45280, + "train_speed(iter/s)": 0.44769 + }, + { + "acc": 0.65071526, + "epoch": 1.1487823439878235, + "grad_norm": 4.9375, + "learning_rate": 4.186958199356624e-06, + "loss": 1.64843178, + "memory(GiB)": 111.15, + "step": 45285, + "train_speed(iter/s)": 0.447716 + }, + { + "acc": 0.64693403, + "epoch": 1.1489091831557585, + "grad_norm": 5.5, + "learning_rate": 4.185923550021185e-06, + "loss": 1.6569416, + "memory(GiB)": 111.15, + "step": 45290, + "train_speed(iter/s)": 0.447742 + }, + { + "acc": 0.65490723, + "epoch": 1.1490360223236935, + "grad_norm": 4.59375, + "learning_rate": 4.1848889364924625e-06, + "loss": 1.61062965, + "memory(GiB)": 111.15, + "step": 45295, + "train_speed(iter/s)": 0.447768 + }, + { + "acc": 0.64903302, + "epoch": 1.1491628614916287, + "grad_norm": 5.71875, + "learning_rate": 4.183854358815962e-06, + "loss": 1.66614342, + "memory(GiB)": 111.15, + "step": 45300, + "train_speed(iter/s)": 0.447795 + }, + { + "epoch": 1.1491628614916287, + "eval_acc": 0.6433487057777637, + "eval_loss": 1.6079432964324951, + "eval_runtime": 114.2185, + "eval_samples_per_second": 55.77, + "eval_steps_per_second": 27.885, + "step": 45300 + }, + { + "acc": 0.64780688, + "epoch": 1.1492897006595637, + "grad_norm": 6.125, + "learning_rate": 4.182819817037192e-06, + "loss": 1.61782818, + "memory(GiB)": 111.15, + "step": 45305, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.64410229, + "epoch": 1.1494165398274987, + "grad_norm": 5.5, + "learning_rate": 4.181785311201655e-06, + "loss": 1.61783447, + "memory(GiB)": 111.15, + "step": 45310, + "train_speed(iter/s)": 0.447321 + }, + { + "acc": 0.64807615, + "epoch": 1.1495433789954337, + "grad_norm": 5.21875, + "learning_rate": 4.1807508413548515e-06, + "loss": 1.67056999, + "memory(GiB)": 111.15, + "step": 45315, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.65412989, + "epoch": 1.149670218163369, + "grad_norm": 5.09375, + "learning_rate": 4.179716407542285e-06, + "loss": 1.55940466, + "memory(GiB)": 111.15, + "step": 45320, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.64235516, + "epoch": 1.1497970573313039, + "grad_norm": 5.375, + "learning_rate": 4.178682009809452e-06, + "loss": 1.64976883, + "memory(GiB)": 111.15, + "step": 45325, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.64815207, + "epoch": 1.1499238964992389, + "grad_norm": 4.40625, + "learning_rate": 4.177647648201854e-06, + "loss": 1.61502342, + "memory(GiB)": 111.15, + "step": 45330, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.64597325, + "epoch": 1.150050735667174, + "grad_norm": 5.28125, + "learning_rate": 4.1766133227649815e-06, + "loss": 1.61486626, + "memory(GiB)": 111.15, + "step": 45335, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.65643806, + "epoch": 1.150177574835109, + "grad_norm": 5.0, + "learning_rate": 4.175579033544332e-06, + "loss": 1.57170238, + "memory(GiB)": 111.15, + "step": 45340, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.63630414, + "epoch": 1.150304414003044, + "grad_norm": 5.34375, + "learning_rate": 4.174544780585395e-06, + "loss": 1.66457138, + "memory(GiB)": 111.15, + "step": 45345, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.63696241, + "epoch": 1.1504312531709793, + "grad_norm": 9.125, + "learning_rate": 4.1735105639336686e-06, + "loss": 1.75595474, + "memory(GiB)": 111.15, + "step": 45350, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.6445446, + "epoch": 1.1505580923389143, + "grad_norm": 4.84375, + "learning_rate": 4.172476383634635e-06, + "loss": 1.66150551, + "memory(GiB)": 111.15, + "step": 45355, + "train_speed(iter/s)": 0.447556 + }, + { + "acc": 0.65328083, + "epoch": 1.1506849315068493, + "grad_norm": 5.75, + "learning_rate": 4.171442239733783e-06, + "loss": 1.6988678, + "memory(GiB)": 111.15, + "step": 45360, + "train_speed(iter/s)": 0.447582 + }, + { + "acc": 0.63236966, + "epoch": 1.1508117706747845, + "grad_norm": 4.84375, + "learning_rate": 4.170408132276603e-06, + "loss": 1.71014271, + "memory(GiB)": 111.15, + "step": 45365, + "train_speed(iter/s)": 0.447608 + }, + { + "acc": 0.6446887, + "epoch": 1.1509386098427195, + "grad_norm": 7.15625, + "learning_rate": 4.1693740613085776e-06, + "loss": 1.65525208, + "memory(GiB)": 111.15, + "step": 45370, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.66644487, + "epoch": 1.1510654490106544, + "grad_norm": 5.34375, + "learning_rate": 4.168340026875188e-06, + "loss": 1.59343204, + "memory(GiB)": 111.15, + "step": 45375, + "train_speed(iter/s)": 0.447661 + }, + { + "acc": 0.65400028, + "epoch": 1.1511922881785894, + "grad_norm": 4.78125, + "learning_rate": 4.167306029021917e-06, + "loss": 1.61788998, + "memory(GiB)": 111.15, + "step": 45380, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.648281, + "epoch": 1.1513191273465246, + "grad_norm": 5.9375, + "learning_rate": 4.166272067794246e-06, + "loss": 1.67480583, + "memory(GiB)": 111.15, + "step": 45385, + "train_speed(iter/s)": 0.447713 + }, + { + "acc": 0.64124966, + "epoch": 1.1514459665144596, + "grad_norm": 4.84375, + "learning_rate": 4.165238143237651e-06, + "loss": 1.67270679, + "memory(GiB)": 111.15, + "step": 45390, + "train_speed(iter/s)": 0.447739 + }, + { + "acc": 0.63464108, + "epoch": 1.1515728056823948, + "grad_norm": 4.90625, + "learning_rate": 4.164204255397608e-06, + "loss": 1.64561214, + "memory(GiB)": 111.15, + "step": 45395, + "train_speed(iter/s)": 0.447765 + }, + { + "acc": 0.63913436, + "epoch": 1.1516996448503298, + "grad_norm": 4.53125, + "learning_rate": 4.163170404319596e-06, + "loss": 1.64672394, + "memory(GiB)": 111.15, + "step": 45400, + "train_speed(iter/s)": 0.447791 + }, + { + "epoch": 1.1516996448503298, + "eval_acc": 0.6433269818776719, + "eval_loss": 1.6078561544418335, + "eval_runtime": 114.3438, + "eval_samples_per_second": 55.709, + "eval_steps_per_second": 27.855, + "step": 45400 + }, + { + "acc": 0.65607281, + "epoch": 1.1518264840182648, + "grad_norm": 5.28125, + "learning_rate": 4.1621365900490825e-06, + "loss": 1.64979572, + "memory(GiB)": 111.15, + "step": 45405, + "train_speed(iter/s)": 0.447292 + }, + { + "acc": 0.65594492, + "epoch": 1.1519533231861998, + "grad_norm": 5.875, + "learning_rate": 4.1611028126315455e-06, + "loss": 1.58164721, + "memory(GiB)": 111.15, + "step": 45410, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.64823623, + "epoch": 1.152080162354135, + "grad_norm": 5.25, + "learning_rate": 4.160069072112451e-06, + "loss": 1.62986145, + "memory(GiB)": 111.15, + "step": 45415, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.65335965, + "epoch": 1.15220700152207, + "grad_norm": 4.5625, + "learning_rate": 4.1590353685372695e-06, + "loss": 1.53595076, + "memory(GiB)": 111.15, + "step": 45420, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.66108227, + "epoch": 1.152333840690005, + "grad_norm": 6.09375, + "learning_rate": 4.158001701951465e-06, + "loss": 1.56029129, + "memory(GiB)": 111.15, + "step": 45425, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.64758368, + "epoch": 1.1524606798579402, + "grad_norm": 5.71875, + "learning_rate": 4.156968072400508e-06, + "loss": 1.64130363, + "memory(GiB)": 111.15, + "step": 45430, + "train_speed(iter/s)": 0.44742 + }, + { + "acc": 0.64628553, + "epoch": 1.1525875190258752, + "grad_norm": 6.15625, + "learning_rate": 4.155934479929858e-06, + "loss": 1.65735149, + "memory(GiB)": 111.15, + "step": 45435, + "train_speed(iter/s)": 0.447446 + }, + { + "acc": 0.67608609, + "epoch": 1.1527143581938102, + "grad_norm": 7.65625, + "learning_rate": 4.154900924584976e-06, + "loss": 1.49414806, + "memory(GiB)": 111.15, + "step": 45440, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.63613725, + "epoch": 1.1528411973617454, + "grad_norm": 5.09375, + "learning_rate": 4.153867406411327e-06, + "loss": 1.69173279, + "memory(GiB)": 111.15, + "step": 45445, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.66287889, + "epoch": 1.1529680365296804, + "grad_norm": 5.46875, + "learning_rate": 4.152833925454367e-06, + "loss": 1.5360528, + "memory(GiB)": 111.15, + "step": 45450, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.65611873, + "epoch": 1.1530948756976154, + "grad_norm": 4.96875, + "learning_rate": 4.1518004817595515e-06, + "loss": 1.57973156, + "memory(GiB)": 111.15, + "step": 45455, + "train_speed(iter/s)": 0.447549 + }, + { + "acc": 0.62621942, + "epoch": 1.1532217148655506, + "grad_norm": 6.1875, + "learning_rate": 4.150767075372338e-06, + "loss": 1.7481163, + "memory(GiB)": 111.15, + "step": 45460, + "train_speed(iter/s)": 0.447575 + }, + { + "acc": 0.65387621, + "epoch": 1.1533485540334856, + "grad_norm": 5.09375, + "learning_rate": 4.149733706338182e-06, + "loss": 1.56980419, + "memory(GiB)": 111.15, + "step": 45465, + "train_speed(iter/s)": 0.447601 + }, + { + "acc": 0.65518246, + "epoch": 1.1534753932014206, + "grad_norm": 5.5625, + "learning_rate": 4.148700374702533e-06, + "loss": 1.6238739, + "memory(GiB)": 111.15, + "step": 45470, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.65790477, + "epoch": 1.1536022323693556, + "grad_norm": 4.78125, + "learning_rate": 4.147667080510841e-06, + "loss": 1.60335903, + "memory(GiB)": 111.15, + "step": 45475, + "train_speed(iter/s)": 0.447652 + }, + { + "acc": 0.65946169, + "epoch": 1.1537290715372908, + "grad_norm": 4.75, + "learning_rate": 4.146633823808557e-06, + "loss": 1.57947006, + "memory(GiB)": 111.15, + "step": 45480, + "train_speed(iter/s)": 0.447678 + }, + { + "acc": 0.63158779, + "epoch": 1.1538559107052258, + "grad_norm": 5.75, + "learning_rate": 4.145600604641127e-06, + "loss": 1.64386711, + "memory(GiB)": 111.15, + "step": 45485, + "train_speed(iter/s)": 0.447704 + }, + { + "acc": 0.65775585, + "epoch": 1.1539827498731607, + "grad_norm": 5.90625, + "learning_rate": 4.1445674230539985e-06, + "loss": 1.60229645, + "memory(GiB)": 111.15, + "step": 45490, + "train_speed(iter/s)": 0.44773 + }, + { + "acc": 0.66226811, + "epoch": 1.154109589041096, + "grad_norm": 5.84375, + "learning_rate": 4.143534279092613e-06, + "loss": 1.58984165, + "memory(GiB)": 111.15, + "step": 45495, + "train_speed(iter/s)": 0.447756 + }, + { + "acc": 0.64538217, + "epoch": 1.154236428209031, + "grad_norm": 6.03125, + "learning_rate": 4.142501172802412e-06, + "loss": 1.69725189, + "memory(GiB)": 111.15, + "step": 45500, + "train_speed(iter/s)": 0.447782 + }, + { + "epoch": 1.154236428209031, + "eval_acc": 0.6433040046756517, + "eval_loss": 1.6078029870986938, + "eval_runtime": 114.9312, + "eval_samples_per_second": 55.424, + "eval_steps_per_second": 27.712, + "step": 45500 + }, + { + "acc": 0.64054174, + "epoch": 1.154363267376966, + "grad_norm": 5.59375, + "learning_rate": 4.14146810422884e-06, + "loss": 1.67145557, + "memory(GiB)": 111.15, + "step": 45505, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.64140358, + "epoch": 1.1544901065449011, + "grad_norm": 5.6875, + "learning_rate": 4.140435073417335e-06, + "loss": 1.59997787, + "memory(GiB)": 111.15, + "step": 45510, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.6527741, + "epoch": 1.1546169457128361, + "grad_norm": 4.8125, + "learning_rate": 4.139402080413331e-06, + "loss": 1.64723988, + "memory(GiB)": 111.15, + "step": 45515, + "train_speed(iter/s)": 0.447332 + }, + { + "acc": 0.65843873, + "epoch": 1.1547437848807711, + "grad_norm": 5.25, + "learning_rate": 4.138369125262266e-06, + "loss": 1.63645401, + "memory(GiB)": 111.15, + "step": 45520, + "train_speed(iter/s)": 0.447358 + }, + { + "acc": 0.65536804, + "epoch": 1.1548706240487063, + "grad_norm": 5.09375, + "learning_rate": 4.137336208009574e-06, + "loss": 1.62714195, + "memory(GiB)": 111.15, + "step": 45525, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.64162169, + "epoch": 1.1549974632166413, + "grad_norm": 5.15625, + "learning_rate": 4.136303328700688e-06, + "loss": 1.66807919, + "memory(GiB)": 111.15, + "step": 45530, + "train_speed(iter/s)": 0.447409 + }, + { + "acc": 0.62809343, + "epoch": 1.1551243023845763, + "grad_norm": 5.375, + "learning_rate": 4.135270487381037e-06, + "loss": 1.66909676, + "memory(GiB)": 111.15, + "step": 45535, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.65639286, + "epoch": 1.1552511415525113, + "grad_norm": 4.78125, + "learning_rate": 4.13423768409605e-06, + "loss": 1.56652775, + "memory(GiB)": 111.15, + "step": 45540, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.64756355, + "epoch": 1.1553779807204465, + "grad_norm": 4.84375, + "learning_rate": 4.133204918891155e-06, + "loss": 1.67517319, + "memory(GiB)": 111.15, + "step": 45545, + "train_speed(iter/s)": 0.447487 + }, + { + "acc": 0.65246973, + "epoch": 1.1555048198883815, + "grad_norm": 5.4375, + "learning_rate": 4.132172191811781e-06, + "loss": 1.6293499, + "memory(GiB)": 111.15, + "step": 45550, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.65187426, + "epoch": 1.1556316590563167, + "grad_norm": 5.03125, + "learning_rate": 4.131139502903345e-06, + "loss": 1.63129807, + "memory(GiB)": 111.15, + "step": 45555, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.64164233, + "epoch": 1.1557584982242517, + "grad_norm": 4.59375, + "learning_rate": 4.130106852211273e-06, + "loss": 1.66203976, + "memory(GiB)": 111.15, + "step": 45560, + "train_speed(iter/s)": 0.447564 + }, + { + "acc": 0.65331469, + "epoch": 1.1558853373921867, + "grad_norm": 5.40625, + "learning_rate": 4.129074239780986e-06, + "loss": 1.57181702, + "memory(GiB)": 111.15, + "step": 45565, + "train_speed(iter/s)": 0.44759 + }, + { + "acc": 0.66523986, + "epoch": 1.1560121765601217, + "grad_norm": 5.65625, + "learning_rate": 4.128041665657903e-06, + "loss": 1.60903664, + "memory(GiB)": 111.15, + "step": 45570, + "train_speed(iter/s)": 0.447616 + }, + { + "acc": 0.65577946, + "epoch": 1.1561390157280569, + "grad_norm": 5.1875, + "learning_rate": 4.127009129887441e-06, + "loss": 1.58849163, + "memory(GiB)": 111.15, + "step": 45575, + "train_speed(iter/s)": 0.447642 + }, + { + "acc": 0.64465618, + "epoch": 1.1562658548959919, + "grad_norm": 5.4375, + "learning_rate": 4.125976632515013e-06, + "loss": 1.63075867, + "memory(GiB)": 111.15, + "step": 45580, + "train_speed(iter/s)": 0.447668 + }, + { + "acc": 0.64686885, + "epoch": 1.1563926940639269, + "grad_norm": 4.84375, + "learning_rate": 4.124944173586036e-06, + "loss": 1.6418396, + "memory(GiB)": 111.15, + "step": 45585, + "train_speed(iter/s)": 0.447693 + }, + { + "acc": 0.64398208, + "epoch": 1.156519533231862, + "grad_norm": 5.75, + "learning_rate": 4.123911753145922e-06, + "loss": 1.66530151, + "memory(GiB)": 111.15, + "step": 45590, + "train_speed(iter/s)": 0.447718 + }, + { + "acc": 0.65579815, + "epoch": 1.156646372399797, + "grad_norm": 5.28125, + "learning_rate": 4.12287937124008e-06, + "loss": 1.62653942, + "memory(GiB)": 111.15, + "step": 45595, + "train_speed(iter/s)": 0.447744 + }, + { + "acc": 0.64812365, + "epoch": 1.156773211567732, + "grad_norm": 7.0625, + "learning_rate": 4.121847027913918e-06, + "loss": 1.67217884, + "memory(GiB)": 111.15, + "step": 45600, + "train_speed(iter/s)": 0.44777 + }, + { + "epoch": 1.156773211567732, + "eval_acc": 0.6433023336064139, + "eval_loss": 1.607658863067627, + "eval_runtime": 114.1318, + "eval_samples_per_second": 55.813, + "eval_steps_per_second": 27.906, + "step": 45600 + }, + { + "acc": 0.65976901, + "epoch": 1.1569000507356673, + "grad_norm": 5.84375, + "learning_rate": 4.1208147232128456e-06, + "loss": 1.61994743, + "memory(GiB)": 111.15, + "step": 45605, + "train_speed(iter/s)": 0.447274 + }, + { + "acc": 0.63547106, + "epoch": 1.1570268899036023, + "grad_norm": 5.5625, + "learning_rate": 4.119782457182267e-06, + "loss": 1.67374229, + "memory(GiB)": 111.15, + "step": 45610, + "train_speed(iter/s)": 0.4473 + }, + { + "acc": 0.65958085, + "epoch": 1.1571537290715372, + "grad_norm": 5.3125, + "learning_rate": 4.118750229867585e-06, + "loss": 1.62442532, + "memory(GiB)": 111.15, + "step": 45615, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.64503198, + "epoch": 1.1572805682394725, + "grad_norm": 5.84375, + "learning_rate": 4.117718041314204e-06, + "loss": 1.65934982, + "memory(GiB)": 111.15, + "step": 45620, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.64340916, + "epoch": 1.1574074074074074, + "grad_norm": 5.0, + "learning_rate": 4.11668589156752e-06, + "loss": 1.63297386, + "memory(GiB)": 111.15, + "step": 45625, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.66505585, + "epoch": 1.1575342465753424, + "grad_norm": 5.34375, + "learning_rate": 4.115653780672937e-06, + "loss": 1.55832024, + "memory(GiB)": 111.15, + "step": 45630, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.64674506, + "epoch": 1.1576610857432774, + "grad_norm": 5.375, + "learning_rate": 4.1146217086758475e-06, + "loss": 1.61876583, + "memory(GiB)": 111.15, + "step": 45635, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.66942453, + "epoch": 1.1577879249112126, + "grad_norm": 5.5, + "learning_rate": 4.113589675621649e-06, + "loss": 1.52456083, + "memory(GiB)": 111.15, + "step": 45640, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.66353416, + "epoch": 1.1579147640791476, + "grad_norm": 4.65625, + "learning_rate": 4.112557681555733e-06, + "loss": 1.56752682, + "memory(GiB)": 111.15, + "step": 45645, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.65083132, + "epoch": 1.1580416032470826, + "grad_norm": 5.375, + "learning_rate": 4.111525726523494e-06, + "loss": 1.64165688, + "memory(GiB)": 111.15, + "step": 45650, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.66319232, + "epoch": 1.1581684424150178, + "grad_norm": 4.59375, + "learning_rate": 4.110493810570319e-06, + "loss": 1.53055897, + "memory(GiB)": 111.15, + "step": 45655, + "train_speed(iter/s)": 0.447532 + }, + { + "acc": 0.64710317, + "epoch": 1.1582952815829528, + "grad_norm": 6.28125, + "learning_rate": 4.109461933741598e-06, + "loss": 1.63489361, + "memory(GiB)": 111.15, + "step": 45660, + "train_speed(iter/s)": 0.447557 + }, + { + "acc": 0.64023962, + "epoch": 1.1584221207508878, + "grad_norm": 6.25, + "learning_rate": 4.108430096082716e-06, + "loss": 1.6917078, + "memory(GiB)": 111.15, + "step": 45665, + "train_speed(iter/s)": 0.447583 + }, + { + "acc": 0.64796915, + "epoch": 1.158548959918823, + "grad_norm": 5.09375, + "learning_rate": 4.107398297639062e-06, + "loss": 1.64578552, + "memory(GiB)": 111.15, + "step": 45670, + "train_speed(iter/s)": 0.447609 + }, + { + "acc": 0.64194632, + "epoch": 1.158675799086758, + "grad_norm": 5.46875, + "learning_rate": 4.106366538456013e-06, + "loss": 1.66036797, + "memory(GiB)": 111.15, + "step": 45675, + "train_speed(iter/s)": 0.447635 + }, + { + "acc": 0.65915813, + "epoch": 1.158802638254693, + "grad_norm": 5.15625, + "learning_rate": 4.105334818578954e-06, + "loss": 1.56280708, + "memory(GiB)": 111.15, + "step": 45680, + "train_speed(iter/s)": 0.447661 + }, + { + "acc": 0.65026851, + "epoch": 1.1589294774226282, + "grad_norm": 5.25, + "learning_rate": 4.104303138053265e-06, + "loss": 1.64805145, + "memory(GiB)": 111.15, + "step": 45685, + "train_speed(iter/s)": 0.447687 + }, + { + "acc": 0.66475115, + "epoch": 1.1590563165905632, + "grad_norm": 6.09375, + "learning_rate": 4.103271496924323e-06, + "loss": 1.59839268, + "memory(GiB)": 111.15, + "step": 45690, + "train_speed(iter/s)": 0.447712 + }, + { + "acc": 0.66686687, + "epoch": 1.1591831557584982, + "grad_norm": 6.8125, + "learning_rate": 4.102239895237503e-06, + "loss": 1.6438694, + "memory(GiB)": 111.15, + "step": 45695, + "train_speed(iter/s)": 0.447738 + }, + { + "acc": 0.66317778, + "epoch": 1.1593099949264332, + "grad_norm": 5.28125, + "learning_rate": 4.101208333038181e-06, + "loss": 1.53769169, + "memory(GiB)": 111.15, + "step": 45700, + "train_speed(iter/s)": 0.447764 + }, + { + "epoch": 1.1593099949264332, + "eval_acc": 0.6432935604929153, + "eval_loss": 1.6078063249588013, + "eval_runtime": 113.6306, + "eval_samples_per_second": 56.059, + "eval_steps_per_second": 28.029, + "step": 45700 + }, + { + "acc": 0.64994173, + "epoch": 1.1594368340943684, + "grad_norm": 6.15625, + "learning_rate": 4.1001768103717285e-06, + "loss": 1.64268684, + "memory(GiB)": 111.15, + "step": 45705, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.64308176, + "epoch": 1.1595636732623034, + "grad_norm": 5.28125, + "learning_rate": 4.09914532728352e-06, + "loss": 1.55346394, + "memory(GiB)": 111.15, + "step": 45710, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.63947086, + "epoch": 1.1596905124302386, + "grad_norm": 5.875, + "learning_rate": 4.09811388381892e-06, + "loss": 1.67384262, + "memory(GiB)": 111.15, + "step": 45715, + "train_speed(iter/s)": 0.447323 + }, + { + "acc": 0.65748339, + "epoch": 1.1598173515981736, + "grad_norm": 6.0625, + "learning_rate": 4.097082480023298e-06, + "loss": 1.62435818, + "memory(GiB)": 111.15, + "step": 45720, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.6362113, + "epoch": 1.1599441907661086, + "grad_norm": 4.96875, + "learning_rate": 4.09605111594202e-06, + "loss": 1.67507992, + "memory(GiB)": 111.15, + "step": 45725, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.65049434, + "epoch": 1.1600710299340435, + "grad_norm": 5.0625, + "learning_rate": 4.095019791620451e-06, + "loss": 1.62788467, + "memory(GiB)": 111.15, + "step": 45730, + "train_speed(iter/s)": 0.447399 + }, + { + "acc": 0.64847665, + "epoch": 1.1601978691019788, + "grad_norm": 5.0625, + "learning_rate": 4.093988507103951e-06, + "loss": 1.60939465, + "memory(GiB)": 111.15, + "step": 45735, + "train_speed(iter/s)": 0.447425 + }, + { + "acc": 0.65839224, + "epoch": 1.1603247082699137, + "grad_norm": 5.71875, + "learning_rate": 4.09295726243788e-06, + "loss": 1.55846262, + "memory(GiB)": 111.15, + "step": 45740, + "train_speed(iter/s)": 0.447451 + }, + { + "acc": 0.66410222, + "epoch": 1.1604515474378487, + "grad_norm": 6.9375, + "learning_rate": 4.091926057667601e-06, + "loss": 1.62758484, + "memory(GiB)": 111.15, + "step": 45745, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.64161997, + "epoch": 1.160578386605784, + "grad_norm": 5.90625, + "learning_rate": 4.0908948928384675e-06, + "loss": 1.58707008, + "memory(GiB)": 111.15, + "step": 45750, + "train_speed(iter/s)": 0.447502 + }, + { + "acc": 0.66184263, + "epoch": 1.160705225773719, + "grad_norm": 5.8125, + "learning_rate": 4.089863767995835e-06, + "loss": 1.6365799, + "memory(GiB)": 111.15, + "step": 45755, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.64182158, + "epoch": 1.160832064941654, + "grad_norm": 5.09375, + "learning_rate": 4.088832683185057e-06, + "loss": 1.62642612, + "memory(GiB)": 111.15, + "step": 45760, + "train_speed(iter/s)": 0.447554 + }, + { + "acc": 0.63541751, + "epoch": 1.1609589041095891, + "grad_norm": 6.0625, + "learning_rate": 4.087801638451485e-06, + "loss": 1.6716898, + "memory(GiB)": 111.15, + "step": 45765, + "train_speed(iter/s)": 0.44758 + }, + { + "acc": 0.65397739, + "epoch": 1.1610857432775241, + "grad_norm": 4.75, + "learning_rate": 4.086770633840472e-06, + "loss": 1.61799469, + "memory(GiB)": 111.15, + "step": 45770, + "train_speed(iter/s)": 0.447605 + }, + { + "acc": 0.64849253, + "epoch": 1.161212582445459, + "grad_norm": 6.1875, + "learning_rate": 4.085739669397362e-06, + "loss": 1.61183987, + "memory(GiB)": 111.15, + "step": 45775, + "train_speed(iter/s)": 0.447631 + }, + { + "acc": 0.65369635, + "epoch": 1.1613394216133943, + "grad_norm": 4.34375, + "learning_rate": 4.084708745167504e-06, + "loss": 1.60100403, + "memory(GiB)": 111.15, + "step": 45780, + "train_speed(iter/s)": 0.447657 + }, + { + "acc": 0.63951945, + "epoch": 1.1614662607813293, + "grad_norm": 5.40625, + "learning_rate": 4.08367786119624e-06, + "loss": 1.73204765, + "memory(GiB)": 111.15, + "step": 45785, + "train_speed(iter/s)": 0.447683 + }, + { + "acc": 0.66286597, + "epoch": 1.1615930999492643, + "grad_norm": 6.09375, + "learning_rate": 4.082647017528918e-06, + "loss": 1.62752571, + "memory(GiB)": 111.15, + "step": 45790, + "train_speed(iter/s)": 0.447709 + }, + { + "acc": 0.64861603, + "epoch": 1.1617199391171993, + "grad_norm": 6.6875, + "learning_rate": 4.081616214210874e-06, + "loss": 1.64290352, + "memory(GiB)": 111.15, + "step": 45795, + "train_speed(iter/s)": 0.447734 + }, + { + "acc": 0.65319643, + "epoch": 1.1618467782851345, + "grad_norm": 4.65625, + "learning_rate": 4.0805854512874485e-06, + "loss": 1.61591434, + "memory(GiB)": 111.15, + "step": 45800, + "train_speed(iter/s)": 0.44776 + }, + { + "epoch": 1.1618467782851345, + "eval_acc": 0.6434848979206468, + "eval_loss": 1.6077747344970703, + "eval_runtime": 114.6564, + "eval_samples_per_second": 55.557, + "eval_steps_per_second": 27.779, + "step": 45800 + }, + { + "acc": 0.65668774, + "epoch": 1.1619736174530695, + "grad_norm": 4.8125, + "learning_rate": 4.079554728803981e-06, + "loss": 1.59205494, + "memory(GiB)": 111.15, + "step": 45805, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.65983052, + "epoch": 1.1621004566210045, + "grad_norm": 5.84375, + "learning_rate": 4.078524046805806e-06, + "loss": 1.6224741, + "memory(GiB)": 111.15, + "step": 45810, + "train_speed(iter/s)": 0.44729 + }, + { + "acc": 0.64889336, + "epoch": 1.1622272957889397, + "grad_norm": 6.125, + "learning_rate": 4.0774934053382576e-06, + "loss": 1.60647774, + "memory(GiB)": 111.15, + "step": 45815, + "train_speed(iter/s)": 0.447316 + }, + { + "acc": 0.66649923, + "epoch": 1.1623541349568747, + "grad_norm": 5.625, + "learning_rate": 4.076462804446667e-06, + "loss": 1.6184082, + "memory(GiB)": 111.15, + "step": 45820, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.64974861, + "epoch": 1.1624809741248097, + "grad_norm": 6.1875, + "learning_rate": 4.0754322441763654e-06, + "loss": 1.66992378, + "memory(GiB)": 111.15, + "step": 45825, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.64464083, + "epoch": 1.1626078132927449, + "grad_norm": 4.75, + "learning_rate": 4.0744017245726834e-06, + "loss": 1.66198196, + "memory(GiB)": 111.15, + "step": 45830, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.64075418, + "epoch": 1.1627346524606799, + "grad_norm": 5.65625, + "learning_rate": 4.073371245680944e-06, + "loss": 1.66034889, + "memory(GiB)": 111.15, + "step": 45835, + "train_speed(iter/s)": 0.447419 + }, + { + "acc": 0.62818413, + "epoch": 1.1628614916286149, + "grad_norm": 5.15625, + "learning_rate": 4.0723408075464754e-06, + "loss": 1.69263763, + "memory(GiB)": 111.15, + "step": 45840, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.63677998, + "epoch": 1.16298833079655, + "grad_norm": 5.5625, + "learning_rate": 4.071310410214598e-06, + "loss": 1.67338676, + "memory(GiB)": 111.15, + "step": 45845, + "train_speed(iter/s)": 0.447471 + }, + { + "acc": 0.64223471, + "epoch": 1.163115169964485, + "grad_norm": 4.78125, + "learning_rate": 4.070280053730639e-06, + "loss": 1.71758041, + "memory(GiB)": 111.15, + "step": 45850, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.65228548, + "epoch": 1.16324200913242, + "grad_norm": 5.5625, + "learning_rate": 4.069249738139911e-06, + "loss": 1.62090607, + "memory(GiB)": 111.15, + "step": 45855, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64800463, + "epoch": 1.163368848300355, + "grad_norm": 6.4375, + "learning_rate": 4.068219463487736e-06, + "loss": 1.63394356, + "memory(GiB)": 111.15, + "step": 45860, + "train_speed(iter/s)": 0.447549 + }, + { + "acc": 0.66189575, + "epoch": 1.1634956874682902, + "grad_norm": 5.21875, + "learning_rate": 4.0671892298194286e-06, + "loss": 1.61612167, + "memory(GiB)": 111.15, + "step": 45865, + "train_speed(iter/s)": 0.447574 + }, + { + "acc": 0.65849876, + "epoch": 1.1636225266362252, + "grad_norm": 5.03125, + "learning_rate": 4.066159037180304e-06, + "loss": 1.60366001, + "memory(GiB)": 111.15, + "step": 45870, + "train_speed(iter/s)": 0.4476 + }, + { + "acc": 0.65422478, + "epoch": 1.1637493658041604, + "grad_norm": 5.21875, + "learning_rate": 4.065128885615674e-06, + "loss": 1.62468624, + "memory(GiB)": 111.15, + "step": 45875, + "train_speed(iter/s)": 0.447626 + }, + { + "acc": 0.62929687, + "epoch": 1.1638762049720954, + "grad_norm": 5.0, + "learning_rate": 4.064098775170849e-06, + "loss": 1.65980148, + "memory(GiB)": 111.15, + "step": 45880, + "train_speed(iter/s)": 0.447652 + }, + { + "acc": 0.64589415, + "epoch": 1.1640030441400304, + "grad_norm": 5.5, + "learning_rate": 4.063068705891139e-06, + "loss": 1.6472477, + "memory(GiB)": 111.15, + "step": 45885, + "train_speed(iter/s)": 0.447678 + }, + { + "acc": 0.643749, + "epoch": 1.1641298833079654, + "grad_norm": 4.5625, + "learning_rate": 4.062038677821852e-06, + "loss": 1.63458862, + "memory(GiB)": 111.15, + "step": 45890, + "train_speed(iter/s)": 0.447703 + }, + { + "acc": 0.65820608, + "epoch": 1.1642567224759006, + "grad_norm": 5.40625, + "learning_rate": 4.061008691008289e-06, + "loss": 1.61216373, + "memory(GiB)": 111.15, + "step": 45895, + "train_speed(iter/s)": 0.447729 + }, + { + "acc": 0.65834498, + "epoch": 1.1643835616438356, + "grad_norm": 5.125, + "learning_rate": 4.059978745495757e-06, + "loss": 1.58290539, + "memory(GiB)": 111.15, + "step": 45900, + "train_speed(iter/s)": 0.447755 + }, + { + "epoch": 1.1643835616438356, + "eval_acc": 0.643369176375927, + "eval_loss": 1.6077266931533813, + "eval_runtime": 114.6908, + "eval_samples_per_second": 55.541, + "eval_steps_per_second": 27.77, + "step": 45900 + }, + { + "acc": 0.65442433, + "epoch": 1.1645104008117706, + "grad_norm": 4.4375, + "learning_rate": 4.058948841329557e-06, + "loss": 1.60035591, + "memory(GiB)": 111.15, + "step": 45905, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.64542093, + "epoch": 1.1646372399797058, + "grad_norm": 5.375, + "learning_rate": 4.057918978554989e-06, + "loss": 1.71731987, + "memory(GiB)": 111.15, + "step": 45910, + "train_speed(iter/s)": 0.447286 + }, + { + "acc": 0.66465864, + "epoch": 1.1647640791476408, + "grad_norm": 5.40625, + "learning_rate": 4.056889157217348e-06, + "loss": 1.57858686, + "memory(GiB)": 111.15, + "step": 45915, + "train_speed(iter/s)": 0.447312 + }, + { + "acc": 0.63360643, + "epoch": 1.1648909183155758, + "grad_norm": 5.96875, + "learning_rate": 4.0558593773619346e-06, + "loss": 1.6993248, + "memory(GiB)": 111.15, + "step": 45920, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.63893189, + "epoch": 1.165017757483511, + "grad_norm": 5.8125, + "learning_rate": 4.05482963903404e-06, + "loss": 1.64424839, + "memory(GiB)": 111.15, + "step": 45925, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.6455472, + "epoch": 1.165144596651446, + "grad_norm": 5.03125, + "learning_rate": 4.05379994227896e-06, + "loss": 1.70573654, + "memory(GiB)": 111.15, + "step": 45930, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.63540254, + "epoch": 1.165271435819381, + "grad_norm": 6.0, + "learning_rate": 4.052770287141981e-06, + "loss": 1.65506001, + "memory(GiB)": 111.15, + "step": 45935, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.65357742, + "epoch": 1.1653982749873162, + "grad_norm": 5.78125, + "learning_rate": 4.051740673668393e-06, + "loss": 1.52992878, + "memory(GiB)": 111.15, + "step": 45940, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.65047092, + "epoch": 1.1655251141552512, + "grad_norm": 5.46875, + "learning_rate": 4.0507111019034855e-06, + "loss": 1.63694744, + "memory(GiB)": 111.15, + "step": 45945, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.64717169, + "epoch": 1.1656519533231862, + "grad_norm": 5.75, + "learning_rate": 4.049681571892543e-06, + "loss": 1.63361263, + "memory(GiB)": 111.15, + "step": 45950, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.6507236, + "epoch": 1.1657787924911212, + "grad_norm": 5.0625, + "learning_rate": 4.048652083680847e-06, + "loss": 1.62778625, + "memory(GiB)": 111.15, + "step": 45955, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.63706827, + "epoch": 1.1659056316590564, + "grad_norm": 4.90625, + "learning_rate": 4.047622637313678e-06, + "loss": 1.74024296, + "memory(GiB)": 111.15, + "step": 45960, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.65976381, + "epoch": 1.1660324708269914, + "grad_norm": 5.25, + "learning_rate": 4.046593232836319e-06, + "loss": 1.59403715, + "memory(GiB)": 111.15, + "step": 45965, + "train_speed(iter/s)": 0.44757 + }, + { + "acc": 0.65821991, + "epoch": 1.1661593099949263, + "grad_norm": 5.34375, + "learning_rate": 4.045563870294047e-06, + "loss": 1.65757065, + "memory(GiB)": 111.15, + "step": 45970, + "train_speed(iter/s)": 0.447596 + }, + { + "acc": 0.66371403, + "epoch": 1.1662861491628616, + "grad_norm": 5.71875, + "learning_rate": 4.044534549732135e-06, + "loss": 1.51997395, + "memory(GiB)": 111.15, + "step": 45975, + "train_speed(iter/s)": 0.447621 + }, + { + "acc": 0.66796732, + "epoch": 1.1664129883307965, + "grad_norm": 5.34375, + "learning_rate": 4.043505271195861e-06, + "loss": 1.54382343, + "memory(GiB)": 111.15, + "step": 45980, + "train_speed(iter/s)": 0.447647 + }, + { + "acc": 0.65047317, + "epoch": 1.1665398274987315, + "grad_norm": 6.375, + "learning_rate": 4.042476034730494e-06, + "loss": 1.68490982, + "memory(GiB)": 111.15, + "step": 45985, + "train_speed(iter/s)": 0.447672 + }, + { + "acc": 0.64315152, + "epoch": 1.1666666666666667, + "grad_norm": 5.65625, + "learning_rate": 4.041446840381309e-06, + "loss": 1.69502392, + "memory(GiB)": 111.15, + "step": 45990, + "train_speed(iter/s)": 0.447698 + }, + { + "acc": 0.65586853, + "epoch": 1.1667935058346017, + "grad_norm": 5.21875, + "learning_rate": 4.040417688193569e-06, + "loss": 1.608494, + "memory(GiB)": 111.15, + "step": 45995, + "train_speed(iter/s)": 0.447723 + }, + { + "acc": 0.65223169, + "epoch": 1.1669203450025367, + "grad_norm": 5.3125, + "learning_rate": 4.039388578212545e-06, + "loss": 1.59654255, + "memory(GiB)": 111.15, + "step": 46000, + "train_speed(iter/s)": 0.447749 + }, + { + "epoch": 1.1669203450025367, + "eval_acc": 0.6433533012181677, + "eval_loss": 1.6076205968856812, + "eval_runtime": 114.8764, + "eval_samples_per_second": 55.451, + "eval_steps_per_second": 27.725, + "step": 46000 + }, + { + "acc": 0.65129728, + "epoch": 1.167047184170472, + "grad_norm": 5.65625, + "learning_rate": 4.0383595104834975e-06, + "loss": 1.62004433, + "memory(GiB)": 111.15, + "step": 46005, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.64594607, + "epoch": 1.167174023338407, + "grad_norm": 5.59375, + "learning_rate": 4.037330485051695e-06, + "loss": 1.64879303, + "memory(GiB)": 111.15, + "step": 46010, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.63482361, + "epoch": 1.167300862506342, + "grad_norm": 4.375, + "learning_rate": 4.0363015019623955e-06, + "loss": 1.63845749, + "memory(GiB)": 111.15, + "step": 46015, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.661023, + "epoch": 1.167427701674277, + "grad_norm": 4.5625, + "learning_rate": 4.0352725612608565e-06, + "loss": 1.59599619, + "memory(GiB)": 111.15, + "step": 46020, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.64989014, + "epoch": 1.167554540842212, + "grad_norm": 6.0, + "learning_rate": 4.0342436629923385e-06, + "loss": 1.64050159, + "memory(GiB)": 111.15, + "step": 46025, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.65782065, + "epoch": 1.167681380010147, + "grad_norm": 5.375, + "learning_rate": 4.033214807202098e-06, + "loss": 1.61310482, + "memory(GiB)": 111.15, + "step": 46030, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.65424347, + "epoch": 1.1678082191780823, + "grad_norm": 5.625, + "learning_rate": 4.032185993935385e-06, + "loss": 1.63646259, + "memory(GiB)": 111.15, + "step": 46035, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.65436678, + "epoch": 1.1679350583460173, + "grad_norm": 5.875, + "learning_rate": 4.031157223237452e-06, + "loss": 1.56024246, + "memory(GiB)": 111.15, + "step": 46040, + "train_speed(iter/s)": 0.447432 + }, + { + "acc": 0.665942, + "epoch": 1.1680618975139523, + "grad_norm": 5.5625, + "learning_rate": 4.0301284951535504e-06, + "loss": 1.54558783, + "memory(GiB)": 111.15, + "step": 46045, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.66312885, + "epoch": 1.1681887366818873, + "grad_norm": 5.03125, + "learning_rate": 4.029099809728929e-06, + "loss": 1.62627144, + "memory(GiB)": 111.15, + "step": 46050, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64464793, + "epoch": 1.1683155758498225, + "grad_norm": 5.71875, + "learning_rate": 4.028071167008831e-06, + "loss": 1.6737175, + "memory(GiB)": 111.15, + "step": 46055, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.65565872, + "epoch": 1.1684424150177575, + "grad_norm": 5.28125, + "learning_rate": 4.027042567038503e-06, + "loss": 1.58504276, + "memory(GiB)": 111.15, + "step": 46060, + "train_speed(iter/s)": 0.447535 + }, + { + "acc": 0.65335665, + "epoch": 1.1685692541856925, + "grad_norm": 4.59375, + "learning_rate": 4.026014009863186e-06, + "loss": 1.65901375, + "memory(GiB)": 111.15, + "step": 46065, + "train_speed(iter/s)": 0.44756 + }, + { + "acc": 0.65064082, + "epoch": 1.1686960933536277, + "grad_norm": 5.59375, + "learning_rate": 4.024985495528124e-06, + "loss": 1.63496971, + "memory(GiB)": 111.15, + "step": 46070, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.65284586, + "epoch": 1.1688229325215627, + "grad_norm": 5.5625, + "learning_rate": 4.023957024078552e-06, + "loss": 1.64229202, + "memory(GiB)": 111.15, + "step": 46075, + "train_speed(iter/s)": 0.447611 + }, + { + "acc": 0.65518007, + "epoch": 1.1689497716894977, + "grad_norm": 6.53125, + "learning_rate": 4.022928595559707e-06, + "loss": 1.56659632, + "memory(GiB)": 111.15, + "step": 46080, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.65121336, + "epoch": 1.1690766108574329, + "grad_norm": 5.28125, + "learning_rate": 4.021900210016824e-06, + "loss": 1.64282799, + "memory(GiB)": 111.15, + "step": 46085, + "train_speed(iter/s)": 0.447662 + }, + { + "acc": 0.65985565, + "epoch": 1.1692034500253679, + "grad_norm": 5.25, + "learning_rate": 4.020871867495139e-06, + "loss": 1.56763611, + "memory(GiB)": 111.15, + "step": 46090, + "train_speed(iter/s)": 0.447688 + }, + { + "acc": 0.66572523, + "epoch": 1.1693302891933028, + "grad_norm": 5.5, + "learning_rate": 4.01984356803988e-06, + "loss": 1.58113136, + "memory(GiB)": 111.15, + "step": 46095, + "train_speed(iter/s)": 0.447713 + }, + { + "acc": 0.65982389, + "epoch": 1.169457128361238, + "grad_norm": 4.9375, + "learning_rate": 4.018815311696274e-06, + "loss": 1.61116714, + "memory(GiB)": 111.15, + "step": 46100, + "train_speed(iter/s)": 0.447739 + }, + { + "epoch": 1.169457128361238, + "eval_acc": 0.6433729362817122, + "eval_loss": 1.6077944040298462, + "eval_runtime": 116.3192, + "eval_samples_per_second": 54.763, + "eval_steps_per_second": 27.382, + "step": 46100 + }, + { + "acc": 0.64126377, + "epoch": 1.169583967529173, + "grad_norm": 6.53125, + "learning_rate": 4.017787098509555e-06, + "loss": 1.63943253, + "memory(GiB)": 111.15, + "step": 46105, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.64962978, + "epoch": 1.169710806697108, + "grad_norm": 5.3125, + "learning_rate": 4.016758928524944e-06, + "loss": 1.59499207, + "memory(GiB)": 111.15, + "step": 46110, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.66243248, + "epoch": 1.169837645865043, + "grad_norm": 5.21875, + "learning_rate": 4.015730801787663e-06, + "loss": 1.53319874, + "memory(GiB)": 111.15, + "step": 46115, + "train_speed(iter/s)": 0.44729 + }, + { + "acc": 0.63749385, + "epoch": 1.1699644850329782, + "grad_norm": 5.25, + "learning_rate": 4.014702718342938e-06, + "loss": 1.6792057, + "memory(GiB)": 111.15, + "step": 46120, + "train_speed(iter/s)": 0.447316 + }, + { + "acc": 0.64103279, + "epoch": 1.1700913242009132, + "grad_norm": 5.125, + "learning_rate": 4.013674678235985e-06, + "loss": 1.61349277, + "memory(GiB)": 111.15, + "step": 46125, + "train_speed(iter/s)": 0.447341 + }, + { + "acc": 0.64736233, + "epoch": 1.1702181633688482, + "grad_norm": 5.25, + "learning_rate": 4.012646681512026e-06, + "loss": 1.6162899, + "memory(GiB)": 111.15, + "step": 46130, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.65025172, + "epoch": 1.1703450025367834, + "grad_norm": 5.46875, + "learning_rate": 4.011618728216271e-06, + "loss": 1.69266548, + "memory(GiB)": 111.15, + "step": 46135, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.64566078, + "epoch": 1.1704718417047184, + "grad_norm": 5.5, + "learning_rate": 4.010590818393938e-06, + "loss": 1.68417892, + "memory(GiB)": 111.15, + "step": 46140, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.67257004, + "epoch": 1.1705986808726534, + "grad_norm": 5.59375, + "learning_rate": 4.009562952090238e-06, + "loss": 1.58950996, + "memory(GiB)": 111.15, + "step": 46145, + "train_speed(iter/s)": 0.447444 + }, + { + "acc": 0.64861879, + "epoch": 1.1707255200405886, + "grad_norm": 5.15625, + "learning_rate": 4.0085351293503825e-06, + "loss": 1.54161606, + "memory(GiB)": 111.15, + "step": 46150, + "train_speed(iter/s)": 0.44747 + }, + { + "acc": 0.64405947, + "epoch": 1.1708523592085236, + "grad_norm": 4.34375, + "learning_rate": 4.007507350219578e-06, + "loss": 1.70124702, + "memory(GiB)": 111.15, + "step": 46155, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.65996575, + "epoch": 1.1709791983764586, + "grad_norm": 5.25, + "learning_rate": 4.0064796147430305e-06, + "loss": 1.55694695, + "memory(GiB)": 111.15, + "step": 46160, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.65849524, + "epoch": 1.1711060375443938, + "grad_norm": 5.21875, + "learning_rate": 4.005451922965946e-06, + "loss": 1.63474541, + "memory(GiB)": 111.15, + "step": 46165, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.65836525, + "epoch": 1.1712328767123288, + "grad_norm": 4.875, + "learning_rate": 4.0044242749335285e-06, + "loss": 1.57953014, + "memory(GiB)": 111.15, + "step": 46170, + "train_speed(iter/s)": 0.447573 + }, + { + "acc": 0.64201012, + "epoch": 1.1713597158802638, + "grad_norm": 4.90625, + "learning_rate": 4.0033966706909735e-06, + "loss": 1.70995865, + "memory(GiB)": 111.15, + "step": 46175, + "train_speed(iter/s)": 0.447598 + }, + { + "acc": 0.6452179, + "epoch": 1.1714865550481988, + "grad_norm": 5.375, + "learning_rate": 4.002369110283482e-06, + "loss": 1.61254444, + "memory(GiB)": 111.15, + "step": 46180, + "train_speed(iter/s)": 0.447624 + }, + { + "acc": 0.63404064, + "epoch": 1.171613394216134, + "grad_norm": 5.0, + "learning_rate": 4.001341593756253e-06, + "loss": 1.66566048, + "memory(GiB)": 111.15, + "step": 46185, + "train_speed(iter/s)": 0.447649 + }, + { + "acc": 0.65483942, + "epoch": 1.171740233384069, + "grad_norm": 6.6875, + "learning_rate": 4.00031412115448e-06, + "loss": 1.65460205, + "memory(GiB)": 111.15, + "step": 46190, + "train_speed(iter/s)": 0.447675 + }, + { + "acc": 0.65235558, + "epoch": 1.1718670725520042, + "grad_norm": 5.5, + "learning_rate": 3.999286692523352e-06, + "loss": 1.5821825, + "memory(GiB)": 111.15, + "step": 46195, + "train_speed(iter/s)": 0.447701 + }, + { + "acc": 0.66271048, + "epoch": 1.1719939117199392, + "grad_norm": 4.71875, + "learning_rate": 3.998259307908065e-06, + "loss": 1.55272045, + "memory(GiB)": 111.15, + "step": 46200, + "train_speed(iter/s)": 0.447726 + }, + { + "epoch": 1.1719939117199392, + "eval_acc": 0.6433996733895174, + "eval_loss": 1.6077109575271606, + "eval_runtime": 115.1298, + "eval_samples_per_second": 55.329, + "eval_steps_per_second": 27.664, + "step": 46200 + }, + { + "acc": 0.64018459, + "epoch": 1.1721207508878742, + "grad_norm": 6.21875, + "learning_rate": 3.997231967353806e-06, + "loss": 1.63310089, + "memory(GiB)": 111.15, + "step": 46205, + "train_speed(iter/s)": 0.447233 + }, + { + "acc": 0.64996023, + "epoch": 1.1722475900558091, + "grad_norm": 4.875, + "learning_rate": 3.996204670905765e-06, + "loss": 1.5922576, + "memory(GiB)": 111.15, + "step": 46210, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.65717735, + "epoch": 1.1723744292237444, + "grad_norm": 5.09375, + "learning_rate": 3.9951774186091195e-06, + "loss": 1.61988792, + "memory(GiB)": 111.15, + "step": 46215, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.67146969, + "epoch": 1.1725012683916793, + "grad_norm": 6.1875, + "learning_rate": 3.9941502105090594e-06, + "loss": 1.55701609, + "memory(GiB)": 111.15, + "step": 46220, + "train_speed(iter/s)": 0.44731 + }, + { + "acc": 0.67363214, + "epoch": 1.1726281075596143, + "grad_norm": 4.9375, + "learning_rate": 3.9931230466507634e-06, + "loss": 1.53920803, + "memory(GiB)": 111.15, + "step": 46225, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.65837116, + "epoch": 1.1727549467275495, + "grad_norm": 4.9375, + "learning_rate": 3.992095927079412e-06, + "loss": 1.65090275, + "memory(GiB)": 111.15, + "step": 46230, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.66441689, + "epoch": 1.1728817858954845, + "grad_norm": 6.25, + "learning_rate": 3.991068851840182e-06, + "loss": 1.55966234, + "memory(GiB)": 111.15, + "step": 46235, + "train_speed(iter/s)": 0.447387 + }, + { + "acc": 0.66298695, + "epoch": 1.1730086250634195, + "grad_norm": 5.6875, + "learning_rate": 3.990041820978246e-06, + "loss": 1.57655296, + "memory(GiB)": 111.15, + "step": 46240, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.66457729, + "epoch": 1.1731354642313547, + "grad_norm": 6.0, + "learning_rate": 3.989014834538782e-06, + "loss": 1.56523371, + "memory(GiB)": 111.15, + "step": 46245, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.66084538, + "epoch": 1.1732623033992897, + "grad_norm": 6.46875, + "learning_rate": 3.987987892566959e-06, + "loss": 1.62080421, + "memory(GiB)": 111.15, + "step": 46250, + "train_speed(iter/s)": 0.447463 + }, + { + "acc": 0.66980453, + "epoch": 1.1733891425672247, + "grad_norm": 4.5625, + "learning_rate": 3.986960995107948e-06, + "loss": 1.58061581, + "memory(GiB)": 111.15, + "step": 46255, + "train_speed(iter/s)": 0.447489 + }, + { + "acc": 0.65931501, + "epoch": 1.17351598173516, + "grad_norm": 5.0, + "learning_rate": 3.985934142206912e-06, + "loss": 1.61051922, + "memory(GiB)": 111.15, + "step": 46260, + "train_speed(iter/s)": 0.447514 + }, + { + "acc": 0.65506196, + "epoch": 1.173642820903095, + "grad_norm": 5.03125, + "learning_rate": 3.984907333909022e-06, + "loss": 1.619557, + "memory(GiB)": 111.15, + "step": 46265, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.64671917, + "epoch": 1.17376966007103, + "grad_norm": 5.46875, + "learning_rate": 3.983880570259441e-06, + "loss": 1.65152912, + "memory(GiB)": 111.15, + "step": 46270, + "train_speed(iter/s)": 0.447565 + }, + { + "acc": 0.65532055, + "epoch": 1.1738964992389649, + "grad_norm": 5.65625, + "learning_rate": 3.982853851303327e-06, + "loss": 1.5661068, + "memory(GiB)": 111.15, + "step": 46275, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.63184781, + "epoch": 1.1740233384069, + "grad_norm": 6.125, + "learning_rate": 3.981827177085842e-06, + "loss": 1.68110352, + "memory(GiB)": 111.15, + "step": 46280, + "train_speed(iter/s)": 0.447617 + }, + { + "acc": 0.64716177, + "epoch": 1.174150177574835, + "grad_norm": 5.90625, + "learning_rate": 3.980800547652143e-06, + "loss": 1.62350597, + "memory(GiB)": 111.15, + "step": 46285, + "train_speed(iter/s)": 0.447642 + }, + { + "acc": 0.63062458, + "epoch": 1.17427701674277, + "grad_norm": 5.0625, + "learning_rate": 3.979773963047388e-06, + "loss": 1.7672123, + "memory(GiB)": 111.15, + "step": 46290, + "train_speed(iter/s)": 0.447668 + }, + { + "acc": 0.65572424, + "epoch": 1.1744038559107053, + "grad_norm": 5.375, + "learning_rate": 3.978747423316729e-06, + "loss": 1.57507458, + "memory(GiB)": 111.15, + "step": 46295, + "train_speed(iter/s)": 0.447693 + }, + { + "acc": 0.64159074, + "epoch": 1.1745306950786403, + "grad_norm": 5.03125, + "learning_rate": 3.977720928505317e-06, + "loss": 1.67647572, + "memory(GiB)": 111.15, + "step": 46300, + "train_speed(iter/s)": 0.447719 + }, + { + "epoch": 1.1745306950786403, + "eval_acc": 0.6433378438277177, + "eval_loss": 1.6077017784118652, + "eval_runtime": 114.1271, + "eval_samples_per_second": 55.815, + "eval_steps_per_second": 27.907, + "step": 46300 + }, + { + "acc": 0.64731064, + "epoch": 1.1746575342465753, + "grad_norm": 7.0, + "learning_rate": 3.976694478658301e-06, + "loss": 1.6575676, + "memory(GiB)": 111.15, + "step": 46305, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.65658069, + "epoch": 1.1747843734145105, + "grad_norm": 6.21875, + "learning_rate": 3.975668073820834e-06, + "loss": 1.60651627, + "memory(GiB)": 111.15, + "step": 46310, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.67964954, + "epoch": 1.1749112125824455, + "grad_norm": 6.1875, + "learning_rate": 3.9746417140380576e-06, + "loss": 1.49111948, + "memory(GiB)": 111.15, + "step": 46315, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.64998856, + "epoch": 1.1750380517503805, + "grad_norm": 6.8125, + "learning_rate": 3.973615399355114e-06, + "loss": 1.65594902, + "memory(GiB)": 111.15, + "step": 46320, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.6516017, + "epoch": 1.1751648909183157, + "grad_norm": 5.0, + "learning_rate": 3.97258912981715e-06, + "loss": 1.65487804, + "memory(GiB)": 111.15, + "step": 46325, + "train_speed(iter/s)": 0.447332 + }, + { + "acc": 0.65798597, + "epoch": 1.1752917300862507, + "grad_norm": 5.6875, + "learning_rate": 3.9715629054693035e-06, + "loss": 1.56240063, + "memory(GiB)": 111.15, + "step": 46330, + "train_speed(iter/s)": 0.447358 + }, + { + "acc": 0.65895128, + "epoch": 1.1754185692541856, + "grad_norm": 5.21875, + "learning_rate": 3.970536726356711e-06, + "loss": 1.59832468, + "memory(GiB)": 111.15, + "step": 46335, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.64337683, + "epoch": 1.1755454084221206, + "grad_norm": 4.46875, + "learning_rate": 3.969510592524509e-06, + "loss": 1.64710369, + "memory(GiB)": 111.15, + "step": 46340, + "train_speed(iter/s)": 0.447409 + }, + { + "acc": 0.65196567, + "epoch": 1.1756722475900558, + "grad_norm": 4.71875, + "learning_rate": 3.968484504017833e-06, + "loss": 1.65350075, + "memory(GiB)": 111.15, + "step": 46345, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.66584568, + "epoch": 1.1757990867579908, + "grad_norm": 6.28125, + "learning_rate": 3.967458460881815e-06, + "loss": 1.58108406, + "memory(GiB)": 111.15, + "step": 46350, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.66394091, + "epoch": 1.175925925925926, + "grad_norm": 5.40625, + "learning_rate": 3.96643246316158e-06, + "loss": 1.56845341, + "memory(GiB)": 111.15, + "step": 46355, + "train_speed(iter/s)": 0.447485 + }, + { + "acc": 0.65722733, + "epoch": 1.176052765093861, + "grad_norm": 4.5625, + "learning_rate": 3.965406510902263e-06, + "loss": 1.56991081, + "memory(GiB)": 111.15, + "step": 46360, + "train_speed(iter/s)": 0.44751 + }, + { + "acc": 0.6644145, + "epoch": 1.176179604261796, + "grad_norm": 5.9375, + "learning_rate": 3.9643806041489855e-06, + "loss": 1.5855032, + "memory(GiB)": 111.15, + "step": 46365, + "train_speed(iter/s)": 0.447536 + }, + { + "acc": 0.64371548, + "epoch": 1.176306443429731, + "grad_norm": 5.125, + "learning_rate": 3.963354742946874e-06, + "loss": 1.61697254, + "memory(GiB)": 111.15, + "step": 46370, + "train_speed(iter/s)": 0.447561 + }, + { + "acc": 0.65071883, + "epoch": 1.1764332825976662, + "grad_norm": 4.34375, + "learning_rate": 3.962328927341048e-06, + "loss": 1.64102936, + "memory(GiB)": 111.15, + "step": 46375, + "train_speed(iter/s)": 0.447587 + }, + { + "acc": 0.66896429, + "epoch": 1.1765601217656012, + "grad_norm": 6.25, + "learning_rate": 3.961303157376628e-06, + "loss": 1.57396908, + "memory(GiB)": 111.15, + "step": 46380, + "train_speed(iter/s)": 0.447612 + }, + { + "acc": 0.64574375, + "epoch": 1.1766869609335362, + "grad_norm": 5.625, + "learning_rate": 3.960277433098734e-06, + "loss": 1.6046608, + "memory(GiB)": 111.15, + "step": 46385, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.6451169, + "epoch": 1.1768138001014714, + "grad_norm": 6.03125, + "learning_rate": 3.959251754552481e-06, + "loss": 1.69671974, + "memory(GiB)": 111.15, + "step": 46390, + "train_speed(iter/s)": 0.447664 + }, + { + "acc": 0.6438715, + "epoch": 1.1769406392694064, + "grad_norm": 4.78125, + "learning_rate": 3.958226121782982e-06, + "loss": 1.64787369, + "memory(GiB)": 111.15, + "step": 46395, + "train_speed(iter/s)": 0.447689 + }, + { + "acc": 0.65169549, + "epoch": 1.1770674784373414, + "grad_norm": 5.1875, + "learning_rate": 3.9572005348353486e-06, + "loss": 1.5928566, + "memory(GiB)": 111.15, + "step": 46400, + "train_speed(iter/s)": 0.447714 + }, + { + "epoch": 1.1770674784373414, + "eval_acc": 0.6434343480762025, + "eval_loss": 1.6076468229293823, + "eval_runtime": 116.1477, + "eval_samples_per_second": 54.844, + "eval_steps_per_second": 27.422, + "step": 46400 + }, + { + "acc": 0.63697977, + "epoch": 1.1771943176052766, + "grad_norm": 6.90625, + "learning_rate": 3.956174993754691e-06, + "loss": 1.66369038, + "memory(GiB)": 111.15, + "step": 46405, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.66791525, + "epoch": 1.1773211567732116, + "grad_norm": 4.4375, + "learning_rate": 3.955149498586119e-06, + "loss": 1.51250553, + "memory(GiB)": 111.15, + "step": 46410, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.64407272, + "epoch": 1.1774479959411466, + "grad_norm": 6.34375, + "learning_rate": 3.954124049374736e-06, + "loss": 1.66300163, + "memory(GiB)": 111.15, + "step": 46415, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.65990496, + "epoch": 1.1775748351090818, + "grad_norm": 4.625, + "learning_rate": 3.9530986461656465e-06, + "loss": 1.54421759, + "memory(GiB)": 111.15, + "step": 46420, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.6535862, + "epoch": 1.1777016742770168, + "grad_norm": 5.8125, + "learning_rate": 3.952073289003953e-06, + "loss": 1.59470539, + "memory(GiB)": 111.15, + "step": 46425, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.66695662, + "epoch": 1.1778285134449518, + "grad_norm": 5.46875, + "learning_rate": 3.9510479779347566e-06, + "loss": 1.54427795, + "memory(GiB)": 111.15, + "step": 46430, + "train_speed(iter/s)": 0.447345 + }, + { + "acc": 0.64840736, + "epoch": 1.1779553526128868, + "grad_norm": 5.46875, + "learning_rate": 3.950022713003151e-06, + "loss": 1.6231411, + "memory(GiB)": 111.15, + "step": 46435, + "train_speed(iter/s)": 0.447371 + }, + { + "acc": 0.64396362, + "epoch": 1.178082191780822, + "grad_norm": 5.15625, + "learning_rate": 3.9489974942542355e-06, + "loss": 1.64093227, + "memory(GiB)": 111.15, + "step": 46440, + "train_speed(iter/s)": 0.447396 + }, + { + "acc": 0.65322843, + "epoch": 1.178209030948757, + "grad_norm": 5.53125, + "learning_rate": 3.947972321733101e-06, + "loss": 1.59906521, + "memory(GiB)": 111.15, + "step": 46445, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.65581274, + "epoch": 1.178335870116692, + "grad_norm": 4.9375, + "learning_rate": 3.946947195484843e-06, + "loss": 1.63855019, + "memory(GiB)": 111.15, + "step": 46450, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.6486804, + "epoch": 1.1784627092846272, + "grad_norm": 5.6875, + "learning_rate": 3.945922115554548e-06, + "loss": 1.65615807, + "memory(GiB)": 111.15, + "step": 46455, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.661484, + "epoch": 1.1785895484525621, + "grad_norm": 4.9375, + "learning_rate": 3.944897081987303e-06, + "loss": 1.6774065, + "memory(GiB)": 111.15, + "step": 46460, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.64812217, + "epoch": 1.1787163876204971, + "grad_norm": 7.40625, + "learning_rate": 3.943872094828197e-06, + "loss": 1.64675961, + "memory(GiB)": 111.15, + "step": 46465, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.66516151, + "epoch": 1.1788432267884323, + "grad_norm": 4.84375, + "learning_rate": 3.942847154122312e-06, + "loss": 1.55543537, + "memory(GiB)": 111.15, + "step": 46470, + "train_speed(iter/s)": 0.44755 + }, + { + "acc": 0.66608601, + "epoch": 1.1789700659563673, + "grad_norm": 6.0, + "learning_rate": 3.941822259914728e-06, + "loss": 1.56402445, + "memory(GiB)": 111.15, + "step": 46475, + "train_speed(iter/s)": 0.447575 + }, + { + "acc": 0.65318666, + "epoch": 1.1790969051243023, + "grad_norm": 4.5, + "learning_rate": 3.940797412250524e-06, + "loss": 1.60932426, + "memory(GiB)": 111.15, + "step": 46480, + "train_speed(iter/s)": 0.447601 + }, + { + "acc": 0.65920687, + "epoch": 1.1792237442922375, + "grad_norm": 4.375, + "learning_rate": 3.93977261117478e-06, + "loss": 1.63443127, + "memory(GiB)": 111.15, + "step": 46485, + "train_speed(iter/s)": 0.447626 + }, + { + "acc": 0.65180326, + "epoch": 1.1793505834601725, + "grad_norm": 4.78125, + "learning_rate": 3.938747856732572e-06, + "loss": 1.61324806, + "memory(GiB)": 111.15, + "step": 46490, + "train_speed(iter/s)": 0.447652 + }, + { + "acc": 0.66408238, + "epoch": 1.1794774226281075, + "grad_norm": 5.40625, + "learning_rate": 3.9377231489689685e-06, + "loss": 1.61795235, + "memory(GiB)": 111.15, + "step": 46495, + "train_speed(iter/s)": 0.447677 + }, + { + "acc": 0.64385667, + "epoch": 1.1796042617960425, + "grad_norm": 5.28125, + "learning_rate": 3.936698487929045e-06, + "loss": 1.63654137, + "memory(GiB)": 111.15, + "step": 46500, + "train_speed(iter/s)": 0.447703 + }, + { + "epoch": 1.1796042617960425, + "eval_acc": 0.6433512123816204, + "eval_loss": 1.6077419519424438, + "eval_runtime": 115.2692, + "eval_samples_per_second": 55.262, + "eval_steps_per_second": 27.631, + "step": 46500 + }, + { + "acc": 0.6648531, + "epoch": 1.1797311009639777, + "grad_norm": 4.84375, + "learning_rate": 3.935673873657868e-06, + "loss": 1.55594215, + "memory(GiB)": 111.15, + "step": 46505, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.64736304, + "epoch": 1.1798579401319127, + "grad_norm": 6.21875, + "learning_rate": 3.934649306200508e-06, + "loss": 1.65629082, + "memory(GiB)": 111.15, + "step": 46510, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.65347881, + "epoch": 1.179984779299848, + "grad_norm": 5.09375, + "learning_rate": 3.933624785602027e-06, + "loss": 1.61072102, + "memory(GiB)": 111.15, + "step": 46515, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.63701611, + "epoch": 1.180111618467783, + "grad_norm": 4.90625, + "learning_rate": 3.932600311907489e-06, + "loss": 1.63558807, + "memory(GiB)": 111.15, + "step": 46520, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.64496632, + "epoch": 1.1802384576357179, + "grad_norm": 4.65625, + "learning_rate": 3.931575885161955e-06, + "loss": 1.65544777, + "memory(GiB)": 111.15, + "step": 46525, + "train_speed(iter/s)": 0.447313 + }, + { + "acc": 0.64037528, + "epoch": 1.1803652968036529, + "grad_norm": 5.25, + "learning_rate": 3.930551505410484e-06, + "loss": 1.62607574, + "memory(GiB)": 111.15, + "step": 46530, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.65760565, + "epoch": 1.180492135971588, + "grad_norm": 5.375, + "learning_rate": 3.929527172698132e-06, + "loss": 1.6124197, + "memory(GiB)": 111.15, + "step": 46535, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.64342122, + "epoch": 1.180618975139523, + "grad_norm": 4.75, + "learning_rate": 3.928502887069954e-06, + "loss": 1.63453407, + "memory(GiB)": 111.15, + "step": 46540, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.67477942, + "epoch": 1.180745814307458, + "grad_norm": 5.09375, + "learning_rate": 3.927478648571003e-06, + "loss": 1.5179882, + "memory(GiB)": 111.15, + "step": 46545, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.64556885, + "epoch": 1.1808726534753933, + "grad_norm": 6.6875, + "learning_rate": 3.926454457246331e-06, + "loss": 1.61301403, + "memory(GiB)": 111.15, + "step": 46550, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.64081244, + "epoch": 1.1809994926433283, + "grad_norm": 5.78125, + "learning_rate": 3.9254303131409834e-06, + "loss": 1.66604939, + "memory(GiB)": 111.15, + "step": 46555, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.64670782, + "epoch": 1.1811263318112633, + "grad_norm": 6.3125, + "learning_rate": 3.924406216300009e-06, + "loss": 1.59214859, + "memory(GiB)": 111.15, + "step": 46560, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.65820322, + "epoch": 1.1812531709791985, + "grad_norm": 5.375, + "learning_rate": 3.92338216676845e-06, + "loss": 1.62463894, + "memory(GiB)": 111.15, + "step": 46565, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.64439669, + "epoch": 1.1813800101471335, + "grad_norm": 5.71875, + "learning_rate": 3.922358164591353e-06, + "loss": 1.68105087, + "memory(GiB)": 111.15, + "step": 46570, + "train_speed(iter/s)": 0.447541 + }, + { + "acc": 0.64659362, + "epoch": 1.1815068493150684, + "grad_norm": 4.90625, + "learning_rate": 3.921334209813752e-06, + "loss": 1.64357643, + "memory(GiB)": 111.15, + "step": 46575, + "train_speed(iter/s)": 0.447566 + }, + { + "acc": 0.64592791, + "epoch": 1.1816336884830037, + "grad_norm": 4.9375, + "learning_rate": 3.92031030248069e-06, + "loss": 1.64956264, + "memory(GiB)": 111.15, + "step": 46580, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.63438368, + "epoch": 1.1817605276509386, + "grad_norm": 5.125, + "learning_rate": 3.9192864426372e-06, + "loss": 1.66991539, + "memory(GiB)": 111.15, + "step": 46585, + "train_speed(iter/s)": 0.447616 + }, + { + "acc": 0.66563625, + "epoch": 1.1818873668188736, + "grad_norm": 5.3125, + "learning_rate": 3.918262630328319e-06, + "loss": 1.55806255, + "memory(GiB)": 111.15, + "step": 46590, + "train_speed(iter/s)": 0.447642 + }, + { + "acc": 0.65755582, + "epoch": 1.1820142059868086, + "grad_norm": 4.65625, + "learning_rate": 3.917238865599077e-06, + "loss": 1.61739502, + "memory(GiB)": 111.15, + "step": 46595, + "train_speed(iter/s)": 0.447667 + }, + { + "acc": 0.67645588, + "epoch": 1.1821410451547438, + "grad_norm": 5.125, + "learning_rate": 3.916215148494502e-06, + "loss": 1.55741835, + "memory(GiB)": 111.15, + "step": 46600, + "train_speed(iter/s)": 0.447692 + }, + { + "epoch": 1.1821410451547438, + "eval_acc": 0.6433340839219327, + "eval_loss": 1.607783555984497, + "eval_runtime": 113.5108, + "eval_samples_per_second": 56.118, + "eval_steps_per_second": 28.059, + "step": 46600 + }, + { + "acc": 0.63433313, + "epoch": 1.1822678843226788, + "grad_norm": 6.15625, + "learning_rate": 3.9151914790596255e-06, + "loss": 1.68751411, + "memory(GiB)": 111.15, + "step": 46605, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.65694666, + "epoch": 1.1823947234906138, + "grad_norm": 5.46875, + "learning_rate": 3.914167857339472e-06, + "loss": 1.59329948, + "memory(GiB)": 111.15, + "step": 46610, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.6377717, + "epoch": 1.182521562658549, + "grad_norm": 5.53125, + "learning_rate": 3.913144283379061e-06, + "loss": 1.65581551, + "memory(GiB)": 111.15, + "step": 46615, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.65438347, + "epoch": 1.182648401826484, + "grad_norm": 5.40625, + "learning_rate": 3.912120757223418e-06, + "loss": 1.68463211, + "memory(GiB)": 111.15, + "step": 46620, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.66205082, + "epoch": 1.182775240994419, + "grad_norm": 5.8125, + "learning_rate": 3.911097278917561e-06, + "loss": 1.60170631, + "memory(GiB)": 111.15, + "step": 46625, + "train_speed(iter/s)": 0.44731 + }, + { + "acc": 0.64891191, + "epoch": 1.1829020801623542, + "grad_norm": 5.71875, + "learning_rate": 3.91007384850651e-06, + "loss": 1.59540462, + "memory(GiB)": 111.15, + "step": 46630, + "train_speed(iter/s)": 0.447335 + }, + { + "acc": 0.65784578, + "epoch": 1.1830289193302892, + "grad_norm": 4.84375, + "learning_rate": 3.909050466035274e-06, + "loss": 1.54751396, + "memory(GiB)": 111.15, + "step": 46635, + "train_speed(iter/s)": 0.44736 + }, + { + "acc": 0.64930005, + "epoch": 1.1831557584982242, + "grad_norm": 4.9375, + "learning_rate": 3.90802713154887e-06, + "loss": 1.61828537, + "memory(GiB)": 111.15, + "step": 46640, + "train_speed(iter/s)": 0.447385 + }, + { + "acc": 0.66440458, + "epoch": 1.1832825976661594, + "grad_norm": 5.34375, + "learning_rate": 3.9070038450923074e-06, + "loss": 1.54139538, + "memory(GiB)": 111.15, + "step": 46645, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.64255705, + "epoch": 1.1834094368340944, + "grad_norm": 5.78125, + "learning_rate": 3.9059806067105985e-06, + "loss": 1.62912636, + "memory(GiB)": 111.15, + "step": 46650, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.64266562, + "epoch": 1.1835362760020294, + "grad_norm": 5.4375, + "learning_rate": 3.904957416448744e-06, + "loss": 1.6719738, + "memory(GiB)": 111.15, + "step": 46655, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.663482, + "epoch": 1.1836631151699644, + "grad_norm": 5.9375, + "learning_rate": 3.903934274351753e-06, + "loss": 1.60834236, + "memory(GiB)": 111.15, + "step": 46660, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.66064787, + "epoch": 1.1837899543378996, + "grad_norm": 5.5625, + "learning_rate": 3.9029111804646245e-06, + "loss": 1.58206253, + "memory(GiB)": 111.15, + "step": 46665, + "train_speed(iter/s)": 0.447511 + }, + { + "acc": 0.64769602, + "epoch": 1.1839167935058346, + "grad_norm": 5.8125, + "learning_rate": 3.9018881348323626e-06, + "loss": 1.61800766, + "memory(GiB)": 111.15, + "step": 46670, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.65959716, + "epoch": 1.1840436326737698, + "grad_norm": 4.5625, + "learning_rate": 3.9008651374999615e-06, + "loss": 1.56911325, + "memory(GiB)": 111.15, + "step": 46675, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.64995155, + "epoch": 1.1841704718417048, + "grad_norm": 4.9375, + "learning_rate": 3.899842188512419e-06, + "loss": 1.59900951, + "memory(GiB)": 111.15, + "step": 46680, + "train_speed(iter/s)": 0.447587 + }, + { + "acc": 0.64757471, + "epoch": 1.1842973110096398, + "grad_norm": 4.5625, + "learning_rate": 3.898819287914729e-06, + "loss": 1.60531654, + "memory(GiB)": 111.15, + "step": 46685, + "train_speed(iter/s)": 0.447612 + }, + { + "acc": 0.64098101, + "epoch": 1.1844241501775747, + "grad_norm": 5.6875, + "learning_rate": 3.897796435751885e-06, + "loss": 1.64190083, + "memory(GiB)": 111.15, + "step": 46690, + "train_speed(iter/s)": 0.447637 + }, + { + "acc": 0.63481112, + "epoch": 1.18455098934551, + "grad_norm": 5.6875, + "learning_rate": 3.896773632068873e-06, + "loss": 1.69057503, + "memory(GiB)": 111.15, + "step": 46695, + "train_speed(iter/s)": 0.447662 + }, + { + "acc": 0.64360027, + "epoch": 1.184677828513445, + "grad_norm": 5.375, + "learning_rate": 3.8957508769106825e-06, + "loss": 1.63173294, + "memory(GiB)": 111.15, + "step": 46700, + "train_speed(iter/s)": 0.447687 + }, + { + "epoch": 1.184677828513445, + "eval_acc": 0.6433741895836406, + "eval_loss": 1.6077251434326172, + "eval_runtime": 114.5728, + "eval_samples_per_second": 55.598, + "eval_steps_per_second": 27.799, + "step": 46700 + }, + { + "acc": 0.65898213, + "epoch": 1.18480466768138, + "grad_norm": 4.625, + "learning_rate": 3.894728170322298e-06, + "loss": 1.62031174, + "memory(GiB)": 111.15, + "step": 46705, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.66822023, + "epoch": 1.1849315068493151, + "grad_norm": 5.1875, + "learning_rate": 3.893705512348705e-06, + "loss": 1.5440136, + "memory(GiB)": 111.15, + "step": 46710, + "train_speed(iter/s)": 0.447226 + }, + { + "acc": 0.66980667, + "epoch": 1.1850583460172501, + "grad_norm": 4.4375, + "learning_rate": 3.89268290303488e-06, + "loss": 1.61800003, + "memory(GiB)": 111.15, + "step": 46715, + "train_speed(iter/s)": 0.447251 + }, + { + "acc": 0.65823202, + "epoch": 1.1851851851851851, + "grad_norm": 5.21875, + "learning_rate": 3.891660342425807e-06, + "loss": 1.53548946, + "memory(GiB)": 111.15, + "step": 46720, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.67416177, + "epoch": 1.1853120243531203, + "grad_norm": 6.1875, + "learning_rate": 3.890637830566459e-06, + "loss": 1.53503933, + "memory(GiB)": 111.15, + "step": 46725, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.6474494, + "epoch": 1.1854388635210553, + "grad_norm": 6.71875, + "learning_rate": 3.889615367501815e-06, + "loss": 1.66917133, + "memory(GiB)": 111.15, + "step": 46730, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.65724311, + "epoch": 1.1855657026889903, + "grad_norm": 4.65625, + "learning_rate": 3.888592953276842e-06, + "loss": 1.62222252, + "memory(GiB)": 111.15, + "step": 46735, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.66495566, + "epoch": 1.1856925418569255, + "grad_norm": 4.625, + "learning_rate": 3.8875705879365135e-06, + "loss": 1.50800276, + "memory(GiB)": 111.15, + "step": 46740, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.65422091, + "epoch": 1.1858193810248605, + "grad_norm": 5.78125, + "learning_rate": 3.886548271525797e-06, + "loss": 1.56123781, + "memory(GiB)": 111.15, + "step": 46745, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.6607801, + "epoch": 1.1859462201927955, + "grad_norm": 5.84375, + "learning_rate": 3.88552600408966e-06, + "loss": 1.55842762, + "memory(GiB)": 111.15, + "step": 46750, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.64826517, + "epoch": 1.1860730593607305, + "grad_norm": 5.15625, + "learning_rate": 3.8845037856730646e-06, + "loss": 1.68134766, + "memory(GiB)": 111.15, + "step": 46755, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.6506875, + "epoch": 1.1861998985286657, + "grad_norm": 5.375, + "learning_rate": 3.883481616320972e-06, + "loss": 1.65179234, + "memory(GiB)": 111.15, + "step": 46760, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.65702763, + "epoch": 1.1863267376966007, + "grad_norm": 5.1875, + "learning_rate": 3.882459496078343e-06, + "loss": 1.64037895, + "memory(GiB)": 111.15, + "step": 46765, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.65741482, + "epoch": 1.1864535768645357, + "grad_norm": 6.9375, + "learning_rate": 3.881437424990137e-06, + "loss": 1.58025188, + "memory(GiB)": 111.15, + "step": 46770, + "train_speed(iter/s)": 0.447529 + }, + { + "acc": 0.65182114, + "epoch": 1.1865804160324709, + "grad_norm": 5.8125, + "learning_rate": 3.880415403101304e-06, + "loss": 1.6149725, + "memory(GiB)": 111.15, + "step": 46775, + "train_speed(iter/s)": 0.447555 + }, + { + "acc": 0.66503105, + "epoch": 1.1867072552004059, + "grad_norm": 5.0625, + "learning_rate": 3.879393430456801e-06, + "loss": 1.54321346, + "memory(GiB)": 111.15, + "step": 46780, + "train_speed(iter/s)": 0.44758 + }, + { + "acc": 0.64614697, + "epoch": 1.1868340943683409, + "grad_norm": 5.1875, + "learning_rate": 3.87837150710158e-06, + "loss": 1.66871834, + "memory(GiB)": 111.15, + "step": 46785, + "train_speed(iter/s)": 0.447605 + }, + { + "acc": 0.66030612, + "epoch": 1.186960933536276, + "grad_norm": 5.90625, + "learning_rate": 3.877349633080587e-06, + "loss": 1.52540379, + "memory(GiB)": 111.15, + "step": 46790, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.65793371, + "epoch": 1.187087772704211, + "grad_norm": 5.40625, + "learning_rate": 3.876327808438767e-06, + "loss": 1.6509758, + "memory(GiB)": 111.15, + "step": 46795, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.64991207, + "epoch": 1.187214611872146, + "grad_norm": 5.25, + "learning_rate": 3.875306033221069e-06, + "loss": 1.54582911, + "memory(GiB)": 111.15, + "step": 46800, + "train_speed(iter/s)": 0.447681 + }, + { + "epoch": 1.187214611872146, + "eval_acc": 0.6433411859661934, + "eval_loss": 1.6078166961669922, + "eval_runtime": 113.7343, + "eval_samples_per_second": 56.008, + "eval_steps_per_second": 28.004, + "step": 46800 + }, + { + "acc": 0.63877516, + "epoch": 1.1873414510400813, + "grad_norm": 5.84375, + "learning_rate": 3.874284307472432e-06, + "loss": 1.69477615, + "memory(GiB)": 111.15, + "step": 46805, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.65301046, + "epoch": 1.1874682902080163, + "grad_norm": 5.28125, + "learning_rate": 3.873262631237799e-06, + "loss": 1.65873566, + "memory(GiB)": 111.15, + "step": 46810, + "train_speed(iter/s)": 0.447225 + }, + { + "acc": 0.64105935, + "epoch": 1.1875951293759512, + "grad_norm": 5.5625, + "learning_rate": 3.872241004562105e-06, + "loss": 1.71467495, + "memory(GiB)": 111.15, + "step": 46815, + "train_speed(iter/s)": 0.44725 + }, + { + "acc": 0.66651692, + "epoch": 1.1877219685438862, + "grad_norm": 6.53125, + "learning_rate": 3.871219427490285e-06, + "loss": 1.54091167, + "memory(GiB)": 111.15, + "step": 46820, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.63732157, + "epoch": 1.1878488077118214, + "grad_norm": 5.96875, + "learning_rate": 3.870197900067276e-06, + "loss": 1.66475563, + "memory(GiB)": 111.15, + "step": 46825, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.64778442, + "epoch": 1.1879756468797564, + "grad_norm": 5.25, + "learning_rate": 3.869176422338009e-06, + "loss": 1.6131958, + "memory(GiB)": 111.15, + "step": 46830, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.65855842, + "epoch": 1.1881024860476916, + "grad_norm": 6.09375, + "learning_rate": 3.868154994347409e-06, + "loss": 1.62099552, + "memory(GiB)": 111.15, + "step": 46835, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.67050047, + "epoch": 1.1882293252156266, + "grad_norm": 6.21875, + "learning_rate": 3.867133616140406e-06, + "loss": 1.53940811, + "memory(GiB)": 111.15, + "step": 46840, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.65345144, + "epoch": 1.1883561643835616, + "grad_norm": 5.8125, + "learning_rate": 3.866112287761926e-06, + "loss": 1.64423256, + "memory(GiB)": 111.15, + "step": 46845, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.6415453, + "epoch": 1.1884830035514966, + "grad_norm": 6.5, + "learning_rate": 3.86509100925689e-06, + "loss": 1.64514389, + "memory(GiB)": 111.15, + "step": 46850, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.65050068, + "epoch": 1.1886098427194318, + "grad_norm": 4.96875, + "learning_rate": 3.8640697806702166e-06, + "loss": 1.58821697, + "memory(GiB)": 111.15, + "step": 46855, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.6538517, + "epoch": 1.1887366818873668, + "grad_norm": 6.09375, + "learning_rate": 3.8630486020468265e-06, + "loss": 1.65535355, + "memory(GiB)": 111.15, + "step": 46860, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.64641705, + "epoch": 1.1888635210553018, + "grad_norm": 5.71875, + "learning_rate": 3.862027473431634e-06, + "loss": 1.61022339, + "memory(GiB)": 111.15, + "step": 46865, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.66413507, + "epoch": 1.188990360223237, + "grad_norm": 6.71875, + "learning_rate": 3.861006394869558e-06, + "loss": 1.57503166, + "memory(GiB)": 111.15, + "step": 46870, + "train_speed(iter/s)": 0.447531 + }, + { + "acc": 0.64258285, + "epoch": 1.189117199391172, + "grad_norm": 5.03125, + "learning_rate": 3.859985366405502e-06, + "loss": 1.66073227, + "memory(GiB)": 111.15, + "step": 46875, + "train_speed(iter/s)": 0.447556 + }, + { + "acc": 0.64807539, + "epoch": 1.189244038559107, + "grad_norm": 4.40625, + "learning_rate": 3.85896438808438e-06, + "loss": 1.61729336, + "memory(GiB)": 111.15, + "step": 46880, + "train_speed(iter/s)": 0.447582 + }, + { + "acc": 0.645749, + "epoch": 1.1893708777270422, + "grad_norm": 5.6875, + "learning_rate": 3.857943459951099e-06, + "loss": 1.641465, + "memory(GiB)": 111.15, + "step": 46885, + "train_speed(iter/s)": 0.447607 + }, + { + "acc": 0.64825687, + "epoch": 1.1894977168949772, + "grad_norm": 5.71875, + "learning_rate": 3.856922582050565e-06, + "loss": 1.65048752, + "memory(GiB)": 111.15, + "step": 46890, + "train_speed(iter/s)": 0.447632 + }, + { + "acc": 0.67235918, + "epoch": 1.1896245560629122, + "grad_norm": 5.46875, + "learning_rate": 3.855901754427678e-06, + "loss": 1.54907627, + "memory(GiB)": 111.15, + "step": 46895, + "train_speed(iter/s)": 0.447658 + }, + { + "acc": 0.62520919, + "epoch": 1.1897513952308474, + "grad_norm": 5.6875, + "learning_rate": 3.854880977127339e-06, + "loss": 1.69621677, + "memory(GiB)": 111.15, + "step": 46900, + "train_speed(iter/s)": 0.447683 + }, + { + "epoch": 1.1897513952308474, + "eval_acc": 0.6433762784201879, + "eval_loss": 1.607796549797058, + "eval_runtime": 113.2496, + "eval_samples_per_second": 56.247, + "eval_steps_per_second": 28.124, + "step": 46900 + }, + { + "acc": 0.64865847, + "epoch": 1.1898782343987824, + "grad_norm": 5.4375, + "learning_rate": 3.8538602501944475e-06, + "loss": 1.5939394, + "memory(GiB)": 111.15, + "step": 46905, + "train_speed(iter/s)": 0.447205 + }, + { + "acc": 0.65164728, + "epoch": 1.1900050735667174, + "grad_norm": 6.34375, + "learning_rate": 3.852839573673902e-06, + "loss": 1.60291367, + "memory(GiB)": 111.15, + "step": 46910, + "train_speed(iter/s)": 0.44723 + }, + { + "acc": 0.65956063, + "epoch": 1.1901319127346524, + "grad_norm": 4.9375, + "learning_rate": 3.851818947610591e-06, + "loss": 1.5942565, + "memory(GiB)": 111.15, + "step": 46915, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.65292206, + "epoch": 1.1902587519025876, + "grad_norm": 5.46875, + "learning_rate": 3.850798372049409e-06, + "loss": 1.60742855, + "memory(GiB)": 111.15, + "step": 46920, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.6544528, + "epoch": 1.1903855910705226, + "grad_norm": 5.21875, + "learning_rate": 3.849777847035246e-06, + "loss": 1.57942448, + "memory(GiB)": 111.15, + "step": 46925, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.66555243, + "epoch": 1.1905124302384575, + "grad_norm": 5.4375, + "learning_rate": 3.84875737261299e-06, + "loss": 1.51958065, + "memory(GiB)": 111.15, + "step": 46930, + "train_speed(iter/s)": 0.447331 + }, + { + "acc": 0.64966784, + "epoch": 1.1906392694063928, + "grad_norm": 5.75, + "learning_rate": 3.847736948827523e-06, + "loss": 1.66825829, + "memory(GiB)": 111.15, + "step": 46935, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.66295085, + "epoch": 1.1907661085743277, + "grad_norm": 5.75, + "learning_rate": 3.846716575723729e-06, + "loss": 1.58078728, + "memory(GiB)": 111.15, + "step": 46940, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.6453351, + "epoch": 1.1908929477422627, + "grad_norm": 4.71875, + "learning_rate": 3.845696253346489e-06, + "loss": 1.66081696, + "memory(GiB)": 111.15, + "step": 46945, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.6588696, + "epoch": 1.191019786910198, + "grad_norm": 5.34375, + "learning_rate": 3.8446759817406835e-06, + "loss": 1.6150034, + "memory(GiB)": 111.15, + "step": 46950, + "train_speed(iter/s)": 0.447431 + }, + { + "acc": 0.65619712, + "epoch": 1.191146626078133, + "grad_norm": 4.71875, + "learning_rate": 3.8436557609511856e-06, + "loss": 1.63253288, + "memory(GiB)": 111.15, + "step": 46955, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.67505159, + "epoch": 1.191273465246068, + "grad_norm": 5.40625, + "learning_rate": 3.842635591022869e-06, + "loss": 1.49584599, + "memory(GiB)": 111.15, + "step": 46960, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.64804583, + "epoch": 1.1914003044140031, + "grad_norm": 6.71875, + "learning_rate": 3.8416154720006065e-06, + "loss": 1.60227165, + "memory(GiB)": 111.15, + "step": 46965, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.6511301, + "epoch": 1.1915271435819381, + "grad_norm": 4.4375, + "learning_rate": 3.840595403929269e-06, + "loss": 1.60385323, + "memory(GiB)": 111.15, + "step": 46970, + "train_speed(iter/s)": 0.447532 + }, + { + "acc": 0.63888903, + "epoch": 1.191653982749873, + "grad_norm": 5.59375, + "learning_rate": 3.839575386853721e-06, + "loss": 1.59445953, + "memory(GiB)": 111.15, + "step": 46975, + "train_speed(iter/s)": 0.447557 + }, + { + "acc": 0.65102034, + "epoch": 1.191780821917808, + "grad_norm": 5.125, + "learning_rate": 3.838555420818827e-06, + "loss": 1.59000874, + "memory(GiB)": 111.15, + "step": 46980, + "train_speed(iter/s)": 0.447582 + }, + { + "acc": 0.65353889, + "epoch": 1.1919076610857433, + "grad_norm": 5.84375, + "learning_rate": 3.837535505869453e-06, + "loss": 1.61040154, + "memory(GiB)": 111.15, + "step": 46985, + "train_speed(iter/s)": 0.447607 + }, + { + "acc": 0.65334883, + "epoch": 1.1920345002536783, + "grad_norm": 4.90625, + "learning_rate": 3.836515642050458e-06, + "loss": 1.62214642, + "memory(GiB)": 111.15, + "step": 46990, + "train_speed(iter/s)": 0.447632 + }, + { + "acc": 0.62742748, + "epoch": 1.1921613394216135, + "grad_norm": 4.875, + "learning_rate": 3.835495829406698e-06, + "loss": 1.64981174, + "memory(GiB)": 111.15, + "step": 46995, + "train_speed(iter/s)": 0.447657 + }, + { + "acc": 0.65620804, + "epoch": 1.1922881785895485, + "grad_norm": 5.59375, + "learning_rate": 3.834476067983031e-06, + "loss": 1.62131004, + "memory(GiB)": 111.15, + "step": 47000, + "train_speed(iter/s)": 0.447683 + }, + { + "epoch": 1.1922881785895485, + "eval_acc": 0.6433336661546232, + "eval_loss": 1.6078417301177979, + "eval_runtime": 113.2386, + "eval_samples_per_second": 56.253, + "eval_steps_per_second": 28.126, + "step": 47000 + }, + { + "acc": 0.64372883, + "epoch": 1.1924150177574835, + "grad_norm": 5.28125, + "learning_rate": 3.83345635782431e-06, + "loss": 1.68912373, + "memory(GiB)": 111.15, + "step": 47005, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.6544651, + "epoch": 1.1925418569254185, + "grad_norm": 5.78125, + "learning_rate": 3.832436698975388e-06, + "loss": 1.64183083, + "memory(GiB)": 111.15, + "step": 47010, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.66400108, + "epoch": 1.1926686960933537, + "grad_norm": 5.21875, + "learning_rate": 3.831417091481111e-06, + "loss": 1.54812889, + "memory(GiB)": 111.15, + "step": 47015, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.67001028, + "epoch": 1.1927955352612887, + "grad_norm": 5.15625, + "learning_rate": 3.830397535386328e-06, + "loss": 1.57026052, + "memory(GiB)": 111.15, + "step": 47020, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.65019093, + "epoch": 1.1929223744292237, + "grad_norm": 5.8125, + "learning_rate": 3.829378030735883e-06, + "loss": 1.64121895, + "memory(GiB)": 111.15, + "step": 47025, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.65588799, + "epoch": 1.1930492135971589, + "grad_norm": 5.0, + "learning_rate": 3.82835857757462e-06, + "loss": 1.61480503, + "memory(GiB)": 111.15, + "step": 47030, + "train_speed(iter/s)": 0.447331 + }, + { + "acc": 0.64989557, + "epoch": 1.1931760527650939, + "grad_norm": 4.4375, + "learning_rate": 3.827339175947378e-06, + "loss": 1.64257927, + "memory(GiB)": 111.15, + "step": 47035, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.64739208, + "epoch": 1.1933028919330289, + "grad_norm": 5.46875, + "learning_rate": 3.826319825898992e-06, + "loss": 1.57474098, + "memory(GiB)": 111.15, + "step": 47040, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.65963697, + "epoch": 1.193429731100964, + "grad_norm": 5.6875, + "learning_rate": 3.825300527474302e-06, + "loss": 1.61434097, + "memory(GiB)": 111.15, + "step": 47045, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.64703913, + "epoch": 1.193556570268899, + "grad_norm": 6.15625, + "learning_rate": 3.824281280718141e-06, + "loss": 1.65535545, + "memory(GiB)": 111.15, + "step": 47050, + "train_speed(iter/s)": 0.447431 + }, + { + "acc": 0.65695896, + "epoch": 1.193683409436834, + "grad_norm": 6.59375, + "learning_rate": 3.823262085675337e-06, + "loss": 1.62028427, + "memory(GiB)": 111.15, + "step": 47055, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64035463, + "epoch": 1.1938102486047693, + "grad_norm": 5.875, + "learning_rate": 3.822242942390718e-06, + "loss": 1.6708004, + "memory(GiB)": 111.15, + "step": 47060, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.65534563, + "epoch": 1.1939370877727042, + "grad_norm": 4.9375, + "learning_rate": 3.821223850909115e-06, + "loss": 1.62452946, + "memory(GiB)": 111.15, + "step": 47065, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.6459826, + "epoch": 1.1940639269406392, + "grad_norm": 5.8125, + "learning_rate": 3.820204811275351e-06, + "loss": 1.66202126, + "memory(GiB)": 111.15, + "step": 47070, + "train_speed(iter/s)": 0.447531 + }, + { + "acc": 0.6526494, + "epoch": 1.1941907661085742, + "grad_norm": 4.90625, + "learning_rate": 3.8191858235342446e-06, + "loss": 1.63852158, + "memory(GiB)": 111.15, + "step": 47075, + "train_speed(iter/s)": 0.447556 + }, + { + "acc": 0.63766775, + "epoch": 1.1943176052765094, + "grad_norm": 6.59375, + "learning_rate": 3.818166887730618e-06, + "loss": 1.65203171, + "memory(GiB)": 111.15, + "step": 47080, + "train_speed(iter/s)": 0.447582 + }, + { + "acc": 0.6425415, + "epoch": 1.1944444444444444, + "grad_norm": 5.25, + "learning_rate": 3.817148003909288e-06, + "loss": 1.68977394, + "memory(GiB)": 111.15, + "step": 47085, + "train_speed(iter/s)": 0.447607 + }, + { + "acc": 0.64908423, + "epoch": 1.1945712836123794, + "grad_norm": 4.8125, + "learning_rate": 3.816129172115073e-06, + "loss": 1.60046387, + "memory(GiB)": 111.15, + "step": 47090, + "train_speed(iter/s)": 0.447632 + }, + { + "acc": 0.63893442, + "epoch": 1.1946981227803146, + "grad_norm": 5.15625, + "learning_rate": 3.815110392392778e-06, + "loss": 1.63302174, + "memory(GiB)": 111.15, + "step": 47095, + "train_speed(iter/s)": 0.447657 + }, + { + "acc": 0.64791975, + "epoch": 1.1948249619482496, + "grad_norm": 5.0625, + "learning_rate": 3.8140916647872204e-06, + "loss": 1.62084579, + "memory(GiB)": 111.15, + "step": 47100, + "train_speed(iter/s)": 0.447683 + }, + { + "epoch": 1.1948249619482496, + "eval_acc": 0.6433152843930071, + "eval_loss": 1.607810616493225, + "eval_runtime": 114.2868, + "eval_samples_per_second": 55.737, + "eval_steps_per_second": 27.868, + "step": 47100 + }, + { + "acc": 0.64452467, + "epoch": 1.1949518011161846, + "grad_norm": 5.65625, + "learning_rate": 3.813072989343205e-06, + "loss": 1.65487766, + "memory(GiB)": 111.15, + "step": 47105, + "train_speed(iter/s)": 0.447202 + }, + { + "acc": 0.65524359, + "epoch": 1.1950786402841198, + "grad_norm": 5.03125, + "learning_rate": 3.812054366105541e-06, + "loss": 1.60009232, + "memory(GiB)": 111.15, + "step": 47110, + "train_speed(iter/s)": 0.447227 + }, + { + "acc": 0.66587877, + "epoch": 1.1952054794520548, + "grad_norm": 4.9375, + "learning_rate": 3.8110357951190284e-06, + "loss": 1.51372757, + "memory(GiB)": 111.15, + "step": 47115, + "train_speed(iter/s)": 0.447252 + }, + { + "acc": 0.62035208, + "epoch": 1.1953323186199898, + "grad_norm": 4.65625, + "learning_rate": 3.8100172764284694e-06, + "loss": 1.69509659, + "memory(GiB)": 111.15, + "step": 47120, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.66741858, + "epoch": 1.195459157787925, + "grad_norm": 5.125, + "learning_rate": 3.8089988100786635e-06, + "loss": 1.57264538, + "memory(GiB)": 111.15, + "step": 47125, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.65391297, + "epoch": 1.19558599695586, + "grad_norm": 5.90625, + "learning_rate": 3.807980396114409e-06, + "loss": 1.58162422, + "memory(GiB)": 111.15, + "step": 47130, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.63872185, + "epoch": 1.195712836123795, + "grad_norm": 5.125, + "learning_rate": 3.8069620345804974e-06, + "loss": 1.62677727, + "memory(GiB)": 111.15, + "step": 47135, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.65035658, + "epoch": 1.19583967529173, + "grad_norm": 5.21875, + "learning_rate": 3.8059437255217214e-06, + "loss": 1.66274986, + "memory(GiB)": 111.15, + "step": 47140, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.66393356, + "epoch": 1.1959665144596652, + "grad_norm": 5.3125, + "learning_rate": 3.8049254689828723e-06, + "loss": 1.59361563, + "memory(GiB)": 111.15, + "step": 47145, + "train_speed(iter/s)": 0.447401 + }, + { + "acc": 0.65213041, + "epoch": 1.1960933536276002, + "grad_norm": 5.875, + "learning_rate": 3.8039072650087377e-06, + "loss": 1.60386372, + "memory(GiB)": 111.15, + "step": 47150, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.65136499, + "epoch": 1.1962201927955354, + "grad_norm": 4.65625, + "learning_rate": 3.8028891136440994e-06, + "loss": 1.56730862, + "memory(GiB)": 111.15, + "step": 47155, + "train_speed(iter/s)": 0.447451 + }, + { + "acc": 0.64433413, + "epoch": 1.1963470319634704, + "grad_norm": 4.71875, + "learning_rate": 3.801871014933744e-06, + "loss": 1.6102293, + "memory(GiB)": 111.15, + "step": 47160, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.64815998, + "epoch": 1.1964738711314054, + "grad_norm": 5.90625, + "learning_rate": 3.8008529689224493e-06, + "loss": 1.61478386, + "memory(GiB)": 111.15, + "step": 47165, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.64651251, + "epoch": 1.1966007102993403, + "grad_norm": 5.78125, + "learning_rate": 3.7998349756549974e-06, + "loss": 1.63293419, + "memory(GiB)": 111.15, + "step": 47170, + "train_speed(iter/s)": 0.447526 + }, + { + "acc": 0.6304461, + "epoch": 1.1967275494672756, + "grad_norm": 5.375, + "learning_rate": 3.79881703517616e-06, + "loss": 1.75762939, + "memory(GiB)": 111.15, + "step": 47175, + "train_speed(iter/s)": 0.447551 + }, + { + "acc": 0.65583591, + "epoch": 1.1968543886352105, + "grad_norm": 6.28125, + "learning_rate": 3.797799147530713e-06, + "loss": 1.60827045, + "memory(GiB)": 111.15, + "step": 47180, + "train_speed(iter/s)": 0.447576 + }, + { + "acc": 0.64915967, + "epoch": 1.1969812278031455, + "grad_norm": 5.84375, + "learning_rate": 3.796781312763425e-06, + "loss": 1.58828526, + "memory(GiB)": 111.15, + "step": 47185, + "train_speed(iter/s)": 0.447601 + }, + { + "acc": 0.65143828, + "epoch": 1.1971080669710807, + "grad_norm": 5.125, + "learning_rate": 3.79576353091907e-06, + "loss": 1.58469896, + "memory(GiB)": 111.15, + "step": 47190, + "train_speed(iter/s)": 0.447626 + }, + { + "acc": 0.66928282, + "epoch": 1.1972349061390157, + "grad_norm": 6.0, + "learning_rate": 3.7947458020424094e-06, + "loss": 1.5973175, + "memory(GiB)": 111.15, + "step": 47195, + "train_speed(iter/s)": 0.447651 + }, + { + "acc": 0.65493622, + "epoch": 1.1973617453069507, + "grad_norm": 4.78125, + "learning_rate": 3.793728126178209e-06, + "loss": 1.62752819, + "memory(GiB)": 111.15, + "step": 47200, + "train_speed(iter/s)": 0.447676 + }, + { + "epoch": 1.1973617453069507, + "eval_acc": 0.6433804560932824, + "eval_loss": 1.6077016592025757, + "eval_runtime": 113.7248, + "eval_samples_per_second": 56.012, + "eval_steps_per_second": 28.006, + "step": 47200 + }, + { + "acc": 0.65599833, + "epoch": 1.197488584474886, + "grad_norm": 5.5625, + "learning_rate": 3.792710503371232e-06, + "loss": 1.548736, + "memory(GiB)": 111.15, + "step": 47205, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.64238148, + "epoch": 1.197615423642821, + "grad_norm": 5.8125, + "learning_rate": 3.7916929336662386e-06, + "loss": 1.75370884, + "memory(GiB)": 111.15, + "step": 47210, + "train_speed(iter/s)": 0.447224 + }, + { + "acc": 0.65672865, + "epoch": 1.197742262810756, + "grad_norm": 5.15625, + "learning_rate": 3.790675417107982e-06, + "loss": 1.58055677, + "memory(GiB)": 111.15, + "step": 47215, + "train_speed(iter/s)": 0.447249 + }, + { + "acc": 0.65181417, + "epoch": 1.1978691019786911, + "grad_norm": 5.40625, + "learning_rate": 3.7896579537412213e-06, + "loss": 1.60296822, + "memory(GiB)": 111.15, + "step": 47220, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.62953358, + "epoch": 1.197995941146626, + "grad_norm": 5.25, + "learning_rate": 3.7886405436107076e-06, + "loss": 1.7062458, + "memory(GiB)": 111.15, + "step": 47225, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.64399862, + "epoch": 1.198122780314561, + "grad_norm": 5.5625, + "learning_rate": 3.7876231867611917e-06, + "loss": 1.61711826, + "memory(GiB)": 111.15, + "step": 47230, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.6409482, + "epoch": 1.198249619482496, + "grad_norm": 5.09375, + "learning_rate": 3.7866058832374197e-06, + "loss": 1.65873909, + "memory(GiB)": 111.15, + "step": 47235, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.6465476, + "epoch": 1.1983764586504313, + "grad_norm": 5.46875, + "learning_rate": 3.7855886330841383e-06, + "loss": 1.65072479, + "memory(GiB)": 111.15, + "step": 47240, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.64928226, + "epoch": 1.1985032978183663, + "grad_norm": 6.3125, + "learning_rate": 3.7845714363460908e-06, + "loss": 1.62824821, + "memory(GiB)": 111.15, + "step": 47245, + "train_speed(iter/s)": 0.447401 + }, + { + "acc": 0.65132589, + "epoch": 1.1986301369863013, + "grad_norm": 5.0625, + "learning_rate": 3.78355429306802e-06, + "loss": 1.6526844, + "memory(GiB)": 111.15, + "step": 47250, + "train_speed(iter/s)": 0.447427 + }, + { + "acc": 0.66083665, + "epoch": 1.1987569761542365, + "grad_norm": 5.71875, + "learning_rate": 3.7825372032946605e-06, + "loss": 1.58583622, + "memory(GiB)": 111.15, + "step": 47255, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.66323175, + "epoch": 1.1988838153221715, + "grad_norm": 5.4375, + "learning_rate": 3.7815201670707502e-06, + "loss": 1.61432533, + "memory(GiB)": 111.15, + "step": 47260, + "train_speed(iter/s)": 0.447477 + }, + { + "acc": 0.64575815, + "epoch": 1.1990106544901065, + "grad_norm": 5.40625, + "learning_rate": 3.7805031844410235e-06, + "loss": 1.71331177, + "memory(GiB)": 111.15, + "step": 47265, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.66020565, + "epoch": 1.1991374936580417, + "grad_norm": 5.09375, + "learning_rate": 3.7794862554502126e-06, + "loss": 1.5668499, + "memory(GiB)": 111.15, + "step": 47270, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.65035701, + "epoch": 1.1992643328259767, + "grad_norm": 5.96875, + "learning_rate": 3.778469380143045e-06, + "loss": 1.58033581, + "memory(GiB)": 111.15, + "step": 47275, + "train_speed(iter/s)": 0.447554 + }, + { + "acc": 0.65289326, + "epoch": 1.1993911719939117, + "grad_norm": 5.3125, + "learning_rate": 3.777452558564246e-06, + "loss": 1.61503963, + "memory(GiB)": 111.15, + "step": 47280, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.65649085, + "epoch": 1.1995180111618469, + "grad_norm": 5.46875, + "learning_rate": 3.776435790758543e-06, + "loss": 1.5854475, + "memory(GiB)": 111.15, + "step": 47285, + "train_speed(iter/s)": 0.447605 + }, + { + "acc": 0.66239305, + "epoch": 1.1996448503297819, + "grad_norm": 5.96875, + "learning_rate": 3.7754190767706577e-06, + "loss": 1.50329571, + "memory(GiB)": 111.15, + "step": 47290, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.64972925, + "epoch": 1.1997716894977168, + "grad_norm": 5.875, + "learning_rate": 3.774402416645307e-06, + "loss": 1.64218693, + "memory(GiB)": 111.15, + "step": 47295, + "train_speed(iter/s)": 0.447656 + }, + { + "acc": 0.64684868, + "epoch": 1.1998985286656518, + "grad_norm": 4.9375, + "learning_rate": 3.77338581042721e-06, + "loss": 1.64996586, + "memory(GiB)": 111.15, + "step": 47300, + "train_speed(iter/s)": 0.447681 + }, + { + "epoch": 1.1998985286656518, + "eval_acc": 0.64337460735095, + "eval_loss": 1.6077994108200073, + "eval_runtime": 113.9927, + "eval_samples_per_second": 55.881, + "eval_steps_per_second": 27.94, + "step": 47300 + }, + { + "acc": 0.66063423, + "epoch": 1.200025367833587, + "grad_norm": 6.25, + "learning_rate": 3.7723692581610817e-06, + "loss": 1.63490067, + "memory(GiB)": 111.15, + "step": 47305, + "train_speed(iter/s)": 0.447204 + }, + { + "acc": 0.65235252, + "epoch": 1.200152207001522, + "grad_norm": 5.09375, + "learning_rate": 3.771352759891637e-06, + "loss": 1.66440639, + "memory(GiB)": 111.15, + "step": 47310, + "train_speed(iter/s)": 0.447229 + }, + { + "acc": 0.64651155, + "epoch": 1.2002790461694572, + "grad_norm": 5.34375, + "learning_rate": 3.7703363156635807e-06, + "loss": 1.70540962, + "memory(GiB)": 111.15, + "step": 47315, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.65964055, + "epoch": 1.2004058853373922, + "grad_norm": 5.21875, + "learning_rate": 3.769319925521624e-06, + "loss": 1.64245872, + "memory(GiB)": 111.15, + "step": 47320, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.65863333, + "epoch": 1.2005327245053272, + "grad_norm": 5.75, + "learning_rate": 3.76830358951047e-06, + "loss": 1.63877335, + "memory(GiB)": 111.15, + "step": 47325, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.65368929, + "epoch": 1.2006595636732622, + "grad_norm": 6.0625, + "learning_rate": 3.767287307674826e-06, + "loss": 1.56880226, + "memory(GiB)": 111.15, + "step": 47330, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.65094099, + "epoch": 1.2007864028411974, + "grad_norm": 5.65625, + "learning_rate": 3.766271080059389e-06, + "loss": 1.63790665, + "memory(GiB)": 111.15, + "step": 47335, + "train_speed(iter/s)": 0.447355 + }, + { + "acc": 0.65386715, + "epoch": 1.2009132420091324, + "grad_norm": 6.21875, + "learning_rate": 3.7652549067088568e-06, + "loss": 1.58669739, + "memory(GiB)": 111.15, + "step": 47340, + "train_speed(iter/s)": 0.44738 + }, + { + "acc": 0.6585969, + "epoch": 1.2010400811770674, + "grad_norm": 6.59375, + "learning_rate": 3.7642387876679275e-06, + "loss": 1.67277603, + "memory(GiB)": 111.15, + "step": 47345, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.64913068, + "epoch": 1.2011669203450026, + "grad_norm": 4.21875, + "learning_rate": 3.7632227229812947e-06, + "loss": 1.61513691, + "memory(GiB)": 111.15, + "step": 47350, + "train_speed(iter/s)": 0.44743 + }, + { + "acc": 0.64804773, + "epoch": 1.2012937595129376, + "grad_norm": 5.9375, + "learning_rate": 3.7622067126936475e-06, + "loss": 1.64707146, + "memory(GiB)": 111.15, + "step": 47355, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.64691219, + "epoch": 1.2014205986808726, + "grad_norm": 5.3125, + "learning_rate": 3.761190756849674e-06, + "loss": 1.60236073, + "memory(GiB)": 111.15, + "step": 47360, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.64791112, + "epoch": 1.2015474378488078, + "grad_norm": 5.71875, + "learning_rate": 3.7601748554940633e-06, + "loss": 1.6262598, + "memory(GiB)": 111.15, + "step": 47365, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.65453033, + "epoch": 1.2016742770167428, + "grad_norm": 5.09375, + "learning_rate": 3.7591590086714984e-06, + "loss": 1.56902199, + "memory(GiB)": 111.15, + "step": 47370, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.64505119, + "epoch": 1.2018011161846778, + "grad_norm": 6.9375, + "learning_rate": 3.7581432164266587e-06, + "loss": 1.58591843, + "memory(GiB)": 111.15, + "step": 47375, + "train_speed(iter/s)": 0.447555 + }, + { + "acc": 0.65155611, + "epoch": 1.201927955352613, + "grad_norm": 5.15625, + "learning_rate": 3.7571274788042255e-06, + "loss": 1.60366821, + "memory(GiB)": 111.15, + "step": 47380, + "train_speed(iter/s)": 0.44758 + }, + { + "acc": 0.64582577, + "epoch": 1.202054794520548, + "grad_norm": 5.46875, + "learning_rate": 3.756111795848874e-06, + "loss": 1.65439987, + "memory(GiB)": 111.15, + "step": 47385, + "train_speed(iter/s)": 0.447605 + }, + { + "acc": 0.64738846, + "epoch": 1.202181633688483, + "grad_norm": 5.5, + "learning_rate": 3.755096167605281e-06, + "loss": 1.6521347, + "memory(GiB)": 111.15, + "step": 47390, + "train_speed(iter/s)": 0.44763 + }, + { + "acc": 0.65458045, + "epoch": 1.202308472856418, + "grad_norm": 5.1875, + "learning_rate": 3.7540805941181165e-06, + "loss": 1.58372736, + "memory(GiB)": 111.15, + "step": 47395, + "train_speed(iter/s)": 0.447655 + }, + { + "acc": 0.65066504, + "epoch": 1.2024353120243532, + "grad_norm": 4.90625, + "learning_rate": 3.7530650754320492e-06, + "loss": 1.66861324, + "memory(GiB)": 111.15, + "step": 47400, + "train_speed(iter/s)": 0.447681 + }, + { + "epoch": 1.2024353120243532, + "eval_acc": 0.6434055221318499, + "eval_loss": 1.6076984405517578, + "eval_runtime": 113.6676, + "eval_samples_per_second": 56.041, + "eval_steps_per_second": 28.02, + "step": 47400 + }, + { + "acc": 0.67682047, + "epoch": 1.2025621511922882, + "grad_norm": 4.71875, + "learning_rate": 3.752049611591746e-06, + "loss": 1.4845768, + "memory(GiB)": 111.15, + "step": 47405, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.65043101, + "epoch": 1.2026889903602231, + "grad_norm": 5.5, + "learning_rate": 3.7510342026418756e-06, + "loss": 1.64132957, + "memory(GiB)": 111.15, + "step": 47410, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.65255251, + "epoch": 1.2028158295281584, + "grad_norm": 5.78125, + "learning_rate": 3.7500188486270948e-06, + "loss": 1.62099266, + "memory(GiB)": 111.15, + "step": 47415, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.64229746, + "epoch": 1.2029426686960933, + "grad_norm": 5.75, + "learning_rate": 3.7490035495920664e-06, + "loss": 1.67899933, + "memory(GiB)": 111.15, + "step": 47420, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.66219988, + "epoch": 1.2030695078640283, + "grad_norm": 4.96875, + "learning_rate": 3.747988305581447e-06, + "loss": 1.55836086, + "memory(GiB)": 111.15, + "step": 47425, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.64597006, + "epoch": 1.2031963470319635, + "grad_norm": 5.4375, + "learning_rate": 3.7469731166398933e-06, + "loss": 1.64388199, + "memory(GiB)": 111.15, + "step": 47430, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.64388843, + "epoch": 1.2033231861998985, + "grad_norm": 5.125, + "learning_rate": 3.745957982812054e-06, + "loss": 1.65910416, + "memory(GiB)": 111.15, + "step": 47435, + "train_speed(iter/s)": 0.447355 + }, + { + "acc": 0.6432899, + "epoch": 1.2034500253678335, + "grad_norm": 5.46875, + "learning_rate": 3.744942904142582e-06, + "loss": 1.6235733, + "memory(GiB)": 111.15, + "step": 47440, + "train_speed(iter/s)": 0.44738 + }, + { + "acc": 0.66812091, + "epoch": 1.2035768645357687, + "grad_norm": 4.59375, + "learning_rate": 3.743927880676125e-06, + "loss": 1.54810514, + "memory(GiB)": 111.15, + "step": 47445, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.66300983, + "epoch": 1.2037037037037037, + "grad_norm": 4.40625, + "learning_rate": 3.742912912457329e-06, + "loss": 1.56910543, + "memory(GiB)": 111.15, + "step": 47450, + "train_speed(iter/s)": 0.44743 + }, + { + "acc": 0.64727383, + "epoch": 1.2038305428716387, + "grad_norm": 5.40625, + "learning_rate": 3.7418979995308336e-06, + "loss": 1.59051266, + "memory(GiB)": 111.15, + "step": 47455, + "train_speed(iter/s)": 0.447455 + }, + { + "acc": 0.65251174, + "epoch": 1.2039573820395737, + "grad_norm": 5.03125, + "learning_rate": 3.740883141941282e-06, + "loss": 1.593297, + "memory(GiB)": 111.15, + "step": 47460, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.6620513, + "epoch": 1.204084221207509, + "grad_norm": 5.53125, + "learning_rate": 3.7398683397333103e-06, + "loss": 1.63043518, + "memory(GiB)": 111.15, + "step": 47465, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.64778442, + "epoch": 1.204211060375444, + "grad_norm": 4.5625, + "learning_rate": 3.7388535929515573e-06, + "loss": 1.59478455, + "memory(GiB)": 111.15, + "step": 47470, + "train_speed(iter/s)": 0.447529 + }, + { + "acc": 0.66306729, + "epoch": 1.204337899543379, + "grad_norm": 6.0, + "learning_rate": 3.737838901640653e-06, + "loss": 1.5375061, + "memory(GiB)": 111.15, + "step": 47475, + "train_speed(iter/s)": 0.447554 + }, + { + "acc": 0.6575891, + "epoch": 1.204464738711314, + "grad_norm": 7.40625, + "learning_rate": 3.736824265845228e-06, + "loss": 1.58714552, + "memory(GiB)": 111.15, + "step": 47480, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.65470543, + "epoch": 1.204591577879249, + "grad_norm": 5.125, + "learning_rate": 3.7358096856099118e-06, + "loss": 1.60038395, + "memory(GiB)": 111.15, + "step": 47485, + "train_speed(iter/s)": 0.447603 + }, + { + "acc": 0.65879736, + "epoch": 1.204718417047184, + "grad_norm": 5.375, + "learning_rate": 3.7347951609793315e-06, + "loss": 1.57160978, + "memory(GiB)": 111.15, + "step": 47490, + "train_speed(iter/s)": 0.447628 + }, + { + "acc": 0.64342966, + "epoch": 1.2048452562151193, + "grad_norm": 5.59375, + "learning_rate": 3.7337806919981077e-06, + "loss": 1.6381731, + "memory(GiB)": 111.15, + "step": 47495, + "train_speed(iter/s)": 0.447653 + }, + { + "acc": 0.6579772, + "epoch": 1.2049720953830543, + "grad_norm": 5.75, + "learning_rate": 3.732766278710861e-06, + "loss": 1.60898933, + "memory(GiB)": 111.15, + "step": 47500, + "train_speed(iter/s)": 0.447678 + }, + { + "epoch": 1.2049720953830543, + "eval_acc": 0.6433980023202797, + "eval_loss": 1.6077312231063843, + "eval_runtime": 114.7563, + "eval_samples_per_second": 55.509, + "eval_steps_per_second": 27.754, + "step": 47500 + }, + { + "acc": 0.6547832, + "epoch": 1.2050989345509893, + "grad_norm": 4.90625, + "learning_rate": 3.7317519211622123e-06, + "loss": 1.62187233, + "memory(GiB)": 111.15, + "step": 47505, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.65012894, + "epoch": 1.2052257737189245, + "grad_norm": 5.15625, + "learning_rate": 3.7307376193967772e-06, + "loss": 1.65923386, + "memory(GiB)": 111.15, + "step": 47510, + "train_speed(iter/s)": 0.447224 + }, + { + "acc": 0.64516664, + "epoch": 1.2053526128868595, + "grad_norm": 4.96875, + "learning_rate": 3.7297233734591664e-06, + "loss": 1.63121109, + "memory(GiB)": 111.15, + "step": 47515, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.64669795, + "epoch": 1.2054794520547945, + "grad_norm": 6.25, + "learning_rate": 3.7287091833939948e-06, + "loss": 1.66650696, + "memory(GiB)": 111.15, + "step": 47520, + "train_speed(iter/s)": 0.447273 + }, + { + "acc": 0.66835423, + "epoch": 1.2056062912227297, + "grad_norm": 4.9375, + "learning_rate": 3.7276950492458675e-06, + "loss": 1.54574051, + "memory(GiB)": 111.15, + "step": 47525, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.65246973, + "epoch": 1.2057331303906647, + "grad_norm": 5.375, + "learning_rate": 3.7266809710593956e-06, + "loss": 1.62127075, + "memory(GiB)": 111.15, + "step": 47530, + "train_speed(iter/s)": 0.447323 + }, + { + "acc": 0.65464272, + "epoch": 1.2058599695585996, + "grad_norm": 5.5625, + "learning_rate": 3.7256669488791763e-06, + "loss": 1.58897781, + "memory(GiB)": 111.15, + "step": 47535, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.65780854, + "epoch": 1.2059868087265349, + "grad_norm": 5.4375, + "learning_rate": 3.7246529827498156e-06, + "loss": 1.52052889, + "memory(GiB)": 111.15, + "step": 47540, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.65141678, + "epoch": 1.2061136478944698, + "grad_norm": 4.9375, + "learning_rate": 3.7236390727159094e-06, + "loss": 1.61569424, + "memory(GiB)": 111.15, + "step": 47545, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.658813, + "epoch": 1.2062404870624048, + "grad_norm": 4.875, + "learning_rate": 3.7226252188220573e-06, + "loss": 1.54030838, + "memory(GiB)": 111.15, + "step": 47550, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.65997934, + "epoch": 1.2063673262303398, + "grad_norm": 5.78125, + "learning_rate": 3.7216114211128505e-06, + "loss": 1.60118923, + "memory(GiB)": 111.15, + "step": 47555, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.65768738, + "epoch": 1.206494165398275, + "grad_norm": 5.6875, + "learning_rate": 3.720597679632879e-06, + "loss": 1.57945309, + "memory(GiB)": 111.15, + "step": 47560, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.64723349, + "epoch": 1.20662100456621, + "grad_norm": 7.15625, + "learning_rate": 3.7195839944267357e-06, + "loss": 1.68111839, + "memory(GiB)": 111.15, + "step": 47565, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.66303234, + "epoch": 1.206747843734145, + "grad_norm": 4.75, + "learning_rate": 3.718570365539006e-06, + "loss": 1.5292613, + "memory(GiB)": 111.15, + "step": 47570, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.65656977, + "epoch": 1.2068746829020802, + "grad_norm": 5.96875, + "learning_rate": 3.717556793014271e-06, + "loss": 1.63489723, + "memory(GiB)": 111.15, + "step": 47575, + "train_speed(iter/s)": 0.447546 + }, + { + "acc": 0.6491662, + "epoch": 1.2070015220700152, + "grad_norm": 4.84375, + "learning_rate": 3.716543276897113e-06, + "loss": 1.66815987, + "memory(GiB)": 111.15, + "step": 47580, + "train_speed(iter/s)": 0.447571 + }, + { + "acc": 0.64680214, + "epoch": 1.2071283612379502, + "grad_norm": 4.75, + "learning_rate": 3.715529817232114e-06, + "loss": 1.61519375, + "memory(GiB)": 111.15, + "step": 47585, + "train_speed(iter/s)": 0.447596 + }, + { + "acc": 0.64570065, + "epoch": 1.2072552004058854, + "grad_norm": 5.75, + "learning_rate": 3.7145164140638483e-06, + "loss": 1.61480827, + "memory(GiB)": 111.15, + "step": 47590, + "train_speed(iter/s)": 0.447621 + }, + { + "acc": 0.64466009, + "epoch": 1.2073820395738204, + "grad_norm": 5.21875, + "learning_rate": 3.713503067436889e-06, + "loss": 1.64471035, + "memory(GiB)": 111.15, + "step": 47595, + "train_speed(iter/s)": 0.447645 + }, + { + "acc": 0.64371881, + "epoch": 1.2075088787417554, + "grad_norm": 5.84375, + "learning_rate": 3.7124897773958084e-06, + "loss": 1.61705055, + "memory(GiB)": 111.15, + "step": 47600, + "train_speed(iter/s)": 0.44767 + }, + { + "epoch": 1.2075088787417554, + "eval_acc": 0.6434051043645405, + "eval_loss": 1.6077505350112915, + "eval_runtime": 113.8312, + "eval_samples_per_second": 55.96, + "eval_steps_per_second": 27.98, + "step": 47600 + }, + { + "acc": 0.64630804, + "epoch": 1.2076357179096906, + "grad_norm": 5.875, + "learning_rate": 3.7114765439851752e-06, + "loss": 1.6979126, + "memory(GiB)": 111.15, + "step": 47605, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.66235161, + "epoch": 1.2077625570776256, + "grad_norm": 4.65625, + "learning_rate": 3.7104633672495584e-06, + "loss": 1.56708546, + "memory(GiB)": 111.15, + "step": 47610, + "train_speed(iter/s)": 0.447221 + }, + { + "acc": 0.66539803, + "epoch": 1.2078893962455606, + "grad_norm": 6.59375, + "learning_rate": 3.709450247233519e-06, + "loss": 1.4865097, + "memory(GiB)": 111.15, + "step": 47615, + "train_speed(iter/s)": 0.447246 + }, + { + "acc": 0.63140802, + "epoch": 1.2080162354134956, + "grad_norm": 8.5625, + "learning_rate": 3.7084371839816204e-06, + "loss": 1.68726883, + "memory(GiB)": 111.15, + "step": 47620, + "train_speed(iter/s)": 0.447271 + }, + { + "acc": 0.66654081, + "epoch": 1.2081430745814308, + "grad_norm": 4.375, + "learning_rate": 3.707424177538419e-06, + "loss": 1.55566435, + "memory(GiB)": 111.15, + "step": 47625, + "train_speed(iter/s)": 0.447296 + }, + { + "acc": 0.65799713, + "epoch": 1.2082699137493658, + "grad_norm": 5.3125, + "learning_rate": 3.7064112279484753e-06, + "loss": 1.59993982, + "memory(GiB)": 111.15, + "step": 47630, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.65974917, + "epoch": 1.208396752917301, + "grad_norm": 6.03125, + "learning_rate": 3.7053983352563407e-06, + "loss": 1.61474133, + "memory(GiB)": 111.15, + "step": 47635, + "train_speed(iter/s)": 0.447345 + }, + { + "acc": 0.65772419, + "epoch": 1.208523592085236, + "grad_norm": 4.5, + "learning_rate": 3.704385499506565e-06, + "loss": 1.58415384, + "memory(GiB)": 111.15, + "step": 47640, + "train_speed(iter/s)": 0.44737 + }, + { + "acc": 0.6479146, + "epoch": 1.208650431253171, + "grad_norm": 4.96875, + "learning_rate": 3.703372720743702e-06, + "loss": 1.61236, + "memory(GiB)": 111.15, + "step": 47645, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.64465985, + "epoch": 1.208777270421106, + "grad_norm": 5.4375, + "learning_rate": 3.7023599990122966e-06, + "loss": 1.60457916, + "memory(GiB)": 111.15, + "step": 47650, + "train_speed(iter/s)": 0.447419 + }, + { + "acc": 0.64368415, + "epoch": 1.2089041095890412, + "grad_norm": 5.3125, + "learning_rate": 3.7013473343568897e-06, + "loss": 1.62619667, + "memory(GiB)": 111.15, + "step": 47655, + "train_speed(iter/s)": 0.447444 + }, + { + "acc": 0.66768713, + "epoch": 1.2090309487569761, + "grad_norm": 5.65625, + "learning_rate": 3.700334726822026e-06, + "loss": 1.57417669, + "memory(GiB)": 111.15, + "step": 47660, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.67035952, + "epoch": 1.2091577879249111, + "grad_norm": 5.75, + "learning_rate": 3.6993221764522435e-06, + "loss": 1.52637615, + "memory(GiB)": 111.15, + "step": 47665, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.65587769, + "epoch": 1.2092846270928463, + "grad_norm": 4.96875, + "learning_rate": 3.6983096832920806e-06, + "loss": 1.60864296, + "memory(GiB)": 111.15, + "step": 47670, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.66529856, + "epoch": 1.2094114662607813, + "grad_norm": 5.4375, + "learning_rate": 3.697297247386066e-06, + "loss": 1.63312664, + "memory(GiB)": 111.15, + "step": 47675, + "train_speed(iter/s)": 0.447543 + }, + { + "acc": 0.65390258, + "epoch": 1.2095383054287163, + "grad_norm": 4.96875, + "learning_rate": 3.6962848687787365e-06, + "loss": 1.6984911, + "memory(GiB)": 111.15, + "step": 47680, + "train_speed(iter/s)": 0.447568 + }, + { + "acc": 0.65136628, + "epoch": 1.2096651445966515, + "grad_norm": 4.875, + "learning_rate": 3.6952725475146183e-06, + "loss": 1.58290577, + "memory(GiB)": 111.15, + "step": 47685, + "train_speed(iter/s)": 0.447592 + }, + { + "acc": 0.65189271, + "epoch": 1.2097919837645865, + "grad_norm": 5.25, + "learning_rate": 3.69426028363824e-06, + "loss": 1.60921211, + "memory(GiB)": 111.15, + "step": 47690, + "train_speed(iter/s)": 0.447617 + }, + { + "acc": 0.65687299, + "epoch": 1.2099188229325215, + "grad_norm": 6.1875, + "learning_rate": 3.6932480771941237e-06, + "loss": 1.55680113, + "memory(GiB)": 111.15, + "step": 47695, + "train_speed(iter/s)": 0.447641 + }, + { + "acc": 0.65945048, + "epoch": 1.2100456621004567, + "grad_norm": 5.40625, + "learning_rate": 3.6922359282267904e-06, + "loss": 1.57848682, + "memory(GiB)": 111.15, + "step": 47700, + "train_speed(iter/s)": 0.447666 + }, + { + "epoch": 1.2100456621004567, + "eval_acc": 0.6433980023202797, + "eval_loss": 1.6077300310134888, + "eval_runtime": 115.4301, + "eval_samples_per_second": 55.185, + "eval_steps_per_second": 27.592, + "step": 47700 + }, + { + "acc": 0.65141521, + "epoch": 1.2101725012683917, + "grad_norm": 4.6875, + "learning_rate": 3.6912238367807606e-06, + "loss": 1.61986694, + "memory(GiB)": 111.15, + "step": 47705, + "train_speed(iter/s)": 0.447187 + }, + { + "acc": 0.63594704, + "epoch": 1.2102993404363267, + "grad_norm": 5.8125, + "learning_rate": 3.6902118029005507e-06, + "loss": 1.65860729, + "memory(GiB)": 111.15, + "step": 47710, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.6520504, + "epoch": 1.2104261796042617, + "grad_norm": 5.46875, + "learning_rate": 3.6891998266306717e-06, + "loss": 1.64313602, + "memory(GiB)": 111.15, + "step": 47715, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.65523911, + "epoch": 1.210553018772197, + "grad_norm": 5.53125, + "learning_rate": 3.688187908015636e-06, + "loss": 1.61708775, + "memory(GiB)": 111.15, + "step": 47720, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.6569603, + "epoch": 1.2106798579401319, + "grad_norm": 6.90625, + "learning_rate": 3.6871760470999546e-06, + "loss": 1.66032448, + "memory(GiB)": 111.15, + "step": 47725, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.65249434, + "epoch": 1.2108066971080669, + "grad_norm": 5.34375, + "learning_rate": 3.6861642439281325e-06, + "loss": 1.58661079, + "memory(GiB)": 111.15, + "step": 47730, + "train_speed(iter/s)": 0.447312 + }, + { + "acc": 0.64394808, + "epoch": 1.210933536276002, + "grad_norm": 5.0, + "learning_rate": 3.6851524985446707e-06, + "loss": 1.66791325, + "memory(GiB)": 111.15, + "step": 47735, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.65306182, + "epoch": 1.211060375443937, + "grad_norm": 5.5625, + "learning_rate": 3.6841408109940737e-06, + "loss": 1.58619537, + "memory(GiB)": 111.15, + "step": 47740, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.64388714, + "epoch": 1.211187214611872, + "grad_norm": 6.1875, + "learning_rate": 3.6831291813208377e-06, + "loss": 1.61109238, + "memory(GiB)": 111.15, + "step": 47745, + "train_speed(iter/s)": 0.447388 + }, + { + "acc": 0.63552046, + "epoch": 1.2113140537798073, + "grad_norm": 5.90625, + "learning_rate": 3.682117609569462e-06, + "loss": 1.66452255, + "memory(GiB)": 111.15, + "step": 47750, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.63942208, + "epoch": 1.2114408929477423, + "grad_norm": 4.84375, + "learning_rate": 3.681106095784436e-06, + "loss": 1.67119751, + "memory(GiB)": 111.15, + "step": 47755, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.67147889, + "epoch": 1.2115677321156773, + "grad_norm": 4.0625, + "learning_rate": 3.6800946400102522e-06, + "loss": 1.5054884, + "memory(GiB)": 111.15, + "step": 47760, + "train_speed(iter/s)": 0.447463 + }, + { + "acc": 0.65671682, + "epoch": 1.2116945712836125, + "grad_norm": 5.5625, + "learning_rate": 3.6790832422913984e-06, + "loss": 1.62184353, + "memory(GiB)": 111.15, + "step": 47765, + "train_speed(iter/s)": 0.447488 + }, + { + "acc": 0.65650301, + "epoch": 1.2118214104515475, + "grad_norm": 6.8125, + "learning_rate": 3.6780719026723632e-06, + "loss": 1.59287777, + "memory(GiB)": 111.15, + "step": 47770, + "train_speed(iter/s)": 0.447513 + }, + { + "acc": 0.64112058, + "epoch": 1.2119482496194824, + "grad_norm": 5.65625, + "learning_rate": 3.677060621197627e-06, + "loss": 1.71733704, + "memory(GiB)": 111.15, + "step": 47775, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.64410601, + "epoch": 1.2120750887874174, + "grad_norm": 4.8125, + "learning_rate": 3.6760493979116696e-06, + "loss": 1.61598816, + "memory(GiB)": 111.15, + "step": 47780, + "train_speed(iter/s)": 0.447563 + }, + { + "acc": 0.65623865, + "epoch": 1.2122019279553526, + "grad_norm": 4.90625, + "learning_rate": 3.6750382328589725e-06, + "loss": 1.60107498, + "memory(GiB)": 111.15, + "step": 47785, + "train_speed(iter/s)": 0.447588 + }, + { + "acc": 0.63281446, + "epoch": 1.2123287671232876, + "grad_norm": 7.03125, + "learning_rate": 3.67402712608401e-06, + "loss": 1.66122513, + "memory(GiB)": 111.15, + "step": 47790, + "train_speed(iter/s)": 0.447613 + }, + { + "acc": 0.65852118, + "epoch": 1.2124556062912228, + "grad_norm": 5.03125, + "learning_rate": 3.673016077631253e-06, + "loss": 1.61151543, + "memory(GiB)": 111.15, + "step": 47795, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.64196701, + "epoch": 1.2125824454591578, + "grad_norm": 5.9375, + "learning_rate": 3.672005087545173e-06, + "loss": 1.61143017, + "memory(GiB)": 111.15, + "step": 47800, + "train_speed(iter/s)": 0.447663 + }, + { + "epoch": 1.2125824454591578, + "eval_acc": 0.6434259927300133, + "eval_loss": 1.6077433824539185, + "eval_runtime": 114.1561, + "eval_samples_per_second": 55.801, + "eval_steps_per_second": 27.9, + "step": 47800 + }, + { + "acc": 0.65422726, + "epoch": 1.2127092846270928, + "grad_norm": 4.9375, + "learning_rate": 3.6709941558702393e-06, + "loss": 1.62622223, + "memory(GiB)": 111.15, + "step": 47805, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.65045505, + "epoch": 1.2128361237950278, + "grad_norm": 5.5, + "learning_rate": 3.6699832826509174e-06, + "loss": 1.63952866, + "memory(GiB)": 111.15, + "step": 47810, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.64424953, + "epoch": 1.212962962962963, + "grad_norm": 5.25, + "learning_rate": 3.6689724679316665e-06, + "loss": 1.68489914, + "memory(GiB)": 111.15, + "step": 47815, + "train_speed(iter/s)": 0.44724 + }, + { + "acc": 0.64231186, + "epoch": 1.213089802130898, + "grad_norm": 5.6875, + "learning_rate": 3.66796171175695e-06, + "loss": 1.67084084, + "memory(GiB)": 111.15, + "step": 47820, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.65276976, + "epoch": 1.213216641298833, + "grad_norm": 5.09375, + "learning_rate": 3.666951014171224e-06, + "loss": 1.61187248, + "memory(GiB)": 111.15, + "step": 47825, + "train_speed(iter/s)": 0.44729 + }, + { + "acc": 0.64309301, + "epoch": 1.2133434804667682, + "grad_norm": 4.71875, + "learning_rate": 3.6659403752189453e-06, + "loss": 1.60531235, + "memory(GiB)": 111.15, + "step": 47830, + "train_speed(iter/s)": 0.447314 + }, + { + "acc": 0.65282698, + "epoch": 1.2134703196347032, + "grad_norm": 5.25, + "learning_rate": 3.664929794944565e-06, + "loss": 1.59204578, + "memory(GiB)": 111.15, + "step": 47835, + "train_speed(iter/s)": 0.447339 + }, + { + "acc": 0.66545134, + "epoch": 1.2135971588026382, + "grad_norm": 5.28125, + "learning_rate": 3.663919273392532e-06, + "loss": 1.57408562, + "memory(GiB)": 111.15, + "step": 47840, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.67837734, + "epoch": 1.2137239979705734, + "grad_norm": 4.65625, + "learning_rate": 3.662908810607294e-06, + "loss": 1.50195189, + "memory(GiB)": 111.15, + "step": 47845, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.65614805, + "epoch": 1.2138508371385084, + "grad_norm": 5.125, + "learning_rate": 3.6618984066332986e-06, + "loss": 1.65863113, + "memory(GiB)": 111.15, + "step": 47850, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.65701585, + "epoch": 1.2139776763064434, + "grad_norm": 6.4375, + "learning_rate": 3.660888061514984e-06, + "loss": 1.6179821, + "memory(GiB)": 111.15, + "step": 47855, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.66167822, + "epoch": 1.2141045154743786, + "grad_norm": 5.8125, + "learning_rate": 3.6598777752967896e-06, + "loss": 1.64049988, + "memory(GiB)": 111.15, + "step": 47860, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.6537859, + "epoch": 1.2142313546423136, + "grad_norm": 5.6875, + "learning_rate": 3.658867548023156e-06, + "loss": 1.64080276, + "memory(GiB)": 111.15, + "step": 47865, + "train_speed(iter/s)": 0.447489 + }, + { + "acc": 0.65503445, + "epoch": 1.2143581938102486, + "grad_norm": 6.34375, + "learning_rate": 3.657857379738515e-06, + "loss": 1.64621544, + "memory(GiB)": 111.15, + "step": 47870, + "train_speed(iter/s)": 0.447514 + }, + { + "acc": 0.65130949, + "epoch": 1.2144850329781836, + "grad_norm": 5.96875, + "learning_rate": 3.656847270487298e-06, + "loss": 1.66771469, + "memory(GiB)": 111.15, + "step": 47875, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.6452878, + "epoch": 1.2146118721461188, + "grad_norm": 5.21875, + "learning_rate": 3.655837220313936e-06, + "loss": 1.61989803, + "memory(GiB)": 111.15, + "step": 47880, + "train_speed(iter/s)": 0.447564 + }, + { + "acc": 0.63829598, + "epoch": 1.2147387113140538, + "grad_norm": 4.25, + "learning_rate": 3.654827229262852e-06, + "loss": 1.67354317, + "memory(GiB)": 111.15, + "step": 47885, + "train_speed(iter/s)": 0.447589 + }, + { + "acc": 0.66510448, + "epoch": 1.2148655504819887, + "grad_norm": 5.59375, + "learning_rate": 3.653817297378476e-06, + "loss": 1.57378635, + "memory(GiB)": 111.15, + "step": 47890, + "train_speed(iter/s)": 0.447614 + }, + { + "acc": 0.66895037, + "epoch": 1.214992389649924, + "grad_norm": 5.125, + "learning_rate": 3.6528074247052225e-06, + "loss": 1.54359665, + "memory(GiB)": 111.15, + "step": 47895, + "train_speed(iter/s)": 0.447639 + }, + { + "acc": 0.65845246, + "epoch": 1.215119228817859, + "grad_norm": 4.3125, + "learning_rate": 3.651797611287514e-06, + "loss": 1.52852211, + "memory(GiB)": 111.15, + "step": 47900, + "train_speed(iter/s)": 0.447663 + }, + { + "epoch": 1.215119228817859, + "eval_acc": 0.6433574788912623, + "eval_loss": 1.6077561378479004, + "eval_runtime": 114.4003, + "eval_samples_per_second": 55.682, + "eval_steps_per_second": 27.841, + "step": 47900 + }, + { + "acc": 0.66938353, + "epoch": 1.215246067985794, + "grad_norm": 5.96875, + "learning_rate": 3.6507878571697646e-06, + "loss": 1.5638504, + "memory(GiB)": 111.15, + "step": 47905, + "train_speed(iter/s)": 0.447191 + }, + { + "acc": 0.65448694, + "epoch": 1.2153729071537291, + "grad_norm": 4.75, + "learning_rate": 3.6497781623963915e-06, + "loss": 1.61602745, + "memory(GiB)": 111.15, + "step": 47910, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.65453444, + "epoch": 1.2154997463216641, + "grad_norm": 4.96875, + "learning_rate": 3.648768527011802e-06, + "loss": 1.62160873, + "memory(GiB)": 111.15, + "step": 47915, + "train_speed(iter/s)": 0.447241 + }, + { + "acc": 0.65369759, + "epoch": 1.2156265854895991, + "grad_norm": 5.15625, + "learning_rate": 3.6477589510604044e-06, + "loss": 1.65145912, + "memory(GiB)": 111.15, + "step": 47920, + "train_speed(iter/s)": 0.447266 + }, + { + "acc": 0.64625349, + "epoch": 1.2157534246575343, + "grad_norm": 4.71875, + "learning_rate": 3.646749434586607e-06, + "loss": 1.62981644, + "memory(GiB)": 111.15, + "step": 47925, + "train_speed(iter/s)": 0.44729 + }, + { + "acc": 0.66276236, + "epoch": 1.2158802638254693, + "grad_norm": 6.65625, + "learning_rate": 3.645739977634811e-06, + "loss": 1.58160534, + "memory(GiB)": 111.15, + "step": 47930, + "train_speed(iter/s)": 0.447315 + }, + { + "acc": 0.65069604, + "epoch": 1.2160071029934043, + "grad_norm": 5.3125, + "learning_rate": 3.6447305802494177e-06, + "loss": 1.61468315, + "memory(GiB)": 111.15, + "step": 47935, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.65809236, + "epoch": 1.2161339421613393, + "grad_norm": 5.46875, + "learning_rate": 3.6437212424748227e-06, + "loss": 1.66192245, + "memory(GiB)": 111.15, + "step": 47940, + "train_speed(iter/s)": 0.447365 + }, + { + "acc": 0.65046949, + "epoch": 1.2162607813292745, + "grad_norm": 6.5625, + "learning_rate": 3.642711964355423e-06, + "loss": 1.66111946, + "memory(GiB)": 111.15, + "step": 47945, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.63981104, + "epoch": 1.2163876204972095, + "grad_norm": 5.65625, + "learning_rate": 3.6417027459356134e-06, + "loss": 1.65926552, + "memory(GiB)": 111.15, + "step": 47950, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.62649975, + "epoch": 1.2165144596651447, + "grad_norm": 5.5, + "learning_rate": 3.640693587259778e-06, + "loss": 1.73616142, + "memory(GiB)": 111.15, + "step": 47955, + "train_speed(iter/s)": 0.44744 + }, + { + "acc": 0.64227886, + "epoch": 1.2166412988330797, + "grad_norm": 4.375, + "learning_rate": 3.6396844883723092e-06, + "loss": 1.69049416, + "memory(GiB)": 111.15, + "step": 47960, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.65092874, + "epoch": 1.2167681380010147, + "grad_norm": 5.0625, + "learning_rate": 3.6386754493175893e-06, + "loss": 1.64562416, + "memory(GiB)": 111.15, + "step": 47965, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.64990368, + "epoch": 1.2168949771689497, + "grad_norm": 4.375, + "learning_rate": 3.637666470140003e-06, + "loss": 1.60775814, + "memory(GiB)": 111.15, + "step": 47970, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.65966177, + "epoch": 1.2170218163368849, + "grad_norm": 5.3125, + "learning_rate": 3.6366575508839265e-06, + "loss": 1.61353378, + "memory(GiB)": 111.15, + "step": 47975, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.64418468, + "epoch": 1.2171486555048199, + "grad_norm": 5.8125, + "learning_rate": 3.635648691593737e-06, + "loss": 1.70213966, + "memory(GiB)": 111.15, + "step": 47980, + "train_speed(iter/s)": 0.447564 + }, + { + "acc": 0.64952192, + "epoch": 1.2172754946727549, + "grad_norm": 5.0, + "learning_rate": 3.6346398923138094e-06, + "loss": 1.63174782, + "memory(GiB)": 111.15, + "step": 47985, + "train_speed(iter/s)": 0.44759 + }, + { + "acc": 0.65460558, + "epoch": 1.21740233384069, + "grad_norm": 4.875, + "learning_rate": 3.633631153088517e-06, + "loss": 1.59307308, + "memory(GiB)": 111.15, + "step": 47990, + "train_speed(iter/s)": 0.447615 + }, + { + "acc": 0.64008503, + "epoch": 1.217529173008625, + "grad_norm": 5.625, + "learning_rate": 3.6326224739622255e-06, + "loss": 1.70322704, + "memory(GiB)": 111.15, + "step": 47995, + "train_speed(iter/s)": 0.447616 + }, + { + "acc": 0.64503198, + "epoch": 1.21765601217656, + "grad_norm": 5.78125, + "learning_rate": 3.6316138549793024e-06, + "loss": 1.59764442, + "memory(GiB)": 111.15, + "step": 48000, + "train_speed(iter/s)": 0.447641 + }, + { + "epoch": 1.21765601217656, + "eval_acc": 0.6433766961874973, + "eval_loss": 1.6077216863632202, + "eval_runtime": 114.5905, + "eval_samples_per_second": 55.589, + "eval_steps_per_second": 27.795, + "step": 48000 + }, + { + "acc": 0.65124836, + "epoch": 1.2177828513444953, + "grad_norm": 5.40625, + "learning_rate": 3.630605296184111e-06, + "loss": 1.57235489, + "memory(GiB)": 111.15, + "step": 48005, + "train_speed(iter/s)": 0.447169 + }, + { + "acc": 0.63286295, + "epoch": 1.2179096905124303, + "grad_norm": 5.46875, + "learning_rate": 3.6295967976210146e-06, + "loss": 1.70333519, + "memory(GiB)": 111.15, + "step": 48010, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.63348994, + "epoch": 1.2180365296803652, + "grad_norm": 5.40625, + "learning_rate": 3.6285883593343685e-06, + "loss": 1.70695877, + "memory(GiB)": 111.15, + "step": 48015, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.65232325, + "epoch": 1.2181633688483005, + "grad_norm": 5.8125, + "learning_rate": 3.6275799813685274e-06, + "loss": 1.61698189, + "memory(GiB)": 111.15, + "step": 48020, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.65834684, + "epoch": 1.2182902080162354, + "grad_norm": 5.3125, + "learning_rate": 3.6265716637678484e-06, + "loss": 1.61616592, + "memory(GiB)": 111.15, + "step": 48025, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.64778399, + "epoch": 1.2184170471841704, + "grad_norm": 6.375, + "learning_rate": 3.62556340657668e-06, + "loss": 1.63860226, + "memory(GiB)": 111.15, + "step": 48030, + "train_speed(iter/s)": 0.447294 + }, + { + "acc": 0.65683537, + "epoch": 1.2185438863521054, + "grad_norm": 5.375, + "learning_rate": 3.6245552098393665e-06, + "loss": 1.63143463, + "memory(GiB)": 111.15, + "step": 48035, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.64933019, + "epoch": 1.2186707255200406, + "grad_norm": 5.34375, + "learning_rate": 3.6235470736002576e-06, + "loss": 1.60983467, + "memory(GiB)": 111.15, + "step": 48040, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.65772276, + "epoch": 1.2187975646879756, + "grad_norm": 4.75, + "learning_rate": 3.622538997903693e-06, + "loss": 1.634832, + "memory(GiB)": 111.15, + "step": 48045, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.64407678, + "epoch": 1.2189244038559106, + "grad_norm": 5.3125, + "learning_rate": 3.621530982794015e-06, + "loss": 1.65220604, + "memory(GiB)": 111.15, + "step": 48050, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.64979753, + "epoch": 1.2190512430238458, + "grad_norm": 5.53125, + "learning_rate": 3.620523028315558e-06, + "loss": 1.68992348, + "memory(GiB)": 111.15, + "step": 48055, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.64577208, + "epoch": 1.2191780821917808, + "grad_norm": 5.71875, + "learning_rate": 3.6195151345126556e-06, + "loss": 1.72460766, + "memory(GiB)": 111.15, + "step": 48060, + "train_speed(iter/s)": 0.447443 + }, + { + "acc": 0.65759902, + "epoch": 1.2193049213597158, + "grad_norm": 4.6875, + "learning_rate": 3.6185073014296425e-06, + "loss": 1.62168846, + "memory(GiB)": 111.15, + "step": 48065, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.65458035, + "epoch": 1.219431760527651, + "grad_norm": 6.28125, + "learning_rate": 3.6174995291108474e-06, + "loss": 1.56610241, + "memory(GiB)": 111.15, + "step": 48070, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.63579173, + "epoch": 1.219558599695586, + "grad_norm": 5.9375, + "learning_rate": 3.6164918176005937e-06, + "loss": 1.69299622, + "memory(GiB)": 111.15, + "step": 48075, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.64314928, + "epoch": 1.219685438863521, + "grad_norm": 5.34375, + "learning_rate": 3.6154841669432062e-06, + "loss": 1.59218025, + "memory(GiB)": 111.15, + "step": 48080, + "train_speed(iter/s)": 0.447542 + }, + { + "acc": 0.65766201, + "epoch": 1.2198122780314562, + "grad_norm": 5.28125, + "learning_rate": 3.614476577183007e-06, + "loss": 1.59719334, + "memory(GiB)": 111.15, + "step": 48085, + "train_speed(iter/s)": 0.447566 + }, + { + "acc": 0.65623488, + "epoch": 1.2199391171993912, + "grad_norm": 4.75, + "learning_rate": 3.6134690483643154e-06, + "loss": 1.57829704, + "memory(GiB)": 111.15, + "step": 48090, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.64585705, + "epoch": 1.2200659563673262, + "grad_norm": 5.71875, + "learning_rate": 3.6124615805314434e-06, + "loss": 1.64439926, + "memory(GiB)": 111.15, + "step": 48095, + "train_speed(iter/s)": 0.447616 + }, + { + "acc": 0.63885517, + "epoch": 1.2201927955352612, + "grad_norm": 5.0625, + "learning_rate": 3.611454173728707e-06, + "loss": 1.63103638, + "memory(GiB)": 111.15, + "step": 48100, + "train_speed(iter/s)": 0.447641 + }, + { + "epoch": 1.2201927955352612, + "eval_acc": 0.6433503768470015, + "eval_loss": 1.6077296733856201, + "eval_runtime": 114.6304, + "eval_samples_per_second": 55.57, + "eval_steps_per_second": 27.785, + "step": 48100 + }, + { + "acc": 0.66144776, + "epoch": 1.2203196347031964, + "grad_norm": 5.09375, + "learning_rate": 3.610446828000414e-06, + "loss": 1.59583511, + "memory(GiB)": 111.15, + "step": 48105, + "train_speed(iter/s)": 0.447169 + }, + { + "acc": 0.64330454, + "epoch": 1.2204464738711314, + "grad_norm": 4.8125, + "learning_rate": 3.609439543390877e-06, + "loss": 1.6276289, + "memory(GiB)": 111.15, + "step": 48110, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.65280132, + "epoch": 1.2205733130390666, + "grad_norm": 5.1875, + "learning_rate": 3.608432319944394e-06, + "loss": 1.65336361, + "memory(GiB)": 111.15, + "step": 48115, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.65101633, + "epoch": 1.2207001522070016, + "grad_norm": 6.59375, + "learning_rate": 3.607425157705271e-06, + "loss": 1.58845482, + "memory(GiB)": 111.15, + "step": 48120, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.64587588, + "epoch": 1.2208269913749366, + "grad_norm": 4.8125, + "learning_rate": 3.6064180567178064e-06, + "loss": 1.63901043, + "memory(GiB)": 111.15, + "step": 48125, + "train_speed(iter/s)": 0.447268 + }, + { + "acc": 0.66341476, + "epoch": 1.2209538305428715, + "grad_norm": 5.0, + "learning_rate": 3.6054110170263002e-06, + "loss": 1.58412638, + "memory(GiB)": 111.15, + "step": 48130, + "train_speed(iter/s)": 0.447292 + }, + { + "acc": 0.63947487, + "epoch": 1.2210806697108068, + "grad_norm": 7.15625, + "learning_rate": 3.6044040386750423e-06, + "loss": 1.62334328, + "memory(GiB)": 111.15, + "step": 48135, + "train_speed(iter/s)": 0.447316 + }, + { + "acc": 0.66491022, + "epoch": 1.2212075088787417, + "grad_norm": 6.21875, + "learning_rate": 3.6033971217083242e-06, + "loss": 1.57158766, + "memory(GiB)": 111.15, + "step": 48140, + "train_speed(iter/s)": 0.447341 + }, + { + "acc": 0.66050806, + "epoch": 1.2213343480466767, + "grad_norm": 4.90625, + "learning_rate": 3.602390266170438e-06, + "loss": 1.57898731, + "memory(GiB)": 111.15, + "step": 48145, + "train_speed(iter/s)": 0.447365 + }, + { + "acc": 0.64657922, + "epoch": 1.221461187214612, + "grad_norm": 7.03125, + "learning_rate": 3.6013834721056683e-06, + "loss": 1.62901497, + "memory(GiB)": 111.15, + "step": 48150, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.65540609, + "epoch": 1.221588026382547, + "grad_norm": 6.09375, + "learning_rate": 3.6003767395582967e-06, + "loss": 1.49912052, + "memory(GiB)": 111.15, + "step": 48155, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.65618868, + "epoch": 1.221714865550482, + "grad_norm": 6.03125, + "learning_rate": 3.599370068572604e-06, + "loss": 1.51594601, + "memory(GiB)": 111.15, + "step": 48160, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.65926113, + "epoch": 1.2218417047184171, + "grad_norm": 5.59375, + "learning_rate": 3.5983634591928705e-06, + "loss": 1.57809429, + "memory(GiB)": 111.15, + "step": 48165, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.66431522, + "epoch": 1.2219685438863521, + "grad_norm": 5.46875, + "learning_rate": 3.5973569114633704e-06, + "loss": 1.61279316, + "memory(GiB)": 111.15, + "step": 48170, + "train_speed(iter/s)": 0.447488 + }, + { + "acc": 0.66413121, + "epoch": 1.222095383054287, + "grad_norm": 6.0625, + "learning_rate": 3.5963504254283743e-06, + "loss": 1.66380348, + "memory(GiB)": 111.15, + "step": 48175, + "train_speed(iter/s)": 0.447513 + }, + { + "acc": 0.65199256, + "epoch": 1.2222222222222223, + "grad_norm": 4.75, + "learning_rate": 3.595344001132154e-06, + "loss": 1.57534294, + "memory(GiB)": 111.15, + "step": 48180, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.64947338, + "epoch": 1.2223490613901573, + "grad_norm": 4.875, + "learning_rate": 3.5943376386189744e-06, + "loss": 1.60487061, + "memory(GiB)": 111.15, + "step": 48185, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.65768194, + "epoch": 1.2224759005580923, + "grad_norm": 5.9375, + "learning_rate": 3.5933313379331047e-06, + "loss": 1.61257477, + "memory(GiB)": 111.15, + "step": 48190, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.63983412, + "epoch": 1.2226027397260273, + "grad_norm": 5.125, + "learning_rate": 3.5923250991188e-06, + "loss": 1.63574295, + "memory(GiB)": 111.15, + "step": 48195, + "train_speed(iter/s)": 0.44761 + }, + { + "acc": 0.64277277, + "epoch": 1.2227295788939625, + "grad_norm": 5.5, + "learning_rate": 3.591318922220324e-06, + "loss": 1.65039158, + "memory(GiB)": 111.15, + "step": 48200, + "train_speed(iter/s)": 0.447635 + }, + { + "epoch": 1.2227295788939625, + "eval_acc": 0.6434105353395634, + "eval_loss": 1.6078860759735107, + "eval_runtime": 116.3517, + "eval_samples_per_second": 54.748, + "eval_steps_per_second": 27.374, + "step": 48200 + }, + { + "acc": 0.65231667, + "epoch": 1.2228564180618975, + "grad_norm": 6.5, + "learning_rate": 3.5903128072819287e-06, + "loss": 1.66494789, + "memory(GiB)": 111.15, + "step": 48205, + "train_speed(iter/s)": 0.447157 + }, + { + "acc": 0.63991966, + "epoch": 1.2229832572298325, + "grad_norm": 5.46875, + "learning_rate": 3.5893067543478733e-06, + "loss": 1.67309532, + "memory(GiB)": 111.15, + "step": 48210, + "train_speed(iter/s)": 0.447182 + }, + { + "acc": 0.63511734, + "epoch": 1.2231100963977677, + "grad_norm": 5.53125, + "learning_rate": 3.5883007634624033e-06, + "loss": 1.69841976, + "memory(GiB)": 111.15, + "step": 48215, + "train_speed(iter/s)": 0.447207 + }, + { + "acc": 0.63905597, + "epoch": 1.2232369355657027, + "grad_norm": 4.84375, + "learning_rate": 3.5872948346697676e-06, + "loss": 1.61943626, + "memory(GiB)": 111.15, + "step": 48220, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.68012533, + "epoch": 1.2233637747336377, + "grad_norm": 5.40625, + "learning_rate": 3.5862889680142133e-06, + "loss": 1.54296989, + "memory(GiB)": 111.15, + "step": 48225, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.65258899, + "epoch": 1.2234906139015729, + "grad_norm": 5.40625, + "learning_rate": 3.5852831635399833e-06, + "loss": 1.63334255, + "memory(GiB)": 111.15, + "step": 48230, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.65408697, + "epoch": 1.2236174530695079, + "grad_norm": 6.0, + "learning_rate": 3.5842774212913144e-06, + "loss": 1.6583849, + "memory(GiB)": 111.15, + "step": 48235, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.6459578, + "epoch": 1.2237442922374429, + "grad_norm": 5.5625, + "learning_rate": 3.583271741312445e-06, + "loss": 1.64370422, + "memory(GiB)": 111.15, + "step": 48240, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.64814959, + "epoch": 1.223871131405378, + "grad_norm": 7.0625, + "learning_rate": 3.58226612364761e-06, + "loss": 1.59727707, + "memory(GiB)": 111.15, + "step": 48245, + "train_speed(iter/s)": 0.447355 + }, + { + "acc": 0.66084051, + "epoch": 1.223997970573313, + "grad_norm": 4.78125, + "learning_rate": 3.581260568341042e-06, + "loss": 1.59529686, + "memory(GiB)": 111.15, + "step": 48250, + "train_speed(iter/s)": 0.44738 + }, + { + "acc": 0.64121184, + "epoch": 1.224124809741248, + "grad_norm": 5.46875, + "learning_rate": 3.580255075436967e-06, + "loss": 1.69327927, + "memory(GiB)": 111.15, + "step": 48255, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.64513178, + "epoch": 1.224251648909183, + "grad_norm": 4.90625, + "learning_rate": 3.5792496449796127e-06, + "loss": 1.62184906, + "memory(GiB)": 111.15, + "step": 48260, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.66440115, + "epoch": 1.2243784880771182, + "grad_norm": 5.0625, + "learning_rate": 3.578244277013201e-06, + "loss": 1.58008366, + "memory(GiB)": 111.15, + "step": 48265, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.66057987, + "epoch": 1.2245053272450532, + "grad_norm": 4.1875, + "learning_rate": 3.5772389715819568e-06, + "loss": 1.60743828, + "memory(GiB)": 111.15, + "step": 48270, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.66055861, + "epoch": 1.2246321664129884, + "grad_norm": 5.375, + "learning_rate": 3.5762337287300925e-06, + "loss": 1.59810162, + "memory(GiB)": 111.15, + "step": 48275, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.6402153, + "epoch": 1.2247590055809234, + "grad_norm": 5.09375, + "learning_rate": 3.575228548501825e-06, + "loss": 1.63626614, + "memory(GiB)": 111.15, + "step": 48280, + "train_speed(iter/s)": 0.447529 + }, + { + "acc": 0.6397727, + "epoch": 1.2248858447488584, + "grad_norm": 5.625, + "learning_rate": 3.574223430941368e-06, + "loss": 1.65031281, + "memory(GiB)": 111.15, + "step": 48285, + "train_speed(iter/s)": 0.447553 + }, + { + "acc": 0.64009237, + "epoch": 1.2250126839167934, + "grad_norm": 6.25, + "learning_rate": 3.573218376092932e-06, + "loss": 1.71342697, + "memory(GiB)": 111.15, + "step": 48290, + "train_speed(iter/s)": 0.447578 + }, + { + "acc": 0.65148759, + "epoch": 1.2251395230847286, + "grad_norm": 6.6875, + "learning_rate": 3.5722133840007197e-06, + "loss": 1.64213409, + "memory(GiB)": 111.15, + "step": 48295, + "train_speed(iter/s)": 0.447603 + }, + { + "acc": 0.64038968, + "epoch": 1.2252663622526636, + "grad_norm": 6.125, + "learning_rate": 3.5712084547089367e-06, + "loss": 1.63766079, + "memory(GiB)": 111.15, + "step": 48300, + "train_speed(iter/s)": 0.447628 + }, + { + "epoch": 1.2252663622526636, + "eval_acc": 0.6433996733895174, + "eval_loss": 1.6077450513839722, + "eval_runtime": 113.8353, + "eval_samples_per_second": 55.958, + "eval_steps_per_second": 27.979, + "step": 48300 + }, + { + "acc": 0.65468769, + "epoch": 1.2253932014205986, + "grad_norm": 5.71875, + "learning_rate": 3.5702035882617857e-06, + "loss": 1.59920387, + "memory(GiB)": 111.15, + "step": 48305, + "train_speed(iter/s)": 0.447162 + }, + { + "acc": 0.63794231, + "epoch": 1.2255200405885338, + "grad_norm": 5.75, + "learning_rate": 3.5691987847034667e-06, + "loss": 1.67073975, + "memory(GiB)": 111.15, + "step": 48310, + "train_speed(iter/s)": 0.447186 + }, + { + "acc": 0.64178362, + "epoch": 1.2256468797564688, + "grad_norm": 5.71875, + "learning_rate": 3.5681940440781705e-06, + "loss": 1.6426281, + "memory(GiB)": 111.15, + "step": 48315, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.66678324, + "epoch": 1.2257737189244038, + "grad_norm": 4.9375, + "learning_rate": 3.5671893664300934e-06, + "loss": 1.55286713, + "memory(GiB)": 111.15, + "step": 48320, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.66302948, + "epoch": 1.225900558092339, + "grad_norm": 4.9375, + "learning_rate": 3.5661847518034244e-06, + "loss": 1.53236265, + "memory(GiB)": 111.15, + "step": 48325, + "train_speed(iter/s)": 0.447261 + }, + { + "acc": 0.63411884, + "epoch": 1.226027397260274, + "grad_norm": 4.9375, + "learning_rate": 3.5651802002423543e-06, + "loss": 1.61926174, + "memory(GiB)": 111.15, + "step": 48330, + "train_speed(iter/s)": 0.447286 + }, + { + "acc": 0.66400146, + "epoch": 1.226154236428209, + "grad_norm": 5.875, + "learning_rate": 3.5641757117910625e-06, + "loss": 1.59985371, + "memory(GiB)": 111.15, + "step": 48335, + "train_speed(iter/s)": 0.447311 + }, + { + "acc": 0.63565955, + "epoch": 1.2262810755961442, + "grad_norm": 5.75, + "learning_rate": 3.563171286493734e-06, + "loss": 1.68648033, + "memory(GiB)": 111.15, + "step": 48340, + "train_speed(iter/s)": 0.447335 + }, + { + "acc": 0.62883015, + "epoch": 1.2264079147640792, + "grad_norm": 5.25, + "learning_rate": 3.5621669243945457e-06, + "loss": 1.63255749, + "memory(GiB)": 111.15, + "step": 48345, + "train_speed(iter/s)": 0.44736 + }, + { + "acc": 0.64603376, + "epoch": 1.2265347539320142, + "grad_norm": 4.75, + "learning_rate": 3.5611626255376785e-06, + "loss": 1.65365391, + "memory(GiB)": 111.15, + "step": 48350, + "train_speed(iter/s)": 0.447385 + }, + { + "acc": 0.66827164, + "epoch": 1.2266615930999492, + "grad_norm": 5.34375, + "learning_rate": 3.560158389967302e-06, + "loss": 1.60416393, + "memory(GiB)": 111.15, + "step": 48355, + "train_speed(iter/s)": 0.44741 + }, + { + "acc": 0.64616065, + "epoch": 1.2267884322678844, + "grad_norm": 6.125, + "learning_rate": 3.559154217727586e-06, + "loss": 1.66946545, + "memory(GiB)": 111.15, + "step": 48360, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.64815755, + "epoch": 1.2269152714358194, + "grad_norm": 5.6875, + "learning_rate": 3.5581501088627026e-06, + "loss": 1.70807571, + "memory(GiB)": 111.15, + "step": 48365, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.65196028, + "epoch": 1.2270421106037543, + "grad_norm": 5.46875, + "learning_rate": 3.557146063416815e-06, + "loss": 1.60556145, + "memory(GiB)": 111.15, + "step": 48370, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.64248352, + "epoch": 1.2271689497716896, + "grad_norm": 5.40625, + "learning_rate": 3.5561420814340843e-06, + "loss": 1.66571541, + "memory(GiB)": 111.15, + "step": 48375, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.64004059, + "epoch": 1.2272957889396245, + "grad_norm": 5.5625, + "learning_rate": 3.555138162958671e-06, + "loss": 1.68097878, + "memory(GiB)": 111.15, + "step": 48380, + "train_speed(iter/s)": 0.447534 + }, + { + "acc": 0.65901833, + "epoch": 1.2274226281075595, + "grad_norm": 6.09375, + "learning_rate": 3.5541343080347325e-06, + "loss": 1.53210182, + "memory(GiB)": 111.15, + "step": 48385, + "train_speed(iter/s)": 0.447559 + }, + { + "acc": 0.65319448, + "epoch": 1.2275494672754947, + "grad_norm": 5.21875, + "learning_rate": 3.5531305167064234e-06, + "loss": 1.61542301, + "memory(GiB)": 111.15, + "step": 48390, + "train_speed(iter/s)": 0.447584 + }, + { + "acc": 0.66001253, + "epoch": 1.2276763064434297, + "grad_norm": 5.5625, + "learning_rate": 3.5521267890178922e-06, + "loss": 1.60195866, + "memory(GiB)": 111.15, + "step": 48395, + "train_speed(iter/s)": 0.447609 + }, + { + "acc": 0.64768724, + "epoch": 1.2278031456113647, + "grad_norm": 4.59375, + "learning_rate": 3.5511231250132905e-06, + "loss": 1.58399429, + "memory(GiB)": 111.15, + "step": 48400, + "train_speed(iter/s)": 0.447633 + }, + { + "epoch": 1.2278031456113647, + "eval_acc": 0.6433378438277177, + "eval_loss": 1.607682228088379, + "eval_runtime": 113.3471, + "eval_samples_per_second": 56.199, + "eval_steps_per_second": 28.1, + "step": 48400 + }, + { + "acc": 0.65700526, + "epoch": 1.2279299847793, + "grad_norm": 4.71875, + "learning_rate": 3.550119524736761e-06, + "loss": 1.68506794, + "memory(GiB)": 111.15, + "step": 48405, + "train_speed(iter/s)": 0.44717 + }, + { + "acc": 0.6405057, + "epoch": 1.228056823947235, + "grad_norm": 5.84375, + "learning_rate": 3.5491159882324513e-06, + "loss": 1.59755764, + "memory(GiB)": 111.15, + "step": 48410, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.64818821, + "epoch": 1.22818366311517, + "grad_norm": 5.21875, + "learning_rate": 3.548112515544495e-06, + "loss": 1.61488247, + "memory(GiB)": 111.15, + "step": 48415, + "train_speed(iter/s)": 0.44722 + }, + { + "acc": 0.65388908, + "epoch": 1.228310502283105, + "grad_norm": 6.8125, + "learning_rate": 3.547109106717034e-06, + "loss": 1.62739983, + "memory(GiB)": 111.15, + "step": 48420, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.65466604, + "epoch": 1.22843734145104, + "grad_norm": 5.21875, + "learning_rate": 3.546105761794199e-06, + "loss": 1.65891991, + "memory(GiB)": 111.15, + "step": 48425, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.64934115, + "epoch": 1.228564180618975, + "grad_norm": 5.5625, + "learning_rate": 3.5451024808201268e-06, + "loss": 1.60069733, + "memory(GiB)": 111.15, + "step": 48430, + "train_speed(iter/s)": 0.447294 + }, + { + "acc": 0.63662152, + "epoch": 1.2286910197869103, + "grad_norm": 5.3125, + "learning_rate": 3.5440992638389417e-06, + "loss": 1.69261818, + "memory(GiB)": 111.15, + "step": 48435, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.66415682, + "epoch": 1.2288178589548453, + "grad_norm": 5.125, + "learning_rate": 3.5430961108947705e-06, + "loss": 1.60691833, + "memory(GiB)": 111.15, + "step": 48440, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.63833866, + "epoch": 1.2289446981227803, + "grad_norm": 4.8125, + "learning_rate": 3.5420930220317373e-06, + "loss": 1.62704468, + "memory(GiB)": 111.15, + "step": 48445, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.6588203, + "epoch": 1.2290715372907153, + "grad_norm": 5.4375, + "learning_rate": 3.541089997293964e-06, + "loss": 1.57238865, + "memory(GiB)": 111.15, + "step": 48450, + "train_speed(iter/s)": 0.447392 + }, + { + "acc": 0.65343819, + "epoch": 1.2291983764586505, + "grad_norm": 6.40625, + "learning_rate": 3.5400870367255635e-06, + "loss": 1.5675498, + "memory(GiB)": 111.15, + "step": 48455, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.64011688, + "epoch": 1.2293252156265855, + "grad_norm": 6.1875, + "learning_rate": 3.539084140370654e-06, + "loss": 1.61940556, + "memory(GiB)": 111.15, + "step": 48460, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.65087452, + "epoch": 1.2294520547945205, + "grad_norm": 5.03125, + "learning_rate": 3.538081308273347e-06, + "loss": 1.64639969, + "memory(GiB)": 111.15, + "step": 48465, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.64846716, + "epoch": 1.2295788939624557, + "grad_norm": 6.0625, + "learning_rate": 3.537078540477752e-06, + "loss": 1.61550369, + "memory(GiB)": 111.15, + "step": 48470, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.63463354, + "epoch": 1.2297057331303907, + "grad_norm": 5.53125, + "learning_rate": 3.5360758370279722e-06, + "loss": 1.66868553, + "memory(GiB)": 111.15, + "step": 48475, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.65601492, + "epoch": 1.2298325722983257, + "grad_norm": 5.125, + "learning_rate": 3.535073197968114e-06, + "loss": 1.57357988, + "memory(GiB)": 111.15, + "step": 48480, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.65463295, + "epoch": 1.2299594114662609, + "grad_norm": 5.625, + "learning_rate": 3.5340706233422763e-06, + "loss": 1.65845432, + "memory(GiB)": 111.15, + "step": 48485, + "train_speed(iter/s)": 0.447563 + }, + { + "acc": 0.64762444, + "epoch": 1.2300862506341959, + "grad_norm": 5.9375, + "learning_rate": 3.5330681131945588e-06, + "loss": 1.61692352, + "memory(GiB)": 111.15, + "step": 48490, + "train_speed(iter/s)": 0.447588 + }, + { + "acc": 0.64837775, + "epoch": 1.2302130898021308, + "grad_norm": 4.9375, + "learning_rate": 3.5320656675690546e-06, + "loss": 1.58299866, + "memory(GiB)": 111.15, + "step": 48495, + "train_speed(iter/s)": 0.447612 + }, + { + "acc": 0.64653273, + "epoch": 1.230339928970066, + "grad_norm": 5.5625, + "learning_rate": 3.531063286509855e-06, + "loss": 1.58698254, + "memory(GiB)": 111.15, + "step": 48500, + "train_speed(iter/s)": 0.447637 + }, + { + "epoch": 1.230339928970066, + "eval_acc": 0.6434142952453484, + "eval_loss": 1.6076165437698364, + "eval_runtime": 113.0784, + "eval_samples_per_second": 56.333, + "eval_steps_per_second": 28.166, + "step": 48500 + }, + { + "acc": 0.65749564, + "epoch": 1.230466768138001, + "grad_norm": 5.78125, + "learning_rate": 3.530060970061051e-06, + "loss": 1.61763039, + "memory(GiB)": 111.15, + "step": 48505, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.64310007, + "epoch": 1.230593607305936, + "grad_norm": 5.4375, + "learning_rate": 3.52905871826673e-06, + "loss": 1.6685524, + "memory(GiB)": 111.15, + "step": 48510, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.65273113, + "epoch": 1.230720446473871, + "grad_norm": 5.25, + "learning_rate": 3.5280565311709725e-06, + "loss": 1.5775939, + "memory(GiB)": 111.15, + "step": 48515, + "train_speed(iter/s)": 0.447225 + }, + { + "acc": 0.63829331, + "epoch": 1.2308472856418062, + "grad_norm": 5.59375, + "learning_rate": 3.5270544088178597e-06, + "loss": 1.68294716, + "memory(GiB)": 111.15, + "step": 48520, + "train_speed(iter/s)": 0.44725 + }, + { + "acc": 0.64387627, + "epoch": 1.2309741248097412, + "grad_norm": 5.1875, + "learning_rate": 3.526052351251471e-06, + "loss": 1.66275482, + "memory(GiB)": 111.15, + "step": 48525, + "train_speed(iter/s)": 0.447274 + }, + { + "acc": 0.65400419, + "epoch": 1.2311009639776762, + "grad_norm": 5.5625, + "learning_rate": 3.5250503585158825e-06, + "loss": 1.65723991, + "memory(GiB)": 111.15, + "step": 48530, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.65357847, + "epoch": 1.2312278031456114, + "grad_norm": 4.9375, + "learning_rate": 3.5240484306551615e-06, + "loss": 1.60450497, + "memory(GiB)": 111.15, + "step": 48535, + "train_speed(iter/s)": 0.447323 + }, + { + "acc": 0.6513917, + "epoch": 1.2313546423135464, + "grad_norm": 6.0, + "learning_rate": 3.5230465677133813e-06, + "loss": 1.68011398, + "memory(GiB)": 111.15, + "step": 48540, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.64177494, + "epoch": 1.2314814814814814, + "grad_norm": 5.59375, + "learning_rate": 3.5220447697346063e-06, + "loss": 1.66310463, + "memory(GiB)": 111.15, + "step": 48545, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.64201803, + "epoch": 1.2316083206494166, + "grad_norm": 6.40625, + "learning_rate": 3.521043036762903e-06, + "loss": 1.6947319, + "memory(GiB)": 111.15, + "step": 48550, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.64817438, + "epoch": 1.2317351598173516, + "grad_norm": 4.875, + "learning_rate": 3.5200413688423284e-06, + "loss": 1.60929165, + "memory(GiB)": 111.15, + "step": 48555, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.64759083, + "epoch": 1.2318619989852866, + "grad_norm": 5.875, + "learning_rate": 3.519039766016943e-06, + "loss": 1.66282883, + "memory(GiB)": 111.15, + "step": 48560, + "train_speed(iter/s)": 0.447446 + }, + { + "acc": 0.64191623, + "epoch": 1.2319888381532218, + "grad_norm": 5.40625, + "learning_rate": 3.5180382283307983e-06, + "loss": 1.73141937, + "memory(GiB)": 111.15, + "step": 48565, + "train_speed(iter/s)": 0.447471 + }, + { + "acc": 0.66150303, + "epoch": 1.2321156773211568, + "grad_norm": 4.75, + "learning_rate": 3.517036755827952e-06, + "loss": 1.56300755, + "memory(GiB)": 111.15, + "step": 48570, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.66354017, + "epoch": 1.2322425164890918, + "grad_norm": 4.875, + "learning_rate": 3.516035348552449e-06, + "loss": 1.60456715, + "memory(GiB)": 111.15, + "step": 48575, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.65221281, + "epoch": 1.2323693556570268, + "grad_norm": 5.375, + "learning_rate": 3.515034006548335e-06, + "loss": 1.6315712, + "memory(GiB)": 111.15, + "step": 48580, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.66313457, + "epoch": 1.232496194824962, + "grad_norm": 6.15625, + "learning_rate": 3.5140327298596565e-06, + "loss": 1.62119503, + "memory(GiB)": 111.15, + "step": 48585, + "train_speed(iter/s)": 0.44757 + }, + { + "acc": 0.65094929, + "epoch": 1.232623033992897, + "grad_norm": 4.75, + "learning_rate": 3.5130315185304547e-06, + "loss": 1.60042706, + "memory(GiB)": 111.15, + "step": 48590, + "train_speed(iter/s)": 0.447595 + }, + { + "acc": 0.64624205, + "epoch": 1.2327498731608322, + "grad_norm": 5.28125, + "learning_rate": 3.5120303726047642e-06, + "loss": 1.62732277, + "memory(GiB)": 111.15, + "step": 48595, + "train_speed(iter/s)": 0.447619 + }, + { + "acc": 0.64343548, + "epoch": 1.2328767123287672, + "grad_norm": 5.1875, + "learning_rate": 3.51102929212662e-06, + "loss": 1.70556946, + "memory(GiB)": 111.15, + "step": 48600, + "train_speed(iter/s)": 0.447644 + }, + { + "epoch": 1.2328767123287672, + "eval_acc": 0.6433775317221162, + "eval_loss": 1.6076533794403076, + "eval_runtime": 113.4499, + "eval_samples_per_second": 56.148, + "eval_steps_per_second": 28.074, + "step": 48600 + }, + { + "acc": 0.64466639, + "epoch": 1.2330035514967022, + "grad_norm": 5.28125, + "learning_rate": 3.5100282771400563e-06, + "loss": 1.67453079, + "memory(GiB)": 111.15, + "step": 48605, + "train_speed(iter/s)": 0.447182 + }, + { + "acc": 0.6420476, + "epoch": 1.2331303906646371, + "grad_norm": 4.6875, + "learning_rate": 3.5090273276891023e-06, + "loss": 1.63339481, + "memory(GiB)": 111.15, + "step": 48610, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.65115547, + "epoch": 1.2332572298325724, + "grad_norm": 5.0, + "learning_rate": 3.5080264438177815e-06, + "loss": 1.59460316, + "memory(GiB)": 111.15, + "step": 48615, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.6451108, + "epoch": 1.2333840690005073, + "grad_norm": 4.9375, + "learning_rate": 3.50702562557012e-06, + "loss": 1.68040962, + "memory(GiB)": 111.15, + "step": 48620, + "train_speed(iter/s)": 0.447255 + }, + { + "acc": 0.64252353, + "epoch": 1.2335109081684423, + "grad_norm": 4.46875, + "learning_rate": 3.506024872990135e-06, + "loss": 1.64815178, + "memory(GiB)": 111.15, + "step": 48625, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.65641232, + "epoch": 1.2336377473363775, + "grad_norm": 5.375, + "learning_rate": 3.5050241861218493e-06, + "loss": 1.59169426, + "memory(GiB)": 111.15, + "step": 48630, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.65107546, + "epoch": 1.2337645865043125, + "grad_norm": 5.15625, + "learning_rate": 3.5040235650092725e-06, + "loss": 1.60214787, + "memory(GiB)": 111.15, + "step": 48635, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.67317133, + "epoch": 1.2338914256722475, + "grad_norm": 5.78125, + "learning_rate": 3.503023009696419e-06, + "loss": 1.50187607, + "memory(GiB)": 111.15, + "step": 48640, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.64834929, + "epoch": 1.2340182648401827, + "grad_norm": 5.34375, + "learning_rate": 3.5020225202272963e-06, + "loss": 1.66751289, + "memory(GiB)": 111.15, + "step": 48645, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.63368406, + "epoch": 1.2341451040081177, + "grad_norm": 4.96875, + "learning_rate": 3.501022096645913e-06, + "loss": 1.66699142, + "memory(GiB)": 111.15, + "step": 48650, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.64129848, + "epoch": 1.2342719431760527, + "grad_norm": 5.625, + "learning_rate": 3.5000217389962685e-06, + "loss": 1.68236542, + "memory(GiB)": 111.15, + "step": 48655, + "train_speed(iter/s)": 0.447425 + }, + { + "acc": 0.64692621, + "epoch": 1.234398782343988, + "grad_norm": 4.78125, + "learning_rate": 3.499021447322365e-06, + "loss": 1.68555813, + "memory(GiB)": 111.15, + "step": 48660, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.65333366, + "epoch": 1.234525621511923, + "grad_norm": 4.75, + "learning_rate": 3.4980212216681997e-06, + "loss": 1.63369255, + "memory(GiB)": 111.15, + "step": 48665, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.65068264, + "epoch": 1.234652460679858, + "grad_norm": 4.625, + "learning_rate": 3.4970210620777687e-06, + "loss": 1.62500362, + "memory(GiB)": 111.15, + "step": 48670, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.66776838, + "epoch": 1.2347792998477929, + "grad_norm": 4.625, + "learning_rate": 3.496020968595059e-06, + "loss": 1.60911884, + "memory(GiB)": 111.15, + "step": 48675, + "train_speed(iter/s)": 0.447522 + }, + { + "acc": 0.66511164, + "epoch": 1.234906139015728, + "grad_norm": 4.40625, + "learning_rate": 3.4950209412640634e-06, + "loss": 1.5661068, + "memory(GiB)": 111.15, + "step": 48680, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.66271653, + "epoch": 1.235032978183663, + "grad_norm": 5.09375, + "learning_rate": 3.494020980128766e-06, + "loss": 1.55168858, + "memory(GiB)": 111.15, + "step": 48685, + "train_speed(iter/s)": 0.447571 + }, + { + "acc": 0.65602899, + "epoch": 1.235159817351598, + "grad_norm": 6.15625, + "learning_rate": 3.4930210852331505e-06, + "loss": 1.53536873, + "memory(GiB)": 111.15, + "step": 48690, + "train_speed(iter/s)": 0.447596 + }, + { + "acc": 0.6568099, + "epoch": 1.2352866565195333, + "grad_norm": 5.78125, + "learning_rate": 3.4920212566211943e-06, + "loss": 1.51329212, + "memory(GiB)": 111.15, + "step": 48695, + "train_speed(iter/s)": 0.44762 + }, + { + "acc": 0.6457448, + "epoch": 1.2354134956874683, + "grad_norm": 5.09375, + "learning_rate": 3.491021494336876e-06, + "loss": 1.60937843, + "memory(GiB)": 111.15, + "step": 48700, + "train_speed(iter/s)": 0.447645 + }, + { + "epoch": 1.2354134956874683, + "eval_acc": 0.6433716829797839, + "eval_loss": 1.6076589822769165, + "eval_runtime": 113.2674, + "eval_samples_per_second": 56.239, + "eval_steps_per_second": 28.119, + "step": 48700 + }, + { + "acc": 0.64705515, + "epoch": 1.2355403348554033, + "grad_norm": 5.25, + "learning_rate": 3.4900217984241692e-06, + "loss": 1.63211498, + "memory(GiB)": 111.15, + "step": 48705, + "train_speed(iter/s)": 0.447185 + }, + { + "acc": 0.64690456, + "epoch": 1.2356671740233385, + "grad_norm": 7.5, + "learning_rate": 3.4890221689270466e-06, + "loss": 1.62593212, + "memory(GiB)": 111.15, + "step": 48710, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.6611805, + "epoch": 1.2357940131912735, + "grad_norm": 5.6875, + "learning_rate": 3.488022605889475e-06, + "loss": 1.61149101, + "memory(GiB)": 111.15, + "step": 48715, + "train_speed(iter/s)": 0.447234 + }, + { + "acc": 0.65489163, + "epoch": 1.2359208523592085, + "grad_norm": 5.5625, + "learning_rate": 3.4870231093554172e-06, + "loss": 1.56131058, + "memory(GiB)": 111.15, + "step": 48720, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.64137802, + "epoch": 1.2360476915271437, + "grad_norm": 4.9375, + "learning_rate": 3.4860236793688407e-06, + "loss": 1.65983181, + "memory(GiB)": 111.15, + "step": 48725, + "train_speed(iter/s)": 0.447283 + }, + { + "acc": 0.64410429, + "epoch": 1.2361745306950787, + "grad_norm": 5.625, + "learning_rate": 3.4850243159737024e-06, + "loss": 1.63901558, + "memory(GiB)": 111.15, + "step": 48730, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.64913092, + "epoch": 1.2363013698630136, + "grad_norm": 5.03125, + "learning_rate": 3.4840250192139574e-06, + "loss": 1.64675503, + "memory(GiB)": 111.15, + "step": 48735, + "train_speed(iter/s)": 0.447332 + }, + { + "acc": 0.66435585, + "epoch": 1.2364282090309486, + "grad_norm": 5.8125, + "learning_rate": 3.4830257891335595e-06, + "loss": 1.54313917, + "memory(GiB)": 111.15, + "step": 48740, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.65373578, + "epoch": 1.2365550481988838, + "grad_norm": 4.625, + "learning_rate": 3.4820266257764613e-06, + "loss": 1.52093887, + "memory(GiB)": 111.15, + "step": 48745, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.66371965, + "epoch": 1.2366818873668188, + "grad_norm": 4.71875, + "learning_rate": 3.4810275291866103e-06, + "loss": 1.57616272, + "memory(GiB)": 111.15, + "step": 48750, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.64770765, + "epoch": 1.236808726534754, + "grad_norm": 5.34375, + "learning_rate": 3.4800284994079487e-06, + "loss": 1.60170422, + "memory(GiB)": 111.15, + "step": 48755, + "train_speed(iter/s)": 0.447431 + }, + { + "acc": 0.6428175, + "epoch": 1.236935565702689, + "grad_norm": 5.125, + "learning_rate": 3.4790295364844207e-06, + "loss": 1.6957243, + "memory(GiB)": 111.15, + "step": 48760, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64462008, + "epoch": 1.237062404870624, + "grad_norm": 5.0, + "learning_rate": 3.4780306404599628e-06, + "loss": 1.64175396, + "memory(GiB)": 111.15, + "step": 48765, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.64029064, + "epoch": 1.237189244038559, + "grad_norm": 6.375, + "learning_rate": 3.4770318113785164e-06, + "loss": 1.71056328, + "memory(GiB)": 111.15, + "step": 48770, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.64581404, + "epoch": 1.2373160832064942, + "grad_norm": 5.96875, + "learning_rate": 3.4760330492840065e-06, + "loss": 1.6722332, + "memory(GiB)": 111.15, + "step": 48775, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.65162172, + "epoch": 1.2374429223744292, + "grad_norm": 5.4375, + "learning_rate": 3.4750343542203684e-06, + "loss": 1.63969269, + "memory(GiB)": 111.15, + "step": 48780, + "train_speed(iter/s)": 0.447554 + }, + { + "acc": 0.64893274, + "epoch": 1.2375697615423642, + "grad_norm": 5.34375, + "learning_rate": 3.474035726231527e-06, + "loss": 1.66862717, + "memory(GiB)": 111.15, + "step": 48785, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.66153164, + "epoch": 1.2376966007102994, + "grad_norm": 4.78125, + "learning_rate": 3.473037165361409e-06, + "loss": 1.64559212, + "memory(GiB)": 111.15, + "step": 48790, + "train_speed(iter/s)": 0.447603 + }, + { + "acc": 0.65718732, + "epoch": 1.2378234398782344, + "grad_norm": 5.78125, + "learning_rate": 3.4720386716539333e-06, + "loss": 1.62682304, + "memory(GiB)": 111.15, + "step": 48795, + "train_speed(iter/s)": 0.447628 + }, + { + "acc": 0.65095978, + "epoch": 1.2379502790461694, + "grad_norm": 4.9375, + "learning_rate": 3.471040245153018e-06, + "loss": 1.63098545, + "memory(GiB)": 111.15, + "step": 48800, + "train_speed(iter/s)": 0.447652 + }, + { + "epoch": 1.2379502790461694, + "eval_acc": 0.643369176375927, + "eval_loss": 1.6076757907867432, + "eval_runtime": 113.9231, + "eval_samples_per_second": 55.915, + "eval_steps_per_second": 27.957, + "step": 48800 + }, + { + "acc": 0.64742002, + "epoch": 1.2380771182141046, + "grad_norm": 5.0, + "learning_rate": 3.4700418859025793e-06, + "loss": 1.65631065, + "memory(GiB)": 111.15, + "step": 48805, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.65148692, + "epoch": 1.2382039573820396, + "grad_norm": 6.15625, + "learning_rate": 3.4690435939465307e-06, + "loss": 1.61284256, + "memory(GiB)": 111.15, + "step": 48810, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.64553642, + "epoch": 1.2383307965499746, + "grad_norm": 4.6875, + "learning_rate": 3.4680453693287786e-06, + "loss": 1.57030849, + "memory(GiB)": 111.15, + "step": 48815, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.6467567, + "epoch": 1.2384576357179098, + "grad_norm": 6.53125, + "learning_rate": 3.4670472120932297e-06, + "loss": 1.6541893, + "memory(GiB)": 111.15, + "step": 48820, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.64933043, + "epoch": 1.2385844748858448, + "grad_norm": 5.59375, + "learning_rate": 3.46604912228379e-06, + "loss": 1.65247498, + "memory(GiB)": 111.15, + "step": 48825, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.65466018, + "epoch": 1.2387113140537798, + "grad_norm": 5.65625, + "learning_rate": 3.46505109994436e-06, + "loss": 1.62160606, + "memory(GiB)": 111.15, + "step": 48830, + "train_speed(iter/s)": 0.447311 + }, + { + "acc": 0.65206327, + "epoch": 1.2388381532217148, + "grad_norm": 5.1875, + "learning_rate": 3.464053145118833e-06, + "loss": 1.54203215, + "memory(GiB)": 111.15, + "step": 48835, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.65799227, + "epoch": 1.23896499238965, + "grad_norm": 6.125, + "learning_rate": 3.4630552578511073e-06, + "loss": 1.61441498, + "memory(GiB)": 111.15, + "step": 48840, + "train_speed(iter/s)": 0.44736 + }, + { + "acc": 0.64305401, + "epoch": 1.239091831557585, + "grad_norm": 5.5, + "learning_rate": 3.4620574381850723e-06, + "loss": 1.63160305, + "memory(GiB)": 111.15, + "step": 48845, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.64893341, + "epoch": 1.23921867072552, + "grad_norm": 5.125, + "learning_rate": 3.4610596861646194e-06, + "loss": 1.64888382, + "memory(GiB)": 111.15, + "step": 48850, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.65980868, + "epoch": 1.2393455098934552, + "grad_norm": 5.6875, + "learning_rate": 3.460062001833632e-06, + "loss": 1.58177738, + "memory(GiB)": 111.15, + "step": 48855, + "train_speed(iter/s)": 0.447433 + }, + { + "acc": 0.6769659, + "epoch": 1.2394723490613901, + "grad_norm": 6.0625, + "learning_rate": 3.459064385235993e-06, + "loss": 1.62156563, + "memory(GiB)": 111.15, + "step": 48860, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.64337158, + "epoch": 1.2395991882293251, + "grad_norm": 5.4375, + "learning_rate": 3.45806683641558e-06, + "loss": 1.59443493, + "memory(GiB)": 111.15, + "step": 48865, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.66179509, + "epoch": 1.2397260273972603, + "grad_norm": 4.75, + "learning_rate": 3.457069355416275e-06, + "loss": 1.55963287, + "memory(GiB)": 111.15, + "step": 48870, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.65426464, + "epoch": 1.2398528665651953, + "grad_norm": 5.90625, + "learning_rate": 3.456071942281947e-06, + "loss": 1.64134731, + "memory(GiB)": 111.15, + "step": 48875, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.63707943, + "epoch": 1.2399797057331303, + "grad_norm": 4.75, + "learning_rate": 3.455074597056467e-06, + "loss": 1.67736721, + "memory(GiB)": 111.15, + "step": 48880, + "train_speed(iter/s)": 0.447554 + }, + { + "acc": 0.65599403, + "epoch": 1.2401065449010655, + "grad_norm": 5.875, + "learning_rate": 3.454077319783705e-06, + "loss": 1.52524014, + "memory(GiB)": 111.15, + "step": 48885, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.63523231, + "epoch": 1.2402333840690005, + "grad_norm": 5.0625, + "learning_rate": 3.4530801105075257e-06, + "loss": 1.68555393, + "memory(GiB)": 111.15, + "step": 48890, + "train_speed(iter/s)": 0.447603 + }, + { + "acc": 0.64404144, + "epoch": 1.2403602232369355, + "grad_norm": 5.46875, + "learning_rate": 3.4520829692717874e-06, + "loss": 1.65428925, + "memory(GiB)": 111.15, + "step": 48895, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.63935881, + "epoch": 1.2404870624048705, + "grad_norm": 4.875, + "learning_rate": 3.451085896120352e-06, + "loss": 1.67799664, + "memory(GiB)": 111.15, + "step": 48900, + "train_speed(iter/s)": 0.447651 + }, + { + "epoch": 1.2404870624048705, + "eval_acc": 0.6433562255893339, + "eval_loss": 1.6077183485031128, + "eval_runtime": 113.1893, + "eval_samples_per_second": 56.277, + "eval_steps_per_second": 28.139, + "step": 48900 + }, + { + "acc": 0.64526196, + "epoch": 1.2406139015728057, + "grad_norm": 4.875, + "learning_rate": 3.450088891097074e-06, + "loss": 1.61622143, + "memory(GiB)": 111.15, + "step": 48905, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.64613528, + "epoch": 1.2407407407407407, + "grad_norm": 5.40625, + "learning_rate": 3.4490919542458085e-06, + "loss": 1.62443962, + "memory(GiB)": 111.15, + "step": 48910, + "train_speed(iter/s)": 0.447217 + }, + { + "acc": 0.6288744, + "epoch": 1.240867579908676, + "grad_norm": 4.90625, + "learning_rate": 3.4480950856104002e-06, + "loss": 1.69640293, + "memory(GiB)": 111.15, + "step": 48915, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.64367266, + "epoch": 1.240994419076611, + "grad_norm": 5.53125, + "learning_rate": 3.4470982852347e-06, + "loss": 1.63576107, + "memory(GiB)": 111.15, + "step": 48920, + "train_speed(iter/s)": 0.447266 + }, + { + "acc": 0.63034573, + "epoch": 1.2411212582445459, + "grad_norm": 5.375, + "learning_rate": 3.44610155316255e-06, + "loss": 1.68067989, + "memory(GiB)": 111.15, + "step": 48925, + "train_speed(iter/s)": 0.447291 + }, + { + "acc": 0.65280385, + "epoch": 1.2412480974124809, + "grad_norm": 5.375, + "learning_rate": 3.4451048894377925e-06, + "loss": 1.62335567, + "memory(GiB)": 111.15, + "step": 48930, + "train_speed(iter/s)": 0.447315 + }, + { + "acc": 0.65480399, + "epoch": 1.241374936580416, + "grad_norm": 5.375, + "learning_rate": 3.444108294104264e-06, + "loss": 1.61859989, + "memory(GiB)": 111.15, + "step": 48935, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.66189294, + "epoch": 1.241501775748351, + "grad_norm": 4.875, + "learning_rate": 3.443111767205797e-06, + "loss": 1.61389885, + "memory(GiB)": 111.15, + "step": 48940, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.64492064, + "epoch": 1.241628614916286, + "grad_norm": 6.1875, + "learning_rate": 3.442115308786227e-06, + "loss": 1.6649395, + "memory(GiB)": 111.15, + "step": 48945, + "train_speed(iter/s)": 0.447388 + }, + { + "acc": 0.64809465, + "epoch": 1.2417554540842213, + "grad_norm": 6.375, + "learning_rate": 3.4411189188893822e-06, + "loss": 1.63323631, + "memory(GiB)": 111.15, + "step": 48950, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.64324331, + "epoch": 1.2418822932521563, + "grad_norm": 5.0, + "learning_rate": 3.4401225975590867e-06, + "loss": 1.6561779, + "memory(GiB)": 111.15, + "step": 48955, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.62490501, + "epoch": 1.2420091324200913, + "grad_norm": 4.3125, + "learning_rate": 3.439126344839163e-06, + "loss": 1.64347038, + "memory(GiB)": 111.15, + "step": 48960, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.65505581, + "epoch": 1.2421359715880265, + "grad_norm": 4.84375, + "learning_rate": 3.438130160773431e-06, + "loss": 1.60981789, + "memory(GiB)": 111.15, + "step": 48965, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.64646158, + "epoch": 1.2422628107559615, + "grad_norm": 4.6875, + "learning_rate": 3.43713404540571e-06, + "loss": 1.68303413, + "memory(GiB)": 111.15, + "step": 48970, + "train_speed(iter/s)": 0.44751 + }, + { + "acc": 0.65450583, + "epoch": 1.2423896499238964, + "grad_norm": 4.8125, + "learning_rate": 3.4361379987798094e-06, + "loss": 1.67326241, + "memory(GiB)": 111.15, + "step": 48975, + "train_speed(iter/s)": 0.447535 + }, + { + "acc": 0.65456457, + "epoch": 1.2425164890918317, + "grad_norm": 5.40625, + "learning_rate": 3.435142020939542e-06, + "loss": 1.60166702, + "memory(GiB)": 111.15, + "step": 48980, + "train_speed(iter/s)": 0.447559 + }, + { + "acc": 0.65886717, + "epoch": 1.2426433282597666, + "grad_norm": 5.84375, + "learning_rate": 3.4341461119287144e-06, + "loss": 1.58966904, + "memory(GiB)": 111.15, + "step": 48985, + "train_speed(iter/s)": 0.447583 + }, + { + "acc": 0.65887871, + "epoch": 1.2427701674277016, + "grad_norm": 6.0, + "learning_rate": 3.433150271791135e-06, + "loss": 1.64013176, + "memory(GiB)": 111.15, + "step": 48990, + "train_speed(iter/s)": 0.447607 + }, + { + "acc": 0.65224743, + "epoch": 1.2428970065956366, + "grad_norm": 5.90625, + "learning_rate": 3.432154500570599e-06, + "loss": 1.57339935, + "memory(GiB)": 111.15, + "step": 48995, + "train_speed(iter/s)": 0.447632 + }, + { + "acc": 0.65456362, + "epoch": 1.2430238457635718, + "grad_norm": 4.65625, + "learning_rate": 3.431158798310909e-06, + "loss": 1.60520725, + "memory(GiB)": 111.15, + "step": 49000, + "train_speed(iter/s)": 0.447656 + }, + { + "epoch": 1.2430238457635718, + "eval_acc": 0.6433775317221162, + "eval_loss": 1.6076040267944336, + "eval_runtime": 113.3308, + "eval_samples_per_second": 56.207, + "eval_steps_per_second": 28.104, + "step": 49000 + }, + { + "acc": 0.64319963, + "epoch": 1.2431506849315068, + "grad_norm": 5.625, + "learning_rate": 3.4301631650558588e-06, + "loss": 1.66964378, + "memory(GiB)": 111.15, + "step": 49005, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.64152808, + "epoch": 1.2432775240994418, + "grad_norm": 5.125, + "learning_rate": 3.4291676008492424e-06, + "loss": 1.6140274, + "memory(GiB)": 111.15, + "step": 49010, + "train_speed(iter/s)": 0.447223 + }, + { + "acc": 0.66194391, + "epoch": 1.243404363267377, + "grad_norm": 5.0, + "learning_rate": 3.428172105734848e-06, + "loss": 1.58728962, + "memory(GiB)": 111.15, + "step": 49015, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.65138168, + "epoch": 1.243531202435312, + "grad_norm": 5.3125, + "learning_rate": 3.4271766797564608e-06, + "loss": 1.66522026, + "memory(GiB)": 111.15, + "step": 49020, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.65441256, + "epoch": 1.243658041603247, + "grad_norm": 6.0625, + "learning_rate": 3.4261813229578665e-06, + "loss": 1.64024277, + "memory(GiB)": 111.15, + "step": 49025, + "train_speed(iter/s)": 0.447296 + }, + { + "acc": 0.66397886, + "epoch": 1.2437848807711822, + "grad_norm": 5.78125, + "learning_rate": 3.425186035382846e-06, + "loss": 1.58736181, + "memory(GiB)": 111.15, + "step": 49030, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.6532177, + "epoch": 1.2439117199391172, + "grad_norm": 5.78125, + "learning_rate": 3.4241908170751727e-06, + "loss": 1.60149059, + "memory(GiB)": 111.15, + "step": 49035, + "train_speed(iter/s)": 0.447344 + }, + { + "acc": 0.65128465, + "epoch": 1.2440385591070522, + "grad_norm": 5.34375, + "learning_rate": 3.4231956680786217e-06, + "loss": 1.55498686, + "memory(GiB)": 111.15, + "step": 49040, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.64979658, + "epoch": 1.2441653982749874, + "grad_norm": 5.125, + "learning_rate": 3.422200588436967e-06, + "loss": 1.64173603, + "memory(GiB)": 111.15, + "step": 49045, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.6390007, + "epoch": 1.2442922374429224, + "grad_norm": 4.34375, + "learning_rate": 3.4212055781939744e-06, + "loss": 1.65371819, + "memory(GiB)": 111.15, + "step": 49050, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.64368839, + "epoch": 1.2444190766108574, + "grad_norm": 6.03125, + "learning_rate": 3.4202106373934085e-06, + "loss": 1.68721008, + "memory(GiB)": 111.15, + "step": 49055, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.64833117, + "epoch": 1.2445459157787924, + "grad_norm": 5.625, + "learning_rate": 3.4192157660790324e-06, + "loss": 1.65247574, + "memory(GiB)": 111.15, + "step": 49060, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.63940954, + "epoch": 1.2446727549467276, + "grad_norm": 4.6875, + "learning_rate": 3.418220964294604e-06, + "loss": 1.65498924, + "memory(GiB)": 111.15, + "step": 49065, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.63992791, + "epoch": 1.2447995941146626, + "grad_norm": 5.125, + "learning_rate": 3.417226232083881e-06, + "loss": 1.65420341, + "memory(GiB)": 111.15, + "step": 49070, + "train_speed(iter/s)": 0.447514 + }, + { + "acc": 0.65059514, + "epoch": 1.2449264332825978, + "grad_norm": 5.4375, + "learning_rate": 3.416231569490615e-06, + "loss": 1.64563332, + "memory(GiB)": 111.15, + "step": 49075, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.68372536, + "epoch": 1.2450532724505328, + "grad_norm": 5.40625, + "learning_rate": 3.4152369765585545e-06, + "loss": 1.45453882, + "memory(GiB)": 111.15, + "step": 49080, + "train_speed(iter/s)": 0.447563 + }, + { + "acc": 0.65816212, + "epoch": 1.2451801116184678, + "grad_norm": 5.75, + "learning_rate": 3.4142424533314474e-06, + "loss": 1.59034824, + "memory(GiB)": 111.15, + "step": 49085, + "train_speed(iter/s)": 0.447587 + }, + { + "acc": 0.6326067, + "epoch": 1.2453069507864027, + "grad_norm": 5.375, + "learning_rate": 3.4132479998530383e-06, + "loss": 1.70315514, + "memory(GiB)": 111.15, + "step": 49090, + "train_speed(iter/s)": 0.447611 + }, + { + "acc": 0.64563961, + "epoch": 1.245433789954338, + "grad_norm": 4.9375, + "learning_rate": 3.4122536161670656e-06, + "loss": 1.63955822, + "memory(GiB)": 111.15, + "step": 49095, + "train_speed(iter/s)": 0.447636 + }, + { + "acc": 0.66026382, + "epoch": 1.245560629122273, + "grad_norm": 6.125, + "learning_rate": 3.411259302317267e-06, + "loss": 1.57469902, + "memory(GiB)": 111.15, + "step": 49100, + "train_speed(iter/s)": 0.44766 + }, + { + "epoch": 1.245560629122273, + "eval_acc": 0.6433817093952108, + "eval_loss": 1.6076385974884033, + "eval_runtime": 114.3337, + "eval_samples_per_second": 55.714, + "eval_steps_per_second": 27.857, + "step": 49100 + }, + { + "acc": 0.64499559, + "epoch": 1.245687468290208, + "grad_norm": 5.1875, + "learning_rate": 3.410265058347378e-06, + "loss": 1.58654079, + "memory(GiB)": 111.15, + "step": 49105, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.65859175, + "epoch": 1.2458143074581431, + "grad_norm": 5.5, + "learning_rate": 3.4092708843011303e-06, + "loss": 1.59958725, + "memory(GiB)": 111.15, + "step": 49110, + "train_speed(iter/s)": 0.447223 + }, + { + "acc": 0.64040709, + "epoch": 1.2459411466260781, + "grad_norm": 5.1875, + "learning_rate": 3.4082767802222493e-06, + "loss": 1.65055237, + "memory(GiB)": 111.15, + "step": 49115, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.64962206, + "epoch": 1.2460679857940131, + "grad_norm": 5.6875, + "learning_rate": 3.4072827461544635e-06, + "loss": 1.62768002, + "memory(GiB)": 111.15, + "step": 49120, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.64899106, + "epoch": 1.2461948249619483, + "grad_norm": 5.34375, + "learning_rate": 3.4062887821414935e-06, + "loss": 1.59830742, + "memory(GiB)": 111.15, + "step": 49125, + "train_speed(iter/s)": 0.447296 + }, + { + "acc": 0.64635859, + "epoch": 1.2463216641298833, + "grad_norm": 5.0625, + "learning_rate": 3.4052948882270585e-06, + "loss": 1.64351921, + "memory(GiB)": 111.15, + "step": 49130, + "train_speed(iter/s)": 0.447321 + }, + { + "acc": 0.64756603, + "epoch": 1.2464485032978183, + "grad_norm": 6.25, + "learning_rate": 3.404301064454873e-06, + "loss": 1.70827904, + "memory(GiB)": 111.15, + "step": 49135, + "train_speed(iter/s)": 0.447345 + }, + { + "acc": 0.64985676, + "epoch": 1.2465753424657535, + "grad_norm": 5.625, + "learning_rate": 3.4033073108686515e-06, + "loss": 1.65390434, + "memory(GiB)": 111.15, + "step": 49140, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.63497066, + "epoch": 1.2467021816336885, + "grad_norm": 4.96875, + "learning_rate": 3.4023136275121026e-06, + "loss": 1.64470329, + "memory(GiB)": 111.15, + "step": 49145, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.63120346, + "epoch": 1.2468290208016235, + "grad_norm": 6.34375, + "learning_rate": 3.401320014428935e-06, + "loss": 1.75466042, + "memory(GiB)": 111.15, + "step": 49150, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.65225224, + "epoch": 1.2469558599695585, + "grad_norm": 5.0, + "learning_rate": 3.40032647166285e-06, + "loss": 1.60120087, + "memory(GiB)": 111.15, + "step": 49155, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.66244841, + "epoch": 1.2470826991374937, + "grad_norm": 4.78125, + "learning_rate": 3.3993329992575473e-06, + "loss": 1.56963711, + "memory(GiB)": 111.15, + "step": 49160, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.63989258, + "epoch": 1.2472095383054287, + "grad_norm": 5.15625, + "learning_rate": 3.3983395972567277e-06, + "loss": 1.64743423, + "memory(GiB)": 111.15, + "step": 49165, + "train_speed(iter/s)": 0.447491 + }, + { + "acc": 0.65267544, + "epoch": 1.2473363774733637, + "grad_norm": 5.375, + "learning_rate": 3.397346265704084e-06, + "loss": 1.57650585, + "memory(GiB)": 111.15, + "step": 49170, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.66418056, + "epoch": 1.2474632166412989, + "grad_norm": 5.9375, + "learning_rate": 3.396353004643306e-06, + "loss": 1.56908693, + "memory(GiB)": 111.15, + "step": 49175, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.64381652, + "epoch": 1.2475900558092339, + "grad_norm": 5.0, + "learning_rate": 3.3953598141180817e-06, + "loss": 1.6576685, + "memory(GiB)": 111.15, + "step": 49180, + "train_speed(iter/s)": 0.447564 + }, + { + "acc": 0.65605912, + "epoch": 1.2477168949771689, + "grad_norm": 7.78125, + "learning_rate": 3.3943666941720978e-06, + "loss": 1.59639292, + "memory(GiB)": 111.15, + "step": 49185, + "train_speed(iter/s)": 0.447588 + }, + { + "acc": 0.64956937, + "epoch": 1.247843734145104, + "grad_norm": 5.125, + "learning_rate": 3.3933736448490363e-06, + "loss": 1.64381027, + "memory(GiB)": 111.15, + "step": 49190, + "train_speed(iter/s)": 0.447612 + }, + { + "acc": 0.65077314, + "epoch": 1.247970573313039, + "grad_norm": 7.3125, + "learning_rate": 3.392380666192573e-06, + "loss": 1.62496681, + "memory(GiB)": 111.15, + "step": 49195, + "train_speed(iter/s)": 0.447636 + }, + { + "acc": 0.64216852, + "epoch": 1.248097412480974, + "grad_norm": 5.09375, + "learning_rate": 3.391387758246386e-06, + "loss": 1.6827528, + "memory(GiB)": 111.15, + "step": 49200, + "train_speed(iter/s)": 0.447661 + }, + { + "epoch": 1.248097412480974, + "eval_acc": 0.6433708474451649, + "eval_loss": 1.6076924800872803, + "eval_runtime": 113.2839, + "eval_samples_per_second": 56.23, + "eval_steps_per_second": 28.115, + "step": 49200 + }, + { + "acc": 0.66762838, + "epoch": 1.2482242516489093, + "grad_norm": 5.15625, + "learning_rate": 3.3903949210541477e-06, + "loss": 1.48552923, + "memory(GiB)": 111.15, + "step": 49205, + "train_speed(iter/s)": 0.447205 + }, + { + "acc": 0.65161018, + "epoch": 1.2483510908168443, + "grad_norm": 4.6875, + "learning_rate": 3.389402154659529e-06, + "loss": 1.62856903, + "memory(GiB)": 111.15, + "step": 49210, + "train_speed(iter/s)": 0.447229 + }, + { + "acc": 0.63972769, + "epoch": 1.2484779299847792, + "grad_norm": 6.0625, + "learning_rate": 3.388409459106192e-06, + "loss": 1.68159561, + "memory(GiB)": 111.15, + "step": 49215, + "train_speed(iter/s)": 0.447253 + }, + { + "acc": 0.64183111, + "epoch": 1.2486047691527142, + "grad_norm": 5.09375, + "learning_rate": 3.3874168344378024e-06, + "loss": 1.59677792, + "memory(GiB)": 111.15, + "step": 49220, + "train_speed(iter/s)": 0.447277 + }, + { + "acc": 0.65485888, + "epoch": 1.2487316083206494, + "grad_norm": 5.75, + "learning_rate": 3.38642428069802e-06, + "loss": 1.58704414, + "memory(GiB)": 111.15, + "step": 49225, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.66307507, + "epoch": 1.2488584474885844, + "grad_norm": 6.59375, + "learning_rate": 3.385431797930503e-06, + "loss": 1.63087082, + "memory(GiB)": 111.15, + "step": 49230, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.64457331, + "epoch": 1.2489852866565196, + "grad_norm": 4.90625, + "learning_rate": 3.3844393861789036e-06, + "loss": 1.68529167, + "memory(GiB)": 111.15, + "step": 49235, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.64283681, + "epoch": 1.2491121258244546, + "grad_norm": 6.40625, + "learning_rate": 3.383447045486872e-06, + "loss": 1.69280281, + "memory(GiB)": 111.15, + "step": 49240, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.63322163, + "epoch": 1.2492389649923896, + "grad_norm": 4.84375, + "learning_rate": 3.382454775898057e-06, + "loss": 1.69146004, + "memory(GiB)": 111.15, + "step": 49245, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.65584135, + "epoch": 1.2493658041603246, + "grad_norm": 6.125, + "learning_rate": 3.381462577456104e-06, + "loss": 1.62672729, + "memory(GiB)": 111.15, + "step": 49250, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.65596514, + "epoch": 1.2494926433282598, + "grad_norm": 5.1875, + "learning_rate": 3.3804704502046527e-06, + "loss": 1.58230486, + "memory(GiB)": 111.15, + "step": 49255, + "train_speed(iter/s)": 0.447446 + }, + { + "acc": 0.66090832, + "epoch": 1.2496194824961948, + "grad_norm": 4.96875, + "learning_rate": 3.3794783941873406e-06, + "loss": 1.58688755, + "memory(GiB)": 111.15, + "step": 49260, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.63651228, + "epoch": 1.2497463216641298, + "grad_norm": 6.8125, + "learning_rate": 3.3784864094478044e-06, + "loss": 1.68297882, + "memory(GiB)": 111.15, + "step": 49265, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.64228449, + "epoch": 1.249873160832065, + "grad_norm": 4.90625, + "learning_rate": 3.377494496029677e-06, + "loss": 1.62428322, + "memory(GiB)": 111.15, + "step": 49270, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.65510368, + "epoch": 1.25, + "grad_norm": 5.4375, + "learning_rate": 3.3765026539765832e-06, + "loss": 1.60106888, + "memory(GiB)": 111.15, + "step": 49275, + "train_speed(iter/s)": 0.447542 + }, + { + "acc": 0.64825206, + "epoch": 1.250126839167935, + "grad_norm": 5.625, + "learning_rate": 3.375510883332152e-06, + "loss": 1.66017799, + "memory(GiB)": 111.15, + "step": 49280, + "train_speed(iter/s)": 0.447566 + }, + { + "acc": 0.63165236, + "epoch": 1.2502536783358702, + "grad_norm": 4.90625, + "learning_rate": 3.3745191841400037e-06, + "loss": 1.63583183, + "memory(GiB)": 111.15, + "step": 49285, + "train_speed(iter/s)": 0.44759 + }, + { + "acc": 0.64153414, + "epoch": 1.2503805175038052, + "grad_norm": 4.78125, + "learning_rate": 3.373527556443762e-06, + "loss": 1.67360916, + "memory(GiB)": 111.15, + "step": 49290, + "train_speed(iter/s)": 0.447614 + }, + { + "acc": 0.64354115, + "epoch": 1.2505073566717402, + "grad_norm": 5.0, + "learning_rate": 3.372536000287038e-06, + "loss": 1.62333031, + "memory(GiB)": 111.15, + "step": 49295, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.6555388, + "epoch": 1.2506341958396754, + "grad_norm": 6.65625, + "learning_rate": 3.3715445157134474e-06, + "loss": 1.59570789, + "memory(GiB)": 111.15, + "step": 49300, + "train_speed(iter/s)": 0.447662 + }, + { + "epoch": 1.2506341958396754, + "eval_acc": 0.6434372724473686, + "eval_loss": 1.607690453529358, + "eval_runtime": 113.0943, + "eval_samples_per_second": 56.325, + "eval_steps_per_second": 28.162, + "step": 49300 + }, + { + "acc": 0.64631157, + "epoch": 1.2507610350076104, + "grad_norm": 5.0625, + "learning_rate": 3.370553102766598e-06, + "loss": 1.58627701, + "memory(GiB)": 111.15, + "step": 49305, + "train_speed(iter/s)": 0.447208 + }, + { + "acc": 0.64021306, + "epoch": 1.2508878741755454, + "grad_norm": 4.75, + "learning_rate": 3.369561761490101e-06, + "loss": 1.65874596, + "memory(GiB)": 111.15, + "step": 49310, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.64964557, + "epoch": 1.2510147133434804, + "grad_norm": 5.59375, + "learning_rate": 3.3685704919275553e-06, + "loss": 1.65846748, + "memory(GiB)": 111.15, + "step": 49315, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.66633897, + "epoch": 1.2511415525114156, + "grad_norm": 5.5, + "learning_rate": 3.3675792941225625e-06, + "loss": 1.53507061, + "memory(GiB)": 111.15, + "step": 49320, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.65372977, + "epoch": 1.2512683916793506, + "grad_norm": 6.1875, + "learning_rate": 3.3665881681187214e-06, + "loss": 1.63561668, + "memory(GiB)": 111.15, + "step": 49325, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.66151257, + "epoch": 1.2513952308472858, + "grad_norm": 6.0, + "learning_rate": 3.3655971139596265e-06, + "loss": 1.63202095, + "memory(GiB)": 111.15, + "step": 49330, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.64431915, + "epoch": 1.2515220700152208, + "grad_norm": 5.1875, + "learning_rate": 3.3646061316888655e-06, + "loss": 1.65445023, + "memory(GiB)": 111.15, + "step": 49335, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.64349756, + "epoch": 1.2516489091831557, + "grad_norm": 6.78125, + "learning_rate": 3.3636152213500295e-06, + "loss": 1.69140091, + "memory(GiB)": 111.15, + "step": 49340, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.64340553, + "epoch": 1.2517757483510907, + "grad_norm": 6.125, + "learning_rate": 3.362624382986702e-06, + "loss": 1.63783226, + "memory(GiB)": 111.15, + "step": 49345, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.65807638, + "epoch": 1.251902587519026, + "grad_norm": 4.75, + "learning_rate": 3.3616336166424653e-06, + "loss": 1.63149414, + "memory(GiB)": 111.15, + "step": 49350, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.66561518, + "epoch": 1.252029426686961, + "grad_norm": 5.5625, + "learning_rate": 3.360642922360895e-06, + "loss": 1.52724924, + "memory(GiB)": 111.15, + "step": 49355, + "train_speed(iter/s)": 0.447448 + }, + { + "acc": 0.63456125, + "epoch": 1.252156265854896, + "grad_norm": 5.375, + "learning_rate": 3.3596523001855684e-06, + "loss": 1.72871857, + "memory(GiB)": 111.15, + "step": 49360, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.6618794, + "epoch": 1.2522831050228311, + "grad_norm": 5.0625, + "learning_rate": 3.358661750160057e-06, + "loss": 1.53018379, + "memory(GiB)": 111.15, + "step": 49365, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.64800406, + "epoch": 1.2524099441907661, + "grad_norm": 4.4375, + "learning_rate": 3.3576712723279326e-06, + "loss": 1.59512291, + "memory(GiB)": 111.15, + "step": 49370, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.64051256, + "epoch": 1.2525367833587011, + "grad_norm": 5.125, + "learning_rate": 3.3566808667327566e-06, + "loss": 1.71679211, + "memory(GiB)": 111.15, + "step": 49375, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.66605201, + "epoch": 1.252663622526636, + "grad_norm": 5.75, + "learning_rate": 3.355690533418091e-06, + "loss": 1.59448929, + "memory(GiB)": 111.15, + "step": 49380, + "train_speed(iter/s)": 0.447569 + }, + { + "acc": 0.64036608, + "epoch": 1.2527904616945713, + "grad_norm": 4.78125, + "learning_rate": 3.354700272427499e-06, + "loss": 1.68655624, + "memory(GiB)": 111.15, + "step": 49385, + "train_speed(iter/s)": 0.447593 + }, + { + "acc": 0.65051999, + "epoch": 1.2529173008625063, + "grad_norm": 5.1875, + "learning_rate": 3.3537100838045356e-06, + "loss": 1.66414127, + "memory(GiB)": 111.15, + "step": 49390, + "train_speed(iter/s)": 0.447617 + }, + { + "acc": 0.65529194, + "epoch": 1.2530441400304415, + "grad_norm": 6.1875, + "learning_rate": 3.3527199675927526e-06, + "loss": 1.64701462, + "memory(GiB)": 111.15, + "step": 49395, + "train_speed(iter/s)": 0.447641 + }, + { + "acc": 0.64048071, + "epoch": 1.2531709791983765, + "grad_norm": 5.59375, + "learning_rate": 3.3517299238356982e-06, + "loss": 1.69436646, + "memory(GiB)": 111.15, + "step": 49400, + "train_speed(iter/s)": 0.447666 + }, + { + "epoch": 1.2531709791983765, + "eval_acc": 0.6433721007470933, + "eval_loss": 1.6077090501785278, + "eval_runtime": 113.4022, + "eval_samples_per_second": 56.172, + "eval_steps_per_second": 28.086, + "step": 49400 + }, + { + "acc": 0.65228214, + "epoch": 1.2532978183663115, + "grad_norm": 4.84375, + "learning_rate": 3.3507399525769214e-06, + "loss": 1.58131704, + "memory(GiB)": 111.15, + "step": 49405, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.638311, + "epoch": 1.2534246575342465, + "grad_norm": 5.875, + "learning_rate": 3.3497500538599664e-06, + "loss": 1.6939537, + "memory(GiB)": 111.15, + "step": 49410, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.64646006, + "epoch": 1.2535514967021817, + "grad_norm": 5.125, + "learning_rate": 3.34876022772837e-06, + "loss": 1.65379753, + "memory(GiB)": 111.15, + "step": 49415, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.65120029, + "epoch": 1.2536783358701167, + "grad_norm": 5.15625, + "learning_rate": 3.347770474225672e-06, + "loss": 1.66930695, + "memory(GiB)": 111.15, + "step": 49420, + "train_speed(iter/s)": 0.447283 + }, + { + "acc": 0.65045428, + "epoch": 1.2538051750380519, + "grad_norm": 5.25, + "learning_rate": 3.3467807933954034e-06, + "loss": 1.60740013, + "memory(GiB)": 111.15, + "step": 49425, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.64233446, + "epoch": 1.2539320142059869, + "grad_norm": 4.625, + "learning_rate": 3.345791185281101e-06, + "loss": 1.62311172, + "memory(GiB)": 111.15, + "step": 49430, + "train_speed(iter/s)": 0.447331 + }, + { + "acc": 0.66269822, + "epoch": 1.2540588533739219, + "grad_norm": 5.5, + "learning_rate": 3.3448016499262836e-06, + "loss": 1.61623363, + "memory(GiB)": 111.15, + "step": 49435, + "train_speed(iter/s)": 0.447355 + }, + { + "acc": 0.643612, + "epoch": 1.2541856925418569, + "grad_norm": 5.5, + "learning_rate": 3.3438121873744812e-06, + "loss": 1.66169243, + "memory(GiB)": 111.15, + "step": 49440, + "train_speed(iter/s)": 0.447379 + }, + { + "acc": 0.64761777, + "epoch": 1.254312531709792, + "grad_norm": 4.875, + "learning_rate": 3.342822797669212e-06, + "loss": 1.59341545, + "memory(GiB)": 111.15, + "step": 49445, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.65080962, + "epoch": 1.254439370877727, + "grad_norm": 5.5625, + "learning_rate": 3.3418334808539966e-06, + "loss": 1.69867363, + "memory(GiB)": 111.15, + "step": 49450, + "train_speed(iter/s)": 0.447427 + }, + { + "acc": 0.63429437, + "epoch": 1.254566210045662, + "grad_norm": 4.96875, + "learning_rate": 3.340844236972347e-06, + "loss": 1.69941521, + "memory(GiB)": 111.15, + "step": 49455, + "train_speed(iter/s)": 0.447451 + }, + { + "acc": 0.652666, + "epoch": 1.2546930492135973, + "grad_norm": 5.875, + "learning_rate": 3.3398550660677748e-06, + "loss": 1.63648682, + "memory(GiB)": 111.15, + "step": 49460, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.63806076, + "epoch": 1.2548198883815322, + "grad_norm": 6.84375, + "learning_rate": 3.3388659681837898e-06, + "loss": 1.66788216, + "memory(GiB)": 111.15, + "step": 49465, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.63527985, + "epoch": 1.2549467275494672, + "grad_norm": 7.0, + "learning_rate": 3.3378769433638965e-06, + "loss": 1.72672253, + "memory(GiB)": 111.15, + "step": 49470, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64112625, + "epoch": 1.2550735667174022, + "grad_norm": 5.6875, + "learning_rate": 3.336887991651595e-06, + "loss": 1.6590641, + "memory(GiB)": 111.15, + "step": 49475, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.64667664, + "epoch": 1.2552004058853374, + "grad_norm": 4.65625, + "learning_rate": 3.3358991130903845e-06, + "loss": 1.63830051, + "memory(GiB)": 111.15, + "step": 49480, + "train_speed(iter/s)": 0.447571 + }, + { + "acc": 0.67880611, + "epoch": 1.2553272450532724, + "grad_norm": 5.28125, + "learning_rate": 3.334910307723761e-06, + "loss": 1.53107319, + "memory(GiB)": 111.15, + "step": 49485, + "train_speed(iter/s)": 0.447595 + }, + { + "acc": 0.65748158, + "epoch": 1.2554540842212076, + "grad_norm": 6.09375, + "learning_rate": 3.333921575595218e-06, + "loss": 1.60756893, + "memory(GiB)": 111.15, + "step": 49490, + "train_speed(iter/s)": 0.447619 + }, + { + "acc": 0.62698889, + "epoch": 1.2555809233891426, + "grad_norm": 5.1875, + "learning_rate": 3.3329329167482404e-06, + "loss": 1.65502281, + "memory(GiB)": 111.15, + "step": 49495, + "train_speed(iter/s)": 0.447643 + }, + { + "acc": 0.66843319, + "epoch": 1.2557077625570776, + "grad_norm": 5.5625, + "learning_rate": 3.331944331226317e-06, + "loss": 1.59684105, + "memory(GiB)": 111.15, + "step": 49500, + "train_speed(iter/s)": 0.447668 + }, + { + "epoch": 1.2557077625570776, + "eval_acc": 0.6434113708741823, + "eval_loss": 1.6076092720031738, + "eval_runtime": 114.1594, + "eval_samples_per_second": 55.799, + "eval_steps_per_second": 27.9, + "step": 49500 + }, + { + "acc": 0.66473956, + "epoch": 1.2558346017250126, + "grad_norm": 4.71875, + "learning_rate": 3.330955819072928e-06, + "loss": 1.60876446, + "memory(GiB)": 111.15, + "step": 49505, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.678582, + "epoch": 1.2559614408929478, + "grad_norm": 5.90625, + "learning_rate": 3.329967380331556e-06, + "loss": 1.54485989, + "memory(GiB)": 111.15, + "step": 49510, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.63830576, + "epoch": 1.2560882800608828, + "grad_norm": 8.25, + "learning_rate": 3.3289790150456737e-06, + "loss": 1.65448608, + "memory(GiB)": 111.15, + "step": 49515, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.64525385, + "epoch": 1.2562151192288178, + "grad_norm": 4.84375, + "learning_rate": 3.327990723258755e-06, + "loss": 1.62041512, + "memory(GiB)": 111.15, + "step": 49520, + "train_speed(iter/s)": 0.447283 + }, + { + "acc": 0.64608965, + "epoch": 1.256341958396753, + "grad_norm": 5.0, + "learning_rate": 3.3270025050142684e-06, + "loss": 1.61028614, + "memory(GiB)": 111.15, + "step": 49525, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.65216894, + "epoch": 1.256468797564688, + "grad_norm": 4.75, + "learning_rate": 3.3260143603556827e-06, + "loss": 1.58564663, + "memory(GiB)": 111.15, + "step": 49530, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.63838272, + "epoch": 1.256595636732623, + "grad_norm": 5.375, + "learning_rate": 3.3250262893264583e-06, + "loss": 1.64035435, + "memory(GiB)": 111.15, + "step": 49535, + "train_speed(iter/s)": 0.447354 + }, + { + "acc": 0.64145112, + "epoch": 1.256722475900558, + "grad_norm": 5.8125, + "learning_rate": 3.3240382919700555e-06, + "loss": 1.61284409, + "memory(GiB)": 111.15, + "step": 49540, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.64376984, + "epoch": 1.2568493150684932, + "grad_norm": 5.59375, + "learning_rate": 3.3230503683299316e-06, + "loss": 1.62973766, + "memory(GiB)": 111.15, + "step": 49545, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.65102882, + "epoch": 1.2569761542364282, + "grad_norm": 6.0625, + "learning_rate": 3.3220625184495404e-06, + "loss": 1.63315163, + "memory(GiB)": 111.15, + "step": 49550, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.64776707, + "epoch": 1.2571029934043634, + "grad_norm": 6.15625, + "learning_rate": 3.3210747423723293e-06, + "loss": 1.64625931, + "memory(GiB)": 111.15, + "step": 49555, + "train_speed(iter/s)": 0.44745 + }, + { + "acc": 0.65332451, + "epoch": 1.2572298325722984, + "grad_norm": 4.9375, + "learning_rate": 3.3200870401417486e-06, + "loss": 1.65684853, + "memory(GiB)": 111.15, + "step": 49560, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.65729885, + "epoch": 1.2573566717402334, + "grad_norm": 4.96875, + "learning_rate": 3.3190994118012387e-06, + "loss": 1.639501, + "memory(GiB)": 111.15, + "step": 49565, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.65120606, + "epoch": 1.2574835109081683, + "grad_norm": 5.625, + "learning_rate": 3.318111857394244e-06, + "loss": 1.65075169, + "memory(GiB)": 111.15, + "step": 49570, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.66114082, + "epoch": 1.2576103500761036, + "grad_norm": 5.25, + "learning_rate": 3.3171243769641957e-06, + "loss": 1.60966644, + "memory(GiB)": 111.15, + "step": 49575, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.64666195, + "epoch": 1.2577371892440385, + "grad_norm": 5.875, + "learning_rate": 3.316136970554532e-06, + "loss": 1.65261974, + "memory(GiB)": 111.15, + "step": 49580, + "train_speed(iter/s)": 0.447569 + }, + { + "acc": 0.66348505, + "epoch": 1.2578640284119738, + "grad_norm": 5.8125, + "learning_rate": 3.315149638208681e-06, + "loss": 1.64910336, + "memory(GiB)": 111.15, + "step": 49585, + "train_speed(iter/s)": 0.447593 + }, + { + "acc": 0.65573225, + "epoch": 1.2579908675799087, + "grad_norm": 6.3125, + "learning_rate": 3.3141623799700738e-06, + "loss": 1.62964745, + "memory(GiB)": 111.15, + "step": 49590, + "train_speed(iter/s)": 0.447616 + }, + { + "acc": 0.63740196, + "epoch": 1.2581177067478437, + "grad_norm": 6.03125, + "learning_rate": 3.3131751958821313e-06, + "loss": 1.70681553, + "memory(GiB)": 111.15, + "step": 49595, + "train_speed(iter/s)": 0.44764 + }, + { + "acc": 0.65528936, + "epoch": 1.2582445459157787, + "grad_norm": 5.8125, + "learning_rate": 3.312188085988273e-06, + "loss": 1.59557323, + "memory(GiB)": 111.15, + "step": 49600, + "train_speed(iter/s)": 0.447664 + }, + { + "epoch": 1.2582445459157787, + "eval_acc": 0.6433499590796921, + "eval_loss": 1.6076806783676147, + "eval_runtime": 113.8192, + "eval_samples_per_second": 55.966, + "eval_steps_per_second": 27.983, + "step": 49600 + }, + { + "acc": 0.65832691, + "epoch": 1.258371385083714, + "grad_norm": 4.59375, + "learning_rate": 3.311201050331919e-06, + "loss": 1.6289156, + "memory(GiB)": 111.15, + "step": 49605, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.66615701, + "epoch": 1.258498224251649, + "grad_norm": 5.28125, + "learning_rate": 3.310214088956485e-06, + "loss": 1.59623337, + "memory(GiB)": 111.15, + "step": 49610, + "train_speed(iter/s)": 0.447233 + }, + { + "acc": 0.67176485, + "epoch": 1.258625063419584, + "grad_norm": 5.4375, + "learning_rate": 3.3092272019053773e-06, + "loss": 1.49739609, + "memory(GiB)": 111.15, + "step": 49615, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.67713718, + "epoch": 1.2587519025875191, + "grad_norm": 4.5, + "learning_rate": 3.308240389222006e-06, + "loss": 1.52206173, + "memory(GiB)": 111.15, + "step": 49620, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.6580256, + "epoch": 1.2588787417554541, + "grad_norm": 4.9375, + "learning_rate": 3.3072536509497762e-06, + "loss": 1.60775108, + "memory(GiB)": 111.15, + "step": 49625, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.65086432, + "epoch": 1.259005580923389, + "grad_norm": 4.71875, + "learning_rate": 3.306266987132089e-06, + "loss": 1.59987583, + "memory(GiB)": 111.15, + "step": 49630, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.67071438, + "epoch": 1.259132420091324, + "grad_norm": 5.125, + "learning_rate": 3.3052803978123405e-06, + "loss": 1.54824343, + "memory(GiB)": 111.15, + "step": 49635, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.66048903, + "epoch": 1.2592592592592593, + "grad_norm": 4.34375, + "learning_rate": 3.3042938830339264e-06, + "loss": 1.57975731, + "memory(GiB)": 111.15, + "step": 49640, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.65136147, + "epoch": 1.2593860984271943, + "grad_norm": 5.15625, + "learning_rate": 3.303307442840238e-06, + "loss": 1.57258816, + "memory(GiB)": 111.15, + "step": 49645, + "train_speed(iter/s)": 0.447401 + }, + { + "acc": 0.66036024, + "epoch": 1.2595129375951295, + "grad_norm": 4.53125, + "learning_rate": 3.302321077274666e-06, + "loss": 1.52835417, + "memory(GiB)": 111.15, + "step": 49650, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.63626728, + "epoch": 1.2596397767630645, + "grad_norm": 4.8125, + "learning_rate": 3.30133478638059e-06, + "loss": 1.69717178, + "memory(GiB)": 111.15, + "step": 49655, + "train_speed(iter/s)": 0.447448 + }, + { + "acc": 0.65202742, + "epoch": 1.2597666159309995, + "grad_norm": 4.5625, + "learning_rate": 3.300348570201395e-06, + "loss": 1.58965607, + "memory(GiB)": 111.15, + "step": 49660, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.63750868, + "epoch": 1.2598934550989345, + "grad_norm": 5.21875, + "learning_rate": 3.299362428780457e-06, + "loss": 1.65280762, + "memory(GiB)": 111.15, + "step": 49665, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.66005955, + "epoch": 1.2600202942668697, + "grad_norm": 5.46875, + "learning_rate": 3.298376362161154e-06, + "loss": 1.46238651, + "memory(GiB)": 111.15, + "step": 49670, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.66012735, + "epoch": 1.2601471334348047, + "grad_norm": 5.65625, + "learning_rate": 3.297390370386856e-06, + "loss": 1.6438961, + "memory(GiB)": 111.15, + "step": 49675, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.63883343, + "epoch": 1.2602739726027397, + "grad_norm": 5.09375, + "learning_rate": 3.2964044535009288e-06, + "loss": 1.67120399, + "memory(GiB)": 111.15, + "step": 49680, + "train_speed(iter/s)": 0.447568 + }, + { + "acc": 0.6512063, + "epoch": 1.2604008117706749, + "grad_norm": 5.3125, + "learning_rate": 3.2954186115467412e-06, + "loss": 1.61113033, + "memory(GiB)": 111.15, + "step": 49685, + "train_speed(iter/s)": 0.447592 + }, + { + "acc": 0.63943386, + "epoch": 1.2605276509386099, + "grad_norm": 5.46875, + "learning_rate": 3.2944328445676543e-06, + "loss": 1.66049576, + "memory(GiB)": 111.15, + "step": 49690, + "train_speed(iter/s)": 0.447615 + }, + { + "acc": 0.63678656, + "epoch": 1.2606544901065448, + "grad_norm": 5.25, + "learning_rate": 3.2934471526070254e-06, + "loss": 1.69283905, + "memory(GiB)": 111.15, + "step": 49695, + "train_speed(iter/s)": 0.447639 + }, + { + "acc": 0.661273, + "epoch": 1.2607813292744798, + "grad_norm": 6.75, + "learning_rate": 3.2924615357082078e-06, + "loss": 1.61276741, + "memory(GiB)": 111.15, + "step": 49700, + "train_speed(iter/s)": 0.447664 + }, + { + "epoch": 1.2607813292744798, + "eval_acc": 0.6433904825087094, + "eval_loss": 1.6076984405517578, + "eval_runtime": 114.6755, + "eval_samples_per_second": 55.548, + "eval_steps_per_second": 27.774, + "step": 49700 + }, + { + "acc": 0.64198875, + "epoch": 1.260908168442415, + "grad_norm": 5.8125, + "learning_rate": 3.2914759939145574e-06, + "loss": 1.65168419, + "memory(GiB)": 111.15, + "step": 49705, + "train_speed(iter/s)": 0.447207 + }, + { + "acc": 0.64660845, + "epoch": 1.26103500761035, + "grad_norm": 6.03125, + "learning_rate": 3.2904905272694214e-06, + "loss": 1.65793076, + "memory(GiB)": 111.15, + "step": 49710, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.66063571, + "epoch": 1.2611618467782852, + "grad_norm": 4.78125, + "learning_rate": 3.289505135816142e-06, + "loss": 1.57222519, + "memory(GiB)": 111.15, + "step": 49715, + "train_speed(iter/s)": 0.447255 + }, + { + "acc": 0.64570284, + "epoch": 1.2612886859462202, + "grad_norm": 5.375, + "learning_rate": 3.2885198195980653e-06, + "loss": 1.64826965, + "memory(GiB)": 111.15, + "step": 49720, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.65148373, + "epoch": 1.2614155251141552, + "grad_norm": 5.375, + "learning_rate": 3.287534578658527e-06, + "loss": 1.62056065, + "memory(GiB)": 111.15, + "step": 49725, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.64085803, + "epoch": 1.2615423642820902, + "grad_norm": 5.625, + "learning_rate": 3.2865494130408657e-06, + "loss": 1.62778339, + "memory(GiB)": 111.15, + "step": 49730, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.64268665, + "epoch": 1.2616692034500254, + "grad_norm": 6.4375, + "learning_rate": 3.2855643227884097e-06, + "loss": 1.65933952, + "memory(GiB)": 111.15, + "step": 49735, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.66326632, + "epoch": 1.2617960426179604, + "grad_norm": 4.8125, + "learning_rate": 3.28457930794449e-06, + "loss": 1.57587681, + "memory(GiB)": 111.15, + "step": 49740, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.66428385, + "epoch": 1.2619228817858956, + "grad_norm": 5.34375, + "learning_rate": 3.283594368552429e-06, + "loss": 1.59262066, + "memory(GiB)": 111.15, + "step": 49745, + "train_speed(iter/s)": 0.447399 + }, + { + "acc": 0.65019517, + "epoch": 1.2620497209538306, + "grad_norm": 5.90625, + "learning_rate": 3.282609504655554e-06, + "loss": 1.65750294, + "memory(GiB)": 111.15, + "step": 49750, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.65336809, + "epoch": 1.2621765601217656, + "grad_norm": 6.6875, + "learning_rate": 3.281624716297179e-06, + "loss": 1.56346664, + "memory(GiB)": 111.15, + "step": 49755, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.64546194, + "epoch": 1.2623033992897006, + "grad_norm": 5.65625, + "learning_rate": 3.28064000352062e-06, + "loss": 1.64669781, + "memory(GiB)": 111.15, + "step": 49760, + "train_speed(iter/s)": 0.447471 + }, + { + "acc": 0.64831867, + "epoch": 1.2624302384576358, + "grad_norm": 5.28125, + "learning_rate": 3.279655366369191e-06, + "loss": 1.60423889, + "memory(GiB)": 111.15, + "step": 49765, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.64295177, + "epoch": 1.2625570776255708, + "grad_norm": 5.03125, + "learning_rate": 3.2786708048862e-06, + "loss": 1.67472649, + "memory(GiB)": 111.15, + "step": 49770, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.66158743, + "epoch": 1.2626839167935058, + "grad_norm": 6.25, + "learning_rate": 3.2776863191149517e-06, + "loss": 1.63628845, + "memory(GiB)": 111.15, + "step": 49775, + "train_speed(iter/s)": 0.447543 + }, + { + "acc": 0.64841137, + "epoch": 1.262810755961441, + "grad_norm": 5.6875, + "learning_rate": 3.2767019090987483e-06, + "loss": 1.67584095, + "memory(GiB)": 111.15, + "step": 49780, + "train_speed(iter/s)": 0.447567 + }, + { + "acc": 0.64042554, + "epoch": 1.262937595129376, + "grad_norm": 4.71875, + "learning_rate": 3.27571757488089e-06, + "loss": 1.71496735, + "memory(GiB)": 111.15, + "step": 49785, + "train_speed(iter/s)": 0.44759 + }, + { + "acc": 0.65434012, + "epoch": 1.263064434297311, + "grad_norm": 5.3125, + "learning_rate": 3.274733316504672e-06, + "loss": 1.59161377, + "memory(GiB)": 111.15, + "step": 49790, + "train_speed(iter/s)": 0.447614 + }, + { + "acc": 0.65235205, + "epoch": 1.263191273465246, + "grad_norm": 4.4375, + "learning_rate": 3.273749134013383e-06, + "loss": 1.5418685, + "memory(GiB)": 111.15, + "step": 49795, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.65715837, + "epoch": 1.2633181126331812, + "grad_norm": 4.90625, + "learning_rate": 3.2727650274503154e-06, + "loss": 1.59599981, + "memory(GiB)": 111.15, + "step": 49800, + "train_speed(iter/s)": 0.447662 + }, + { + "epoch": 1.2633181126331812, + "eval_acc": 0.6434243216607755, + "eval_loss": 1.6075972318649292, + "eval_runtime": 113.795, + "eval_samples_per_second": 55.978, + "eval_steps_per_second": 27.989, + "step": 49800 + }, + { + "acc": 0.65128951, + "epoch": 1.2634449518011162, + "grad_norm": 6.0625, + "learning_rate": 3.2717809968587523e-06, + "loss": 1.58955154, + "memory(GiB)": 111.15, + "step": 49805, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.64873362, + "epoch": 1.2635717909690514, + "grad_norm": 5.71875, + "learning_rate": 3.270797042281979e-06, + "loss": 1.60637321, + "memory(GiB)": 111.15, + "step": 49810, + "train_speed(iter/s)": 0.447234 + }, + { + "acc": 0.63749685, + "epoch": 1.2636986301369864, + "grad_norm": 4.65625, + "learning_rate": 3.269813163763271e-06, + "loss": 1.7353672, + "memory(GiB)": 111.15, + "step": 49815, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.6500968, + "epoch": 1.2638254693049213, + "grad_norm": 5.71875, + "learning_rate": 3.268829361345904e-06, + "loss": 1.6207653, + "memory(GiB)": 111.15, + "step": 49820, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.65395808, + "epoch": 1.2639523084728563, + "grad_norm": 5.375, + "learning_rate": 3.2678456350731526e-06, + "loss": 1.59151535, + "memory(GiB)": 111.15, + "step": 49825, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.67056036, + "epoch": 1.2640791476407915, + "grad_norm": 6.0, + "learning_rate": 3.266861984988283e-06, + "loss": 1.55549793, + "memory(GiB)": 111.15, + "step": 49830, + "train_speed(iter/s)": 0.447329 + }, + { + "acc": 0.66144791, + "epoch": 1.2642059868087265, + "grad_norm": 6.4375, + "learning_rate": 3.2658784111345614e-06, + "loss": 1.61143227, + "memory(GiB)": 111.15, + "step": 49835, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.64963503, + "epoch": 1.2643328259766615, + "grad_norm": 5.1875, + "learning_rate": 3.2648949135552482e-06, + "loss": 1.5689086, + "memory(GiB)": 111.15, + "step": 49840, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.64650221, + "epoch": 1.2644596651445967, + "grad_norm": 7.96875, + "learning_rate": 3.2639114922936045e-06, + "loss": 1.66356239, + "memory(GiB)": 111.15, + "step": 49845, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.66122494, + "epoch": 1.2645865043125317, + "grad_norm": 6.8125, + "learning_rate": 3.2629281473928855e-06, + "loss": 1.54492779, + "memory(GiB)": 111.15, + "step": 49850, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.65090766, + "epoch": 1.2647133434804667, + "grad_norm": 4.5625, + "learning_rate": 3.26194487889634e-06, + "loss": 1.60642681, + "memory(GiB)": 111.15, + "step": 49855, + "train_speed(iter/s)": 0.447448 + }, + { + "acc": 0.65949221, + "epoch": 1.2648401826484017, + "grad_norm": 6.21875, + "learning_rate": 3.2609616868472192e-06, + "loss": 1.56322803, + "memory(GiB)": 111.15, + "step": 49860, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.65364795, + "epoch": 1.264967021816337, + "grad_norm": 5.96875, + "learning_rate": 3.259978571288767e-06, + "loss": 1.58070297, + "memory(GiB)": 111.15, + "step": 49865, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.6398654, + "epoch": 1.265093860984272, + "grad_norm": 4.8125, + "learning_rate": 3.2589955322642293e-06, + "loss": 1.68270073, + "memory(GiB)": 111.15, + "step": 49870, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.64357014, + "epoch": 1.2652207001522071, + "grad_norm": 4.3125, + "learning_rate": 3.2580125698168376e-06, + "loss": 1.58939495, + "memory(GiB)": 111.15, + "step": 49875, + "train_speed(iter/s)": 0.447543 + }, + { + "acc": 0.64326792, + "epoch": 1.265347539320142, + "grad_norm": 5.1875, + "learning_rate": 3.2570296839898314e-06, + "loss": 1.67074432, + "memory(GiB)": 111.15, + "step": 49880, + "train_speed(iter/s)": 0.447567 + }, + { + "acc": 0.65201669, + "epoch": 1.265474378488077, + "grad_norm": 5.90625, + "learning_rate": 3.2560468748264405e-06, + "loss": 1.59917355, + "memory(GiB)": 111.15, + "step": 49885, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.64176607, + "epoch": 1.265601217656012, + "grad_norm": 5.34375, + "learning_rate": 3.2550641423698965e-06, + "loss": 1.66149101, + "memory(GiB)": 111.15, + "step": 49890, + "train_speed(iter/s)": 0.447614 + }, + { + "acc": 0.62546263, + "epoch": 1.2657280568239473, + "grad_norm": 5.03125, + "learning_rate": 3.2540814866634206e-06, + "loss": 1.70406952, + "memory(GiB)": 111.15, + "step": 49895, + "train_speed(iter/s)": 0.447638 + }, + { + "acc": 0.64594569, + "epoch": 1.2658548959918823, + "grad_norm": 5.375, + "learning_rate": 3.2530989077502355e-06, + "loss": 1.60852127, + "memory(GiB)": 111.15, + "step": 49900, + "train_speed(iter/s)": 0.447662 + }, + { + "epoch": 1.2658548959918823, + "eval_acc": 0.6433921535779472, + "eval_loss": 1.6075422763824463, + "eval_runtime": 114.6682, + "eval_samples_per_second": 55.552, + "eval_steps_per_second": 27.776, + "step": 49900 + }, + { + "acc": 0.66452971, + "epoch": 1.2659817351598175, + "grad_norm": 6.34375, + "learning_rate": 3.252116405673561e-06, + "loss": 1.56108389, + "memory(GiB)": 111.15, + "step": 49905, + "train_speed(iter/s)": 0.447207 + }, + { + "acc": 0.66003122, + "epoch": 1.2661085743277525, + "grad_norm": 6.0, + "learning_rate": 3.2511339804766107e-06, + "loss": 1.56979637, + "memory(GiB)": 111.15, + "step": 49910, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.6620224, + "epoch": 1.2662354134956875, + "grad_norm": 5.78125, + "learning_rate": 3.250151632202596e-06, + "loss": 1.58418846, + "memory(GiB)": 111.15, + "step": 49915, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.66003041, + "epoch": 1.2663622526636225, + "grad_norm": 6.4375, + "learning_rate": 3.249169360894724e-06, + "loss": 1.61683006, + "memory(GiB)": 111.15, + "step": 49920, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.63450928, + "epoch": 1.2664890918315577, + "grad_norm": 6.1875, + "learning_rate": 3.2481871665962006e-06, + "loss": 1.7150898, + "memory(GiB)": 111.15, + "step": 49925, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.66307163, + "epoch": 1.2666159309994927, + "grad_norm": 5.21875, + "learning_rate": 3.2472050493502282e-06, + "loss": 1.56990261, + "memory(GiB)": 111.15, + "step": 49930, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.66680093, + "epoch": 1.2667427701674276, + "grad_norm": 5.21875, + "learning_rate": 3.2462230092000017e-06, + "loss": 1.55625334, + "memory(GiB)": 111.15, + "step": 49935, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.65933809, + "epoch": 1.2668696093353629, + "grad_norm": 5.28125, + "learning_rate": 3.2452410461887184e-06, + "loss": 1.56375885, + "memory(GiB)": 111.15, + "step": 49940, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.66160002, + "epoch": 1.2669964485032978, + "grad_norm": 5.0625, + "learning_rate": 3.244259160359567e-06, + "loss": 1.58950901, + "memory(GiB)": 111.15, + "step": 49945, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.65437698, + "epoch": 1.2671232876712328, + "grad_norm": 5.28125, + "learning_rate": 3.2432773517557385e-06, + "loss": 1.62954597, + "memory(GiB)": 111.15, + "step": 49950, + "train_speed(iter/s)": 0.44742 + }, + { + "acc": 0.6426115, + "epoch": 1.2672501268391678, + "grad_norm": 4.4375, + "learning_rate": 3.2422956204204147e-06, + "loss": 1.58818636, + "memory(GiB)": 111.15, + "step": 49955, + "train_speed(iter/s)": 0.447444 + }, + { + "acc": 0.65747938, + "epoch": 1.267376966007103, + "grad_norm": 5.78125, + "learning_rate": 3.2413139663967763e-06, + "loss": 1.64875641, + "memory(GiB)": 111.15, + "step": 49960, + "train_speed(iter/s)": 0.447468 + }, + { + "acc": 0.66909957, + "epoch": 1.267503805175038, + "grad_norm": 5.40625, + "learning_rate": 3.2403323897280013e-06, + "loss": 1.51689234, + "memory(GiB)": 111.15, + "step": 49965, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.64251814, + "epoch": 1.2676306443429732, + "grad_norm": 5.75, + "learning_rate": 3.2393508904572663e-06, + "loss": 1.58810606, + "memory(GiB)": 111.15, + "step": 49970, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.64263349, + "epoch": 1.2677574835109082, + "grad_norm": 4.46875, + "learning_rate": 3.2383694686277382e-06, + "loss": 1.66216736, + "memory(GiB)": 111.15, + "step": 49975, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.64217749, + "epoch": 1.2678843226788432, + "grad_norm": 4.84375, + "learning_rate": 3.2373881242825857e-06, + "loss": 1.63963165, + "memory(GiB)": 111.15, + "step": 49980, + "train_speed(iter/s)": 0.447563 + }, + { + "acc": 0.65042772, + "epoch": 1.2680111618467782, + "grad_norm": 5.0, + "learning_rate": 3.236406857464973e-06, + "loss": 1.59802704, + "memory(GiB)": 111.15, + "step": 49985, + "train_speed(iter/s)": 0.447587 + }, + { + "acc": 0.64161005, + "epoch": 1.2681380010147134, + "grad_norm": 5.03125, + "learning_rate": 3.235425668218063e-06, + "loss": 1.64140434, + "memory(GiB)": 111.15, + "step": 49990, + "train_speed(iter/s)": 0.447611 + }, + { + "acc": 0.64202852, + "epoch": 1.2682648401826484, + "grad_norm": 4.8125, + "learning_rate": 3.234444556585007e-06, + "loss": 1.63023472, + "memory(GiB)": 111.15, + "step": 49995, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.64770455, + "epoch": 1.2683916793505834, + "grad_norm": 5.1875, + "learning_rate": 3.233463522608964e-06, + "loss": 1.63790684, + "memory(GiB)": 111.15, + "step": 50000, + "train_speed(iter/s)": 0.447658 + }, + { + "epoch": 1.2683916793505834, + "eval_acc": 0.6433950779491134, + "eval_loss": 1.6075520515441895, + "eval_runtime": 113.3333, + "eval_samples_per_second": 56.206, + "eval_steps_per_second": 28.103, + "step": 50000 + }, + { + "acc": 0.66678371, + "epoch": 1.2685185185185186, + "grad_norm": 5.5625, + "learning_rate": 3.2324825663330818e-06, + "loss": 1.61853924, + "memory(GiB)": 111.15, + "step": 50005, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.6590148, + "epoch": 1.2686453576864536, + "grad_norm": 6.0625, + "learning_rate": 3.231501687800509e-06, + "loss": 1.56504583, + "memory(GiB)": 111.15, + "step": 50010, + "train_speed(iter/s)": 0.447233 + }, + { + "acc": 0.65355253, + "epoch": 1.2687721968543886, + "grad_norm": 5.28125, + "learning_rate": 3.2305208870543857e-06, + "loss": 1.60804939, + "memory(GiB)": 111.15, + "step": 50015, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.65685959, + "epoch": 1.2688990360223236, + "grad_norm": 6.21875, + "learning_rate": 3.2295401641378544e-06, + "loss": 1.61958733, + "memory(GiB)": 111.15, + "step": 50020, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.65463109, + "epoch": 1.2690258751902588, + "grad_norm": 5.84375, + "learning_rate": 3.2285595190940513e-06, + "loss": 1.62055893, + "memory(GiB)": 111.15, + "step": 50025, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.64562864, + "epoch": 1.2691527143581938, + "grad_norm": 5.6875, + "learning_rate": 3.2275789519661103e-06, + "loss": 1.70457592, + "memory(GiB)": 111.15, + "step": 50030, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.65711589, + "epoch": 1.269279553526129, + "grad_norm": 5.25, + "learning_rate": 3.2265984627971595e-06, + "loss": 1.62931366, + "memory(GiB)": 111.15, + "step": 50035, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.63208475, + "epoch": 1.269406392694064, + "grad_norm": 5.25, + "learning_rate": 3.225618051630326e-06, + "loss": 1.6696455, + "memory(GiB)": 111.15, + "step": 50040, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.65890055, + "epoch": 1.269533231861999, + "grad_norm": 5.0625, + "learning_rate": 3.2246377185087325e-06, + "loss": 1.65114021, + "memory(GiB)": 111.15, + "step": 50045, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.64868407, + "epoch": 1.269660071029934, + "grad_norm": 6.0, + "learning_rate": 3.2236574634755003e-06, + "loss": 1.67368011, + "memory(GiB)": 111.15, + "step": 50050, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.63104134, + "epoch": 1.2697869101978692, + "grad_norm": 5.09375, + "learning_rate": 3.222677286573742e-06, + "loss": 1.75072517, + "memory(GiB)": 111.15, + "step": 50055, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.64478464, + "epoch": 1.2699137493658041, + "grad_norm": 5.28125, + "learning_rate": 3.221697187846571e-06, + "loss": 1.65889206, + "memory(GiB)": 111.15, + "step": 50060, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.65687389, + "epoch": 1.2700405885337394, + "grad_norm": 6.46875, + "learning_rate": 3.2207171673370984e-06, + "loss": 1.67837906, + "memory(GiB)": 111.15, + "step": 50065, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.64970446, + "epoch": 1.2701674277016743, + "grad_norm": 5.0, + "learning_rate": 3.2197372250884295e-06, + "loss": 1.65374565, + "memory(GiB)": 111.15, + "step": 50070, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.6524261, + "epoch": 1.2702942668696093, + "grad_norm": 5.625, + "learning_rate": 3.218757361143664e-06, + "loss": 1.65465393, + "memory(GiB)": 111.15, + "step": 50075, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.6502749, + "epoch": 1.2704211060375443, + "grad_norm": 5.46875, + "learning_rate": 3.2177775755459034e-06, + "loss": 1.6404068, + "memory(GiB)": 111.15, + "step": 50080, + "train_speed(iter/s)": 0.447563 + }, + { + "acc": 0.6484786, + "epoch": 1.2705479452054795, + "grad_norm": 4.875, + "learning_rate": 3.216797868338241e-06, + "loss": 1.64378738, + "memory(GiB)": 111.15, + "step": 50085, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.6538517, + "epoch": 1.2706747843734145, + "grad_norm": 5.09375, + "learning_rate": 3.215818239563773e-06, + "loss": 1.62832222, + "memory(GiB)": 111.15, + "step": 50090, + "train_speed(iter/s)": 0.44761 + }, + { + "acc": 0.6560288, + "epoch": 1.2708016235413495, + "grad_norm": 4.9375, + "learning_rate": 3.2148386892655814e-06, + "loss": 1.56558037, + "memory(GiB)": 111.15, + "step": 50095, + "train_speed(iter/s)": 0.447634 + }, + { + "acc": 0.64755011, + "epoch": 1.2709284627092847, + "grad_norm": 5.53125, + "learning_rate": 3.2138592174867556e-06, + "loss": 1.68871346, + "memory(GiB)": 111.15, + "step": 50100, + "train_speed(iter/s)": 0.447657 + }, + { + "epoch": 1.2709284627092847, + "eval_acc": 0.6433921535779472, + "eval_loss": 1.6076438426971436, + "eval_runtime": 114.9238, + "eval_samples_per_second": 55.428, + "eval_steps_per_second": 27.714, + "step": 50100 + }, + { + "acc": 0.65881186, + "epoch": 1.2710553018772197, + "grad_norm": 6.3125, + "learning_rate": 3.2128798242703745e-06, + "loss": 1.6033308, + "memory(GiB)": 111.15, + "step": 50105, + "train_speed(iter/s)": 0.447203 + }, + { + "acc": 0.65730991, + "epoch": 1.2711821410451547, + "grad_norm": 4.9375, + "learning_rate": 3.2119005096595203e-06, + "loss": 1.61953011, + "memory(GiB)": 111.15, + "step": 50110, + "train_speed(iter/s)": 0.447227 + }, + { + "acc": 0.66475868, + "epoch": 1.2713089802130897, + "grad_norm": 4.65625, + "learning_rate": 3.2109212736972636e-06, + "loss": 1.59062853, + "memory(GiB)": 111.15, + "step": 50115, + "train_speed(iter/s)": 0.44725 + }, + { + "acc": 0.65360527, + "epoch": 1.271435819381025, + "grad_norm": 5.5, + "learning_rate": 3.2099421164266758e-06, + "loss": 1.54943886, + "memory(GiB)": 111.15, + "step": 50120, + "train_speed(iter/s)": 0.447273 + }, + { + "acc": 0.65443211, + "epoch": 1.27156265854896, + "grad_norm": 6.6875, + "learning_rate": 3.2089630378908264e-06, + "loss": 1.62703686, + "memory(GiB)": 111.15, + "step": 50125, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.66433592, + "epoch": 1.271689497716895, + "grad_norm": 4.34375, + "learning_rate": 3.207984038132781e-06, + "loss": 1.56784687, + "memory(GiB)": 111.15, + "step": 50130, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.6558876, + "epoch": 1.27181633688483, + "grad_norm": 4.875, + "learning_rate": 3.2070051171955966e-06, + "loss": 1.62478352, + "memory(GiB)": 111.15, + "step": 50135, + "train_speed(iter/s)": 0.447344 + }, + { + "acc": 0.67122297, + "epoch": 1.271943176052765, + "grad_norm": 4.9375, + "learning_rate": 3.206026275122332e-06, + "loss": 1.49917755, + "memory(GiB)": 111.15, + "step": 50140, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.67349014, + "epoch": 1.2720700152207, + "grad_norm": 5.5, + "learning_rate": 3.205047511956042e-06, + "loss": 1.56338634, + "memory(GiB)": 111.15, + "step": 50145, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.65452662, + "epoch": 1.2721968543886353, + "grad_norm": 5.40625, + "learning_rate": 3.204068827739777e-06, + "loss": 1.68532066, + "memory(GiB)": 111.15, + "step": 50150, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.65037675, + "epoch": 1.2723236935565703, + "grad_norm": 5.34375, + "learning_rate": 3.2030902225165814e-06, + "loss": 1.65561619, + "memory(GiB)": 111.15, + "step": 50155, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.65763569, + "epoch": 1.2724505327245053, + "grad_norm": 4.375, + "learning_rate": 3.2021116963295016e-06, + "loss": 1.60267143, + "memory(GiB)": 111.15, + "step": 50160, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.64391456, + "epoch": 1.2725773718924405, + "grad_norm": 4.8125, + "learning_rate": 3.2011332492215753e-06, + "loss": 1.65124207, + "memory(GiB)": 111.15, + "step": 50165, + "train_speed(iter/s)": 0.447485 + }, + { + "acc": 0.6535676, + "epoch": 1.2727042110603755, + "grad_norm": 5.53125, + "learning_rate": 3.200154881235842e-06, + "loss": 1.61708755, + "memory(GiB)": 111.15, + "step": 50170, + "train_speed(iter/s)": 0.447508 + }, + { + "acc": 0.6567946, + "epoch": 1.2728310502283104, + "grad_norm": 5.53125, + "learning_rate": 3.1991765924153316e-06, + "loss": 1.62024879, + "memory(GiB)": 111.15, + "step": 50175, + "train_speed(iter/s)": 0.447532 + }, + { + "acc": 0.64222221, + "epoch": 1.2729578893962454, + "grad_norm": 5.0625, + "learning_rate": 3.198198382803075e-06, + "loss": 1.61872482, + "memory(GiB)": 111.15, + "step": 50180, + "train_speed(iter/s)": 0.447556 + }, + { + "acc": 0.63628402, + "epoch": 1.2730847285641806, + "grad_norm": 5.78125, + "learning_rate": 3.197220252442097e-06, + "loss": 1.63465767, + "memory(GiB)": 111.15, + "step": 50185, + "train_speed(iter/s)": 0.447579 + }, + { + "acc": 0.641926, + "epoch": 1.2732115677321156, + "grad_norm": 5.09375, + "learning_rate": 3.1962422013754237e-06, + "loss": 1.63595409, + "memory(GiB)": 111.15, + "step": 50190, + "train_speed(iter/s)": 0.447603 + }, + { + "acc": 0.66985378, + "epoch": 1.2733384069000508, + "grad_norm": 5.25, + "learning_rate": 3.1952642296460696e-06, + "loss": 1.53056126, + "memory(GiB)": 111.15, + "step": 50195, + "train_speed(iter/s)": 0.447627 + }, + { + "acc": 0.63048429, + "epoch": 1.2734652460679858, + "grad_norm": 5.40625, + "learning_rate": 3.194286337297051e-06, + "loss": 1.66971245, + "memory(GiB)": 111.15, + "step": 50200, + "train_speed(iter/s)": 0.44765 + }, + { + "epoch": 1.2734652460679858, + "eval_acc": 0.6434351836108213, + "eval_loss": 1.6076834201812744, + "eval_runtime": 113.9471, + "eval_samples_per_second": 55.903, + "eval_steps_per_second": 27.952, + "step": 50200 + }, + { + "acc": 0.64885273, + "epoch": 1.2735920852359208, + "grad_norm": 8.0625, + "learning_rate": 3.1933085243713837e-06, + "loss": 1.58140507, + "memory(GiB)": 111.15, + "step": 50205, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.65860119, + "epoch": 1.2737189244038558, + "grad_norm": 5.21875, + "learning_rate": 3.1923307909120736e-06, + "loss": 1.65360985, + "memory(GiB)": 111.15, + "step": 50210, + "train_speed(iter/s)": 0.447224 + }, + { + "acc": 0.65188298, + "epoch": 1.273845763571791, + "grad_norm": 4.90625, + "learning_rate": 3.1913531369621243e-06, + "loss": 1.62068176, + "memory(GiB)": 111.15, + "step": 50215, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.65496397, + "epoch": 1.273972602739726, + "grad_norm": 5.46875, + "learning_rate": 3.19037556256454e-06, + "loss": 1.61668816, + "memory(GiB)": 111.15, + "step": 50220, + "train_speed(iter/s)": 0.447271 + }, + { + "acc": 0.64114323, + "epoch": 1.2740994419076612, + "grad_norm": 5.90625, + "learning_rate": 3.189398067762318e-06, + "loss": 1.67316666, + "memory(GiB)": 111.15, + "step": 50225, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.65216227, + "epoch": 1.2742262810755962, + "grad_norm": 6.09375, + "learning_rate": 3.1884206525984535e-06, + "loss": 1.66436729, + "memory(GiB)": 111.15, + "step": 50230, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.64429302, + "epoch": 1.2743531202435312, + "grad_norm": 5.40625, + "learning_rate": 3.1874433171159348e-06, + "loss": 1.67595749, + "memory(GiB)": 111.15, + "step": 50235, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.652425, + "epoch": 1.2744799594114662, + "grad_norm": 4.875, + "learning_rate": 3.1864660613577523e-06, + "loss": 1.64321938, + "memory(GiB)": 111.15, + "step": 50240, + "train_speed(iter/s)": 0.447366 + }, + { + "acc": 0.65790148, + "epoch": 1.2746067985794014, + "grad_norm": 5.9375, + "learning_rate": 3.185488885366889e-06, + "loss": 1.54103336, + "memory(GiB)": 111.15, + "step": 50245, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.64930525, + "epoch": 1.2747336377473364, + "grad_norm": 4.65625, + "learning_rate": 3.1845117891863274e-06, + "loss": 1.56862106, + "memory(GiB)": 111.15, + "step": 50250, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.64608421, + "epoch": 1.2748604769152714, + "grad_norm": 4.96875, + "learning_rate": 3.1835347728590414e-06, + "loss": 1.62138138, + "memory(GiB)": 111.15, + "step": 50255, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.64177723, + "epoch": 1.2749873160832066, + "grad_norm": 5.375, + "learning_rate": 3.1825578364280064e-06, + "loss": 1.59158344, + "memory(GiB)": 111.15, + "step": 50260, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.65852103, + "epoch": 1.2751141552511416, + "grad_norm": 5.0625, + "learning_rate": 3.181580979936192e-06, + "loss": 1.60603333, + "memory(GiB)": 111.15, + "step": 50265, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64020967, + "epoch": 1.2752409944190766, + "grad_norm": 5.09375, + "learning_rate": 3.1806042034265656e-06, + "loss": 1.6715271, + "memory(GiB)": 111.15, + "step": 50270, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.6485239, + "epoch": 1.2753678335870116, + "grad_norm": 5.8125, + "learning_rate": 3.179627506942089e-06, + "loss": 1.63843193, + "memory(GiB)": 111.15, + "step": 50275, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.66278744, + "epoch": 1.2754946727549468, + "grad_norm": 5.125, + "learning_rate": 3.17865089052572e-06, + "loss": 1.6263195, + "memory(GiB)": 111.15, + "step": 50280, + "train_speed(iter/s)": 0.447553 + }, + { + "acc": 0.65776415, + "epoch": 1.2756215119228818, + "grad_norm": 4.875, + "learning_rate": 3.177674354220418e-06, + "loss": 1.59134083, + "memory(GiB)": 111.15, + "step": 50285, + "train_speed(iter/s)": 0.447576 + }, + { + "acc": 0.65003228, + "epoch": 1.275748351090817, + "grad_norm": 5.4375, + "learning_rate": 3.1766978980691355e-06, + "loss": 1.60378475, + "memory(GiB)": 111.15, + "step": 50290, + "train_speed(iter/s)": 0.4476 + }, + { + "acc": 0.66763821, + "epoch": 1.275875190258752, + "grad_norm": 4.75, + "learning_rate": 3.1757215221148173e-06, + "loss": 1.60213184, + "memory(GiB)": 111.15, + "step": 50295, + "train_speed(iter/s)": 0.447623 + }, + { + "acc": 0.64865694, + "epoch": 1.276002029426687, + "grad_norm": 5.0625, + "learning_rate": 3.174745226400413e-06, + "loss": 1.62512321, + "memory(GiB)": 111.15, + "step": 50300, + "train_speed(iter/s)": 0.447647 + }, + { + "epoch": 1.276002029426687, + "eval_acc": 0.6433687586086176, + "eval_loss": 1.6076085567474365, + "eval_runtime": 114.6834, + "eval_samples_per_second": 55.544, + "eval_steps_per_second": 27.772, + "step": 50300 + }, + { + "acc": 0.66438375, + "epoch": 1.276128868594622, + "grad_norm": 5.40625, + "learning_rate": 3.1737690109688613e-06, + "loss": 1.57721672, + "memory(GiB)": 111.15, + "step": 50305, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.64973593, + "epoch": 1.2762557077625571, + "grad_norm": 5.28125, + "learning_rate": 3.1727928758631054e-06, + "loss": 1.58800011, + "memory(GiB)": 111.15, + "step": 50310, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.64813776, + "epoch": 1.2763825469304921, + "grad_norm": 5.375, + "learning_rate": 3.1718168211260734e-06, + "loss": 1.69037819, + "memory(GiB)": 111.15, + "step": 50315, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.64027934, + "epoch": 1.2765093860984271, + "grad_norm": 6.65625, + "learning_rate": 3.1708408468007014e-06, + "loss": 1.67929554, + "memory(GiB)": 111.15, + "step": 50320, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.6445962, + "epoch": 1.2766362252663623, + "grad_norm": 4.8125, + "learning_rate": 3.169864952929914e-06, + "loss": 1.6591568, + "memory(GiB)": 111.15, + "step": 50325, + "train_speed(iter/s)": 0.447289 + }, + { + "acc": 0.64590302, + "epoch": 1.2767630644342973, + "grad_norm": 4.65625, + "learning_rate": 3.168889139556639e-06, + "loss": 1.6343956, + "memory(GiB)": 111.15, + "step": 50330, + "train_speed(iter/s)": 0.447312 + }, + { + "acc": 0.64891787, + "epoch": 1.2768899036022323, + "grad_norm": 4.46875, + "learning_rate": 3.1679134067237942e-06, + "loss": 1.62290688, + "memory(GiB)": 111.15, + "step": 50335, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.65314322, + "epoch": 1.2770167427701673, + "grad_norm": 5.65625, + "learning_rate": 3.166937754474296e-06, + "loss": 1.60902157, + "memory(GiB)": 111.15, + "step": 50340, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.66177564, + "epoch": 1.2771435819381025, + "grad_norm": 5.25, + "learning_rate": 3.165962182851061e-06, + "loss": 1.59812059, + "memory(GiB)": 111.15, + "step": 50345, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.6567729, + "epoch": 1.2772704211060375, + "grad_norm": 5.75, + "learning_rate": 3.1649866918969984e-06, + "loss": 1.60915451, + "memory(GiB)": 111.15, + "step": 50350, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.64549398, + "epoch": 1.2773972602739727, + "grad_norm": 5.125, + "learning_rate": 3.164011281655013e-06, + "loss": 1.64837074, + "memory(GiB)": 111.15, + "step": 50355, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.63942294, + "epoch": 1.2775240994419077, + "grad_norm": 5.65625, + "learning_rate": 3.163035952168007e-06, + "loss": 1.66848755, + "memory(GiB)": 111.15, + "step": 50360, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.66552134, + "epoch": 1.2776509386098427, + "grad_norm": 7.5625, + "learning_rate": 3.1620607034788832e-06, + "loss": 1.70404453, + "memory(GiB)": 111.15, + "step": 50365, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.64841681, + "epoch": 1.2777777777777777, + "grad_norm": 4.78125, + "learning_rate": 3.1610855356305354e-06, + "loss": 1.68361511, + "memory(GiB)": 111.15, + "step": 50370, + "train_speed(iter/s)": 0.4475 + }, + { + "acc": 0.64920897, + "epoch": 1.277904616945713, + "grad_norm": 5.625, + "learning_rate": 3.160110448665854e-06, + "loss": 1.64806404, + "memory(GiB)": 111.15, + "step": 50375, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.65678291, + "epoch": 1.2780314561136479, + "grad_norm": 4.53125, + "learning_rate": 3.159135442627731e-06, + "loss": 1.56589661, + "memory(GiB)": 111.15, + "step": 50380, + "train_speed(iter/s)": 0.447546 + }, + { + "acc": 0.66540689, + "epoch": 1.278158295281583, + "grad_norm": 5.03125, + "learning_rate": 3.158160517559049e-06, + "loss": 1.59690685, + "memory(GiB)": 111.15, + "step": 50385, + "train_speed(iter/s)": 0.447569 + }, + { + "acc": 0.65143428, + "epoch": 1.278285134449518, + "grad_norm": 5.15625, + "learning_rate": 3.157185673502693e-06, + "loss": 1.6384306, + "memory(GiB)": 111.15, + "step": 50390, + "train_speed(iter/s)": 0.447592 + }, + { + "acc": 0.64414091, + "epoch": 1.278411973617453, + "grad_norm": 6.40625, + "learning_rate": 3.156210910501537e-06, + "loss": 1.6053072, + "memory(GiB)": 111.15, + "step": 50395, + "train_speed(iter/s)": 0.447615 + }, + { + "acc": 0.64345655, + "epoch": 1.278538812785388, + "grad_norm": 4.625, + "learning_rate": 3.155236228598457e-06, + "loss": 1.61986542, + "memory(GiB)": 111.15, + "step": 50400, + "train_speed(iter/s)": 0.447638 + }, + { + "epoch": 1.278538812785388, + "eval_acc": 0.6433591499605001, + "eval_loss": 1.6075512170791626, + "eval_runtime": 114.4829, + "eval_samples_per_second": 55.641, + "eval_steps_per_second": 27.821, + "step": 50400 + }, + { + "acc": 0.6458962, + "epoch": 1.2786656519533233, + "grad_norm": 5.84375, + "learning_rate": 3.1542616278363238e-06, + "loss": 1.63003159, + "memory(GiB)": 111.15, + "step": 50405, + "train_speed(iter/s)": 0.447188 + }, + { + "acc": 0.6450963, + "epoch": 1.2787924911212583, + "grad_norm": 5.6875, + "learning_rate": 3.1532871082580064e-06, + "loss": 1.65014191, + "memory(GiB)": 111.15, + "step": 50410, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.65298405, + "epoch": 1.2789193302891932, + "grad_norm": 5.40625, + "learning_rate": 3.152312669906366e-06, + "loss": 1.62638969, + "memory(GiB)": 111.15, + "step": 50415, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.66571107, + "epoch": 1.2790461694571285, + "grad_norm": 5.15625, + "learning_rate": 3.1513383128242624e-06, + "loss": 1.55525541, + "memory(GiB)": 111.15, + "step": 50420, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.66219096, + "epoch": 1.2791730086250634, + "grad_norm": 6.28125, + "learning_rate": 3.150364037054555e-06, + "loss": 1.65317535, + "memory(GiB)": 111.15, + "step": 50425, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.64776382, + "epoch": 1.2792998477929984, + "grad_norm": 5.59375, + "learning_rate": 3.149389842640096e-06, + "loss": 1.58974571, + "memory(GiB)": 111.15, + "step": 50430, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.66061811, + "epoch": 1.2794266869609334, + "grad_norm": 6.125, + "learning_rate": 3.148415729623732e-06, + "loss": 1.61202927, + "memory(GiB)": 111.15, + "step": 50435, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.66365891, + "epoch": 1.2795535261288686, + "grad_norm": 4.5, + "learning_rate": 3.1474416980483126e-06, + "loss": 1.66959534, + "memory(GiB)": 111.15, + "step": 50440, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.66404982, + "epoch": 1.2796803652968036, + "grad_norm": 6.4375, + "learning_rate": 3.1464677479566774e-06, + "loss": 1.61831131, + "memory(GiB)": 111.15, + "step": 50445, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.65554304, + "epoch": 1.2798072044647388, + "grad_norm": 5.46875, + "learning_rate": 3.1454938793916677e-06, + "loss": 1.66830254, + "memory(GiB)": 111.15, + "step": 50450, + "train_speed(iter/s)": 0.447399 + }, + { + "acc": 0.64016218, + "epoch": 1.2799340436326738, + "grad_norm": 5.4375, + "learning_rate": 3.144520092396115e-06, + "loss": 1.64622402, + "memory(GiB)": 111.15, + "step": 50455, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.64364614, + "epoch": 1.2800608828006088, + "grad_norm": 4.71875, + "learning_rate": 3.1435463870128536e-06, + "loss": 1.65517769, + "memory(GiB)": 111.15, + "step": 50460, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.66514854, + "epoch": 1.2801877219685438, + "grad_norm": 5.4375, + "learning_rate": 3.142572763284709e-06, + "loss": 1.56312637, + "memory(GiB)": 111.15, + "step": 50465, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.63974895, + "epoch": 1.280314561136479, + "grad_norm": 5.375, + "learning_rate": 3.14159922125451e-06, + "loss": 1.66173534, + "memory(GiB)": 111.15, + "step": 50470, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.65010772, + "epoch": 1.280441400304414, + "grad_norm": 4.28125, + "learning_rate": 3.1406257609650724e-06, + "loss": 1.61982365, + "memory(GiB)": 111.15, + "step": 50475, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.65566902, + "epoch": 1.280568239472349, + "grad_norm": 6.78125, + "learning_rate": 3.139652382459215e-06, + "loss": 1.65226421, + "memory(GiB)": 111.15, + "step": 50480, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.64348707, + "epoch": 1.2806950786402842, + "grad_norm": 5.53125, + "learning_rate": 3.1386790857797535e-06, + "loss": 1.67672119, + "memory(GiB)": 111.15, + "step": 50485, + "train_speed(iter/s)": 0.447563 + }, + { + "acc": 0.66968789, + "epoch": 1.2808219178082192, + "grad_norm": 6.6875, + "learning_rate": 3.1377058709694957e-06, + "loss": 1.59041128, + "memory(GiB)": 111.15, + "step": 50490, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.64699817, + "epoch": 1.2809487569761542, + "grad_norm": 4.71875, + "learning_rate": 3.1367327380712477e-06, + "loss": 1.59387321, + "memory(GiB)": 111.15, + "step": 50495, + "train_speed(iter/s)": 0.447609 + }, + { + "acc": 0.64478717, + "epoch": 1.2810755961440892, + "grad_norm": 4.71875, + "learning_rate": 3.135759687127812e-06, + "loss": 1.61802063, + "memory(GiB)": 111.15, + "step": 50500, + "train_speed(iter/s)": 0.447633 + }, + { + "epoch": 1.2810755961440892, + "eval_acc": 0.6433750251182595, + "eval_loss": 1.6075987815856934, + "eval_runtime": 113.8571, + "eval_samples_per_second": 55.947, + "eval_steps_per_second": 27.974, + "step": 50500 + }, + { + "acc": 0.65162392, + "epoch": 1.2812024353120244, + "grad_norm": 4.28125, + "learning_rate": 3.134786718181989e-06, + "loss": 1.617519, + "memory(GiB)": 111.15, + "step": 50505, + "train_speed(iter/s)": 0.447186 + }, + { + "acc": 0.65153871, + "epoch": 1.2813292744799594, + "grad_norm": 5.65625, + "learning_rate": 3.133813831276575e-06, + "loss": 1.62299995, + "memory(GiB)": 111.15, + "step": 50510, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.65000887, + "epoch": 1.2814561136478946, + "grad_norm": 6.21875, + "learning_rate": 3.1328410264543585e-06, + "loss": 1.64032192, + "memory(GiB)": 111.15, + "step": 50515, + "train_speed(iter/s)": 0.447233 + }, + { + "acc": 0.65616465, + "epoch": 1.2815829528158296, + "grad_norm": 5.5625, + "learning_rate": 3.131868303758131e-06, + "loss": 1.58066502, + "memory(GiB)": 111.15, + "step": 50520, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.65110846, + "epoch": 1.2817097919837646, + "grad_norm": 6.03125, + "learning_rate": 3.1308956632306754e-06, + "loss": 1.58990259, + "memory(GiB)": 111.15, + "step": 50525, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.65341525, + "epoch": 1.2818366311516995, + "grad_norm": 5.21875, + "learning_rate": 3.129923104914776e-06, + "loss": 1.63359623, + "memory(GiB)": 111.15, + "step": 50530, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.65989337, + "epoch": 1.2819634703196348, + "grad_norm": 5.875, + "learning_rate": 3.1289506288532045e-06, + "loss": 1.4942646, + "memory(GiB)": 111.15, + "step": 50535, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.63742547, + "epoch": 1.2820903094875697, + "grad_norm": 6.03125, + "learning_rate": 3.12797823508874e-06, + "loss": 1.64737492, + "memory(GiB)": 111.15, + "step": 50540, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.65663347, + "epoch": 1.282217148655505, + "grad_norm": 5.0625, + "learning_rate": 3.127005923664149e-06, + "loss": 1.59725389, + "memory(GiB)": 111.15, + "step": 50545, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.6539197, + "epoch": 1.28234398782344, + "grad_norm": 4.9375, + "learning_rate": 3.1260336946222014e-06, + "loss": 1.57346001, + "memory(GiB)": 111.15, + "step": 50550, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.65294638, + "epoch": 1.282470826991375, + "grad_norm": 5.75, + "learning_rate": 3.1250615480056584e-06, + "loss": 1.53654938, + "memory(GiB)": 111.15, + "step": 50555, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.66913114, + "epoch": 1.28259766615931, + "grad_norm": 4.84375, + "learning_rate": 3.124089483857278e-06, + "loss": 1.52652807, + "memory(GiB)": 111.15, + "step": 50560, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.64907255, + "epoch": 1.2827245053272451, + "grad_norm": 7.4375, + "learning_rate": 3.123117502219819e-06, + "loss": 1.65857506, + "memory(GiB)": 111.15, + "step": 50565, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.65550747, + "epoch": 1.2828513444951801, + "grad_norm": 5.75, + "learning_rate": 3.122145603136032e-06, + "loss": 1.59926786, + "memory(GiB)": 111.15, + "step": 50570, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.65806832, + "epoch": 1.2829781836631151, + "grad_norm": 6.1875, + "learning_rate": 3.1211737866486653e-06, + "loss": 1.66119041, + "memory(GiB)": 111.15, + "step": 50575, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.65358038, + "epoch": 1.2831050228310503, + "grad_norm": 4.6875, + "learning_rate": 3.1202020528004627e-06, + "loss": 1.58854828, + "memory(GiB)": 111.15, + "step": 50580, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.64953403, + "epoch": 1.2832318619989853, + "grad_norm": 5.03125, + "learning_rate": 3.119230401634167e-06, + "loss": 1.60453911, + "memory(GiB)": 111.15, + "step": 50585, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.65336227, + "epoch": 1.2833587011669203, + "grad_norm": 5.1875, + "learning_rate": 3.118258833192517e-06, + "loss": 1.56720285, + "memory(GiB)": 111.15, + "step": 50590, + "train_speed(iter/s)": 0.447586 + }, + { + "acc": 0.6413856, + "epoch": 1.2834855403348553, + "grad_norm": 5.0625, + "learning_rate": 3.117287347518242e-06, + "loss": 1.61443825, + "memory(GiB)": 111.15, + "step": 50595, + "train_speed(iter/s)": 0.44761 + }, + { + "acc": 0.65168505, + "epoch": 1.2836123795027905, + "grad_norm": 6.1875, + "learning_rate": 3.116315944654077e-06, + "loss": 1.72866554, + "memory(GiB)": 111.15, + "step": 50600, + "train_speed(iter/s)": 0.447633 + }, + { + "epoch": 1.2836123795027905, + "eval_acc": 0.6433704296778554, + "eval_loss": 1.6075800657272339, + "eval_runtime": 113.4854, + "eval_samples_per_second": 56.131, + "eval_steps_per_second": 28.065, + "step": 50600 + }, + { + "acc": 0.64321423, + "epoch": 1.2837392186707255, + "grad_norm": 5.65625, + "learning_rate": 3.115344624642745e-06, + "loss": 1.65816727, + "memory(GiB)": 111.15, + "step": 50605, + "train_speed(iter/s)": 0.447189 + }, + { + "acc": 0.66692905, + "epoch": 1.2838660578386607, + "grad_norm": 6.125, + "learning_rate": 3.1143733875269734e-06, + "loss": 1.64355583, + "memory(GiB)": 111.15, + "step": 50610, + "train_speed(iter/s)": 0.447213 + }, + { + "acc": 0.67056551, + "epoch": 1.2839928970065957, + "grad_norm": 5.59375, + "learning_rate": 3.1134022333494774e-06, + "loss": 1.5481576, + "memory(GiB)": 111.15, + "step": 50615, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.66538715, + "epoch": 1.2841197361745307, + "grad_norm": 5.25, + "learning_rate": 3.1124311621529744e-06, + "loss": 1.56469498, + "memory(GiB)": 111.15, + "step": 50620, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.65030689, + "epoch": 1.2842465753424657, + "grad_norm": 6.125, + "learning_rate": 3.111460173980175e-06, + "loss": 1.67110977, + "memory(GiB)": 111.15, + "step": 50625, + "train_speed(iter/s)": 0.447283 + }, + { + "acc": 0.64902205, + "epoch": 1.2843734145104009, + "grad_norm": 5.09375, + "learning_rate": 3.110489268873792e-06, + "loss": 1.58635387, + "memory(GiB)": 111.15, + "step": 50630, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.65909595, + "epoch": 1.2845002536783359, + "grad_norm": 5.875, + "learning_rate": 3.1095184468765248e-06, + "loss": 1.65138359, + "memory(GiB)": 111.15, + "step": 50635, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.64538879, + "epoch": 1.2846270928462709, + "grad_norm": 4.5, + "learning_rate": 3.1085477080310766e-06, + "loss": 1.6245882, + "memory(GiB)": 111.15, + "step": 50640, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.66140027, + "epoch": 1.284753932014206, + "grad_norm": 5.4375, + "learning_rate": 3.1075770523801453e-06, + "loss": 1.54511099, + "memory(GiB)": 111.15, + "step": 50645, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.65383897, + "epoch": 1.284880771182141, + "grad_norm": 4.96875, + "learning_rate": 3.106606479966426e-06, + "loss": 1.59237652, + "memory(GiB)": 111.15, + "step": 50650, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.67240281, + "epoch": 1.285007610350076, + "grad_norm": 5.65625, + "learning_rate": 3.1056359908326044e-06, + "loss": 1.60171757, + "memory(GiB)": 111.15, + "step": 50655, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.64167595, + "epoch": 1.285134449518011, + "grad_norm": 5.4375, + "learning_rate": 3.1046655850213707e-06, + "loss": 1.61559563, + "memory(GiB)": 111.15, + "step": 50660, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.64164371, + "epoch": 1.2852612886859462, + "grad_norm": 5.625, + "learning_rate": 3.103695262575407e-06, + "loss": 1.64900932, + "memory(GiB)": 111.15, + "step": 50665, + "train_speed(iter/s)": 0.447471 + }, + { + "acc": 0.6503665, + "epoch": 1.2853881278538812, + "grad_norm": 5.1875, + "learning_rate": 3.102725023537393e-06, + "loss": 1.63714943, + "memory(GiB)": 111.15, + "step": 50670, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.65457869, + "epoch": 1.2855149670218164, + "grad_norm": 5.0, + "learning_rate": 3.1017548679500008e-06, + "loss": 1.52845592, + "memory(GiB)": 111.15, + "step": 50675, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.64397726, + "epoch": 1.2856418061897514, + "grad_norm": 6.59375, + "learning_rate": 3.1007847958559057e-06, + "loss": 1.62584267, + "memory(GiB)": 111.15, + "step": 50680, + "train_speed(iter/s)": 0.447541 + }, + { + "acc": 0.64530878, + "epoch": 1.2857686453576864, + "grad_norm": 5.3125, + "learning_rate": 3.099814807297774e-06, + "loss": 1.66882706, + "memory(GiB)": 111.15, + "step": 50685, + "train_speed(iter/s)": 0.447564 + }, + { + "acc": 0.66609206, + "epoch": 1.2858954845256214, + "grad_norm": 5.25, + "learning_rate": 3.098844902318272e-06, + "loss": 1.55015755, + "memory(GiB)": 111.15, + "step": 50690, + "train_speed(iter/s)": 0.447588 + }, + { + "acc": 0.66356387, + "epoch": 1.2860223236935566, + "grad_norm": 5.1875, + "learning_rate": 3.0978750809600596e-06, + "loss": 1.61108189, + "memory(GiB)": 111.15, + "step": 50695, + "train_speed(iter/s)": 0.447611 + }, + { + "acc": 0.66041918, + "epoch": 1.2861491628614916, + "grad_norm": 5.71875, + "learning_rate": 3.0969053432657913e-06, + "loss": 1.57763882, + "memory(GiB)": 111.15, + "step": 50700, + "train_speed(iter/s)": 0.447634 + }, + { + "epoch": 1.2861491628614916, + "eval_acc": 0.6433658342374514, + "eval_loss": 1.6075608730316162, + "eval_runtime": 113.2198, + "eval_samples_per_second": 56.262, + "eval_steps_per_second": 28.131, + "step": 50700 + }, + { + "acc": 0.65405664, + "epoch": 1.2862760020294268, + "grad_norm": 5.34375, + "learning_rate": 3.0959356892781246e-06, + "loss": 1.58841686, + "memory(GiB)": 111.15, + "step": 50705, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.66551991, + "epoch": 1.2864028411973618, + "grad_norm": 6.15625, + "learning_rate": 3.0949661190397072e-06, + "loss": 1.60581379, + "memory(GiB)": 111.15, + "step": 50710, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.65257435, + "epoch": 1.2865296803652968, + "grad_norm": 5.59375, + "learning_rate": 3.0939966325931852e-06, + "loss": 1.68313103, + "memory(GiB)": 111.15, + "step": 50715, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.63156157, + "epoch": 1.2866565195332318, + "grad_norm": 4.90625, + "learning_rate": 3.0930272299812e-06, + "loss": 1.68850803, + "memory(GiB)": 111.15, + "step": 50720, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.64759459, + "epoch": 1.286783358701167, + "grad_norm": 5.03125, + "learning_rate": 3.0920579112463916e-06, + "loss": 1.59468002, + "memory(GiB)": 111.15, + "step": 50725, + "train_speed(iter/s)": 0.447286 + }, + { + "acc": 0.64817781, + "epoch": 1.286910197869102, + "grad_norm": 5.21875, + "learning_rate": 3.0910886764313964e-06, + "loss": 1.70804024, + "memory(GiB)": 111.15, + "step": 50730, + "train_speed(iter/s)": 0.44731 + }, + { + "acc": 0.66556273, + "epoch": 1.287037037037037, + "grad_norm": 4.96875, + "learning_rate": 3.0901195255788406e-06, + "loss": 1.57845955, + "memory(GiB)": 111.15, + "step": 50735, + "train_speed(iter/s)": 0.447333 + }, + { + "acc": 0.66189728, + "epoch": 1.2871638762049722, + "grad_norm": 5.3125, + "learning_rate": 3.089150458731357e-06, + "loss": 1.5540432, + "memory(GiB)": 111.15, + "step": 50740, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.67607899, + "epoch": 1.2872907153729072, + "grad_norm": 5.28125, + "learning_rate": 3.0881814759315666e-06, + "loss": 1.51502686, + "memory(GiB)": 111.15, + "step": 50745, + "train_speed(iter/s)": 0.44738 + }, + { + "acc": 0.64930615, + "epoch": 1.2874175545408422, + "grad_norm": 5.21875, + "learning_rate": 3.0872125772220934e-06, + "loss": 1.66312637, + "memory(GiB)": 111.15, + "step": 50750, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.65339389, + "epoch": 1.2875443937087772, + "grad_norm": 5.78125, + "learning_rate": 3.0862437626455483e-06, + "loss": 1.65199776, + "memory(GiB)": 111.15, + "step": 50755, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.66248226, + "epoch": 1.2876712328767124, + "grad_norm": 5.9375, + "learning_rate": 3.0852750322445473e-06, + "loss": 1.54205647, + "memory(GiB)": 111.15, + "step": 50760, + "train_speed(iter/s)": 0.44745 + }, + { + "acc": 0.65875573, + "epoch": 1.2877980720446474, + "grad_norm": 5.5, + "learning_rate": 3.0843063860616982e-06, + "loss": 1.58181, + "memory(GiB)": 111.15, + "step": 50765, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.65223918, + "epoch": 1.2879249112125826, + "grad_norm": 5.46875, + "learning_rate": 3.0833378241396094e-06, + "loss": 1.62725563, + "memory(GiB)": 111.15, + "step": 50770, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.63703032, + "epoch": 1.2880517503805176, + "grad_norm": 4.90625, + "learning_rate": 3.0823693465208794e-06, + "loss": 1.64224434, + "memory(GiB)": 111.15, + "step": 50775, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.64605174, + "epoch": 1.2881785895484525, + "grad_norm": 6.3125, + "learning_rate": 3.081400953248106e-06, + "loss": 1.59280167, + "memory(GiB)": 111.15, + "step": 50780, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.64998856, + "epoch": 1.2883054287163875, + "grad_norm": 7.40625, + "learning_rate": 3.0804326443638854e-06, + "loss": 1.62623043, + "memory(GiB)": 111.15, + "step": 50785, + "train_speed(iter/s)": 0.447567 + }, + { + "acc": 0.64588404, + "epoch": 1.2884322678843227, + "grad_norm": 5.625, + "learning_rate": 3.0794644199108087e-06, + "loss": 1.68599758, + "memory(GiB)": 111.15, + "step": 50790, + "train_speed(iter/s)": 0.44759 + }, + { + "acc": 0.65105944, + "epoch": 1.2885591070522577, + "grad_norm": 5.5625, + "learning_rate": 3.07849627993146e-06, + "loss": 1.62836037, + "memory(GiB)": 111.15, + "step": 50795, + "train_speed(iter/s)": 0.447614 + }, + { + "acc": 0.65090694, + "epoch": 1.2886859462201927, + "grad_norm": 4.875, + "learning_rate": 3.0775282244684233e-06, + "loss": 1.57931175, + "memory(GiB)": 111.15, + "step": 50800, + "train_speed(iter/s)": 0.447637 + }, + { + "epoch": 1.2886859462201927, + "eval_acc": 0.6434673516936495, + "eval_loss": 1.607621192932129, + "eval_runtime": 113.0982, + "eval_samples_per_second": 56.323, + "eval_steps_per_second": 28.161, + "step": 50800 + }, + { + "acc": 0.64493876, + "epoch": 1.288812785388128, + "grad_norm": 5.0, + "learning_rate": 3.076560253564279e-06, + "loss": 1.64364929, + "memory(GiB)": 111.15, + "step": 50805, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.64325299, + "epoch": 1.288939624556063, + "grad_norm": 7.09375, + "learning_rate": 3.0755923672616038e-06, + "loss": 1.65190392, + "memory(GiB)": 111.15, + "step": 50810, + "train_speed(iter/s)": 0.44722 + }, + { + "acc": 0.63952608, + "epoch": 1.289066463723998, + "grad_norm": 5.40625, + "learning_rate": 3.074624565602966e-06, + "loss": 1.7016037, + "memory(GiB)": 111.15, + "step": 50815, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.64864073, + "epoch": 1.289193302891933, + "grad_norm": 5.59375, + "learning_rate": 3.073656848630937e-06, + "loss": 1.62224617, + "memory(GiB)": 111.15, + "step": 50820, + "train_speed(iter/s)": 0.447266 + }, + { + "acc": 0.64522543, + "epoch": 1.2893201420598681, + "grad_norm": 4.59375, + "learning_rate": 3.0726892163880784e-06, + "loss": 1.60075684, + "memory(GiB)": 111.15, + "step": 50825, + "train_speed(iter/s)": 0.447289 + }, + { + "acc": 0.64214211, + "epoch": 1.289446981227803, + "grad_norm": 4.625, + "learning_rate": 3.071721668916956e-06, + "loss": 1.61937923, + "memory(GiB)": 111.15, + "step": 50830, + "train_speed(iter/s)": 0.447312 + }, + { + "acc": 0.66349182, + "epoch": 1.2895738203957383, + "grad_norm": 5.875, + "learning_rate": 3.0707542062601225e-06, + "loss": 1.54023037, + "memory(GiB)": 111.15, + "step": 50835, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.64415512, + "epoch": 1.2897006595636733, + "grad_norm": 5.03125, + "learning_rate": 3.0697868284601323e-06, + "loss": 1.63431454, + "memory(GiB)": 111.15, + "step": 50840, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.67142735, + "epoch": 1.2898274987316083, + "grad_norm": 6.0, + "learning_rate": 3.068819535559534e-06, + "loss": 1.55343952, + "memory(GiB)": 111.15, + "step": 50845, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.65067644, + "epoch": 1.2899543378995433, + "grad_norm": 5.5, + "learning_rate": 3.0678523276008774e-06, + "loss": 1.60008812, + "memory(GiB)": 111.15, + "step": 50850, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.65633278, + "epoch": 1.2900811770674785, + "grad_norm": 5.4375, + "learning_rate": 3.0668852046267e-06, + "loss": 1.57363434, + "memory(GiB)": 111.15, + "step": 50855, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.65746851, + "epoch": 1.2902080162354135, + "grad_norm": 5.84375, + "learning_rate": 3.0659181666795413e-06, + "loss": 1.59942074, + "memory(GiB)": 111.15, + "step": 50860, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.65090275, + "epoch": 1.2903348554033487, + "grad_norm": 5.34375, + "learning_rate": 3.0649512138019376e-06, + "loss": 1.6375246, + "memory(GiB)": 111.15, + "step": 50865, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.65456181, + "epoch": 1.2904616945712837, + "grad_norm": 5.28125, + "learning_rate": 3.0639843460364203e-06, + "loss": 1.60088177, + "memory(GiB)": 111.15, + "step": 50870, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.63481236, + "epoch": 1.2905885337392187, + "grad_norm": 5.25, + "learning_rate": 3.0630175634255134e-06, + "loss": 1.68277283, + "memory(GiB)": 111.15, + "step": 50875, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.64174032, + "epoch": 1.2907153729071537, + "grad_norm": 4.875, + "learning_rate": 3.062050866011742e-06, + "loss": 1.6920908, + "memory(GiB)": 111.15, + "step": 50880, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.63693657, + "epoch": 1.2908422120750889, + "grad_norm": 5.625, + "learning_rate": 3.0610842538376264e-06, + "loss": 1.72194042, + "memory(GiB)": 111.15, + "step": 50885, + "train_speed(iter/s)": 0.447567 + }, + { + "acc": 0.66001205, + "epoch": 1.2909690512430239, + "grad_norm": 5.09375, + "learning_rate": 3.060117726945683e-06, + "loss": 1.63972969, + "memory(GiB)": 111.15, + "step": 50890, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.62433224, + "epoch": 1.2910958904109588, + "grad_norm": 5.1875, + "learning_rate": 3.059151285378421e-06, + "loss": 1.69562798, + "memory(GiB)": 111.15, + "step": 50895, + "train_speed(iter/s)": 0.447614 + }, + { + "acc": 0.65456376, + "epoch": 1.291222729578894, + "grad_norm": 5.375, + "learning_rate": 3.0581849291783518e-06, + "loss": 1.56094208, + "memory(GiB)": 111.15, + "step": 50900, + "train_speed(iter/s)": 0.447637 + }, + { + "epoch": 1.291222729578894, + "eval_acc": 0.643359985495119, + "eval_loss": 1.6074568033218384, + "eval_runtime": 114.3307, + "eval_samples_per_second": 55.716, + "eval_steps_per_second": 27.858, + "step": 50900 + }, + { + "acc": 0.65253792, + "epoch": 1.291349568746829, + "grad_norm": 5.53125, + "learning_rate": 3.057218658387977e-06, + "loss": 1.60202541, + "memory(GiB)": 111.15, + "step": 50905, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.65202971, + "epoch": 1.291476407914764, + "grad_norm": 6.8125, + "learning_rate": 3.0562524730498023e-06, + "loss": 1.64425011, + "memory(GiB)": 111.15, + "step": 50910, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.63757029, + "epoch": 1.291603247082699, + "grad_norm": 6.0, + "learning_rate": 3.055286373206321e-06, + "loss": 1.69563828, + "memory(GiB)": 111.15, + "step": 50915, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.65594072, + "epoch": 1.2917300862506342, + "grad_norm": 4.71875, + "learning_rate": 3.054320358900027e-06, + "loss": 1.64868145, + "memory(GiB)": 111.15, + "step": 50920, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.63741455, + "epoch": 1.2918569254185692, + "grad_norm": 5.46875, + "learning_rate": 3.053354430173411e-06, + "loss": 1.68188438, + "memory(GiB)": 111.15, + "step": 50925, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.66659493, + "epoch": 1.2919837645865044, + "grad_norm": 6.9375, + "learning_rate": 3.0523885870689595e-06, + "loss": 1.55800953, + "memory(GiB)": 111.15, + "step": 50930, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.66385279, + "epoch": 1.2921106037544394, + "grad_norm": 6.0, + "learning_rate": 3.051422829629152e-06, + "loss": 1.5559844, + "memory(GiB)": 111.15, + "step": 50935, + "train_speed(iter/s)": 0.447331 + }, + { + "acc": 0.6458786, + "epoch": 1.2922374429223744, + "grad_norm": 4.53125, + "learning_rate": 3.050457157896467e-06, + "loss": 1.6168684, + "memory(GiB)": 111.15, + "step": 50940, + "train_speed(iter/s)": 0.447354 + }, + { + "acc": 0.6515089, + "epoch": 1.2923642820903094, + "grad_norm": 4.53125, + "learning_rate": 3.049491571913382e-06, + "loss": 1.56255598, + "memory(GiB)": 111.15, + "step": 50945, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.64557652, + "epoch": 1.2924911212582446, + "grad_norm": 5.3125, + "learning_rate": 3.048526071722367e-06, + "loss": 1.63697262, + "memory(GiB)": 111.15, + "step": 50950, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.65332165, + "epoch": 1.2926179604261796, + "grad_norm": 5.53125, + "learning_rate": 3.047560657365886e-06, + "loss": 1.61530457, + "memory(GiB)": 111.15, + "step": 50955, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.63692188, + "epoch": 1.2927447995941146, + "grad_norm": 5.53125, + "learning_rate": 3.046595328886405e-06, + "loss": 1.70753365, + "memory(GiB)": 111.15, + "step": 50960, + "train_speed(iter/s)": 0.447446 + }, + { + "acc": 0.65631351, + "epoch": 1.2928716387620498, + "grad_norm": 5.28125, + "learning_rate": 3.0456300863263815e-06, + "loss": 1.57765007, + "memory(GiB)": 111.15, + "step": 50965, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.65483809, + "epoch": 1.2929984779299848, + "grad_norm": 4.8125, + "learning_rate": 3.044664929728276e-06, + "loss": 1.56392441, + "memory(GiB)": 111.15, + "step": 50970, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.66515799, + "epoch": 1.2931253170979198, + "grad_norm": 5.6875, + "learning_rate": 3.0436998591345336e-06, + "loss": 1.53311777, + "memory(GiB)": 111.15, + "step": 50975, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.64834862, + "epoch": 1.2932521562658548, + "grad_norm": 6.625, + "learning_rate": 3.042734874587607e-06, + "loss": 1.64389076, + "memory(GiB)": 111.15, + "step": 50980, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.64102259, + "epoch": 1.29337899543379, + "grad_norm": 5.3125, + "learning_rate": 3.041769976129938e-06, + "loss": 1.59390564, + "memory(GiB)": 111.15, + "step": 50985, + "train_speed(iter/s)": 0.447562 + }, + { + "acc": 0.66142001, + "epoch": 1.293505834601725, + "grad_norm": 4.6875, + "learning_rate": 3.0408051638039697e-06, + "loss": 1.63047581, + "memory(GiB)": 111.15, + "step": 50990, + "train_speed(iter/s)": 0.447585 + }, + { + "acc": 0.62736588, + "epoch": 1.2936326737696602, + "grad_norm": 8.375, + "learning_rate": 3.039840437652137e-06, + "loss": 1.72599869, + "memory(GiB)": 111.15, + "step": 50995, + "train_speed(iter/s)": 0.447608 + }, + { + "acc": 0.65045147, + "epoch": 1.2937595129375952, + "grad_norm": 5.6875, + "learning_rate": 3.0388757977168724e-06, + "loss": 1.67511177, + "memory(GiB)": 111.15, + "step": 51000, + "train_speed(iter/s)": 0.447631 + }, + { + "epoch": 1.2937595129375952, + "eval_acc": 0.6433808738605918, + "eval_loss": 1.607554316520691, + "eval_runtime": 113.7933, + "eval_samples_per_second": 55.979, + "eval_steps_per_second": 27.989, + "step": 51000 + }, + { + "acc": 0.65808597, + "epoch": 1.2938863521055302, + "grad_norm": 6.125, + "learning_rate": 3.0379112440406066e-06, + "loss": 1.56366348, + "memory(GiB)": 111.15, + "step": 51005, + "train_speed(iter/s)": 0.447189 + }, + { + "acc": 0.65496979, + "epoch": 1.2940131912734651, + "grad_norm": 4.96875, + "learning_rate": 3.036946776665766e-06, + "loss": 1.60323639, + "memory(GiB)": 111.15, + "step": 51010, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.65254555, + "epoch": 1.2941400304414004, + "grad_norm": 5.71875, + "learning_rate": 3.0359823956347695e-06, + "loss": 1.56458521, + "memory(GiB)": 111.15, + "step": 51015, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.64948702, + "epoch": 1.2942668696093353, + "grad_norm": 4.8125, + "learning_rate": 3.0350181009900347e-06, + "loss": 1.66063881, + "memory(GiB)": 111.15, + "step": 51020, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.66753263, + "epoch": 1.2943937087772706, + "grad_norm": 5.375, + "learning_rate": 3.0340538927739784e-06, + "loss": 1.55555143, + "memory(GiB)": 111.15, + "step": 51025, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.66981926, + "epoch": 1.2945205479452055, + "grad_norm": 5.3125, + "learning_rate": 3.0330897710290093e-06, + "loss": 1.63424473, + "memory(GiB)": 111.15, + "step": 51030, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.64267168, + "epoch": 1.2946473871131405, + "grad_norm": 5.125, + "learning_rate": 3.032125735797532e-06, + "loss": 1.62619781, + "memory(GiB)": 111.15, + "step": 51035, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.64279094, + "epoch": 1.2947742262810755, + "grad_norm": 6.0625, + "learning_rate": 3.031161787121952e-06, + "loss": 1.69595528, + "memory(GiB)": 111.15, + "step": 51040, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.66395597, + "epoch": 1.2949010654490107, + "grad_norm": 4.4375, + "learning_rate": 3.0301979250446655e-06, + "loss": 1.64902306, + "memory(GiB)": 111.15, + "step": 51045, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.67098384, + "epoch": 1.2950279046169457, + "grad_norm": 7.5625, + "learning_rate": 3.029234149608071e-06, + "loss": 1.57896805, + "memory(GiB)": 111.15, + "step": 51050, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.66464968, + "epoch": 1.2951547437848807, + "grad_norm": 5.875, + "learning_rate": 3.0282704608545566e-06, + "loss": 1.57159786, + "memory(GiB)": 111.15, + "step": 51055, + "train_speed(iter/s)": 0.447421 + }, + { + "acc": 0.66302652, + "epoch": 1.295281582952816, + "grad_norm": 5.4375, + "learning_rate": 3.0273068588265097e-06, + "loss": 1.61451683, + "memory(GiB)": 111.15, + "step": 51060, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.64180274, + "epoch": 1.295408422120751, + "grad_norm": 4.65625, + "learning_rate": 3.0263433435663143e-06, + "loss": 1.65819473, + "memory(GiB)": 111.15, + "step": 51065, + "train_speed(iter/s)": 0.447468 + }, + { + "acc": 0.64135962, + "epoch": 1.295535261288686, + "grad_norm": 4.71875, + "learning_rate": 3.0253799151163522e-06, + "loss": 1.67711563, + "memory(GiB)": 111.15, + "step": 51070, + "train_speed(iter/s)": 0.447491 + }, + { + "acc": 0.64158831, + "epoch": 1.295662100456621, + "grad_norm": 5.0, + "learning_rate": 3.0244165735189967e-06, + "loss": 1.63629131, + "memory(GiB)": 111.15, + "step": 51075, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.64085808, + "epoch": 1.295788939624556, + "grad_norm": 5.1875, + "learning_rate": 3.023453318816619e-06, + "loss": 1.68171844, + "memory(GiB)": 111.15, + "step": 51080, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.64380584, + "epoch": 1.295915778792491, + "grad_norm": 5.84375, + "learning_rate": 3.022490151051591e-06, + "loss": 1.61244259, + "memory(GiB)": 111.15, + "step": 51085, + "train_speed(iter/s)": 0.44756 + }, + { + "acc": 0.65129166, + "epoch": 1.2960426179604263, + "grad_norm": 6.875, + "learning_rate": 3.0215270702662753e-06, + "loss": 1.64658031, + "memory(GiB)": 111.15, + "step": 51090, + "train_speed(iter/s)": 0.447583 + }, + { + "acc": 0.66951828, + "epoch": 1.2961694571283613, + "grad_norm": 5.28125, + "learning_rate": 3.020564076503031e-06, + "loss": 1.50924854, + "memory(GiB)": 111.15, + "step": 51095, + "train_speed(iter/s)": 0.447606 + }, + { + "acc": 0.65400076, + "epoch": 1.2962962962962963, + "grad_norm": 4.28125, + "learning_rate": 3.019601169804216e-06, + "loss": 1.61943207, + "memory(GiB)": 111.15, + "step": 51100, + "train_speed(iter/s)": 0.447629 + }, + { + "epoch": 1.2962962962962963, + "eval_acc": 0.6433821271625203, + "eval_loss": 1.607573390007019, + "eval_runtime": 113.8976, + "eval_samples_per_second": 55.927, + "eval_steps_per_second": 27.964, + "step": 51100 + }, + { + "acc": 0.65587549, + "epoch": 1.2964231354642313, + "grad_norm": 4.34375, + "learning_rate": 3.018638350212184e-06, + "loss": 1.64298611, + "memory(GiB)": 111.15, + "step": 51105, + "train_speed(iter/s)": 0.447188 + }, + { + "acc": 0.64406147, + "epoch": 1.2965499746321665, + "grad_norm": 4.71875, + "learning_rate": 3.0176756177692845e-06, + "loss": 1.60268402, + "memory(GiB)": 111.15, + "step": 51110, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.64480138, + "epoch": 1.2966768138001015, + "grad_norm": 5.21875, + "learning_rate": 3.01671297251786e-06, + "loss": 1.65557671, + "memory(GiB)": 111.15, + "step": 51115, + "train_speed(iter/s)": 0.447233 + }, + { + "acc": 0.65870914, + "epoch": 1.2968036529680365, + "grad_norm": 4.78125, + "learning_rate": 3.0157504145002546e-06, + "loss": 1.5925478, + "memory(GiB)": 111.15, + "step": 51120, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.63934231, + "epoch": 1.2969304921359717, + "grad_norm": 5.40625, + "learning_rate": 3.0147879437588046e-06, + "loss": 1.6415163, + "memory(GiB)": 111.15, + "step": 51125, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.65003233, + "epoch": 1.2970573313039067, + "grad_norm": 6.46875, + "learning_rate": 3.013825560335845e-06, + "loss": 1.61503487, + "memory(GiB)": 111.15, + "step": 51130, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.6559659, + "epoch": 1.2971841704718416, + "grad_norm": 6.0, + "learning_rate": 3.0128632642737044e-06, + "loss": 1.55975714, + "memory(GiB)": 111.15, + "step": 51135, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.66094742, + "epoch": 1.2973110096397766, + "grad_norm": 6.09375, + "learning_rate": 3.0119010556147088e-06, + "loss": 1.59061565, + "memory(GiB)": 111.15, + "step": 51140, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.66174603, + "epoch": 1.2974378488077118, + "grad_norm": 5.0, + "learning_rate": 3.0109389344011813e-06, + "loss": 1.59737034, + "memory(GiB)": 111.15, + "step": 51145, + "train_speed(iter/s)": 0.44737 + }, + { + "acc": 0.65427437, + "epoch": 1.2975646879756468, + "grad_norm": 5.125, + "learning_rate": 3.0099769006754415e-06, + "loss": 1.59399719, + "memory(GiB)": 111.15, + "step": 51150, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.64752107, + "epoch": 1.297691527143582, + "grad_norm": 5.375, + "learning_rate": 3.0090149544798007e-06, + "loss": 1.60930977, + "memory(GiB)": 111.15, + "step": 51155, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.63717909, + "epoch": 1.297818366311517, + "grad_norm": 5.53125, + "learning_rate": 3.0080530958565712e-06, + "loss": 1.65434799, + "memory(GiB)": 111.15, + "step": 51160, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.65723095, + "epoch": 1.297945205479452, + "grad_norm": 5.5, + "learning_rate": 3.0070913248480602e-06, + "loss": 1.61227608, + "memory(GiB)": 111.15, + "step": 51165, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.65341158, + "epoch": 1.298072044647387, + "grad_norm": 4.84375, + "learning_rate": 3.0061296414965724e-06, + "loss": 1.61293755, + "memory(GiB)": 111.15, + "step": 51170, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.66788125, + "epoch": 1.2981988838153222, + "grad_norm": 4.625, + "learning_rate": 3.005168045844402e-06, + "loss": 1.60451813, + "memory(GiB)": 111.15, + "step": 51175, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.64781542, + "epoch": 1.2983257229832572, + "grad_norm": 5.8125, + "learning_rate": 3.0042065379338486e-06, + "loss": 1.64085159, + "memory(GiB)": 111.15, + "step": 51180, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.64861593, + "epoch": 1.2984525621511924, + "grad_norm": 6.0, + "learning_rate": 3.003245117807201e-06, + "loss": 1.64908619, + "memory(GiB)": 111.15, + "step": 51185, + "train_speed(iter/s)": 0.447553 + }, + { + "acc": 0.66848979, + "epoch": 1.2985794013191274, + "grad_norm": 4.71875, + "learning_rate": 3.0022837855067514e-06, + "loss": 1.5220768, + "memory(GiB)": 111.15, + "step": 51190, + "train_speed(iter/s)": 0.447576 + }, + { + "acc": 0.65437288, + "epoch": 1.2987062404870624, + "grad_norm": 4.875, + "learning_rate": 3.0013225410747772e-06, + "loss": 1.56430683, + "memory(GiB)": 111.15, + "step": 51195, + "train_speed(iter/s)": 0.447599 + }, + { + "acc": 0.63677311, + "epoch": 1.2988330796549974, + "grad_norm": 5.875, + "learning_rate": 3.0003613845535617e-06, + "loss": 1.63965836, + "memory(GiB)": 111.15, + "step": 51200, + "train_speed(iter/s)": 0.447622 + }, + { + "epoch": 1.2988330796549974, + "eval_acc": 0.6434660983917212, + "eval_loss": 1.6076191663742065, + "eval_runtime": 114.3373, + "eval_samples_per_second": 55.712, + "eval_steps_per_second": 27.856, + "step": 51200 + }, + { + "acc": 0.65167313, + "epoch": 1.2989599188229326, + "grad_norm": 5.78125, + "learning_rate": 2.9994003159853793e-06, + "loss": 1.63172073, + "memory(GiB)": 111.15, + "step": 51205, + "train_speed(iter/s)": 0.44718 + }, + { + "acc": 0.65560074, + "epoch": 1.2990867579908676, + "grad_norm": 5.4375, + "learning_rate": 2.998439335412505e-06, + "loss": 1.56675167, + "memory(GiB)": 111.15, + "step": 51210, + "train_speed(iter/s)": 0.447203 + }, + { + "acc": 0.65206189, + "epoch": 1.2992135971588026, + "grad_norm": 4.40625, + "learning_rate": 2.9974784428772043e-06, + "loss": 1.58930407, + "memory(GiB)": 111.15, + "step": 51215, + "train_speed(iter/s)": 0.447226 + }, + { + "acc": 0.65716453, + "epoch": 1.2993404363267378, + "grad_norm": 4.875, + "learning_rate": 2.996517638421741e-06, + "loss": 1.6391983, + "memory(GiB)": 111.15, + "step": 51220, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.65369239, + "epoch": 1.2994672754946728, + "grad_norm": 4.96875, + "learning_rate": 2.9955569220883777e-06, + "loss": 1.59920931, + "memory(GiB)": 111.15, + "step": 51225, + "train_speed(iter/s)": 0.447271 + }, + { + "acc": 0.64544249, + "epoch": 1.2995941146626078, + "grad_norm": 5.6875, + "learning_rate": 2.9945962939193718e-06, + "loss": 1.64459209, + "memory(GiB)": 111.15, + "step": 51230, + "train_speed(iter/s)": 0.447294 + }, + { + "acc": 0.66017451, + "epoch": 1.2997209538305428, + "grad_norm": 4.5, + "learning_rate": 2.9936357539569728e-06, + "loss": 1.57557335, + "memory(GiB)": 111.15, + "step": 51235, + "train_speed(iter/s)": 0.447317 + }, + { + "acc": 0.66131592, + "epoch": 1.299847792998478, + "grad_norm": 5.875, + "learning_rate": 2.9926753022434306e-06, + "loss": 1.55244493, + "memory(GiB)": 111.15, + "step": 51240, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.67075272, + "epoch": 1.299974632166413, + "grad_norm": 4.96875, + "learning_rate": 2.9917149388209908e-06, + "loss": 1.60667553, + "memory(GiB)": 111.15, + "step": 51245, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.65166702, + "epoch": 1.3001014713343482, + "grad_norm": 4.96875, + "learning_rate": 2.9907546637318964e-06, + "loss": 1.62053967, + "memory(GiB)": 111.15, + "step": 51250, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.66032076, + "epoch": 1.3002283105022832, + "grad_norm": 8.0, + "learning_rate": 2.98979447701838e-06, + "loss": 1.62873192, + "memory(GiB)": 111.15, + "step": 51255, + "train_speed(iter/s)": 0.447409 + }, + { + "acc": 0.66048927, + "epoch": 1.3003551496702181, + "grad_norm": 6.34375, + "learning_rate": 2.988834378722679e-06, + "loss": 1.63615723, + "memory(GiB)": 111.15, + "step": 51260, + "train_speed(iter/s)": 0.447432 + }, + { + "acc": 0.648067, + "epoch": 1.3004819888381531, + "grad_norm": 4.875, + "learning_rate": 2.9878743688870193e-06, + "loss": 1.62589264, + "memory(GiB)": 111.15, + "step": 51265, + "train_speed(iter/s)": 0.447455 + }, + { + "acc": 0.64674358, + "epoch": 1.3006088280060883, + "grad_norm": 4.75, + "learning_rate": 2.9869144475536306e-06, + "loss": 1.63633385, + "memory(GiB)": 111.15, + "step": 51270, + "train_speed(iter/s)": 0.447477 + }, + { + "acc": 0.64220552, + "epoch": 1.3007356671740233, + "grad_norm": 4.84375, + "learning_rate": 2.9859546147647316e-06, + "loss": 1.69746208, + "memory(GiB)": 111.15, + "step": 51275, + "train_speed(iter/s)": 0.4475 + }, + { + "acc": 0.64280567, + "epoch": 1.3008625063419583, + "grad_norm": 5.0625, + "learning_rate": 2.984994870562541e-06, + "loss": 1.64469452, + "memory(GiB)": 111.15, + "step": 51280, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64972091, + "epoch": 1.3009893455098935, + "grad_norm": 5.125, + "learning_rate": 2.9840352149892703e-06, + "loss": 1.61731987, + "memory(GiB)": 111.15, + "step": 51285, + "train_speed(iter/s)": 0.447546 + }, + { + "acc": 0.66771164, + "epoch": 1.3011161846778285, + "grad_norm": 5.4375, + "learning_rate": 2.9830756480871343e-06, + "loss": 1.61184082, + "memory(GiB)": 111.15, + "step": 51290, + "train_speed(iter/s)": 0.447568 + }, + { + "acc": 0.64773607, + "epoch": 1.3012430238457635, + "grad_norm": 5.5625, + "learning_rate": 2.9821161698983347e-06, + "loss": 1.63943443, + "memory(GiB)": 111.15, + "step": 51295, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.64648056, + "epoch": 1.3013698630136985, + "grad_norm": 5.625, + "learning_rate": 2.9811567804650747e-06, + "loss": 1.62170029, + "memory(GiB)": 111.15, + "step": 51300, + "train_speed(iter/s)": 0.447614 + }, + { + "epoch": 1.3013698630136985, + "eval_acc": 0.6434013444587553, + "eval_loss": 1.6075541973114014, + "eval_runtime": 114.3444, + "eval_samples_per_second": 55.709, + "eval_steps_per_second": 27.854, + "step": 51300 + }, + { + "acc": 0.64982181, + "epoch": 1.3014967021816337, + "grad_norm": 5.625, + "learning_rate": 2.980197479829554e-06, + "loss": 1.6378624, + "memory(GiB)": 111.15, + "step": 51305, + "train_speed(iter/s)": 0.447173 + }, + { + "acc": 0.66028318, + "epoch": 1.3016235413495687, + "grad_norm": 5.6875, + "learning_rate": 2.9792382680339666e-06, + "loss": 1.56594849, + "memory(GiB)": 111.15, + "step": 51310, + "train_speed(iter/s)": 0.447196 + }, + { + "acc": 0.63617945, + "epoch": 1.301750380517504, + "grad_norm": 6.0, + "learning_rate": 2.9782791451205e-06, + "loss": 1.64544029, + "memory(GiB)": 111.15, + "step": 51315, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.63051672, + "epoch": 1.301877219685439, + "grad_norm": 6.3125, + "learning_rate": 2.9773201111313444e-06, + "loss": 1.68770714, + "memory(GiB)": 111.15, + "step": 51320, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.66358595, + "epoch": 1.302004058853374, + "grad_norm": 5.625, + "learning_rate": 2.9763611661086806e-06, + "loss": 1.59983025, + "memory(GiB)": 111.15, + "step": 51325, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.66721387, + "epoch": 1.3021308980213089, + "grad_norm": 5.875, + "learning_rate": 2.975402310094689e-06, + "loss": 1.56919937, + "memory(GiB)": 111.15, + "step": 51330, + "train_speed(iter/s)": 0.447288 + }, + { + "acc": 0.64364901, + "epoch": 1.302257737189244, + "grad_norm": 4.5625, + "learning_rate": 2.9744435431315403e-06, + "loss": 1.64272766, + "memory(GiB)": 111.15, + "step": 51335, + "train_speed(iter/s)": 0.447311 + }, + { + "acc": 0.63917465, + "epoch": 1.302384576357179, + "grad_norm": 6.34375, + "learning_rate": 2.9734848652614097e-06, + "loss": 1.72054749, + "memory(GiB)": 111.15, + "step": 51340, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.65636702, + "epoch": 1.3025114155251143, + "grad_norm": 6.0625, + "learning_rate": 2.972526276526461e-06, + "loss": 1.62535706, + "memory(GiB)": 111.15, + "step": 51345, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.66368456, + "epoch": 1.3026382546930493, + "grad_norm": 5.5, + "learning_rate": 2.971567776968861e-06, + "loss": 1.59600468, + "memory(GiB)": 111.15, + "step": 51350, + "train_speed(iter/s)": 0.44738 + }, + { + "acc": 0.67166414, + "epoch": 1.3027650938609843, + "grad_norm": 5.28125, + "learning_rate": 2.9706093666307654e-06, + "loss": 1.58377113, + "memory(GiB)": 111.15, + "step": 51355, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.65111675, + "epoch": 1.3028919330289193, + "grad_norm": 5.09375, + "learning_rate": 2.969651045554329e-06, + "loss": 1.60181274, + "memory(GiB)": 111.15, + "step": 51360, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.64884739, + "epoch": 1.3030187721968545, + "grad_norm": 5.21875, + "learning_rate": 2.968692813781706e-06, + "loss": 1.63060913, + "memory(GiB)": 111.15, + "step": 51365, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.63389487, + "epoch": 1.3031456113647895, + "grad_norm": 5.6875, + "learning_rate": 2.9677346713550437e-06, + "loss": 1.63097477, + "memory(GiB)": 111.15, + "step": 51370, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.67029266, + "epoch": 1.3032724505327244, + "grad_norm": 4.125, + "learning_rate": 2.966776618316482e-06, + "loss": 1.52730551, + "memory(GiB)": 111.15, + "step": 51375, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.64372253, + "epoch": 1.3033992897006597, + "grad_norm": 5.25, + "learning_rate": 2.9658186547081612e-06, + "loss": 1.65298615, + "memory(GiB)": 111.15, + "step": 51380, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.66167803, + "epoch": 1.3035261288685946, + "grad_norm": 5.78125, + "learning_rate": 2.9648607805722197e-06, + "loss": 1.60375595, + "memory(GiB)": 111.15, + "step": 51385, + "train_speed(iter/s)": 0.447542 + }, + { + "acc": 0.65363221, + "epoch": 1.3036529680365296, + "grad_norm": 5.53125, + "learning_rate": 2.963902995950788e-06, + "loss": 1.56821194, + "memory(GiB)": 111.15, + "step": 51390, + "train_speed(iter/s)": 0.447565 + }, + { + "acc": 0.64070525, + "epoch": 1.3037798072044646, + "grad_norm": 6.0, + "learning_rate": 2.962945300885991e-06, + "loss": 1.68458481, + "memory(GiB)": 111.15, + "step": 51395, + "train_speed(iter/s)": 0.447588 + }, + { + "acc": 0.64909096, + "epoch": 1.3039066463723998, + "grad_norm": 6.40625, + "learning_rate": 2.9619876954199557e-06, + "loss": 1.60121555, + "memory(GiB)": 111.15, + "step": 51400, + "train_speed(iter/s)": 0.447611 + }, + { + "epoch": 1.3039066463723998, + "eval_acc": 0.6434142952453484, + "eval_loss": 1.6076395511627197, + "eval_runtime": 113.256, + "eval_samples_per_second": 56.244, + "eval_steps_per_second": 28.122, + "step": 51400 + }, + { + "acc": 0.6693121, + "epoch": 1.3040334855403348, + "grad_norm": 4.59375, + "learning_rate": 2.9610301795947992e-06, + "loss": 1.61637802, + "memory(GiB)": 111.15, + "step": 51405, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.64564199, + "epoch": 1.30416032470827, + "grad_norm": 5.59375, + "learning_rate": 2.9600727534526417e-06, + "loss": 1.66766052, + "memory(GiB)": 111.15, + "step": 51410, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.65081158, + "epoch": 1.304287163876205, + "grad_norm": 5.90625, + "learning_rate": 2.9591154170355895e-06, + "loss": 1.65795174, + "memory(GiB)": 111.15, + "step": 51415, + "train_speed(iter/s)": 0.447222 + }, + { + "acc": 0.63602657, + "epoch": 1.30441400304414, + "grad_norm": 4.96875, + "learning_rate": 2.9581581703857545e-06, + "loss": 1.70293846, + "memory(GiB)": 111.15, + "step": 51420, + "train_speed(iter/s)": 0.447245 + }, + { + "acc": 0.65043082, + "epoch": 1.304540842212075, + "grad_norm": 5.53125, + "learning_rate": 2.9572010135452377e-06, + "loss": 1.58666611, + "memory(GiB)": 111.15, + "step": 51425, + "train_speed(iter/s)": 0.447268 + }, + { + "acc": 0.65420389, + "epoch": 1.3046676813800102, + "grad_norm": 5.3125, + "learning_rate": 2.9562439465561425e-06, + "loss": 1.66190643, + "memory(GiB)": 111.15, + "step": 51430, + "train_speed(iter/s)": 0.447291 + }, + { + "acc": 0.66052356, + "epoch": 1.3047945205479452, + "grad_norm": 5.71875, + "learning_rate": 2.955286969460563e-06, + "loss": 1.5842371, + "memory(GiB)": 111.15, + "step": 51435, + "train_speed(iter/s)": 0.447314 + }, + { + "acc": 0.643993, + "epoch": 1.3049213597158802, + "grad_norm": 5.46875, + "learning_rate": 2.9543300823005903e-06, + "loss": 1.68179245, + "memory(GiB)": 111.15, + "step": 51440, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.63876963, + "epoch": 1.3050481988838154, + "grad_norm": 4.75, + "learning_rate": 2.953373285118315e-06, + "loss": 1.64612083, + "memory(GiB)": 111.15, + "step": 51445, + "train_speed(iter/s)": 0.44736 + }, + { + "acc": 0.64398565, + "epoch": 1.3051750380517504, + "grad_norm": 5.375, + "learning_rate": 2.9524165779558206e-06, + "loss": 1.68176422, + "memory(GiB)": 111.15, + "step": 51450, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.6537035, + "epoch": 1.3053018772196854, + "grad_norm": 5.65625, + "learning_rate": 2.9514599608551865e-06, + "loss": 1.59751301, + "memory(GiB)": 111.15, + "step": 51455, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.64413204, + "epoch": 1.3054287163876204, + "grad_norm": 4.84375, + "learning_rate": 2.9505034338584882e-06, + "loss": 1.74033737, + "memory(GiB)": 111.15, + "step": 51460, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.67083492, + "epoch": 1.3055555555555556, + "grad_norm": 5.59375, + "learning_rate": 2.9495469970078e-06, + "loss": 1.59854794, + "memory(GiB)": 111.15, + "step": 51465, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.64594526, + "epoch": 1.3056823947234906, + "grad_norm": 5.53125, + "learning_rate": 2.9485906503451907e-06, + "loss": 1.68954506, + "memory(GiB)": 111.15, + "step": 51470, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.63535032, + "epoch": 1.3058092338914258, + "grad_norm": 5.1875, + "learning_rate": 2.9476343939127217e-06, + "loss": 1.65026054, + "memory(GiB)": 111.15, + "step": 51475, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.6656208, + "epoch": 1.3059360730593608, + "grad_norm": 5.0, + "learning_rate": 2.9466782277524554e-06, + "loss": 1.51059504, + "memory(GiB)": 111.15, + "step": 51480, + "train_speed(iter/s)": 0.447522 + }, + { + "acc": 0.65013747, + "epoch": 1.3060629122272958, + "grad_norm": 5.71875, + "learning_rate": 2.9457221519064477e-06, + "loss": 1.67445679, + "memory(GiB)": 111.15, + "step": 51485, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.66107087, + "epoch": 1.3061897513952307, + "grad_norm": 5.53125, + "learning_rate": 2.944766166416754e-06, + "loss": 1.59287291, + "memory(GiB)": 111.15, + "step": 51490, + "train_speed(iter/s)": 0.447568 + }, + { + "acc": 0.65122519, + "epoch": 1.306316590563166, + "grad_norm": 5.03125, + "learning_rate": 2.943810271325418e-06, + "loss": 1.61386909, + "memory(GiB)": 111.15, + "step": 51495, + "train_speed(iter/s)": 0.447591 + }, + { + "acc": 0.64727736, + "epoch": 1.306443429731101, + "grad_norm": 5.21875, + "learning_rate": 2.9428544666744873e-06, + "loss": 1.59625702, + "memory(GiB)": 111.15, + "step": 51500, + "train_speed(iter/s)": 0.447613 + }, + { + "epoch": 1.306443429731101, + "eval_acc": 0.6433658342374514, + "eval_loss": 1.6076680421829224, + "eval_runtime": 113.0946, + "eval_samples_per_second": 56.325, + "eval_steps_per_second": 28.162, + "step": 51500 + }, + { + "acc": 0.63194475, + "epoch": 1.3065702688990362, + "grad_norm": 5.53125, + "learning_rate": 2.9418987525060004e-06, + "loss": 1.76627884, + "memory(GiB)": 111.15, + "step": 51505, + "train_speed(iter/s)": 0.447179 + }, + { + "acc": 0.66167421, + "epoch": 1.3066971080669711, + "grad_norm": 6.28125, + "learning_rate": 2.9409431288619973e-06, + "loss": 1.62824326, + "memory(GiB)": 111.15, + "step": 51510, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.64405284, + "epoch": 1.3068239472349061, + "grad_norm": 5.15625, + "learning_rate": 2.939987595784507e-06, + "loss": 1.67999573, + "memory(GiB)": 111.15, + "step": 51515, + "train_speed(iter/s)": 0.447224 + }, + { + "acc": 0.64822855, + "epoch": 1.3069507864028411, + "grad_norm": 5.34375, + "learning_rate": 2.9390321533155585e-06, + "loss": 1.70967178, + "memory(GiB)": 111.15, + "step": 51520, + "train_speed(iter/s)": 0.447247 + }, + { + "acc": 0.65721874, + "epoch": 1.3070776255707763, + "grad_norm": 5.125, + "learning_rate": 2.9380768014971794e-06, + "loss": 1.64087906, + "memory(GiB)": 111.15, + "step": 51525, + "train_speed(iter/s)": 0.44727 + }, + { + "acc": 0.66057014, + "epoch": 1.3072044647387113, + "grad_norm": 5.78125, + "learning_rate": 2.937121540371389e-06, + "loss": 1.53796501, + "memory(GiB)": 111.15, + "step": 51530, + "train_speed(iter/s)": 0.447293 + }, + { + "acc": 0.65827589, + "epoch": 1.3073313039066463, + "grad_norm": 6.15625, + "learning_rate": 2.9361663699802007e-06, + "loss": 1.60662041, + "memory(GiB)": 111.15, + "step": 51535, + "train_speed(iter/s)": 0.447315 + }, + { + "acc": 0.64967151, + "epoch": 1.3074581430745815, + "grad_norm": 4.78125, + "learning_rate": 2.9352112903656315e-06, + "loss": 1.59344339, + "memory(GiB)": 111.15, + "step": 51540, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.64899497, + "epoch": 1.3075849822425165, + "grad_norm": 4.75, + "learning_rate": 2.9342563015696866e-06, + "loss": 1.6919302, + "memory(GiB)": 111.15, + "step": 51545, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.6706521, + "epoch": 1.3077118214104515, + "grad_norm": 5.1875, + "learning_rate": 2.9333014036343765e-06, + "loss": 1.55650682, + "memory(GiB)": 111.15, + "step": 51550, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.64905415, + "epoch": 1.3078386605783865, + "grad_norm": 5.09375, + "learning_rate": 2.932346596601694e-06, + "loss": 1.62649632, + "memory(GiB)": 111.15, + "step": 51555, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.64209538, + "epoch": 1.3079654997463217, + "grad_norm": 5.09375, + "learning_rate": 2.931391880513641e-06, + "loss": 1.6057724, + "memory(GiB)": 111.15, + "step": 51560, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.64285784, + "epoch": 1.3080923389142567, + "grad_norm": 5.90625, + "learning_rate": 2.9304372554122074e-06, + "loss": 1.66431847, + "memory(GiB)": 111.15, + "step": 51565, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.63632884, + "epoch": 1.308219178082192, + "grad_norm": 5.40625, + "learning_rate": 2.9294827213393857e-06, + "loss": 1.66739502, + "memory(GiB)": 111.15, + "step": 51570, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.66618891, + "epoch": 1.308346017250127, + "grad_norm": 4.96875, + "learning_rate": 2.9285282783371567e-06, + "loss": 1.53918743, + "memory(GiB)": 111.15, + "step": 51575, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.65359783, + "epoch": 1.3084728564180619, + "grad_norm": 5.71875, + "learning_rate": 2.9275739264475013e-06, + "loss": 1.5899766, + "memory(GiB)": 111.15, + "step": 51580, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.64497957, + "epoch": 1.3085996955859969, + "grad_norm": 5.1875, + "learning_rate": 2.926619665712399e-06, + "loss": 1.59155684, + "memory(GiB)": 111.15, + "step": 51585, + "train_speed(iter/s)": 0.447543 + }, + { + "acc": 0.64139585, + "epoch": 1.308726534753932, + "grad_norm": 6.5625, + "learning_rate": 2.9256654961738217e-06, + "loss": 1.64278851, + "memory(GiB)": 111.15, + "step": 51590, + "train_speed(iter/s)": 0.447566 + }, + { + "acc": 0.66258092, + "epoch": 1.308853373921867, + "grad_norm": 5.28125, + "learning_rate": 2.9247114178737356e-06, + "loss": 1.56495075, + "memory(GiB)": 111.15, + "step": 51595, + "train_speed(iter/s)": 0.447589 + }, + { + "acc": 0.65272865, + "epoch": 1.308980213089802, + "grad_norm": 4.6875, + "learning_rate": 2.9237574308541063e-06, + "loss": 1.60139027, + "memory(GiB)": 111.15, + "step": 51600, + "train_speed(iter/s)": 0.447612 + }, + { + "epoch": 1.308980213089802, + "eval_acc": 0.6434297526357984, + "eval_loss": 1.6075140237808228, + "eval_runtime": 113.4411, + "eval_samples_per_second": 56.152, + "eval_steps_per_second": 28.076, + "step": 51600 + }, + { + "acc": 0.64269319, + "epoch": 1.3091070522577373, + "grad_norm": 4.75, + "learning_rate": 2.9228035351568955e-06, + "loss": 1.65584755, + "memory(GiB)": 111.15, + "step": 51605, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.66475806, + "epoch": 1.3092338914256723, + "grad_norm": 6.21875, + "learning_rate": 2.92184973082406e-06, + "loss": 1.5263031, + "memory(GiB)": 111.15, + "step": 51610, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.65817032, + "epoch": 1.3093607305936072, + "grad_norm": 4.9375, + "learning_rate": 2.920896017897551e-06, + "loss": 1.59793606, + "memory(GiB)": 111.15, + "step": 51615, + "train_speed(iter/s)": 0.447221 + }, + { + "acc": 0.63960896, + "epoch": 1.3094875697615422, + "grad_norm": 4.46875, + "learning_rate": 2.9199423964193176e-06, + "loss": 1.63207474, + "memory(GiB)": 111.15, + "step": 51620, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.65686569, + "epoch": 1.3096144089294774, + "grad_norm": 5.25, + "learning_rate": 2.9189888664313045e-06, + "loss": 1.64334068, + "memory(GiB)": 111.15, + "step": 51625, + "train_speed(iter/s)": 0.447266 + }, + { + "acc": 0.64752941, + "epoch": 1.3097412480974124, + "grad_norm": 5.03125, + "learning_rate": 2.9180354279754517e-06, + "loss": 1.63963852, + "memory(GiB)": 111.15, + "step": 51630, + "train_speed(iter/s)": 0.447289 + }, + { + "acc": 0.65374889, + "epoch": 1.3098680872653476, + "grad_norm": 5.75, + "learning_rate": 2.9170820810936968e-06, + "loss": 1.59423599, + "memory(GiB)": 111.15, + "step": 51635, + "train_speed(iter/s)": 0.447311 + }, + { + "acc": 0.63992691, + "epoch": 1.3099949264332826, + "grad_norm": 5.21875, + "learning_rate": 2.9161288258279715e-06, + "loss": 1.64783859, + "memory(GiB)": 111.15, + "step": 51640, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.65144973, + "epoch": 1.3101217656012176, + "grad_norm": 5.21875, + "learning_rate": 2.9151756622202037e-06, + "loss": 1.60871296, + "memory(GiB)": 111.15, + "step": 51645, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.64545317, + "epoch": 1.3102486047691526, + "grad_norm": 6.65625, + "learning_rate": 2.914222590312319e-06, + "loss": 1.65289078, + "memory(GiB)": 111.15, + "step": 51650, + "train_speed(iter/s)": 0.447379 + }, + { + "acc": 0.64236231, + "epoch": 1.3103754439370878, + "grad_norm": 7.5625, + "learning_rate": 2.9132696101462366e-06, + "loss": 1.64614773, + "memory(GiB)": 111.15, + "step": 51655, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.64048662, + "epoch": 1.3105022831050228, + "grad_norm": 5.84375, + "learning_rate": 2.912316721763874e-06, + "loss": 1.72400589, + "memory(GiB)": 111.15, + "step": 51660, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.66082258, + "epoch": 1.310629122272958, + "grad_norm": 4.96875, + "learning_rate": 2.9113639252071395e-06, + "loss": 1.54874249, + "memory(GiB)": 111.15, + "step": 51665, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.64150276, + "epoch": 1.310755961440893, + "grad_norm": 5.375, + "learning_rate": 2.91041122051795e-06, + "loss": 1.63816566, + "memory(GiB)": 111.15, + "step": 51670, + "train_speed(iter/s)": 0.44747 + }, + { + "acc": 0.63902278, + "epoch": 1.310882800608828, + "grad_norm": 4.71875, + "learning_rate": 2.9094586077382016e-06, + "loss": 1.65815659, + "memory(GiB)": 111.15, + "step": 51675, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.6475256, + "epoch": 1.311009639776763, + "grad_norm": 6.53125, + "learning_rate": 2.9085060869097977e-06, + "loss": 1.59861851, + "memory(GiB)": 111.15, + "step": 51680, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.65805998, + "epoch": 1.3111364789446982, + "grad_norm": 5.03125, + "learning_rate": 2.907553658074631e-06, + "loss": 1.5536068, + "memory(GiB)": 111.15, + "step": 51685, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.64234972, + "epoch": 1.3112633181126332, + "grad_norm": 4.875, + "learning_rate": 2.906601321274601e-06, + "loss": 1.68426228, + "memory(GiB)": 111.15, + "step": 51690, + "train_speed(iter/s)": 0.44756 + }, + { + "acc": 0.66437454, + "epoch": 1.3113901572805682, + "grad_norm": 4.84375, + "learning_rate": 2.90564907655159e-06, + "loss": 1.63536587, + "memory(GiB)": 111.15, + "step": 51695, + "train_speed(iter/s)": 0.447583 + }, + { + "acc": 0.65887194, + "epoch": 1.3115169964485034, + "grad_norm": 5.15625, + "learning_rate": 2.9046969239474808e-06, + "loss": 1.55727673, + "memory(GiB)": 111.15, + "step": 51700, + "train_speed(iter/s)": 0.447605 + }, + { + "epoch": 1.3115169964485034, + "eval_acc": 0.6433771139548068, + "eval_loss": 1.607596755027771, + "eval_runtime": 114.9666, + "eval_samples_per_second": 55.407, + "eval_steps_per_second": 27.704, + "step": 51700 + }, + { + "acc": 0.6495625, + "epoch": 1.3116438356164384, + "grad_norm": 5.03125, + "learning_rate": 2.9037448635041574e-06, + "loss": 1.62425137, + "memory(GiB)": 111.15, + "step": 51705, + "train_speed(iter/s)": 0.447165 + }, + { + "acc": 0.65078993, + "epoch": 1.3117706747843734, + "grad_norm": 4.96875, + "learning_rate": 2.9027928952634964e-06, + "loss": 1.60145664, + "memory(GiB)": 111.15, + "step": 51710, + "train_speed(iter/s)": 0.447187 + }, + { + "acc": 0.65241737, + "epoch": 1.3118975139523084, + "grad_norm": 4.96875, + "learning_rate": 2.901841019267363e-06, + "loss": 1.54820175, + "memory(GiB)": 111.15, + "step": 51715, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.65326233, + "epoch": 1.3120243531202436, + "grad_norm": 5.375, + "learning_rate": 2.900889235557631e-06, + "loss": 1.63316212, + "memory(GiB)": 111.15, + "step": 51720, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.66227541, + "epoch": 1.3121511922881786, + "grad_norm": 5.4375, + "learning_rate": 2.8999375441761627e-06, + "loss": 1.60789108, + "memory(GiB)": 111.15, + "step": 51725, + "train_speed(iter/s)": 0.447255 + }, + { + "acc": 0.65843096, + "epoch": 1.3122780314561138, + "grad_norm": 4.9375, + "learning_rate": 2.8989859451648193e-06, + "loss": 1.5714077, + "memory(GiB)": 111.15, + "step": 51730, + "train_speed(iter/s)": 0.447277 + }, + { + "acc": 0.6576479, + "epoch": 1.3124048706240488, + "grad_norm": 7.03125, + "learning_rate": 2.89803443856545e-06, + "loss": 1.5491168, + "memory(GiB)": 111.15, + "step": 51735, + "train_speed(iter/s)": 0.4473 + }, + { + "acc": 0.65274453, + "epoch": 1.3125317097919837, + "grad_norm": 5.09375, + "learning_rate": 2.897083024419913e-06, + "loss": 1.64148827, + "memory(GiB)": 111.15, + "step": 51740, + "train_speed(iter/s)": 0.447323 + }, + { + "acc": 0.64789381, + "epoch": 1.3126585489599187, + "grad_norm": 6.03125, + "learning_rate": 2.8961317027700534e-06, + "loss": 1.62172852, + "memory(GiB)": 111.15, + "step": 51745, + "train_speed(iter/s)": 0.447345 + }, + { + "acc": 0.64860587, + "epoch": 1.312785388127854, + "grad_norm": 5.90625, + "learning_rate": 2.8951804736577148e-06, + "loss": 1.62198906, + "memory(GiB)": 111.15, + "step": 51750, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.66691604, + "epoch": 1.312912227295789, + "grad_norm": 5.875, + "learning_rate": 2.894229337124736e-06, + "loss": 1.58112278, + "memory(GiB)": 111.15, + "step": 51755, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.63719645, + "epoch": 1.313039066463724, + "grad_norm": 4.90625, + "learning_rate": 2.8932782932129524e-06, + "loss": 1.62769852, + "memory(GiB)": 111.15, + "step": 51760, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.65849047, + "epoch": 1.3131659056316591, + "grad_norm": 5.75, + "learning_rate": 2.8923273419641956e-06, + "loss": 1.54769497, + "memory(GiB)": 111.15, + "step": 51765, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.65600033, + "epoch": 1.3132927447995941, + "grad_norm": 5.84375, + "learning_rate": 2.891376483420292e-06, + "loss": 1.6164526, + "memory(GiB)": 111.15, + "step": 51770, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.65261326, + "epoch": 1.3134195839675291, + "grad_norm": 6.0, + "learning_rate": 2.8904257176230655e-06, + "loss": 1.61816788, + "memory(GiB)": 111.15, + "step": 51775, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.66023989, + "epoch": 1.313546423135464, + "grad_norm": 5.5, + "learning_rate": 2.8894750446143345e-06, + "loss": 1.54078121, + "memory(GiB)": 111.15, + "step": 51780, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.64474773, + "epoch": 1.3136732623033993, + "grad_norm": 5.875, + "learning_rate": 2.8885244644359134e-06, + "loss": 1.61117249, + "memory(GiB)": 111.15, + "step": 51785, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.63870802, + "epoch": 1.3138001014713343, + "grad_norm": 6.34375, + "learning_rate": 2.887573977129614e-06, + "loss": 1.68238945, + "memory(GiB)": 111.15, + "step": 51790, + "train_speed(iter/s)": 0.447551 + }, + { + "acc": 0.66421838, + "epoch": 1.3139269406392695, + "grad_norm": 6.5, + "learning_rate": 2.886623582737242e-06, + "loss": 1.61451263, + "memory(GiB)": 111.15, + "step": 51795, + "train_speed(iter/s)": 0.447574 + }, + { + "acc": 0.65350685, + "epoch": 1.3140537798072045, + "grad_norm": 5.03125, + "learning_rate": 2.8856732813006007e-06, + "loss": 1.65663033, + "memory(GiB)": 111.15, + "step": 51800, + "train_speed(iter/s)": 0.447596 + }, + { + "epoch": 1.3140537798072045, + "eval_acc": 0.6434122064088011, + "eval_loss": 1.6075665950775146, + "eval_runtime": 114.3258, + "eval_samples_per_second": 55.718, + "eval_steps_per_second": 27.859, + "step": 51800 + }, + { + "acc": 0.65035648, + "epoch": 1.3141806189751395, + "grad_norm": 5.15625, + "learning_rate": 2.8847230728614854e-06, + "loss": 1.6826149, + "memory(GiB)": 111.15, + "step": 51805, + "train_speed(iter/s)": 0.44716 + }, + { + "acc": 0.66183357, + "epoch": 1.3143074581430745, + "grad_norm": 5.3125, + "learning_rate": 2.883772957461698e-06, + "loss": 1.54368267, + "memory(GiB)": 111.15, + "step": 51810, + "train_speed(iter/s)": 0.447182 + }, + { + "acc": 0.64122658, + "epoch": 1.3144342973110097, + "grad_norm": 6.03125, + "learning_rate": 2.8828229351430224e-06, + "loss": 1.70447197, + "memory(GiB)": 111.15, + "step": 51815, + "train_speed(iter/s)": 0.447205 + }, + { + "acc": 0.66266041, + "epoch": 1.3145611364789447, + "grad_norm": 5.78125, + "learning_rate": 2.881873005947247e-06, + "loss": 1.57068291, + "memory(GiB)": 111.15, + "step": 51820, + "train_speed(iter/s)": 0.447228 + }, + { + "acc": 0.64804506, + "epoch": 1.31468797564688, + "grad_norm": 4.90625, + "learning_rate": 2.88092316991615e-06, + "loss": 1.6391716, + "memory(GiB)": 111.15, + "step": 51825, + "train_speed(iter/s)": 0.447251 + }, + { + "acc": 0.66284857, + "epoch": 1.3148148148148149, + "grad_norm": 5.4375, + "learning_rate": 2.879973427091518e-06, + "loss": 1.58616505, + "memory(GiB)": 111.15, + "step": 51830, + "train_speed(iter/s)": 0.447274 + }, + { + "acc": 0.66085758, + "epoch": 1.3149416539827499, + "grad_norm": 5.75, + "learning_rate": 2.879023777515118e-06, + "loss": 1.56278687, + "memory(GiB)": 111.15, + "step": 51835, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.63702478, + "epoch": 1.3150684931506849, + "grad_norm": 5.3125, + "learning_rate": 2.8780742212287192e-06, + "loss": 1.61218395, + "memory(GiB)": 111.15, + "step": 51840, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.63968267, + "epoch": 1.31519533231862, + "grad_norm": 5.28125, + "learning_rate": 2.8771247582740924e-06, + "loss": 1.66425629, + "memory(GiB)": 111.15, + "step": 51845, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.6571692, + "epoch": 1.315322171486555, + "grad_norm": 5.25, + "learning_rate": 2.876175388692999e-06, + "loss": 1.5996418, + "memory(GiB)": 111.15, + "step": 51850, + "train_speed(iter/s)": 0.447365 + }, + { + "acc": 0.63747807, + "epoch": 1.31544901065449, + "grad_norm": 6.0, + "learning_rate": 2.875226112527192e-06, + "loss": 1.69977417, + "memory(GiB)": 111.15, + "step": 51855, + "train_speed(iter/s)": 0.447388 + }, + { + "acc": 0.65503407, + "epoch": 1.3155758498224253, + "grad_norm": 4.96875, + "learning_rate": 2.8742769298184246e-06, + "loss": 1.64841194, + "memory(GiB)": 111.15, + "step": 51860, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.65693464, + "epoch": 1.3157026889903602, + "grad_norm": 4.375, + "learning_rate": 2.8733278406084507e-06, + "loss": 1.60958366, + "memory(GiB)": 111.15, + "step": 51865, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.65114684, + "epoch": 1.3158295281582952, + "grad_norm": 5.46875, + "learning_rate": 2.872378844939015e-06, + "loss": 1.65011406, + "memory(GiB)": 111.15, + "step": 51870, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.65357294, + "epoch": 1.3159563673262302, + "grad_norm": 5.34375, + "learning_rate": 2.871429942851853e-06, + "loss": 1.67586231, + "memory(GiB)": 111.15, + "step": 51875, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.65496798, + "epoch": 1.3160832064941654, + "grad_norm": 6.59375, + "learning_rate": 2.8704811343887075e-06, + "loss": 1.58533173, + "memory(GiB)": 111.15, + "step": 51880, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.65483913, + "epoch": 1.3162100456621004, + "grad_norm": 5.84375, + "learning_rate": 2.86953241959131e-06, + "loss": 1.59632101, + "memory(GiB)": 111.15, + "step": 51885, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.66229196, + "epoch": 1.3163368848300356, + "grad_norm": 5.375, + "learning_rate": 2.8685837985013874e-06, + "loss": 1.57527828, + "memory(GiB)": 111.15, + "step": 51890, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.6572937, + "epoch": 1.3164637239979706, + "grad_norm": 5.5625, + "learning_rate": 2.867635271160666e-06, + "loss": 1.584828, + "memory(GiB)": 111.15, + "step": 51895, + "train_speed(iter/s)": 0.44755 + }, + { + "acc": 0.63994889, + "epoch": 1.3165905631659056, + "grad_norm": 5.6875, + "learning_rate": 2.8666868376108658e-06, + "loss": 1.74312363, + "memory(GiB)": 111.15, + "step": 51900, + "train_speed(iter/s)": 0.447573 + }, + { + "epoch": 1.3165905631659056, + "eval_acc": 0.6433503768470015, + "eval_loss": 1.607569694519043, + "eval_runtime": 115.5308, + "eval_samples_per_second": 55.137, + "eval_steps_per_second": 27.568, + "step": 51900 + }, + { + "acc": 0.64654427, + "epoch": 1.3167174023338406, + "grad_norm": 6.21875, + "learning_rate": 2.865738497893703e-06, + "loss": 1.62039337, + "memory(GiB)": 111.15, + "step": 51905, + "train_speed(iter/s)": 0.447133 + }, + { + "acc": 0.64462137, + "epoch": 1.3168442415017758, + "grad_norm": 5.5625, + "learning_rate": 2.8647902520508896e-06, + "loss": 1.62269764, + "memory(GiB)": 111.15, + "step": 51910, + "train_speed(iter/s)": 0.447155 + }, + { + "acc": 0.6435369, + "epoch": 1.3169710806697108, + "grad_norm": 5.125, + "learning_rate": 2.8638421001241346e-06, + "loss": 1.63888664, + "memory(GiB)": 111.15, + "step": 51915, + "train_speed(iter/s)": 0.447178 + }, + { + "acc": 0.64915962, + "epoch": 1.3170979198376458, + "grad_norm": 6.25, + "learning_rate": 2.8628940421551404e-06, + "loss": 1.63322506, + "memory(GiB)": 111.15, + "step": 51920, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.63892069, + "epoch": 1.317224759005581, + "grad_norm": 5.4375, + "learning_rate": 2.861946078185608e-06, + "loss": 1.68192825, + "memory(GiB)": 111.15, + "step": 51925, + "train_speed(iter/s)": 0.447223 + }, + { + "acc": 0.65514231, + "epoch": 1.317351598173516, + "grad_norm": 5.09375, + "learning_rate": 2.860998208257233e-06, + "loss": 1.59708624, + "memory(GiB)": 111.15, + "step": 51930, + "train_speed(iter/s)": 0.447246 + }, + { + "acc": 0.64644699, + "epoch": 1.317478437341451, + "grad_norm": 5.03125, + "learning_rate": 2.860050432411707e-06, + "loss": 1.67630711, + "memory(GiB)": 111.15, + "step": 51935, + "train_speed(iter/s)": 0.447268 + }, + { + "acc": 0.64157829, + "epoch": 1.317605276509386, + "grad_norm": 4.9375, + "learning_rate": 2.8591027506907167e-06, + "loss": 1.67157173, + "memory(GiB)": 111.15, + "step": 51940, + "train_speed(iter/s)": 0.447291 + }, + { + "acc": 0.66497402, + "epoch": 1.3177321156773212, + "grad_norm": 4.875, + "learning_rate": 2.858155163135946e-06, + "loss": 1.60823212, + "memory(GiB)": 111.15, + "step": 51945, + "train_speed(iter/s)": 0.447313 + }, + { + "acc": 0.63527889, + "epoch": 1.3178589548452562, + "grad_norm": 4.96875, + "learning_rate": 2.857207669789074e-06, + "loss": 1.66968784, + "memory(GiB)": 111.15, + "step": 51950, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.6362504, + "epoch": 1.3179857940131914, + "grad_norm": 5.28125, + "learning_rate": 2.8562602706917754e-06, + "loss": 1.6605875, + "memory(GiB)": 111.15, + "step": 51955, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.6603786, + "epoch": 1.3181126331811264, + "grad_norm": 4.875, + "learning_rate": 2.8553129658857215e-06, + "loss": 1.62357159, + "memory(GiB)": 111.15, + "step": 51960, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.64153538, + "epoch": 1.3182394723490614, + "grad_norm": 5.125, + "learning_rate": 2.854365755412576e-06, + "loss": 1.63620148, + "memory(GiB)": 111.15, + "step": 51965, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.66151729, + "epoch": 1.3183663115169963, + "grad_norm": 5.875, + "learning_rate": 2.8534186393140083e-06, + "loss": 1.57221575, + "memory(GiB)": 111.15, + "step": 51970, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.64632158, + "epoch": 1.3184931506849316, + "grad_norm": 5.09375, + "learning_rate": 2.8524716176316715e-06, + "loss": 1.5725317, + "memory(GiB)": 111.15, + "step": 51975, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.62907853, + "epoch": 1.3186199898528665, + "grad_norm": 6.65625, + "learning_rate": 2.851524690407218e-06, + "loss": 1.71485596, + "memory(GiB)": 111.15, + "step": 51980, + "train_speed(iter/s)": 0.447471 + }, + { + "acc": 0.65423727, + "epoch": 1.3187468290208018, + "grad_norm": 4.40625, + "learning_rate": 2.8505778576823036e-06, + "loss": 1.62629776, + "memory(GiB)": 111.15, + "step": 51985, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.64625306, + "epoch": 1.3188736681887367, + "grad_norm": 6.21875, + "learning_rate": 2.849631119498573e-06, + "loss": 1.67956314, + "memory(GiB)": 111.15, + "step": 51990, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.63755603, + "epoch": 1.3190005073566717, + "grad_norm": 4.9375, + "learning_rate": 2.8486844758976652e-06, + "loss": 1.66462173, + "memory(GiB)": 111.15, + "step": 51995, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.65300665, + "epoch": 1.3191273465246067, + "grad_norm": 5.625, + "learning_rate": 2.8477379269212157e-06, + "loss": 1.60286407, + "memory(GiB)": 111.15, + "step": 52000, + "train_speed(iter/s)": 0.447562 + }, + { + "epoch": 1.3191273465246067, + "eval_acc": 0.6434172196165147, + "eval_loss": 1.6075531244277954, + "eval_runtime": 113.9646, + "eval_samples_per_second": 55.895, + "eval_steps_per_second": 27.947, + "step": 52000 + }, + { + "acc": 0.65164657, + "epoch": 1.319254185692542, + "grad_norm": 5.09375, + "learning_rate": 2.846791472610865e-06, + "loss": 1.60270329, + "memory(GiB)": 111.15, + "step": 52005, + "train_speed(iter/s)": 0.447128 + }, + { + "acc": 0.63437414, + "epoch": 1.319381024860477, + "grad_norm": 5.34375, + "learning_rate": 2.845845113008239e-06, + "loss": 1.67756348, + "memory(GiB)": 111.15, + "step": 52010, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.65123463, + "epoch": 1.319507864028412, + "grad_norm": 5.40625, + "learning_rate": 2.84489884815496e-06, + "loss": 1.62159195, + "memory(GiB)": 111.15, + "step": 52015, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.64134455, + "epoch": 1.3196347031963471, + "grad_norm": 4.5625, + "learning_rate": 2.843952678092653e-06, + "loss": 1.59502592, + "memory(GiB)": 111.15, + "step": 52020, + "train_speed(iter/s)": 0.447196 + }, + { + "acc": 0.63990612, + "epoch": 1.3197615423642821, + "grad_norm": 5.5, + "learning_rate": 2.8430066028629328e-06, + "loss": 1.68054352, + "memory(GiB)": 111.15, + "step": 52025, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.65490832, + "epoch": 1.319888381532217, + "grad_norm": 5.125, + "learning_rate": 2.842060622507415e-06, + "loss": 1.59156752, + "memory(GiB)": 111.15, + "step": 52030, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.65409184, + "epoch": 1.320015220700152, + "grad_norm": 5.28125, + "learning_rate": 2.841114737067702e-06, + "loss": 1.5509882, + "memory(GiB)": 111.15, + "step": 52035, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.67078614, + "epoch": 1.3201420598680873, + "grad_norm": 4.71875, + "learning_rate": 2.840168946585402e-06, + "loss": 1.50965061, + "memory(GiB)": 111.15, + "step": 52040, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.66275196, + "epoch": 1.3202688990360223, + "grad_norm": 5.28125, + "learning_rate": 2.8392232511021158e-06, + "loss": 1.58553171, + "memory(GiB)": 111.15, + "step": 52045, + "train_speed(iter/s)": 0.44731 + }, + { + "acc": 0.64519625, + "epoch": 1.3203957382039575, + "grad_norm": 5.34375, + "learning_rate": 2.8382776506594385e-06, + "loss": 1.67092457, + "memory(GiB)": 111.15, + "step": 52050, + "train_speed(iter/s)": 0.447332 + }, + { + "acc": 0.64908171, + "epoch": 1.3205225773718925, + "grad_norm": 5.3125, + "learning_rate": 2.837332145298961e-06, + "loss": 1.62069817, + "memory(GiB)": 111.15, + "step": 52055, + "train_speed(iter/s)": 0.447355 + }, + { + "acc": 0.65596228, + "epoch": 1.3206494165398275, + "grad_norm": 6.90625, + "learning_rate": 2.836386735062271e-06, + "loss": 1.64731159, + "memory(GiB)": 111.15, + "step": 52060, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.67140579, + "epoch": 1.3207762557077625, + "grad_norm": 5.5, + "learning_rate": 2.835441419990953e-06, + "loss": 1.50725746, + "memory(GiB)": 111.15, + "step": 52065, + "train_speed(iter/s)": 0.447401 + }, + { + "acc": 0.63423309, + "epoch": 1.3209030948756977, + "grad_norm": 4.90625, + "learning_rate": 2.834496200126585e-06, + "loss": 1.6593874, + "memory(GiB)": 111.15, + "step": 52070, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.63486576, + "epoch": 1.3210299340436327, + "grad_norm": 5.6875, + "learning_rate": 2.8335510755107426e-06, + "loss": 1.70788918, + "memory(GiB)": 111.15, + "step": 52075, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.65876217, + "epoch": 1.3211567732115677, + "grad_norm": 6.0, + "learning_rate": 2.8326060461849966e-06, + "loss": 1.60739155, + "memory(GiB)": 111.15, + "step": 52080, + "train_speed(iter/s)": 0.44747 + }, + { + "acc": 0.66561136, + "epoch": 1.3212836123795029, + "grad_norm": 6.28125, + "learning_rate": 2.8316611121909126e-06, + "loss": 1.59585323, + "memory(GiB)": 111.15, + "step": 52085, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.65133753, + "epoch": 1.3214104515474379, + "grad_norm": 5.125, + "learning_rate": 2.8307162735700544e-06, + "loss": 1.74217644, + "memory(GiB)": 111.15, + "step": 52090, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.65980291, + "epoch": 1.3215372907153728, + "grad_norm": 5.59375, + "learning_rate": 2.8297715303639796e-06, + "loss": 1.55058298, + "memory(GiB)": 111.15, + "step": 52095, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.63457909, + "epoch": 1.3216641298833078, + "grad_norm": 4.65625, + "learning_rate": 2.8288268826142423e-06, + "loss": 1.7179203, + "memory(GiB)": 111.15, + "step": 52100, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.3216641298833078, + "eval_acc": 0.6434067754337782, + "eval_loss": 1.6075010299682617, + "eval_runtime": 113.3935, + "eval_samples_per_second": 56.176, + "eval_steps_per_second": 28.088, + "step": 52100 + }, + { + "acc": 0.65409746, + "epoch": 1.321790969051243, + "grad_norm": 5.75, + "learning_rate": 2.8278823303623905e-06, + "loss": 1.62761879, + "memory(GiB)": 111.15, + "step": 52105, + "train_speed(iter/s)": 0.44713 + }, + { + "acc": 0.63894682, + "epoch": 1.321917808219178, + "grad_norm": 5.3125, + "learning_rate": 2.8269378736499754e-06, + "loss": 1.66500874, + "memory(GiB)": 111.15, + "step": 52110, + "train_speed(iter/s)": 0.447152 + }, + { + "acc": 0.65619936, + "epoch": 1.3220446473871132, + "grad_norm": 5.625, + "learning_rate": 2.8259935125185323e-06, + "loss": 1.66425705, + "memory(GiB)": 111.15, + "step": 52115, + "train_speed(iter/s)": 0.447175 + }, + { + "acc": 0.63502359, + "epoch": 1.3221714865550482, + "grad_norm": 5.90625, + "learning_rate": 2.8250492470096008e-06, + "loss": 1.69116516, + "memory(GiB)": 111.15, + "step": 52120, + "train_speed(iter/s)": 0.447198 + }, + { + "acc": 0.64987698, + "epoch": 1.3222983257229832, + "grad_norm": 6.34375, + "learning_rate": 2.824105077164712e-06, + "loss": 1.60672798, + "memory(GiB)": 111.15, + "step": 52125, + "train_speed(iter/s)": 0.44722 + }, + { + "acc": 0.6460454, + "epoch": 1.3224251648909182, + "grad_norm": 4.90625, + "learning_rate": 2.823161003025401e-06, + "loss": 1.66500053, + "memory(GiB)": 111.15, + "step": 52130, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.650423, + "epoch": 1.3225520040588534, + "grad_norm": 7.21875, + "learning_rate": 2.822217024633186e-06, + "loss": 1.68834362, + "memory(GiB)": 111.15, + "step": 52135, + "train_speed(iter/s)": 0.447266 + }, + { + "acc": 0.65094814, + "epoch": 1.3226788432267884, + "grad_norm": 5.21875, + "learning_rate": 2.821273142029587e-06, + "loss": 1.63238335, + "memory(GiB)": 111.15, + "step": 52140, + "train_speed(iter/s)": 0.447288 + }, + { + "acc": 0.64054031, + "epoch": 1.3228056823947236, + "grad_norm": 5.6875, + "learning_rate": 2.820329355256124e-06, + "loss": 1.61056023, + "memory(GiB)": 111.15, + "step": 52145, + "train_speed(iter/s)": 0.447311 + }, + { + "acc": 0.65471077, + "epoch": 1.3229325215626586, + "grad_norm": 5.375, + "learning_rate": 2.8193856643543106e-06, + "loss": 1.61375923, + "memory(GiB)": 111.15, + "step": 52150, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.64766092, + "epoch": 1.3230593607305936, + "grad_norm": 6.90625, + "learning_rate": 2.8184420693656468e-06, + "loss": 1.62207355, + "memory(GiB)": 111.15, + "step": 52155, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.65813999, + "epoch": 1.3231861998985286, + "grad_norm": 4.6875, + "learning_rate": 2.817498570331643e-06, + "loss": 1.58204803, + "memory(GiB)": 111.15, + "step": 52160, + "train_speed(iter/s)": 0.447379 + }, + { + "acc": 0.65937376, + "epoch": 1.3233130390664638, + "grad_norm": 4.90625, + "learning_rate": 2.816555167293795e-06, + "loss": 1.58489618, + "memory(GiB)": 111.15, + "step": 52165, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.65080757, + "epoch": 1.3234398782343988, + "grad_norm": 5.0625, + "learning_rate": 2.815611860293603e-06, + "loss": 1.68309841, + "memory(GiB)": 111.15, + "step": 52170, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.65599265, + "epoch": 1.3235667174023338, + "grad_norm": 5.09375, + "learning_rate": 2.814668649372549e-06, + "loss": 1.62099228, + "memory(GiB)": 111.15, + "step": 52175, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.64593887, + "epoch": 1.323693556570269, + "grad_norm": 6.03125, + "learning_rate": 2.8137255345721266e-06, + "loss": 1.67100029, + "memory(GiB)": 111.15, + "step": 52180, + "train_speed(iter/s)": 0.44747 + }, + { + "acc": 0.65187764, + "epoch": 1.323820395738204, + "grad_norm": 6.125, + "learning_rate": 2.8127825159338163e-06, + "loss": 1.67342834, + "memory(GiB)": 111.15, + "step": 52185, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.64188695, + "epoch": 1.323947234906139, + "grad_norm": 5.75, + "learning_rate": 2.8118395934990962e-06, + "loss": 1.72190323, + "memory(GiB)": 111.15, + "step": 52190, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.63349895, + "epoch": 1.324074074074074, + "grad_norm": 5.15625, + "learning_rate": 2.81089676730944e-06, + "loss": 1.65689182, + "memory(GiB)": 111.15, + "step": 52195, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.66355257, + "epoch": 1.3242009132420092, + "grad_norm": 4.59375, + "learning_rate": 2.8099540374063185e-06, + "loss": 1.57959805, + "memory(GiB)": 111.15, + "step": 52200, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.3242009132420092, + "eval_acc": 0.6434280815665606, + "eval_loss": 1.6076656579971313, + "eval_runtime": 114.4271, + "eval_samples_per_second": 55.669, + "eval_steps_per_second": 27.834, + "step": 52200 + }, + { + "acc": 0.63358693, + "epoch": 1.3243277524099442, + "grad_norm": 4.5, + "learning_rate": 2.8090114038311956e-06, + "loss": 1.68698921, + "memory(GiB)": 111.15, + "step": 52205, + "train_speed(iter/s)": 0.447126 + }, + { + "acc": 0.64548616, + "epoch": 1.3244545915778794, + "grad_norm": 5.71875, + "learning_rate": 2.8080688666255328e-06, + "loss": 1.62653389, + "memory(GiB)": 111.15, + "step": 52210, + "train_speed(iter/s)": 0.447149 + }, + { + "acc": 0.6441947, + "epoch": 1.3245814307458144, + "grad_norm": 5.03125, + "learning_rate": 2.8071264258307884e-06, + "loss": 1.60196915, + "memory(GiB)": 111.15, + "step": 52215, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.66268568, + "epoch": 1.3247082699137493, + "grad_norm": 4.96875, + "learning_rate": 2.8061840814884133e-06, + "loss": 1.58032045, + "memory(GiB)": 111.15, + "step": 52220, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.66499128, + "epoch": 1.3248351090816843, + "grad_norm": 4.625, + "learning_rate": 2.805241833639858e-06, + "loss": 1.59556236, + "memory(GiB)": 111.15, + "step": 52225, + "train_speed(iter/s)": 0.447217 + }, + { + "acc": 0.64801416, + "epoch": 1.3249619482496195, + "grad_norm": 4.8125, + "learning_rate": 2.804299682326565e-06, + "loss": 1.64481983, + "memory(GiB)": 111.15, + "step": 52230, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.65199566, + "epoch": 1.3250887874175545, + "grad_norm": 5.21875, + "learning_rate": 2.8033576275899752e-06, + "loss": 1.64422684, + "memory(GiB)": 111.15, + "step": 52235, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.64881196, + "epoch": 1.3252156265854895, + "grad_norm": 5.03125, + "learning_rate": 2.8024156694715242e-06, + "loss": 1.65006638, + "memory(GiB)": 111.15, + "step": 52240, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.6359581, + "epoch": 1.3253424657534247, + "grad_norm": 5.4375, + "learning_rate": 2.8014738080126424e-06, + "loss": 1.6880949, + "memory(GiB)": 111.15, + "step": 52245, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.65246181, + "epoch": 1.3254693049213597, + "grad_norm": 4.34375, + "learning_rate": 2.8005320432547612e-06, + "loss": 1.63177299, + "memory(GiB)": 111.15, + "step": 52250, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.63821292, + "epoch": 1.3255961440892947, + "grad_norm": 5.75, + "learning_rate": 2.7995903752392993e-06, + "loss": 1.64787979, + "memory(GiB)": 111.15, + "step": 52255, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.66979837, + "epoch": 1.3257229832572297, + "grad_norm": 5.96875, + "learning_rate": 2.7986488040076764e-06, + "loss": 1.53770504, + "memory(GiB)": 111.15, + "step": 52260, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.64499645, + "epoch": 1.325849822425165, + "grad_norm": 6.5625, + "learning_rate": 2.797707329601306e-06, + "loss": 1.70375977, + "memory(GiB)": 111.15, + "step": 52265, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.64760723, + "epoch": 1.3259766615931, + "grad_norm": 5.90625, + "learning_rate": 2.7967659520616032e-06, + "loss": 1.60324993, + "memory(GiB)": 111.15, + "step": 52270, + "train_speed(iter/s)": 0.447421 + }, + { + "acc": 0.65784149, + "epoch": 1.3261035007610351, + "grad_norm": 4.71875, + "learning_rate": 2.7958246714299685e-06, + "loss": 1.48627863, + "memory(GiB)": 111.15, + "step": 52275, + "train_speed(iter/s)": 0.447443 + }, + { + "acc": 0.64107094, + "epoch": 1.32623033992897, + "grad_norm": 4.46875, + "learning_rate": 2.7948834877478035e-06, + "loss": 1.64859982, + "memory(GiB)": 111.15, + "step": 52280, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.63146014, + "epoch": 1.326357179096905, + "grad_norm": 6.1875, + "learning_rate": 2.7939424010565107e-06, + "loss": 1.69948101, + "memory(GiB)": 111.15, + "step": 52285, + "train_speed(iter/s)": 0.447488 + }, + { + "acc": 0.65651646, + "epoch": 1.32648401826484, + "grad_norm": 5.09375, + "learning_rate": 2.793001411397482e-06, + "loss": 1.6648901, + "memory(GiB)": 111.15, + "step": 52290, + "train_speed(iter/s)": 0.447511 + }, + { + "acc": 0.64474525, + "epoch": 1.3266108574327753, + "grad_norm": 5.9375, + "learning_rate": 2.792060518812103e-06, + "loss": 1.65687714, + "memory(GiB)": 111.15, + "step": 52295, + "train_speed(iter/s)": 0.447534 + }, + { + "acc": 0.65726428, + "epoch": 1.3267376966007103, + "grad_norm": 6.15625, + "learning_rate": 2.7911197233417574e-06, + "loss": 1.61543655, + "memory(GiB)": 111.15, + "step": 52300, + "train_speed(iter/s)": 0.447556 + }, + { + "epoch": 1.3267376966007103, + "eval_acc": 0.6433917358106378, + "eval_loss": 1.6075632572174072, + "eval_runtime": 113.8104, + "eval_samples_per_second": 55.97, + "eval_steps_per_second": 27.985, + "step": 52300 + }, + { + "acc": 0.64864416, + "epoch": 1.3268645357686455, + "grad_norm": 5.59375, + "learning_rate": 2.790179025027831e-06, + "loss": 1.66857777, + "memory(GiB)": 111.15, + "step": 52305, + "train_speed(iter/s)": 0.447126 + }, + { + "acc": 0.634446, + "epoch": 1.3269913749365805, + "grad_norm": 6.375, + "learning_rate": 2.789238423911699e-06, + "loss": 1.66079559, + "memory(GiB)": 111.15, + "step": 52310, + "train_speed(iter/s)": 0.447149 + }, + { + "acc": 0.66116157, + "epoch": 1.3271182141045155, + "grad_norm": 5.34375, + "learning_rate": 2.788297920034727e-06, + "loss": 1.58883286, + "memory(GiB)": 111.15, + "step": 52315, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.65766153, + "epoch": 1.3272450532724505, + "grad_norm": 6.0, + "learning_rate": 2.78735751343829e-06, + "loss": 1.63275776, + "memory(GiB)": 111.15, + "step": 52320, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.6500865, + "epoch": 1.3273718924403857, + "grad_norm": 5.8125, + "learning_rate": 2.786417204163748e-06, + "loss": 1.63558235, + "memory(GiB)": 111.15, + "step": 52325, + "train_speed(iter/s)": 0.447217 + }, + { + "acc": 0.64498119, + "epoch": 1.3274987316083207, + "grad_norm": 4.71875, + "learning_rate": 2.7854769922524593e-06, + "loss": 1.61316528, + "memory(GiB)": 111.15, + "step": 52330, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.6551404, + "epoch": 1.3276255707762556, + "grad_norm": 5.4375, + "learning_rate": 2.7845368777457803e-06, + "loss": 1.64711761, + "memory(GiB)": 111.15, + "step": 52335, + "train_speed(iter/s)": 0.447261 + }, + { + "acc": 0.65340734, + "epoch": 1.3277524099441909, + "grad_norm": 5.25, + "learning_rate": 2.7835968606850616e-06, + "loss": 1.67016792, + "memory(GiB)": 111.15, + "step": 52340, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.6383204, + "epoch": 1.3278792491121258, + "grad_norm": 5.90625, + "learning_rate": 2.782656941111648e-06, + "loss": 1.63859749, + "memory(GiB)": 111.15, + "step": 52345, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.66359091, + "epoch": 1.3280060882800608, + "grad_norm": 6.75, + "learning_rate": 2.7817171190668812e-06, + "loss": 1.52025909, + "memory(GiB)": 111.15, + "step": 52350, + "train_speed(iter/s)": 0.447329 + }, + { + "acc": 0.64622078, + "epoch": 1.3281329274479958, + "grad_norm": 5.625, + "learning_rate": 2.7807773945921e-06, + "loss": 1.63784294, + "memory(GiB)": 111.15, + "step": 52355, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.66414585, + "epoch": 1.328259766615931, + "grad_norm": 4.84375, + "learning_rate": 2.7798377677286363e-06, + "loss": 1.59609432, + "memory(GiB)": 111.15, + "step": 52360, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.65351386, + "epoch": 1.328386605783866, + "grad_norm": 5.4375, + "learning_rate": 2.778898238517821e-06, + "loss": 1.55859947, + "memory(GiB)": 111.15, + "step": 52365, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.64565239, + "epoch": 1.3285134449518012, + "grad_norm": 5.09375, + "learning_rate": 2.7779588070009767e-06, + "loss": 1.61928024, + "memory(GiB)": 111.15, + "step": 52370, + "train_speed(iter/s)": 0.447419 + }, + { + "acc": 0.65534434, + "epoch": 1.3286402841197362, + "grad_norm": 5.09375, + "learning_rate": 2.7770194732194256e-06, + "loss": 1.6037468, + "memory(GiB)": 111.15, + "step": 52375, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.6443974, + "epoch": 1.3287671232876712, + "grad_norm": 4.53125, + "learning_rate": 2.7760802372144825e-06, + "loss": 1.67486916, + "memory(GiB)": 111.15, + "step": 52380, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.66042356, + "epoch": 1.3288939624556062, + "grad_norm": 6.09375, + "learning_rate": 2.7751410990274596e-06, + "loss": 1.60646439, + "memory(GiB)": 111.15, + "step": 52385, + "train_speed(iter/s)": 0.447487 + }, + { + "acc": 0.65320311, + "epoch": 1.3290208016235414, + "grad_norm": 5.625, + "learning_rate": 2.774202058699664e-06, + "loss": 1.53985758, + "memory(GiB)": 111.15, + "step": 52390, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.64811983, + "epoch": 1.3291476407914764, + "grad_norm": 4.65625, + "learning_rate": 2.7732631162724005e-06, + "loss": 1.59303188, + "memory(GiB)": 111.15, + "step": 52395, + "train_speed(iter/s)": 0.447532 + }, + { + "acc": 0.65029635, + "epoch": 1.3292744799594114, + "grad_norm": 4.4375, + "learning_rate": 2.772324271786966e-06, + "loss": 1.63236294, + "memory(GiB)": 111.15, + "step": 52400, + "train_speed(iter/s)": 0.447554 + }, + { + "epoch": 1.3292744799594114, + "eval_acc": 0.643400091156827, + "eval_loss": 1.607512354850769, + "eval_runtime": 113.2952, + "eval_samples_per_second": 56.225, + "eval_steps_per_second": 28.112, + "step": 52400 + }, + { + "acc": 0.65643592, + "epoch": 1.3294013191273466, + "grad_norm": 5.625, + "learning_rate": 2.7713855252846545e-06, + "loss": 1.62707596, + "memory(GiB)": 111.15, + "step": 52405, + "train_speed(iter/s)": 0.447127 + }, + { + "acc": 0.65401134, + "epoch": 1.3295281582952816, + "grad_norm": 5.1875, + "learning_rate": 2.7704468768067616e-06, + "loss": 1.56421375, + "memory(GiB)": 111.15, + "step": 52410, + "train_speed(iter/s)": 0.447149 + }, + { + "acc": 0.63932686, + "epoch": 1.3296549974632166, + "grad_norm": 6.90625, + "learning_rate": 2.7695083263945664e-06, + "loss": 1.69855385, + "memory(GiB)": 111.15, + "step": 52415, + "train_speed(iter/s)": 0.447172 + }, + { + "acc": 0.64069662, + "epoch": 1.3297818366311516, + "grad_norm": 5.46875, + "learning_rate": 2.7685698740893516e-06, + "loss": 1.69707184, + "memory(GiB)": 111.15, + "step": 52420, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.65318012, + "epoch": 1.3299086757990868, + "grad_norm": 5.96875, + "learning_rate": 2.7676315199323995e-06, + "loss": 1.63118629, + "memory(GiB)": 111.15, + "step": 52425, + "train_speed(iter/s)": 0.447217 + }, + { + "acc": 0.66689315, + "epoch": 1.3300355149670218, + "grad_norm": 5.75, + "learning_rate": 2.7666932639649814e-06, + "loss": 1.547118, + "memory(GiB)": 111.15, + "step": 52430, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.64043655, + "epoch": 1.330162354134957, + "grad_norm": 5.09375, + "learning_rate": 2.765755106228362e-06, + "loss": 1.62142429, + "memory(GiB)": 111.15, + "step": 52435, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.6451354, + "epoch": 1.330289193302892, + "grad_norm": 5.90625, + "learning_rate": 2.764817046763807e-06, + "loss": 1.61909657, + "memory(GiB)": 111.15, + "step": 52440, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.65412731, + "epoch": 1.330416032470827, + "grad_norm": 5.375, + "learning_rate": 2.7638790856125786e-06, + "loss": 1.65015774, + "memory(GiB)": 111.15, + "step": 52445, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.64888215, + "epoch": 1.330542871638762, + "grad_norm": 5.28125, + "learning_rate": 2.7629412228159346e-06, + "loss": 1.66187363, + "memory(GiB)": 111.15, + "step": 52450, + "train_speed(iter/s)": 0.447329 + }, + { + "acc": 0.64430885, + "epoch": 1.3306697108066972, + "grad_norm": 6.25, + "learning_rate": 2.762003458415119e-06, + "loss": 1.65838203, + "memory(GiB)": 111.15, + "step": 52455, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.65810924, + "epoch": 1.3307965499746321, + "grad_norm": 5.15625, + "learning_rate": 2.7610657924513853e-06, + "loss": 1.62168617, + "memory(GiB)": 111.15, + "step": 52460, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.65891066, + "epoch": 1.3309233891425674, + "grad_norm": 5.5, + "learning_rate": 2.7601282249659737e-06, + "loss": 1.64532299, + "memory(GiB)": 111.15, + "step": 52465, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.63913946, + "epoch": 1.3310502283105023, + "grad_norm": 5.8125, + "learning_rate": 2.759190756000126e-06, + "loss": 1.69640503, + "memory(GiB)": 111.15, + "step": 52470, + "train_speed(iter/s)": 0.44742 + }, + { + "acc": 0.64560795, + "epoch": 1.3311770674784373, + "grad_norm": 6.5625, + "learning_rate": 2.7582533855950687e-06, + "loss": 1.66519241, + "memory(GiB)": 111.15, + "step": 52475, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.64546876, + "epoch": 1.3313039066463723, + "grad_norm": 5.1875, + "learning_rate": 2.757316113792038e-06, + "loss": 1.66235714, + "memory(GiB)": 111.15, + "step": 52480, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.63341217, + "epoch": 1.3314307458143075, + "grad_norm": 4.8125, + "learning_rate": 2.756378940632258e-06, + "loss": 1.66120186, + "memory(GiB)": 111.15, + "step": 52485, + "train_speed(iter/s)": 0.447487 + }, + { + "acc": 0.63737078, + "epoch": 1.3315575849822425, + "grad_norm": 6.125, + "learning_rate": 2.755441866156949e-06, + "loss": 1.70366135, + "memory(GiB)": 111.15, + "step": 52490, + "train_speed(iter/s)": 0.44751 + }, + { + "acc": 0.65611734, + "epoch": 1.3316844241501775, + "grad_norm": 5.125, + "learning_rate": 2.7545048904073278e-06, + "loss": 1.6495306, + "memory(GiB)": 111.15, + "step": 52495, + "train_speed(iter/s)": 0.447533 + }, + { + "acc": 0.65985022, + "epoch": 1.3318112633181127, + "grad_norm": 5.0625, + "learning_rate": 2.7535680134246067e-06, + "loss": 1.61899147, + "memory(GiB)": 111.15, + "step": 52500, + "train_speed(iter/s)": 0.447555 + }, + { + "epoch": 1.3318112633181127, + "eval_acc": 0.6434109531068728, + "eval_loss": 1.6075339317321777, + "eval_runtime": 114.0734, + "eval_samples_per_second": 55.841, + "eval_steps_per_second": 27.921, + "step": 52500 + }, + { + "acc": 0.66929626, + "epoch": 1.3319381024860477, + "grad_norm": 5.0625, + "learning_rate": 2.752631235249995e-06, + "loss": 1.62384148, + "memory(GiB)": 111.15, + "step": 52505, + "train_speed(iter/s)": 0.447125 + }, + { + "acc": 0.65940037, + "epoch": 1.3320649416539827, + "grad_norm": 5.90625, + "learning_rate": 2.7516945559246945e-06, + "loss": 1.53961611, + "memory(GiB)": 111.15, + "step": 52510, + "train_speed(iter/s)": 0.447148 + }, + { + "acc": 0.64912672, + "epoch": 1.3321917808219177, + "grad_norm": 5.28125, + "learning_rate": 2.7507579754899053e-06, + "loss": 1.5987112, + "memory(GiB)": 111.15, + "step": 52515, + "train_speed(iter/s)": 0.44717 + }, + { + "acc": 0.66625929, + "epoch": 1.332318619989853, + "grad_norm": 5.625, + "learning_rate": 2.749821493986823e-06, + "loss": 1.53242416, + "memory(GiB)": 111.15, + "step": 52520, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.66253343, + "epoch": 1.332445459157788, + "grad_norm": 5.96875, + "learning_rate": 2.748885111456637e-06, + "loss": 1.62852077, + "memory(GiB)": 111.15, + "step": 52525, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.6605175, + "epoch": 1.332572298325723, + "grad_norm": 6.15625, + "learning_rate": 2.7479488279405354e-06, + "loss": 1.6123806, + "memory(GiB)": 111.15, + "step": 52530, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.65388393, + "epoch": 1.332699137493658, + "grad_norm": 5.96875, + "learning_rate": 2.7470126434796984e-06, + "loss": 1.60558472, + "memory(GiB)": 111.15, + "step": 52535, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.63599892, + "epoch": 1.332825976661593, + "grad_norm": 5.4375, + "learning_rate": 2.746076558115304e-06, + "loss": 1.67281132, + "memory(GiB)": 111.15, + "step": 52540, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.66483622, + "epoch": 1.332952815829528, + "grad_norm": 5.0625, + "learning_rate": 2.7451405718885237e-06, + "loss": 1.58534985, + "memory(GiB)": 111.15, + "step": 52545, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.64917707, + "epoch": 1.3330796549974633, + "grad_norm": 5.3125, + "learning_rate": 2.7442046848405328e-06, + "loss": 1.57776146, + "memory(GiB)": 111.15, + "step": 52550, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.66196413, + "epoch": 1.3332064941653983, + "grad_norm": 5.03125, + "learning_rate": 2.743268897012489e-06, + "loss": 1.56166677, + "memory(GiB)": 111.15, + "step": 52555, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.64267707, + "epoch": 1.3333333333333333, + "grad_norm": 5.90625, + "learning_rate": 2.7423332084455543e-06, + "loss": 1.64629021, + "memory(GiB)": 111.15, + "step": 52560, + "train_speed(iter/s)": 0.447372 + }, + { + "acc": 0.66486483, + "epoch": 1.3334601725012685, + "grad_norm": 5.6875, + "learning_rate": 2.741397619180883e-06, + "loss": 1.51922836, + "memory(GiB)": 111.15, + "step": 52565, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.64411945, + "epoch": 1.3335870116692035, + "grad_norm": 5.21875, + "learning_rate": 2.740462129259633e-06, + "loss": 1.69503899, + "memory(GiB)": 111.15, + "step": 52570, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.64473686, + "epoch": 1.3337138508371384, + "grad_norm": 5.03125, + "learning_rate": 2.739526738722944e-06, + "loss": 1.68766899, + "memory(GiB)": 111.15, + "step": 52575, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.65661812, + "epoch": 1.3338406900050734, + "grad_norm": 4.96875, + "learning_rate": 2.738591447611959e-06, + "loss": 1.57193947, + "memory(GiB)": 111.15, + "step": 52580, + "train_speed(iter/s)": 0.447463 + }, + { + "acc": 0.63269653, + "epoch": 1.3339675291730086, + "grad_norm": 5.53125, + "learning_rate": 2.7376562559678214e-06, + "loss": 1.67139969, + "memory(GiB)": 111.15, + "step": 52585, + "train_speed(iter/s)": 0.447485 + }, + { + "acc": 0.65087762, + "epoch": 1.3340943683409436, + "grad_norm": 6.3125, + "learning_rate": 2.7367211638316637e-06, + "loss": 1.68400345, + "memory(GiB)": 111.15, + "step": 52590, + "train_speed(iter/s)": 0.447508 + }, + { + "acc": 0.65782366, + "epoch": 1.3342212075088788, + "grad_norm": 4.84375, + "learning_rate": 2.735786171244611e-06, + "loss": 1.62124252, + "memory(GiB)": 111.15, + "step": 52595, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.63886976, + "epoch": 1.3343480466768138, + "grad_norm": 5.25, + "learning_rate": 2.7348512782477922e-06, + "loss": 1.62755547, + "memory(GiB)": 111.15, + "step": 52600, + "train_speed(iter/s)": 0.447553 + }, + { + "epoch": 1.3343480466768138, + "eval_acc": 0.6434610851840077, + "eval_loss": 1.607561707496643, + "eval_runtime": 113.5926, + "eval_samples_per_second": 56.078, + "eval_steps_per_second": 28.039, + "step": 52600 + }, + { + "acc": 0.65200443, + "epoch": 1.3344748858447488, + "grad_norm": 6.0625, + "learning_rate": 2.7339164848823287e-06, + "loss": 1.60147133, + "memory(GiB)": 111.15, + "step": 52605, + "train_speed(iter/s)": 0.447126 + }, + { + "acc": 0.6489748, + "epoch": 1.3346017250126838, + "grad_norm": 5.71875, + "learning_rate": 2.7329817911893365e-06, + "loss": 1.66016006, + "memory(GiB)": 111.15, + "step": 52610, + "train_speed(iter/s)": 0.447148 + }, + { + "acc": 0.65182171, + "epoch": 1.334728564180619, + "grad_norm": 5.59375, + "learning_rate": 2.7320471972099226e-06, + "loss": 1.61597233, + "memory(GiB)": 111.15, + "step": 52615, + "train_speed(iter/s)": 0.44717 + }, + { + "acc": 0.65115376, + "epoch": 1.334855403348554, + "grad_norm": 4.90625, + "learning_rate": 2.7311127029852007e-06, + "loss": 1.61048679, + "memory(GiB)": 111.15, + "step": 52620, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.63780551, + "epoch": 1.3349822425164892, + "grad_norm": 5.1875, + "learning_rate": 2.7301783085562726e-06, + "loss": 1.69403648, + "memory(GiB)": 111.15, + "step": 52625, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.6502841, + "epoch": 1.3351090816844242, + "grad_norm": 5.25, + "learning_rate": 2.7292440139642364e-06, + "loss": 1.65515251, + "memory(GiB)": 111.15, + "step": 52630, + "train_speed(iter/s)": 0.447238 + }, + { + "acc": 0.64561157, + "epoch": 1.3352359208523592, + "grad_norm": 5.28125, + "learning_rate": 2.7283098192501855e-06, + "loss": 1.60465851, + "memory(GiB)": 111.15, + "step": 52635, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.65191488, + "epoch": 1.3353627600202942, + "grad_norm": 5.5, + "learning_rate": 2.7273757244552124e-06, + "loss": 1.54540596, + "memory(GiB)": 111.15, + "step": 52640, + "train_speed(iter/s)": 0.447283 + }, + { + "acc": 0.65430346, + "epoch": 1.3354895991882294, + "grad_norm": 5.15625, + "learning_rate": 2.726441729620401e-06, + "loss": 1.59909019, + "memory(GiB)": 111.15, + "step": 52645, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.65815258, + "epoch": 1.3356164383561644, + "grad_norm": 5.03125, + "learning_rate": 2.725507834786833e-06, + "loss": 1.64427319, + "memory(GiB)": 111.15, + "step": 52650, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.6541852, + "epoch": 1.3357432775240994, + "grad_norm": 5.8125, + "learning_rate": 2.7245740399955857e-06, + "loss": 1.62635994, + "memory(GiB)": 111.15, + "step": 52655, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.65479126, + "epoch": 1.3358701166920346, + "grad_norm": 5.0, + "learning_rate": 2.72364034528773e-06, + "loss": 1.61537666, + "memory(GiB)": 111.15, + "step": 52660, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.65765753, + "epoch": 1.3359969558599696, + "grad_norm": 5.3125, + "learning_rate": 2.722706750704337e-06, + "loss": 1.61468487, + "memory(GiB)": 111.15, + "step": 52665, + "train_speed(iter/s)": 0.447396 + }, + { + "acc": 0.64497232, + "epoch": 1.3361237950279046, + "grad_norm": 5.4375, + "learning_rate": 2.7217732562864673e-06, + "loss": 1.64504128, + "memory(GiB)": 111.15, + "step": 52670, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.63576684, + "epoch": 1.3362506341958396, + "grad_norm": 6.8125, + "learning_rate": 2.720839862075181e-06, + "loss": 1.66652126, + "memory(GiB)": 111.15, + "step": 52675, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.65718546, + "epoch": 1.3363774733637748, + "grad_norm": 4.59375, + "learning_rate": 2.7199065681115344e-06, + "loss": 1.62274837, + "memory(GiB)": 111.15, + "step": 52680, + "train_speed(iter/s)": 0.447463 + }, + { + "acc": 0.6507391, + "epoch": 1.3365043125317098, + "grad_norm": 5.875, + "learning_rate": 2.7189733744365742e-06, + "loss": 1.61160049, + "memory(GiB)": 111.15, + "step": 52685, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.64688416, + "epoch": 1.336631151699645, + "grad_norm": 5.28125, + "learning_rate": 2.718040281091353e-06, + "loss": 1.57212868, + "memory(GiB)": 111.15, + "step": 52690, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.65037251, + "epoch": 1.33675799086758, + "grad_norm": 5.34375, + "learning_rate": 2.717107288116906e-06, + "loss": 1.59334316, + "memory(GiB)": 111.15, + "step": 52695, + "train_speed(iter/s)": 0.447531 + }, + { + "acc": 0.65334435, + "epoch": 1.336884830035515, + "grad_norm": 5.0625, + "learning_rate": 2.716174395554274e-06, + "loss": 1.62640648, + "memory(GiB)": 111.15, + "step": 52700, + "train_speed(iter/s)": 0.447554 + }, + { + "epoch": 1.336884830035515, + "eval_acc": 0.6434025977606836, + "eval_loss": 1.6076854467391968, + "eval_runtime": 114.2529, + "eval_samples_per_second": 55.753, + "eval_steps_per_second": 27.877, + "step": 52700 + }, + { + "acc": 0.65734053, + "epoch": 1.33701166920345, + "grad_norm": 4.75, + "learning_rate": 2.715241603444486e-06, + "loss": 1.65646095, + "memory(GiB)": 111.15, + "step": 52705, + "train_speed(iter/s)": 0.447125 + }, + { + "acc": 0.65683508, + "epoch": 1.3371385083713851, + "grad_norm": 5.1875, + "learning_rate": 2.714308911828577e-06, + "loss": 1.57499371, + "memory(GiB)": 111.15, + "step": 52710, + "train_speed(iter/s)": 0.447147 + }, + { + "acc": 0.64546719, + "epoch": 1.3372653475393201, + "grad_norm": 5.5, + "learning_rate": 2.713376320747565e-06, + "loss": 1.68561363, + "memory(GiB)": 111.15, + "step": 52715, + "train_speed(iter/s)": 0.447169 + }, + { + "acc": 0.64315352, + "epoch": 1.3373921867072551, + "grad_norm": 5.4375, + "learning_rate": 2.7124438302424696e-06, + "loss": 1.62076626, + "memory(GiB)": 111.15, + "step": 52720, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.64043331, + "epoch": 1.3375190258751903, + "grad_norm": 6.40625, + "learning_rate": 2.711511440354309e-06, + "loss": 1.62980652, + "memory(GiB)": 111.15, + "step": 52725, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.65095749, + "epoch": 1.3376458650431253, + "grad_norm": 6.0625, + "learning_rate": 2.710579151124095e-06, + "loss": 1.65370846, + "memory(GiB)": 111.15, + "step": 52730, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.64828339, + "epoch": 1.3377727042110603, + "grad_norm": 5.15625, + "learning_rate": 2.70964696259283e-06, + "loss": 1.60217934, + "memory(GiB)": 111.15, + "step": 52735, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.6442688, + "epoch": 1.3378995433789953, + "grad_norm": 5.09375, + "learning_rate": 2.7087148748015146e-06, + "loss": 1.6166399, + "memory(GiB)": 111.15, + "step": 52740, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.62793598, + "epoch": 1.3380263825469305, + "grad_norm": 5.125, + "learning_rate": 2.7077828877911517e-06, + "loss": 1.76788788, + "memory(GiB)": 111.15, + "step": 52745, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.64898701, + "epoch": 1.3381532217148655, + "grad_norm": 6.75, + "learning_rate": 2.706851001602733e-06, + "loss": 1.56905594, + "memory(GiB)": 111.15, + "step": 52750, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.64697104, + "epoch": 1.3382800608828007, + "grad_norm": 4.78125, + "learning_rate": 2.7059192162772407e-06, + "loss": 1.63563023, + "memory(GiB)": 111.15, + "step": 52755, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.65281076, + "epoch": 1.3384069000507357, + "grad_norm": 6.1875, + "learning_rate": 2.704987531855666e-06, + "loss": 1.63738575, + "memory(GiB)": 111.15, + "step": 52760, + "train_speed(iter/s)": 0.447371 + }, + { + "acc": 0.66172066, + "epoch": 1.3385337392186707, + "grad_norm": 5.1875, + "learning_rate": 2.704055948378986e-06, + "loss": 1.64370937, + "memory(GiB)": 111.15, + "step": 52765, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.64877129, + "epoch": 1.3386605783866057, + "grad_norm": 5.59375, + "learning_rate": 2.7031244658881773e-06, + "loss": 1.56079273, + "memory(GiB)": 111.15, + "step": 52770, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.65590582, + "epoch": 1.338787417554541, + "grad_norm": 6.21875, + "learning_rate": 2.7021930844242085e-06, + "loss": 1.63638515, + "memory(GiB)": 111.15, + "step": 52775, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.65453339, + "epoch": 1.3389142567224759, + "grad_norm": 6.375, + "learning_rate": 2.7012618040280463e-06, + "loss": 1.63513641, + "memory(GiB)": 111.15, + "step": 52780, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.66635132, + "epoch": 1.339041095890411, + "grad_norm": 5.90625, + "learning_rate": 2.7003306247406536e-06, + "loss": 1.57220821, + "memory(GiB)": 111.15, + "step": 52785, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.63315563, + "epoch": 1.339167935058346, + "grad_norm": 5.34375, + "learning_rate": 2.6993995466029877e-06, + "loss": 1.71693153, + "memory(GiB)": 111.15, + "step": 52790, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.63633857, + "epoch": 1.339294774226281, + "grad_norm": 4.65625, + "learning_rate": 2.6984685696560002e-06, + "loss": 1.64617271, + "memory(GiB)": 111.15, + "step": 52795, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.64936905, + "epoch": 1.339421613394216, + "grad_norm": 4.90625, + "learning_rate": 2.6975376939406418e-06, + "loss": 1.61023178, + "memory(GiB)": 111.15, + "step": 52800, + "train_speed(iter/s)": 0.447549 + }, + { + "epoch": 1.339421613394216, + "eval_acc": 0.643429334868489, + "eval_loss": 1.6076077222824097, + "eval_runtime": 113.8288, + "eval_samples_per_second": 55.961, + "eval_steps_per_second": 27.981, + "step": 52800 + }, + { + "acc": 0.65409808, + "epoch": 1.3395484525621513, + "grad_norm": 4.78125, + "learning_rate": 2.6966069194978537e-06, + "loss": 1.5957345, + "memory(GiB)": 111.15, + "step": 52805, + "train_speed(iter/s)": 0.447123 + }, + { + "acc": 0.63743877, + "epoch": 1.3396752917300863, + "grad_norm": 5.59375, + "learning_rate": 2.6956762463685787e-06, + "loss": 1.67296562, + "memory(GiB)": 111.15, + "step": 52810, + "train_speed(iter/s)": 0.447145 + }, + { + "acc": 0.65170846, + "epoch": 1.3398021308980212, + "grad_norm": 4.84375, + "learning_rate": 2.69474567459375e-06, + "loss": 1.65565567, + "memory(GiB)": 111.15, + "step": 52815, + "train_speed(iter/s)": 0.447167 + }, + { + "acc": 0.6531374, + "epoch": 1.3399289700659565, + "grad_norm": 6.4375, + "learning_rate": 2.693815204214299e-06, + "loss": 1.64371262, + "memory(GiB)": 111.15, + "step": 52820, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.65587006, + "epoch": 1.3400558092338914, + "grad_norm": 5.28125, + "learning_rate": 2.692884835271151e-06, + "loss": 1.62098656, + "memory(GiB)": 111.15, + "step": 52825, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.66897106, + "epoch": 1.3401826484018264, + "grad_norm": 6.75, + "learning_rate": 2.6919545678052296e-06, + "loss": 1.61869164, + "memory(GiB)": 111.15, + "step": 52830, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.66709929, + "epoch": 1.3403094875697614, + "grad_norm": 6.15625, + "learning_rate": 2.69102440185745e-06, + "loss": 1.5184948, + "memory(GiB)": 111.15, + "step": 52835, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.63171692, + "epoch": 1.3404363267376966, + "grad_norm": 4.96875, + "learning_rate": 2.690094337468726e-06, + "loss": 1.76163521, + "memory(GiB)": 111.15, + "step": 52840, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.63918438, + "epoch": 1.3405631659056316, + "grad_norm": 5.8125, + "learning_rate": 2.6891643746799643e-06, + "loss": 1.65637531, + "memory(GiB)": 111.15, + "step": 52845, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.6195014, + "epoch": 1.3406900050735668, + "grad_norm": 4.875, + "learning_rate": 2.6882345135320753e-06, + "loss": 1.67640152, + "memory(GiB)": 111.15, + "step": 52850, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.65539799, + "epoch": 1.3408168442415018, + "grad_norm": 5.125, + "learning_rate": 2.68730475406595e-06, + "loss": 1.65568085, + "memory(GiB)": 111.15, + "step": 52855, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.63879261, + "epoch": 1.3409436834094368, + "grad_norm": 5.34375, + "learning_rate": 2.6863750963224867e-06, + "loss": 1.68573666, + "memory(GiB)": 111.15, + "step": 52860, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.63584709, + "epoch": 1.3410705225773718, + "grad_norm": 5.21875, + "learning_rate": 2.685445540342577e-06, + "loss": 1.68384724, + "memory(GiB)": 111.15, + "step": 52865, + "train_speed(iter/s)": 0.447392 + }, + { + "acc": 0.63756437, + "epoch": 1.341197361745307, + "grad_norm": 5.8125, + "learning_rate": 2.6845160861671094e-06, + "loss": 1.7170454, + "memory(GiB)": 111.15, + "step": 52870, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.64868603, + "epoch": 1.341324200913242, + "grad_norm": 5.5, + "learning_rate": 2.6835867338369593e-06, + "loss": 1.57301254, + "memory(GiB)": 111.15, + "step": 52875, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.65434341, + "epoch": 1.341451040081177, + "grad_norm": 5.28125, + "learning_rate": 2.6826574833930053e-06, + "loss": 1.55393143, + "memory(GiB)": 111.15, + "step": 52880, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.64401407, + "epoch": 1.3415778792491122, + "grad_norm": 6.3125, + "learning_rate": 2.681728334876123e-06, + "loss": 1.64500122, + "memory(GiB)": 111.15, + "step": 52885, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.66139364, + "epoch": 1.3417047184170472, + "grad_norm": 6.21875, + "learning_rate": 2.6807992883271806e-06, + "loss": 1.53819666, + "memory(GiB)": 111.15, + "step": 52890, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.65319109, + "epoch": 1.3418315575849822, + "grad_norm": 6.25, + "learning_rate": 2.6798703437870364e-06, + "loss": 1.60216579, + "memory(GiB)": 111.15, + "step": 52895, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.65750523, + "epoch": 1.3419583967529172, + "grad_norm": 5.5625, + "learning_rate": 2.678941501296555e-06, + "loss": 1.61751633, + "memory(GiB)": 111.15, + "step": 52900, + "train_speed(iter/s)": 0.447549 + }, + { + "epoch": 1.3419583967529172, + "eval_acc": 0.6433792027913541, + "eval_loss": 1.607611060142517, + "eval_runtime": 114.6158, + "eval_samples_per_second": 55.577, + "eval_steps_per_second": 27.788, + "step": 52900 + }, + { + "acc": 0.66590586, + "epoch": 1.3420852359208524, + "grad_norm": 5.15625, + "learning_rate": 2.6780127608965896e-06, + "loss": 1.58479519, + "memory(GiB)": 111.15, + "step": 52905, + "train_speed(iter/s)": 0.447121 + }, + { + "acc": 0.65700407, + "epoch": 1.3422120750887874, + "grad_norm": 5.15625, + "learning_rate": 2.677084122627991e-06, + "loss": 1.63002396, + "memory(GiB)": 111.15, + "step": 52910, + "train_speed(iter/s)": 0.447143 + }, + { + "acc": 0.6463037, + "epoch": 1.3423389142567226, + "grad_norm": 4.28125, + "learning_rate": 2.6761555865316003e-06, + "loss": 1.67876854, + "memory(GiB)": 111.15, + "step": 52915, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.66267214, + "epoch": 1.3424657534246576, + "grad_norm": 5.625, + "learning_rate": 2.6752271526482644e-06, + "loss": 1.64342308, + "memory(GiB)": 111.15, + "step": 52920, + "train_speed(iter/s)": 0.447188 + }, + { + "acc": 0.63780928, + "epoch": 1.3425925925925926, + "grad_norm": 4.96875, + "learning_rate": 2.674298821018817e-06, + "loss": 1.6771225, + "memory(GiB)": 111.15, + "step": 52925, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.64751873, + "epoch": 1.3427194317605275, + "grad_norm": 5.875, + "learning_rate": 2.673370591684091e-06, + "loss": 1.67398415, + "memory(GiB)": 111.15, + "step": 52930, + "train_speed(iter/s)": 0.447234 + }, + { + "acc": 0.64302788, + "epoch": 1.3428462709284628, + "grad_norm": 5.0625, + "learning_rate": 2.672442464684915e-06, + "loss": 1.70577488, + "memory(GiB)": 111.15, + "step": 52935, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.65356364, + "epoch": 1.3429731100963977, + "grad_norm": 4.5625, + "learning_rate": 2.671514440062111e-06, + "loss": 1.60000916, + "memory(GiB)": 111.15, + "step": 52940, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.65313587, + "epoch": 1.343099949264333, + "grad_norm": 5.5, + "learning_rate": 2.6705865178564973e-06, + "loss": 1.61751003, + "memory(GiB)": 111.15, + "step": 52945, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.6412715, + "epoch": 1.343226788432268, + "grad_norm": 5.0, + "learning_rate": 2.6696586981088886e-06, + "loss": 1.66012287, + "memory(GiB)": 111.15, + "step": 52950, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.63567204, + "epoch": 1.343353627600203, + "grad_norm": 6.4375, + "learning_rate": 2.6687309808600947e-06, + "loss": 1.67595291, + "memory(GiB)": 111.15, + "step": 52955, + "train_speed(iter/s)": 0.447346 + }, + { + "acc": 0.64755268, + "epoch": 1.343480466768138, + "grad_norm": 5.375, + "learning_rate": 2.6678033661509208e-06, + "loss": 1.62432899, + "memory(GiB)": 111.15, + "step": 52960, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.64341068, + "epoch": 1.3436073059360731, + "grad_norm": 5.375, + "learning_rate": 2.6668758540221665e-06, + "loss": 1.67823982, + "memory(GiB)": 111.15, + "step": 52965, + "train_speed(iter/s)": 0.447392 + }, + { + "acc": 0.64505272, + "epoch": 1.3437341451040081, + "grad_norm": 6.09375, + "learning_rate": 2.66594844451463e-06, + "loss": 1.52192097, + "memory(GiB)": 111.15, + "step": 52970, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.65243387, + "epoch": 1.3438609842719431, + "grad_norm": 5.25, + "learning_rate": 2.6650211376691008e-06, + "loss": 1.69794369, + "memory(GiB)": 111.15, + "step": 52975, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.6491446, + "epoch": 1.3439878234398783, + "grad_norm": 4.5625, + "learning_rate": 2.664093933526368e-06, + "loss": 1.61785164, + "memory(GiB)": 111.15, + "step": 52980, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.64398189, + "epoch": 1.3441146626078133, + "grad_norm": 5.46875, + "learning_rate": 2.6631668321272097e-06, + "loss": 1.64575062, + "memory(GiB)": 111.15, + "step": 52985, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.64207001, + "epoch": 1.3442415017757483, + "grad_norm": 6.28125, + "learning_rate": 2.6622398335124122e-06, + "loss": 1.67398491, + "memory(GiB)": 111.15, + "step": 52990, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.64653335, + "epoch": 1.3443683409436833, + "grad_norm": 5.09375, + "learning_rate": 2.661312937722742e-06, + "loss": 1.61974354, + "memory(GiB)": 111.15, + "step": 52995, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.64991488, + "epoch": 1.3444951801116185, + "grad_norm": 4.8125, + "learning_rate": 2.6603861447989703e-06, + "loss": 1.60675316, + "memory(GiB)": 111.15, + "step": 53000, + "train_speed(iter/s)": 0.447549 + }, + { + "epoch": 1.3444951801116185, + "eval_acc": 0.6433904825087094, + "eval_loss": 1.6075634956359863, + "eval_runtime": 113.876, + "eval_samples_per_second": 55.938, + "eval_steps_per_second": 27.969, + "step": 53000 + }, + { + "acc": 0.65199599, + "epoch": 1.3446220192795535, + "grad_norm": 5.25, + "learning_rate": 2.65945945478186e-06, + "loss": 1.58602638, + "memory(GiB)": 111.15, + "step": 53005, + "train_speed(iter/s)": 0.447124 + }, + { + "acc": 0.65855322, + "epoch": 1.3447488584474887, + "grad_norm": 5.28125, + "learning_rate": 2.658532867712176e-06, + "loss": 1.62017555, + "memory(GiB)": 111.15, + "step": 53010, + "train_speed(iter/s)": 0.447147 + }, + { + "acc": 0.65495934, + "epoch": 1.3448756976154237, + "grad_norm": 5.375, + "learning_rate": 2.6576063836306687e-06, + "loss": 1.58591518, + "memory(GiB)": 111.15, + "step": 53015, + "train_speed(iter/s)": 0.447169 + }, + { + "acc": 0.65101709, + "epoch": 1.3450025367833587, + "grad_norm": 5.0625, + "learning_rate": 2.656680002578088e-06, + "loss": 1.65584145, + "memory(GiB)": 111.15, + "step": 53020, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.65269284, + "epoch": 1.3451293759512937, + "grad_norm": 5.0, + "learning_rate": 2.655753724595186e-06, + "loss": 1.5785408, + "memory(GiB)": 111.15, + "step": 53025, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.64311099, + "epoch": 1.3452562151192289, + "grad_norm": 5.40625, + "learning_rate": 2.6548275497227028e-06, + "loss": 1.61233158, + "memory(GiB)": 111.15, + "step": 53030, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.65003047, + "epoch": 1.3453830542871639, + "grad_norm": 4.84375, + "learning_rate": 2.6539014780013707e-06, + "loss": 1.59166851, + "memory(GiB)": 111.15, + "step": 53035, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.64481211, + "epoch": 1.3455098934550989, + "grad_norm": 5.8125, + "learning_rate": 2.6529755094719276e-06, + "loss": 1.68209095, + "memory(GiB)": 111.15, + "step": 53040, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.64038134, + "epoch": 1.345636732623034, + "grad_norm": 5.625, + "learning_rate": 2.652049644175101e-06, + "loss": 1.63699493, + "memory(GiB)": 111.15, + "step": 53045, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.65308857, + "epoch": 1.345763571790969, + "grad_norm": 4.65625, + "learning_rate": 2.6511238821516154e-06, + "loss": 1.608358, + "memory(GiB)": 111.15, + "step": 53050, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.65285807, + "epoch": 1.345890410958904, + "grad_norm": 4.6875, + "learning_rate": 2.650198223442185e-06, + "loss": 1.56640749, + "memory(GiB)": 111.15, + "step": 53055, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.64280272, + "epoch": 1.346017250126839, + "grad_norm": 5.34375, + "learning_rate": 2.6492726680875296e-06, + "loss": 1.66745243, + "memory(GiB)": 111.15, + "step": 53060, + "train_speed(iter/s)": 0.447371 + }, + { + "acc": 0.66888981, + "epoch": 1.3461440892947742, + "grad_norm": 5.5, + "learning_rate": 2.6483472161283576e-06, + "loss": 1.58021488, + "memory(GiB)": 111.15, + "step": 53065, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.6429472, + "epoch": 1.3462709284627092, + "grad_norm": 5.71875, + "learning_rate": 2.647421867605374e-06, + "loss": 1.60597687, + "memory(GiB)": 111.15, + "step": 53070, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.6515121, + "epoch": 1.3463977676306444, + "grad_norm": 5.3125, + "learning_rate": 2.6464966225592804e-06, + "loss": 1.61368389, + "memory(GiB)": 111.15, + "step": 53075, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.64693584, + "epoch": 1.3465246067985794, + "grad_norm": 5.3125, + "learning_rate": 2.645571481030773e-06, + "loss": 1.66502113, + "memory(GiB)": 111.15, + "step": 53080, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.66005626, + "epoch": 1.3466514459665144, + "grad_norm": 7.15625, + "learning_rate": 2.6446464430605434e-06, + "loss": 1.63581238, + "memory(GiB)": 111.15, + "step": 53085, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64274697, + "epoch": 1.3467782851344494, + "grad_norm": 4.96875, + "learning_rate": 2.6437215086892797e-06, + "loss": 1.61910553, + "memory(GiB)": 111.15, + "step": 53090, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.66252785, + "epoch": 1.3469051243023846, + "grad_norm": 5.53125, + "learning_rate": 2.642796677957664e-06, + "loss": 1.58016968, + "memory(GiB)": 111.15, + "step": 53095, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.64148493, + "epoch": 1.3470319634703196, + "grad_norm": 4.84375, + "learning_rate": 2.641871950906374e-06, + "loss": 1.67986031, + "memory(GiB)": 111.15, + "step": 53100, + "train_speed(iter/s)": 0.44755 + }, + { + "epoch": 1.3470319634703196, + "eval_acc": 0.6434264104973227, + "eval_loss": 1.6076393127441406, + "eval_runtime": 113.68, + "eval_samples_per_second": 56.034, + "eval_steps_per_second": 28.017, + "step": 53100 + }, + { + "acc": 0.6478404, + "epoch": 1.3471588026382548, + "grad_norm": 4.59375, + "learning_rate": 2.6409473275760843e-06, + "loss": 1.65549488, + "memory(GiB)": 111.15, + "step": 53105, + "train_speed(iter/s)": 0.447126 + }, + { + "acc": 0.65561476, + "epoch": 1.3472856418061898, + "grad_norm": 5.65625, + "learning_rate": 2.640022808007463e-06, + "loss": 1.68477936, + "memory(GiB)": 111.15, + "step": 53110, + "train_speed(iter/s)": 0.447148 + }, + { + "acc": 0.64887276, + "epoch": 1.3474124809741248, + "grad_norm": 5.09375, + "learning_rate": 2.639098392241176e-06, + "loss": 1.67058449, + "memory(GiB)": 111.15, + "step": 53115, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.63394504, + "epoch": 1.3475393201420598, + "grad_norm": 5.21875, + "learning_rate": 2.6381740803178826e-06, + "loss": 1.63873444, + "memory(GiB)": 111.15, + "step": 53120, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.64811258, + "epoch": 1.347666159309995, + "grad_norm": 5.21875, + "learning_rate": 2.6372498722782346e-06, + "loss": 1.60014172, + "memory(GiB)": 111.15, + "step": 53125, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.6443253, + "epoch": 1.34779299847793, + "grad_norm": 4.96875, + "learning_rate": 2.6363257681628907e-06, + "loss": 1.63096981, + "memory(GiB)": 111.15, + "step": 53130, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.65902195, + "epoch": 1.347919837645865, + "grad_norm": 6.09375, + "learning_rate": 2.635401768012491e-06, + "loss": 1.5810812, + "memory(GiB)": 111.15, + "step": 53135, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.65356193, + "epoch": 1.3480466768138002, + "grad_norm": 6.3125, + "learning_rate": 2.6344778718676783e-06, + "loss": 1.58186607, + "memory(GiB)": 111.15, + "step": 53140, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.64250717, + "epoch": 1.3481735159817352, + "grad_norm": 6.40625, + "learning_rate": 2.6335540797690886e-06, + "loss": 1.66270561, + "memory(GiB)": 111.15, + "step": 53145, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.66930809, + "epoch": 1.3483003551496702, + "grad_norm": 4.4375, + "learning_rate": 2.632630391757359e-06, + "loss": 1.56048155, + "memory(GiB)": 111.15, + "step": 53150, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.63815503, + "epoch": 1.3484271943176052, + "grad_norm": 5.59375, + "learning_rate": 2.6317068078731126e-06, + "loss": 1.58026638, + "memory(GiB)": 111.15, + "step": 53155, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.66023602, + "epoch": 1.3485540334855404, + "grad_norm": 5.0, + "learning_rate": 2.630783328156973e-06, + "loss": 1.54257603, + "memory(GiB)": 111.15, + "step": 53160, + "train_speed(iter/s)": 0.447371 + }, + { + "acc": 0.6424696, + "epoch": 1.3486808726534754, + "grad_norm": 5.46875, + "learning_rate": 2.629859952649562e-06, + "loss": 1.58299923, + "memory(GiB)": 111.15, + "step": 53165, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.67080746, + "epoch": 1.3488077118214106, + "grad_norm": 5.5, + "learning_rate": 2.628936681391494e-06, + "loss": 1.5610589, + "memory(GiB)": 111.15, + "step": 53170, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.65207176, + "epoch": 1.3489345509893456, + "grad_norm": 5.8125, + "learning_rate": 2.628013514423375e-06, + "loss": 1.64970875, + "memory(GiB)": 111.15, + "step": 53175, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.64146738, + "epoch": 1.3490613901572805, + "grad_norm": 5.4375, + "learning_rate": 2.6270904517858102e-06, + "loss": 1.67596741, + "memory(GiB)": 111.15, + "step": 53180, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.65317636, + "epoch": 1.3491882293252155, + "grad_norm": 6.09375, + "learning_rate": 2.6261674935194036e-06, + "loss": 1.64295731, + "memory(GiB)": 111.15, + "step": 53185, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.66222601, + "epoch": 1.3493150684931507, + "grad_norm": 5.65625, + "learning_rate": 2.6252446396647503e-06, + "loss": 1.51576424, + "memory(GiB)": 111.15, + "step": 53190, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.63153086, + "epoch": 1.3494419076610857, + "grad_norm": 5.34375, + "learning_rate": 2.6243218902624367e-06, + "loss": 1.72067013, + "memory(GiB)": 111.15, + "step": 53195, + "train_speed(iter/s)": 0.447525 + }, + { + "acc": 0.65056601, + "epoch": 1.3495687468290207, + "grad_norm": 5.59375, + "learning_rate": 2.6233992453530555e-06, + "loss": 1.61020775, + "memory(GiB)": 111.15, + "step": 53200, + "train_speed(iter/s)": 0.447547 + }, + { + "epoch": 1.3495687468290207, + "eval_acc": 0.6434280815665606, + "eval_loss": 1.6075913906097412, + "eval_runtime": 114.1926, + "eval_samples_per_second": 55.783, + "eval_steps_per_second": 27.891, + "step": 53200 + }, + { + "acc": 0.6672781, + "epoch": 1.349695585996956, + "grad_norm": 4.6875, + "learning_rate": 2.6224767049771856e-06, + "loss": 1.54008656, + "memory(GiB)": 111.15, + "step": 53205, + "train_speed(iter/s)": 0.447122 + }, + { + "acc": 0.62648487, + "epoch": 1.349822425164891, + "grad_norm": 6.625, + "learning_rate": 2.621554269175405e-06, + "loss": 1.70552254, + "memory(GiB)": 111.15, + "step": 53210, + "train_speed(iter/s)": 0.447144 + }, + { + "acc": 0.63169775, + "epoch": 1.349949264332826, + "grad_norm": 5.59375, + "learning_rate": 2.620631937988287e-06, + "loss": 1.72790146, + "memory(GiB)": 111.15, + "step": 53215, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.64441576, + "epoch": 1.350076103500761, + "grad_norm": 5.28125, + "learning_rate": 2.6197097114564e-06, + "loss": 1.61711159, + "memory(GiB)": 111.15, + "step": 53220, + "train_speed(iter/s)": 0.447188 + }, + { + "acc": 0.63913388, + "epoch": 1.3502029426686961, + "grad_norm": 5.40625, + "learning_rate": 2.618787589620306e-06, + "loss": 1.66379089, + "memory(GiB)": 111.15, + "step": 53225, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.66073771, + "epoch": 1.350329781836631, + "grad_norm": 6.0, + "learning_rate": 2.6178655725205653e-06, + "loss": 1.57526197, + "memory(GiB)": 111.15, + "step": 53230, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.66777515, + "epoch": 1.3504566210045663, + "grad_norm": 5.34375, + "learning_rate": 2.6169436601977325e-06, + "loss": 1.56533852, + "memory(GiB)": 111.15, + "step": 53235, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.64882898, + "epoch": 1.3505834601725013, + "grad_norm": 4.875, + "learning_rate": 2.6160218526923576e-06, + "loss": 1.65935173, + "memory(GiB)": 111.15, + "step": 53240, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.63170443, + "epoch": 1.3507102993404363, + "grad_norm": 6.1875, + "learning_rate": 2.6151001500449847e-06, + "loss": 1.67393684, + "memory(GiB)": 111.15, + "step": 53245, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.65574436, + "epoch": 1.3508371385083713, + "grad_norm": 4.84375, + "learning_rate": 2.614178552296155e-06, + "loss": 1.6366663, + "memory(GiB)": 111.15, + "step": 53250, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.64754243, + "epoch": 1.3509639776763065, + "grad_norm": 5.6875, + "learning_rate": 2.6132570594864047e-06, + "loss": 1.63735886, + "memory(GiB)": 111.15, + "step": 53255, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.67179976, + "epoch": 1.3510908168442415, + "grad_norm": 4.9375, + "learning_rate": 2.612335671656265e-06, + "loss": 1.62637558, + "memory(GiB)": 111.15, + "step": 53260, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.68238764, + "epoch": 1.3512176560121767, + "grad_norm": 5.375, + "learning_rate": 2.6114143888462607e-06, + "loss": 1.56880817, + "memory(GiB)": 111.15, + "step": 53265, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.65750904, + "epoch": 1.3513444951801117, + "grad_norm": 5.8125, + "learning_rate": 2.6104932110969195e-06, + "loss": 1.63726063, + "memory(GiB)": 111.15, + "step": 53270, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.65673599, + "epoch": 1.3514713343480467, + "grad_norm": 4.71875, + "learning_rate": 2.609572138448753e-06, + "loss": 1.55596294, + "memory(GiB)": 111.15, + "step": 53275, + "train_speed(iter/s)": 0.44743 + }, + { + "acc": 0.65587797, + "epoch": 1.3515981735159817, + "grad_norm": 4.90625, + "learning_rate": 2.608651170942277e-06, + "loss": 1.62563305, + "memory(GiB)": 111.15, + "step": 53280, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.64840479, + "epoch": 1.3517250126839169, + "grad_norm": 5.0625, + "learning_rate": 2.607730308617997e-06, + "loss": 1.60432777, + "memory(GiB)": 111.15, + "step": 53285, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.67096157, + "epoch": 1.3518518518518519, + "grad_norm": 5.59375, + "learning_rate": 2.6068095515164226e-06, + "loss": 1.56484928, + "memory(GiB)": 111.15, + "step": 53290, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.64149776, + "epoch": 1.3519786910197868, + "grad_norm": 5.21875, + "learning_rate": 2.605888899678047e-06, + "loss": 1.6483696, + "memory(GiB)": 111.15, + "step": 53295, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.65417848, + "epoch": 1.352105530187722, + "grad_norm": 5.90625, + "learning_rate": 2.6049683531433645e-06, + "loss": 1.63271313, + "memory(GiB)": 111.15, + "step": 53300, + "train_speed(iter/s)": 0.44754 + }, + { + "epoch": 1.352105530187722, + "eval_acc": 0.6434034332953026, + "eval_loss": 1.60764479637146, + "eval_runtime": 113.5262, + "eval_samples_per_second": 56.11, + "eval_steps_per_second": 28.055, + "step": 53300 + }, + { + "acc": 0.64832907, + "epoch": 1.352232369355657, + "grad_norm": 5.84375, + "learning_rate": 2.6040479119528683e-06, + "loss": 1.66874542, + "memory(GiB)": 111.15, + "step": 53305, + "train_speed(iter/s)": 0.447118 + }, + { + "acc": 0.65657959, + "epoch": 1.352359208523592, + "grad_norm": 5.15625, + "learning_rate": 2.6031275761470447e-06, + "loss": 1.64475307, + "memory(GiB)": 111.15, + "step": 53310, + "train_speed(iter/s)": 0.44714 + }, + { + "acc": 0.6316802, + "epoch": 1.352486047691527, + "grad_norm": 5.46875, + "learning_rate": 2.60220734576637e-06, + "loss": 1.68786926, + "memory(GiB)": 111.15, + "step": 53315, + "train_speed(iter/s)": 0.447162 + }, + { + "acc": 0.63557687, + "epoch": 1.3526128868594622, + "grad_norm": 5.5625, + "learning_rate": 2.601287220851319e-06, + "loss": 1.75430508, + "memory(GiB)": 111.15, + "step": 53320, + "train_speed(iter/s)": 0.447184 + }, + { + "acc": 0.67131796, + "epoch": 1.3527397260273972, + "grad_norm": 7.21875, + "learning_rate": 2.6003672014423677e-06, + "loss": 1.61812401, + "memory(GiB)": 111.15, + "step": 53325, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.64943476, + "epoch": 1.3528665651953324, + "grad_norm": 4.875, + "learning_rate": 2.5994472875799827e-06, + "loss": 1.66734524, + "memory(GiB)": 111.15, + "step": 53330, + "train_speed(iter/s)": 0.447228 + }, + { + "acc": 0.65356474, + "epoch": 1.3529934043632674, + "grad_norm": 4.09375, + "learning_rate": 2.598527479304619e-06, + "loss": 1.64034004, + "memory(GiB)": 111.15, + "step": 53335, + "train_speed(iter/s)": 0.44725 + }, + { + "acc": 0.66501927, + "epoch": 1.3531202435312024, + "grad_norm": 4.5625, + "learning_rate": 2.597607776656741e-06, + "loss": 1.57319202, + "memory(GiB)": 111.15, + "step": 53340, + "train_speed(iter/s)": 0.447273 + }, + { + "acc": 0.65428295, + "epoch": 1.3532470826991374, + "grad_norm": 5.5625, + "learning_rate": 2.5966881796767984e-06, + "loss": 1.61468925, + "memory(GiB)": 111.15, + "step": 53345, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.64293804, + "epoch": 1.3533739218670726, + "grad_norm": 8.375, + "learning_rate": 2.5957686884052423e-06, + "loss": 1.65622559, + "memory(GiB)": 111.15, + "step": 53350, + "train_speed(iter/s)": 0.447317 + }, + { + "acc": 0.64383144, + "epoch": 1.3535007610350076, + "grad_norm": 6.21875, + "learning_rate": 2.5948493028825093e-06, + "loss": 1.68648109, + "memory(GiB)": 111.15, + "step": 53355, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.65368242, + "epoch": 1.3536276002029426, + "grad_norm": 5.5, + "learning_rate": 2.593930023149044e-06, + "loss": 1.64864445, + "memory(GiB)": 111.15, + "step": 53360, + "train_speed(iter/s)": 0.447362 + }, + { + "acc": 0.64722157, + "epoch": 1.3537544393708778, + "grad_norm": 5.0625, + "learning_rate": 2.593010849245279e-06, + "loss": 1.61587334, + "memory(GiB)": 111.15, + "step": 53365, + "train_speed(iter/s)": 0.447385 + }, + { + "acc": 0.64364052, + "epoch": 1.3538812785388128, + "grad_norm": 5.5, + "learning_rate": 2.592091781211643e-06, + "loss": 1.63288441, + "memory(GiB)": 111.15, + "step": 53370, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.678578, + "epoch": 1.3540081177067478, + "grad_norm": 5.78125, + "learning_rate": 2.591172819088561e-06, + "loss": 1.51388626, + "memory(GiB)": 111.15, + "step": 53375, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.64251118, + "epoch": 1.3541349568746828, + "grad_norm": 4.40625, + "learning_rate": 2.590253962916453e-06, + "loss": 1.65375633, + "memory(GiB)": 111.15, + "step": 53380, + "train_speed(iter/s)": 0.447451 + }, + { + "acc": 0.65023251, + "epoch": 1.354261796042618, + "grad_norm": 5.03125, + "learning_rate": 2.5893352127357347e-06, + "loss": 1.54686375, + "memory(GiB)": 111.15, + "step": 53385, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.65224581, + "epoch": 1.354388635210553, + "grad_norm": 5.53125, + "learning_rate": 2.5884165685868164e-06, + "loss": 1.59068441, + "memory(GiB)": 111.15, + "step": 53390, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.63696012, + "epoch": 1.3545154743784882, + "grad_norm": 5.53125, + "learning_rate": 2.5874980305101045e-06, + "loss": 1.67543106, + "memory(GiB)": 111.15, + "step": 53395, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.648281, + "epoch": 1.3546423135464232, + "grad_norm": 5.59375, + "learning_rate": 2.586579598546e-06, + "loss": 1.65890579, + "memory(GiB)": 111.15, + "step": 53400, + "train_speed(iter/s)": 0.44754 + }, + { + "epoch": 1.3546423135464232, + "eval_acc": 0.6434142952453484, + "eval_loss": 1.6076840162277222, + "eval_runtime": 113.4108, + "eval_samples_per_second": 56.168, + "eval_steps_per_second": 28.084, + "step": 53400 + }, + { + "acc": 0.66322336, + "epoch": 1.3547691527143582, + "grad_norm": 6.28125, + "learning_rate": 2.5856612727348995e-06, + "loss": 1.54684649, + "memory(GiB)": 111.15, + "step": 53405, + "train_speed(iter/s)": 0.447121 + }, + { + "acc": 0.6635088, + "epoch": 1.3548959918822931, + "grad_norm": 5.40625, + "learning_rate": 2.584743053117196e-06, + "loss": 1.57237129, + "memory(GiB)": 111.15, + "step": 53410, + "train_speed(iter/s)": 0.447143 + }, + { + "acc": 0.66092834, + "epoch": 1.3550228310502284, + "grad_norm": 5.09375, + "learning_rate": 2.583824939733277e-06, + "loss": 1.527565, + "memory(GiB)": 111.15, + "step": 53415, + "train_speed(iter/s)": 0.447165 + }, + { + "acc": 0.64567366, + "epoch": 1.3551496702181633, + "grad_norm": 5.1875, + "learning_rate": 2.5829069326235234e-06, + "loss": 1.66879406, + "memory(GiB)": 111.15, + "step": 53420, + "train_speed(iter/s)": 0.447187 + }, + { + "acc": 0.63293648, + "epoch": 1.3552765093860986, + "grad_norm": 5.8125, + "learning_rate": 2.5819890318283137e-06, + "loss": 1.60069275, + "memory(GiB)": 111.15, + "step": 53425, + "train_speed(iter/s)": 0.447209 + }, + { + "acc": 0.65008459, + "epoch": 1.3554033485540335, + "grad_norm": 5.53125, + "learning_rate": 2.5810712373880253e-06, + "loss": 1.6520813, + "memory(GiB)": 111.15, + "step": 53430, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.66005378, + "epoch": 1.3555301877219685, + "grad_norm": 4.59375, + "learning_rate": 2.5801535493430215e-06, + "loss": 1.61153336, + "memory(GiB)": 111.15, + "step": 53435, + "train_speed(iter/s)": 0.447253 + }, + { + "acc": 0.66663046, + "epoch": 1.3556570268899035, + "grad_norm": 5.25, + "learning_rate": 2.5792359677336685e-06, + "loss": 1.58611488, + "memory(GiB)": 111.15, + "step": 53440, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.65495749, + "epoch": 1.3557838660578387, + "grad_norm": 5.3125, + "learning_rate": 2.5783184926003237e-06, + "loss": 1.63121815, + "memory(GiB)": 111.15, + "step": 53445, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.64005728, + "epoch": 1.3559107052257737, + "grad_norm": 4.96875, + "learning_rate": 2.5774011239833473e-06, + "loss": 1.65347252, + "memory(GiB)": 111.15, + "step": 53450, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.65646095, + "epoch": 1.3560375443937087, + "grad_norm": 6.3125, + "learning_rate": 2.5764838619230843e-06, + "loss": 1.6189085, + "memory(GiB)": 111.15, + "step": 53455, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.66733546, + "epoch": 1.356164383561644, + "grad_norm": 5.34375, + "learning_rate": 2.575566706459879e-06, + "loss": 1.56068649, + "memory(GiB)": 111.15, + "step": 53460, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.64607682, + "epoch": 1.356291222729579, + "grad_norm": 5.5625, + "learning_rate": 2.574649657634076e-06, + "loss": 1.66322021, + "memory(GiB)": 111.15, + "step": 53465, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.65745649, + "epoch": 1.356418061897514, + "grad_norm": 6.4375, + "learning_rate": 2.5737327154860116e-06, + "loss": 1.65252609, + "memory(GiB)": 111.15, + "step": 53470, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.6363029, + "epoch": 1.356544901065449, + "grad_norm": 5.34375, + "learning_rate": 2.572815880056011e-06, + "loss": 1.68347244, + "memory(GiB)": 111.15, + "step": 53475, + "train_speed(iter/s)": 0.447431 + }, + { + "acc": 0.66196041, + "epoch": 1.356671740233384, + "grad_norm": 5.625, + "learning_rate": 2.571899151384406e-06, + "loss": 1.52347069, + "memory(GiB)": 111.15, + "step": 53480, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.65426426, + "epoch": 1.356798579401319, + "grad_norm": 5.03125, + "learning_rate": 2.5709825295115178e-06, + "loss": 1.65075359, + "memory(GiB)": 111.15, + "step": 53485, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.63336825, + "epoch": 1.3569254185692543, + "grad_norm": 4.96875, + "learning_rate": 2.5700660144776647e-06, + "loss": 1.6345993, + "memory(GiB)": 111.15, + "step": 53490, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.64962382, + "epoch": 1.3570522577371893, + "grad_norm": 5.53125, + "learning_rate": 2.5691496063231527e-06, + "loss": 1.62113476, + "memory(GiB)": 111.15, + "step": 53495, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.65524778, + "epoch": 1.3571790969051243, + "grad_norm": 5.46875, + "learning_rate": 2.568233305088296e-06, + "loss": 1.5629961, + "memory(GiB)": 111.15, + "step": 53500, + "train_speed(iter/s)": 0.447541 + }, + { + "epoch": 1.3571790969051243, + "eval_acc": 0.643370011910546, + "eval_loss": 1.6076256036758423, + "eval_runtime": 114.3182, + "eval_samples_per_second": 55.722, + "eval_steps_per_second": 27.861, + "step": 53500 + }, + { + "acc": 0.65749879, + "epoch": 1.3573059360730593, + "grad_norm": 5.15625, + "learning_rate": 2.5673171108133956e-06, + "loss": 1.56520119, + "memory(GiB)": 111.15, + "step": 53505, + "train_speed(iter/s)": 0.447118 + }, + { + "acc": 0.63625317, + "epoch": 1.3574327752409945, + "grad_norm": 4.84375, + "learning_rate": 2.5664010235387503e-06, + "loss": 1.63184319, + "memory(GiB)": 111.15, + "step": 53510, + "train_speed(iter/s)": 0.44714 + }, + { + "acc": 0.65131688, + "epoch": 1.3575596144089295, + "grad_norm": 6.71875, + "learning_rate": 2.565485043304653e-06, + "loss": 1.62371159, + "memory(GiB)": 111.15, + "step": 53515, + "train_speed(iter/s)": 0.447162 + }, + { + "acc": 0.63942871, + "epoch": 1.3576864535768645, + "grad_norm": 5.34375, + "learning_rate": 2.564569170151392e-06, + "loss": 1.64945965, + "memory(GiB)": 111.15, + "step": 53520, + "train_speed(iter/s)": 0.447184 + }, + { + "acc": 0.64950275, + "epoch": 1.3578132927447997, + "grad_norm": 4.96875, + "learning_rate": 2.5636534041192534e-06, + "loss": 1.55946865, + "memory(GiB)": 111.15, + "step": 53525, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.65202751, + "epoch": 1.3579401319127347, + "grad_norm": 4.96875, + "learning_rate": 2.5627377452485153e-06, + "loss": 1.59545994, + "memory(GiB)": 111.15, + "step": 53530, + "train_speed(iter/s)": 0.447229 + }, + { + "acc": 0.65634274, + "epoch": 1.3580669710806696, + "grad_norm": 4.5625, + "learning_rate": 2.561822193579453e-06, + "loss": 1.54839096, + "memory(GiB)": 111.15, + "step": 53535, + "train_speed(iter/s)": 0.447251 + }, + { + "acc": 0.66794696, + "epoch": 1.3581938102486046, + "grad_norm": 5.46875, + "learning_rate": 2.560906749152335e-06, + "loss": 1.62551422, + "memory(GiB)": 111.15, + "step": 53540, + "train_speed(iter/s)": 0.447273 + }, + { + "acc": 0.64571419, + "epoch": 1.3583206494165398, + "grad_norm": 5.65625, + "learning_rate": 2.55999141200743e-06, + "loss": 1.61927948, + "memory(GiB)": 111.15, + "step": 53545, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.65883193, + "epoch": 1.3584474885844748, + "grad_norm": 4.65625, + "learning_rate": 2.5590761821849954e-06, + "loss": 1.5165554, + "memory(GiB)": 111.15, + "step": 53550, + "train_speed(iter/s)": 0.447317 + }, + { + "acc": 0.66283417, + "epoch": 1.35857432775241, + "grad_norm": 5.71875, + "learning_rate": 2.5581610597252883e-06, + "loss": 1.60564079, + "memory(GiB)": 111.15, + "step": 53555, + "train_speed(iter/s)": 0.447339 + }, + { + "acc": 0.64904094, + "epoch": 1.358701166920345, + "grad_norm": 5.0, + "learning_rate": 2.5572460446685593e-06, + "loss": 1.57299128, + "memory(GiB)": 111.15, + "step": 53560, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.64697952, + "epoch": 1.35882800608828, + "grad_norm": 5.21875, + "learning_rate": 2.5563311370550535e-06, + "loss": 1.63833733, + "memory(GiB)": 111.15, + "step": 53565, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.64462385, + "epoch": 1.358954845256215, + "grad_norm": 5.53125, + "learning_rate": 2.5554163369250194e-06, + "loss": 1.67696571, + "memory(GiB)": 111.15, + "step": 53570, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.66370239, + "epoch": 1.3590816844241502, + "grad_norm": 5.71875, + "learning_rate": 2.5545016443186867e-06, + "loss": 1.56147881, + "memory(GiB)": 111.15, + "step": 53575, + "train_speed(iter/s)": 0.447427 + }, + { + "acc": 0.64815383, + "epoch": 1.3592085235920852, + "grad_norm": 5.875, + "learning_rate": 2.55358705927629e-06, + "loss": 1.61112595, + "memory(GiB)": 111.15, + "step": 53580, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.64872837, + "epoch": 1.3593353627600204, + "grad_norm": 4.9375, + "learning_rate": 2.552672581838055e-06, + "loss": 1.62052326, + "memory(GiB)": 111.15, + "step": 53585, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.67622762, + "epoch": 1.3594622019279554, + "grad_norm": 6.03125, + "learning_rate": 2.5517582120442095e-06, + "loss": 1.57533073, + "memory(GiB)": 111.15, + "step": 53590, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.63873091, + "epoch": 1.3595890410958904, + "grad_norm": 5.1875, + "learning_rate": 2.5508439499349675e-06, + "loss": 1.63436165, + "memory(GiB)": 111.15, + "step": 53595, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.63463273, + "epoch": 1.3597158802638254, + "grad_norm": 5.03125, + "learning_rate": 2.549929795550541e-06, + "loss": 1.65779533, + "memory(GiB)": 111.15, + "step": 53600, + "train_speed(iter/s)": 0.447538 + }, + { + "epoch": 1.3597158802638254, + "eval_acc": 0.6433980023202797, + "eval_loss": 1.6076401472091675, + "eval_runtime": 113.2922, + "eval_samples_per_second": 56.226, + "eval_steps_per_second": 28.113, + "step": 53600 + }, + { + "acc": 0.66123066, + "epoch": 1.3598427194317606, + "grad_norm": 6.8125, + "learning_rate": 2.549015748931143e-06, + "loss": 1.58687363, + "memory(GiB)": 111.15, + "step": 53605, + "train_speed(iter/s)": 0.44712 + }, + { + "acc": 0.65715103, + "epoch": 1.3599695585996956, + "grad_norm": 5.625, + "learning_rate": 2.5481018101169763e-06, + "loss": 1.63794823, + "memory(GiB)": 111.15, + "step": 53610, + "train_speed(iter/s)": 0.447142 + }, + { + "acc": 0.65161114, + "epoch": 1.3600963977676306, + "grad_norm": 5.5625, + "learning_rate": 2.547187979148238e-06, + "loss": 1.5825819, + "memory(GiB)": 111.15, + "step": 53615, + "train_speed(iter/s)": 0.447164 + }, + { + "acc": 0.65964866, + "epoch": 1.3602232369355658, + "grad_norm": 4.59375, + "learning_rate": 2.546274256065121e-06, + "loss": 1.61942024, + "memory(GiB)": 111.15, + "step": 53620, + "train_speed(iter/s)": 0.447186 + }, + { + "acc": 0.64865513, + "epoch": 1.3603500761035008, + "grad_norm": 5.0625, + "learning_rate": 2.545360640907819e-06, + "loss": 1.61151295, + "memory(GiB)": 111.15, + "step": 53625, + "train_speed(iter/s)": 0.447208 + }, + { + "acc": 0.6361639, + "epoch": 1.3604769152714358, + "grad_norm": 5.25, + "learning_rate": 2.544447133716518e-06, + "loss": 1.58780022, + "memory(GiB)": 111.15, + "step": 53630, + "train_speed(iter/s)": 0.44723 + }, + { + "acc": 0.63074188, + "epoch": 1.3606037544393708, + "grad_norm": 5.0, + "learning_rate": 2.5435337345313904e-06, + "loss": 1.65205383, + "memory(GiB)": 111.15, + "step": 53635, + "train_speed(iter/s)": 0.447252 + }, + { + "acc": 0.63920035, + "epoch": 1.360730593607306, + "grad_norm": 5.71875, + "learning_rate": 2.5426204433926194e-06, + "loss": 1.62534332, + "memory(GiB)": 111.15, + "step": 53640, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.64779468, + "epoch": 1.360857432775241, + "grad_norm": 4.90625, + "learning_rate": 2.541707260340372e-06, + "loss": 1.66536655, + "memory(GiB)": 111.15, + "step": 53645, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.65298939, + "epoch": 1.3609842719431762, + "grad_norm": 4.65625, + "learning_rate": 2.5407941854148156e-06, + "loss": 1.63516502, + "memory(GiB)": 111.15, + "step": 53650, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.63578048, + "epoch": 1.3611111111111112, + "grad_norm": 5.1875, + "learning_rate": 2.5398812186561095e-06, + "loss": 1.70741196, + "memory(GiB)": 111.15, + "step": 53655, + "train_speed(iter/s)": 0.447341 + }, + { + "acc": 0.63648758, + "epoch": 1.3612379502790461, + "grad_norm": 4.5625, + "learning_rate": 2.5389683601044114e-06, + "loss": 1.63265629, + "memory(GiB)": 111.15, + "step": 53660, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.65695601, + "epoch": 1.3613647894469811, + "grad_norm": 6.3125, + "learning_rate": 2.538055609799873e-06, + "loss": 1.60659142, + "memory(GiB)": 111.15, + "step": 53665, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.64847775, + "epoch": 1.3614916286149163, + "grad_norm": 5.375, + "learning_rate": 2.5371429677826397e-06, + "loss": 1.62755318, + "memory(GiB)": 111.15, + "step": 53670, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.67299066, + "epoch": 1.3616184677828513, + "grad_norm": 5.96875, + "learning_rate": 2.5362304340928556e-06, + "loss": 1.55818338, + "memory(GiB)": 111.15, + "step": 53675, + "train_speed(iter/s)": 0.44743 + }, + { + "acc": 0.64032192, + "epoch": 1.3617453069507863, + "grad_norm": 5.65625, + "learning_rate": 2.535318008770656e-06, + "loss": 1.65353355, + "memory(GiB)": 111.15, + "step": 53680, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.65938268, + "epoch": 1.3618721461187215, + "grad_norm": 5.3125, + "learning_rate": 2.534405691856175e-06, + "loss": 1.58861389, + "memory(GiB)": 111.15, + "step": 53685, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.65935035, + "epoch": 1.3619989852866565, + "grad_norm": 4.71875, + "learning_rate": 2.5334934833895396e-06, + "loss": 1.59112034, + "memory(GiB)": 111.15, + "step": 53690, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.64794111, + "epoch": 1.3621258244545915, + "grad_norm": 4.875, + "learning_rate": 2.5325813834108724e-06, + "loss": 1.65006695, + "memory(GiB)": 111.15, + "step": 53695, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.66493464, + "epoch": 1.3622526636225265, + "grad_norm": 5.09375, + "learning_rate": 2.531669391960293e-06, + "loss": 1.61649971, + "memory(GiB)": 111.15, + "step": 53700, + "train_speed(iter/s)": 0.44754 + }, + { + "epoch": 1.3622526636225265, + "eval_acc": 0.6432931427256058, + "eval_loss": 1.6076061725616455, + "eval_runtime": 112.5629, + "eval_samples_per_second": 56.591, + "eval_steps_per_second": 28.295, + "step": 53700 + }, + { + "acc": 0.65131669, + "epoch": 1.3623795027904617, + "grad_norm": 4.71875, + "learning_rate": 2.5307575090779125e-06, + "loss": 1.68075237, + "memory(GiB)": 111.15, + "step": 53705, + "train_speed(iter/s)": 0.447125 + }, + { + "acc": 0.64756422, + "epoch": 1.3625063419583967, + "grad_norm": 5.4375, + "learning_rate": 2.529845734803844e-06, + "loss": 1.62017136, + "memory(GiB)": 111.15, + "step": 53710, + "train_speed(iter/s)": 0.447147 + }, + { + "acc": 0.65245414, + "epoch": 1.362633181126332, + "grad_norm": 5.6875, + "learning_rate": 2.5289340691781872e-06, + "loss": 1.65525627, + "memory(GiB)": 111.15, + "step": 53715, + "train_speed(iter/s)": 0.44717 + }, + { + "acc": 0.64950771, + "epoch": 1.362760020294267, + "grad_norm": 5.84375, + "learning_rate": 2.528022512241042e-06, + "loss": 1.57923899, + "memory(GiB)": 111.15, + "step": 53720, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.65749311, + "epoch": 1.362886859462202, + "grad_norm": 5.1875, + "learning_rate": 2.5271110640325013e-06, + "loss": 1.60904694, + "memory(GiB)": 111.15, + "step": 53725, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.65147429, + "epoch": 1.3630136986301369, + "grad_norm": 4.90625, + "learning_rate": 2.5261997245926612e-06, + "loss": 1.59437771, + "memory(GiB)": 111.15, + "step": 53730, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.64182878, + "epoch": 1.363140537798072, + "grad_norm": 5.09375, + "learning_rate": 2.5252884939615995e-06, + "loss": 1.61992874, + "memory(GiB)": 111.15, + "step": 53735, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.64842548, + "epoch": 1.363267376966007, + "grad_norm": 4.5625, + "learning_rate": 2.5243773721793973e-06, + "loss": 1.60432358, + "memory(GiB)": 111.15, + "step": 53740, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.6591629, + "epoch": 1.3633942161339423, + "grad_norm": 5.46875, + "learning_rate": 2.5234663592861325e-06, + "loss": 1.59207602, + "memory(GiB)": 111.15, + "step": 53745, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.65095701, + "epoch": 1.3635210553018773, + "grad_norm": 5.5625, + "learning_rate": 2.522555455321876e-06, + "loss": 1.60262756, + "memory(GiB)": 111.15, + "step": 53750, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.65170269, + "epoch": 1.3636478944698123, + "grad_norm": 5.875, + "learning_rate": 2.52164466032669e-06, + "loss": 1.58231144, + "memory(GiB)": 111.15, + "step": 53755, + "train_speed(iter/s)": 0.447346 + }, + { + "acc": 0.66010265, + "epoch": 1.3637747336377473, + "grad_norm": 4.75, + "learning_rate": 2.5207339743406344e-06, + "loss": 1.62863102, + "memory(GiB)": 111.15, + "step": 53760, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.63215275, + "epoch": 1.3639015728056825, + "grad_norm": 4.21875, + "learning_rate": 2.5198233974037705e-06, + "loss": 1.63572845, + "memory(GiB)": 111.15, + "step": 53765, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.64354935, + "epoch": 1.3640284119736175, + "grad_norm": 6.0625, + "learning_rate": 2.5189129295561486e-06, + "loss": 1.68065243, + "memory(GiB)": 111.15, + "step": 53770, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.64947109, + "epoch": 1.3641552511415524, + "grad_norm": 4.78125, + "learning_rate": 2.518002570837809e-06, + "loss": 1.64671555, + "memory(GiB)": 111.15, + "step": 53775, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.64212036, + "epoch": 1.3642820903094877, + "grad_norm": 5.96875, + "learning_rate": 2.5170923212887997e-06, + "loss": 1.71752129, + "memory(GiB)": 111.15, + "step": 53780, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.67178898, + "epoch": 1.3644089294774226, + "grad_norm": 5.46875, + "learning_rate": 2.5161821809491554e-06, + "loss": 1.58943415, + "memory(GiB)": 111.15, + "step": 53785, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.65954351, + "epoch": 1.3645357686453576, + "grad_norm": 5.09375, + "learning_rate": 2.5152721498589104e-06, + "loss": 1.61051807, + "memory(GiB)": 111.15, + "step": 53790, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.63769784, + "epoch": 1.3646626078132926, + "grad_norm": 6.78125, + "learning_rate": 2.514362228058086e-06, + "loss": 1.62113857, + "memory(GiB)": 111.15, + "step": 53795, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64498138, + "epoch": 1.3647894469812278, + "grad_norm": 5.71875, + "learning_rate": 2.51345241558671e-06, + "loss": 1.6742878, + "memory(GiB)": 111.15, + "step": 53800, + "train_speed(iter/s)": 0.447545 + }, + { + "epoch": 1.3647894469812278, + "eval_acc": 0.6434702760648158, + "eval_loss": 1.6076200008392334, + "eval_runtime": 113.8288, + "eval_samples_per_second": 55.961, + "eval_steps_per_second": 27.981, + "step": 53800 + }, + { + "acc": 0.63967981, + "epoch": 1.3649162861491628, + "grad_norm": 5.6875, + "learning_rate": 2.5125427124847985e-06, + "loss": 1.68268661, + "memory(GiB)": 111.15, + "step": 53805, + "train_speed(iter/s)": 0.447127 + }, + { + "acc": 0.65533047, + "epoch": 1.365043125317098, + "grad_norm": 5.5, + "learning_rate": 2.5116331187923645e-06, + "loss": 1.58075695, + "memory(GiB)": 111.15, + "step": 53810, + "train_speed(iter/s)": 0.447149 + }, + { + "acc": 0.65239658, + "epoch": 1.365169964485033, + "grad_norm": 5.875, + "learning_rate": 2.510723634549415e-06, + "loss": 1.63485966, + "memory(GiB)": 111.15, + "step": 53815, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.65630302, + "epoch": 1.365296803652968, + "grad_norm": 6.375, + "learning_rate": 2.509814259795954e-06, + "loss": 1.63906174, + "memory(GiB)": 111.15, + "step": 53820, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.66072731, + "epoch": 1.365423642820903, + "grad_norm": 5.65625, + "learning_rate": 2.50890499457198e-06, + "loss": 1.59450607, + "memory(GiB)": 111.15, + "step": 53825, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.66513567, + "epoch": 1.3655504819888382, + "grad_norm": 5.0625, + "learning_rate": 2.5079958389174865e-06, + "loss": 1.53741465, + "memory(GiB)": 111.15, + "step": 53830, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.63600407, + "epoch": 1.3656773211567732, + "grad_norm": 5.5625, + "learning_rate": 2.5070867928724618e-06, + "loss": 1.73357239, + "memory(GiB)": 111.15, + "step": 53835, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.64953756, + "epoch": 1.3658041603247082, + "grad_norm": 4.9375, + "learning_rate": 2.50617785647689e-06, + "loss": 1.5646286, + "memory(GiB)": 111.15, + "step": 53840, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.64405594, + "epoch": 1.3659309994926434, + "grad_norm": 5.75, + "learning_rate": 2.5052690297707506e-06, + "loss": 1.58428173, + "memory(GiB)": 111.15, + "step": 53845, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.65732603, + "epoch": 1.3660578386605784, + "grad_norm": 5.4375, + "learning_rate": 2.5043603127940164e-06, + "loss": 1.52856522, + "memory(GiB)": 111.15, + "step": 53850, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.64463425, + "epoch": 1.3661846778285134, + "grad_norm": 5.53125, + "learning_rate": 2.503451705586659e-06, + "loss": 1.62363186, + "memory(GiB)": 111.15, + "step": 53855, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.65845661, + "epoch": 1.3663115169964484, + "grad_norm": 4.9375, + "learning_rate": 2.5025432081886412e-06, + "loss": 1.57045841, + "memory(GiB)": 111.15, + "step": 53860, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.65020275, + "epoch": 1.3664383561643836, + "grad_norm": 5.8125, + "learning_rate": 2.5016348206399215e-06, + "loss": 1.59868774, + "memory(GiB)": 111.15, + "step": 53865, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.64941826, + "epoch": 1.3665651953323186, + "grad_norm": 6.4375, + "learning_rate": 2.500726542980461e-06, + "loss": 1.67974243, + "memory(GiB)": 111.15, + "step": 53870, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.65533905, + "epoch": 1.3666920345002538, + "grad_norm": 4.8125, + "learning_rate": 2.499818375250204e-06, + "loss": 1.61846981, + "memory(GiB)": 111.15, + "step": 53875, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.65063362, + "epoch": 1.3668188736681888, + "grad_norm": 7.0625, + "learning_rate": 2.4989103174890946e-06, + "loss": 1.64481182, + "memory(GiB)": 111.15, + "step": 53880, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.65091419, + "epoch": 1.3669457128361238, + "grad_norm": 4.78125, + "learning_rate": 2.498002369737078e-06, + "loss": 1.66141624, + "memory(GiB)": 111.15, + "step": 53885, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.65422978, + "epoch": 1.3670725520040587, + "grad_norm": 6.5, + "learning_rate": 2.49709453203409e-06, + "loss": 1.60266876, + "memory(GiB)": 111.15, + "step": 53890, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.64642415, + "epoch": 1.367199391171994, + "grad_norm": 4.6875, + "learning_rate": 2.496186804420057e-06, + "loss": 1.66190796, + "memory(GiB)": 111.15, + "step": 53895, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.6501009, + "epoch": 1.367326230339929, + "grad_norm": 4.5, + "learning_rate": 2.4952791869349056e-06, + "loss": 1.62598152, + "memory(GiB)": 111.15, + "step": 53900, + "train_speed(iter/s)": 0.447545 + }, + { + "epoch": 1.367326230339929, + "eval_acc": 0.6434126241761107, + "eval_loss": 1.607625961303711, + "eval_runtime": 114.0855, + "eval_samples_per_second": 55.835, + "eval_steps_per_second": 27.918, + "step": 53900 + }, + { + "acc": 0.66053495, + "epoch": 1.3674530695078642, + "grad_norm": 7.84375, + "learning_rate": 2.4943716796185603e-06, + "loss": 1.6291832, + "memory(GiB)": 111.15, + "step": 53905, + "train_speed(iter/s)": 0.447126 + }, + { + "acc": 0.65272274, + "epoch": 1.3675799086757991, + "grad_norm": 5.8125, + "learning_rate": 2.493464282510937e-06, + "loss": 1.66543503, + "memory(GiB)": 111.15, + "step": 53910, + "train_speed(iter/s)": 0.447148 + }, + { + "acc": 0.67182198, + "epoch": 1.3677067478437341, + "grad_norm": 5.25, + "learning_rate": 2.4925569956519414e-06, + "loss": 1.58488197, + "memory(GiB)": 111.15, + "step": 53915, + "train_speed(iter/s)": 0.44717 + }, + { + "acc": 0.64124699, + "epoch": 1.3678335870116691, + "grad_norm": 5.28125, + "learning_rate": 2.491649819081486e-06, + "loss": 1.69223747, + "memory(GiB)": 111.15, + "step": 53920, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.66097775, + "epoch": 1.3679604261796043, + "grad_norm": 5.0625, + "learning_rate": 2.490742752839471e-06, + "loss": 1.64284267, + "memory(GiB)": 111.15, + "step": 53925, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.65012608, + "epoch": 1.3680872653475393, + "grad_norm": 5.8125, + "learning_rate": 2.4898357969657943e-06, + "loss": 1.59399204, + "memory(GiB)": 111.15, + "step": 53930, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.66440682, + "epoch": 1.3682141045154743, + "grad_norm": 5.40625, + "learning_rate": 2.4889289515003425e-06, + "loss": 1.59462557, + "memory(GiB)": 111.15, + "step": 53935, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.66493983, + "epoch": 1.3683409436834095, + "grad_norm": 5.625, + "learning_rate": 2.4880222164830085e-06, + "loss": 1.58884668, + "memory(GiB)": 111.15, + "step": 53940, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.65550346, + "epoch": 1.3684677828513445, + "grad_norm": 4.71875, + "learning_rate": 2.4871155919536725e-06, + "loss": 1.61865578, + "memory(GiB)": 111.15, + "step": 53945, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.66767015, + "epoch": 1.3685946220192795, + "grad_norm": 6.875, + "learning_rate": 2.486209077952212e-06, + "loss": 1.57417145, + "memory(GiB)": 111.15, + "step": 53950, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.66276217, + "epoch": 1.3687214611872145, + "grad_norm": 5.96875, + "learning_rate": 2.4853026745185e-06, + "loss": 1.60117989, + "memory(GiB)": 111.15, + "step": 53955, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.66225986, + "epoch": 1.3688483003551497, + "grad_norm": 4.90625, + "learning_rate": 2.4843963816924035e-06, + "loss": 1.58685398, + "memory(GiB)": 111.15, + "step": 53960, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.64460812, + "epoch": 1.3689751395230847, + "grad_norm": 4.9375, + "learning_rate": 2.483490199513785e-06, + "loss": 1.57389116, + "memory(GiB)": 111.15, + "step": 53965, + "train_speed(iter/s)": 0.447392 + }, + { + "acc": 0.6641758, + "epoch": 1.36910197869102, + "grad_norm": 6.34375, + "learning_rate": 2.4825841280225033e-06, + "loss": 1.62485905, + "memory(GiB)": 111.15, + "step": 53970, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.64341431, + "epoch": 1.369228817858955, + "grad_norm": 5.46875, + "learning_rate": 2.4816781672584107e-06, + "loss": 1.63893394, + "memory(GiB)": 111.15, + "step": 53975, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.63590364, + "epoch": 1.3693556570268899, + "grad_norm": 5.09375, + "learning_rate": 2.480772317261356e-06, + "loss": 1.64935303, + "memory(GiB)": 111.15, + "step": 53980, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.65358453, + "epoch": 1.3694824961948249, + "grad_norm": 5.375, + "learning_rate": 2.479866578071183e-06, + "loss": 1.64516487, + "memory(GiB)": 111.15, + "step": 53985, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.64769468, + "epoch": 1.36960933536276, + "grad_norm": 6.53125, + "learning_rate": 2.4789609497277284e-06, + "loss": 1.62047577, + "memory(GiB)": 111.15, + "step": 53990, + "train_speed(iter/s)": 0.447502 + }, + { + "acc": 0.65165801, + "epoch": 1.369736174530695, + "grad_norm": 6.15625, + "learning_rate": 2.478055432270828e-06, + "loss": 1.63498802, + "memory(GiB)": 111.15, + "step": 53995, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.68386297, + "epoch": 1.36986301369863, + "grad_norm": 4.3125, + "learning_rate": 2.4771500257403086e-06, + "loss": 1.45399466, + "memory(GiB)": 111.15, + "step": 54000, + "train_speed(iter/s)": 0.447546 + }, + { + "epoch": 1.36986301369863, + "eval_acc": 0.643359985495119, + "eval_loss": 1.6075286865234375, + "eval_runtime": 114.0531, + "eval_samples_per_second": 55.851, + "eval_steps_per_second": 27.926, + "step": 54000 + }, + { + "acc": 0.65283713, + "epoch": 1.3699898528665653, + "grad_norm": 4.71875, + "learning_rate": 2.476244730175993e-06, + "loss": 1.63605785, + "memory(GiB)": 111.15, + "step": 54005, + "train_speed(iter/s)": 0.447129 + }, + { + "acc": 0.65631208, + "epoch": 1.3701166920345003, + "grad_norm": 5.53125, + "learning_rate": 2.4753395456177056e-06, + "loss": 1.62325039, + "memory(GiB)": 111.15, + "step": 54010, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.64412069, + "epoch": 1.3702435312024352, + "grad_norm": 5.0, + "learning_rate": 2.474434472105255e-06, + "loss": 1.65310822, + "memory(GiB)": 111.15, + "step": 54015, + "train_speed(iter/s)": 0.447172 + }, + { + "acc": 0.64262562, + "epoch": 1.3703703703703702, + "grad_norm": 5.625, + "learning_rate": 2.473529509678452e-06, + "loss": 1.59661636, + "memory(GiB)": 111.15, + "step": 54020, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.64984269, + "epoch": 1.3704972095383054, + "grad_norm": 5.5625, + "learning_rate": 2.4726246583770996e-06, + "loss": 1.67539902, + "memory(GiB)": 111.15, + "step": 54025, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.65169973, + "epoch": 1.3706240487062404, + "grad_norm": 4.625, + "learning_rate": 2.4717199182410025e-06, + "loss": 1.55737724, + "memory(GiB)": 111.15, + "step": 54030, + "train_speed(iter/s)": 0.447238 + }, + { + "acc": 0.66279774, + "epoch": 1.3707508878741756, + "grad_norm": 5.6875, + "learning_rate": 2.4708152893099493e-06, + "loss": 1.57991695, + "memory(GiB)": 111.15, + "step": 54035, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.66502361, + "epoch": 1.3708777270421106, + "grad_norm": 6.28125, + "learning_rate": 2.4699107716237293e-06, + "loss": 1.58999081, + "memory(GiB)": 111.15, + "step": 54040, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.65244412, + "epoch": 1.3710045662100456, + "grad_norm": 4.78125, + "learning_rate": 2.469006365222132e-06, + "loss": 1.59948092, + "memory(GiB)": 111.15, + "step": 54045, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.64311647, + "epoch": 1.3711314053779806, + "grad_norm": 5.40625, + "learning_rate": 2.4681020701449365e-06, + "loss": 1.64587307, + "memory(GiB)": 111.15, + "step": 54050, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.64383817, + "epoch": 1.3712582445459158, + "grad_norm": 5.375, + "learning_rate": 2.4671978864319123e-06, + "loss": 1.61419506, + "memory(GiB)": 111.15, + "step": 54055, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.64757047, + "epoch": 1.3713850837138508, + "grad_norm": 5.90625, + "learning_rate": 2.466293814122835e-06, + "loss": 1.65925941, + "memory(GiB)": 111.15, + "step": 54060, + "train_speed(iter/s)": 0.44737 + }, + { + "acc": 0.67053723, + "epoch": 1.371511922881786, + "grad_norm": 6.78125, + "learning_rate": 2.4653898532574684e-06, + "loss": 1.55502205, + "memory(GiB)": 111.15, + "step": 54065, + "train_speed(iter/s)": 0.447392 + }, + { + "acc": 0.66027589, + "epoch": 1.371638762049721, + "grad_norm": 5.34375, + "learning_rate": 2.4644860038755737e-06, + "loss": 1.60154572, + "memory(GiB)": 111.15, + "step": 54070, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.64132648, + "epoch": 1.371765601217656, + "grad_norm": 5.4375, + "learning_rate": 2.4635822660169007e-06, + "loss": 1.66392365, + "memory(GiB)": 111.15, + "step": 54075, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.64989734, + "epoch": 1.371892440385591, + "grad_norm": 6.21875, + "learning_rate": 2.4626786397212065e-06, + "loss": 1.67736282, + "memory(GiB)": 111.15, + "step": 54080, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.66153569, + "epoch": 1.3720192795535262, + "grad_norm": 7.0, + "learning_rate": 2.461775125028234e-06, + "loss": 1.62993965, + "memory(GiB)": 111.15, + "step": 54085, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.63310614, + "epoch": 1.3721461187214612, + "grad_norm": 5.875, + "learning_rate": 2.4608717219777236e-06, + "loss": 1.65197239, + "memory(GiB)": 111.15, + "step": 54090, + "train_speed(iter/s)": 0.447502 + }, + { + "acc": 0.65307555, + "epoch": 1.3722729578893962, + "grad_norm": 5.03125, + "learning_rate": 2.459968430609411e-06, + "loss": 1.61922131, + "memory(GiB)": 111.15, + "step": 54095, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.65075665, + "epoch": 1.3723997970573314, + "grad_norm": 5.21875, + "learning_rate": 2.459065250963028e-06, + "loss": 1.56514893, + "memory(GiB)": 111.15, + "step": 54100, + "train_speed(iter/s)": 0.447546 + }, + { + "epoch": 1.3723997970573314, + "eval_acc": 0.6434326770069646, + "eval_loss": 1.607506513595581, + "eval_runtime": 113.8065, + "eval_samples_per_second": 55.972, + "eval_steps_per_second": 27.986, + "step": 54100 + }, + { + "acc": 0.66506958, + "epoch": 1.3725266362252664, + "grad_norm": 5.65625, + "learning_rate": 2.458162183078299e-06, + "loss": 1.57335291, + "memory(GiB)": 111.15, + "step": 54105, + "train_speed(iter/s)": 0.44713 + }, + { + "acc": 0.67396069, + "epoch": 1.3726534753932014, + "grad_norm": 5.34375, + "learning_rate": 2.4572592269949464e-06, + "loss": 1.52641773, + "memory(GiB)": 111.15, + "step": 54110, + "train_speed(iter/s)": 0.447152 + }, + { + "acc": 0.64913712, + "epoch": 1.3727803145611364, + "grad_norm": 5.625, + "learning_rate": 2.4563563827526848e-06, + "loss": 1.62573986, + "memory(GiB)": 111.15, + "step": 54115, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.64341526, + "epoch": 1.3729071537290716, + "grad_norm": 5.6875, + "learning_rate": 2.455453650391226e-06, + "loss": 1.60747776, + "memory(GiB)": 111.15, + "step": 54120, + "train_speed(iter/s)": 0.447196 + }, + { + "acc": 0.66006231, + "epoch": 1.3730339928970066, + "grad_norm": 5.40625, + "learning_rate": 2.454551029950277e-06, + "loss": 1.61412067, + "memory(GiB)": 111.15, + "step": 54125, + "train_speed(iter/s)": 0.447218 + }, + { + "acc": 0.65793524, + "epoch": 1.3731608320649418, + "grad_norm": 5.125, + "learning_rate": 2.4536485214695377e-06, + "loss": 1.53926353, + "memory(GiB)": 111.15, + "step": 54130, + "train_speed(iter/s)": 0.447241 + }, + { + "acc": 0.65194378, + "epoch": 1.3732876712328768, + "grad_norm": 4.4375, + "learning_rate": 2.4527461249887054e-06, + "loss": 1.59197626, + "memory(GiB)": 111.15, + "step": 54135, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.64801779, + "epoch": 1.3734145104008117, + "grad_norm": 4.96875, + "learning_rate": 2.451843840547471e-06, + "loss": 1.61186676, + "memory(GiB)": 111.15, + "step": 54140, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.65309134, + "epoch": 1.3735413495687467, + "grad_norm": 5.25, + "learning_rate": 2.4509416681855193e-06, + "loss": 1.65496292, + "memory(GiB)": 111.15, + "step": 54145, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.65230007, + "epoch": 1.373668188736682, + "grad_norm": 6.0, + "learning_rate": 2.4500396079425377e-06, + "loss": 1.64339371, + "memory(GiB)": 111.15, + "step": 54150, + "train_speed(iter/s)": 0.447329 + }, + { + "acc": 0.6514914, + "epoch": 1.373795027904617, + "grad_norm": 6.3125, + "learning_rate": 2.4491376598581967e-06, + "loss": 1.63499184, + "memory(GiB)": 111.15, + "step": 54155, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.65102873, + "epoch": 1.373921867072552, + "grad_norm": 6.1875, + "learning_rate": 2.4482358239721704e-06, + "loss": 1.65550671, + "memory(GiB)": 111.15, + "step": 54160, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.65483875, + "epoch": 1.3740487062404871, + "grad_norm": 5.25, + "learning_rate": 2.4473341003241234e-06, + "loss": 1.62768974, + "memory(GiB)": 111.15, + "step": 54165, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.65897627, + "epoch": 1.3741755454084221, + "grad_norm": 5.4375, + "learning_rate": 2.446432488953724e-06, + "loss": 1.47019701, + "memory(GiB)": 111.15, + "step": 54170, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.64088774, + "epoch": 1.3743023845763571, + "grad_norm": 6.8125, + "learning_rate": 2.445530989900622e-06, + "loss": 1.61296196, + "memory(GiB)": 111.15, + "step": 54175, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.65318742, + "epoch": 1.374429223744292, + "grad_norm": 5.875, + "learning_rate": 2.4446296032044697e-06, + "loss": 1.65706158, + "memory(GiB)": 111.15, + "step": 54180, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.65252695, + "epoch": 1.3745560629122273, + "grad_norm": 5.5625, + "learning_rate": 2.443728328904919e-06, + "loss": 1.60684261, + "memory(GiB)": 111.15, + "step": 54185, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.63892503, + "epoch": 1.3746829020801623, + "grad_norm": 6.0625, + "learning_rate": 2.442827167041611e-06, + "loss": 1.65832977, + "memory(GiB)": 111.15, + "step": 54190, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.66288891, + "epoch": 1.3748097412480975, + "grad_norm": 5.78125, + "learning_rate": 2.441926117654179e-06, + "loss": 1.61495571, + "memory(GiB)": 111.15, + "step": 54195, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.64723091, + "epoch": 1.3749365804160325, + "grad_norm": 6.125, + "learning_rate": 2.4410251807822555e-06, + "loss": 1.68451042, + "memory(GiB)": 111.15, + "step": 54200, + "train_speed(iter/s)": 0.447549 + }, + { + "epoch": 1.3749365804160325, + "eval_acc": 0.6433737718163312, + "eval_loss": 1.6075576543807983, + "eval_runtime": 114.6949, + "eval_samples_per_second": 55.539, + "eval_steps_per_second": 27.769, + "step": 54200 + }, + { + "acc": 0.65749006, + "epoch": 1.3750634195839675, + "grad_norm": 5.21875, + "learning_rate": 2.4401243564654713e-06, + "loss": 1.59885588, + "memory(GiB)": 111.15, + "step": 54205, + "train_speed(iter/s)": 0.447131 + }, + { + "acc": 0.65639372, + "epoch": 1.3751902587519025, + "grad_norm": 5.40625, + "learning_rate": 2.4392236447434494e-06, + "loss": 1.58142643, + "memory(GiB)": 111.15, + "step": 54210, + "train_speed(iter/s)": 0.447153 + }, + { + "acc": 0.6563168, + "epoch": 1.3753170979198377, + "grad_norm": 5.09375, + "learning_rate": 2.4383230456558005e-06, + "loss": 1.55810089, + "memory(GiB)": 111.15, + "step": 54215, + "train_speed(iter/s)": 0.447175 + }, + { + "acc": 0.65408735, + "epoch": 1.3754439370877727, + "grad_norm": 5.8125, + "learning_rate": 2.437422559242143e-06, + "loss": 1.62198868, + "memory(GiB)": 111.15, + "step": 54220, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.65834684, + "epoch": 1.375570776255708, + "grad_norm": 5.15625, + "learning_rate": 2.4365221855420822e-06, + "loss": 1.63757401, + "memory(GiB)": 111.15, + "step": 54225, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.64717593, + "epoch": 1.3756976154236429, + "grad_norm": 5.3125, + "learning_rate": 2.435621924595221e-06, + "loss": 1.58079052, + "memory(GiB)": 111.15, + "step": 54230, + "train_speed(iter/s)": 0.447241 + }, + { + "acc": 0.65896749, + "epoch": 1.3758244545915779, + "grad_norm": 5.59375, + "learning_rate": 2.4347217764411567e-06, + "loss": 1.65701065, + "memory(GiB)": 111.15, + "step": 54235, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.66342549, + "epoch": 1.3759512937595129, + "grad_norm": 5.625, + "learning_rate": 2.433821741119482e-06, + "loss": 1.55835285, + "memory(GiB)": 111.15, + "step": 54240, + "train_speed(iter/s)": 0.447286 + }, + { + "acc": 0.6351398, + "epoch": 1.376078132927448, + "grad_norm": 5.4375, + "learning_rate": 2.432921818669784e-06, + "loss": 1.64769554, + "memory(GiB)": 111.15, + "step": 54245, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.66411724, + "epoch": 1.376204972095383, + "grad_norm": 5.125, + "learning_rate": 2.432022009131646e-06, + "loss": 1.57920542, + "memory(GiB)": 111.15, + "step": 54250, + "train_speed(iter/s)": 0.447329 + }, + { + "acc": 0.64469128, + "epoch": 1.376331811263318, + "grad_norm": 5.40625, + "learning_rate": 2.4311223125446447e-06, + "loss": 1.71905022, + "memory(GiB)": 111.15, + "step": 54255, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.64162774, + "epoch": 1.3764586504312533, + "grad_norm": 5.0, + "learning_rate": 2.4302227289483537e-06, + "loss": 1.7233799, + "memory(GiB)": 111.15, + "step": 54260, + "train_speed(iter/s)": 0.447372 + }, + { + "acc": 0.65267153, + "epoch": 1.3765854895991883, + "grad_norm": 5.75, + "learning_rate": 2.42932325838234e-06, + "loss": 1.65109024, + "memory(GiB)": 111.15, + "step": 54265, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.66144891, + "epoch": 1.3767123287671232, + "grad_norm": 5.1875, + "learning_rate": 2.4284239008861665e-06, + "loss": 1.61261978, + "memory(GiB)": 111.15, + "step": 54270, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.65323009, + "epoch": 1.3768391679350582, + "grad_norm": 5.90625, + "learning_rate": 2.4275246564993917e-06, + "loss": 1.64087257, + "memory(GiB)": 111.15, + "step": 54275, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.65506334, + "epoch": 1.3769660071029934, + "grad_norm": 5.5, + "learning_rate": 2.426625525261567e-06, + "loss": 1.5868, + "memory(GiB)": 111.15, + "step": 54280, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.65450735, + "epoch": 1.3770928462709284, + "grad_norm": 4.125, + "learning_rate": 2.425726507212242e-06, + "loss": 1.59628048, + "memory(GiB)": 111.15, + "step": 54285, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.65414705, + "epoch": 1.3772196854388636, + "grad_norm": 5.125, + "learning_rate": 2.424827602390958e-06, + "loss": 1.61867752, + "memory(GiB)": 111.15, + "step": 54290, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.64450226, + "epoch": 1.3773465246067986, + "grad_norm": 4.71875, + "learning_rate": 2.4239288108372534e-06, + "loss": 1.63585472, + "memory(GiB)": 111.15, + "step": 54295, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.65800261, + "epoch": 1.3774733637747336, + "grad_norm": 6.3125, + "learning_rate": 2.4230301325906606e-06, + "loss": 1.60416584, + "memory(GiB)": 111.15, + "step": 54300, + "train_speed(iter/s)": 0.447546 + }, + { + "epoch": 1.3774733637747336, + "eval_acc": 0.6433837982317581, + "eval_loss": 1.607613205909729, + "eval_runtime": 113.5121, + "eval_samples_per_second": 56.117, + "eval_steps_per_second": 28.059, + "step": 54300 + }, + { + "acc": 0.6450376, + "epoch": 1.3776002029426686, + "grad_norm": 4.875, + "learning_rate": 2.4221315676907066e-06, + "loss": 1.67018299, + "memory(GiB)": 111.15, + "step": 54305, + "train_speed(iter/s)": 0.447133 + }, + { + "acc": 0.64392834, + "epoch": 1.3777270421106038, + "grad_norm": 5.75, + "learning_rate": 2.4212331161769194e-06, + "loss": 1.70141411, + "memory(GiB)": 111.15, + "step": 54310, + "train_speed(iter/s)": 0.447154 + }, + { + "acc": 0.64372301, + "epoch": 1.3778538812785388, + "grad_norm": 4.96875, + "learning_rate": 2.420334778088811e-06, + "loss": 1.62558765, + "memory(GiB)": 111.15, + "step": 54315, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.65850372, + "epoch": 1.3779807204464738, + "grad_norm": 4.875, + "learning_rate": 2.4194365534658944e-06, + "loss": 1.62272205, + "memory(GiB)": 111.15, + "step": 54320, + "train_speed(iter/s)": 0.447198 + }, + { + "acc": 0.63821692, + "epoch": 1.378107559614409, + "grad_norm": 5.1875, + "learning_rate": 2.4185384423476817e-06, + "loss": 1.6870163, + "memory(GiB)": 111.15, + "step": 54325, + "train_speed(iter/s)": 0.44722 + }, + { + "acc": 0.63756447, + "epoch": 1.378234398782344, + "grad_norm": 6.0, + "learning_rate": 2.4176404447736758e-06, + "loss": 1.67064953, + "memory(GiB)": 111.15, + "step": 54330, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.64170799, + "epoch": 1.378361237950279, + "grad_norm": 6.40625, + "learning_rate": 2.41674256078337e-06, + "loss": 1.62566433, + "memory(GiB)": 111.15, + "step": 54335, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.63791261, + "epoch": 1.378488077118214, + "grad_norm": 5.40625, + "learning_rate": 2.4158447904162585e-06, + "loss": 1.71883507, + "memory(GiB)": 111.15, + "step": 54340, + "train_speed(iter/s)": 0.447286 + }, + { + "acc": 0.64991484, + "epoch": 1.3786149162861492, + "grad_norm": 5.65625, + "learning_rate": 2.414947133711832e-06, + "loss": 1.64734802, + "memory(GiB)": 111.15, + "step": 54345, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.64000006, + "epoch": 1.3787417554540842, + "grad_norm": 4.875, + "learning_rate": 2.414049590709574e-06, + "loss": 1.62274704, + "memory(GiB)": 111.15, + "step": 54350, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.63573432, + "epoch": 1.3788685946220194, + "grad_norm": 5.3125, + "learning_rate": 2.4131521614489567e-06, + "loss": 1.67726822, + "memory(GiB)": 111.15, + "step": 54355, + "train_speed(iter/s)": 0.447351 + }, + { + "acc": 0.66043658, + "epoch": 1.3789954337899544, + "grad_norm": 4.6875, + "learning_rate": 2.412254845969459e-06, + "loss": 1.54086018, + "memory(GiB)": 111.15, + "step": 54360, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.66526546, + "epoch": 1.3791222729578894, + "grad_norm": 4.625, + "learning_rate": 2.4113576443105464e-06, + "loss": 1.56086617, + "memory(GiB)": 111.15, + "step": 54365, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.64108496, + "epoch": 1.3792491121258244, + "grad_norm": 5.25, + "learning_rate": 2.410460556511684e-06, + "loss": 1.58796272, + "memory(GiB)": 111.15, + "step": 54370, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.65421605, + "epoch": 1.3793759512937596, + "grad_norm": 5.21875, + "learning_rate": 2.4095635826123235e-06, + "loss": 1.58763866, + "memory(GiB)": 111.15, + "step": 54375, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.66396832, + "epoch": 1.3795027904616946, + "grad_norm": 5.0625, + "learning_rate": 2.4086667226519245e-06, + "loss": 1.60019321, + "memory(GiB)": 111.15, + "step": 54380, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.65088072, + "epoch": 1.3796296296296298, + "grad_norm": 5.46875, + "learning_rate": 2.4077699766699323e-06, + "loss": 1.66221428, + "memory(GiB)": 111.15, + "step": 54385, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.65890799, + "epoch": 1.3797564687975648, + "grad_norm": 4.71875, + "learning_rate": 2.4068733447057903e-06, + "loss": 1.60825596, + "memory(GiB)": 111.15, + "step": 54390, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.62435875, + "epoch": 1.3798833079654997, + "grad_norm": 5.1875, + "learning_rate": 2.405976826798936e-06, + "loss": 1.71643028, + "memory(GiB)": 111.15, + "step": 54395, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.66107168, + "epoch": 1.3800101471334347, + "grad_norm": 4.71875, + "learning_rate": 2.405080422988802e-06, + "loss": 1.59396954, + "memory(GiB)": 111.15, + "step": 54400, + "train_speed(iter/s)": 0.447549 + }, + { + "epoch": 1.3800101471334347, + "eval_acc": 0.6434326770069646, + "eval_loss": 1.6075164079666138, + "eval_runtime": 114.4328, + "eval_samples_per_second": 55.666, + "eval_steps_per_second": 27.833, + "step": 54400 + }, + { + "acc": 0.64809713, + "epoch": 1.38013698630137, + "grad_norm": 5.1875, + "learning_rate": 2.404184133314817e-06, + "loss": 1.67884521, + "memory(GiB)": 111.15, + "step": 54405, + "train_speed(iter/s)": 0.447133 + }, + { + "acc": 0.65976915, + "epoch": 1.380263825469305, + "grad_norm": 5.9375, + "learning_rate": 2.4032879578164027e-06, + "loss": 1.59476967, + "memory(GiB)": 111.15, + "step": 54410, + "train_speed(iter/s)": 0.447154 + }, + { + "acc": 0.65071726, + "epoch": 1.38039066463724, + "grad_norm": 5.65625, + "learning_rate": 2.402391896532978e-06, + "loss": 1.62358932, + "memory(GiB)": 111.15, + "step": 54415, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.63930559, + "epoch": 1.3805175038051751, + "grad_norm": 5.375, + "learning_rate": 2.4014959495039548e-06, + "loss": 1.71121559, + "memory(GiB)": 111.15, + "step": 54420, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.6498456, + "epoch": 1.3806443429731101, + "grad_norm": 5.28125, + "learning_rate": 2.4006001167687416e-06, + "loss": 1.64641342, + "memory(GiB)": 111.15, + "step": 54425, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.6600049, + "epoch": 1.380771182141045, + "grad_norm": 4.9375, + "learning_rate": 2.39970439836674e-06, + "loss": 1.59635162, + "memory(GiB)": 111.15, + "step": 54430, + "train_speed(iter/s)": 0.44724 + }, + { + "acc": 0.64323215, + "epoch": 1.38089802130898, + "grad_norm": 4.96875, + "learning_rate": 2.3988087943373497e-06, + "loss": 1.63957348, + "memory(GiB)": 111.15, + "step": 54435, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.66327171, + "epoch": 1.3810248604769153, + "grad_norm": 4.875, + "learning_rate": 2.397913304719961e-06, + "loss": 1.5317564, + "memory(GiB)": 111.15, + "step": 54440, + "train_speed(iter/s)": 0.447283 + }, + { + "acc": 0.65372186, + "epoch": 1.3811516996448503, + "grad_norm": 5.90625, + "learning_rate": 2.397017929553961e-06, + "loss": 1.59062748, + "memory(GiB)": 111.15, + "step": 54445, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.6593924, + "epoch": 1.3812785388127855, + "grad_norm": 5.34375, + "learning_rate": 2.396122668878738e-06, + "loss": 1.60315037, + "memory(GiB)": 111.15, + "step": 54450, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.63230972, + "epoch": 1.3814053779807205, + "grad_norm": 5.375, + "learning_rate": 2.3952275227336636e-06, + "loss": 1.74860439, + "memory(GiB)": 111.15, + "step": 54455, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.6544405, + "epoch": 1.3815322171486555, + "grad_norm": 5.0625, + "learning_rate": 2.3943324911581117e-06, + "loss": 1.59573669, + "memory(GiB)": 111.15, + "step": 54460, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.65410528, + "epoch": 1.3816590563165905, + "grad_norm": 4.8125, + "learning_rate": 2.393437574191449e-06, + "loss": 1.61146431, + "memory(GiB)": 111.15, + "step": 54465, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.65370698, + "epoch": 1.3817858954845257, + "grad_norm": 5.03125, + "learning_rate": 2.3925427718730426e-06, + "loss": 1.61598873, + "memory(GiB)": 111.15, + "step": 54470, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.65521479, + "epoch": 1.3819127346524607, + "grad_norm": 6.03125, + "learning_rate": 2.391648084242245e-06, + "loss": 1.57316408, + "memory(GiB)": 111.15, + "step": 54475, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.64912281, + "epoch": 1.3820395738203957, + "grad_norm": 5.09375, + "learning_rate": 2.3907535113384084e-06, + "loss": 1.68617134, + "memory(GiB)": 111.15, + "step": 54480, + "train_speed(iter/s)": 0.447455 + }, + { + "acc": 0.64956026, + "epoch": 1.3821664129883309, + "grad_norm": 5.5625, + "learning_rate": 2.389859053200883e-06, + "loss": 1.64678841, + "memory(GiB)": 111.15, + "step": 54485, + "train_speed(iter/s)": 0.447477 + }, + { + "acc": 0.6673707, + "epoch": 1.3822932521562659, + "grad_norm": 5.5625, + "learning_rate": 2.3889647098690127e-06, + "loss": 1.61290779, + "memory(GiB)": 111.15, + "step": 54490, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.65189824, + "epoch": 1.3824200913242009, + "grad_norm": 5.375, + "learning_rate": 2.3880704813821275e-06, + "loss": 1.62739487, + "memory(GiB)": 111.15, + "step": 54495, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.64360433, + "epoch": 1.3825469304921358, + "grad_norm": 4.84375, + "learning_rate": 2.3871763677795656e-06, + "loss": 1.64381294, + "memory(GiB)": 111.15, + "step": 54500, + "train_speed(iter/s)": 0.447542 + }, + { + "epoch": 1.3825469304921358, + "eval_acc": 0.64342849933387, + "eval_loss": 1.6075654029846191, + "eval_runtime": 113.5854, + "eval_samples_per_second": 56.081, + "eval_steps_per_second": 28.041, + "step": 54500 + }, + { + "acc": 0.65638885, + "epoch": 1.382673769660071, + "grad_norm": 5.84375, + "learning_rate": 2.386282369100653e-06, + "loss": 1.59132185, + "memory(GiB)": 111.15, + "step": 54505, + "train_speed(iter/s)": 0.44713 + }, + { + "acc": 0.65127239, + "epoch": 1.382800608828006, + "grad_norm": 5.15625, + "learning_rate": 2.385388485384713e-06, + "loss": 1.67811546, + "memory(GiB)": 111.15, + "step": 54510, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.65806646, + "epoch": 1.3829274479959413, + "grad_norm": 5.53125, + "learning_rate": 2.384494716671057e-06, + "loss": 1.61329479, + "memory(GiB)": 111.15, + "step": 54515, + "train_speed(iter/s)": 0.447173 + }, + { + "acc": 0.672545, + "epoch": 1.3830542871638762, + "grad_norm": 5.8125, + "learning_rate": 2.3836010629990027e-06, + "loss": 1.4973649, + "memory(GiB)": 111.15, + "step": 54520, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.65482564, + "epoch": 1.3831811263318112, + "grad_norm": 5.9375, + "learning_rate": 2.382707524407855e-06, + "loss": 1.62267399, + "memory(GiB)": 111.15, + "step": 54525, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.64949207, + "epoch": 1.3833079654997462, + "grad_norm": 5.5, + "learning_rate": 2.3818141009369155e-06, + "loss": 1.64203396, + "memory(GiB)": 111.15, + "step": 54530, + "train_speed(iter/s)": 0.447238 + }, + { + "acc": 0.65520215, + "epoch": 1.3834348046676814, + "grad_norm": 4.78125, + "learning_rate": 2.3809207926254813e-06, + "loss": 1.63779392, + "memory(GiB)": 111.15, + "step": 54535, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.66122842, + "epoch": 1.3835616438356164, + "grad_norm": 5.25, + "learning_rate": 2.380027599512844e-06, + "loss": 1.60229111, + "memory(GiB)": 111.15, + "step": 54540, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.64745474, + "epoch": 1.3836884830035516, + "grad_norm": 6.5625, + "learning_rate": 2.3791345216382906e-06, + "loss": 1.63672523, + "memory(GiB)": 111.15, + "step": 54545, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.65147991, + "epoch": 1.3838153221714866, + "grad_norm": 5.4375, + "learning_rate": 2.378241559041102e-06, + "loss": 1.64390697, + "memory(GiB)": 111.15, + "step": 54550, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.65547714, + "epoch": 1.3839421613394216, + "grad_norm": 5.59375, + "learning_rate": 2.377348711760555e-06, + "loss": 1.68051758, + "memory(GiB)": 111.15, + "step": 54555, + "train_speed(iter/s)": 0.447346 + }, + { + "acc": 0.64207006, + "epoch": 1.3840690005073566, + "grad_norm": 5.65625, + "learning_rate": 2.3764559798359204e-06, + "loss": 1.66483994, + "memory(GiB)": 111.15, + "step": 54560, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.66335669, + "epoch": 1.3841958396752918, + "grad_norm": 4.78125, + "learning_rate": 2.3755633633064658e-06, + "loss": 1.54986076, + "memory(GiB)": 111.15, + "step": 54565, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.62395768, + "epoch": 1.3843226788432268, + "grad_norm": 4.625, + "learning_rate": 2.374670862211451e-06, + "loss": 1.71314468, + "memory(GiB)": 111.15, + "step": 54570, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.65833616, + "epoch": 1.3844495180111618, + "grad_norm": 5.8125, + "learning_rate": 2.373778476590134e-06, + "loss": 1.58416939, + "memory(GiB)": 111.15, + "step": 54575, + "train_speed(iter/s)": 0.447432 + }, + { + "acc": 0.65811334, + "epoch": 1.384576357179097, + "grad_norm": 5.25, + "learning_rate": 2.372886206481764e-06, + "loss": 1.61228466, + "memory(GiB)": 111.15, + "step": 54580, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.6383656, + "epoch": 1.384703196347032, + "grad_norm": 4.1875, + "learning_rate": 2.3719940519255864e-06, + "loss": 1.65537605, + "memory(GiB)": 111.15, + "step": 54585, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.64007697, + "epoch": 1.384830035514967, + "grad_norm": 5.46875, + "learning_rate": 2.371102012960847e-06, + "loss": 1.64312706, + "memory(GiB)": 111.15, + "step": 54590, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.65482688, + "epoch": 1.384956874682902, + "grad_norm": 7.40625, + "learning_rate": 2.3702100896267767e-06, + "loss": 1.61087494, + "memory(GiB)": 111.15, + "step": 54595, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.65495806, + "epoch": 1.3850837138508372, + "grad_norm": 4.5625, + "learning_rate": 2.3693182819626077e-06, + "loss": 1.6300602, + "memory(GiB)": 111.15, + "step": 54600, + "train_speed(iter/s)": 0.447541 + }, + { + "epoch": 1.3850837138508372, + "eval_acc": 0.6434213972896092, + "eval_loss": 1.607495903968811, + "eval_runtime": 113.6642, + "eval_samples_per_second": 56.042, + "eval_steps_per_second": 28.021, + "step": 54600 + }, + { + "acc": 0.64413443, + "epoch": 1.3852105530187722, + "grad_norm": 4.875, + "learning_rate": 2.3684265900075637e-06, + "loss": 1.62472839, + "memory(GiB)": 111.15, + "step": 54605, + "train_speed(iter/s)": 0.447129 + }, + { + "acc": 0.65278025, + "epoch": 1.3853373921867074, + "grad_norm": 4.90625, + "learning_rate": 2.3675350138008714e-06, + "loss": 1.62760048, + "memory(GiB)": 111.15, + "step": 54610, + "train_speed(iter/s)": 0.44715 + }, + { + "acc": 0.65307603, + "epoch": 1.3854642313546424, + "grad_norm": 5.0625, + "learning_rate": 2.3666435533817406e-06, + "loss": 1.6288929, + "memory(GiB)": 111.15, + "step": 54615, + "train_speed(iter/s)": 0.447172 + }, + { + "acc": 0.65403109, + "epoch": 1.3855910705225774, + "grad_norm": 6.59375, + "learning_rate": 2.3657522087893806e-06, + "loss": 1.65198021, + "memory(GiB)": 111.15, + "step": 54620, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.66682835, + "epoch": 1.3857179096905123, + "grad_norm": 4.75, + "learning_rate": 2.3648609800630022e-06, + "loss": 1.5866703, + "memory(GiB)": 111.15, + "step": 54625, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.64418054, + "epoch": 1.3858447488584476, + "grad_norm": 4.84375, + "learning_rate": 2.363969867241805e-06, + "loss": 1.62624779, + "memory(GiB)": 111.15, + "step": 54630, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.63954716, + "epoch": 1.3859715880263825, + "grad_norm": 5.84375, + "learning_rate": 2.36307887036498e-06, + "loss": 1.64616871, + "memory(GiB)": 111.15, + "step": 54635, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.64608278, + "epoch": 1.3860984271943175, + "grad_norm": 4.90625, + "learning_rate": 2.3621879894717177e-06, + "loss": 1.607967, + "memory(GiB)": 111.15, + "step": 54640, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.67078714, + "epoch": 1.3862252663622527, + "grad_norm": 5.75, + "learning_rate": 2.361297224601206e-06, + "loss": 1.57571125, + "memory(GiB)": 111.15, + "step": 54645, + "train_speed(iter/s)": 0.4473 + }, + { + "acc": 0.67344413, + "epoch": 1.3863521055301877, + "grad_norm": 5.5, + "learning_rate": 2.360406575792625e-06, + "loss": 1.54310455, + "memory(GiB)": 111.15, + "step": 54650, + "train_speed(iter/s)": 0.447322 + }, + { + "acc": 0.66326408, + "epoch": 1.3864789446981227, + "grad_norm": 5.1875, + "learning_rate": 2.3595160430851445e-06, + "loss": 1.5397747, + "memory(GiB)": 111.15, + "step": 54655, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.64254022, + "epoch": 1.3866057838660577, + "grad_norm": 6.0625, + "learning_rate": 2.3586256265179392e-06, + "loss": 1.66351585, + "memory(GiB)": 111.15, + "step": 54660, + "train_speed(iter/s)": 0.447365 + }, + { + "acc": 0.65499001, + "epoch": 1.386732623033993, + "grad_norm": 5.75, + "learning_rate": 2.3577353261301715e-06, + "loss": 1.5509016, + "memory(GiB)": 111.15, + "step": 54665, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.64582405, + "epoch": 1.386859462201928, + "grad_norm": 5.96875, + "learning_rate": 2.356845141961001e-06, + "loss": 1.66930466, + "memory(GiB)": 111.15, + "step": 54670, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.64140797, + "epoch": 1.3869863013698631, + "grad_norm": 5.4375, + "learning_rate": 2.355955074049582e-06, + "loss": 1.62228298, + "memory(GiB)": 111.15, + "step": 54675, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.66235514, + "epoch": 1.387113140537798, + "grad_norm": 5.34375, + "learning_rate": 2.355065122435064e-06, + "loss": 1.61223392, + "memory(GiB)": 111.15, + "step": 54680, + "train_speed(iter/s)": 0.44745 + }, + { + "acc": 0.64747496, + "epoch": 1.387239979705733, + "grad_norm": 6.0, + "learning_rate": 2.3541752871565902e-06, + "loss": 1.66372871, + "memory(GiB)": 111.15, + "step": 54685, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.65175772, + "epoch": 1.387366818873668, + "grad_norm": 4.6875, + "learning_rate": 2.3532855682533003e-06, + "loss": 1.61401672, + "memory(GiB)": 111.15, + "step": 54690, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.65787735, + "epoch": 1.3874936580416033, + "grad_norm": 5.875, + "learning_rate": 2.352395965764328e-06, + "loss": 1.62926369, + "memory(GiB)": 111.15, + "step": 54695, + "train_speed(iter/s)": 0.447514 + }, + { + "acc": 0.65182753, + "epoch": 1.3876204972095383, + "grad_norm": 4.75, + "learning_rate": 2.3515064797288013e-06, + "loss": 1.62496624, + "memory(GiB)": 111.15, + "step": 54700, + "train_speed(iter/s)": 0.447536 + }, + { + "epoch": 1.3876204972095383, + "eval_acc": 0.6434356013781308, + "eval_loss": 1.6075246334075928, + "eval_runtime": 113.5641, + "eval_samples_per_second": 56.092, + "eval_steps_per_second": 28.046, + "step": 54700 + }, + { + "acc": 0.64899197, + "epoch": 1.3877473363774735, + "grad_norm": 4.53125, + "learning_rate": 2.350617110185845e-06, + "loss": 1.65967941, + "memory(GiB)": 111.15, + "step": 54705, + "train_speed(iter/s)": 0.447125 + }, + { + "acc": 0.64957457, + "epoch": 1.3878741755454085, + "grad_norm": 4.875, + "learning_rate": 2.3497278571745763e-06, + "loss": 1.62741013, + "memory(GiB)": 111.15, + "step": 54710, + "train_speed(iter/s)": 0.447146 + }, + { + "acc": 0.6589107, + "epoch": 1.3880010147133435, + "grad_norm": 5.46875, + "learning_rate": 2.348838720734109e-06, + "loss": 1.55465355, + "memory(GiB)": 111.15, + "step": 54715, + "train_speed(iter/s)": 0.447167 + }, + { + "acc": 0.66221108, + "epoch": 1.3881278538812785, + "grad_norm": 5.90625, + "learning_rate": 2.347949700903552e-06, + "loss": 1.6439146, + "memory(GiB)": 111.15, + "step": 54720, + "train_speed(iter/s)": 0.447189 + }, + { + "acc": 0.6477808, + "epoch": 1.3882546930492137, + "grad_norm": 5.625, + "learning_rate": 2.3470607977220066e-06, + "loss": 1.60640831, + "memory(GiB)": 111.15, + "step": 54725, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.64855957, + "epoch": 1.3883815322171487, + "grad_norm": 4.90625, + "learning_rate": 2.346172011228573e-06, + "loss": 1.65097847, + "memory(GiB)": 111.15, + "step": 54730, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.66187353, + "epoch": 1.3885083713850837, + "grad_norm": 5.1875, + "learning_rate": 2.345283341462342e-06, + "loss": 1.5785387, + "memory(GiB)": 111.15, + "step": 54735, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.64422159, + "epoch": 1.3886352105530189, + "grad_norm": 5.375, + "learning_rate": 2.3443947884624026e-06, + "loss": 1.66559448, + "memory(GiB)": 111.15, + "step": 54740, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.65672851, + "epoch": 1.3887620497209539, + "grad_norm": 5.40625, + "learning_rate": 2.3435063522678346e-06, + "loss": 1.55562334, + "memory(GiB)": 111.15, + "step": 54745, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.65175314, + "epoch": 1.3888888888888888, + "grad_norm": 4.65625, + "learning_rate": 2.3426180329177217e-06, + "loss": 1.64230595, + "memory(GiB)": 111.15, + "step": 54750, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.64142323, + "epoch": 1.3890157280568238, + "grad_norm": 4.875, + "learning_rate": 2.3417298304511297e-06, + "loss": 1.61486931, + "memory(GiB)": 111.15, + "step": 54755, + "train_speed(iter/s)": 0.447341 + }, + { + "acc": 0.62748523, + "epoch": 1.389142567224759, + "grad_norm": 5.5, + "learning_rate": 2.340841744907127e-06, + "loss": 1.7429863, + "memory(GiB)": 111.15, + "step": 54760, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.65717201, + "epoch": 1.389269406392694, + "grad_norm": 5.71875, + "learning_rate": 2.3399537763247783e-06, + "loss": 1.55564003, + "memory(GiB)": 111.15, + "step": 54765, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.63951383, + "epoch": 1.3893962455606292, + "grad_norm": 5.125, + "learning_rate": 2.3390659247431404e-06, + "loss": 1.62931862, + "memory(GiB)": 111.15, + "step": 54770, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.65612378, + "epoch": 1.3895230847285642, + "grad_norm": 5.75, + "learning_rate": 2.338178190201261e-06, + "loss": 1.58162451, + "memory(GiB)": 111.15, + "step": 54775, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.6476398, + "epoch": 1.3896499238964992, + "grad_norm": 5.34375, + "learning_rate": 2.3372905727381877e-06, + "loss": 1.64012775, + "memory(GiB)": 111.15, + "step": 54780, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.66441069, + "epoch": 1.3897767630644342, + "grad_norm": 5.21875, + "learning_rate": 2.3364030723929647e-06, + "loss": 1.55012369, + "memory(GiB)": 111.15, + "step": 54785, + "train_speed(iter/s)": 0.447471 + }, + { + "acc": 0.66532369, + "epoch": 1.3899036022323694, + "grad_norm": 5.21875, + "learning_rate": 2.335515689204629e-06, + "loss": 1.57364197, + "memory(GiB)": 111.15, + "step": 54790, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.64783525, + "epoch": 1.3900304414003044, + "grad_norm": 5.71875, + "learning_rate": 2.334628423212206e-06, + "loss": 1.63448849, + "memory(GiB)": 111.15, + "step": 54795, + "train_speed(iter/s)": 0.447514 + }, + { + "acc": 0.65659075, + "epoch": 1.3901572805682394, + "grad_norm": 5.96875, + "learning_rate": 2.3337412744547256e-06, + "loss": 1.62704849, + "memory(GiB)": 111.15, + "step": 54800, + "train_speed(iter/s)": 0.447536 + }, + { + "epoch": 1.3901572805682394, + "eval_acc": 0.643433094774274, + "eval_loss": 1.6075447797775269, + "eval_runtime": 113.1756, + "eval_samples_per_second": 56.284, + "eval_steps_per_second": 28.142, + "step": 54800 + }, + { + "acc": 0.6657032, + "epoch": 1.3902841197361746, + "grad_norm": 5.09375, + "learning_rate": 2.332854242971209e-06, + "loss": 1.58884039, + "memory(GiB)": 111.15, + "step": 54805, + "train_speed(iter/s)": 0.447128 + }, + { + "acc": 0.64376736, + "epoch": 1.3904109589041096, + "grad_norm": 6.84375, + "learning_rate": 2.331967328800672e-06, + "loss": 1.60647125, + "memory(GiB)": 111.15, + "step": 54810, + "train_speed(iter/s)": 0.447149 + }, + { + "acc": 0.65672731, + "epoch": 1.3905377980720446, + "grad_norm": 5.75, + "learning_rate": 2.33108053198212e-06, + "loss": 1.55529099, + "memory(GiB)": 111.15, + "step": 54815, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.65332303, + "epoch": 1.3906646372399796, + "grad_norm": 5.3125, + "learning_rate": 2.330193852554564e-06, + "loss": 1.59610481, + "memory(GiB)": 111.15, + "step": 54820, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.66074114, + "epoch": 1.3907914764079148, + "grad_norm": 6.875, + "learning_rate": 2.3293072905570024e-06, + "loss": 1.63630447, + "memory(GiB)": 111.15, + "step": 54825, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.66354952, + "epoch": 1.3909183155758498, + "grad_norm": 4.5, + "learning_rate": 2.3284208460284303e-06, + "loss": 1.61565628, + "memory(GiB)": 111.15, + "step": 54830, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.66817188, + "epoch": 1.391045154743785, + "grad_norm": 6.25, + "learning_rate": 2.3275345190078364e-06, + "loss": 1.64965973, + "memory(GiB)": 111.15, + "step": 54835, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.63878031, + "epoch": 1.39117199391172, + "grad_norm": 5.15625, + "learning_rate": 2.3266483095342064e-06, + "loss": 1.64956131, + "memory(GiB)": 111.15, + "step": 54840, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.65426354, + "epoch": 1.391298833079655, + "grad_norm": 5.4375, + "learning_rate": 2.3257622176465194e-06, + "loss": 1.57891293, + "memory(GiB)": 111.15, + "step": 54845, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.64944115, + "epoch": 1.39142567224759, + "grad_norm": 6.09375, + "learning_rate": 2.3248762433837494e-06, + "loss": 1.66123409, + "memory(GiB)": 111.15, + "step": 54850, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.64330759, + "epoch": 1.3915525114155252, + "grad_norm": 4.9375, + "learning_rate": 2.323990386784867e-06, + "loss": 1.60456104, + "memory(GiB)": 111.15, + "step": 54855, + "train_speed(iter/s)": 0.447345 + }, + { + "acc": 0.6554965, + "epoch": 1.3916793505834602, + "grad_norm": 6.03125, + "learning_rate": 2.3231046478888335e-06, + "loss": 1.63987579, + "memory(GiB)": 111.15, + "step": 54860, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.64334521, + "epoch": 1.3918061897513954, + "grad_norm": 5.65625, + "learning_rate": 2.3222190267346094e-06, + "loss": 1.62673912, + "memory(GiB)": 111.15, + "step": 54865, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.65283642, + "epoch": 1.3919330289193304, + "grad_norm": 5.625, + "learning_rate": 2.3213335233611484e-06, + "loss": 1.61944389, + "memory(GiB)": 111.15, + "step": 54870, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.65002565, + "epoch": 1.3920598680872653, + "grad_norm": 6.59375, + "learning_rate": 2.320448137807398e-06, + "loss": 1.6193058, + "memory(GiB)": 111.15, + "step": 54875, + "train_speed(iter/s)": 0.447433 + }, + { + "acc": 0.64873419, + "epoch": 1.3921867072552003, + "grad_norm": 5.25, + "learning_rate": 2.3195628701123017e-06, + "loss": 1.67553062, + "memory(GiB)": 111.15, + "step": 54880, + "train_speed(iter/s)": 0.447455 + }, + { + "acc": 0.65818777, + "epoch": 1.3923135464231355, + "grad_norm": 4.96875, + "learning_rate": 2.3186777203147964e-06, + "loss": 1.58723354, + "memory(GiB)": 111.15, + "step": 54885, + "train_speed(iter/s)": 0.447477 + }, + { + "acc": 0.6607152, + "epoch": 1.3924403855910705, + "grad_norm": 5.25, + "learning_rate": 2.3177926884538193e-06, + "loss": 1.63881016, + "memory(GiB)": 111.15, + "step": 54890, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.65617328, + "epoch": 1.3925672247590055, + "grad_norm": 5.8125, + "learning_rate": 2.3169077745682933e-06, + "loss": 1.60396385, + "memory(GiB)": 111.15, + "step": 54895, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.6475872, + "epoch": 1.3926940639269407, + "grad_norm": 5.6875, + "learning_rate": 2.316022978697143e-06, + "loss": 1.62557583, + "memory(GiB)": 111.15, + "step": 54900, + "train_speed(iter/s)": 0.447542 + }, + { + "epoch": 1.3926940639269407, + "eval_acc": 0.6434314237050363, + "eval_loss": 1.6075263023376465, + "eval_runtime": 113.603, + "eval_samples_per_second": 56.072, + "eval_steps_per_second": 28.036, + "step": 54900 + }, + { + "acc": 0.6423377, + "epoch": 1.3928209030948757, + "grad_norm": 4.84375, + "learning_rate": 2.3151383008792826e-06, + "loss": 1.63268089, + "memory(GiB)": 111.15, + "step": 54905, + "train_speed(iter/s)": 0.447133 + }, + { + "acc": 0.65516534, + "epoch": 1.3929477422628107, + "grad_norm": 5.46875, + "learning_rate": 2.314253741153631e-06, + "loss": 1.60871296, + "memory(GiB)": 111.15, + "step": 54910, + "train_speed(iter/s)": 0.447154 + }, + { + "acc": 0.64471183, + "epoch": 1.3930745814307457, + "grad_norm": 5.125, + "learning_rate": 2.313369299559088e-06, + "loss": 1.55853024, + "memory(GiB)": 111.15, + "step": 54915, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.65644975, + "epoch": 1.393201420598681, + "grad_norm": 4.625, + "learning_rate": 2.3124849761345576e-06, + "loss": 1.52087555, + "memory(GiB)": 111.15, + "step": 54920, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.66084452, + "epoch": 1.393328259766616, + "grad_norm": 4.59375, + "learning_rate": 2.311600770918938e-06, + "loss": 1.62343616, + "memory(GiB)": 111.15, + "step": 54925, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.65948701, + "epoch": 1.393455098934551, + "grad_norm": 4.65625, + "learning_rate": 2.310716683951122e-06, + "loss": 1.60135612, + "memory(GiB)": 111.15, + "step": 54930, + "train_speed(iter/s)": 0.44724 + }, + { + "acc": 0.64936323, + "epoch": 1.393581938102486, + "grad_norm": 6.6875, + "learning_rate": 2.3098327152699884e-06, + "loss": 1.68810062, + "memory(GiB)": 111.15, + "step": 54935, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.65774803, + "epoch": 1.393708777270421, + "grad_norm": 5.1875, + "learning_rate": 2.308948864914425e-06, + "loss": 1.63990288, + "memory(GiB)": 111.15, + "step": 54940, + "train_speed(iter/s)": 0.447283 + }, + { + "acc": 0.64700813, + "epoch": 1.393835616438356, + "grad_norm": 5.15625, + "learning_rate": 2.308065132923305e-06, + "loss": 1.66170807, + "memory(GiB)": 111.15, + "step": 54945, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.65739832, + "epoch": 1.3939624556062913, + "grad_norm": 5.03125, + "learning_rate": 2.3071815193355005e-06, + "loss": 1.56487083, + "memory(GiB)": 111.15, + "step": 54950, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.6588006, + "epoch": 1.3940892947742263, + "grad_norm": 5.0625, + "learning_rate": 2.3062980241898725e-06, + "loss": 1.56086254, + "memory(GiB)": 111.15, + "step": 54955, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.64591894, + "epoch": 1.3942161339421613, + "grad_norm": 6.40625, + "learning_rate": 2.3054146475252852e-06, + "loss": 1.64110088, + "memory(GiB)": 111.15, + "step": 54960, + "train_speed(iter/s)": 0.44737 + }, + { + "acc": 0.664363, + "epoch": 1.3943429731100965, + "grad_norm": 5.25, + "learning_rate": 2.3045313893805926e-06, + "loss": 1.46369457, + "memory(GiB)": 111.15, + "step": 54965, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.65449085, + "epoch": 1.3944698122780315, + "grad_norm": 7.40625, + "learning_rate": 2.303648249794644e-06, + "loss": 1.65266304, + "memory(GiB)": 111.15, + "step": 54970, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.64992313, + "epoch": 1.3945966514459665, + "grad_norm": 5.375, + "learning_rate": 2.302765228806283e-06, + "loss": 1.61147232, + "memory(GiB)": 111.15, + "step": 54975, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.64690418, + "epoch": 1.3947234906139014, + "grad_norm": 5.46875, + "learning_rate": 2.30188232645435e-06, + "loss": 1.68412018, + "memory(GiB)": 111.15, + "step": 54980, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64164972, + "epoch": 1.3948503297818367, + "grad_norm": 5.65625, + "learning_rate": 2.300999542777678e-06, + "loss": 1.696138, + "memory(GiB)": 111.15, + "step": 54985, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.65270224, + "epoch": 1.3949771689497716, + "grad_norm": 5.5, + "learning_rate": 2.300116877815097e-06, + "loss": 1.57406178, + "memory(GiB)": 111.15, + "step": 54990, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.64173155, + "epoch": 1.3951040081177069, + "grad_norm": 5.40625, + "learning_rate": 2.2992343316054296e-06, + "loss": 1.67154045, + "memory(GiB)": 111.15, + "step": 54995, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.64343715, + "epoch": 1.3952308472856418, + "grad_norm": 4.8125, + "learning_rate": 2.298351904187494e-06, + "loss": 1.57504244, + "memory(GiB)": 111.15, + "step": 55000, + "train_speed(iter/s)": 0.447542 + }, + { + "epoch": 1.3952308472856418, + "eval_acc": 0.643457743045532, + "eval_loss": 1.6074556112289429, + "eval_runtime": 113.1056, + "eval_samples_per_second": 56.319, + "eval_steps_per_second": 28.16, + "step": 55000 + }, + { + "acc": 0.6473381, + "epoch": 1.3953576864535768, + "grad_norm": 5.1875, + "learning_rate": 2.297469595600104e-06, + "loss": 1.61186733, + "memory(GiB)": 111.15, + "step": 55005, + "train_speed(iter/s)": 0.447135 + }, + { + "acc": 0.63960719, + "epoch": 1.3954845256215118, + "grad_norm": 5.09375, + "learning_rate": 2.2965874058820668e-06, + "loss": 1.66384621, + "memory(GiB)": 111.15, + "step": 55010, + "train_speed(iter/s)": 0.447157 + }, + { + "acc": 0.64895201, + "epoch": 1.395611364789447, + "grad_norm": 6.96875, + "learning_rate": 2.2957053350721857e-06, + "loss": 1.61066055, + "memory(GiB)": 111.15, + "step": 55015, + "train_speed(iter/s)": 0.447179 + }, + { + "acc": 0.64946985, + "epoch": 1.395738203957382, + "grad_norm": 5.46875, + "learning_rate": 2.294823383209258e-06, + "loss": 1.63018131, + "memory(GiB)": 111.15, + "step": 55020, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.65105171, + "epoch": 1.3958650431253172, + "grad_norm": 5.3125, + "learning_rate": 2.2939415503320733e-06, + "loss": 1.62948704, + "memory(GiB)": 111.15, + "step": 55025, + "train_speed(iter/s)": 0.447222 + }, + { + "acc": 0.62818089, + "epoch": 1.3959918822932522, + "grad_norm": 5.0625, + "learning_rate": 2.293059836479425e-06, + "loss": 1.68856258, + "memory(GiB)": 111.15, + "step": 55030, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.65504732, + "epoch": 1.3961187214611872, + "grad_norm": 5.15625, + "learning_rate": 2.2921782416900883e-06, + "loss": 1.56294823, + "memory(GiB)": 111.15, + "step": 55035, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.66158495, + "epoch": 1.3962455606291222, + "grad_norm": 6.09375, + "learning_rate": 2.2912967660028425e-06, + "loss": 1.59866343, + "memory(GiB)": 111.15, + "step": 55040, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.65680614, + "epoch": 1.3963723997970574, + "grad_norm": 4.09375, + "learning_rate": 2.2904154094564568e-06, + "loss": 1.56556759, + "memory(GiB)": 111.15, + "step": 55045, + "train_speed(iter/s)": 0.447309 + }, + { + "acc": 0.64282265, + "epoch": 1.3964992389649924, + "grad_norm": 5.9375, + "learning_rate": 2.2895341720897018e-06, + "loss": 1.6341898, + "memory(GiB)": 111.15, + "step": 55050, + "train_speed(iter/s)": 0.447331 + }, + { + "acc": 0.64759293, + "epoch": 1.3966260781329274, + "grad_norm": 5.71875, + "learning_rate": 2.2886530539413336e-06, + "loss": 1.67841644, + "memory(GiB)": 111.15, + "step": 55055, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.66211872, + "epoch": 1.3967529173008626, + "grad_norm": 5.40625, + "learning_rate": 2.2877720550501082e-06, + "loss": 1.62133732, + "memory(GiB)": 111.15, + "step": 55060, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.66987629, + "epoch": 1.3968797564687976, + "grad_norm": 5.8125, + "learning_rate": 2.2868911754547783e-06, + "loss": 1.55506783, + "memory(GiB)": 111.15, + "step": 55065, + "train_speed(iter/s)": 0.447396 + }, + { + "acc": 0.64942927, + "epoch": 1.3970065956367326, + "grad_norm": 5.34375, + "learning_rate": 2.28601041519409e-06, + "loss": 1.60248528, + "memory(GiB)": 111.15, + "step": 55070, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.64305277, + "epoch": 1.3971334348046676, + "grad_norm": 6.15625, + "learning_rate": 2.2851297743067786e-06, + "loss": 1.62245159, + "memory(GiB)": 111.15, + "step": 55075, + "train_speed(iter/s)": 0.44744 + }, + { + "acc": 0.65701847, + "epoch": 1.3972602739726028, + "grad_norm": 5.3125, + "learning_rate": 2.2842492528315784e-06, + "loss": 1.5451685, + "memory(GiB)": 111.15, + "step": 55080, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.65278912, + "epoch": 1.3973871131405378, + "grad_norm": 4.53125, + "learning_rate": 2.283368850807223e-06, + "loss": 1.57256165, + "memory(GiB)": 111.15, + "step": 55085, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.66075153, + "epoch": 1.397513952308473, + "grad_norm": 5.3125, + "learning_rate": 2.282488568272437e-06, + "loss": 1.5908432, + "memory(GiB)": 111.15, + "step": 55090, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.64385486, + "epoch": 1.397640791476408, + "grad_norm": 5.375, + "learning_rate": 2.2816084052659326e-06, + "loss": 1.6499527, + "memory(GiB)": 111.15, + "step": 55095, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.65068898, + "epoch": 1.397767630644343, + "grad_norm": 4.9375, + "learning_rate": 2.2807283618264288e-06, + "loss": 1.63947334, + "memory(GiB)": 111.15, + "step": 55100, + "train_speed(iter/s)": 0.447549 + }, + { + "epoch": 1.397767630644343, + "eval_acc": 0.6434122064088011, + "eval_loss": 1.6075583696365356, + "eval_runtime": 113.6077, + "eval_samples_per_second": 56.07, + "eval_steps_per_second": 28.035, + "step": 55100 + }, + { + "acc": 0.65081344, + "epoch": 1.397894469812278, + "grad_norm": 5.75, + "learning_rate": 2.2798484379926324e-06, + "loss": 1.59877777, + "memory(GiB)": 111.15, + "step": 55105, + "train_speed(iter/s)": 0.447141 + }, + { + "acc": 0.6597518, + "epoch": 1.3980213089802132, + "grad_norm": 5.4375, + "learning_rate": 2.278968633803246e-06, + "loss": 1.55130692, + "memory(GiB)": 111.15, + "step": 55110, + "train_speed(iter/s)": 0.447163 + }, + { + "acc": 0.65268641, + "epoch": 1.3981481481481481, + "grad_norm": 5.1875, + "learning_rate": 2.2780889492969684e-06, + "loss": 1.60199585, + "memory(GiB)": 111.15, + "step": 55115, + "train_speed(iter/s)": 0.447184 + }, + { + "acc": 0.66184564, + "epoch": 1.3982749873160831, + "grad_norm": 6.15625, + "learning_rate": 2.277209384512491e-06, + "loss": 1.57674732, + "memory(GiB)": 111.15, + "step": 55120, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.63805552, + "epoch": 1.3984018264840183, + "grad_norm": 7.46875, + "learning_rate": 2.2763299394885013e-06, + "loss": 1.64926376, + "memory(GiB)": 111.15, + "step": 55125, + "train_speed(iter/s)": 0.447228 + }, + { + "acc": 0.65462427, + "epoch": 1.3985286656519533, + "grad_norm": 4.8125, + "learning_rate": 2.2754506142636808e-06, + "loss": 1.6433672, + "memory(GiB)": 111.15, + "step": 55130, + "train_speed(iter/s)": 0.44725 + }, + { + "acc": 0.66788521, + "epoch": 1.3986555048198883, + "grad_norm": 4.8125, + "learning_rate": 2.274571408876707e-06, + "loss": 1.59453583, + "memory(GiB)": 111.15, + "step": 55135, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.66430168, + "epoch": 1.3987823439878233, + "grad_norm": 5.78125, + "learning_rate": 2.2736923233662504e-06, + "loss": 1.60728226, + "memory(GiB)": 111.15, + "step": 55140, + "train_speed(iter/s)": 0.447294 + }, + { + "acc": 0.64991536, + "epoch": 1.3989091831557585, + "grad_norm": 4.59375, + "learning_rate": 2.2728133577709776e-06, + "loss": 1.68083019, + "memory(GiB)": 111.15, + "step": 55145, + "train_speed(iter/s)": 0.447315 + }, + { + "acc": 0.65153842, + "epoch": 1.3990360223236935, + "grad_norm": 6.125, + "learning_rate": 2.2719345121295495e-06, + "loss": 1.65125751, + "memory(GiB)": 111.15, + "step": 55150, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.64291792, + "epoch": 1.3991628614916287, + "grad_norm": 5.125, + "learning_rate": 2.2710557864806214e-06, + "loss": 1.70624542, + "memory(GiB)": 111.15, + "step": 55155, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.64929333, + "epoch": 1.3992897006595637, + "grad_norm": 5.375, + "learning_rate": 2.2701771808628438e-06, + "loss": 1.66490841, + "memory(GiB)": 111.15, + "step": 55160, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.64578362, + "epoch": 1.3994165398274987, + "grad_norm": 5.15625, + "learning_rate": 2.269298695314861e-06, + "loss": 1.67416286, + "memory(GiB)": 111.15, + "step": 55165, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.6508193, + "epoch": 1.3995433789954337, + "grad_norm": 5.53125, + "learning_rate": 2.268420329875314e-06, + "loss": 1.65117683, + "memory(GiB)": 111.15, + "step": 55170, + "train_speed(iter/s)": 0.447425 + }, + { + "acc": 0.655901, + "epoch": 1.399670218163369, + "grad_norm": 5.15625, + "learning_rate": 2.2675420845828363e-06, + "loss": 1.5449028, + "memory(GiB)": 111.15, + "step": 55175, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.64606695, + "epoch": 1.3997970573313039, + "grad_norm": 6.28125, + "learning_rate": 2.266663959476057e-06, + "loss": 1.612743, + "memory(GiB)": 111.15, + "step": 55180, + "train_speed(iter/s)": 0.447468 + }, + { + "acc": 0.65505619, + "epoch": 1.399923896499239, + "grad_norm": 6.0, + "learning_rate": 2.265785954593598e-06, + "loss": 1.61517601, + "memory(GiB)": 111.15, + "step": 55185, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.6420579, + "epoch": 1.400050735667174, + "grad_norm": 5.6875, + "learning_rate": 2.264908069974085e-06, + "loss": 1.62605286, + "memory(GiB)": 111.15, + "step": 55190, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.65596619, + "epoch": 1.400177574835109, + "grad_norm": 4.21875, + "learning_rate": 2.2640303056561236e-06, + "loss": 1.54183416, + "memory(GiB)": 111.15, + "step": 55195, + "train_speed(iter/s)": 0.447534 + }, + { + "acc": 0.63990536, + "epoch": 1.400304414003044, + "grad_norm": 7.125, + "learning_rate": 2.2631526616783234e-06, + "loss": 1.60768051, + "memory(GiB)": 111.15, + "step": 55200, + "train_speed(iter/s)": 0.447556 + }, + { + "epoch": 1.400304414003044, + "eval_acc": 0.6433670875393798, + "eval_loss": 1.607593297958374, + "eval_runtime": 113.1287, + "eval_samples_per_second": 56.308, + "eval_steps_per_second": 28.154, + "step": 55200 + }, + { + "acc": 0.66301341, + "epoch": 1.4004312531709793, + "grad_norm": 5.28125, + "learning_rate": 2.2622751380792896e-06, + "loss": 1.58845654, + "memory(GiB)": 111.15, + "step": 55205, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.63432169, + "epoch": 1.4005580923389143, + "grad_norm": 5.25, + "learning_rate": 2.26139773489762e-06, + "loss": 1.66355934, + "memory(GiB)": 111.15, + "step": 55210, + "train_speed(iter/s)": 0.447173 + }, + { + "acc": 0.65619445, + "epoch": 1.4006849315068493, + "grad_norm": 4.625, + "learning_rate": 2.260520452171904e-06, + "loss": 1.59360943, + "memory(GiB)": 111.15, + "step": 55215, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.64894266, + "epoch": 1.4008117706747845, + "grad_norm": 4.875, + "learning_rate": 2.259643289940727e-06, + "loss": 1.58452129, + "memory(GiB)": 111.15, + "step": 55220, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.65393767, + "epoch": 1.4009386098427195, + "grad_norm": 4.90625, + "learning_rate": 2.2587662482426748e-06, + "loss": 1.61127453, + "memory(GiB)": 111.15, + "step": 55225, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.65080543, + "epoch": 1.4010654490106544, + "grad_norm": 6.78125, + "learning_rate": 2.2578893271163234e-06, + "loss": 1.67061043, + "memory(GiB)": 111.15, + "step": 55230, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.64757204, + "epoch": 1.4011922881785894, + "grad_norm": 5.5, + "learning_rate": 2.2570125266002385e-06, + "loss": 1.61983376, + "memory(GiB)": 111.15, + "step": 55235, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.65525036, + "epoch": 1.4013191273465246, + "grad_norm": 5.5625, + "learning_rate": 2.2561358467329907e-06, + "loss": 1.59737177, + "memory(GiB)": 111.15, + "step": 55240, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.65722089, + "epoch": 1.4014459665144596, + "grad_norm": 5.125, + "learning_rate": 2.2552592875531397e-06, + "loss": 1.5977561, + "memory(GiB)": 111.15, + "step": 55245, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.65912437, + "epoch": 1.4015728056823948, + "grad_norm": 5.125, + "learning_rate": 2.254382849099241e-06, + "loss": 1.60434227, + "memory(GiB)": 111.15, + "step": 55250, + "train_speed(iter/s)": 0.447346 + }, + { + "acc": 0.65998697, + "epoch": 1.4016996448503298, + "grad_norm": 5.40625, + "learning_rate": 2.253506531409839e-06, + "loss": 1.5549181, + "memory(GiB)": 111.15, + "step": 55255, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.6412231, + "epoch": 1.4018264840182648, + "grad_norm": 6.40625, + "learning_rate": 2.252630334523484e-06, + "loss": 1.66369858, + "memory(GiB)": 111.15, + "step": 55260, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.64555492, + "epoch": 1.4019533231861998, + "grad_norm": 6.28125, + "learning_rate": 2.2517542584787134e-06, + "loss": 1.70186577, + "memory(GiB)": 111.15, + "step": 55265, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.65826564, + "epoch": 1.402080162354135, + "grad_norm": 5.34375, + "learning_rate": 2.2508783033140596e-06, + "loss": 1.6135807, + "memory(GiB)": 111.15, + "step": 55270, + "train_speed(iter/s)": 0.447432 + }, + { + "acc": 0.63344717, + "epoch": 1.40220700152207, + "grad_norm": 6.1875, + "learning_rate": 2.2500024690680528e-06, + "loss": 1.6980217, + "memory(GiB)": 111.15, + "step": 55275, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.664363, + "epoch": 1.402333840690005, + "grad_norm": 5.3125, + "learning_rate": 2.249126755779215e-06, + "loss": 1.54853458, + "memory(GiB)": 111.15, + "step": 55280, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.64863424, + "epoch": 1.4024606798579402, + "grad_norm": 5.0625, + "learning_rate": 2.2482511634860645e-06, + "loss": 1.60577488, + "memory(GiB)": 111.15, + "step": 55285, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.63879094, + "epoch": 1.4025875190258752, + "grad_norm": 6.0, + "learning_rate": 2.247375692227113e-06, + "loss": 1.65206032, + "memory(GiB)": 111.15, + "step": 55290, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.65609078, + "epoch": 1.4027143581938102, + "grad_norm": 6.03125, + "learning_rate": 2.2465003420408683e-06, + "loss": 1.62589588, + "memory(GiB)": 111.15, + "step": 55295, + "train_speed(iter/s)": 0.447541 + }, + { + "acc": 0.65455475, + "epoch": 1.4028411973617452, + "grad_norm": 5.375, + "learning_rate": 2.2456251129658325e-06, + "loss": 1.61592922, + "memory(GiB)": 111.15, + "step": 55300, + "train_speed(iter/s)": 0.447562 + }, + { + "epoch": 1.4028411973617452, + "eval_acc": 0.6433758606528784, + "eval_loss": 1.607513427734375, + "eval_runtime": 113.9297, + "eval_samples_per_second": 55.912, + "eval_steps_per_second": 27.956, + "step": 55300 + }, + { + "acc": 0.63301153, + "epoch": 1.4029680365296804, + "grad_norm": 5.5625, + "learning_rate": 2.2447500050405008e-06, + "loss": 1.6759903, + "memory(GiB)": 111.15, + "step": 55305, + "train_speed(iter/s)": 0.447155 + }, + { + "acc": 0.65614877, + "epoch": 1.4030948756976154, + "grad_norm": 5.84375, + "learning_rate": 2.2438750183033657e-06, + "loss": 1.62799282, + "memory(GiB)": 111.15, + "step": 55310, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.6560463, + "epoch": 1.4032217148655506, + "grad_norm": 5.125, + "learning_rate": 2.2430001527929123e-06, + "loss": 1.61211605, + "memory(GiB)": 111.15, + "step": 55315, + "train_speed(iter/s)": 0.447198 + }, + { + "acc": 0.64624987, + "epoch": 1.4033485540334856, + "grad_norm": 6.40625, + "learning_rate": 2.242125408547622e-06, + "loss": 1.69457817, + "memory(GiB)": 111.15, + "step": 55320, + "train_speed(iter/s)": 0.44722 + }, + { + "acc": 0.65175247, + "epoch": 1.4034753932014206, + "grad_norm": 6.625, + "learning_rate": 2.2412507856059667e-06, + "loss": 1.66687088, + "memory(GiB)": 111.15, + "step": 55325, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.64298263, + "epoch": 1.4036022323693556, + "grad_norm": 5.3125, + "learning_rate": 2.2403762840064223e-06, + "loss": 1.60305824, + "memory(GiB)": 111.15, + "step": 55330, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.66607862, + "epoch": 1.4037290715372908, + "grad_norm": 4.53125, + "learning_rate": 2.239501903787448e-06, + "loss": 1.61206703, + "memory(GiB)": 111.15, + "step": 55335, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.64239702, + "epoch": 1.4038559107052258, + "grad_norm": 6.75, + "learning_rate": 2.2386276449875057e-06, + "loss": 1.65491104, + "memory(GiB)": 111.15, + "step": 55340, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.64531026, + "epoch": 1.403982749873161, + "grad_norm": 6.0, + "learning_rate": 2.2377535076450452e-06, + "loss": 1.65138969, + "memory(GiB)": 111.15, + "step": 55345, + "train_speed(iter/s)": 0.447329 + }, + { + "acc": 0.65241423, + "epoch": 1.404109589041096, + "grad_norm": 5.75, + "learning_rate": 2.236879491798522e-06, + "loss": 1.58985825, + "memory(GiB)": 111.15, + "step": 55350, + "train_speed(iter/s)": 0.44735 + }, + { + "acc": 0.65431728, + "epoch": 1.404236428209031, + "grad_norm": 5.78125, + "learning_rate": 2.2360055974863737e-06, + "loss": 1.64613342, + "memory(GiB)": 111.15, + "step": 55355, + "train_speed(iter/s)": 0.447372 + }, + { + "acc": 0.66218414, + "epoch": 1.404363267376966, + "grad_norm": 4.96875, + "learning_rate": 2.2351318247470376e-06, + "loss": 1.60053043, + "memory(GiB)": 111.15, + "step": 55360, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.64293733, + "epoch": 1.4044901065449011, + "grad_norm": 4.9375, + "learning_rate": 2.2342581736189496e-06, + "loss": 1.63935032, + "memory(GiB)": 111.15, + "step": 55365, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.65222206, + "epoch": 1.4046169457128361, + "grad_norm": 5.03125, + "learning_rate": 2.233384644140537e-06, + "loss": 1.61668377, + "memory(GiB)": 111.15, + "step": 55370, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.66723175, + "epoch": 1.4047437848807711, + "grad_norm": 5.4375, + "learning_rate": 2.2325112363502167e-06, + "loss": 1.5490449, + "memory(GiB)": 111.15, + "step": 55375, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.66540251, + "epoch": 1.4048706240487063, + "grad_norm": 4.71875, + "learning_rate": 2.23163795028641e-06, + "loss": 1.57871809, + "memory(GiB)": 111.15, + "step": 55380, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.65160475, + "epoch": 1.4049974632166413, + "grad_norm": 5.0625, + "learning_rate": 2.230764785987526e-06, + "loss": 1.58741741, + "memory(GiB)": 111.15, + "step": 55385, + "train_speed(iter/s)": 0.447502 + }, + { + "acc": 0.645895, + "epoch": 1.4051243023845763, + "grad_norm": 6.5, + "learning_rate": 2.229891743491972e-06, + "loss": 1.57967415, + "memory(GiB)": 111.15, + "step": 55390, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.6462369, + "epoch": 1.4052511415525113, + "grad_norm": 5.125, + "learning_rate": 2.2290188228381434e-06, + "loss": 1.60953979, + "memory(GiB)": 111.15, + "step": 55395, + "train_speed(iter/s)": 0.447545 + }, + { + "acc": 0.65442829, + "epoch": 1.4053779807204465, + "grad_norm": 4.78125, + "learning_rate": 2.2281460240644397e-06, + "loss": 1.61536102, + "memory(GiB)": 111.15, + "step": 55400, + "train_speed(iter/s)": 0.447567 + }, + { + "epoch": 1.4053779807204465, + "eval_acc": 0.6433883936721622, + "eval_loss": 1.6076085567474365, + "eval_runtime": 113.5378, + "eval_samples_per_second": 56.105, + "eval_steps_per_second": 28.052, + "step": 55400 + }, + { + "acc": 0.63874483, + "epoch": 1.4055048198883815, + "grad_norm": 5.28125, + "learning_rate": 2.22727334720925e-06, + "loss": 1.65986176, + "memory(GiB)": 111.15, + "step": 55405, + "train_speed(iter/s)": 0.447161 + }, + { + "acc": 0.65849752, + "epoch": 1.4056316590563167, + "grad_norm": 4.875, + "learning_rate": 2.2264007923109575e-06, + "loss": 1.5765337, + "memory(GiB)": 111.15, + "step": 55410, + "train_speed(iter/s)": 0.447183 + }, + { + "acc": 0.65588531, + "epoch": 1.4057584982242517, + "grad_norm": 6.53125, + "learning_rate": 2.225528359407942e-06, + "loss": 1.62608204, + "memory(GiB)": 111.15, + "step": 55415, + "train_speed(iter/s)": 0.447205 + }, + { + "acc": 0.64721603, + "epoch": 1.4058853373921867, + "grad_norm": 5.40625, + "learning_rate": 2.2246560485385756e-06, + "loss": 1.57760105, + "memory(GiB)": 111.15, + "step": 55420, + "train_speed(iter/s)": 0.447226 + }, + { + "acc": 0.6463644, + "epoch": 1.4060121765601217, + "grad_norm": 5.03125, + "learning_rate": 2.2237838597412277e-06, + "loss": 1.63391647, + "memory(GiB)": 111.15, + "step": 55425, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.66413679, + "epoch": 1.4061390157280569, + "grad_norm": 4.90625, + "learning_rate": 2.22291179305426e-06, + "loss": 1.51885939, + "memory(GiB)": 111.15, + "step": 55430, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.65942345, + "epoch": 1.4062658548959919, + "grad_norm": 5.59375, + "learning_rate": 2.222039848516031e-06, + "loss": 1.58188457, + "memory(GiB)": 111.15, + "step": 55435, + "train_speed(iter/s)": 0.447291 + }, + { + "acc": 0.64248114, + "epoch": 1.4063926940639269, + "grad_norm": 5.625, + "learning_rate": 2.2211680261648918e-06, + "loss": 1.63077087, + "memory(GiB)": 111.15, + "step": 55440, + "train_speed(iter/s)": 0.447312 + }, + { + "acc": 0.65193343, + "epoch": 1.406519533231862, + "grad_norm": 5.15625, + "learning_rate": 2.22029632603919e-06, + "loss": 1.60009689, + "memory(GiB)": 111.15, + "step": 55445, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.65488253, + "epoch": 1.406646372399797, + "grad_norm": 5.71875, + "learning_rate": 2.2194247481772652e-06, + "loss": 1.62145634, + "memory(GiB)": 111.15, + "step": 55450, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.65203609, + "epoch": 1.406773211567732, + "grad_norm": 5.53125, + "learning_rate": 2.218553292617455e-06, + "loss": 1.62728157, + "memory(GiB)": 111.15, + "step": 55455, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.6527369, + "epoch": 1.406900050735667, + "grad_norm": 4.96875, + "learning_rate": 2.2176819593980892e-06, + "loss": 1.5747366, + "memory(GiB)": 111.15, + "step": 55460, + "train_speed(iter/s)": 0.447399 + }, + { + "acc": 0.64417782, + "epoch": 1.4070268899036023, + "grad_norm": 6.84375, + "learning_rate": 2.2168107485574914e-06, + "loss": 1.71052895, + "memory(GiB)": 111.15, + "step": 55465, + "train_speed(iter/s)": 0.44742 + }, + { + "acc": 0.64029555, + "epoch": 1.4071537290715372, + "grad_norm": 5.6875, + "learning_rate": 2.215939660133986e-06, + "loss": 1.62950211, + "memory(GiB)": 111.15, + "step": 55470, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.65946722, + "epoch": 1.4072805682394725, + "grad_norm": 5.375, + "learning_rate": 2.215068694165883e-06, + "loss": 1.61622353, + "memory(GiB)": 111.15, + "step": 55475, + "train_speed(iter/s)": 0.447463 + }, + { + "acc": 0.65615654, + "epoch": 1.4074074074074074, + "grad_norm": 5.03125, + "learning_rate": 2.2141978506914922e-06, + "loss": 1.52622824, + "memory(GiB)": 111.15, + "step": 55480, + "train_speed(iter/s)": 0.447485 + }, + { + "acc": 0.63226757, + "epoch": 1.4075342465753424, + "grad_norm": 5.25, + "learning_rate": 2.2133271297491165e-06, + "loss": 1.74187679, + "memory(GiB)": 111.15, + "step": 55485, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.65884361, + "epoch": 1.4076610857432774, + "grad_norm": 5.40625, + "learning_rate": 2.2124565313770584e-06, + "loss": 1.59279099, + "memory(GiB)": 111.15, + "step": 55490, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.65984268, + "epoch": 1.4077879249112126, + "grad_norm": 4.75, + "learning_rate": 2.211586055613606e-06, + "loss": 1.58330212, + "memory(GiB)": 111.15, + "step": 55495, + "train_speed(iter/s)": 0.44755 + }, + { + "acc": 0.64155674, + "epoch": 1.4079147640791476, + "grad_norm": 4.84375, + "learning_rate": 2.210715702497046e-06, + "loss": 1.61848145, + "memory(GiB)": 111.15, + "step": 55500, + "train_speed(iter/s)": 0.447572 + }, + { + "epoch": 1.4079147640791476, + "eval_acc": 0.6434126241761107, + "eval_loss": 1.6075310707092285, + "eval_runtime": 113.584, + "eval_samples_per_second": 56.082, + "eval_steps_per_second": 28.041, + "step": 55500 + }, + { + "acc": 0.65825548, + "epoch": 1.4080416032470828, + "grad_norm": 5.84375, + "learning_rate": 2.2098454720656647e-06, + "loss": 1.65343399, + "memory(GiB)": 111.15, + "step": 55505, + "train_speed(iter/s)": 0.447167 + }, + { + "acc": 0.65523133, + "epoch": 1.4081684424150178, + "grad_norm": 6.40625, + "learning_rate": 2.2089753643577384e-06, + "loss": 1.60590897, + "memory(GiB)": 111.15, + "step": 55510, + "train_speed(iter/s)": 0.447189 + }, + { + "acc": 0.65470314, + "epoch": 1.4082952815829528, + "grad_norm": 5.15625, + "learning_rate": 2.208105379411535e-06, + "loss": 1.58342228, + "memory(GiB)": 111.15, + "step": 55515, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.64794884, + "epoch": 1.4084221207508878, + "grad_norm": 5.5, + "learning_rate": 2.2072355172653197e-06, + "loss": 1.56788177, + "memory(GiB)": 111.15, + "step": 55520, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.64734392, + "epoch": 1.408548959918823, + "grad_norm": 5.59375, + "learning_rate": 2.2063657779573573e-06, + "loss": 1.58637848, + "memory(GiB)": 111.15, + "step": 55525, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.65586443, + "epoch": 1.408675799086758, + "grad_norm": 5.3125, + "learning_rate": 2.2054961615259023e-06, + "loss": 1.59346323, + "memory(GiB)": 111.15, + "step": 55530, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.66836076, + "epoch": 1.408802638254693, + "grad_norm": 4.6875, + "learning_rate": 2.2046266680091994e-06, + "loss": 1.54648476, + "memory(GiB)": 111.15, + "step": 55535, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.64313984, + "epoch": 1.4089294774226282, + "grad_norm": 5.3125, + "learning_rate": 2.2037572974454974e-06, + "loss": 1.71497726, + "memory(GiB)": 111.15, + "step": 55540, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.65777221, + "epoch": 1.4090563165905632, + "grad_norm": 5.65625, + "learning_rate": 2.202888049873034e-06, + "loss": 1.60096874, + "memory(GiB)": 111.15, + "step": 55545, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.64226389, + "epoch": 1.4091831557584982, + "grad_norm": 5.9375, + "learning_rate": 2.2020189253300428e-06, + "loss": 1.75291557, + "memory(GiB)": 111.15, + "step": 55550, + "train_speed(iter/s)": 0.447362 + }, + { + "acc": 0.64295897, + "epoch": 1.4093099949264332, + "grad_norm": 5.1875, + "learning_rate": 2.2011499238547506e-06, + "loss": 1.67860756, + "memory(GiB)": 111.15, + "step": 55555, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.64488106, + "epoch": 1.4094368340943684, + "grad_norm": 6.625, + "learning_rate": 2.2002810454853813e-06, + "loss": 1.68024406, + "memory(GiB)": 111.15, + "step": 55560, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.6506753, + "epoch": 1.4095636732623034, + "grad_norm": 6.21875, + "learning_rate": 2.1994122902601513e-06, + "loss": 1.62848759, + "memory(GiB)": 111.15, + "step": 55565, + "train_speed(iter/s)": 0.447427 + }, + { + "acc": 0.64538755, + "epoch": 1.4096905124302386, + "grad_norm": 5.40625, + "learning_rate": 2.1985436582172724e-06, + "loss": 1.59540272, + "memory(GiB)": 111.15, + "step": 55570, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.64573588, + "epoch": 1.4098173515981736, + "grad_norm": 5.25, + "learning_rate": 2.1976751493949512e-06, + "loss": 1.66329079, + "memory(GiB)": 111.15, + "step": 55575, + "train_speed(iter/s)": 0.44745 + }, + { + "acc": 0.64529719, + "epoch": 1.4099441907661086, + "grad_norm": 5.875, + "learning_rate": 2.196806763831388e-06, + "loss": 1.60727959, + "memory(GiB)": 111.15, + "step": 55580, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.65486465, + "epoch": 1.4100710299340435, + "grad_norm": 5.3125, + "learning_rate": 2.1959385015647794e-06, + "loss": 1.65519657, + "memory(GiB)": 111.15, + "step": 55585, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.64587941, + "epoch": 1.4101978691019788, + "grad_norm": 4.96875, + "learning_rate": 2.195070362633314e-06, + "loss": 1.63699932, + "memory(GiB)": 111.15, + "step": 55590, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.65396824, + "epoch": 1.4103247082699137, + "grad_norm": 6.625, + "learning_rate": 2.194202347075178e-06, + "loss": 1.64205894, + "memory(GiB)": 111.15, + "step": 55595, + "train_speed(iter/s)": 0.447536 + }, + { + "acc": 0.64511509, + "epoch": 1.4104515474378487, + "grad_norm": 5.53125, + "learning_rate": 2.1933344549285485e-06, + "loss": 1.67041264, + "memory(GiB)": 111.15, + "step": 55600, + "train_speed(iter/s)": 0.447558 + }, + { + "epoch": 1.4104515474378487, + "eval_acc": 0.643418472918443, + "eval_loss": 1.6074844598770142, + "eval_runtime": 112.9386, + "eval_samples_per_second": 56.402, + "eval_steps_per_second": 28.201, + "step": 55600 + }, + { + "acc": 0.64482756, + "epoch": 1.410578386605784, + "grad_norm": 6.8125, + "learning_rate": 2.1924666862316015e-06, + "loss": 1.65318069, + "memory(GiB)": 111.15, + "step": 55605, + "train_speed(iter/s)": 0.447156 + }, + { + "acc": 0.66012735, + "epoch": 1.410705225773719, + "grad_norm": 5.8125, + "learning_rate": 2.191599041022504e-06, + "loss": 1.60366592, + "memory(GiB)": 111.15, + "step": 55610, + "train_speed(iter/s)": 0.447178 + }, + { + "acc": 0.64893961, + "epoch": 1.410832064941654, + "grad_norm": 4.9375, + "learning_rate": 2.19073151933942e-06, + "loss": 1.54841022, + "memory(GiB)": 111.15, + "step": 55615, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.64925661, + "epoch": 1.410958904109589, + "grad_norm": 5.40625, + "learning_rate": 2.1898641212205053e-06, + "loss": 1.59838352, + "memory(GiB)": 111.15, + "step": 55620, + "train_speed(iter/s)": 0.447221 + }, + { + "acc": 0.63824735, + "epoch": 1.4110857432775241, + "grad_norm": 5.09375, + "learning_rate": 2.1889968467039114e-06, + "loss": 1.64966335, + "memory(GiB)": 111.15, + "step": 55625, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.64872193, + "epoch": 1.411212582445459, + "grad_norm": 5.3125, + "learning_rate": 2.1881296958277897e-06, + "loss": 1.61591339, + "memory(GiB)": 111.15, + "step": 55630, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.66292744, + "epoch": 1.4113394216133943, + "grad_norm": 5.59375, + "learning_rate": 2.1872626686302767e-06, + "loss": 1.56117191, + "memory(GiB)": 111.15, + "step": 55635, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.65988712, + "epoch": 1.4114662607813293, + "grad_norm": 5.90625, + "learning_rate": 2.186395765149508e-06, + "loss": 1.52137938, + "memory(GiB)": 111.15, + "step": 55640, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.64180212, + "epoch": 1.4115930999492643, + "grad_norm": 5.1875, + "learning_rate": 2.1855289854236165e-06, + "loss": 1.68515625, + "memory(GiB)": 111.15, + "step": 55645, + "train_speed(iter/s)": 0.44733 + }, + { + "acc": 0.6556931, + "epoch": 1.4117199391171993, + "grad_norm": 4.8125, + "learning_rate": 2.184662329490728e-06, + "loss": 1.65344219, + "memory(GiB)": 111.15, + "step": 55650, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.6523912, + "epoch": 1.4118467782851345, + "grad_norm": 5.8125, + "learning_rate": 2.1837957973889584e-06, + "loss": 1.62503223, + "memory(GiB)": 111.15, + "step": 55655, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.63564577, + "epoch": 1.4119736174530695, + "grad_norm": 6.25, + "learning_rate": 2.1829293891564212e-06, + "loss": 1.68575764, + "memory(GiB)": 111.15, + "step": 55660, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.64855986, + "epoch": 1.4121004566210047, + "grad_norm": 5.125, + "learning_rate": 2.182063104831228e-06, + "loss": 1.61404915, + "memory(GiB)": 111.15, + "step": 55665, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.65448475, + "epoch": 1.4122272957889397, + "grad_norm": 4.9375, + "learning_rate": 2.181196944451483e-06, + "loss": 1.59914274, + "memory(GiB)": 111.15, + "step": 55670, + "train_speed(iter/s)": 0.447439 + }, + { + "acc": 0.6529541, + "epoch": 1.4123541349568747, + "grad_norm": 6.0, + "learning_rate": 2.180330908055278e-06, + "loss": 1.69774704, + "memory(GiB)": 111.15, + "step": 55675, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.64324355, + "epoch": 1.4124809741248097, + "grad_norm": 5.09375, + "learning_rate": 2.1794649956807094e-06, + "loss": 1.61355476, + "memory(GiB)": 111.15, + "step": 55680, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.65975399, + "epoch": 1.4126078132927449, + "grad_norm": 5.5625, + "learning_rate": 2.178599207365864e-06, + "loss": 1.52956505, + "memory(GiB)": 111.15, + "step": 55685, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.6630415, + "epoch": 1.4127346524606799, + "grad_norm": 4.65625, + "learning_rate": 2.1777335431488234e-06, + "loss": 1.58070717, + "memory(GiB)": 111.15, + "step": 55690, + "train_speed(iter/s)": 0.447526 + }, + { + "acc": 0.66378756, + "epoch": 1.4128614916286149, + "grad_norm": 5.625, + "learning_rate": 2.176868003067659e-06, + "loss": 1.63196793, + "memory(GiB)": 111.15, + "step": 55695, + "train_speed(iter/s)": 0.447548 + }, + { + "acc": 0.65093727, + "epoch": 1.41298833079655, + "grad_norm": 5.5, + "learning_rate": 2.1760025871604456e-06, + "loss": 1.62896614, + "memory(GiB)": 111.15, + "step": 55700, + "train_speed(iter/s)": 0.447569 + }, + { + "epoch": 1.41298833079655, + "eval_acc": 0.6433340839219327, + "eval_loss": 1.6075607538223267, + "eval_runtime": 113.0098, + "eval_samples_per_second": 56.367, + "eval_steps_per_second": 28.183, + "step": 55700 + }, + { + "acc": 0.65946612, + "epoch": 1.413115169964485, + "grad_norm": 6.3125, + "learning_rate": 2.175137295465247e-06, + "loss": 1.6286274, + "memory(GiB)": 111.15, + "step": 55705, + "train_speed(iter/s)": 0.447168 + }, + { + "acc": 0.65116796, + "epoch": 1.41324200913242, + "grad_norm": 4.53125, + "learning_rate": 2.174272128020122e-06, + "loss": 1.68964424, + "memory(GiB)": 111.15, + "step": 55710, + "train_speed(iter/s)": 0.447189 + }, + { + "acc": 0.66525183, + "epoch": 1.413368848300355, + "grad_norm": 5.8125, + "learning_rate": 2.1734070848631245e-06, + "loss": 1.59700413, + "memory(GiB)": 111.15, + "step": 55715, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.65805111, + "epoch": 1.4134956874682902, + "grad_norm": 5.96875, + "learning_rate": 2.172542166032303e-06, + "loss": 1.54883137, + "memory(GiB)": 111.15, + "step": 55720, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.64495516, + "epoch": 1.4136225266362252, + "grad_norm": 5.0625, + "learning_rate": 2.171677371565701e-06, + "loss": 1.6520462, + "memory(GiB)": 111.15, + "step": 55725, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.65134163, + "epoch": 1.4137493658041604, + "grad_norm": 5.8125, + "learning_rate": 2.1708127015013565e-06, + "loss": 1.66394157, + "memory(GiB)": 111.15, + "step": 55730, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.64003301, + "epoch": 1.4138762049720954, + "grad_norm": 5.21875, + "learning_rate": 2.1699481558773e-06, + "loss": 1.61516914, + "memory(GiB)": 111.15, + "step": 55735, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.64439421, + "epoch": 1.4140030441400304, + "grad_norm": 4.21875, + "learning_rate": 2.1690837347315597e-06, + "loss": 1.63409767, + "memory(GiB)": 111.15, + "step": 55740, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.66308403, + "epoch": 1.4141298833079654, + "grad_norm": 6.1875, + "learning_rate": 2.168219438102155e-06, + "loss": 1.59377232, + "memory(GiB)": 111.15, + "step": 55745, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.64211521, + "epoch": 1.4142567224759006, + "grad_norm": 6.375, + "learning_rate": 2.167355266027103e-06, + "loss": 1.62712307, + "memory(GiB)": 111.15, + "step": 55750, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.6564775, + "epoch": 1.4143835616438356, + "grad_norm": 5.34375, + "learning_rate": 2.1664912185444127e-06, + "loss": 1.55020275, + "memory(GiB)": 111.15, + "step": 55755, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.6486495, + "epoch": 1.4145104008117706, + "grad_norm": 4.90625, + "learning_rate": 2.16562729569209e-06, + "loss": 1.71447601, + "memory(GiB)": 111.15, + "step": 55760, + "train_speed(iter/s)": 0.447404 + }, + { + "acc": 0.65903311, + "epoch": 1.4146372399797058, + "grad_norm": 5.28125, + "learning_rate": 2.164763497508131e-06, + "loss": 1.56397743, + "memory(GiB)": 111.15, + "step": 55765, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.66544304, + "epoch": 1.4147640791476408, + "grad_norm": 5.5, + "learning_rate": 2.1638998240305355e-06, + "loss": 1.58231039, + "memory(GiB)": 111.15, + "step": 55770, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.63512707, + "epoch": 1.4148909183155758, + "grad_norm": 5.75, + "learning_rate": 2.163036275297286e-06, + "loss": 1.66073227, + "memory(GiB)": 111.15, + "step": 55775, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.65333109, + "epoch": 1.4150177574835108, + "grad_norm": 5.25, + "learning_rate": 2.162172851346368e-06, + "loss": 1.59874229, + "memory(GiB)": 111.15, + "step": 55780, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.66392636, + "epoch": 1.415144596651446, + "grad_norm": 5.9375, + "learning_rate": 2.1613095522157557e-06, + "loss": 1.588445, + "memory(GiB)": 111.15, + "step": 55785, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.65414338, + "epoch": 1.415271435819381, + "grad_norm": 5.03125, + "learning_rate": 2.1604463779434267e-06, + "loss": 1.58975925, + "memory(GiB)": 111.15, + "step": 55790, + "train_speed(iter/s)": 0.447533 + }, + { + "acc": 0.6469841, + "epoch": 1.4153982749873162, + "grad_norm": 5.53125, + "learning_rate": 2.159583328567342e-06, + "loss": 1.64264107, + "memory(GiB)": 111.15, + "step": 55795, + "train_speed(iter/s)": 0.447555 + }, + { + "acc": 0.65557861, + "epoch": 1.4155251141552512, + "grad_norm": 5.34375, + "learning_rate": 2.158720404125462e-06, + "loss": 1.61578388, + "memory(GiB)": 111.15, + "step": 55800, + "train_speed(iter/s)": 0.447577 + }, + { + "epoch": 1.4155251141552512, + "eval_acc": 0.6434498054666523, + "eval_loss": 1.6074860095977783, + "eval_runtime": 114.0197, + "eval_samples_per_second": 55.868, + "eval_steps_per_second": 27.934, + "step": 55800 + }, + { + "acc": 0.65538845, + "epoch": 1.4156519533231862, + "grad_norm": 5.125, + "learning_rate": 2.1578576046557463e-06, + "loss": 1.59876461, + "memory(GiB)": 111.15, + "step": 55805, + "train_speed(iter/s)": 0.447172 + }, + { + "acc": 0.64056053, + "epoch": 1.4157787924911212, + "grad_norm": 5.3125, + "learning_rate": 2.156994930196144e-06, + "loss": 1.62557125, + "memory(GiB)": 111.15, + "step": 55810, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.65917277, + "epoch": 1.4159056316590564, + "grad_norm": 5.40625, + "learning_rate": 2.156132380784594e-06, + "loss": 1.55431423, + "memory(GiB)": 111.15, + "step": 55815, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.64639225, + "epoch": 1.4160324708269914, + "grad_norm": 6.28125, + "learning_rate": 2.155269956459041e-06, + "loss": 1.62074814, + "memory(GiB)": 111.15, + "step": 55820, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.66090841, + "epoch": 1.4161593099949266, + "grad_norm": 5.53125, + "learning_rate": 2.1544076572574156e-06, + "loss": 1.60202198, + "memory(GiB)": 111.15, + "step": 55825, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.64345016, + "epoch": 1.4162861491628616, + "grad_norm": 4.5, + "learning_rate": 2.1535454832176482e-06, + "loss": 1.62321281, + "memory(GiB)": 111.15, + "step": 55830, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.63962955, + "epoch": 1.4164129883307965, + "grad_norm": 5.65625, + "learning_rate": 2.1526834343776556e-06, + "loss": 1.6854557, + "memory(GiB)": 111.15, + "step": 55835, + "train_speed(iter/s)": 0.4473 + }, + { + "acc": 0.62745762, + "epoch": 1.4165398274987315, + "grad_norm": 7.46875, + "learning_rate": 2.1518215107753593e-06, + "loss": 1.69691544, + "memory(GiB)": 111.15, + "step": 55840, + "train_speed(iter/s)": 0.447321 + }, + { + "acc": 0.63900466, + "epoch": 1.4166666666666667, + "grad_norm": 5.3125, + "learning_rate": 2.1509597124486693e-06, + "loss": 1.65631008, + "memory(GiB)": 111.15, + "step": 55845, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.66439447, + "epoch": 1.4167935058346017, + "grad_norm": 4.875, + "learning_rate": 2.1500980394354907e-06, + "loss": 1.60967579, + "memory(GiB)": 111.15, + "step": 55850, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.64850578, + "epoch": 1.4169203450025367, + "grad_norm": 5.0, + "learning_rate": 2.1492364917737252e-06, + "loss": 1.6218523, + "memory(GiB)": 111.15, + "step": 55855, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.65642681, + "epoch": 1.417047184170472, + "grad_norm": 5.625, + "learning_rate": 2.148375069501266e-06, + "loss": 1.60242004, + "memory(GiB)": 111.15, + "step": 55860, + "train_speed(iter/s)": 0.447405 + }, + { + "acc": 0.65095196, + "epoch": 1.417174023338407, + "grad_norm": 4.1875, + "learning_rate": 2.147513772656003e-06, + "loss": 1.60890732, + "memory(GiB)": 111.15, + "step": 55865, + "train_speed(iter/s)": 0.447427 + }, + { + "acc": 0.63584042, + "epoch": 1.417300862506342, + "grad_norm": 6.78125, + "learning_rate": 2.1466526012758194e-06, + "loss": 1.6269907, + "memory(GiB)": 111.15, + "step": 55870, + "train_speed(iter/s)": 0.447448 + }, + { + "acc": 0.65745797, + "epoch": 1.417427701674277, + "grad_norm": 5.09375, + "learning_rate": 2.145791555398594e-06, + "loss": 1.57455444, + "memory(GiB)": 111.15, + "step": 55875, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.65565324, + "epoch": 1.417554540842212, + "grad_norm": 5.90625, + "learning_rate": 2.144930635062199e-06, + "loss": 1.5324192, + "memory(GiB)": 111.15, + "step": 55880, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.63588448, + "epoch": 1.417681380010147, + "grad_norm": 5.875, + "learning_rate": 2.144069840304502e-06, + "loss": 1.69857292, + "memory(GiB)": 111.15, + "step": 55885, + "train_speed(iter/s)": 0.447511 + }, + { + "acc": 0.65857897, + "epoch": 1.4178082191780823, + "grad_norm": 5.71875, + "learning_rate": 2.1432091711633634e-06, + "loss": 1.62534294, + "memory(GiB)": 111.15, + "step": 55890, + "train_speed(iter/s)": 0.447532 + }, + { + "acc": 0.65040827, + "epoch": 1.4179350583460173, + "grad_norm": 5.375, + "learning_rate": 2.142348627676641e-06, + "loss": 1.62272053, + "memory(GiB)": 111.15, + "step": 55895, + "train_speed(iter/s)": 0.447553 + }, + { + "acc": 0.65005922, + "epoch": 1.4180618975139523, + "grad_norm": 5.28125, + "learning_rate": 2.1414882098821836e-06, + "loss": 1.60534019, + "memory(GiB)": 111.15, + "step": 55900, + "train_speed(iter/s)": 0.447574 + }, + { + "epoch": 1.4180618975139523, + "eval_acc": 0.6434046865972309, + "eval_loss": 1.6075425148010254, + "eval_runtime": 113.8189, + "eval_samples_per_second": 55.966, + "eval_steps_per_second": 27.983, + "step": 55900 + }, + { + "acc": 0.65743289, + "epoch": 1.4181887366818873, + "grad_norm": 5.4375, + "learning_rate": 2.1406279178178355e-06, + "loss": 1.64108944, + "memory(GiB)": 111.15, + "step": 55905, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.652671, + "epoch": 1.4183155758498225, + "grad_norm": 5.71875, + "learning_rate": 2.1397677515214422e-06, + "loss": 1.63358154, + "memory(GiB)": 111.15, + "step": 55910, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.64822826, + "epoch": 1.4184424150177575, + "grad_norm": 6.5625, + "learning_rate": 2.1389077110308304e-06, + "loss": 1.662854, + "memory(GiB)": 111.15, + "step": 55915, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.64673843, + "epoch": 1.4185692541856925, + "grad_norm": 6.25, + "learning_rate": 2.138047796383832e-06, + "loss": 1.66949406, + "memory(GiB)": 111.15, + "step": 55920, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.64664612, + "epoch": 1.4186960933536277, + "grad_norm": 7.4375, + "learning_rate": 2.1371880076182666e-06, + "loss": 1.67883682, + "memory(GiB)": 111.15, + "step": 55925, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.63709292, + "epoch": 1.4188229325215627, + "grad_norm": 6.03125, + "learning_rate": 2.1363283447719584e-06, + "loss": 1.67796459, + "memory(GiB)": 111.15, + "step": 55930, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.65979586, + "epoch": 1.4189497716894977, + "grad_norm": 6.90625, + "learning_rate": 2.135468807882713e-06, + "loss": 1.59852047, + "memory(GiB)": 111.15, + "step": 55935, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.64526138, + "epoch": 1.4190766108574326, + "grad_norm": 5.9375, + "learning_rate": 2.1346093969883367e-06, + "loss": 1.63534527, + "memory(GiB)": 111.15, + "step": 55940, + "train_speed(iter/s)": 0.447321 + }, + { + "acc": 0.6540925, + "epoch": 1.4192034500253679, + "grad_norm": 5.34375, + "learning_rate": 2.1337501121266345e-06, + "loss": 1.59941463, + "memory(GiB)": 111.15, + "step": 55945, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.63602238, + "epoch": 1.4193302891933028, + "grad_norm": 6.46875, + "learning_rate": 2.132890953335401e-06, + "loss": 1.68423481, + "memory(GiB)": 111.15, + "step": 55950, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.64360447, + "epoch": 1.419457128361238, + "grad_norm": 5.28125, + "learning_rate": 2.1320319206524215e-06, + "loss": 1.70478153, + "memory(GiB)": 111.15, + "step": 55955, + "train_speed(iter/s)": 0.447385 + }, + { + "acc": 0.65257988, + "epoch": 1.419583967529173, + "grad_norm": 4.8125, + "learning_rate": 2.1311730141154813e-06, + "loss": 1.62999821, + "memory(GiB)": 111.15, + "step": 55960, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.65277996, + "epoch": 1.419710806697108, + "grad_norm": 4.9375, + "learning_rate": 2.1303142337623623e-06, + "loss": 1.64737759, + "memory(GiB)": 111.15, + "step": 55965, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.65106091, + "epoch": 1.419837645865043, + "grad_norm": 5.90625, + "learning_rate": 2.1294555796308375e-06, + "loss": 1.58642406, + "memory(GiB)": 111.15, + "step": 55970, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.65375066, + "epoch": 1.4199644850329782, + "grad_norm": 5.28125, + "learning_rate": 2.1285970517586686e-06, + "loss": 1.70657806, + "memory(GiB)": 111.15, + "step": 55975, + "train_speed(iter/s)": 0.447471 + }, + { + "acc": 0.6387711, + "epoch": 1.4200913242009132, + "grad_norm": 5.25, + "learning_rate": 2.127738650183623e-06, + "loss": 1.66443481, + "memory(GiB)": 111.15, + "step": 55980, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.64868693, + "epoch": 1.4202181633688484, + "grad_norm": 5.09375, + "learning_rate": 2.1268803749434546e-06, + "loss": 1.6312603, + "memory(GiB)": 111.15, + "step": 55985, + "train_speed(iter/s)": 0.447513 + }, + { + "acc": 0.62956281, + "epoch": 1.4203450025367834, + "grad_norm": 6.4375, + "learning_rate": 2.1260222260759158e-06, + "loss": 1.60396194, + "memory(GiB)": 111.15, + "step": 55990, + "train_speed(iter/s)": 0.447535 + }, + { + "acc": 0.65884132, + "epoch": 1.4204718417047184, + "grad_norm": 5.21875, + "learning_rate": 2.1251642036187502e-06, + "loss": 1.62932148, + "memory(GiB)": 111.15, + "step": 55995, + "train_speed(iter/s)": 0.447556 + }, + { + "acc": 0.66040487, + "epoch": 1.4205986808726534, + "grad_norm": 5.28125, + "learning_rate": 2.124306307609699e-06, + "loss": 1.63530197, + "memory(GiB)": 111.15, + "step": 56000, + "train_speed(iter/s)": 0.447577 + }, + { + "epoch": 1.4205986808726534, + "eval_acc": 0.6434055221318499, + "eval_loss": 1.607498049736023, + "eval_runtime": 113.6952, + "eval_samples_per_second": 56.027, + "eval_steps_per_second": 28.014, + "step": 56000 + }, + { + "acc": 0.62808676, + "epoch": 1.4207255200405886, + "grad_norm": 5.46875, + "learning_rate": 2.123448538086495e-06, + "loss": 1.70467453, + "memory(GiB)": 111.15, + "step": 56005, + "train_speed(iter/s)": 0.447175 + }, + { + "acc": 0.64478292, + "epoch": 1.4208523592085236, + "grad_norm": 6.34375, + "learning_rate": 2.122590895086867e-06, + "loss": 1.70288353, + "memory(GiB)": 111.15, + "step": 56010, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.64343605, + "epoch": 1.4209791983764586, + "grad_norm": 5.5625, + "learning_rate": 2.1217333786485385e-06, + "loss": 1.60869198, + "memory(GiB)": 111.15, + "step": 56015, + "train_speed(iter/s)": 0.447218 + }, + { + "acc": 0.6472868, + "epoch": 1.4211060375443938, + "grad_norm": 5.6875, + "learning_rate": 2.120875988809226e-06, + "loss": 1.64239616, + "memory(GiB)": 111.15, + "step": 56020, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.64988785, + "epoch": 1.4212328767123288, + "grad_norm": 5.8125, + "learning_rate": 2.1200187256066425e-06, + "loss": 1.59436636, + "memory(GiB)": 111.15, + "step": 56025, + "train_speed(iter/s)": 0.447261 + }, + { + "acc": 0.66384902, + "epoch": 1.4213597158802638, + "grad_norm": 6.09375, + "learning_rate": 2.119161589078493e-06, + "loss": 1.61093311, + "memory(GiB)": 111.15, + "step": 56030, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.63542433, + "epoch": 1.4214865550481988, + "grad_norm": 5.34375, + "learning_rate": 2.118304579262479e-06, + "loss": 1.70800896, + "memory(GiB)": 111.15, + "step": 56035, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.66646814, + "epoch": 1.421613394216134, + "grad_norm": 4.6875, + "learning_rate": 2.1174476961962957e-06, + "loss": 1.6258667, + "memory(GiB)": 111.15, + "step": 56040, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.64053922, + "epoch": 1.421740233384069, + "grad_norm": 5.875, + "learning_rate": 2.1165909399176328e-06, + "loss": 1.67121792, + "memory(GiB)": 111.15, + "step": 56045, + "train_speed(iter/s)": 0.447346 + }, + { + "acc": 0.63953443, + "epoch": 1.4218670725520042, + "grad_norm": 5.125, + "learning_rate": 2.1157343104641733e-06, + "loss": 1.7051506, + "memory(GiB)": 111.15, + "step": 56050, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.65037708, + "epoch": 1.4219939117199392, + "grad_norm": 5.125, + "learning_rate": 2.114877807873596e-06, + "loss": 1.55959854, + "memory(GiB)": 111.15, + "step": 56055, + "train_speed(iter/s)": 0.447388 + }, + { + "acc": 0.63940735, + "epoch": 1.4221207508878742, + "grad_norm": 5.0625, + "learning_rate": 2.114021432183574e-06, + "loss": 1.61257248, + "memory(GiB)": 111.15, + "step": 56060, + "train_speed(iter/s)": 0.44741 + }, + { + "acc": 0.64273763, + "epoch": 1.4222475900558091, + "grad_norm": 5.0, + "learning_rate": 2.113165183431773e-06, + "loss": 1.64214554, + "memory(GiB)": 111.15, + "step": 56065, + "train_speed(iter/s)": 0.447431 + }, + { + "acc": 0.66959219, + "epoch": 1.4223744292237444, + "grad_norm": 6.3125, + "learning_rate": 2.112309061655859e-06, + "loss": 1.57030935, + "memory(GiB)": 111.15, + "step": 56070, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.67592497, + "epoch": 1.4225012683916793, + "grad_norm": 5.8125, + "learning_rate": 2.1114530668934836e-06, + "loss": 1.58166075, + "memory(GiB)": 111.15, + "step": 56075, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.63672018, + "epoch": 1.4226281075596143, + "grad_norm": 5.53125, + "learning_rate": 2.1105971991822966e-06, + "loss": 1.65288696, + "memory(GiB)": 111.15, + "step": 56080, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.63924494, + "epoch": 1.4227549467275495, + "grad_norm": 5.96875, + "learning_rate": 2.1097414585599474e-06, + "loss": 1.67434578, + "memory(GiB)": 111.15, + "step": 56085, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.65964413, + "epoch": 1.4228817858954845, + "grad_norm": 5.71875, + "learning_rate": 2.1088858450640743e-06, + "loss": 1.64543381, + "memory(GiB)": 111.15, + "step": 56090, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.65669146, + "epoch": 1.4230086250634195, + "grad_norm": 6.65625, + "learning_rate": 2.108030358732308e-06, + "loss": 1.56034384, + "memory(GiB)": 111.15, + "step": 56095, + "train_speed(iter/s)": 0.447559 + }, + { + "acc": 0.65065193, + "epoch": 1.4231354642313545, + "grad_norm": 4.9375, + "learning_rate": 2.107174999602277e-06, + "loss": 1.66710873, + "memory(GiB)": 111.15, + "step": 56100, + "train_speed(iter/s)": 0.447581 + }, + { + "epoch": 1.4231354642313545, + "eval_acc": 0.6434088642703255, + "eval_loss": 1.6075249910354614, + "eval_runtime": 113.7494, + "eval_samples_per_second": 56.0, + "eval_steps_per_second": 28.0, + "step": 56100 + }, + { + "acc": 0.66101942, + "epoch": 1.4232623033992897, + "grad_norm": 5.84375, + "learning_rate": 2.106319767711606e-06, + "loss": 1.59294548, + "memory(GiB)": 111.15, + "step": 56105, + "train_speed(iter/s)": 0.447179 + }, + { + "acc": 0.64050493, + "epoch": 1.4233891425672247, + "grad_norm": 6.15625, + "learning_rate": 2.105464663097913e-06, + "loss": 1.66906052, + "memory(GiB)": 111.15, + "step": 56110, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.66142788, + "epoch": 1.42351598173516, + "grad_norm": 5.5625, + "learning_rate": 2.1046096857988033e-06, + "loss": 1.57787371, + "memory(GiB)": 111.15, + "step": 56115, + "train_speed(iter/s)": 0.447221 + }, + { + "acc": 0.64583778, + "epoch": 1.423642820903095, + "grad_norm": 4.75, + "learning_rate": 2.103754835851889e-06, + "loss": 1.61744537, + "memory(GiB)": 111.15, + "step": 56120, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.66171122, + "epoch": 1.42376966007103, + "grad_norm": 5.21875, + "learning_rate": 2.102900113294768e-06, + "loss": 1.58314018, + "memory(GiB)": 111.15, + "step": 56125, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.64692249, + "epoch": 1.4238964992389649, + "grad_norm": 5.1875, + "learning_rate": 2.1020455181650356e-06, + "loss": 1.61493015, + "memory(GiB)": 111.15, + "step": 56130, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.65494623, + "epoch": 1.4240233384069, + "grad_norm": 5.75, + "learning_rate": 2.101191050500277e-06, + "loss": 1.61068916, + "memory(GiB)": 111.15, + "step": 56135, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.63782477, + "epoch": 1.424150177574835, + "grad_norm": 5.84375, + "learning_rate": 2.1003367103380797e-06, + "loss": 1.64089012, + "memory(GiB)": 111.15, + "step": 56140, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.65194902, + "epoch": 1.4242770167427703, + "grad_norm": 6.3125, + "learning_rate": 2.09948249771602e-06, + "loss": 1.69724331, + "memory(GiB)": 111.15, + "step": 56145, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.65150332, + "epoch": 1.4244038559107053, + "grad_norm": 6.15625, + "learning_rate": 2.09862841267167e-06, + "loss": 1.55631924, + "memory(GiB)": 111.15, + "step": 56150, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.64779973, + "epoch": 1.4245306950786403, + "grad_norm": 5.90625, + "learning_rate": 2.097774455242596e-06, + "loss": 1.63004532, + "memory(GiB)": 111.15, + "step": 56155, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.64808569, + "epoch": 1.4246575342465753, + "grad_norm": 5.25, + "learning_rate": 2.096920625466359e-06, + "loss": 1.59002266, + "memory(GiB)": 111.15, + "step": 56160, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.63841166, + "epoch": 1.4247843734145105, + "grad_norm": 5.375, + "learning_rate": 2.0960669233805136e-06, + "loss": 1.68767281, + "memory(GiB)": 111.15, + "step": 56165, + "train_speed(iter/s)": 0.447433 + }, + { + "acc": 0.65113573, + "epoch": 1.4249112125824455, + "grad_norm": 4.5625, + "learning_rate": 2.0952133490226106e-06, + "loss": 1.69092484, + "memory(GiB)": 111.15, + "step": 56170, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.6470686, + "epoch": 1.4250380517503805, + "grad_norm": 5.96875, + "learning_rate": 2.0943599024301935e-06, + "loss": 1.66196938, + "memory(GiB)": 111.15, + "step": 56175, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.65314741, + "epoch": 1.4251648909183157, + "grad_norm": 6.0, + "learning_rate": 2.093506583640801e-06, + "loss": 1.57658138, + "memory(GiB)": 111.15, + "step": 56180, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.65054379, + "epoch": 1.4252917300862507, + "grad_norm": 5.59375, + "learning_rate": 2.092653392691965e-06, + "loss": 1.5835083, + "memory(GiB)": 111.15, + "step": 56185, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.65392218, + "epoch": 1.4254185692541856, + "grad_norm": 5.34375, + "learning_rate": 2.091800329621213e-06, + "loss": 1.62041016, + "memory(GiB)": 111.15, + "step": 56190, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.66769147, + "epoch": 1.4255454084221206, + "grad_norm": 4.75, + "learning_rate": 2.0909473944660667e-06, + "loss": 1.57650433, + "memory(GiB)": 111.15, + "step": 56195, + "train_speed(iter/s)": 0.44756 + }, + { + "acc": 0.64713049, + "epoch": 1.4256722475900558, + "grad_norm": 4.8125, + "learning_rate": 2.0900945872640427e-06, + "loss": 1.6606987, + "memory(GiB)": 111.15, + "step": 56200, + "train_speed(iter/s)": 0.447581 + }, + { + "epoch": 1.4256722475900558, + "eval_acc": 0.6434786314110049, + "eval_loss": 1.6075265407562256, + "eval_runtime": 113.8713, + "eval_samples_per_second": 55.94, + "eval_steps_per_second": 27.97, + "step": 56200 + }, + { + "acc": 0.66239533, + "epoch": 1.4257990867579908, + "grad_norm": 4.75, + "learning_rate": 2.0892419080526484e-06, + "loss": 1.5885891, + "memory(GiB)": 111.15, + "step": 56205, + "train_speed(iter/s)": 0.44718 + }, + { + "acc": 0.6594717, + "epoch": 1.425925925925926, + "grad_norm": 5.1875, + "learning_rate": 2.0883893568693935e-06, + "loss": 1.61490097, + "memory(GiB)": 111.15, + "step": 56210, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.66051083, + "epoch": 1.426052765093861, + "grad_norm": 5.09375, + "learning_rate": 2.087536933751773e-06, + "loss": 1.63276901, + "memory(GiB)": 111.15, + "step": 56215, + "train_speed(iter/s)": 0.447222 + }, + { + "acc": 0.65409012, + "epoch": 1.426179604261796, + "grad_norm": 7.0, + "learning_rate": 2.0866846387372814e-06, + "loss": 1.65277519, + "memory(GiB)": 111.15, + "step": 56220, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.63516855, + "epoch": 1.426306443429731, + "grad_norm": 5.5625, + "learning_rate": 2.085832471863404e-06, + "loss": 1.72760124, + "memory(GiB)": 111.15, + "step": 56225, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.64074574, + "epoch": 1.4264332825976662, + "grad_norm": 5.75, + "learning_rate": 2.08498043316763e-06, + "loss": 1.67741909, + "memory(GiB)": 111.15, + "step": 56230, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.66876621, + "epoch": 1.4265601217656012, + "grad_norm": 5.59375, + "learning_rate": 2.0841285226874282e-06, + "loss": 1.61654797, + "memory(GiB)": 111.15, + "step": 56235, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.64060631, + "epoch": 1.4266869609335362, + "grad_norm": 5.125, + "learning_rate": 2.083276740460271e-06, + "loss": 1.63308296, + "memory(GiB)": 111.15, + "step": 56240, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.65725193, + "epoch": 1.4268138001014714, + "grad_norm": 5.0, + "learning_rate": 2.082425086523627e-06, + "loss": 1.55801449, + "memory(GiB)": 111.15, + "step": 56245, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.65592918, + "epoch": 1.4269406392694064, + "grad_norm": 5.71875, + "learning_rate": 2.0815735609149556e-06, + "loss": 1.600807, + "memory(GiB)": 111.15, + "step": 56250, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.67181673, + "epoch": 1.4270674784373414, + "grad_norm": 5.03125, + "learning_rate": 2.080722163671705e-06, + "loss": 1.58532734, + "memory(GiB)": 111.15, + "step": 56255, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.66039009, + "epoch": 1.4271943176052764, + "grad_norm": 5.1875, + "learning_rate": 2.079870894831329e-06, + "loss": 1.62940102, + "memory(GiB)": 111.15, + "step": 56260, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.64691315, + "epoch": 1.4273211567732116, + "grad_norm": 5.15625, + "learning_rate": 2.0790197544312683e-06, + "loss": 1.67488785, + "memory(GiB)": 111.15, + "step": 56265, + "train_speed(iter/s)": 0.447432 + }, + { + "acc": 0.641008, + "epoch": 1.4274479959411466, + "grad_norm": 5.0625, + "learning_rate": 2.0781687425089624e-06, + "loss": 1.65846405, + "memory(GiB)": 111.15, + "step": 56270, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.64795876, + "epoch": 1.4275748351090818, + "grad_norm": 5.78125, + "learning_rate": 2.0773178591018363e-06, + "loss": 1.66823692, + "memory(GiB)": 111.15, + "step": 56275, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.67021265, + "epoch": 1.4277016742770168, + "grad_norm": 6.0625, + "learning_rate": 2.076467104247322e-06, + "loss": 1.53036594, + "memory(GiB)": 111.15, + "step": 56280, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.66429024, + "epoch": 1.4278285134449518, + "grad_norm": 4.875, + "learning_rate": 2.0756164779828365e-06, + "loss": 1.55899296, + "memory(GiB)": 111.15, + "step": 56285, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.65059943, + "epoch": 1.4279553526128868, + "grad_norm": 4.71875, + "learning_rate": 2.0747659803457943e-06, + "loss": 1.59098063, + "memory(GiB)": 111.15, + "step": 56290, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.66356449, + "epoch": 1.428082191780822, + "grad_norm": 4.9375, + "learning_rate": 2.073915611373605e-06, + "loss": 1.57309752, + "memory(GiB)": 111.15, + "step": 56295, + "train_speed(iter/s)": 0.447559 + }, + { + "acc": 0.64546528, + "epoch": 1.428209030948757, + "grad_norm": 4.5625, + "learning_rate": 2.0730653711036713e-06, + "loss": 1.61351242, + "memory(GiB)": 111.15, + "step": 56300, + "train_speed(iter/s)": 0.44758 + }, + { + "epoch": 1.428209030948757, + "eval_acc": 0.6434401968185348, + "eval_loss": 1.6075465679168701, + "eval_runtime": 114.096, + "eval_samples_per_second": 55.83, + "eval_steps_per_second": 27.915, + "step": 56300 + }, + { + "acc": 0.65789051, + "epoch": 1.4283358701166922, + "grad_norm": 5.46875, + "learning_rate": 2.0722152595733903e-06, + "loss": 1.53278694, + "memory(GiB)": 111.15, + "step": 56305, + "train_speed(iter/s)": 0.447179 + }, + { + "acc": 0.65670028, + "epoch": 1.4284627092846272, + "grad_norm": 5.375, + "learning_rate": 2.0713652768201536e-06, + "loss": 1.60789089, + "memory(GiB)": 111.15, + "step": 56310, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.65497379, + "epoch": 1.4285895484525621, + "grad_norm": 6.59375, + "learning_rate": 2.0705154228813477e-06, + "loss": 1.61775513, + "memory(GiB)": 111.15, + "step": 56315, + "train_speed(iter/s)": 0.447222 + }, + { + "acc": 0.66225386, + "epoch": 1.4287163876204971, + "grad_norm": 4.96875, + "learning_rate": 2.0696656977943524e-06, + "loss": 1.54751701, + "memory(GiB)": 111.15, + "step": 56320, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.64495106, + "epoch": 1.4288432267884323, + "grad_norm": 5.21875, + "learning_rate": 2.068816101596543e-06, + "loss": 1.6476511, + "memory(GiB)": 111.15, + "step": 56325, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.65630407, + "epoch": 1.4289700659563673, + "grad_norm": 4.71875, + "learning_rate": 2.067966634325288e-06, + "loss": 1.59987946, + "memory(GiB)": 111.15, + "step": 56330, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.65740976, + "epoch": 1.4290969051243023, + "grad_norm": 5.84375, + "learning_rate": 2.0671172960179513e-06, + "loss": 1.59469872, + "memory(GiB)": 111.15, + "step": 56335, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.66053982, + "epoch": 1.4292237442922375, + "grad_norm": 4.90625, + "learning_rate": 2.06626808671189e-06, + "loss": 1.55951223, + "memory(GiB)": 111.15, + "step": 56340, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.65270319, + "epoch": 1.4293505834601725, + "grad_norm": 6.09375, + "learning_rate": 2.065419006444455e-06, + "loss": 1.62475777, + "memory(GiB)": 111.15, + "step": 56345, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.65622506, + "epoch": 1.4294774226281075, + "grad_norm": 5.3125, + "learning_rate": 2.0645700552529973e-06, + "loss": 1.61214867, + "memory(GiB)": 111.15, + "step": 56350, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.62975945, + "epoch": 1.4296042617960425, + "grad_norm": 5.03125, + "learning_rate": 2.063721233174853e-06, + "loss": 1.71917877, + "memory(GiB)": 111.15, + "step": 56355, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.66241837, + "epoch": 1.4297311009639777, + "grad_norm": 5.71875, + "learning_rate": 2.0628725402473574e-06, + "loss": 1.52194138, + "memory(GiB)": 111.15, + "step": 56360, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.65256386, + "epoch": 1.4298579401319127, + "grad_norm": 5.8125, + "learning_rate": 2.0620239765078404e-06, + "loss": 1.59550629, + "memory(GiB)": 111.15, + "step": 56365, + "train_speed(iter/s)": 0.447432 + }, + { + "acc": 0.64713469, + "epoch": 1.429984779299848, + "grad_norm": 5.15625, + "learning_rate": 2.0611755419936285e-06, + "loss": 1.62247086, + "memory(GiB)": 111.15, + "step": 56370, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.65388432, + "epoch": 1.430111618467783, + "grad_norm": 4.78125, + "learning_rate": 2.0603272367420357e-06, + "loss": 1.59977341, + "memory(GiB)": 111.15, + "step": 56375, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.65278945, + "epoch": 1.4302384576357179, + "grad_norm": 5.75, + "learning_rate": 2.0594790607903743e-06, + "loss": 1.60655613, + "memory(GiB)": 111.15, + "step": 56380, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.65640945, + "epoch": 1.4303652968036529, + "grad_norm": 4.625, + "learning_rate": 2.0586310141759534e-06, + "loss": 1.60628338, + "memory(GiB)": 111.15, + "step": 56385, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.64474239, + "epoch": 1.430492135971588, + "grad_norm": 5.5, + "learning_rate": 2.0577830969360753e-06, + "loss": 1.65681114, + "memory(GiB)": 111.15, + "step": 56390, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.66217451, + "epoch": 1.430618975139523, + "grad_norm": 5.75, + "learning_rate": 2.0569353091080304e-06, + "loss": 1.6247654, + "memory(GiB)": 111.15, + "step": 56395, + "train_speed(iter/s)": 0.447558 + }, + { + "acc": 0.65902281, + "epoch": 1.430745814307458, + "grad_norm": 5.46875, + "learning_rate": 2.056087650729109e-06, + "loss": 1.57480679, + "memory(GiB)": 111.15, + "step": 56400, + "train_speed(iter/s)": 0.447579 + }, + { + "epoch": 1.430745814307458, + "eval_acc": 0.6434276637992511, + "eval_loss": 1.6075377464294434, + "eval_runtime": 113.7677, + "eval_samples_per_second": 55.991, + "eval_steps_per_second": 27.996, + "step": 56400 + }, + { + "acc": 0.63944511, + "epoch": 1.4308726534753933, + "grad_norm": 5.125, + "learning_rate": 2.0552401218365975e-06, + "loss": 1.6833725, + "memory(GiB)": 111.15, + "step": 56405, + "train_speed(iter/s)": 0.44718 + }, + { + "acc": 0.64763603, + "epoch": 1.4309994926433283, + "grad_norm": 4.59375, + "learning_rate": 2.054392722467775e-06, + "loss": 1.59413548, + "memory(GiB)": 111.15, + "step": 56410, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.63542919, + "epoch": 1.4311263318112633, + "grad_norm": 5.15625, + "learning_rate": 2.0535454526599086e-06, + "loss": 1.72192993, + "memory(GiB)": 111.15, + "step": 56415, + "train_speed(iter/s)": 0.447222 + }, + { + "acc": 0.65711613, + "epoch": 1.4312531709791982, + "grad_norm": 5.8125, + "learning_rate": 2.0526983124502692e-06, + "loss": 1.55512476, + "memory(GiB)": 111.15, + "step": 56420, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.64058762, + "epoch": 1.4313800101471335, + "grad_norm": 6.15625, + "learning_rate": 2.051851301876117e-06, + "loss": 1.66548424, + "memory(GiB)": 111.15, + "step": 56425, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.64299226, + "epoch": 1.4315068493150684, + "grad_norm": 6.09375, + "learning_rate": 2.0510044209747078e-06, + "loss": 1.69187527, + "memory(GiB)": 111.15, + "step": 56430, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.64415483, + "epoch": 1.4316336884830037, + "grad_norm": 5.5, + "learning_rate": 2.05015766978329e-06, + "loss": 1.6315691, + "memory(GiB)": 111.15, + "step": 56435, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.64331279, + "epoch": 1.4317605276509386, + "grad_norm": 5.125, + "learning_rate": 2.049311048339108e-06, + "loss": 1.68607941, + "memory(GiB)": 111.15, + "step": 56440, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.64830732, + "epoch": 1.4318873668188736, + "grad_norm": 5.125, + "learning_rate": 2.0484645566793993e-06, + "loss": 1.67121162, + "memory(GiB)": 111.15, + "step": 56445, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.64098206, + "epoch": 1.4320142059868086, + "grad_norm": 5.90625, + "learning_rate": 2.0476181948413975e-06, + "loss": 1.66928539, + "memory(GiB)": 111.15, + "step": 56450, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.64654937, + "epoch": 1.4321410451547438, + "grad_norm": 5.75, + "learning_rate": 2.0467719628623293e-06, + "loss": 1.61688461, + "memory(GiB)": 111.15, + "step": 56455, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.65947962, + "epoch": 1.4322678843226788, + "grad_norm": 5.1875, + "learning_rate": 2.045925860779415e-06, + "loss": 1.56720285, + "memory(GiB)": 111.15, + "step": 56460, + "train_speed(iter/s)": 0.447411 + }, + { + "acc": 0.64383035, + "epoch": 1.432394723490614, + "grad_norm": 5.53125, + "learning_rate": 2.0450798886298707e-06, + "loss": 1.63727951, + "memory(GiB)": 111.15, + "step": 56465, + "train_speed(iter/s)": 0.447432 + }, + { + "acc": 0.65142817, + "epoch": 1.432521562658549, + "grad_norm": 5.09375, + "learning_rate": 2.044234046450905e-06, + "loss": 1.64156075, + "memory(GiB)": 111.15, + "step": 56470, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.65518799, + "epoch": 1.432648401826484, + "grad_norm": 5.15625, + "learning_rate": 2.0433883342797233e-06, + "loss": 1.56515579, + "memory(GiB)": 111.15, + "step": 56475, + "train_speed(iter/s)": 0.447474 + }, + { + "acc": 0.66218023, + "epoch": 1.432775240994419, + "grad_norm": 5.125, + "learning_rate": 2.042542752153522e-06, + "loss": 1.59656525, + "memory(GiB)": 111.15, + "step": 56480, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.65294275, + "epoch": 1.4329020801623542, + "grad_norm": 6.28125, + "learning_rate": 2.0416973001094953e-06, + "loss": 1.63570709, + "memory(GiB)": 111.15, + "step": 56485, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.6558785, + "epoch": 1.4330289193302892, + "grad_norm": 5.25, + "learning_rate": 2.0408519781848292e-06, + "loss": 1.61863956, + "memory(GiB)": 111.15, + "step": 56490, + "train_speed(iter/s)": 0.447537 + }, + { + "acc": 0.67041578, + "epoch": 1.4331557584982242, + "grad_norm": 4.53125, + "learning_rate": 2.0400067864167044e-06, + "loss": 1.54550133, + "memory(GiB)": 111.15, + "step": 56495, + "train_speed(iter/s)": 0.447558 + }, + { + "acc": 0.66338701, + "epoch": 1.4332825976661594, + "grad_norm": 5.40625, + "learning_rate": 2.0391617248422967e-06, + "loss": 1.58215618, + "memory(GiB)": 111.15, + "step": 56500, + "train_speed(iter/s)": 0.447579 + }, + { + "epoch": 1.4332825976661594, + "eval_acc": 0.6433959134837324, + "eval_loss": 1.607515811920166, + "eval_runtime": 113.4564, + "eval_samples_per_second": 56.145, + "eval_steps_per_second": 28.072, + "step": 56500 + }, + { + "acc": 0.64353447, + "epoch": 1.4334094368340944, + "grad_norm": 5.84375, + "learning_rate": 2.038316793498774e-06, + "loss": 1.66714821, + "memory(GiB)": 111.15, + "step": 56505, + "train_speed(iter/s)": 0.447181 + }, + { + "acc": 0.64771976, + "epoch": 1.4335362760020294, + "grad_norm": 6.84375, + "learning_rate": 2.037471992423305e-06, + "loss": 1.6101614, + "memory(GiB)": 111.15, + "step": 56510, + "train_speed(iter/s)": 0.447202 + }, + { + "acc": 0.63411345, + "epoch": 1.4336631151699644, + "grad_norm": 6.0, + "learning_rate": 2.036627321653043e-06, + "loss": 1.70233135, + "memory(GiB)": 111.15, + "step": 56515, + "train_speed(iter/s)": 0.447223 + }, + { + "acc": 0.63389907, + "epoch": 1.4337899543378996, + "grad_norm": 5.34375, + "learning_rate": 2.0357827812251403e-06, + "loss": 1.68496552, + "memory(GiB)": 111.15, + "step": 56520, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.65481272, + "epoch": 1.4339167935058346, + "grad_norm": 5.625, + "learning_rate": 2.0349383711767463e-06, + "loss": 1.64196815, + "memory(GiB)": 111.15, + "step": 56525, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.63592782, + "epoch": 1.4340436326737698, + "grad_norm": 5.4375, + "learning_rate": 2.0340940915450026e-06, + "loss": 1.61801834, + "memory(GiB)": 111.15, + "step": 56530, + "train_speed(iter/s)": 0.447286 + }, + { + "acc": 0.6396358, + "epoch": 1.4341704718417048, + "grad_norm": 5.46875, + "learning_rate": 2.033249942367041e-06, + "loss": 1.6550457, + "memory(GiB)": 111.15, + "step": 56535, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.65600352, + "epoch": 1.4342973110096398, + "grad_norm": 5.25, + "learning_rate": 2.032405923679991e-06, + "loss": 1.57881546, + "memory(GiB)": 111.15, + "step": 56540, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.64532666, + "epoch": 1.4344241501775747, + "grad_norm": 4.71875, + "learning_rate": 2.0315620355209792e-06, + "loss": 1.68657684, + "memory(GiB)": 111.15, + "step": 56545, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.64520202, + "epoch": 1.43455098934551, + "grad_norm": 7.28125, + "learning_rate": 2.030718277927124e-06, + "loss": 1.63995609, + "memory(GiB)": 111.15, + "step": 56550, + "train_speed(iter/s)": 0.44737 + }, + { + "acc": 0.66766462, + "epoch": 1.434677828513445, + "grad_norm": 5.0, + "learning_rate": 2.0298746509355326e-06, + "loss": 1.56143131, + "memory(GiB)": 111.15, + "step": 56555, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.6582552, + "epoch": 1.43480466768138, + "grad_norm": 4.65625, + "learning_rate": 2.0290311545833166e-06, + "loss": 1.60119686, + "memory(GiB)": 111.15, + "step": 56560, + "train_speed(iter/s)": 0.447412 + }, + { + "acc": 0.66940804, + "epoch": 1.4349315068493151, + "grad_norm": 4.96875, + "learning_rate": 2.028187788907574e-06, + "loss": 1.63571739, + "memory(GiB)": 111.15, + "step": 56565, + "train_speed(iter/s)": 0.447433 + }, + { + "acc": 0.66137009, + "epoch": 1.4350583460172501, + "grad_norm": 4.34375, + "learning_rate": 2.027344553945403e-06, + "loss": 1.6387886, + "memory(GiB)": 111.15, + "step": 56570, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.65407505, + "epoch": 1.4351851851851851, + "grad_norm": 4.9375, + "learning_rate": 2.0265014497338868e-06, + "loss": 1.55361671, + "memory(GiB)": 111.15, + "step": 56575, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.63815889, + "epoch": 1.43531202435312, + "grad_norm": 6.15625, + "learning_rate": 2.0256584763101145e-06, + "loss": 1.64658012, + "memory(GiB)": 111.15, + "step": 56580, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.64419231, + "epoch": 1.4354388635210553, + "grad_norm": 5.15625, + "learning_rate": 2.024815633711162e-06, + "loss": 1.66254826, + "memory(GiB)": 111.15, + "step": 56585, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.65264087, + "epoch": 1.4355657026889903, + "grad_norm": 5.03125, + "learning_rate": 2.0239729219741005e-06, + "loss": 1.61051216, + "memory(GiB)": 111.15, + "step": 56590, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.65684943, + "epoch": 1.4356925418569255, + "grad_norm": 5.6875, + "learning_rate": 2.0231303411359975e-06, + "loss": 1.58785858, + "memory(GiB)": 111.15, + "step": 56595, + "train_speed(iter/s)": 0.447558 + }, + { + "acc": 0.64535484, + "epoch": 1.4358193810248605, + "grad_norm": 6.5, + "learning_rate": 2.0222878912339127e-06, + "loss": 1.65299473, + "memory(GiB)": 111.15, + "step": 56600, + "train_speed(iter/s)": 0.44758 + }, + { + "epoch": 1.4358193810248605, + "eval_acc": 0.6434322592396552, + "eval_loss": 1.6075458526611328, + "eval_runtime": 114.3673, + "eval_samples_per_second": 55.698, + "eval_steps_per_second": 27.849, + "step": 56600 + }, + { + "acc": 0.65809593, + "epoch": 1.4359462201927955, + "grad_norm": 5.28125, + "learning_rate": 2.021445572304901e-06, + "loss": 1.62986164, + "memory(GiB)": 111.15, + "step": 56605, + "train_speed(iter/s)": 0.44718 + }, + { + "acc": 0.65908947, + "epoch": 1.4360730593607305, + "grad_norm": 6.21875, + "learning_rate": 2.0206033843860113e-06, + "loss": 1.58276501, + "memory(GiB)": 111.15, + "step": 56610, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.65504947, + "epoch": 1.4361998985286657, + "grad_norm": 5.09375, + "learning_rate": 2.0197613275142868e-06, + "loss": 1.6228138, + "memory(GiB)": 111.15, + "step": 56615, + "train_speed(iter/s)": 0.447221 + }, + { + "acc": 0.65788169, + "epoch": 1.4363267376966007, + "grad_norm": 5.28125, + "learning_rate": 2.018919401726765e-06, + "loss": 1.59430122, + "memory(GiB)": 111.15, + "step": 56620, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.6511898, + "epoch": 1.436453576864536, + "grad_norm": 4.65625, + "learning_rate": 2.0180776070604773e-06, + "loss": 1.63236427, + "memory(GiB)": 111.15, + "step": 56625, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.66122479, + "epoch": 1.4365804160324709, + "grad_norm": 5.28125, + "learning_rate": 2.0172359435524497e-06, + "loss": 1.65969696, + "memory(GiB)": 111.15, + "step": 56630, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.65232987, + "epoch": 1.4367072552004059, + "grad_norm": 6.40625, + "learning_rate": 2.0163944112397027e-06, + "loss": 1.58979216, + "memory(GiB)": 111.15, + "step": 56635, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.64494572, + "epoch": 1.4368340943683409, + "grad_norm": 4.9375, + "learning_rate": 2.01555301015925e-06, + "loss": 1.65109253, + "memory(GiB)": 111.15, + "step": 56640, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.64573798, + "epoch": 1.436960933536276, + "grad_norm": 4.625, + "learning_rate": 2.0147117403480994e-06, + "loss": 1.65340958, + "memory(GiB)": 111.15, + "step": 56645, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.65003667, + "epoch": 1.437087772704211, + "grad_norm": 4.96875, + "learning_rate": 2.0138706018432576e-06, + "loss": 1.57567253, + "memory(GiB)": 111.15, + "step": 56650, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.64317465, + "epoch": 1.437214611872146, + "grad_norm": 4.28125, + "learning_rate": 2.0130295946817176e-06, + "loss": 1.62601471, + "memory(GiB)": 111.15, + "step": 56655, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.65970554, + "epoch": 1.4373414510400813, + "grad_norm": 4.96875, + "learning_rate": 2.0121887189004713e-06, + "loss": 1.58519096, + "memory(GiB)": 111.15, + "step": 56660, + "train_speed(iter/s)": 0.44741 + }, + { + "acc": 0.65759144, + "epoch": 1.4374682902080163, + "grad_norm": 4.71875, + "learning_rate": 2.0113479745365033e-06, + "loss": 1.56526403, + "memory(GiB)": 111.15, + "step": 56665, + "train_speed(iter/s)": 0.44743 + }, + { + "acc": 0.64424639, + "epoch": 1.4375951293759512, + "grad_norm": 5.5625, + "learning_rate": 2.0105073616267984e-06, + "loss": 1.67965317, + "memory(GiB)": 111.15, + "step": 56670, + "train_speed(iter/s)": 0.447451 + }, + { + "acc": 0.65694513, + "epoch": 1.4377219685438862, + "grad_norm": 6.21875, + "learning_rate": 2.0096668802083254e-06, + "loss": 1.67875824, + "memory(GiB)": 111.15, + "step": 56675, + "train_speed(iter/s)": 0.447472 + }, + { + "acc": 0.65654688, + "epoch": 1.4378488077118214, + "grad_norm": 5.125, + "learning_rate": 2.0088265303180516e-06, + "loss": 1.59722958, + "memory(GiB)": 111.15, + "step": 56680, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.64735746, + "epoch": 1.4379756468797564, + "grad_norm": 5.84375, + "learning_rate": 2.0079863119929434e-06, + "loss": 1.63470154, + "memory(GiB)": 111.15, + "step": 56685, + "train_speed(iter/s)": 0.447514 + }, + { + "acc": 0.64923768, + "epoch": 1.4381024860476916, + "grad_norm": 5.0, + "learning_rate": 2.0071462252699575e-06, + "loss": 1.68177719, + "memory(GiB)": 111.15, + "step": 56690, + "train_speed(iter/s)": 0.447535 + }, + { + "acc": 0.64351025, + "epoch": 1.4382293252156266, + "grad_norm": 5.125, + "learning_rate": 2.006306270186039e-06, + "loss": 1.59158058, + "memory(GiB)": 111.15, + "step": 56695, + "train_speed(iter/s)": 0.447556 + }, + { + "acc": 0.6475174, + "epoch": 1.4383561643835616, + "grad_norm": 4.9375, + "learning_rate": 2.005466446778139e-06, + "loss": 1.62039852, + "memory(GiB)": 111.15, + "step": 56700, + "train_speed(iter/s)": 0.447577 + }, + { + "epoch": 1.4383561643835616, + "eval_acc": 0.6434063576664688, + "eval_loss": 1.6075323820114136, + "eval_runtime": 115.3513, + "eval_samples_per_second": 55.223, + "eval_steps_per_second": 27.611, + "step": 56700 + }, + { + "acc": 0.64056325, + "epoch": 1.4384830035514966, + "grad_norm": 5.375, + "learning_rate": 2.0046267550831935e-06, + "loss": 1.70133648, + "memory(GiB)": 111.15, + "step": 56705, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.65311384, + "epoch": 1.4386098427194318, + "grad_norm": 5.15625, + "learning_rate": 2.003787195138139e-06, + "loss": 1.64374523, + "memory(GiB)": 111.15, + "step": 56710, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.6494544, + "epoch": 1.4387366818873668, + "grad_norm": 5.3125, + "learning_rate": 2.002947766979897e-06, + "loss": 1.64897461, + "memory(GiB)": 111.15, + "step": 56715, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.65482483, + "epoch": 1.4388635210553018, + "grad_norm": 5.03125, + "learning_rate": 2.0021084706453945e-06, + "loss": 1.58454914, + "memory(GiB)": 111.15, + "step": 56720, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.64432235, + "epoch": 1.438990360223237, + "grad_norm": 4.90625, + "learning_rate": 2.0012693061715467e-06, + "loss": 1.59713697, + "memory(GiB)": 111.15, + "step": 56725, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.66688561, + "epoch": 1.439117199391172, + "grad_norm": 5.5625, + "learning_rate": 2.000430273595263e-06, + "loss": 1.56731024, + "memory(GiB)": 111.15, + "step": 56730, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.63814063, + "epoch": 1.439244038559107, + "grad_norm": 5.84375, + "learning_rate": 1.9995913729534477e-06, + "loss": 1.6311018, + "memory(GiB)": 111.15, + "step": 56735, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.6404705, + "epoch": 1.439370877727042, + "grad_norm": 4.6875, + "learning_rate": 1.9987526042830003e-06, + "loss": 1.65270348, + "memory(GiB)": 111.15, + "step": 56740, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.65910568, + "epoch": 1.4394977168949772, + "grad_norm": 4.46875, + "learning_rate": 1.9979139676208124e-06, + "loss": 1.59200039, + "memory(GiB)": 111.15, + "step": 56745, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.64856091, + "epoch": 1.4396245560629122, + "grad_norm": 5.96875, + "learning_rate": 1.9970754630037718e-06, + "loss": 1.60078926, + "memory(GiB)": 111.15, + "step": 56750, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.6597373, + "epoch": 1.4397513952308474, + "grad_norm": 4.9375, + "learning_rate": 1.9962370904687596e-06, + "loss": 1.55937986, + "memory(GiB)": 111.15, + "step": 56755, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.6426609, + "epoch": 1.4398782343987824, + "grad_norm": 6.90625, + "learning_rate": 1.9953988500526506e-06, + "loss": 1.68455696, + "memory(GiB)": 111.15, + "step": 56760, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.65446444, + "epoch": 1.4400050735667174, + "grad_norm": 5.09375, + "learning_rate": 1.994560741792315e-06, + "loss": 1.60796909, + "memory(GiB)": 111.15, + "step": 56765, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.65525913, + "epoch": 1.4401319127346524, + "grad_norm": 5.5, + "learning_rate": 1.993722765724616e-06, + "loss": 1.62190437, + "memory(GiB)": 111.15, + "step": 56770, + "train_speed(iter/s)": 0.447444 + }, + { + "acc": 0.64828057, + "epoch": 1.4402587519025876, + "grad_norm": 5.5, + "learning_rate": 1.992884921886412e-06, + "loss": 1.64701996, + "memory(GiB)": 111.15, + "step": 56775, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.64642334, + "epoch": 1.4403855910705226, + "grad_norm": 5.40625, + "learning_rate": 1.9920472103145555e-06, + "loss": 1.66976357, + "memory(GiB)": 111.15, + "step": 56780, + "train_speed(iter/s)": 0.447485 + }, + { + "acc": 0.64859357, + "epoch": 1.4405124302384578, + "grad_norm": 4.71875, + "learning_rate": 1.99120963104589e-06, + "loss": 1.6243969, + "memory(GiB)": 111.15, + "step": 56785, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.64745483, + "epoch": 1.4406392694063928, + "grad_norm": 6.09375, + "learning_rate": 1.990372184117262e-06, + "loss": 1.66411285, + "memory(GiB)": 111.15, + "step": 56790, + "train_speed(iter/s)": 0.447527 + }, + { + "acc": 0.64378219, + "epoch": 1.4407661085743277, + "grad_norm": 5.65625, + "learning_rate": 1.9895348695655e-06, + "loss": 1.62332611, + "memory(GiB)": 111.15, + "step": 56795, + "train_speed(iter/s)": 0.447548 + }, + { + "acc": 0.66731977, + "epoch": 1.4408929477422627, + "grad_norm": 5.09375, + "learning_rate": 1.9886976874274356e-06, + "loss": 1.53933744, + "memory(GiB)": 111.15, + "step": 56800, + "train_speed(iter/s)": 0.447568 + }, + { + "epoch": 1.4408929477422627, + "eval_acc": 0.643369176375927, + "eval_loss": 1.6075506210327148, + "eval_runtime": 113.5016, + "eval_samples_per_second": 56.123, + "eval_steps_per_second": 28.061, + "step": 56800 + }, + { + "acc": 0.66198969, + "epoch": 1.441019786910198, + "grad_norm": 5.125, + "learning_rate": 1.9878606377398895e-06, + "loss": 1.52272739, + "memory(GiB)": 111.15, + "step": 56805, + "train_speed(iter/s)": 0.447173 + }, + { + "acc": 0.65528135, + "epoch": 1.441146626078133, + "grad_norm": 4.09375, + "learning_rate": 1.9870237205396844e-06, + "loss": 1.6083889, + "memory(GiB)": 111.15, + "step": 56810, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.64702649, + "epoch": 1.441273465246068, + "grad_norm": 5.40625, + "learning_rate": 1.986186935863626e-06, + "loss": 1.66459332, + "memory(GiB)": 111.15, + "step": 56815, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.6456769, + "epoch": 1.4414003044140031, + "grad_norm": 5.15625, + "learning_rate": 1.9853502837485207e-06, + "loss": 1.63061008, + "memory(GiB)": 111.15, + "step": 56820, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.64796247, + "epoch": 1.4415271435819381, + "grad_norm": 5.28125, + "learning_rate": 1.9845137642311707e-06, + "loss": 1.61414566, + "memory(GiB)": 111.15, + "step": 56825, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.6600975, + "epoch": 1.441653982749873, + "grad_norm": 5.71875, + "learning_rate": 1.9836773773483704e-06, + "loss": 1.59804344, + "memory(GiB)": 111.15, + "step": 56830, + "train_speed(iter/s)": 0.447277 + }, + { + "acc": 0.63944416, + "epoch": 1.441780821917808, + "grad_norm": 5.40625, + "learning_rate": 1.982841123136904e-06, + "loss": 1.66620121, + "memory(GiB)": 111.15, + "step": 56835, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.64071178, + "epoch": 1.4419076610857433, + "grad_norm": 5.15625, + "learning_rate": 1.982005001633554e-06, + "loss": 1.64978695, + "memory(GiB)": 111.15, + "step": 56840, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.63474436, + "epoch": 1.4420345002536783, + "grad_norm": 5.0625, + "learning_rate": 1.9811690128751002e-06, + "loss": 1.75585556, + "memory(GiB)": 111.15, + "step": 56845, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.64836197, + "epoch": 1.4421613394216135, + "grad_norm": 5.0625, + "learning_rate": 1.980333156898313e-06, + "loss": 1.59285355, + "memory(GiB)": 111.15, + "step": 56850, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.63888187, + "epoch": 1.4422881785895485, + "grad_norm": 5.0, + "learning_rate": 1.979497433739952e-06, + "loss": 1.67020111, + "memory(GiB)": 111.15, + "step": 56855, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.65479908, + "epoch": 1.4424150177574835, + "grad_norm": 5.65625, + "learning_rate": 1.9786618434367814e-06, + "loss": 1.57785673, + "memory(GiB)": 111.15, + "step": 56860, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.65248652, + "epoch": 1.4425418569254185, + "grad_norm": 5.84375, + "learning_rate": 1.977826386025552e-06, + "loss": 1.60627861, + "memory(GiB)": 111.15, + "step": 56865, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.6481039, + "epoch": 1.4426686960933537, + "grad_norm": 5.125, + "learning_rate": 1.976991061543011e-06, + "loss": 1.61572533, + "memory(GiB)": 111.15, + "step": 56870, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.64054351, + "epoch": 1.4427955352612887, + "grad_norm": 5.65625, + "learning_rate": 1.9761558700259e-06, + "loss": 1.651408, + "memory(GiB)": 111.15, + "step": 56875, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.65256906, + "epoch": 1.4429223744292237, + "grad_norm": 5.84375, + "learning_rate": 1.9753208115109546e-06, + "loss": 1.60000076, + "memory(GiB)": 111.15, + "step": 56880, + "train_speed(iter/s)": 0.447487 + }, + { + "acc": 0.64867039, + "epoch": 1.4430492135971589, + "grad_norm": 5.4375, + "learning_rate": 1.9744858860349043e-06, + "loss": 1.63248177, + "memory(GiB)": 111.15, + "step": 56885, + "train_speed(iter/s)": 0.447508 + }, + { + "acc": 0.65133514, + "epoch": 1.4431760527650939, + "grad_norm": 5.0625, + "learning_rate": 1.9736510936344723e-06, + "loss": 1.63177567, + "memory(GiB)": 111.15, + "step": 56890, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.63836756, + "epoch": 1.4433028919330289, + "grad_norm": 5.375, + "learning_rate": 1.9728164343463764e-06, + "loss": 1.67910004, + "memory(GiB)": 111.15, + "step": 56895, + "train_speed(iter/s)": 0.44755 + }, + { + "acc": 0.67412658, + "epoch": 1.4434297311009638, + "grad_norm": 6.59375, + "learning_rate": 1.97198190820733e-06, + "loss": 1.59129467, + "memory(GiB)": 111.15, + "step": 56900, + "train_speed(iter/s)": 0.44757 + }, + { + "epoch": 1.4434297311009638, + "eval_acc": 0.6434013444587553, + "eval_loss": 1.6075855493545532, + "eval_runtime": 113.989, + "eval_samples_per_second": 55.883, + "eval_steps_per_second": 27.941, + "step": 56900 + }, + { + "acc": 0.64020996, + "epoch": 1.443556570268899, + "grad_norm": 4.59375, + "learning_rate": 1.9711475152540376e-06, + "loss": 1.67322884, + "memory(GiB)": 111.15, + "step": 56905, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.64844127, + "epoch": 1.443683409436834, + "grad_norm": 5.25, + "learning_rate": 1.9703132555232007e-06, + "loss": 1.68262291, + "memory(GiB)": 111.15, + "step": 56910, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.66647205, + "epoch": 1.4438102486047693, + "grad_norm": 5.28125, + "learning_rate": 1.9694791290515135e-06, + "loss": 1.58567629, + "memory(GiB)": 111.15, + "step": 56915, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.65324631, + "epoch": 1.4439370877727042, + "grad_norm": 5.75, + "learning_rate": 1.968645135875665e-06, + "loss": 1.65529823, + "memory(GiB)": 111.15, + "step": 56920, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.64236403, + "epoch": 1.4440639269406392, + "grad_norm": 5.0625, + "learning_rate": 1.967811276032335e-06, + "loss": 1.61119099, + "memory(GiB)": 111.15, + "step": 56925, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.64524031, + "epoch": 1.4441907661085742, + "grad_norm": 5.34375, + "learning_rate": 1.966977549558206e-06, + "loss": 1.65274162, + "memory(GiB)": 111.15, + "step": 56930, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.66571665, + "epoch": 1.4443176052765094, + "grad_norm": 4.65625, + "learning_rate": 1.966143956489945e-06, + "loss": 1.59788532, + "memory(GiB)": 111.15, + "step": 56935, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.66842146, + "epoch": 1.4444444444444444, + "grad_norm": 5.0625, + "learning_rate": 1.965310496864217e-06, + "loss": 1.62504044, + "memory(GiB)": 111.15, + "step": 56940, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.66158938, + "epoch": 1.4445712836123796, + "grad_norm": 5.78125, + "learning_rate": 1.9644771707176813e-06, + "loss": 1.59354334, + "memory(GiB)": 111.15, + "step": 56945, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.64075575, + "epoch": 1.4446981227803146, + "grad_norm": 4.59375, + "learning_rate": 1.963643978086996e-06, + "loss": 1.61149273, + "memory(GiB)": 111.15, + "step": 56950, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.65259781, + "epoch": 1.4448249619482496, + "grad_norm": 5.71875, + "learning_rate": 1.9628109190088023e-06, + "loss": 1.57842932, + "memory(GiB)": 111.15, + "step": 56955, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.64971294, + "epoch": 1.4449518011161846, + "grad_norm": 6.125, + "learning_rate": 1.961977993519743e-06, + "loss": 1.66366444, + "memory(GiB)": 111.15, + "step": 56960, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.65363889, + "epoch": 1.4450786402841198, + "grad_norm": 5.1875, + "learning_rate": 1.9611452016564574e-06, + "loss": 1.56871452, + "memory(GiB)": 111.15, + "step": 56965, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.64530449, + "epoch": 1.4452054794520548, + "grad_norm": 5.21875, + "learning_rate": 1.960312543455575e-06, + "loss": 1.66209736, + "memory(GiB)": 111.15, + "step": 56970, + "train_speed(iter/s)": 0.447444 + }, + { + "acc": 0.65417342, + "epoch": 1.4453323186199898, + "grad_norm": 5.59375, + "learning_rate": 1.959480018953716e-06, + "loss": 1.63293495, + "memory(GiB)": 111.15, + "step": 56975, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.62637539, + "epoch": 1.445459157787925, + "grad_norm": 4.0625, + "learning_rate": 1.9586476281874994e-06, + "loss": 1.67549324, + "memory(GiB)": 111.15, + "step": 56980, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.63939805, + "epoch": 1.44558599695586, + "grad_norm": 6.40625, + "learning_rate": 1.9578153711935403e-06, + "loss": 1.64371605, + "memory(GiB)": 111.15, + "step": 56985, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.63728876, + "epoch": 1.445712836123795, + "grad_norm": 6.125, + "learning_rate": 1.9569832480084456e-06, + "loss": 1.68571625, + "memory(GiB)": 111.15, + "step": 56990, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.64277649, + "epoch": 1.44583967529173, + "grad_norm": 7.3125, + "learning_rate": 1.9561512586688096e-06, + "loss": 1.68007393, + "memory(GiB)": 111.15, + "step": 56995, + "train_speed(iter/s)": 0.447549 + }, + { + "acc": 0.64973354, + "epoch": 1.4459665144596652, + "grad_norm": 5.5625, + "learning_rate": 1.9553194032112334e-06, + "loss": 1.59950418, + "memory(GiB)": 111.15, + "step": 57000, + "train_speed(iter/s)": 0.44757 + }, + { + "epoch": 1.4459665144596652, + "eval_acc": 0.6433461991739069, + "eval_loss": 1.6074669361114502, + "eval_runtime": 113.8421, + "eval_samples_per_second": 55.955, + "eval_steps_per_second": 27.977, + "step": 57000 + }, + { + "acc": 0.65603008, + "epoch": 1.4460933536276002, + "grad_norm": 5.84375, + "learning_rate": 1.954487681672303e-06, + "loss": 1.63647251, + "memory(GiB)": 111.15, + "step": 57005, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.66027069, + "epoch": 1.4462201927955354, + "grad_norm": 7.1875, + "learning_rate": 1.9536560940886033e-06, + "loss": 1.58417358, + "memory(GiB)": 111.15, + "step": 57010, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.63318644, + "epoch": 1.4463470319634704, + "grad_norm": 5.3125, + "learning_rate": 1.9528246404967067e-06, + "loss": 1.70260258, + "memory(GiB)": 111.15, + "step": 57015, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.65190401, + "epoch": 1.4464738711314054, + "grad_norm": 5.78125, + "learning_rate": 1.951993320933188e-06, + "loss": 1.6621357, + "memory(GiB)": 111.15, + "step": 57020, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.64989109, + "epoch": 1.4466007102993403, + "grad_norm": 6.28125, + "learning_rate": 1.951162135434612e-06, + "loss": 1.70973511, + "memory(GiB)": 111.15, + "step": 57025, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.65904312, + "epoch": 1.4467275494672756, + "grad_norm": 5.3125, + "learning_rate": 1.9503310840375374e-06, + "loss": 1.63295383, + "memory(GiB)": 111.15, + "step": 57030, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.6522481, + "epoch": 1.4468543886352105, + "grad_norm": 5.9375, + "learning_rate": 1.949500166778517e-06, + "loss": 1.69368668, + "memory(GiB)": 111.15, + "step": 57035, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.64824867, + "epoch": 1.4469812278031455, + "grad_norm": 5.34375, + "learning_rate": 1.948669383694099e-06, + "loss": 1.62688751, + "memory(GiB)": 111.15, + "step": 57040, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.65446587, + "epoch": 1.4471080669710807, + "grad_norm": 5.09375, + "learning_rate": 1.947838734820825e-06, + "loss": 1.59561558, + "memory(GiB)": 111.15, + "step": 57045, + "train_speed(iter/s)": 0.447341 + }, + { + "acc": 0.65527697, + "epoch": 1.4472349061390157, + "grad_norm": 5.6875, + "learning_rate": 1.94700822019523e-06, + "loss": 1.59206133, + "memory(GiB)": 111.15, + "step": 57050, + "train_speed(iter/s)": 0.447362 + }, + { + "acc": 0.66867943, + "epoch": 1.4473617453069507, + "grad_norm": 6.59375, + "learning_rate": 1.9461778398538447e-06, + "loss": 1.53930407, + "memory(GiB)": 111.15, + "step": 57055, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.62266827, + "epoch": 1.4474885844748857, + "grad_norm": 5.34375, + "learning_rate": 1.945347593833191e-06, + "loss": 1.76260414, + "memory(GiB)": 111.15, + "step": 57060, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.64169655, + "epoch": 1.447615423642821, + "grad_norm": 5.46875, + "learning_rate": 1.9445174821697893e-06, + "loss": 1.66545849, + "memory(GiB)": 111.15, + "step": 57065, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.62768173, + "epoch": 1.447742262810756, + "grad_norm": 4.90625, + "learning_rate": 1.94368750490015e-06, + "loss": 1.70575848, + "memory(GiB)": 111.15, + "step": 57070, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.66729565, + "epoch": 1.4478691019786911, + "grad_norm": 5.84375, + "learning_rate": 1.94285766206078e-06, + "loss": 1.55959349, + "memory(GiB)": 111.15, + "step": 57075, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.65060263, + "epoch": 1.447995941146626, + "grad_norm": 6.40625, + "learning_rate": 1.9420279536881794e-06, + "loss": 1.66106796, + "memory(GiB)": 111.15, + "step": 57080, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.65359993, + "epoch": 1.448122780314561, + "grad_norm": 4.78125, + "learning_rate": 1.9411983798188398e-06, + "loss": 1.67618504, + "memory(GiB)": 111.15, + "step": 57085, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.63780527, + "epoch": 1.448249619482496, + "grad_norm": 5.28125, + "learning_rate": 1.940368940489256e-06, + "loss": 1.64427109, + "memory(GiB)": 111.15, + "step": 57090, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.65601301, + "epoch": 1.4483764586504313, + "grad_norm": 5.9375, + "learning_rate": 1.939539635735905e-06, + "loss": 1.5990284, + "memory(GiB)": 111.15, + "step": 57095, + "train_speed(iter/s)": 0.447548 + }, + { + "acc": 0.66606302, + "epoch": 1.4485032978183663, + "grad_norm": 5.65625, + "learning_rate": 1.9387104655952625e-06, + "loss": 1.60883656, + "memory(GiB)": 111.15, + "step": 57100, + "train_speed(iter/s)": 0.447569 + }, + { + "epoch": 1.4485032978183663, + "eval_acc": 0.6433863048356149, + "eval_loss": 1.6075040102005005, + "eval_runtime": 113.7828, + "eval_samples_per_second": 55.984, + "eval_steps_per_second": 27.992, + "step": 57100 + }, + { + "acc": 0.65280499, + "epoch": 1.4486301369863015, + "grad_norm": 5.5625, + "learning_rate": 1.9378814301038033e-06, + "loss": 1.58057957, + "memory(GiB)": 111.15, + "step": 57105, + "train_speed(iter/s)": 0.447175 + }, + { + "acc": 0.64783735, + "epoch": 1.4487569761542365, + "grad_norm": 5.4375, + "learning_rate": 1.937052529297992e-06, + "loss": 1.60144558, + "memory(GiB)": 111.15, + "step": 57110, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.66537275, + "epoch": 1.4488838153221715, + "grad_norm": 5.71875, + "learning_rate": 1.9362237632142838e-06, + "loss": 1.60240154, + "memory(GiB)": 111.15, + "step": 57115, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.66642818, + "epoch": 1.4490106544901065, + "grad_norm": 5.3125, + "learning_rate": 1.9353951318891313e-06, + "loss": 1.53960133, + "memory(GiB)": 111.15, + "step": 57120, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.66481085, + "epoch": 1.4491374936580417, + "grad_norm": 6.09375, + "learning_rate": 1.9345666353589855e-06, + "loss": 1.60342941, + "memory(GiB)": 111.15, + "step": 57125, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.66136632, + "epoch": 1.4492643328259767, + "grad_norm": 6.21875, + "learning_rate": 1.9337382736602868e-06, + "loss": 1.58587847, + "memory(GiB)": 111.15, + "step": 57130, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.65220418, + "epoch": 1.4493911719939117, + "grad_norm": 5.5625, + "learning_rate": 1.9329100468294646e-06, + "loss": 1.61272869, + "memory(GiB)": 111.15, + "step": 57135, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.65887418, + "epoch": 1.4495180111618469, + "grad_norm": 4.6875, + "learning_rate": 1.9320819549029546e-06, + "loss": 1.56940556, + "memory(GiB)": 111.15, + "step": 57140, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.65373087, + "epoch": 1.4496448503297819, + "grad_norm": 5.03125, + "learning_rate": 1.9312539979171774e-06, + "loss": 1.60157146, + "memory(GiB)": 111.15, + "step": 57145, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.64528956, + "epoch": 1.4497716894977168, + "grad_norm": 4.5, + "learning_rate": 1.9304261759085525e-06, + "loss": 1.61926193, + "memory(GiB)": 111.15, + "step": 57150, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.66012144, + "epoch": 1.4498985286656518, + "grad_norm": 5.90625, + "learning_rate": 1.929598488913485e-06, + "loss": 1.59968643, + "memory(GiB)": 111.15, + "step": 57155, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.66273212, + "epoch": 1.450025367833587, + "grad_norm": 4.4375, + "learning_rate": 1.928770936968386e-06, + "loss": 1.57273979, + "memory(GiB)": 111.15, + "step": 57160, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.67191162, + "epoch": 1.450152207001522, + "grad_norm": 4.84375, + "learning_rate": 1.927943520109653e-06, + "loss": 1.55349407, + "memory(GiB)": 111.15, + "step": 57165, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.6543891, + "epoch": 1.4502790461694572, + "grad_norm": 6.0, + "learning_rate": 1.9271162383736804e-06, + "loss": 1.67722282, + "memory(GiB)": 111.15, + "step": 57170, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.65683236, + "epoch": 1.4504058853373922, + "grad_norm": 6.125, + "learning_rate": 1.9262890917968547e-06, + "loss": 1.6054678, + "memory(GiB)": 111.15, + "step": 57175, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.63546562, + "epoch": 1.4505327245053272, + "grad_norm": 5.65625, + "learning_rate": 1.925462080415558e-06, + "loss": 1.73894405, + "memory(GiB)": 111.15, + "step": 57180, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.63642511, + "epoch": 1.4506595636732622, + "grad_norm": 5.40625, + "learning_rate": 1.924635204266166e-06, + "loss": 1.65890999, + "memory(GiB)": 111.15, + "step": 57185, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.65928516, + "epoch": 1.4507864028411974, + "grad_norm": 4.5625, + "learning_rate": 1.923808463385048e-06, + "loss": 1.59265041, + "memory(GiB)": 111.15, + "step": 57190, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.64421806, + "epoch": 1.4509132420091324, + "grad_norm": 4.75, + "learning_rate": 1.922981857808568e-06, + "loss": 1.65882874, + "memory(GiB)": 111.15, + "step": 57195, + "train_speed(iter/s)": 0.447549 + }, + { + "acc": 0.64633408, + "epoch": 1.4510400811770674, + "grad_norm": 4.46875, + "learning_rate": 1.9221553875730835e-06, + "loss": 1.65597801, + "memory(GiB)": 111.15, + "step": 57200, + "train_speed(iter/s)": 0.447569 + }, + { + "epoch": 1.4510400811770674, + "eval_acc": 0.6434401968185348, + "eval_loss": 1.6074659824371338, + "eval_runtime": 113.7389, + "eval_samples_per_second": 56.005, + "eval_steps_per_second": 28.003, + "step": 57200 + }, + { + "acc": 0.64301829, + "epoch": 1.4511669203450026, + "grad_norm": 4.875, + "learning_rate": 1.921329052714947e-06, + "loss": 1.62289066, + "memory(GiB)": 111.15, + "step": 57205, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.6539175, + "epoch": 1.4512937595129376, + "grad_norm": 6.1875, + "learning_rate": 1.920502853270504e-06, + "loss": 1.59929714, + "memory(GiB)": 111.15, + "step": 57210, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.65015087, + "epoch": 1.4514205986808726, + "grad_norm": 5.03125, + "learning_rate": 1.919676789276094e-06, + "loss": 1.6314188, + "memory(GiB)": 111.15, + "step": 57215, + "train_speed(iter/s)": 0.447218 + }, + { + "acc": 0.64427729, + "epoch": 1.4515474378488076, + "grad_norm": 6.34375, + "learning_rate": 1.918850860768052e-06, + "loss": 1.6805727, + "memory(GiB)": 111.15, + "step": 57220, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.65371623, + "epoch": 1.4516742770167428, + "grad_norm": 6.0, + "learning_rate": 1.918025067782704e-06, + "loss": 1.60866146, + "memory(GiB)": 111.15, + "step": 57225, + "train_speed(iter/s)": 0.447259 + }, + { + "acc": 0.65287333, + "epoch": 1.4518011161846778, + "grad_norm": 5.15625, + "learning_rate": 1.9171994103563766e-06, + "loss": 1.64839096, + "memory(GiB)": 111.15, + "step": 57230, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.63488035, + "epoch": 1.451927955352613, + "grad_norm": 5.4375, + "learning_rate": 1.916373888525381e-06, + "loss": 1.73293438, + "memory(GiB)": 111.15, + "step": 57235, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.64199224, + "epoch": 1.452054794520548, + "grad_norm": 5.875, + "learning_rate": 1.9155485023260294e-06, + "loss": 1.67591934, + "memory(GiB)": 111.15, + "step": 57240, + "train_speed(iter/s)": 0.447321 + }, + { + "acc": 0.64426212, + "epoch": 1.452181633688483, + "grad_norm": 5.03125, + "learning_rate": 1.914723251794624e-06, + "loss": 1.63148327, + "memory(GiB)": 111.15, + "step": 57245, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.65125299, + "epoch": 1.452308472856418, + "grad_norm": 5.8125, + "learning_rate": 1.9138981369674688e-06, + "loss": 1.67420292, + "memory(GiB)": 111.15, + "step": 57250, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.64112096, + "epoch": 1.4524353120243532, + "grad_norm": 4.78125, + "learning_rate": 1.9130731578808493e-06, + "loss": 1.660812, + "memory(GiB)": 111.15, + "step": 57255, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.64221549, + "epoch": 1.4525621511922882, + "grad_norm": 4.96875, + "learning_rate": 1.912248314571053e-06, + "loss": 1.64714546, + "memory(GiB)": 111.15, + "step": 57260, + "train_speed(iter/s)": 0.447404 + }, + { + "acc": 0.66421165, + "epoch": 1.4526889903602234, + "grad_norm": 6.125, + "learning_rate": 1.9114236070743638e-06, + "loss": 1.6043972, + "memory(GiB)": 111.15, + "step": 57265, + "train_speed(iter/s)": 0.447425 + }, + { + "acc": 0.65499945, + "epoch": 1.4528158295281584, + "grad_norm": 5.3125, + "learning_rate": 1.910599035427055e-06, + "loss": 1.61851807, + "memory(GiB)": 111.15, + "step": 57270, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.64870872, + "epoch": 1.4529426686960933, + "grad_norm": 6.28125, + "learning_rate": 1.909774599665392e-06, + "loss": 1.58237238, + "memory(GiB)": 111.15, + "step": 57275, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.66529951, + "epoch": 1.4530695078640283, + "grad_norm": 6.96875, + "learning_rate": 1.9089502998256382e-06, + "loss": 1.56070814, + "memory(GiB)": 111.15, + "step": 57280, + "train_speed(iter/s)": 0.447487 + }, + { + "acc": 0.65869255, + "epoch": 1.4531963470319635, + "grad_norm": 5.875, + "learning_rate": 1.9081261359440517e-06, + "loss": 1.60358238, + "memory(GiB)": 111.15, + "step": 57285, + "train_speed(iter/s)": 0.447507 + }, + { + "acc": 0.66820431, + "epoch": 1.4533231861998985, + "grad_norm": 6.09375, + "learning_rate": 1.9073021080568837e-06, + "loss": 1.60153294, + "memory(GiB)": 111.15, + "step": 57290, + "train_speed(iter/s)": 0.447528 + }, + { + "acc": 0.65425167, + "epoch": 1.4534500253678335, + "grad_norm": 6.15625, + "learning_rate": 1.9064782162003737e-06, + "loss": 1.55728512, + "memory(GiB)": 111.15, + "step": 57295, + "train_speed(iter/s)": 0.447549 + }, + { + "acc": 0.65088987, + "epoch": 1.4535768645357687, + "grad_norm": 5.0625, + "learning_rate": 1.9056544604107646e-06, + "loss": 1.59003239, + "memory(GiB)": 111.15, + "step": 57300, + "train_speed(iter/s)": 0.44757 + }, + { + "epoch": 1.4535768645357687, + "eval_acc": 0.6433670875393798, + "eval_loss": 1.6075439453125, + "eval_runtime": 113.7974, + "eval_samples_per_second": 55.977, + "eval_steps_per_second": 27.988, + "step": 57300 + }, + { + "acc": 0.64925666, + "epoch": 1.4537037037037037, + "grad_norm": 6.125, + "learning_rate": 1.9048308407242882e-06, + "loss": 1.61447945, + "memory(GiB)": 111.15, + "step": 57305, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.64498558, + "epoch": 1.4538305428716387, + "grad_norm": 6.46875, + "learning_rate": 1.90400735717717e-06, + "loss": 1.65086136, + "memory(GiB)": 111.15, + "step": 57310, + "train_speed(iter/s)": 0.447197 + }, + { + "acc": 0.64301796, + "epoch": 1.4539573820395737, + "grad_norm": 5.3125, + "learning_rate": 1.903184009805631e-06, + "loss": 1.59810762, + "memory(GiB)": 111.15, + "step": 57315, + "train_speed(iter/s)": 0.447217 + }, + { + "acc": 0.64348025, + "epoch": 1.454084221207509, + "grad_norm": 4.8125, + "learning_rate": 1.9023607986458854e-06, + "loss": 1.70465202, + "memory(GiB)": 111.15, + "step": 57320, + "train_speed(iter/s)": 0.447237 + }, + { + "acc": 0.65344834, + "epoch": 1.454211060375444, + "grad_norm": 5.0, + "learning_rate": 1.901537723734142e-06, + "loss": 1.62563038, + "memory(GiB)": 111.15, + "step": 57325, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.65876985, + "epoch": 1.454337899543379, + "grad_norm": 4.46875, + "learning_rate": 1.9007147851066031e-06, + "loss": 1.5971837, + "memory(GiB)": 111.15, + "step": 57330, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.6691875, + "epoch": 1.454464738711314, + "grad_norm": 4.59375, + "learning_rate": 1.8998919827994654e-06, + "loss": 1.52112961, + "memory(GiB)": 111.15, + "step": 57335, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.65528107, + "epoch": 1.454591577879249, + "grad_norm": 5.4375, + "learning_rate": 1.899069316848919e-06, + "loss": 1.66771011, + "memory(GiB)": 111.15, + "step": 57340, + "train_speed(iter/s)": 0.447319 + }, + { + "acc": 0.64749784, + "epoch": 1.454718417047184, + "grad_norm": 4.65625, + "learning_rate": 1.8982467872911486e-06, + "loss": 1.70237694, + "memory(GiB)": 111.15, + "step": 57345, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.64585094, + "epoch": 1.4548452562151193, + "grad_norm": 4.96875, + "learning_rate": 1.8974243941623332e-06, + "loss": 1.6398819, + "memory(GiB)": 111.15, + "step": 57350, + "train_speed(iter/s)": 0.44736 + }, + { + "acc": 0.64195871, + "epoch": 1.4549720953830543, + "grad_norm": 6.21875, + "learning_rate": 1.896602137498645e-06, + "loss": 1.70341148, + "memory(GiB)": 111.15, + "step": 57355, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.64730744, + "epoch": 1.4550989345509893, + "grad_norm": 5.25, + "learning_rate": 1.89578001733625e-06, + "loss": 1.63075714, + "memory(GiB)": 111.15, + "step": 57360, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.66042013, + "epoch": 1.4552257737189245, + "grad_norm": 4.8125, + "learning_rate": 1.8949580337113078e-06, + "loss": 1.56956081, + "memory(GiB)": 111.15, + "step": 57365, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.63235435, + "epoch": 1.4553526128868595, + "grad_norm": 5.5, + "learning_rate": 1.8941361866599778e-06, + "loss": 1.6798233, + "memory(GiB)": 111.15, + "step": 57370, + "train_speed(iter/s)": 0.447443 + }, + { + "acc": 0.64484215, + "epoch": 1.4554794520547945, + "grad_norm": 4.625, + "learning_rate": 1.893314476218403e-06, + "loss": 1.66152782, + "memory(GiB)": 111.15, + "step": 57375, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.64816008, + "epoch": 1.4556062912227294, + "grad_norm": 4.8125, + "learning_rate": 1.8924929024227279e-06, + "loss": 1.63946152, + "memory(GiB)": 111.15, + "step": 57380, + "train_speed(iter/s)": 0.447485 + }, + { + "acc": 0.63754416, + "epoch": 1.4557331303906647, + "grad_norm": 5.6875, + "learning_rate": 1.8916714653090874e-06, + "loss": 1.67251472, + "memory(GiB)": 111.15, + "step": 57385, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.6441309, + "epoch": 1.4558599695585996, + "grad_norm": 4.75, + "learning_rate": 1.8908501649136174e-06, + "loss": 1.64039154, + "memory(GiB)": 111.15, + "step": 57390, + "train_speed(iter/s)": 0.447526 + }, + { + "acc": 0.65354185, + "epoch": 1.4559868087265349, + "grad_norm": 5.5, + "learning_rate": 1.8900290012724358e-06, + "loss": 1.59603462, + "memory(GiB)": 111.15, + "step": 57395, + "train_speed(iter/s)": 0.447547 + }, + { + "acc": 0.65859303, + "epoch": 1.4561136478944698, + "grad_norm": 5.6875, + "learning_rate": 1.889207974421663e-06, + "loss": 1.55156975, + "memory(GiB)": 111.15, + "step": 57400, + "train_speed(iter/s)": 0.447568 + }, + { + "epoch": 1.4561136478944698, + "eval_acc": 0.6433683408413081, + "eval_loss": 1.607506513595581, + "eval_runtime": 114.2065, + "eval_samples_per_second": 55.776, + "eval_steps_per_second": 27.888, + "step": 57400 + }, + { + "acc": 0.66466932, + "epoch": 1.4562404870624048, + "grad_norm": 6.84375, + "learning_rate": 1.8883870843974134e-06, + "loss": 1.54925547, + "memory(GiB)": 111.15, + "step": 57405, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.63045692, + "epoch": 1.4563673262303398, + "grad_norm": 5.75, + "learning_rate": 1.887566331235794e-06, + "loss": 1.74767876, + "memory(GiB)": 111.15, + "step": 57410, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.63587465, + "epoch": 1.456494165398275, + "grad_norm": 5.96875, + "learning_rate": 1.8867457149729013e-06, + "loss": 1.75127449, + "memory(GiB)": 111.15, + "step": 57415, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.64759932, + "epoch": 1.45662100456621, + "grad_norm": 5.25, + "learning_rate": 1.8859252356448305e-06, + "loss": 1.6350071, + "memory(GiB)": 111.15, + "step": 57420, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.66206875, + "epoch": 1.4567478437341452, + "grad_norm": 4.6875, + "learning_rate": 1.8851048932876725e-06, + "loss": 1.55607271, + "memory(GiB)": 111.15, + "step": 57425, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.66576705, + "epoch": 1.4568746829020802, + "grad_norm": 5.3125, + "learning_rate": 1.8842846879375104e-06, + "loss": 1.60824852, + "memory(GiB)": 111.15, + "step": 57430, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.66053228, + "epoch": 1.4570015220700152, + "grad_norm": 6.03125, + "learning_rate": 1.8834646196304146e-06, + "loss": 1.60882263, + "memory(GiB)": 111.15, + "step": 57435, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.65212469, + "epoch": 1.4571283612379502, + "grad_norm": 6.71875, + "learning_rate": 1.8826446884024612e-06, + "loss": 1.70460758, + "memory(GiB)": 111.15, + "step": 57440, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.65475717, + "epoch": 1.4572552004058854, + "grad_norm": 4.625, + "learning_rate": 1.8818248942897122e-06, + "loss": 1.5622345, + "memory(GiB)": 111.15, + "step": 57445, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.64496832, + "epoch": 1.4573820395738204, + "grad_norm": 5.65625, + "learning_rate": 1.8810052373282277e-06, + "loss": 1.63361931, + "memory(GiB)": 111.15, + "step": 57450, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.64508553, + "epoch": 1.4575088787417554, + "grad_norm": 5.59375, + "learning_rate": 1.880185717554055e-06, + "loss": 1.6635376, + "memory(GiB)": 111.15, + "step": 57455, + "train_speed(iter/s)": 0.44738 + }, + { + "acc": 0.63890724, + "epoch": 1.4576357179096906, + "grad_norm": 5.21875, + "learning_rate": 1.879366335003245e-06, + "loss": 1.66537476, + "memory(GiB)": 111.15, + "step": 57460, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.64046335, + "epoch": 1.4577625570776256, + "grad_norm": 5.96875, + "learning_rate": 1.8785470897118362e-06, + "loss": 1.65967312, + "memory(GiB)": 111.15, + "step": 57465, + "train_speed(iter/s)": 0.447421 + }, + { + "acc": 0.65390992, + "epoch": 1.4578893962455606, + "grad_norm": 5.5, + "learning_rate": 1.8777279817158627e-06, + "loss": 1.59753265, + "memory(GiB)": 111.15, + "step": 57470, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.65168238, + "epoch": 1.4580162354134956, + "grad_norm": 4.84375, + "learning_rate": 1.8769090110513522e-06, + "loss": 1.61164627, + "memory(GiB)": 111.15, + "step": 57475, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.64162035, + "epoch": 1.4581430745814308, + "grad_norm": 5.625, + "learning_rate": 1.8760901777543273e-06, + "loss": 1.67379036, + "memory(GiB)": 111.15, + "step": 57480, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.65615044, + "epoch": 1.4582699137493658, + "grad_norm": 5.40625, + "learning_rate": 1.8752714818608036e-06, + "loss": 1.61937904, + "memory(GiB)": 111.15, + "step": 57485, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.65808616, + "epoch": 1.458396752917301, + "grad_norm": 5.0625, + "learning_rate": 1.874452923406791e-06, + "loss": 1.58524914, + "memory(GiB)": 111.15, + "step": 57490, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64861183, + "epoch": 1.458523592085236, + "grad_norm": 5.0625, + "learning_rate": 1.8736345024282937e-06, + "loss": 1.64928322, + "memory(GiB)": 111.15, + "step": 57495, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.65042677, + "epoch": 1.458650431253171, + "grad_norm": 5.90625, + "learning_rate": 1.8728162189613085e-06, + "loss": 1.60293503, + "memory(GiB)": 111.15, + "step": 57500, + "train_speed(iter/s)": 0.447565 + }, + { + "epoch": 1.458650431253171, + "eval_acc": 0.6433549722874056, + "eval_loss": 1.607520341873169, + "eval_runtime": 113.5043, + "eval_samples_per_second": 56.121, + "eval_steps_per_second": 28.061, + "step": 57500 + }, + { + "acc": 0.65655947, + "epoch": 1.458777270421106, + "grad_norm": 5.53125, + "learning_rate": 1.8719980730418285e-06, + "loss": 1.627145, + "memory(GiB)": 111.15, + "step": 57505, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.64585533, + "epoch": 1.4589041095890412, + "grad_norm": 6.46875, + "learning_rate": 1.8711800647058388e-06, + "loss": 1.61016884, + "memory(GiB)": 111.15, + "step": 57510, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.63342481, + "epoch": 1.4590309487569761, + "grad_norm": 5.40625, + "learning_rate": 1.8703621939893185e-06, + "loss": 1.73023872, + "memory(GiB)": 111.15, + "step": 57515, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.64386749, + "epoch": 1.4591577879249111, + "grad_norm": 6.03125, + "learning_rate": 1.869544460928242e-06, + "loss": 1.60654602, + "memory(GiB)": 111.15, + "step": 57520, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.6448503, + "epoch": 1.4592846270928463, + "grad_norm": 5.25, + "learning_rate": 1.868726865558575e-06, + "loss": 1.63131981, + "memory(GiB)": 111.15, + "step": 57525, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.65288887, + "epoch": 1.4594114662607813, + "grad_norm": 4.6875, + "learning_rate": 1.8679094079162835e-06, + "loss": 1.59185581, + "memory(GiB)": 111.15, + "step": 57530, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.64922504, + "epoch": 1.4595383054287163, + "grad_norm": 6.53125, + "learning_rate": 1.867092088037319e-06, + "loss": 1.59167852, + "memory(GiB)": 111.15, + "step": 57535, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.63674946, + "epoch": 1.4596651445966513, + "grad_norm": 5.5, + "learning_rate": 1.8662749059576296e-06, + "loss": 1.67718029, + "memory(GiB)": 111.15, + "step": 57540, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.65251236, + "epoch": 1.4597919837645865, + "grad_norm": 6.1875, + "learning_rate": 1.865457861713163e-06, + "loss": 1.67348251, + "memory(GiB)": 111.15, + "step": 57545, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.66447468, + "epoch": 1.4599188229325215, + "grad_norm": 5.09375, + "learning_rate": 1.8646409553398558e-06, + "loss": 1.55884151, + "memory(GiB)": 111.15, + "step": 57550, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.65708494, + "epoch": 1.4600456621004567, + "grad_norm": 5.21875, + "learning_rate": 1.8638241868736367e-06, + "loss": 1.60554981, + "memory(GiB)": 111.15, + "step": 57555, + "train_speed(iter/s)": 0.44738 + }, + { + "acc": 0.65220523, + "epoch": 1.4601725012683917, + "grad_norm": 5.90625, + "learning_rate": 1.8630075563504297e-06, + "loss": 1.68808594, + "memory(GiB)": 111.15, + "step": 57560, + "train_speed(iter/s)": 0.4474 + }, + { + "acc": 0.65040331, + "epoch": 1.4602993404363267, + "grad_norm": 5.15625, + "learning_rate": 1.8621910638061575e-06, + "loss": 1.62485542, + "memory(GiB)": 111.15, + "step": 57565, + "train_speed(iter/s)": 0.447421 + }, + { + "acc": 0.64118776, + "epoch": 1.4604261796042617, + "grad_norm": 5.5625, + "learning_rate": 1.8613747092767336e-06, + "loss": 1.63242035, + "memory(GiB)": 111.15, + "step": 57570, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.65582967, + "epoch": 1.460553018772197, + "grad_norm": 5.25, + "learning_rate": 1.8605584927980596e-06, + "loss": 1.62695656, + "memory(GiB)": 111.15, + "step": 57575, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.64907789, + "epoch": 1.4606798579401319, + "grad_norm": 5.75, + "learning_rate": 1.859742414406041e-06, + "loss": 1.56931229, + "memory(GiB)": 111.15, + "step": 57580, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.67471881, + "epoch": 1.460806697108067, + "grad_norm": 5.28125, + "learning_rate": 1.8589264741365714e-06, + "loss": 1.46527777, + "memory(GiB)": 111.15, + "step": 57585, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.64273338, + "epoch": 1.460933536276002, + "grad_norm": 4.46875, + "learning_rate": 1.8581106720255414e-06, + "loss": 1.6806118, + "memory(GiB)": 111.15, + "step": 57590, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.6541564, + "epoch": 1.461060375443937, + "grad_norm": 5.8125, + "learning_rate": 1.8572950081088282e-06, + "loss": 1.60863304, + "memory(GiB)": 111.15, + "step": 57595, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.64759073, + "epoch": 1.461187214611872, + "grad_norm": 5.375, + "learning_rate": 1.856479482422313e-06, + "loss": 1.59547472, + "memory(GiB)": 111.15, + "step": 57600, + "train_speed(iter/s)": 0.447565 + }, + { + "epoch": 1.461187214611872, + "eval_acc": 0.6433374260604083, + "eval_loss": 1.6075332164764404, + "eval_runtime": 113.4242, + "eval_samples_per_second": 56.161, + "eval_steps_per_second": 28.08, + "step": 57600 + }, + { + "acc": 0.65651979, + "epoch": 1.4613140537798073, + "grad_norm": 6.28125, + "learning_rate": 1.8556640950018651e-06, + "loss": 1.5722393, + "memory(GiB)": 111.15, + "step": 57605, + "train_speed(iter/s)": 0.447175 + }, + { + "acc": 0.65833702, + "epoch": 1.4614408929477423, + "grad_norm": 6.4375, + "learning_rate": 1.8548488458833485e-06, + "loss": 1.58300858, + "memory(GiB)": 111.15, + "step": 57610, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.65495529, + "epoch": 1.4615677321156773, + "grad_norm": 4.875, + "learning_rate": 1.854033735102622e-06, + "loss": 1.62655067, + "memory(GiB)": 111.15, + "step": 57615, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.64791455, + "epoch": 1.4616945712836125, + "grad_norm": 6.375, + "learning_rate": 1.8532187626955377e-06, + "loss": 1.65172806, + "memory(GiB)": 111.15, + "step": 57620, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.65525351, + "epoch": 1.4618214104515475, + "grad_norm": 4.75, + "learning_rate": 1.8524039286979417e-06, + "loss": 1.63540611, + "memory(GiB)": 111.15, + "step": 57625, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.65439615, + "epoch": 1.4619482496194824, + "grad_norm": 5.375, + "learning_rate": 1.8515892331456736e-06, + "loss": 1.64452248, + "memory(GiB)": 111.15, + "step": 57630, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.68318605, + "epoch": 1.4620750887874174, + "grad_norm": 5.3125, + "learning_rate": 1.850774676074568e-06, + "loss": 1.51823483, + "memory(GiB)": 111.15, + "step": 57635, + "train_speed(iter/s)": 0.447296 + }, + { + "acc": 0.64382086, + "epoch": 1.4622019279553526, + "grad_norm": 5.09375, + "learning_rate": 1.8499602575204522e-06, + "loss": 1.66702194, + "memory(GiB)": 111.15, + "step": 57640, + "train_speed(iter/s)": 0.447316 + }, + { + "acc": 0.6465766, + "epoch": 1.4623287671232876, + "grad_norm": 4.78125, + "learning_rate": 1.8491459775191484e-06, + "loss": 1.62736549, + "memory(GiB)": 111.15, + "step": 57645, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.63828163, + "epoch": 1.4624556062912228, + "grad_norm": 5.53125, + "learning_rate": 1.8483318361064716e-06, + "loss": 1.6704319, + "memory(GiB)": 111.15, + "step": 57650, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.66191673, + "epoch": 1.4625824454591578, + "grad_norm": 5.1875, + "learning_rate": 1.847517833318232e-06, + "loss": 1.57927427, + "memory(GiB)": 111.15, + "step": 57655, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.66698966, + "epoch": 1.4627092846270928, + "grad_norm": 5.5, + "learning_rate": 1.8467039691902334e-06, + "loss": 1.51500797, + "memory(GiB)": 111.15, + "step": 57660, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.66032219, + "epoch": 1.4628361237950278, + "grad_norm": 5.28125, + "learning_rate": 1.8458902437582705e-06, + "loss": 1.57239523, + "memory(GiB)": 111.15, + "step": 57665, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.65156269, + "epoch": 1.462962962962963, + "grad_norm": 4.9375, + "learning_rate": 1.8450766570581402e-06, + "loss": 1.65529575, + "memory(GiB)": 111.15, + "step": 57670, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.63909864, + "epoch": 1.463089802130898, + "grad_norm": 5.28125, + "learning_rate": 1.8442632091256223e-06, + "loss": 1.63863525, + "memory(GiB)": 111.15, + "step": 57675, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.65167637, + "epoch": 1.463216641298833, + "grad_norm": 4.84375, + "learning_rate": 1.8434498999964983e-06, + "loss": 1.6121769, + "memory(GiB)": 111.15, + "step": 57680, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.66416717, + "epoch": 1.4633434804667682, + "grad_norm": 5.3125, + "learning_rate": 1.8426367297065384e-06, + "loss": 1.59483767, + "memory(GiB)": 111.15, + "step": 57685, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.63740964, + "epoch": 1.4634703196347032, + "grad_norm": 5.0625, + "learning_rate": 1.841823698291516e-06, + "loss": 1.69871101, + "memory(GiB)": 111.15, + "step": 57690, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.66007061, + "epoch": 1.4635971588026382, + "grad_norm": 5.125, + "learning_rate": 1.8410108057871851e-06, + "loss": 1.56418171, + "memory(GiB)": 111.15, + "step": 57695, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.6575088, + "epoch": 1.4637239979705732, + "grad_norm": 5.9375, + "learning_rate": 1.8401980522293017e-06, + "loss": 1.61183014, + "memory(GiB)": 111.15, + "step": 57700, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.4637239979705732, + "eval_acc": 0.6433345016892421, + "eval_loss": 1.6075726747512817, + "eval_runtime": 114.0121, + "eval_samples_per_second": 55.871, + "eval_steps_per_second": 27.936, + "step": 57700 + }, + { + "acc": 0.63428593, + "epoch": 1.4638508371385084, + "grad_norm": 7.28125, + "learning_rate": 1.839385437653617e-06, + "loss": 1.68566666, + "memory(GiB)": 111.15, + "step": 57705, + "train_speed(iter/s)": 0.447169 + }, + { + "acc": 0.65248227, + "epoch": 1.4639776763064434, + "grad_norm": 4.28125, + "learning_rate": 1.8385729620958731e-06, + "loss": 1.66219482, + "memory(GiB)": 111.15, + "step": 57710, + "train_speed(iter/s)": 0.447189 + }, + { + "acc": 0.65317822, + "epoch": 1.4641045154743786, + "grad_norm": 6.09375, + "learning_rate": 1.8377606255918024e-06, + "loss": 1.63625641, + "memory(GiB)": 111.15, + "step": 57715, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.64402876, + "epoch": 1.4642313546423136, + "grad_norm": 5.3125, + "learning_rate": 1.8369484281771388e-06, + "loss": 1.67070045, + "memory(GiB)": 111.15, + "step": 57720, + "train_speed(iter/s)": 0.44723 + }, + { + "acc": 0.64764795, + "epoch": 1.4643581938102486, + "grad_norm": 4.5625, + "learning_rate": 1.836136369887606e-06, + "loss": 1.61327248, + "memory(GiB)": 111.15, + "step": 57725, + "train_speed(iter/s)": 0.447251 + }, + { + "acc": 0.65790629, + "epoch": 1.4644850329781836, + "grad_norm": 8.25, + "learning_rate": 1.8353244507589225e-06, + "loss": 1.59055138, + "memory(GiB)": 111.15, + "step": 57730, + "train_speed(iter/s)": 0.447271 + }, + { + "acc": 0.64307823, + "epoch": 1.4646118721461188, + "grad_norm": 5.125, + "learning_rate": 1.8345126708267958e-06, + "loss": 1.64421024, + "memory(GiB)": 111.15, + "step": 57735, + "train_speed(iter/s)": 0.447292 + }, + { + "acc": 0.66652503, + "epoch": 1.4647387113140538, + "grad_norm": 5.90625, + "learning_rate": 1.8337010301269364e-06, + "loss": 1.56656971, + "memory(GiB)": 111.15, + "step": 57740, + "train_speed(iter/s)": 0.447312 + }, + { + "acc": 0.64833593, + "epoch": 1.464865550481989, + "grad_norm": 6.5, + "learning_rate": 1.8328895286950422e-06, + "loss": 1.64156952, + "memory(GiB)": 111.15, + "step": 57745, + "train_speed(iter/s)": 0.447333 + }, + { + "acc": 0.64749017, + "epoch": 1.464992389649924, + "grad_norm": 5.21875, + "learning_rate": 1.8320781665668063e-06, + "loss": 1.62897873, + "memory(GiB)": 111.15, + "step": 57750, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.63028374, + "epoch": 1.465119228817859, + "grad_norm": 5.40625, + "learning_rate": 1.8312669437779167e-06, + "loss": 1.71531029, + "memory(GiB)": 111.15, + "step": 57755, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.64770665, + "epoch": 1.465246067985794, + "grad_norm": 5.34375, + "learning_rate": 1.8304558603640544e-06, + "loss": 1.65581627, + "memory(GiB)": 111.15, + "step": 57760, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.64196596, + "epoch": 1.4653729071537291, + "grad_norm": 4.84375, + "learning_rate": 1.8296449163608942e-06, + "loss": 1.71336193, + "memory(GiB)": 111.15, + "step": 57765, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.66232786, + "epoch": 1.4654997463216641, + "grad_norm": 4.4375, + "learning_rate": 1.8288341118041052e-06, + "loss": 1.49666853, + "memory(GiB)": 111.15, + "step": 57770, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.65680428, + "epoch": 1.4656265854895991, + "grad_norm": 5.90625, + "learning_rate": 1.82802344672935e-06, + "loss": 1.64602337, + "memory(GiB)": 111.15, + "step": 57775, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64453201, + "epoch": 1.4657534246575343, + "grad_norm": 4.625, + "learning_rate": 1.8272129211722855e-06, + "loss": 1.67505074, + "memory(GiB)": 111.15, + "step": 57780, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.65065703, + "epoch": 1.4658802638254693, + "grad_norm": 6.0625, + "learning_rate": 1.8264025351685627e-06, + "loss": 1.65402184, + "memory(GiB)": 111.15, + "step": 57785, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.6484972, + "epoch": 1.4660071029934043, + "grad_norm": 5.3125, + "learning_rate": 1.8255922887538251e-06, + "loss": 1.59810429, + "memory(GiB)": 111.15, + "step": 57790, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.65564089, + "epoch": 1.4661339421613393, + "grad_norm": 6.6875, + "learning_rate": 1.8247821819637112e-06, + "loss": 1.64300652, + "memory(GiB)": 111.15, + "step": 57795, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.65485392, + "epoch": 1.4662607813292745, + "grad_norm": 5.90625, + "learning_rate": 1.8239722148338534e-06, + "loss": 1.57899895, + "memory(GiB)": 111.15, + "step": 57800, + "train_speed(iter/s)": 0.447559 + }, + { + "epoch": 1.4662607813292745, + "eval_acc": 0.6434193084530619, + "eval_loss": 1.6074469089508057, + "eval_runtime": 113.7545, + "eval_samples_per_second": 55.998, + "eval_steps_per_second": 27.999, + "step": 57800 + }, + { + "acc": 0.64636335, + "epoch": 1.4663876204972095, + "grad_norm": 5.0625, + "learning_rate": 1.823162387399876e-06, + "loss": 1.64918175, + "memory(GiB)": 111.15, + "step": 57805, + "train_speed(iter/s)": 0.44717 + }, + { + "acc": 0.66974745, + "epoch": 1.4665144596651447, + "grad_norm": 5.5625, + "learning_rate": 1.822352699697404e-06, + "loss": 1.58471031, + "memory(GiB)": 111.15, + "step": 57810, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.66008291, + "epoch": 1.4666412988330797, + "grad_norm": 5.09375, + "learning_rate": 1.8215431517620452e-06, + "loss": 1.60256004, + "memory(GiB)": 111.15, + "step": 57815, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.63970795, + "epoch": 1.4667681380010147, + "grad_norm": 5.46875, + "learning_rate": 1.8207337436294097e-06, + "loss": 1.74554558, + "memory(GiB)": 111.15, + "step": 57820, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.6547965, + "epoch": 1.4668949771689497, + "grad_norm": 5.8125, + "learning_rate": 1.819924475335097e-06, + "loss": 1.61086006, + "memory(GiB)": 111.15, + "step": 57825, + "train_speed(iter/s)": 0.447252 + }, + { + "acc": 0.65027876, + "epoch": 1.4670218163368849, + "grad_norm": 4.875, + "learning_rate": 1.8191153469147065e-06, + "loss": 1.61279278, + "memory(GiB)": 111.15, + "step": 57830, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.64033175, + "epoch": 1.4671486555048199, + "grad_norm": 5.03125, + "learning_rate": 1.8183063584038236e-06, + "loss": 1.70126114, + "memory(GiB)": 111.15, + "step": 57835, + "train_speed(iter/s)": 0.447293 + }, + { + "acc": 0.64404669, + "epoch": 1.4672754946727549, + "grad_norm": 5.46875, + "learning_rate": 1.8174975098380304e-06, + "loss": 1.65759411, + "memory(GiB)": 111.15, + "step": 57840, + "train_speed(iter/s)": 0.447314 + }, + { + "acc": 0.660848, + "epoch": 1.46740233384069, + "grad_norm": 5.375, + "learning_rate": 1.8166888012529078e-06, + "loss": 1.63211746, + "memory(GiB)": 111.15, + "step": 57845, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.64545183, + "epoch": 1.467529173008625, + "grad_norm": 5.46875, + "learning_rate": 1.8158802326840252e-06, + "loss": 1.59385719, + "memory(GiB)": 111.15, + "step": 57850, + "train_speed(iter/s)": 0.447355 + }, + { + "acc": 0.6661046, + "epoch": 1.46765601217656, + "grad_norm": 4.40625, + "learning_rate": 1.8150718041669447e-06, + "loss": 1.56079617, + "memory(GiB)": 111.15, + "step": 57855, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.65358682, + "epoch": 1.467782851344495, + "grad_norm": 5.5625, + "learning_rate": 1.814263515737224e-06, + "loss": 1.65320282, + "memory(GiB)": 111.15, + "step": 57860, + "train_speed(iter/s)": 0.447396 + }, + { + "acc": 0.64469972, + "epoch": 1.4679096905124303, + "grad_norm": 4.34375, + "learning_rate": 1.813455367430419e-06, + "loss": 1.56950665, + "memory(GiB)": 111.15, + "step": 57865, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.64630013, + "epoch": 1.4680365296803652, + "grad_norm": 5.0, + "learning_rate": 1.812647359282076e-06, + "loss": 1.72361088, + "memory(GiB)": 111.15, + "step": 57870, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.63455687, + "epoch": 1.4681633688483005, + "grad_norm": 4.96875, + "learning_rate": 1.8118394913277287e-06, + "loss": 1.62319393, + "memory(GiB)": 111.15, + "step": 57875, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.64015775, + "epoch": 1.4682902080162354, + "grad_norm": 5.71875, + "learning_rate": 1.8110317636029162e-06, + "loss": 1.62401428, + "memory(GiB)": 111.15, + "step": 57880, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.65634174, + "epoch": 1.4684170471841704, + "grad_norm": 5.78125, + "learning_rate": 1.810224176143165e-06, + "loss": 1.63373547, + "memory(GiB)": 111.15, + "step": 57885, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.65182405, + "epoch": 1.4685438863521054, + "grad_norm": 5.46875, + "learning_rate": 1.8094167289839953e-06, + "loss": 1.52193079, + "memory(GiB)": 111.15, + "step": 57890, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.6459816, + "epoch": 1.4686707255200406, + "grad_norm": 6.0625, + "learning_rate": 1.808609422160923e-06, + "loss": 1.70517731, + "memory(GiB)": 111.15, + "step": 57895, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.64571857, + "epoch": 1.4687975646879756, + "grad_norm": 5.90625, + "learning_rate": 1.8078022557094571e-06, + "loss": 1.62033043, + "memory(GiB)": 111.15, + "step": 57900, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.4687975646879756, + "eval_acc": 0.6434042688299215, + "eval_loss": 1.6075515747070312, + "eval_runtime": 113.9373, + "eval_samples_per_second": 55.908, + "eval_steps_per_second": 27.954, + "step": 57900 + }, + { + "acc": 0.63651934, + "epoch": 1.4689244038559108, + "grad_norm": 5.75, + "learning_rate": 1.8069952296651e-06, + "loss": 1.67583351, + "memory(GiB)": 111.15, + "step": 57905, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.64516478, + "epoch": 1.4690512430238458, + "grad_norm": 6.75, + "learning_rate": 1.8061883440633481e-06, + "loss": 1.65473671, + "memory(GiB)": 111.15, + "step": 57910, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.66393442, + "epoch": 1.4691780821917808, + "grad_norm": 5.53125, + "learning_rate": 1.8053815989396927e-06, + "loss": 1.61953354, + "memory(GiB)": 111.15, + "step": 57915, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.65989981, + "epoch": 1.4693049213597158, + "grad_norm": 5.28125, + "learning_rate": 1.8045749943296171e-06, + "loss": 1.56159496, + "memory(GiB)": 111.15, + "step": 57920, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.65260477, + "epoch": 1.469431760527651, + "grad_norm": 5.53125, + "learning_rate": 1.8037685302686003e-06, + "loss": 1.66801186, + "memory(GiB)": 111.15, + "step": 57925, + "train_speed(iter/s)": 0.447253 + }, + { + "acc": 0.64416833, + "epoch": 1.469558599695586, + "grad_norm": 5.71875, + "learning_rate": 1.8029622067921133e-06, + "loss": 1.69989433, + "memory(GiB)": 111.15, + "step": 57930, + "train_speed(iter/s)": 0.447273 + }, + { + "acc": 0.66122675, + "epoch": 1.469685438863521, + "grad_norm": 6.0, + "learning_rate": 1.8021560239356223e-06, + "loss": 1.56434994, + "memory(GiB)": 111.15, + "step": 57935, + "train_speed(iter/s)": 0.447293 + }, + { + "acc": 0.66273861, + "epoch": 1.4698122780314562, + "grad_norm": 4.78125, + "learning_rate": 1.8013499817345865e-06, + "loss": 1.60410919, + "memory(GiB)": 111.15, + "step": 57940, + "train_speed(iter/s)": 0.447314 + }, + { + "acc": 0.65635219, + "epoch": 1.4699391171993912, + "grad_norm": 5.3125, + "learning_rate": 1.8005440802244595e-06, + "loss": 1.61758556, + "memory(GiB)": 111.15, + "step": 57945, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.65343552, + "epoch": 1.4700659563673262, + "grad_norm": 4.625, + "learning_rate": 1.7997383194406887e-06, + "loss": 1.5699295, + "memory(GiB)": 111.15, + "step": 57950, + "train_speed(iter/s)": 0.447355 + }, + { + "acc": 0.65881462, + "epoch": 1.4701927955352612, + "grad_norm": 4.78125, + "learning_rate": 1.7989326994187146e-06, + "loss": 1.61006527, + "memory(GiB)": 111.15, + "step": 57955, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.66283321, + "epoch": 1.4703196347031964, + "grad_norm": 5.875, + "learning_rate": 1.798127220193972e-06, + "loss": 1.61613007, + "memory(GiB)": 111.15, + "step": 57960, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.64601994, + "epoch": 1.4704464738711314, + "grad_norm": 6.09375, + "learning_rate": 1.7973218818018878e-06, + "loss": 1.64478722, + "memory(GiB)": 111.15, + "step": 57965, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.64834318, + "epoch": 1.4705733130390666, + "grad_norm": 5.0625, + "learning_rate": 1.7965166842778897e-06, + "loss": 1.61521797, + "memory(GiB)": 111.15, + "step": 57970, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.66106191, + "epoch": 1.4707001522070016, + "grad_norm": 4.78125, + "learning_rate": 1.7957116276573888e-06, + "loss": 1.62552776, + "memory(GiB)": 111.15, + "step": 57975, + "train_speed(iter/s)": 0.447457 + }, + { + "acc": 0.65128312, + "epoch": 1.4708269913749366, + "grad_norm": 5.46875, + "learning_rate": 1.7949067119757951e-06, + "loss": 1.58285294, + "memory(GiB)": 111.15, + "step": 57980, + "train_speed(iter/s)": 0.447477 + }, + { + "acc": 0.6336905, + "epoch": 1.4709538305428715, + "grad_norm": 5.6875, + "learning_rate": 1.7941019372685154e-06, + "loss": 1.67761574, + "memory(GiB)": 111.15, + "step": 57985, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.64565945, + "epoch": 1.4710806697108068, + "grad_norm": 6.3125, + "learning_rate": 1.7932973035709471e-06, + "loss": 1.53968563, + "memory(GiB)": 111.15, + "step": 57990, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.65308003, + "epoch": 1.4712075088787417, + "grad_norm": 5.9375, + "learning_rate": 1.792492810918479e-06, + "loss": 1.61027184, + "memory(GiB)": 111.15, + "step": 57995, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.64064445, + "epoch": 1.4713343480466767, + "grad_norm": 5.0, + "learning_rate": 1.7916884593464957e-06, + "loss": 1.66607552, + "memory(GiB)": 111.15, + "step": 58000, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.4713343480466767, + "eval_acc": 0.6434406145858443, + "eval_loss": 1.6075291633605957, + "eval_runtime": 114.271, + "eval_samples_per_second": 55.745, + "eval_steps_per_second": 27.872, + "step": 58000 + }, + { + "acc": 0.65243092, + "epoch": 1.471461187214612, + "grad_norm": 5.78125, + "learning_rate": 1.79088424889038e-06, + "loss": 1.66656609, + "memory(GiB)": 111.15, + "step": 58005, + "train_speed(iter/s)": 0.447169 + }, + { + "acc": 0.65252295, + "epoch": 1.471588026382547, + "grad_norm": 6.5, + "learning_rate": 1.7900801795855043e-06, + "loss": 1.60830536, + "memory(GiB)": 111.15, + "step": 58010, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.64745378, + "epoch": 1.471714865550482, + "grad_norm": 5.96875, + "learning_rate": 1.7892762514672303e-06, + "loss": 1.60676422, + "memory(GiB)": 111.15, + "step": 58015, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.6419374, + "epoch": 1.471841704718417, + "grad_norm": 5.59375, + "learning_rate": 1.7884724645709228e-06, + "loss": 1.65562019, + "memory(GiB)": 111.15, + "step": 58020, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.65520372, + "epoch": 1.4719685438863521, + "grad_norm": 5.09375, + "learning_rate": 1.7876688189319353e-06, + "loss": 1.63726845, + "memory(GiB)": 111.15, + "step": 58025, + "train_speed(iter/s)": 0.447251 + }, + { + "acc": 0.66013737, + "epoch": 1.472095383054287, + "grad_norm": 4.40625, + "learning_rate": 1.7868653145856163e-06, + "loss": 1.56042633, + "memory(GiB)": 111.15, + "step": 58030, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.67318954, + "epoch": 1.4722222222222223, + "grad_norm": 6.09375, + "learning_rate": 1.7860619515673034e-06, + "loss": 1.48575382, + "memory(GiB)": 111.15, + "step": 58035, + "train_speed(iter/s)": 0.447292 + }, + { + "acc": 0.66508455, + "epoch": 1.4723490613901573, + "grad_norm": 4.875, + "learning_rate": 1.785258729912337e-06, + "loss": 1.6196579, + "memory(GiB)": 111.15, + "step": 58040, + "train_speed(iter/s)": 0.447313 + }, + { + "acc": 0.6358552, + "epoch": 1.4724759005580923, + "grad_norm": 5.25, + "learning_rate": 1.784455649656044e-06, + "loss": 1.73560295, + "memory(GiB)": 111.15, + "step": 58045, + "train_speed(iter/s)": 0.447333 + }, + { + "acc": 0.65767322, + "epoch": 1.4726027397260273, + "grad_norm": 4.59375, + "learning_rate": 1.7836527108337482e-06, + "loss": 1.65537186, + "memory(GiB)": 111.15, + "step": 58050, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.6474618, + "epoch": 1.4727295788939625, + "grad_norm": 4.40625, + "learning_rate": 1.782849913480766e-06, + "loss": 1.5721571, + "memory(GiB)": 111.15, + "step": 58055, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.64921436, + "epoch": 1.4728564180618975, + "grad_norm": 4.9375, + "learning_rate": 1.7820472576324078e-06, + "loss": 1.62607441, + "memory(GiB)": 111.15, + "step": 58060, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.66671238, + "epoch": 1.4729832572298327, + "grad_norm": 5.28125, + "learning_rate": 1.7812447433239789e-06, + "loss": 1.5835165, + "memory(GiB)": 111.15, + "step": 58065, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.64661627, + "epoch": 1.4731100963977677, + "grad_norm": 5.0625, + "learning_rate": 1.7804423705907764e-06, + "loss": 1.75292282, + "memory(GiB)": 111.15, + "step": 58070, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.65741658, + "epoch": 1.4732369355657027, + "grad_norm": 4.875, + "learning_rate": 1.779640139468093e-06, + "loss": 1.58164654, + "memory(GiB)": 111.15, + "step": 58075, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.65454621, + "epoch": 1.4733637747336377, + "grad_norm": 5.375, + "learning_rate": 1.778838049991214e-06, + "loss": 1.64990768, + "memory(GiB)": 111.15, + "step": 58080, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.63925085, + "epoch": 1.4734906139015729, + "grad_norm": 5.25, + "learning_rate": 1.778036102195419e-06, + "loss": 1.60106049, + "memory(GiB)": 111.15, + "step": 58085, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.64884443, + "epoch": 1.4736174530695079, + "grad_norm": 5.3125, + "learning_rate": 1.7772342961159817e-06, + "loss": 1.62945175, + "memory(GiB)": 111.15, + "step": 58090, + "train_speed(iter/s)": 0.447517 + }, + { + "acc": 0.65488653, + "epoch": 1.4737442922374429, + "grad_norm": 5.90625, + "learning_rate": 1.7764326317881681e-06, + "loss": 1.63148022, + "memory(GiB)": 111.15, + "step": 58095, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.66004686, + "epoch": 1.473871131405378, + "grad_norm": 6.59375, + "learning_rate": 1.77563110924724e-06, + "loss": 1.62857094, + "memory(GiB)": 111.15, + "step": 58100, + "train_speed(iter/s)": 0.447558 + }, + { + "epoch": 1.473871131405378, + "eval_acc": 0.6434422856550821, + "eval_loss": 1.6075628995895386, + "eval_runtime": 113.8446, + "eval_samples_per_second": 55.953, + "eval_steps_per_second": 27.977, + "step": 58100 + }, + { + "acc": 0.64455619, + "epoch": 1.473997970573313, + "grad_norm": 5.28125, + "learning_rate": 1.7748297285284494e-06, + "loss": 1.62297344, + "memory(GiB)": 111.15, + "step": 58105, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.65427437, + "epoch": 1.474124809741248, + "grad_norm": 5.875, + "learning_rate": 1.7740284896670507e-06, + "loss": 1.55753374, + "memory(GiB)": 111.15, + "step": 58110, + "train_speed(iter/s)": 0.447191 + }, + { + "acc": 0.65848312, + "epoch": 1.474251648909183, + "grad_norm": 5.46875, + "learning_rate": 1.7732273926982796e-06, + "loss": 1.61133423, + "memory(GiB)": 111.15, + "step": 58115, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.65760756, + "epoch": 1.4743784880771182, + "grad_norm": 4.8125, + "learning_rate": 1.7724264376573747e-06, + "loss": 1.57222137, + "memory(GiB)": 111.15, + "step": 58120, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.64078588, + "epoch": 1.4745053272450532, + "grad_norm": 5.25, + "learning_rate": 1.7716256245795631e-06, + "loss": 1.68827305, + "memory(GiB)": 111.15, + "step": 58125, + "train_speed(iter/s)": 0.447251 + }, + { + "acc": 0.63952546, + "epoch": 1.4746321664129884, + "grad_norm": 5.34375, + "learning_rate": 1.7708249535000737e-06, + "loss": 1.64110184, + "memory(GiB)": 111.15, + "step": 58130, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.63487406, + "epoch": 1.4747590055809234, + "grad_norm": 5.03125, + "learning_rate": 1.7700244244541182e-06, + "loss": 1.64078331, + "memory(GiB)": 111.15, + "step": 58135, + "train_speed(iter/s)": 0.447292 + }, + { + "acc": 0.66445098, + "epoch": 1.4748858447488584, + "grad_norm": 5.8125, + "learning_rate": 1.7692240374769081e-06, + "loss": 1.6231287, + "memory(GiB)": 111.15, + "step": 58140, + "train_speed(iter/s)": 0.447312 + }, + { + "acc": 0.65545626, + "epoch": 1.4750126839167934, + "grad_norm": 5.4375, + "learning_rate": 1.7684237926036507e-06, + "loss": 1.68126259, + "memory(GiB)": 111.15, + "step": 58145, + "train_speed(iter/s)": 0.447333 + }, + { + "acc": 0.64123259, + "epoch": 1.4751395230847286, + "grad_norm": 5.0625, + "learning_rate": 1.7676236898695442e-06, + "loss": 1.65162907, + "memory(GiB)": 111.15, + "step": 58150, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.64478512, + "epoch": 1.4752663622526636, + "grad_norm": 5.15625, + "learning_rate": 1.7668237293097762e-06, + "loss": 1.65856705, + "memory(GiB)": 111.15, + "step": 58155, + "train_speed(iter/s)": 0.447373 + }, + { + "acc": 0.65980134, + "epoch": 1.4753932014205986, + "grad_norm": 5.25, + "learning_rate": 1.7660239109595374e-06, + "loss": 1.58383846, + "memory(GiB)": 111.15, + "step": 58160, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.63060603, + "epoch": 1.4755200405885338, + "grad_norm": 6.5, + "learning_rate": 1.7652242348540056e-06, + "loss": 1.71057816, + "memory(GiB)": 111.15, + "step": 58165, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.6524229, + "epoch": 1.4756468797564688, + "grad_norm": 5.03125, + "learning_rate": 1.764424701028356e-06, + "loss": 1.55293636, + "memory(GiB)": 111.15, + "step": 58170, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.63688283, + "epoch": 1.4757737189244038, + "grad_norm": 5.40625, + "learning_rate": 1.7636253095177507e-06, + "loss": 1.70025902, + "memory(GiB)": 111.15, + "step": 58175, + "train_speed(iter/s)": 0.447455 + }, + { + "acc": 0.63268723, + "epoch": 1.4759005580923388, + "grad_norm": 4.84375, + "learning_rate": 1.762826060357355e-06, + "loss": 1.63297615, + "memory(GiB)": 111.15, + "step": 58180, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.65657501, + "epoch": 1.476027397260274, + "grad_norm": 5.25, + "learning_rate": 1.762026953582322e-06, + "loss": 1.60180187, + "memory(GiB)": 111.15, + "step": 58185, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.65074739, + "epoch": 1.476154236428209, + "grad_norm": 6.25, + "learning_rate": 1.7612279892278006e-06, + "loss": 1.59531555, + "memory(GiB)": 111.15, + "step": 58190, + "train_speed(iter/s)": 0.447516 + }, + { + "acc": 0.65543537, + "epoch": 1.4762810755961442, + "grad_norm": 5.21875, + "learning_rate": 1.7604291673289314e-06, + "loss": 1.61665039, + "memory(GiB)": 111.15, + "step": 58195, + "train_speed(iter/s)": 0.447536 + }, + { + "acc": 0.65638566, + "epoch": 1.4764079147640792, + "grad_norm": 5.375, + "learning_rate": 1.759630487920852e-06, + "loss": 1.55604658, + "memory(GiB)": 111.15, + "step": 58200, + "train_speed(iter/s)": 0.447557 + }, + { + "epoch": 1.4764079147640792, + "eval_acc": 0.6433879759048526, + "eval_loss": 1.6074656248092651, + "eval_runtime": 113.2712, + "eval_samples_per_second": 56.237, + "eval_steps_per_second": 28.118, + "step": 58200 + }, + { + "acc": 0.65556593, + "epoch": 1.4765347539320142, + "grad_norm": 4.78125, + "learning_rate": 1.7588319510386903e-06, + "loss": 1.59348459, + "memory(GiB)": 111.15, + "step": 58205, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.64962611, + "epoch": 1.4766615930999492, + "grad_norm": 4.78125, + "learning_rate": 1.7580335567175704e-06, + "loss": 1.62083473, + "memory(GiB)": 111.15, + "step": 58210, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.63872461, + "epoch": 1.4767884322678844, + "grad_norm": 4.8125, + "learning_rate": 1.7572353049926094e-06, + "loss": 1.71400948, + "memory(GiB)": 111.15, + "step": 58215, + "train_speed(iter/s)": 0.447212 + }, + { + "acc": 0.66591396, + "epoch": 1.4769152714358194, + "grad_norm": 4.84375, + "learning_rate": 1.7564371958989173e-06, + "loss": 1.55014515, + "memory(GiB)": 111.15, + "step": 58220, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.65837526, + "epoch": 1.4770421106037546, + "grad_norm": 6.21875, + "learning_rate": 1.7556392294715984e-06, + "loss": 1.62041397, + "memory(GiB)": 111.15, + "step": 58225, + "train_speed(iter/s)": 0.447252 + }, + { + "acc": 0.64168434, + "epoch": 1.4771689497716896, + "grad_norm": 5.6875, + "learning_rate": 1.7548414057457518e-06, + "loss": 1.62254372, + "memory(GiB)": 111.15, + "step": 58230, + "train_speed(iter/s)": 0.447273 + }, + { + "acc": 0.65857558, + "epoch": 1.4772957889396245, + "grad_norm": 4.8125, + "learning_rate": 1.7540437247564685e-06, + "loss": 1.62731152, + "memory(GiB)": 111.15, + "step": 58235, + "train_speed(iter/s)": 0.447293 + }, + { + "acc": 0.66837826, + "epoch": 1.4774226281075595, + "grad_norm": 4.53125, + "learning_rate": 1.7532461865388345e-06, + "loss": 1.54287205, + "memory(GiB)": 111.15, + "step": 58240, + "train_speed(iter/s)": 0.447313 + }, + { + "acc": 0.6482295, + "epoch": 1.4775494672754947, + "grad_norm": 5.53125, + "learning_rate": 1.752448791127927e-06, + "loss": 1.61667824, + "memory(GiB)": 111.15, + "step": 58245, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.65805483, + "epoch": 1.4776763064434297, + "grad_norm": 5.6875, + "learning_rate": 1.7516515385588245e-06, + "loss": 1.62177048, + "memory(GiB)": 111.15, + "step": 58250, + "train_speed(iter/s)": 0.447354 + }, + { + "acc": 0.63581829, + "epoch": 1.4778031456113647, + "grad_norm": 6.0, + "learning_rate": 1.7508544288665885e-06, + "loss": 1.65450401, + "memory(GiB)": 111.15, + "step": 58255, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.66169558, + "epoch": 1.4779299847793, + "grad_norm": 4.8125, + "learning_rate": 1.750057462086281e-06, + "loss": 1.59355145, + "memory(GiB)": 111.15, + "step": 58260, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.65215521, + "epoch": 1.478056823947235, + "grad_norm": 5.03125, + "learning_rate": 1.7492606382529542e-06, + "loss": 1.5924017, + "memory(GiB)": 111.15, + "step": 58265, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.66226811, + "epoch": 1.47818366311517, + "grad_norm": 5.5625, + "learning_rate": 1.748463957401662e-06, + "loss": 1.65245895, + "memory(GiB)": 111.15, + "step": 58270, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.64803314, + "epoch": 1.478310502283105, + "grad_norm": 5.59375, + "learning_rate": 1.7476674195674404e-06, + "loss": 1.64765358, + "memory(GiB)": 111.15, + "step": 58275, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.65740004, + "epoch": 1.47843734145104, + "grad_norm": 6.125, + "learning_rate": 1.7468710247853244e-06, + "loss": 1.65339813, + "memory(GiB)": 111.15, + "step": 58280, + "train_speed(iter/s)": 0.447477 + }, + { + "acc": 0.64943376, + "epoch": 1.478564180618975, + "grad_norm": 5.15625, + "learning_rate": 1.7460747730903466e-06, + "loss": 1.56628866, + "memory(GiB)": 111.15, + "step": 58285, + "train_speed(iter/s)": 0.447497 + }, + { + "acc": 0.64941235, + "epoch": 1.4786910197869103, + "grad_norm": 6.03125, + "learning_rate": 1.7452786645175297e-06, + "loss": 1.61458778, + "memory(GiB)": 111.15, + "step": 58290, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.6476408, + "epoch": 1.4788178589548453, + "grad_norm": 5.0625, + "learning_rate": 1.7444826991018864e-06, + "loss": 1.6823616, + "memory(GiB)": 111.15, + "step": 58295, + "train_speed(iter/s)": 0.447538 + }, + { + "acc": 0.65026379, + "epoch": 1.4789446981227803, + "grad_norm": 5.25, + "learning_rate": 1.7436868768784276e-06, + "loss": 1.63072071, + "memory(GiB)": 111.15, + "step": 58300, + "train_speed(iter/s)": 0.447558 + }, + { + "epoch": 1.4789446981227803, + "eval_acc": 0.6433888114394716, + "eval_loss": 1.6075495481491089, + "eval_runtime": 113.3233, + "eval_samples_per_second": 56.211, + "eval_steps_per_second": 28.105, + "step": 58300 + }, + { + "acc": 0.65594077, + "epoch": 1.4790715372907153, + "grad_norm": 6.40625, + "learning_rate": 1.7428911978821594e-06, + "loss": 1.59324818, + "memory(GiB)": 111.15, + "step": 58305, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.63910542, + "epoch": 1.4791983764586505, + "grad_norm": 5.78125, + "learning_rate": 1.7420956621480806e-06, + "loss": 1.65394993, + "memory(GiB)": 111.15, + "step": 58310, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.65935478, + "epoch": 1.4793252156265855, + "grad_norm": 4.6875, + "learning_rate": 1.7413002697111765e-06, + "loss": 1.53878422, + "memory(GiB)": 111.15, + "step": 58315, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.65939264, + "epoch": 1.4794520547945205, + "grad_norm": 5.59375, + "learning_rate": 1.7405050206064372e-06, + "loss": 1.62786446, + "memory(GiB)": 111.15, + "step": 58320, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.6469347, + "epoch": 1.4795788939624557, + "grad_norm": 4.6875, + "learning_rate": 1.73970991486884e-06, + "loss": 1.63285789, + "memory(GiB)": 111.15, + "step": 58325, + "train_speed(iter/s)": 0.447255 + }, + { + "acc": 0.63600731, + "epoch": 1.4797057331303907, + "grad_norm": 6.375, + "learning_rate": 1.7389149525333565e-06, + "loss": 1.68682938, + "memory(GiB)": 111.15, + "step": 58330, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.62610197, + "epoch": 1.4798325722983257, + "grad_norm": 5.0, + "learning_rate": 1.7381201336349535e-06, + "loss": 1.62766609, + "memory(GiB)": 111.15, + "step": 58335, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.64607267, + "epoch": 1.4799594114662606, + "grad_norm": 5.375, + "learning_rate": 1.7373254582085896e-06, + "loss": 1.63782921, + "memory(GiB)": 111.15, + "step": 58340, + "train_speed(iter/s)": 0.447316 + }, + { + "acc": 0.65470257, + "epoch": 1.4800862506341959, + "grad_norm": 6.21875, + "learning_rate": 1.7365309262892194e-06, + "loss": 1.58465099, + "memory(GiB)": 111.15, + "step": 58345, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.62747574, + "epoch": 1.4802130898021308, + "grad_norm": 6.40625, + "learning_rate": 1.735736537911789e-06, + "loss": 1.70026073, + "memory(GiB)": 111.15, + "step": 58350, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.63447533, + "epoch": 1.480339928970066, + "grad_norm": 4.75, + "learning_rate": 1.7349422931112403e-06, + "loss": 1.68418846, + "memory(GiB)": 111.15, + "step": 58355, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.66041017, + "epoch": 1.480466768138001, + "grad_norm": 4.5, + "learning_rate": 1.7341481919225062e-06, + "loss": 1.60844879, + "memory(GiB)": 111.15, + "step": 58360, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.65298777, + "epoch": 1.480593607305936, + "grad_norm": 4.6875, + "learning_rate": 1.733354234380516e-06, + "loss": 1.58539066, + "memory(GiB)": 111.15, + "step": 58365, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.64562836, + "epoch": 1.480720446473871, + "grad_norm": 5.03125, + "learning_rate": 1.7325604205201912e-06, + "loss": 1.57029438, + "memory(GiB)": 111.15, + "step": 58370, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.6428997, + "epoch": 1.4808472856418062, + "grad_norm": 4.6875, + "learning_rate": 1.7317667503764468e-06, + "loss": 1.57957144, + "memory(GiB)": 111.15, + "step": 58375, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.6374732, + "epoch": 1.4809741248097412, + "grad_norm": 5.21875, + "learning_rate": 1.7309732239841926e-06, + "loss": 1.7527523, + "memory(GiB)": 111.15, + "step": 58380, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.659554, + "epoch": 1.4811009639776764, + "grad_norm": 5.25, + "learning_rate": 1.730179841378331e-06, + "loss": 1.57223225, + "memory(GiB)": 111.15, + "step": 58385, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.6576242, + "epoch": 1.4812278031456114, + "grad_norm": 4.78125, + "learning_rate": 1.7293866025937589e-06, + "loss": 1.60860596, + "memory(GiB)": 111.15, + "step": 58390, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.65528603, + "epoch": 1.4813546423135464, + "grad_norm": 4.9375, + "learning_rate": 1.7285935076653659e-06, + "loss": 1.54409523, + "memory(GiB)": 111.15, + "step": 58395, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.66362906, + "epoch": 1.4814814814814814, + "grad_norm": 5.8125, + "learning_rate": 1.7278005566280365e-06, + "loss": 1.58631506, + "memory(GiB)": 111.15, + "step": 58400, + "train_speed(iter/s)": 0.447559 + }, + { + "epoch": 1.4814814814814814, + "eval_acc": 0.6434790491783143, + "eval_loss": 1.6074994802474976, + "eval_runtime": 113.1257, + "eval_samples_per_second": 56.309, + "eval_steps_per_second": 28.155, + "step": 58400 + }, + { + "acc": 0.63569055, + "epoch": 1.4816083206494166, + "grad_norm": 5.71875, + "learning_rate": 1.727007749516646e-06, + "loss": 1.68203392, + "memory(GiB)": 111.15, + "step": 58405, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.64839907, + "epoch": 1.4817351598173516, + "grad_norm": 4.84375, + "learning_rate": 1.7262150863660709e-06, + "loss": 1.69707088, + "memory(GiB)": 111.15, + "step": 58410, + "train_speed(iter/s)": 0.447196 + }, + { + "acc": 0.64876556, + "epoch": 1.4818619989852866, + "grad_norm": 5.09375, + "learning_rate": 1.7254225672111713e-06, + "loss": 1.6409153, + "memory(GiB)": 111.15, + "step": 58415, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.64870543, + "epoch": 1.4819888381532218, + "grad_norm": 6.34375, + "learning_rate": 1.7246301920868052e-06, + "loss": 1.61809883, + "memory(GiB)": 111.15, + "step": 58420, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.65386925, + "epoch": 1.4821156773211568, + "grad_norm": 5.78125, + "learning_rate": 1.723837961027829e-06, + "loss": 1.53111248, + "memory(GiB)": 111.15, + "step": 58425, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.65757084, + "epoch": 1.4822425164890918, + "grad_norm": 6.34375, + "learning_rate": 1.723045874069087e-06, + "loss": 1.61952648, + "memory(GiB)": 111.15, + "step": 58430, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.65925589, + "epoch": 1.4823693556570268, + "grad_norm": 5.03125, + "learning_rate": 1.7222539312454167e-06, + "loss": 1.6399292, + "memory(GiB)": 111.15, + "step": 58435, + "train_speed(iter/s)": 0.447296 + }, + { + "acc": 0.64644279, + "epoch": 1.482496194824962, + "grad_norm": 6.0, + "learning_rate": 1.7214621325916515e-06, + "loss": 1.66069756, + "memory(GiB)": 111.15, + "step": 58440, + "train_speed(iter/s)": 0.447317 + }, + { + "acc": 0.66636057, + "epoch": 1.482623033992897, + "grad_norm": 7.4375, + "learning_rate": 1.7206704781426204e-06, + "loss": 1.57163591, + "memory(GiB)": 111.15, + "step": 58445, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.66229782, + "epoch": 1.4827498731608322, + "grad_norm": 5.71875, + "learning_rate": 1.7198789679331445e-06, + "loss": 1.57930679, + "memory(GiB)": 111.15, + "step": 58450, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.6611557, + "epoch": 1.4828767123287672, + "grad_norm": 4.0, + "learning_rate": 1.7190876019980329e-06, + "loss": 1.54644833, + "memory(GiB)": 111.15, + "step": 58455, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.64931006, + "epoch": 1.4830035514967022, + "grad_norm": 5.21875, + "learning_rate": 1.7182963803720987e-06, + "loss": 1.61827431, + "memory(GiB)": 111.15, + "step": 58460, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.63233333, + "epoch": 1.4831303906646371, + "grad_norm": 6.0625, + "learning_rate": 1.7175053030901418e-06, + "loss": 1.6987751, + "memory(GiB)": 111.15, + "step": 58465, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.65638218, + "epoch": 1.4832572298325724, + "grad_norm": 5.59375, + "learning_rate": 1.7167143701869582e-06, + "loss": 1.61285973, + "memory(GiB)": 111.15, + "step": 58470, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.6474638, + "epoch": 1.4833840690005073, + "grad_norm": 6.0, + "learning_rate": 1.7159235816973318e-06, + "loss": 1.61614189, + "memory(GiB)": 111.15, + "step": 58475, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.65271401, + "epoch": 1.4835109081684423, + "grad_norm": 4.84375, + "learning_rate": 1.7151329376560506e-06, + "loss": 1.65035954, + "memory(GiB)": 111.15, + "step": 58480, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.64618397, + "epoch": 1.4836377473363775, + "grad_norm": 4.40625, + "learning_rate": 1.7143424380978885e-06, + "loss": 1.62648659, + "memory(GiB)": 111.15, + "step": 58485, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.64870901, + "epoch": 1.4837645865043125, + "grad_norm": 5.75, + "learning_rate": 1.7135520830576157e-06, + "loss": 1.67320175, + "memory(GiB)": 111.15, + "step": 58490, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.65418048, + "epoch": 1.4838914256722475, + "grad_norm": 4.78125, + "learning_rate": 1.712761872569995e-06, + "loss": 1.60712719, + "memory(GiB)": 111.15, + "step": 58495, + "train_speed(iter/s)": 0.44754 + }, + { + "acc": 0.64903831, + "epoch": 1.4840182648401825, + "grad_norm": 4.90625, + "learning_rate": 1.7119718066697838e-06, + "loss": 1.64217968, + "memory(GiB)": 111.15, + "step": 58500, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.4840182648401825, + "eval_acc": 0.6434443744916294, + "eval_loss": 1.6075260639190674, + "eval_runtime": 114.1357, + "eval_samples_per_second": 55.811, + "eval_steps_per_second": 27.905, + "step": 58500 + }, + { + "acc": 0.64305115, + "epoch": 1.4841451040081177, + "grad_norm": 5.0625, + "learning_rate": 1.7111818853917323e-06, + "loss": 1.61544876, + "memory(GiB)": 111.15, + "step": 58505, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.64864855, + "epoch": 1.4842719431760527, + "grad_norm": 5.0, + "learning_rate": 1.710392108770585e-06, + "loss": 1.6819603, + "memory(GiB)": 111.15, + "step": 58510, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.65075321, + "epoch": 1.484398782343988, + "grad_norm": 5.125, + "learning_rate": 1.7096024768410796e-06, + "loss": 1.65263882, + "memory(GiB)": 111.15, + "step": 58515, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.637534, + "epoch": 1.484525621511923, + "grad_norm": 4.5625, + "learning_rate": 1.7088129896379484e-06, + "loss": 1.67136574, + "memory(GiB)": 111.15, + "step": 58520, + "train_speed(iter/s)": 0.447234 + }, + { + "acc": 0.63932772, + "epoch": 1.484652460679858, + "grad_norm": 4.96875, + "learning_rate": 1.7080236471959155e-06, + "loss": 1.60586166, + "memory(GiB)": 111.15, + "step": 58525, + "train_speed(iter/s)": 0.447255 + }, + { + "acc": 0.65553565, + "epoch": 1.4847792998477929, + "grad_norm": 5.6875, + "learning_rate": 1.7072344495497007e-06, + "loss": 1.63078747, + "memory(GiB)": 111.15, + "step": 58530, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.68273687, + "epoch": 1.484906139015728, + "grad_norm": 5.09375, + "learning_rate": 1.7064453967340155e-06, + "loss": 1.54023848, + "memory(GiB)": 111.15, + "step": 58535, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.64928274, + "epoch": 1.485032978183663, + "grad_norm": 5.4375, + "learning_rate": 1.7056564887835664e-06, + "loss": 1.6009819, + "memory(GiB)": 111.15, + "step": 58540, + "train_speed(iter/s)": 0.447316 + }, + { + "acc": 0.63362703, + "epoch": 1.4851598173515983, + "grad_norm": 5.5625, + "learning_rate": 1.704867725733052e-06, + "loss": 1.71206017, + "memory(GiB)": 111.15, + "step": 58545, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.65298996, + "epoch": 1.4852866565195333, + "grad_norm": 5.65625, + "learning_rate": 1.7040791076171692e-06, + "loss": 1.61970043, + "memory(GiB)": 111.15, + "step": 58550, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.6494379, + "epoch": 1.4854134956874683, + "grad_norm": 4.71875, + "learning_rate": 1.7032906344706017e-06, + "loss": 1.61238956, + "memory(GiB)": 111.15, + "step": 58555, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.63166046, + "epoch": 1.4855403348554033, + "grad_norm": 5.09375, + "learning_rate": 1.7025023063280306e-06, + "loss": 1.66852989, + "memory(GiB)": 111.15, + "step": 58560, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.64479008, + "epoch": 1.4856671740233385, + "grad_norm": 4.90625, + "learning_rate": 1.701714123224128e-06, + "loss": 1.65162582, + "memory(GiB)": 111.15, + "step": 58565, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.63925319, + "epoch": 1.4857940131912735, + "grad_norm": 5.3125, + "learning_rate": 1.7009260851935684e-06, + "loss": 1.6378746, + "memory(GiB)": 111.15, + "step": 58570, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.63723459, + "epoch": 1.4859208523592085, + "grad_norm": 4.6875, + "learning_rate": 1.7001381922710064e-06, + "loss": 1.70705681, + "memory(GiB)": 111.15, + "step": 58575, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.64663901, + "epoch": 1.4860476915271437, + "grad_norm": 5.6875, + "learning_rate": 1.699350444491098e-06, + "loss": 1.60357533, + "memory(GiB)": 111.15, + "step": 58580, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.64261107, + "epoch": 1.4861745306950787, + "grad_norm": 4.90625, + "learning_rate": 1.6985628418884953e-06, + "loss": 1.66579819, + "memory(GiB)": 111.15, + "step": 58585, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.66025181, + "epoch": 1.4863013698630136, + "grad_norm": 5.6875, + "learning_rate": 1.6977753844978406e-06, + "loss": 1.63229198, + "memory(GiB)": 111.15, + "step": 58590, + "train_speed(iter/s)": 0.447518 + }, + { + "acc": 0.65551567, + "epoch": 1.4864282090309486, + "grad_norm": 5.53125, + "learning_rate": 1.696988072353764e-06, + "loss": 1.60150375, + "memory(GiB)": 111.15, + "step": 58595, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.64394646, + "epoch": 1.4865550481988838, + "grad_norm": 5.15625, + "learning_rate": 1.6962009054909007e-06, + "loss": 1.64184475, + "memory(GiB)": 111.15, + "step": 58600, + "train_speed(iter/s)": 0.447559 + }, + { + "epoch": 1.4865550481988838, + "eval_acc": 0.6433574788912623, + "eval_loss": 1.6075457334518433, + "eval_runtime": 113.997, + "eval_samples_per_second": 55.879, + "eval_steps_per_second": 27.939, + "step": 58600 + }, + { + "acc": 0.65527039, + "epoch": 1.4866818873668188, + "grad_norm": 4.4375, + "learning_rate": 1.6954138839438723e-06, + "loss": 1.588626, + "memory(GiB)": 111.15, + "step": 58605, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.65062342, + "epoch": 1.486808726534754, + "grad_norm": 5.8125, + "learning_rate": 1.6946270077472966e-06, + "loss": 1.66311703, + "memory(GiB)": 111.15, + "step": 58610, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.66131325, + "epoch": 1.486935565702689, + "grad_norm": 5.0, + "learning_rate": 1.6938402769357787e-06, + "loss": 1.55253773, + "memory(GiB)": 111.15, + "step": 58615, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.65563374, + "epoch": 1.487062404870624, + "grad_norm": 6.21875, + "learning_rate": 1.6930536915439288e-06, + "loss": 1.61415482, + "memory(GiB)": 111.15, + "step": 58620, + "train_speed(iter/s)": 0.447234 + }, + { + "acc": 0.64014258, + "epoch": 1.487189244038559, + "grad_norm": 6.34375, + "learning_rate": 1.6922672516063415e-06, + "loss": 1.62640533, + "memory(GiB)": 111.15, + "step": 58625, + "train_speed(iter/s)": 0.447255 + }, + { + "acc": 0.64583349, + "epoch": 1.4873160832064942, + "grad_norm": 4.6875, + "learning_rate": 1.6914809571576086e-06, + "loss": 1.62142868, + "memory(GiB)": 111.15, + "step": 58630, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.63985858, + "epoch": 1.4874429223744292, + "grad_norm": 5.1875, + "learning_rate": 1.6906948082323149e-06, + "loss": 1.70493793, + "memory(GiB)": 111.15, + "step": 58635, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.65323324, + "epoch": 1.4875697615423642, + "grad_norm": 5.40625, + "learning_rate": 1.689908804865038e-06, + "loss": 1.63549042, + "memory(GiB)": 111.15, + "step": 58640, + "train_speed(iter/s)": 0.447316 + }, + { + "acc": 0.6395916, + "epoch": 1.4876966007102994, + "grad_norm": 5.125, + "learning_rate": 1.6891229470903509e-06, + "loss": 1.66685715, + "memory(GiB)": 111.15, + "step": 58645, + "train_speed(iter/s)": 0.447336 + }, + { + "acc": 0.63920946, + "epoch": 1.4878234398782344, + "grad_norm": 5.90625, + "learning_rate": 1.6883372349428184e-06, + "loss": 1.70119934, + "memory(GiB)": 111.15, + "step": 58650, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.67151098, + "epoch": 1.4879502790461694, + "grad_norm": 6.125, + "learning_rate": 1.6875516684569999e-06, + "loss": 1.59584064, + "memory(GiB)": 111.15, + "step": 58655, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.65644789, + "epoch": 1.4880771182141044, + "grad_norm": 5.125, + "learning_rate": 1.686766247667448e-06, + "loss": 1.65553761, + "memory(GiB)": 111.15, + "step": 58660, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.65011902, + "epoch": 1.4882039573820396, + "grad_norm": 5.6875, + "learning_rate": 1.685980972608709e-06, + "loss": 1.62721691, + "memory(GiB)": 111.15, + "step": 58665, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.6438447, + "epoch": 1.4883307965499746, + "grad_norm": 4.84375, + "learning_rate": 1.6851958433153227e-06, + "loss": 1.5963026, + "memory(GiB)": 111.15, + "step": 58670, + "train_speed(iter/s)": 0.447438 + }, + { + "acc": 0.64123945, + "epoch": 1.4884576357179098, + "grad_norm": 5.90625, + "learning_rate": 1.684410859821823e-06, + "loss": 1.66902733, + "memory(GiB)": 111.15, + "step": 58675, + "train_speed(iter/s)": 0.447459 + }, + { + "acc": 0.66636167, + "epoch": 1.4885844748858448, + "grad_norm": 4.8125, + "learning_rate": 1.6836260221627364e-06, + "loss": 1.5628582, + "memory(GiB)": 111.15, + "step": 58680, + "train_speed(iter/s)": 0.447479 + }, + { + "acc": 0.6486165, + "epoch": 1.4887113140537798, + "grad_norm": 4.8125, + "learning_rate": 1.682841330372582e-06, + "loss": 1.60411892, + "memory(GiB)": 111.15, + "step": 58685, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.6537324, + "epoch": 1.4888381532217148, + "grad_norm": 4.46875, + "learning_rate": 1.6820567844858798e-06, + "loss": 1.55579014, + "memory(GiB)": 111.15, + "step": 58690, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.64398146, + "epoch": 1.48896499238965, + "grad_norm": 5.21875, + "learning_rate": 1.681272384537132e-06, + "loss": 1.61049881, + "memory(GiB)": 111.15, + "step": 58695, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.66264067, + "epoch": 1.489091831557585, + "grad_norm": 4.84375, + "learning_rate": 1.6804881305608423e-06, + "loss": 1.55516663, + "memory(GiB)": 111.15, + "step": 58700, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.489091831557585, + "eval_acc": 0.643339932664265, + "eval_loss": 1.6074788570404053, + "eval_runtime": 114.4897, + "eval_samples_per_second": 55.638, + "eval_steps_per_second": 27.819, + "step": 58700 + }, + { + "acc": 0.64075365, + "epoch": 1.4892186707255202, + "grad_norm": 5.3125, + "learning_rate": 1.679704022591503e-06, + "loss": 1.70118904, + "memory(GiB)": 111.15, + "step": 58705, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.64247441, + "epoch": 1.4893455098934552, + "grad_norm": 5.25, + "learning_rate": 1.678920060663608e-06, + "loss": 1.62132587, + "memory(GiB)": 111.15, + "step": 58710, + "train_speed(iter/s)": 0.447194 + }, + { + "acc": 0.64934163, + "epoch": 1.4894723490613901, + "grad_norm": 5.09375, + "learning_rate": 1.6781362448116344e-06, + "loss": 1.56671505, + "memory(GiB)": 111.15, + "step": 58715, + "train_speed(iter/s)": 0.447214 + }, + { + "acc": 0.64348011, + "epoch": 1.4895991882293251, + "grad_norm": 4.625, + "learning_rate": 1.6773525750700586e-06, + "loss": 1.6283226, + "memory(GiB)": 111.15, + "step": 58720, + "train_speed(iter/s)": 0.447234 + }, + { + "acc": 0.64833007, + "epoch": 1.4897260273972603, + "grad_norm": 6.0, + "learning_rate": 1.676569051473353e-06, + "loss": 1.62301273, + "memory(GiB)": 111.15, + "step": 58725, + "train_speed(iter/s)": 0.447254 + }, + { + "acc": 0.64574203, + "epoch": 1.4898528665651953, + "grad_norm": 5.09375, + "learning_rate": 1.6757856740559796e-06, + "loss": 1.61193886, + "memory(GiB)": 111.15, + "step": 58730, + "train_speed(iter/s)": 0.447274 + }, + { + "acc": 0.66367202, + "epoch": 1.4899797057331303, + "grad_norm": 4.78125, + "learning_rate": 1.6750024428523926e-06, + "loss": 1.55381012, + "memory(GiB)": 111.15, + "step": 58735, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.65655317, + "epoch": 1.4901065449010655, + "grad_norm": 4.34375, + "learning_rate": 1.6742193578970418e-06, + "loss": 1.61333961, + "memory(GiB)": 111.15, + "step": 58740, + "train_speed(iter/s)": 0.447315 + }, + { + "acc": 0.66071963, + "epoch": 1.4902333840690005, + "grad_norm": 6.0625, + "learning_rate": 1.673436419224373e-06, + "loss": 1.62763367, + "memory(GiB)": 111.15, + "step": 58745, + "train_speed(iter/s)": 0.447335 + }, + { + "acc": 0.63775368, + "epoch": 1.4903602232369355, + "grad_norm": 5.125, + "learning_rate": 1.6726536268688248e-06, + "loss": 1.69716396, + "memory(GiB)": 111.15, + "step": 58750, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.64823203, + "epoch": 1.4904870624048705, + "grad_norm": 5.84375, + "learning_rate": 1.671870980864822e-06, + "loss": 1.64590149, + "memory(GiB)": 111.15, + "step": 58755, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.65980506, + "epoch": 1.4906139015728057, + "grad_norm": 5.6875, + "learning_rate": 1.6710884812467943e-06, + "loss": 1.53621368, + "memory(GiB)": 111.15, + "step": 58760, + "train_speed(iter/s)": 0.447396 + }, + { + "acc": 0.65105319, + "epoch": 1.4907407407407407, + "grad_norm": 5.25, + "learning_rate": 1.6703061280491579e-06, + "loss": 1.64742012, + "memory(GiB)": 111.15, + "step": 58765, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.65881882, + "epoch": 1.490867579908676, + "grad_norm": 5.28125, + "learning_rate": 1.6695239213063237e-06, + "loss": 1.59869938, + "memory(GiB)": 111.15, + "step": 58770, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.65538664, + "epoch": 1.490994419076611, + "grad_norm": 6.25, + "learning_rate": 1.6687418610526972e-06, + "loss": 1.61019821, + "memory(GiB)": 111.15, + "step": 58775, + "train_speed(iter/s)": 0.447458 + }, + { + "acc": 0.66794777, + "epoch": 1.4911212582445459, + "grad_norm": 6.0, + "learning_rate": 1.6679599473226766e-06, + "loss": 1.51760597, + "memory(GiB)": 111.15, + "step": 58780, + "train_speed(iter/s)": 0.447478 + }, + { + "acc": 0.6487709, + "epoch": 1.4912480974124809, + "grad_norm": 5.78125, + "learning_rate": 1.6671781801506536e-06, + "loss": 1.68501759, + "memory(GiB)": 111.15, + "step": 58785, + "train_speed(iter/s)": 0.447498 + }, + { + "acc": 0.64713392, + "epoch": 1.491374936580416, + "grad_norm": 4.375, + "learning_rate": 1.6663965595710147e-06, + "loss": 1.60506897, + "memory(GiB)": 111.15, + "step": 58790, + "train_speed(iter/s)": 0.447519 + }, + { + "acc": 0.65909615, + "epoch": 1.491501775748351, + "grad_norm": 5.84375, + "learning_rate": 1.6656150856181386e-06, + "loss": 1.60577431, + "memory(GiB)": 111.15, + "step": 58795, + "train_speed(iter/s)": 0.447539 + }, + { + "acc": 0.64573736, + "epoch": 1.491628614916286, + "grad_norm": 5.1875, + "learning_rate": 1.6648337583263974e-06, + "loss": 1.67303524, + "memory(GiB)": 111.15, + "step": 58800, + "train_speed(iter/s)": 0.44756 + }, + { + "epoch": 1.491628614916286, + "eval_acc": 0.643380038325973, + "eval_loss": 1.607542872428894, + "eval_runtime": 114.2892, + "eval_samples_per_second": 55.736, + "eval_steps_per_second": 27.868, + "step": 58800 + }, + { + "acc": 0.6528266, + "epoch": 1.4917554540842213, + "grad_norm": 5.0625, + "learning_rate": 1.6640525777301586e-06, + "loss": 1.66997089, + "memory(GiB)": 111.15, + "step": 58805, + "train_speed(iter/s)": 0.447175 + }, + { + "acc": 0.65600142, + "epoch": 1.4918822932521563, + "grad_norm": 5.6875, + "learning_rate": 1.663271543863781e-06, + "loss": 1.61500511, + "memory(GiB)": 111.15, + "step": 58810, + "train_speed(iter/s)": 0.447195 + }, + { + "acc": 0.65778313, + "epoch": 1.4920091324200913, + "grad_norm": 5.6875, + "learning_rate": 1.6624906567616183e-06, + "loss": 1.65025902, + "memory(GiB)": 111.15, + "step": 58815, + "train_speed(iter/s)": 0.447216 + }, + { + "acc": 0.65066013, + "epoch": 1.4921359715880262, + "grad_norm": 5.25, + "learning_rate": 1.6617099164580175e-06, + "loss": 1.59101543, + "memory(GiB)": 111.15, + "step": 58820, + "train_speed(iter/s)": 0.447236 + }, + { + "acc": 0.64922814, + "epoch": 1.4922628107559615, + "grad_norm": 5.96875, + "learning_rate": 1.660929322987319e-06, + "loss": 1.60522156, + "memory(GiB)": 111.15, + "step": 58825, + "train_speed(iter/s)": 0.447256 + }, + { + "acc": 0.64697371, + "epoch": 1.4923896499238964, + "grad_norm": 6.15625, + "learning_rate": 1.660148876383857e-06, + "loss": 1.63818665, + "memory(GiB)": 111.15, + "step": 58830, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.66043024, + "epoch": 1.4925164890918317, + "grad_norm": 5.65625, + "learning_rate": 1.6593685766819584e-06, + "loss": 1.59930058, + "memory(GiB)": 111.15, + "step": 58835, + "train_speed(iter/s)": 0.447297 + }, + { + "acc": 0.64272523, + "epoch": 1.4926433282597666, + "grad_norm": 5.125, + "learning_rate": 1.658588423915945e-06, + "loss": 1.59380198, + "memory(GiB)": 111.15, + "step": 58840, + "train_speed(iter/s)": 0.447317 + }, + { + "acc": 0.65860901, + "epoch": 1.4927701674277016, + "grad_norm": 4.6875, + "learning_rate": 1.6578084181201293e-06, + "loss": 1.58588657, + "memory(GiB)": 111.15, + "step": 58845, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.64958196, + "epoch": 1.4928970065956366, + "grad_norm": 6.21875, + "learning_rate": 1.6570285593288242e-06, + "loss": 1.65560074, + "memory(GiB)": 111.15, + "step": 58850, + "train_speed(iter/s)": 0.447358 + }, + { + "acc": 0.64440637, + "epoch": 1.4930238457635718, + "grad_norm": 5.25, + "learning_rate": 1.6562488475763267e-06, + "loss": 1.65183945, + "memory(GiB)": 111.15, + "step": 58855, + "train_speed(iter/s)": 0.447378 + }, + { + "acc": 0.65378857, + "epoch": 1.4931506849315068, + "grad_norm": 6.96875, + "learning_rate": 1.6554692828969321e-06, + "loss": 1.63293591, + "memory(GiB)": 111.15, + "step": 58860, + "train_speed(iter/s)": 0.447399 + }, + { + "acc": 0.63899813, + "epoch": 1.493277524099442, + "grad_norm": 6.3125, + "learning_rate": 1.6546898653249326e-06, + "loss": 1.64469757, + "memory(GiB)": 111.15, + "step": 58865, + "train_speed(iter/s)": 0.447419 + }, + { + "acc": 0.64053564, + "epoch": 1.493404363267377, + "grad_norm": 5.09375, + "learning_rate": 1.65391059489461e-06, + "loss": 1.6877182, + "memory(GiB)": 111.15, + "step": 58870, + "train_speed(iter/s)": 0.44744 + }, + { + "acc": 0.64146304, + "epoch": 1.493531202435312, + "grad_norm": 4.5, + "learning_rate": 1.6531314716402369e-06, + "loss": 1.61634178, + "memory(GiB)": 111.15, + "step": 58875, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.64574404, + "epoch": 1.493658041603247, + "grad_norm": 5.4375, + "learning_rate": 1.652352495596083e-06, + "loss": 1.61793995, + "memory(GiB)": 111.15, + "step": 58880, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.65077391, + "epoch": 1.4937848807711822, + "grad_norm": 5.03125, + "learning_rate": 1.6515736667964144e-06, + "loss": 1.6186512, + "memory(GiB)": 111.15, + "step": 58885, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.65200624, + "epoch": 1.4939117199391172, + "grad_norm": 5.65625, + "learning_rate": 1.6507949852754867e-06, + "loss": 1.67626705, + "memory(GiB)": 111.15, + "step": 58890, + "train_speed(iter/s)": 0.447521 + }, + { + "acc": 0.65060053, + "epoch": 1.4940385591070522, + "grad_norm": 4.8125, + "learning_rate": 1.6500164510675453e-06, + "loss": 1.62553272, + "memory(GiB)": 111.15, + "step": 58895, + "train_speed(iter/s)": 0.447541 + }, + { + "acc": 0.66150885, + "epoch": 1.4941653982749874, + "grad_norm": 5.75, + "learning_rate": 1.649238064206839e-06, + "loss": 1.58504591, + "memory(GiB)": 111.15, + "step": 58900, + "train_speed(iter/s)": 0.447562 + }, + { + "epoch": 1.4941653982749874, + "eval_acc": 0.6433729362817122, + "eval_loss": 1.6075701713562012, + "eval_runtime": 114.0901, + "eval_samples_per_second": 55.833, + "eval_steps_per_second": 27.917, + "step": 58900 + }, + { + "acc": 0.65175886, + "epoch": 1.4942922374429224, + "grad_norm": 4.90625, + "learning_rate": 1.6484598247276023e-06, + "loss": 1.62185745, + "memory(GiB)": 111.15, + "step": 58905, + "train_speed(iter/s)": 0.447179 + }, + { + "acc": 0.63379216, + "epoch": 1.4944190766108574, + "grad_norm": 5.59375, + "learning_rate": 1.6476817326640682e-06, + "loss": 1.64575539, + "memory(GiB)": 111.15, + "step": 58910, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.66452446, + "epoch": 1.4945459157787924, + "grad_norm": 5.8125, + "learning_rate": 1.646903788050455e-06, + "loss": 1.57178249, + "memory(GiB)": 111.15, + "step": 58915, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.65218811, + "epoch": 1.4946727549467276, + "grad_norm": 4.84375, + "learning_rate": 1.6461259909209853e-06, + "loss": 1.59897509, + "memory(GiB)": 111.15, + "step": 58920, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.66757855, + "epoch": 1.4947995941146626, + "grad_norm": 5.15625, + "learning_rate": 1.6453483413098687e-06, + "loss": 1.62772751, + "memory(GiB)": 111.15, + "step": 58925, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.65187626, + "epoch": 1.4949264332825978, + "grad_norm": 5.28125, + "learning_rate": 1.6445708392513093e-06, + "loss": 1.67433796, + "memory(GiB)": 111.15, + "step": 58930, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.63122959, + "epoch": 1.4950532724505328, + "grad_norm": 5.03125, + "learning_rate": 1.643793484779505e-06, + "loss": 1.68151588, + "memory(GiB)": 111.15, + "step": 58935, + "train_speed(iter/s)": 0.4473 + }, + { + "acc": 0.66642361, + "epoch": 1.4951801116184678, + "grad_norm": 7.1875, + "learning_rate": 1.6430162779286484e-06, + "loss": 1.59751701, + "memory(GiB)": 111.15, + "step": 58940, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.65998974, + "epoch": 1.4953069507864027, + "grad_norm": 4.625, + "learning_rate": 1.6422392187329233e-06, + "loss": 1.55916405, + "memory(GiB)": 111.15, + "step": 58945, + "train_speed(iter/s)": 0.447341 + }, + { + "acc": 0.63830013, + "epoch": 1.495433789954338, + "grad_norm": 6.84375, + "learning_rate": 1.6414623072265085e-06, + "loss": 1.6973381, + "memory(GiB)": 111.15, + "step": 58950, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.65500784, + "epoch": 1.495560629122273, + "grad_norm": 5.4375, + "learning_rate": 1.6406855434435765e-06, + "loss": 1.602981, + "memory(GiB)": 111.15, + "step": 58955, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.63981161, + "epoch": 1.495687468290208, + "grad_norm": 4.8125, + "learning_rate": 1.6399089274182922e-06, + "loss": 1.64110203, + "memory(GiB)": 111.15, + "step": 58960, + "train_speed(iter/s)": 0.447401 + }, + { + "acc": 0.66140385, + "epoch": 1.4958143074581431, + "grad_norm": 4.5625, + "learning_rate": 1.6391324591848156e-06, + "loss": 1.60298843, + "memory(GiB)": 111.15, + "step": 58965, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.65684719, + "epoch": 1.4959411466260781, + "grad_norm": 4.90625, + "learning_rate": 1.6383561387772984e-06, + "loss": 1.5963357, + "memory(GiB)": 111.15, + "step": 58970, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.64651155, + "epoch": 1.4960679857940131, + "grad_norm": 4.96875, + "learning_rate": 1.6375799662298868e-06, + "loss": 1.63111916, + "memory(GiB)": 111.15, + "step": 58975, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.63888979, + "epoch": 1.4961948249619481, + "grad_norm": 6.0, + "learning_rate": 1.6368039415767201e-06, + "loss": 1.66575413, + "memory(GiB)": 111.15, + "step": 58980, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.64384251, + "epoch": 1.4963216641298833, + "grad_norm": 5.5625, + "learning_rate": 1.6360280648519305e-06, + "loss": 1.63611355, + "memory(GiB)": 111.15, + "step": 58985, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.66336031, + "epoch": 1.4964485032978183, + "grad_norm": 5.5, + "learning_rate": 1.6352523360896488e-06, + "loss": 1.5744853, + "memory(GiB)": 111.15, + "step": 58990, + "train_speed(iter/s)": 0.447523 + }, + { + "acc": 0.64527378, + "epoch": 1.4965753424657535, + "grad_norm": 4.90625, + "learning_rate": 1.63447675532399e-06, + "loss": 1.64541473, + "memory(GiB)": 111.15, + "step": 58995, + "train_speed(iter/s)": 0.447543 + }, + { + "acc": 0.65704832, + "epoch": 1.4967021816336885, + "grad_norm": 5.53125, + "learning_rate": 1.6337013225890698e-06, + "loss": 1.67099915, + "memory(GiB)": 111.15, + "step": 59000, + "train_speed(iter/s)": 0.447564 + }, + { + "epoch": 1.4967021816336885, + "eval_acc": 0.6434243216607755, + "eval_loss": 1.6074501276016235, + "eval_runtime": 114.2733, + "eval_samples_per_second": 55.744, + "eval_steps_per_second": 27.872, + "step": 59000 + }, + { + "acc": 0.66162496, + "epoch": 1.4968290208016235, + "grad_norm": 4.65625, + "learning_rate": 1.6329260379189932e-06, + "loss": 1.60008945, + "memory(GiB)": 111.15, + "step": 59005, + "train_speed(iter/s)": 0.447181 + }, + { + "acc": 0.64277115, + "epoch": 1.4969558599695585, + "grad_norm": 5.34375, + "learning_rate": 1.6321509013478653e-06, + "loss": 1.60472832, + "memory(GiB)": 111.15, + "step": 59010, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.6540834, + "epoch": 1.4970826991374937, + "grad_norm": 5.0, + "learning_rate": 1.6313759129097757e-06, + "loss": 1.64834995, + "memory(GiB)": 111.15, + "step": 59015, + "train_speed(iter/s)": 0.447221 + }, + { + "acc": 0.64342494, + "epoch": 1.4972095383054287, + "grad_norm": 5.1875, + "learning_rate": 1.6306010726388117e-06, + "loss": 1.64771881, + "memory(GiB)": 111.15, + "step": 59020, + "train_speed(iter/s)": 0.447241 + }, + { + "acc": 0.65083895, + "epoch": 1.497336377473364, + "grad_norm": 6.1875, + "learning_rate": 1.6298263805690573e-06, + "loss": 1.69162369, + "memory(GiB)": 111.15, + "step": 59025, + "train_speed(iter/s)": 0.447261 + }, + { + "acc": 0.64993916, + "epoch": 1.4974632166412989, + "grad_norm": 5.09375, + "learning_rate": 1.629051836734587e-06, + "loss": 1.60448723, + "memory(GiB)": 111.15, + "step": 59030, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.67214923, + "epoch": 1.4975900558092339, + "grad_norm": 5.8125, + "learning_rate": 1.6282774411694641e-06, + "loss": 1.57025127, + "memory(GiB)": 111.15, + "step": 59035, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.65586977, + "epoch": 1.4977168949771689, + "grad_norm": 6.3125, + "learning_rate": 1.6275031939077545e-06, + "loss": 1.58758488, + "memory(GiB)": 111.15, + "step": 59040, + "train_speed(iter/s)": 0.447322 + }, + { + "acc": 0.66174603, + "epoch": 1.497843734145104, + "grad_norm": 5.0, + "learning_rate": 1.6267290949835119e-06, + "loss": 1.58515339, + "memory(GiB)": 111.15, + "step": 59045, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.640312, + "epoch": 1.497970573313039, + "grad_norm": 4.6875, + "learning_rate": 1.6259551444307852e-06, + "loss": 1.60288696, + "memory(GiB)": 111.15, + "step": 59050, + "train_speed(iter/s)": 0.447362 + }, + { + "acc": 0.63710012, + "epoch": 1.498097412480974, + "grad_norm": 5.3125, + "learning_rate": 1.6251813422836127e-06, + "loss": 1.65021019, + "memory(GiB)": 111.15, + "step": 59055, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.64737806, + "epoch": 1.4982242516489093, + "grad_norm": 5.8125, + "learning_rate": 1.6244076885760334e-06, + "loss": 1.63079109, + "memory(GiB)": 111.15, + "step": 59060, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.64907961, + "epoch": 1.4983510908168443, + "grad_norm": 5.25, + "learning_rate": 1.6236341833420755e-06, + "loss": 1.66874638, + "memory(GiB)": 111.15, + "step": 59065, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.6475049, + "epoch": 1.4984779299847792, + "grad_norm": 5.21875, + "learning_rate": 1.6228608266157596e-06, + "loss": 1.59829865, + "memory(GiB)": 111.15, + "step": 59070, + "train_speed(iter/s)": 0.447444 + }, + { + "acc": 0.65697751, + "epoch": 1.4986047691527142, + "grad_norm": 5.03125, + "learning_rate": 1.6220876184311034e-06, + "loss": 1.57717991, + "memory(GiB)": 111.15, + "step": 59075, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.64087934, + "epoch": 1.4987316083206494, + "grad_norm": 6.375, + "learning_rate": 1.6213145588221146e-06, + "loss": 1.61150875, + "memory(GiB)": 111.15, + "step": 59080, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.64939518, + "epoch": 1.4988584474885844, + "grad_norm": 6.125, + "learning_rate": 1.620541647822796e-06, + "loss": 1.57067347, + "memory(GiB)": 111.15, + "step": 59085, + "train_speed(iter/s)": 0.447504 + }, + { + "acc": 0.65348129, + "epoch": 1.4989852866565196, + "grad_norm": 5.90625, + "learning_rate": 1.6197688854671444e-06, + "loss": 1.65342979, + "memory(GiB)": 111.15, + "step": 59090, + "train_speed(iter/s)": 0.447524 + }, + { + "acc": 0.66224031, + "epoch": 1.4991121258244546, + "grad_norm": 5.0625, + "learning_rate": 1.6189962717891484e-06, + "loss": 1.56247854, + "memory(GiB)": 111.15, + "step": 59095, + "train_speed(iter/s)": 0.447544 + }, + { + "acc": 0.63340626, + "epoch": 1.4992389649923896, + "grad_norm": 5.3125, + "learning_rate": 1.6182238068227917e-06, + "loss": 1.7024189, + "memory(GiB)": 111.15, + "step": 59100, + "train_speed(iter/s)": 0.447565 + }, + { + "epoch": 1.4992389649923896, + "eval_acc": 0.6434076109683972, + "eval_loss": 1.6075934171676636, + "eval_runtime": 113.5484, + "eval_samples_per_second": 56.099, + "eval_steps_per_second": 28.05, + "step": 59100 + }, + { + "acc": 0.64163122, + "epoch": 1.4993658041603246, + "grad_norm": 5.15625, + "learning_rate": 1.6174514906020505e-06, + "loss": 1.63408623, + "memory(GiB)": 111.15, + "step": 59105, + "train_speed(iter/s)": 0.447185 + }, + { + "acc": 0.66418266, + "epoch": 1.4994926433282598, + "grad_norm": 4.5, + "learning_rate": 1.6166793231608952e-06, + "loss": 1.5339241, + "memory(GiB)": 111.15, + "step": 59110, + "train_speed(iter/s)": 0.447205 + }, + { + "acc": 0.64294286, + "epoch": 1.4996194824961948, + "grad_norm": 5.40625, + "learning_rate": 1.615907304533288e-06, + "loss": 1.67296543, + "memory(GiB)": 111.15, + "step": 59115, + "train_speed(iter/s)": 0.447225 + }, + { + "acc": 0.64718866, + "epoch": 1.4997463216641298, + "grad_norm": 5.15625, + "learning_rate": 1.6151354347531868e-06, + "loss": 1.65730438, + "memory(GiB)": 111.15, + "step": 59120, + "train_speed(iter/s)": 0.447245 + }, + { + "acc": 0.65538254, + "epoch": 1.499873160832065, + "grad_norm": 6.0, + "learning_rate": 1.61436371385454e-06, + "loss": 1.65555267, + "memory(GiB)": 111.15, + "step": 59125, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.65258832, + "epoch": 1.5, + "grad_norm": 5.8125, + "learning_rate": 1.6135921418712959e-06, + "loss": 1.63677311, + "memory(GiB)": 111.15, + "step": 59130, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.65723395, + "epoch": 1.500126839167935, + "grad_norm": 5.875, + "learning_rate": 1.6128207188373867e-06, + "loss": 1.63065567, + "memory(GiB)": 111.15, + "step": 59135, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.64340911, + "epoch": 1.50025367833587, + "grad_norm": 4.53125, + "learning_rate": 1.6120494447867451e-06, + "loss": 1.64091949, + "memory(GiB)": 111.15, + "step": 59140, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.65249686, + "epoch": 1.5003805175038052, + "grad_norm": 7.6875, + "learning_rate": 1.6112783197532932e-06, + "loss": 1.53547192, + "memory(GiB)": 111.15, + "step": 59145, + "train_speed(iter/s)": 0.447346 + }, + { + "acc": 0.64765186, + "epoch": 1.5005073566717404, + "grad_norm": 5.34375, + "learning_rate": 1.6105073437709545e-06, + "loss": 1.62546616, + "memory(GiB)": 111.15, + "step": 59150, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.63824434, + "epoch": 1.5006341958396754, + "grad_norm": 5.5625, + "learning_rate": 1.6097365168736335e-06, + "loss": 1.66942368, + "memory(GiB)": 111.15, + "step": 59155, + "train_speed(iter/s)": 0.447387 + }, + { + "acc": 0.6428421, + "epoch": 1.5007610350076104, + "grad_norm": 5.15625, + "learning_rate": 1.6089658390952351e-06, + "loss": 1.62978115, + "memory(GiB)": 111.15, + "step": 59160, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.65597262, + "epoch": 1.5008878741755454, + "grad_norm": 5.0625, + "learning_rate": 1.6081953104696612e-06, + "loss": 1.63064842, + "memory(GiB)": 111.15, + "step": 59165, + "train_speed(iter/s)": 0.447427 + }, + { + "acc": 0.66427107, + "epoch": 1.5010147133434804, + "grad_norm": 5.03125, + "learning_rate": 1.6074249310308021e-06, + "loss": 1.62294178, + "memory(GiB)": 111.15, + "step": 59170, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.65725474, + "epoch": 1.5011415525114156, + "grad_norm": 5.53125, + "learning_rate": 1.6066547008125399e-06, + "loss": 1.60215893, + "memory(GiB)": 111.15, + "step": 59175, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.65086551, + "epoch": 1.5012683916793506, + "grad_norm": 5.46875, + "learning_rate": 1.6058846198487522e-06, + "loss": 1.6136219, + "memory(GiB)": 111.15, + "step": 59180, + "train_speed(iter/s)": 0.447488 + }, + { + "acc": 0.63740582, + "epoch": 1.5013952308472858, + "grad_norm": 5.0, + "learning_rate": 1.6051146881733142e-06, + "loss": 1.57896976, + "memory(GiB)": 111.15, + "step": 59185, + "train_speed(iter/s)": 0.447508 + }, + { + "acc": 0.64791689, + "epoch": 1.5015220700152208, + "grad_norm": 6.65625, + "learning_rate": 1.6043449058200916e-06, + "loss": 1.65219879, + "memory(GiB)": 111.15, + "step": 59190, + "train_speed(iter/s)": 0.44751 + }, + { + "acc": 0.65278883, + "epoch": 1.5016489091831557, + "grad_norm": 5.1875, + "learning_rate": 1.6035752728229364e-06, + "loss": 1.60292187, + "memory(GiB)": 111.15, + "step": 59195, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.6440002, + "epoch": 1.5017757483510907, + "grad_norm": 4.53125, + "learning_rate": 1.6028057892157067e-06, + "loss": 1.71200867, + "memory(GiB)": 111.15, + "step": 59200, + "train_speed(iter/s)": 0.44755 + }, + { + "epoch": 1.5017757483510907, + "eval_acc": 0.6434230683588471, + "eval_loss": 1.6075836420059204, + "eval_runtime": 114.3424, + "eval_samples_per_second": 55.71, + "eval_steps_per_second": 27.855, + "step": 59200 + }, + { + "acc": 0.65478663, + "epoch": 1.5019025875190257, + "grad_norm": 5.09375, + "learning_rate": 1.602036455032246e-06, + "loss": 1.59546013, + "memory(GiB)": 111.15, + "step": 59205, + "train_speed(iter/s)": 0.447168 + }, + { + "acc": 0.65146456, + "epoch": 1.502029426686961, + "grad_norm": 5.84375, + "learning_rate": 1.6012672703063925e-06, + "loss": 1.65096436, + "memory(GiB)": 111.15, + "step": 59210, + "train_speed(iter/s)": 0.447188 + }, + { + "acc": 0.64565725, + "epoch": 1.5021562658548961, + "grad_norm": 5.90625, + "learning_rate": 1.600498235071979e-06, + "loss": 1.67426682, + "memory(GiB)": 111.15, + "step": 59215, + "train_speed(iter/s)": 0.447208 + }, + { + "acc": 0.62971859, + "epoch": 1.5022831050228311, + "grad_norm": 5.4375, + "learning_rate": 1.5997293493628301e-06, + "loss": 1.68267365, + "memory(GiB)": 111.15, + "step": 59220, + "train_speed(iter/s)": 0.447228 + }, + { + "acc": 0.64857264, + "epoch": 1.5024099441907661, + "grad_norm": 5.8125, + "learning_rate": 1.598960613212766e-06, + "loss": 1.6305584, + "memory(GiB)": 111.15, + "step": 59225, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.65766907, + "epoch": 1.5025367833587011, + "grad_norm": 5.21875, + "learning_rate": 1.598192026655599e-06, + "loss": 1.65459366, + "memory(GiB)": 111.15, + "step": 59230, + "train_speed(iter/s)": 0.447268 + }, + { + "acc": 0.63888845, + "epoch": 1.502663622526636, + "grad_norm": 5.40625, + "learning_rate": 1.5974235897251344e-06, + "loss": 1.62067547, + "memory(GiB)": 111.15, + "step": 59235, + "train_speed(iter/s)": 0.447288 + }, + { + "acc": 0.64546776, + "epoch": 1.5027904616945713, + "grad_norm": 6.625, + "learning_rate": 1.5966553024551717e-06, + "loss": 1.66576271, + "memory(GiB)": 111.15, + "step": 59240, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.64655609, + "epoch": 1.5029173008625063, + "grad_norm": 5.34375, + "learning_rate": 1.5958871648795032e-06, + "loss": 1.64487381, + "memory(GiB)": 111.15, + "step": 59245, + "train_speed(iter/s)": 0.447329 + }, + { + "acc": 0.66194792, + "epoch": 1.5030441400304415, + "grad_norm": 5.21875, + "learning_rate": 1.5951191770319164e-06, + "loss": 1.60104122, + "memory(GiB)": 111.15, + "step": 59250, + "train_speed(iter/s)": 0.447349 + }, + { + "acc": 0.65135441, + "epoch": 1.5031709791983765, + "grad_norm": 5.90625, + "learning_rate": 1.594351338946189e-06, + "loss": 1.62612762, + "memory(GiB)": 111.15, + "step": 59255, + "train_speed(iter/s)": 0.447369 + }, + { + "acc": 0.66140075, + "epoch": 1.5032978183663115, + "grad_norm": 5.71875, + "learning_rate": 1.5935836506560953e-06, + "loss": 1.62480736, + "memory(GiB)": 111.15, + "step": 59260, + "train_speed(iter/s)": 0.447389 + }, + { + "acc": 0.65046701, + "epoch": 1.5034246575342465, + "grad_norm": 4.6875, + "learning_rate": 1.5928161121954012e-06, + "loss": 1.57120018, + "memory(GiB)": 111.15, + "step": 59265, + "train_speed(iter/s)": 0.447409 + }, + { + "acc": 0.66082315, + "epoch": 1.5035514967021817, + "grad_norm": 4.90625, + "learning_rate": 1.592048723597866e-06, + "loss": 1.5599577, + "memory(GiB)": 111.15, + "step": 59270, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.67174711, + "epoch": 1.5036783358701167, + "grad_norm": 6.34375, + "learning_rate": 1.591281484897244e-06, + "loss": 1.56362839, + "memory(GiB)": 111.15, + "step": 59275, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.6585897, + "epoch": 1.5038051750380519, + "grad_norm": 5.90625, + "learning_rate": 1.5905143961272807e-06, + "loss": 1.63690643, + "memory(GiB)": 111.15, + "step": 59280, + "train_speed(iter/s)": 0.44747 + }, + { + "acc": 0.65825686, + "epoch": 1.5039320142059869, + "grad_norm": 5.75, + "learning_rate": 1.5897474573217153e-06, + "loss": 1.52825489, + "memory(GiB)": 111.15, + "step": 59285, + "train_speed(iter/s)": 0.44749 + }, + { + "acc": 0.6467422, + "epoch": 1.5040588533739219, + "grad_norm": 6.34375, + "learning_rate": 1.588980668514285e-06, + "loss": 1.64120712, + "memory(GiB)": 111.15, + "step": 59290, + "train_speed(iter/s)": 0.44751 + }, + { + "acc": 0.6395689, + "epoch": 1.5041856925418569, + "grad_norm": 4.96875, + "learning_rate": 1.5882140297387127e-06, + "loss": 1.59772911, + "memory(GiB)": 111.15, + "step": 59295, + "train_speed(iter/s)": 0.44753 + }, + { + "acc": 0.64974747, + "epoch": 1.5043125317097918, + "grad_norm": 5.09375, + "learning_rate": 1.5874475410287189e-06, + "loss": 1.66569443, + "memory(GiB)": 111.15, + "step": 59300, + "train_speed(iter/s)": 0.44755 + }, + { + "epoch": 1.5043125317097918, + "eval_acc": 0.6434197262203715, + "eval_loss": 1.607558012008667, + "eval_runtime": 115.0951, + "eval_samples_per_second": 55.346, + "eval_steps_per_second": 27.673, + "step": 59300 + }, + { + "acc": 0.65725527, + "epoch": 1.504439370877727, + "grad_norm": 5.3125, + "learning_rate": 1.586681202418019e-06, + "loss": 1.64224167, + "memory(GiB)": 111.15, + "step": 59305, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.64735832, + "epoch": 1.5045662100456623, + "grad_norm": 5.78125, + "learning_rate": 1.5859150139403212e-06, + "loss": 1.71872864, + "memory(GiB)": 111.15, + "step": 59310, + "train_speed(iter/s)": 0.447185 + }, + { + "acc": 0.66001391, + "epoch": 1.5046930492135973, + "grad_norm": 5.1875, + "learning_rate": 1.585148975629322e-06, + "loss": 1.56666279, + "memory(GiB)": 111.15, + "step": 59315, + "train_speed(iter/s)": 0.447205 + }, + { + "acc": 0.64274001, + "epoch": 1.5048198883815322, + "grad_norm": 5.625, + "learning_rate": 1.5843830875187155e-06, + "loss": 1.62955971, + "memory(GiB)": 111.15, + "step": 59320, + "train_speed(iter/s)": 0.447225 + }, + { + "acc": 0.6589869, + "epoch": 1.5049467275494672, + "grad_norm": 5.65625, + "learning_rate": 1.583617349642192e-06, + "loss": 1.66563454, + "memory(GiB)": 111.15, + "step": 59325, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.64371648, + "epoch": 1.5050735667174022, + "grad_norm": 4.59375, + "learning_rate": 1.5828517620334322e-06, + "loss": 1.66821556, + "memory(GiB)": 111.15, + "step": 59330, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.65946574, + "epoch": 1.5052004058853374, + "grad_norm": 5.03125, + "learning_rate": 1.5820863247261054e-06, + "loss": 1.61371861, + "memory(GiB)": 111.15, + "step": 59335, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.6594264, + "epoch": 1.5053272450532724, + "grad_norm": 5.1875, + "learning_rate": 1.5813210377538834e-06, + "loss": 1.60673714, + "memory(GiB)": 111.15, + "step": 59340, + "train_speed(iter/s)": 0.447303 + }, + { + "acc": 0.64570751, + "epoch": 1.5054540842212076, + "grad_norm": 4.78125, + "learning_rate": 1.5805559011504252e-06, + "loss": 1.65724392, + "memory(GiB)": 111.15, + "step": 59345, + "train_speed(iter/s)": 0.447323 + }, + { + "acc": 0.65474682, + "epoch": 1.5055809233891426, + "grad_norm": 5.09375, + "learning_rate": 1.5797909149493873e-06, + "loss": 1.53960524, + "memory(GiB)": 111.15, + "step": 59350, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.6342164, + "epoch": 1.5057077625570776, + "grad_norm": 5.0, + "learning_rate": 1.5790260791844114e-06, + "loss": 1.70316658, + "memory(GiB)": 111.15, + "step": 59355, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.66093311, + "epoch": 1.5058346017250126, + "grad_norm": 4.875, + "learning_rate": 1.5782613938891438e-06, + "loss": 1.60034142, + "memory(GiB)": 111.15, + "step": 59360, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.67324772, + "epoch": 1.5059614408929476, + "grad_norm": 6.8125, + "learning_rate": 1.5774968590972172e-06, + "loss": 1.59896221, + "memory(GiB)": 111.15, + "step": 59365, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.6498364, + "epoch": 1.5060882800608828, + "grad_norm": 5.375, + "learning_rate": 1.5767324748422592e-06, + "loss": 1.61642685, + "memory(GiB)": 111.15, + "step": 59370, + "train_speed(iter/s)": 0.447421 + }, + { + "acc": 0.65688219, + "epoch": 1.506215119228818, + "grad_norm": 6.28125, + "learning_rate": 1.5759682411578909e-06, + "loss": 1.62163429, + "memory(GiB)": 111.15, + "step": 59375, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.65494771, + "epoch": 1.506341958396753, + "grad_norm": 5.25, + "learning_rate": 1.575204158077726e-06, + "loss": 1.66645641, + "memory(GiB)": 111.15, + "step": 59380, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.66841612, + "epoch": 1.506468797564688, + "grad_norm": 5.46875, + "learning_rate": 1.574440225635373e-06, + "loss": 1.58660507, + "memory(GiB)": 111.15, + "step": 59385, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.65463037, + "epoch": 1.506595636732623, + "grad_norm": 4.78125, + "learning_rate": 1.5736764438644332e-06, + "loss": 1.58031406, + "memory(GiB)": 111.15, + "step": 59390, + "train_speed(iter/s)": 0.4475 + }, + { + "acc": 0.65850039, + "epoch": 1.506722475900558, + "grad_norm": 6.0625, + "learning_rate": 1.5729128127985004e-06, + "loss": 1.57293482, + "memory(GiB)": 111.15, + "step": 59395, + "train_speed(iter/s)": 0.44752 + }, + { + "acc": 0.63598795, + "epoch": 1.5068493150684932, + "grad_norm": 4.90625, + "learning_rate": 1.5721493324711633e-06, + "loss": 1.64439545, + "memory(GiB)": 111.15, + "step": 59400, + "train_speed(iter/s)": 0.44754 + }, + { + "epoch": 1.5068493150684932, + "eval_acc": 0.6433996733895174, + "eval_loss": 1.6075469255447388, + "eval_runtime": 113.8077, + "eval_samples_per_second": 55.972, + "eval_steps_per_second": 27.986, + "step": 59400 + }, + { + "acc": 0.6501956, + "epoch": 1.5069761542364282, + "grad_norm": 5.125, + "learning_rate": 1.5713860029160028e-06, + "loss": 1.62249107, + "memory(GiB)": 111.15, + "step": 59405, + "train_speed(iter/s)": 0.44716 + }, + { + "acc": 0.63989925, + "epoch": 1.5071029934043634, + "grad_norm": 5.65625, + "learning_rate": 1.5706228241665932e-06, + "loss": 1.65936127, + "memory(GiB)": 111.15, + "step": 59410, + "train_speed(iter/s)": 0.44718 + }, + { + "acc": 0.65798922, + "epoch": 1.5072298325722984, + "grad_norm": 4.59375, + "learning_rate": 1.5698597962565032e-06, + "loss": 1.63347626, + "memory(GiB)": 111.15, + "step": 59415, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.66332407, + "epoch": 1.5073566717402334, + "grad_norm": 5.78125, + "learning_rate": 1.5690969192192933e-06, + "loss": 1.54536333, + "memory(GiB)": 111.15, + "step": 59420, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.65890369, + "epoch": 1.5074835109081683, + "grad_norm": 4.28125, + "learning_rate": 1.5683341930885183e-06, + "loss": 1.57902336, + "memory(GiB)": 111.15, + "step": 59425, + "train_speed(iter/s)": 0.447239 + }, + { + "acc": 0.64195957, + "epoch": 1.5076103500761036, + "grad_norm": 5.59375, + "learning_rate": 1.567571617897729e-06, + "loss": 1.67013035, + "memory(GiB)": 111.15, + "step": 59430, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.65936074, + "epoch": 1.5077371892440385, + "grad_norm": 4.5, + "learning_rate": 1.566809193680463e-06, + "loss": 1.54908953, + "memory(GiB)": 111.15, + "step": 59435, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.64404936, + "epoch": 1.5078640284119738, + "grad_norm": 6.0, + "learning_rate": 1.566046920470257e-06, + "loss": 1.71743851, + "memory(GiB)": 111.15, + "step": 59440, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.66367149, + "epoch": 1.5079908675799087, + "grad_norm": 4.59375, + "learning_rate": 1.5652847983006376e-06, + "loss": 1.57582626, + "memory(GiB)": 111.15, + "step": 59445, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.65568686, + "epoch": 1.5081177067478437, + "grad_norm": 4.53125, + "learning_rate": 1.564522827205131e-06, + "loss": 1.60464535, + "memory(GiB)": 111.15, + "step": 59450, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.64674211, + "epoch": 1.5082445459157787, + "grad_norm": 5.28125, + "learning_rate": 1.5637610072172464e-06, + "loss": 1.62248478, + "memory(GiB)": 111.15, + "step": 59455, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.64776793, + "epoch": 1.5083713850837137, + "grad_norm": 5.4375, + "learning_rate": 1.5629993383704933e-06, + "loss": 1.72595825, + "memory(GiB)": 111.15, + "step": 59460, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.65464034, + "epoch": 1.508498224251649, + "grad_norm": 5.90625, + "learning_rate": 1.5622378206983764e-06, + "loss": 1.65297661, + "memory(GiB)": 111.15, + "step": 59465, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.65673046, + "epoch": 1.5086250634195841, + "grad_norm": 5.5625, + "learning_rate": 1.5614764542343896e-06, + "loss": 1.58422899, + "memory(GiB)": 111.15, + "step": 59470, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.65106068, + "epoch": 1.5087519025875191, + "grad_norm": 5.09375, + "learning_rate": 1.5607152390120173e-06, + "loss": 1.58888397, + "memory(GiB)": 111.15, + "step": 59475, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.64625835, + "epoch": 1.5088787417554541, + "grad_norm": 4.96875, + "learning_rate": 1.5599541750647457e-06, + "loss": 1.64378624, + "memory(GiB)": 111.15, + "step": 59480, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64896345, + "epoch": 1.509005580923389, + "grad_norm": 5.125, + "learning_rate": 1.559193262426048e-06, + "loss": 1.58713417, + "memory(GiB)": 111.15, + "step": 59485, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.65284748, + "epoch": 1.509132420091324, + "grad_norm": 4.9375, + "learning_rate": 1.5584325011293943e-06, + "loss": 1.61546173, + "memory(GiB)": 111.15, + "step": 59490, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.64500017, + "epoch": 1.5092592592592593, + "grad_norm": 5.28125, + "learning_rate": 1.5576718912082417e-06, + "loss": 1.66601677, + "memory(GiB)": 111.15, + "step": 59495, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.6545754, + "epoch": 1.5093860984271943, + "grad_norm": 5.5, + "learning_rate": 1.5569114326960494e-06, + "loss": 1.64743404, + "memory(GiB)": 111.15, + "step": 59500, + "train_speed(iter/s)": 0.447535 + }, + { + "epoch": 1.5093860984271943, + "eval_acc": 0.6434059398991593, + "eval_loss": 1.6074860095977783, + "eval_runtime": 113.8913, + "eval_samples_per_second": 55.931, + "eval_steps_per_second": 27.965, + "step": 59500 + }, + { + "acc": 0.65407767, + "epoch": 1.5095129375951295, + "grad_norm": 5.53125, + "learning_rate": 1.5561511256262651e-06, + "loss": 1.61545963, + "memory(GiB)": 111.15, + "step": 59505, + "train_speed(iter/s)": 0.447156 + }, + { + "acc": 0.63908911, + "epoch": 1.5096397767630645, + "grad_norm": 5.03125, + "learning_rate": 1.55539097003233e-06, + "loss": 1.64478874, + "memory(GiB)": 111.15, + "step": 59510, + "train_speed(iter/s)": 0.447176 + }, + { + "acc": 0.65877566, + "epoch": 1.5097666159309995, + "grad_norm": 5.0, + "learning_rate": 1.5546309659476788e-06, + "loss": 1.60434875, + "memory(GiB)": 111.15, + "step": 59515, + "train_speed(iter/s)": 0.447196 + }, + { + "acc": 0.66534214, + "epoch": 1.5098934550989345, + "grad_norm": 4.9375, + "learning_rate": 1.55387111340574e-06, + "loss": 1.60407372, + "memory(GiB)": 111.15, + "step": 59520, + "train_speed(iter/s)": 0.447215 + }, + { + "acc": 0.6587709, + "epoch": 1.5100202942668695, + "grad_norm": 5.65625, + "learning_rate": 1.553111412439936e-06, + "loss": 1.60242386, + "memory(GiB)": 111.15, + "step": 59525, + "train_speed(iter/s)": 0.447235 + }, + { + "acc": 0.64898419, + "epoch": 1.5101471334348047, + "grad_norm": 5.15625, + "learning_rate": 1.5523518630836809e-06, + "loss": 1.6708046, + "memory(GiB)": 111.15, + "step": 59530, + "train_speed(iter/s)": 0.447255 + }, + { + "acc": 0.66300507, + "epoch": 1.5102739726027399, + "grad_norm": 6.03125, + "learning_rate": 1.551592465370384e-06, + "loss": 1.53205509, + "memory(GiB)": 111.15, + "step": 59535, + "train_speed(iter/s)": 0.447275 + }, + { + "acc": 0.63613396, + "epoch": 1.5104008117706749, + "grad_norm": 5.5625, + "learning_rate": 1.5508332193334457e-06, + "loss": 1.60356712, + "memory(GiB)": 111.15, + "step": 59540, + "train_speed(iter/s)": 0.447294 + }, + { + "acc": 0.64130106, + "epoch": 1.5105276509386099, + "grad_norm": 5.25, + "learning_rate": 1.5500741250062628e-06, + "loss": 1.69565334, + "memory(GiB)": 111.15, + "step": 59545, + "train_speed(iter/s)": 0.447314 + }, + { + "acc": 0.67670002, + "epoch": 1.5106544901065448, + "grad_norm": 5.6875, + "learning_rate": 1.549315182422222e-06, + "loss": 1.56400166, + "memory(GiB)": 111.15, + "step": 59550, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.66014585, + "epoch": 1.5107813292744798, + "grad_norm": 4.46875, + "learning_rate": 1.5485563916147062e-06, + "loss": 1.63977718, + "memory(GiB)": 111.15, + "step": 59555, + "train_speed(iter/s)": 0.447354 + }, + { + "acc": 0.63057241, + "epoch": 1.510908168442415, + "grad_norm": 5.71875, + "learning_rate": 1.5477977526170895e-06, + "loss": 1.63327465, + "memory(GiB)": 111.15, + "step": 59560, + "train_speed(iter/s)": 0.447374 + }, + { + "acc": 0.6490202, + "epoch": 1.51103500761035, + "grad_norm": 5.75, + "learning_rate": 1.5470392654627392e-06, + "loss": 1.65638237, + "memory(GiB)": 111.15, + "step": 59565, + "train_speed(iter/s)": 0.447393 + }, + { + "acc": 0.65727386, + "epoch": 1.5111618467782852, + "grad_norm": 5.125, + "learning_rate": 1.5462809301850212e-06, + "loss": 1.63469582, + "memory(GiB)": 111.15, + "step": 59570, + "train_speed(iter/s)": 0.447413 + }, + { + "acc": 0.6488399, + "epoch": 1.5112886859462202, + "grad_norm": 5.0, + "learning_rate": 1.5455227468172862e-06, + "loss": 1.64031868, + "memory(GiB)": 111.15, + "step": 59575, + "train_speed(iter/s)": 0.447433 + }, + { + "acc": 0.64597101, + "epoch": 1.5114155251141552, + "grad_norm": 6.125, + "learning_rate": 1.5447647153928842e-06, + "loss": 1.68319378, + "memory(GiB)": 111.15, + "step": 59580, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.64783216, + "epoch": 1.5115423642820902, + "grad_norm": 6.03125, + "learning_rate": 1.5440068359451548e-06, + "loss": 1.64066067, + "memory(GiB)": 111.15, + "step": 59585, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.66351919, + "epoch": 1.5116692034500254, + "grad_norm": 9.0, + "learning_rate": 1.5432491085074381e-06, + "loss": 1.63448334, + "memory(GiB)": 111.15, + "step": 59590, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.63733082, + "epoch": 1.5117960426179604, + "grad_norm": 4.84375, + "learning_rate": 1.5424915331130568e-06, + "loss": 1.67175331, + "memory(GiB)": 111.15, + "step": 59595, + "train_speed(iter/s)": 0.447512 + }, + { + "acc": 0.66003532, + "epoch": 1.5119228817858956, + "grad_norm": 6.53125, + "learning_rate": 1.5417341097953332e-06, + "loss": 1.55326042, + "memory(GiB)": 111.15, + "step": 59600, + "train_speed(iter/s)": 0.447532 + }, + { + "epoch": 1.5119228817858956, + "eval_acc": 0.643380038325973, + "eval_loss": 1.6075928211212158, + "eval_runtime": 112.3157, + "eval_samples_per_second": 56.715, + "eval_steps_per_second": 28.358, + "step": 59600 + }, + { + "acc": 0.66619644, + "epoch": 1.5120497209538306, + "grad_norm": 7.03125, + "learning_rate": 1.540976838587585e-06, + "loss": 1.61836014, + "memory(GiB)": 111.15, + "step": 59605, + "train_speed(iter/s)": 0.447159 + }, + { + "acc": 0.65142288, + "epoch": 1.5121765601217656, + "grad_norm": 5.0625, + "learning_rate": 1.5402197195231205e-06, + "loss": 1.60947266, + "memory(GiB)": 111.15, + "step": 59610, + "train_speed(iter/s)": 0.447179 + }, + { + "acc": 0.65635772, + "epoch": 1.5123033992897006, + "grad_norm": 5.09375, + "learning_rate": 1.5394627526352379e-06, + "loss": 1.58271618, + "memory(GiB)": 111.15, + "step": 59615, + "train_speed(iter/s)": 0.447199 + }, + { + "acc": 0.65267363, + "epoch": 1.5124302384576356, + "grad_norm": 5.625, + "learning_rate": 1.5387059379572322e-06, + "loss": 1.62455559, + "memory(GiB)": 111.15, + "step": 59620, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.67252941, + "epoch": 1.5125570776255708, + "grad_norm": 5.90625, + "learning_rate": 1.537949275522394e-06, + "loss": 1.57196264, + "memory(GiB)": 111.15, + "step": 59625, + "train_speed(iter/s)": 0.447238 + }, + { + "acc": 0.6365407, + "epoch": 1.512683916793506, + "grad_norm": 6.5, + "learning_rate": 1.5371927653640056e-06, + "loss": 1.67821159, + "memory(GiB)": 111.15, + "step": 59630, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.65380173, + "epoch": 1.512810755961441, + "grad_norm": 4.90625, + "learning_rate": 1.5364364075153366e-06, + "loss": 1.5871542, + "memory(GiB)": 111.15, + "step": 59635, + "train_speed(iter/s)": 0.447278 + }, + { + "acc": 0.65698128, + "epoch": 1.512937595129376, + "grad_norm": 5.28125, + "learning_rate": 1.5356802020096595e-06, + "loss": 1.58151703, + "memory(GiB)": 111.15, + "step": 59640, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.65728555, + "epoch": 1.513064434297311, + "grad_norm": 5.125, + "learning_rate": 1.5349241488802346e-06, + "loss": 1.59607382, + "memory(GiB)": 111.15, + "step": 59645, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.64914389, + "epoch": 1.513191273465246, + "grad_norm": 5.3125, + "learning_rate": 1.5341682481603155e-06, + "loss": 1.64512959, + "memory(GiB)": 111.15, + "step": 59650, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.65927014, + "epoch": 1.5133181126331812, + "grad_norm": 5.8125, + "learning_rate": 1.5334124998831512e-06, + "loss": 1.58811779, + "memory(GiB)": 111.15, + "step": 59655, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.62713332, + "epoch": 1.5134449518011162, + "grad_norm": 4.8125, + "learning_rate": 1.532656904081982e-06, + "loss": 1.67415237, + "memory(GiB)": 111.15, + "step": 59660, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.65575657, + "epoch": 1.5135717909690514, + "grad_norm": 5.25, + "learning_rate": 1.5319014607900428e-06, + "loss": 1.57562122, + "memory(GiB)": 111.15, + "step": 59665, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.65832605, + "epoch": 1.5136986301369864, + "grad_norm": 5.34375, + "learning_rate": 1.5311461700405617e-06, + "loss": 1.61154346, + "memory(GiB)": 111.15, + "step": 59670, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.64813786, + "epoch": 1.5138254693049213, + "grad_norm": 6.1875, + "learning_rate": 1.5303910318667586e-06, + "loss": 1.64115486, + "memory(GiB)": 111.15, + "step": 59675, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.6441844, + "epoch": 1.5139523084728563, + "grad_norm": 5.125, + "learning_rate": 1.529636046301849e-06, + "loss": 1.64704628, + "memory(GiB)": 111.15, + "step": 59680, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.63008699, + "epoch": 1.5140791476407913, + "grad_norm": 5.78125, + "learning_rate": 1.5288812133790405e-06, + "loss": 1.70643044, + "memory(GiB)": 111.15, + "step": 59685, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.64239058, + "epoch": 1.5142059868087265, + "grad_norm": 5.0625, + "learning_rate": 1.5281265331315332e-06, + "loss": 1.63183231, + "memory(GiB)": 111.15, + "step": 59690, + "train_speed(iter/s)": 0.447496 + }, + { + "acc": 0.65393629, + "epoch": 1.5143328259766617, + "grad_norm": 5.03125, + "learning_rate": 1.5273720055925217e-06, + "loss": 1.60926552, + "memory(GiB)": 111.15, + "step": 59695, + "train_speed(iter/s)": 0.447515 + }, + { + "acc": 0.65483088, + "epoch": 1.5144596651445967, + "grad_norm": 4.75, + "learning_rate": 1.5266176307951936e-06, + "loss": 1.6413559, + "memory(GiB)": 111.15, + "step": 59700, + "train_speed(iter/s)": 0.447535 + }, + { + "epoch": 1.5144596651445967, + "eval_acc": 0.6433904825087094, + "eval_loss": 1.607596755027771, + "eval_runtime": 114.9775, + "eval_samples_per_second": 55.402, + "eval_steps_per_second": 27.701, + "step": 59700 + }, + { + "acc": 0.62640305, + "epoch": 1.5145865043125317, + "grad_norm": 4.875, + "learning_rate": 1.5258634087727298e-06, + "loss": 1.61743603, + "memory(GiB)": 111.15, + "step": 59705, + "train_speed(iter/s)": 0.447154 + }, + { + "acc": 0.65359969, + "epoch": 1.5147133434804667, + "grad_norm": 5.28125, + "learning_rate": 1.5251093395583045e-06, + "loss": 1.64204884, + "memory(GiB)": 111.15, + "step": 59710, + "train_speed(iter/s)": 0.447174 + }, + { + "acc": 0.65808525, + "epoch": 1.5148401826484017, + "grad_norm": 5.09375, + "learning_rate": 1.5243554231850843e-06, + "loss": 1.60448513, + "memory(GiB)": 111.15, + "step": 59715, + "train_speed(iter/s)": 0.447193 + }, + { + "acc": 0.64710035, + "epoch": 1.514967021816337, + "grad_norm": 4.875, + "learning_rate": 1.5236016596862302e-06, + "loss": 1.59945507, + "memory(GiB)": 111.15, + "step": 59720, + "train_speed(iter/s)": 0.447213 + }, + { + "acc": 0.64972887, + "epoch": 1.515093860984272, + "grad_norm": 4.53125, + "learning_rate": 1.5228480490948943e-06, + "loss": 1.58401117, + "memory(GiB)": 111.15, + "step": 59725, + "train_speed(iter/s)": 0.447232 + }, + { + "acc": 0.65685673, + "epoch": 1.5152207001522071, + "grad_norm": 5.40625, + "learning_rate": 1.5220945914442292e-06, + "loss": 1.66196442, + "memory(GiB)": 111.15, + "step": 59730, + "train_speed(iter/s)": 0.447252 + }, + { + "acc": 0.64917827, + "epoch": 1.515347539320142, + "grad_norm": 5.0, + "learning_rate": 1.52134128676737e-06, + "loss": 1.57322769, + "memory(GiB)": 111.15, + "step": 59735, + "train_speed(iter/s)": 0.447272 + }, + { + "acc": 0.64983125, + "epoch": 1.515474378488077, + "grad_norm": 6.09375, + "learning_rate": 1.5205881350974504e-06, + "loss": 1.72152596, + "memory(GiB)": 111.15, + "step": 59740, + "train_speed(iter/s)": 0.447291 + }, + { + "acc": 0.63896265, + "epoch": 1.515601217656012, + "grad_norm": 6.65625, + "learning_rate": 1.5198351364676012e-06, + "loss": 1.60947647, + "memory(GiB)": 111.15, + "step": 59745, + "train_speed(iter/s)": 0.447311 + }, + { + "acc": 0.65435438, + "epoch": 1.5157280568239473, + "grad_norm": 5.375, + "learning_rate": 1.5190822909109415e-06, + "loss": 1.62179985, + "memory(GiB)": 111.15, + "step": 59750, + "train_speed(iter/s)": 0.447331 + }, + { + "acc": 0.65049467, + "epoch": 1.5158548959918823, + "grad_norm": 4.59375, + "learning_rate": 1.5183295984605824e-06, + "loss": 1.65684185, + "memory(GiB)": 111.15, + "step": 59755, + "train_speed(iter/s)": 0.44735 + }, + { + "acc": 0.65793595, + "epoch": 1.5159817351598175, + "grad_norm": 4.875, + "learning_rate": 1.5175770591496303e-06, + "loss": 1.61157627, + "memory(GiB)": 111.15, + "step": 59760, + "train_speed(iter/s)": 0.44737 + }, + { + "acc": 0.64831781, + "epoch": 1.5161085743277525, + "grad_norm": 4.65625, + "learning_rate": 1.5168246730111892e-06, + "loss": 1.62989426, + "memory(GiB)": 111.15, + "step": 59765, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.67648869, + "epoch": 1.5162354134956875, + "grad_norm": 5.15625, + "learning_rate": 1.5160724400783511e-06, + "loss": 1.59406624, + "memory(GiB)": 111.15, + "step": 59770, + "train_speed(iter/s)": 0.44741 + }, + { + "acc": 0.65006742, + "epoch": 1.5163622526636225, + "grad_norm": 4.84375, + "learning_rate": 1.5153203603841992e-06, + "loss": 1.61998291, + "memory(GiB)": 111.15, + "step": 59775, + "train_speed(iter/s)": 0.44743 + }, + { + "acc": 0.64968948, + "epoch": 1.5164890918315574, + "grad_norm": 5.21875, + "learning_rate": 1.5145684339618172e-06, + "loss": 1.69063301, + "memory(GiB)": 111.15, + "step": 59780, + "train_speed(iter/s)": 0.447449 + }, + { + "acc": 0.66691999, + "epoch": 1.5166159309994927, + "grad_norm": 5.5625, + "learning_rate": 1.5138166608442768e-06, + "loss": 1.53781958, + "memory(GiB)": 111.15, + "step": 59785, + "train_speed(iter/s)": 0.447469 + }, + { + "acc": 0.64984303, + "epoch": 1.5167427701674279, + "grad_norm": 5.96875, + "learning_rate": 1.5130650410646452e-06, + "loss": 1.58414965, + "memory(GiB)": 111.15, + "step": 59790, + "train_speed(iter/s)": 0.447489 + }, + { + "acc": 0.65290098, + "epoch": 1.5168696093353629, + "grad_norm": 4.34375, + "learning_rate": 1.5123135746559792e-06, + "loss": 1.61110249, + "memory(GiB)": 111.15, + "step": 59795, + "train_speed(iter/s)": 0.447509 + }, + { + "acc": 0.66783056, + "epoch": 1.5169964485032978, + "grad_norm": 6.15625, + "learning_rate": 1.5115622616513343e-06, + "loss": 1.56646309, + "memory(GiB)": 111.15, + "step": 59800, + "train_speed(iter/s)": 0.447529 + }, + { + "epoch": 1.5169964485032978, + "eval_acc": 0.643413877478039, + "eval_loss": 1.6075533628463745, + "eval_runtime": 114.1522, + "eval_samples_per_second": 55.803, + "eval_steps_per_second": 27.901, + "step": 59800 + }, + { + "acc": 0.64027271, + "epoch": 1.5171232876712328, + "grad_norm": 5.03125, + "learning_rate": 1.5108111020837564e-06, + "loss": 1.71249428, + "memory(GiB)": 111.15, + "step": 59805, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.64660177, + "epoch": 1.5172501268391678, + "grad_norm": 5.0, + "learning_rate": 1.5100600959862838e-06, + "loss": 1.59993439, + "memory(GiB)": 111.15, + "step": 59810, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.65249014, + "epoch": 1.517376966007103, + "grad_norm": 5.3125, + "learning_rate": 1.5093092433919497e-06, + "loss": 1.61603374, + "memory(GiB)": 111.15, + "step": 59815, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.64412746, + "epoch": 1.517503805175038, + "grad_norm": 5.34375, + "learning_rate": 1.5085585443337803e-06, + "loss": 1.59153976, + "memory(GiB)": 111.15, + "step": 59820, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.65250721, + "epoch": 1.5176306443429732, + "grad_norm": 5.9375, + "learning_rate": 1.507807998844794e-06, + "loss": 1.72072392, + "memory(GiB)": 111.15, + "step": 59825, + "train_speed(iter/s)": 0.44723 + }, + { + "acc": 0.64310398, + "epoch": 1.5177574835109082, + "grad_norm": 5.8125, + "learning_rate": 1.5070576069580039e-06, + "loss": 1.59251385, + "memory(GiB)": 111.15, + "step": 59830, + "train_speed(iter/s)": 0.447249 + }, + { + "acc": 0.65321069, + "epoch": 1.5178843226788432, + "grad_norm": 4.59375, + "learning_rate": 1.5063073687064144e-06, + "loss": 1.66049366, + "memory(GiB)": 111.15, + "step": 59835, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.66143446, + "epoch": 1.5180111618467782, + "grad_norm": 5.25, + "learning_rate": 1.5055572841230253e-06, + "loss": 1.6189167, + "memory(GiB)": 111.15, + "step": 59840, + "train_speed(iter/s)": 0.447289 + }, + { + "acc": 0.6575695, + "epoch": 1.5181380010147132, + "grad_norm": 4.9375, + "learning_rate": 1.5048073532408287e-06, + "loss": 1.62347946, + "memory(GiB)": 111.15, + "step": 59845, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.65974474, + "epoch": 1.5182648401826484, + "grad_norm": 4.53125, + "learning_rate": 1.5040575760928094e-06, + "loss": 1.58132677, + "memory(GiB)": 111.15, + "step": 59850, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.65526834, + "epoch": 1.5183916793505836, + "grad_norm": 5.125, + "learning_rate": 1.5033079527119466e-06, + "loss": 1.67370796, + "memory(GiB)": 111.15, + "step": 59855, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.66112509, + "epoch": 1.5185185185185186, + "grad_norm": 5.625, + "learning_rate": 1.5025584831312112e-06, + "loss": 1.61700745, + "memory(GiB)": 111.15, + "step": 59860, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.64592996, + "epoch": 1.5186453576864536, + "grad_norm": 6.34375, + "learning_rate": 1.5018091673835667e-06, + "loss": 1.63819504, + "memory(GiB)": 111.15, + "step": 59865, + "train_speed(iter/s)": 0.447388 + }, + { + "acc": 0.65680008, + "epoch": 1.5187721968543886, + "grad_norm": 5.28125, + "learning_rate": 1.501060005501977e-06, + "loss": 1.60709934, + "memory(GiB)": 111.15, + "step": 59870, + "train_speed(iter/s)": 0.447408 + }, + { + "acc": 0.6579277, + "epoch": 1.5188990360223236, + "grad_norm": 5.65625, + "learning_rate": 1.500310997519388e-06, + "loss": 1.64171982, + "memory(GiB)": 111.15, + "step": 59875, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.65919666, + "epoch": 1.5190258751902588, + "grad_norm": 5.03125, + "learning_rate": 1.4995621434687468e-06, + "loss": 1.59225483, + "memory(GiB)": 111.15, + "step": 59880, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.64306755, + "epoch": 1.5191527143581938, + "grad_norm": 5.5, + "learning_rate": 1.4988134433829892e-06, + "loss": 1.68221149, + "memory(GiB)": 111.15, + "step": 59885, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.64452968, + "epoch": 1.519279553526129, + "grad_norm": 5.28125, + "learning_rate": 1.4980648972950507e-06, + "loss": 1.61323433, + "memory(GiB)": 111.15, + "step": 59890, + "train_speed(iter/s)": 0.447487 + }, + { + "acc": 0.64893699, + "epoch": 1.519406392694064, + "grad_norm": 6.1875, + "learning_rate": 1.4973165052378518e-06, + "loss": 1.59929953, + "memory(GiB)": 111.15, + "step": 59895, + "train_speed(iter/s)": 0.447506 + }, + { + "acc": 0.64583421, + "epoch": 1.519533231861999, + "grad_norm": 5.5625, + "learning_rate": 1.49656826724431e-06, + "loss": 1.63631744, + "memory(GiB)": 111.15, + "step": 59900, + "train_speed(iter/s)": 0.447526 + }, + { + "epoch": 1.519533231861999, + "eval_acc": 0.6434744537379103, + "eval_loss": 1.607487678527832, + "eval_runtime": 113.4759, + "eval_samples_per_second": 56.135, + "eval_steps_per_second": 28.068, + "step": 59900 + }, + { + "acc": 0.65246658, + "epoch": 1.519660071029934, + "grad_norm": 5.1875, + "learning_rate": 1.4958201833473386e-06, + "loss": 1.60695076, + "memory(GiB)": 111.15, + "step": 59905, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.63448753, + "epoch": 1.5197869101978692, + "grad_norm": 5.125, + "learning_rate": 1.4950722535798423e-06, + "loss": 1.6476593, + "memory(GiB)": 111.15, + "step": 59910, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.65107889, + "epoch": 1.5199137493658041, + "grad_norm": 5.25, + "learning_rate": 1.4943244779747134e-06, + "loss": 1.63481407, + "memory(GiB)": 111.15, + "step": 59915, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.64174247, + "epoch": 1.5200405885337394, + "grad_norm": 4.6875, + "learning_rate": 1.4935768565648478e-06, + "loss": 1.60233002, + "memory(GiB)": 111.15, + "step": 59920, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.64108324, + "epoch": 1.5201674277016743, + "grad_norm": 7.40625, + "learning_rate": 1.4928293893831265e-06, + "loss": 1.69445343, + "memory(GiB)": 111.15, + "step": 59925, + "train_speed(iter/s)": 0.44723 + }, + { + "acc": 0.64460678, + "epoch": 1.5202942668696093, + "grad_norm": 6.125, + "learning_rate": 1.4920820764624288e-06, + "loss": 1.62856789, + "memory(GiB)": 111.15, + "step": 59930, + "train_speed(iter/s)": 0.447249 + }, + { + "acc": 0.64615221, + "epoch": 1.5204211060375443, + "grad_norm": 5.5625, + "learning_rate": 1.4913349178356202e-06, + "loss": 1.64554367, + "memory(GiB)": 111.15, + "step": 59935, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.64068298, + "epoch": 1.5205479452054793, + "grad_norm": 6.40625, + "learning_rate": 1.4905879135355684e-06, + "loss": 1.68960514, + "memory(GiB)": 111.15, + "step": 59940, + "train_speed(iter/s)": 0.447289 + }, + { + "acc": 0.64109359, + "epoch": 1.5206747843734145, + "grad_norm": 6.125, + "learning_rate": 1.4898410635951282e-06, + "loss": 1.60842915, + "memory(GiB)": 111.15, + "step": 59945, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.6584897, + "epoch": 1.5208016235413497, + "grad_norm": 5.40625, + "learning_rate": 1.4890943680471503e-06, + "loss": 1.61734123, + "memory(GiB)": 111.15, + "step": 59950, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.64938145, + "epoch": 1.5209284627092847, + "grad_norm": 5.15625, + "learning_rate": 1.4883478269244766e-06, + "loss": 1.61212883, + "memory(GiB)": 111.15, + "step": 59955, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.65076618, + "epoch": 1.5210553018772197, + "grad_norm": 4.75, + "learning_rate": 1.4876014402599443e-06, + "loss": 1.68238602, + "memory(GiB)": 111.15, + "step": 59960, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.63586426, + "epoch": 1.5211821410451547, + "grad_norm": 5.375, + "learning_rate": 1.4868552080863824e-06, + "loss": 1.67603455, + "memory(GiB)": 111.15, + "step": 59965, + "train_speed(iter/s)": 0.447387 + }, + { + "acc": 0.66906247, + "epoch": 1.5213089802130897, + "grad_norm": 6.71875, + "learning_rate": 1.4861091304366139e-06, + "loss": 1.5238534, + "memory(GiB)": 111.15, + "step": 59970, + "train_speed(iter/s)": 0.447407 + }, + { + "acc": 0.66376019, + "epoch": 1.521435819381025, + "grad_norm": 5.3125, + "learning_rate": 1.4853632073434533e-06, + "loss": 1.52662411, + "memory(GiB)": 111.15, + "step": 59975, + "train_speed(iter/s)": 0.447427 + }, + { + "acc": 0.64240322, + "epoch": 1.52156265854896, + "grad_norm": 7.8125, + "learning_rate": 1.484617438839711e-06, + "loss": 1.68637867, + "memory(GiB)": 111.15, + "step": 59980, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.6546061, + "epoch": 1.521689497716895, + "grad_norm": 6.53125, + "learning_rate": 1.483871824958189e-06, + "loss": 1.57913227, + "memory(GiB)": 111.15, + "step": 59985, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.64424324, + "epoch": 1.52181633688483, + "grad_norm": 5.4375, + "learning_rate": 1.483126365731682e-06, + "loss": 1.60780773, + "memory(GiB)": 111.15, + "step": 59990, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.65878992, + "epoch": 1.521943176052765, + "grad_norm": 4.9375, + "learning_rate": 1.4823810611929795e-06, + "loss": 1.52101269, + "memory(GiB)": 111.15, + "step": 59995, + "train_speed(iter/s)": 0.447505 + }, + { + "acc": 0.67096891, + "epoch": 1.5220700152207, + "grad_norm": 4.6875, + "learning_rate": 1.481635911374863e-06, + "loss": 1.51707954, + "memory(GiB)": 111.15, + "step": 60000, + "train_speed(iter/s)": 0.447525 + }, + { + "epoch": 1.5220700152207, + "eval_acc": 0.6434172196165147, + "eval_loss": 1.6075445413589478, + "eval_runtime": 114.6389, + "eval_samples_per_second": 55.566, + "eval_steps_per_second": 27.783, + "step": 60000 + }, + { + "acc": 0.65110869, + "epoch": 1.522196854388635, + "grad_norm": 5.375, + "learning_rate": 1.480890916310106e-06, + "loss": 1.63620872, + "memory(GiB)": 111.15, + "step": 60005, + "train_speed(iter/s)": 0.447147 + }, + { + "acc": 0.66446009, + "epoch": 1.5223236935565703, + "grad_norm": 5.4375, + "learning_rate": 1.4801460760314811e-06, + "loss": 1.55722103, + "memory(GiB)": 111.15, + "step": 60010, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.64216995, + "epoch": 1.5224505327245055, + "grad_norm": 6.5, + "learning_rate": 1.4794013905717453e-06, + "loss": 1.65406723, + "memory(GiB)": 111.15, + "step": 60015, + "train_speed(iter/s)": 0.447186 + }, + { + "acc": 0.64937954, + "epoch": 1.5225773718924405, + "grad_norm": 4.84375, + "learning_rate": 1.4786568599636548e-06, + "loss": 1.62575989, + "memory(GiB)": 111.15, + "step": 60020, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.64973106, + "epoch": 1.5227042110603755, + "grad_norm": 4.90625, + "learning_rate": 1.4779124842399556e-06, + "loss": 1.67492599, + "memory(GiB)": 111.15, + "step": 60025, + "train_speed(iter/s)": 0.447226 + }, + { + "acc": 0.6547081, + "epoch": 1.5228310502283104, + "grad_norm": 5.78125, + "learning_rate": 1.4771682634333933e-06, + "loss": 1.57540712, + "memory(GiB)": 111.15, + "step": 60030, + "train_speed(iter/s)": 0.447245 + }, + { + "acc": 0.6621995, + "epoch": 1.5229578893962454, + "grad_norm": 6.40625, + "learning_rate": 1.4764241975766975e-06, + "loss": 1.56417866, + "memory(GiB)": 111.15, + "step": 60035, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.65636997, + "epoch": 1.5230847285641806, + "grad_norm": 5.90625, + "learning_rate": 1.475680286702596e-06, + "loss": 1.60341358, + "memory(GiB)": 111.15, + "step": 60040, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.6577528, + "epoch": 1.5232115677321156, + "grad_norm": 4.875, + "learning_rate": 1.474936530843812e-06, + "loss": 1.56296558, + "memory(GiB)": 111.15, + "step": 60045, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.65523987, + "epoch": 1.5233384069000508, + "grad_norm": 5.65625, + "learning_rate": 1.4741929300330588e-06, + "loss": 1.61075649, + "memory(GiB)": 111.15, + "step": 60050, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.64703512, + "epoch": 1.5234652460679858, + "grad_norm": 5.34375, + "learning_rate": 1.4734494843030405e-06, + "loss": 1.62727375, + "memory(GiB)": 111.15, + "step": 60055, + "train_speed(iter/s)": 0.447344 + }, + { + "acc": 0.65093341, + "epoch": 1.5235920852359208, + "grad_norm": 5.34375, + "learning_rate": 1.4727061936864573e-06, + "loss": 1.61760902, + "memory(GiB)": 111.15, + "step": 60060, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.64155054, + "epoch": 1.5237189244038558, + "grad_norm": 5.34375, + "learning_rate": 1.4719630582160056e-06, + "loss": 1.69251347, + "memory(GiB)": 111.15, + "step": 60065, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.65225792, + "epoch": 1.523845763571791, + "grad_norm": 5.3125, + "learning_rate": 1.4712200779243718e-06, + "loss": 1.62834473, + "memory(GiB)": 111.15, + "step": 60070, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.65185499, + "epoch": 1.523972602739726, + "grad_norm": 5.5625, + "learning_rate": 1.4704772528442308e-06, + "loss": 1.63532677, + "memory(GiB)": 111.15, + "step": 60075, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.6614459, + "epoch": 1.5240994419076612, + "grad_norm": 4.9375, + "learning_rate": 1.46973458300826e-06, + "loss": 1.52279797, + "memory(GiB)": 111.15, + "step": 60080, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.64750071, + "epoch": 1.5242262810755962, + "grad_norm": 5.375, + "learning_rate": 1.4689920684491232e-06, + "loss": 1.62920609, + "memory(GiB)": 111.15, + "step": 60085, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.65803118, + "epoch": 1.5243531202435312, + "grad_norm": 5.5625, + "learning_rate": 1.4682497091994807e-06, + "loss": 1.61584015, + "memory(GiB)": 111.15, + "step": 60090, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.64887266, + "epoch": 1.5244799594114662, + "grad_norm": 5.1875, + "learning_rate": 1.467507505291984e-06, + "loss": 1.60319443, + "memory(GiB)": 111.15, + "step": 60095, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.63338642, + "epoch": 1.5246067985794012, + "grad_norm": 5.75, + "learning_rate": 1.4667654567592781e-06, + "loss": 1.70607681, + "memory(GiB)": 111.15, + "step": 60100, + "train_speed(iter/s)": 0.447521 + }, + { + "epoch": 1.5246067985794012, + "eval_acc": 0.6434447922589389, + "eval_loss": 1.6075071096420288, + "eval_runtime": 114.5428, + "eval_samples_per_second": 55.612, + "eval_steps_per_second": 27.806, + "step": 60100 + }, + { + "acc": 0.64505849, + "epoch": 1.5247336377473364, + "grad_norm": 5.25, + "learning_rate": 1.4660235636340025e-06, + "loss": 1.62753487, + "memory(GiB)": 111.15, + "step": 60105, + "train_speed(iter/s)": 0.447144 + }, + { + "acc": 0.64228182, + "epoch": 1.5248604769152716, + "grad_norm": 6.40625, + "learning_rate": 1.465281825948789e-06, + "loss": 1.64089813, + "memory(GiB)": 111.15, + "step": 60110, + "train_speed(iter/s)": 0.447163 + }, + { + "acc": 0.63351345, + "epoch": 1.5249873160832066, + "grad_norm": 5.25, + "learning_rate": 1.464540243736262e-06, + "loss": 1.63434219, + "memory(GiB)": 111.15, + "step": 60115, + "train_speed(iter/s)": 0.447183 + }, + { + "acc": 0.6527328, + "epoch": 1.5251141552511416, + "grad_norm": 5.46875, + "learning_rate": 1.4637988170290396e-06, + "loss": 1.57619629, + "memory(GiB)": 111.15, + "step": 60120, + "train_speed(iter/s)": 0.447203 + }, + { + "acc": 0.65855098, + "epoch": 1.5252409944190766, + "grad_norm": 4.96875, + "learning_rate": 1.4630575458597334e-06, + "loss": 1.57010021, + "memory(GiB)": 111.15, + "step": 60125, + "train_speed(iter/s)": 0.447223 + }, + { + "acc": 0.66835794, + "epoch": 1.5253678335870116, + "grad_norm": 6.28125, + "learning_rate": 1.4623164302609472e-06, + "loss": 1.54288425, + "memory(GiB)": 111.15, + "step": 60130, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.64482741, + "epoch": 1.5254946727549468, + "grad_norm": 5.65625, + "learning_rate": 1.4615754702652796e-06, + "loss": 1.70495834, + "memory(GiB)": 111.15, + "step": 60135, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.66458416, + "epoch": 1.5256215119228818, + "grad_norm": 5.15625, + "learning_rate": 1.4608346659053208e-06, + "loss": 1.54164658, + "memory(GiB)": 111.15, + "step": 60140, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.64952674, + "epoch": 1.525748351090817, + "grad_norm": 5.28125, + "learning_rate": 1.4600940172136541e-06, + "loss": 1.60344162, + "memory(GiB)": 111.15, + "step": 60145, + "train_speed(iter/s)": 0.447302 + }, + { + "acc": 0.65903339, + "epoch": 1.525875190258752, + "grad_norm": 6.03125, + "learning_rate": 1.4593535242228575e-06, + "loss": 1.57191954, + "memory(GiB)": 111.15, + "step": 60150, + "train_speed(iter/s)": 0.447322 + }, + { + "acc": 0.64487262, + "epoch": 1.526002029426687, + "grad_norm": 6.21875, + "learning_rate": 1.4586131869655001e-06, + "loss": 1.61865196, + "memory(GiB)": 111.15, + "step": 60155, + "train_speed(iter/s)": 0.447342 + }, + { + "acc": 0.63595924, + "epoch": 1.526128868594622, + "grad_norm": 5.40625, + "learning_rate": 1.4578730054741462e-06, + "loss": 1.67954826, + "memory(GiB)": 111.15, + "step": 60160, + "train_speed(iter/s)": 0.447361 + }, + { + "acc": 0.65163183, + "epoch": 1.526255707762557, + "grad_norm": 5.5625, + "learning_rate": 1.4571329797813511e-06, + "loss": 1.59331341, + "memory(GiB)": 111.15, + "step": 60165, + "train_speed(iter/s)": 0.447381 + }, + { + "acc": 0.65993242, + "epoch": 1.5263825469304921, + "grad_norm": 5.8125, + "learning_rate": 1.4563931099196678e-06, + "loss": 1.57999134, + "memory(GiB)": 111.15, + "step": 60170, + "train_speed(iter/s)": 0.447401 + }, + { + "acc": 0.66109571, + "epoch": 1.5265093860984273, + "grad_norm": 5.40625, + "learning_rate": 1.455653395921635e-06, + "loss": 1.56358833, + "memory(GiB)": 111.15, + "step": 60175, + "train_speed(iter/s)": 0.447421 + }, + { + "acc": 0.66217899, + "epoch": 1.5266362252663623, + "grad_norm": 6.3125, + "learning_rate": 1.4549138378197891e-06, + "loss": 1.62584648, + "memory(GiB)": 111.15, + "step": 60180, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.64837251, + "epoch": 1.5267630644342973, + "grad_norm": 5.46875, + "learning_rate": 1.4541744356466615e-06, + "loss": 1.64731483, + "memory(GiB)": 111.15, + "step": 60185, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.65330997, + "epoch": 1.5268899036022323, + "grad_norm": 4.6875, + "learning_rate": 1.4534351894347748e-06, + "loss": 1.55314684, + "memory(GiB)": 111.15, + "step": 60190, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.6551218, + "epoch": 1.5270167427701673, + "grad_norm": 5.3125, + "learning_rate": 1.4526960992166412e-06, + "loss": 1.62011528, + "memory(GiB)": 111.15, + "step": 60195, + "train_speed(iter/s)": 0.4475 + }, + { + "acc": 0.65360889, + "epoch": 1.5271435819381025, + "grad_norm": 4.875, + "learning_rate": 1.4519571650247687e-06, + "loss": 1.59396973, + "memory(GiB)": 111.15, + "step": 60200, + "train_speed(iter/s)": 0.44752 + }, + { + "epoch": 1.5271435819381025, + "eval_acc": 0.6433917358106378, + "eval_loss": 1.6075494289398193, + "eval_runtime": 113.4476, + "eval_samples_per_second": 56.149, + "eval_steps_per_second": 28.075, + "step": 60200 + }, + { + "acc": 0.657621, + "epoch": 1.5272704211060375, + "grad_norm": 4.40625, + "learning_rate": 1.4512183868916629e-06, + "loss": 1.61167946, + "memory(GiB)": 111.15, + "step": 60205, + "train_speed(iter/s)": 0.447147 + }, + { + "acc": 0.65549994, + "epoch": 1.5273972602739727, + "grad_norm": 5.59375, + "learning_rate": 1.4504797648498186e-06, + "loss": 1.6262022, + "memory(GiB)": 111.15, + "step": 60210, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.65158944, + "epoch": 1.5275240994419077, + "grad_norm": 5.09375, + "learning_rate": 1.4497412989317184e-06, + "loss": 1.58597889, + "memory(GiB)": 111.15, + "step": 60215, + "train_speed(iter/s)": 0.447186 + }, + { + "acc": 0.65262198, + "epoch": 1.5276509386098427, + "grad_norm": 5.40625, + "learning_rate": 1.4490029891698476e-06, + "loss": 1.69272404, + "memory(GiB)": 111.15, + "step": 60220, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.64808321, + "epoch": 1.5277777777777777, + "grad_norm": 5.03125, + "learning_rate": 1.44826483559668e-06, + "loss": 1.54462891, + "memory(GiB)": 111.15, + "step": 60225, + "train_speed(iter/s)": 0.447225 + }, + { + "acc": 0.65730619, + "epoch": 1.527904616945713, + "grad_norm": 6.21875, + "learning_rate": 1.4475268382446833e-06, + "loss": 1.54425621, + "memory(GiB)": 111.15, + "step": 60230, + "train_speed(iter/s)": 0.447245 + }, + { + "acc": 0.65481677, + "epoch": 1.5280314561136479, + "grad_norm": 5.5625, + "learning_rate": 1.4467889971463144e-06, + "loss": 1.61895733, + "memory(GiB)": 111.15, + "step": 60235, + "train_speed(iter/s)": 0.447264 + }, + { + "acc": 0.65245991, + "epoch": 1.528158295281583, + "grad_norm": 4.53125, + "learning_rate": 1.4460513123340308e-06, + "loss": 1.6269455, + "memory(GiB)": 111.15, + "step": 60240, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.654842, + "epoch": 1.528285134449518, + "grad_norm": 6.4375, + "learning_rate": 1.4453137838402775e-06, + "loss": 1.62777328, + "memory(GiB)": 111.15, + "step": 60245, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.65356379, + "epoch": 1.528411973617453, + "grad_norm": 5.34375, + "learning_rate": 1.4445764116974948e-06, + "loss": 1.59951363, + "memory(GiB)": 111.15, + "step": 60250, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.65906739, + "epoch": 1.528538812785388, + "grad_norm": 5.34375, + "learning_rate": 1.4438391959381149e-06, + "loss": 1.5980217, + "memory(GiB)": 111.15, + "step": 60255, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.65023389, + "epoch": 1.528665651953323, + "grad_norm": 5.6875, + "learning_rate": 1.4431021365945647e-06, + "loss": 1.62337761, + "memory(GiB)": 111.15, + "step": 60260, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.64621735, + "epoch": 1.5287924911212583, + "grad_norm": 5.375, + "learning_rate": 1.4423652336992627e-06, + "loss": 1.64377537, + "memory(GiB)": 111.15, + "step": 60265, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.65309391, + "epoch": 1.5289193302891935, + "grad_norm": 6.21875, + "learning_rate": 1.4416284872846215e-06, + "loss": 1.64602566, + "memory(GiB)": 111.15, + "step": 60270, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.66436605, + "epoch": 1.5290461694571285, + "grad_norm": 5.625, + "learning_rate": 1.440891897383046e-06, + "loss": 1.48402786, + "memory(GiB)": 111.15, + "step": 60275, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.67041979, + "epoch": 1.5291730086250634, + "grad_norm": 6.25, + "learning_rate": 1.4401554640269354e-06, + "loss": 1.47563505, + "memory(GiB)": 111.15, + "step": 60280, + "train_speed(iter/s)": 0.447442 + }, + { + "acc": 0.64079428, + "epoch": 1.5292998477929984, + "grad_norm": 5.75, + "learning_rate": 1.4394191872486812e-06, + "loss": 1.66140499, + "memory(GiB)": 111.15, + "step": 60285, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.64304314, + "epoch": 1.5294266869609334, + "grad_norm": 5.9375, + "learning_rate": 1.4386830670806684e-06, + "loss": 1.65826836, + "memory(GiB)": 111.15, + "step": 60290, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.65595779, + "epoch": 1.5295535261288686, + "grad_norm": 5.40625, + "learning_rate": 1.4379471035552738e-06, + "loss": 1.60607758, + "memory(GiB)": 111.15, + "step": 60295, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.64735532, + "epoch": 1.5296803652968036, + "grad_norm": 5.5625, + "learning_rate": 1.437211296704869e-06, + "loss": 1.66688595, + "memory(GiB)": 111.15, + "step": 60300, + "train_speed(iter/s)": 0.447521 + }, + { + "epoch": 1.5296803652968036, + "eval_acc": 0.6433980023202797, + "eval_loss": 1.6075321435928345, + "eval_runtime": 113.5207, + "eval_samples_per_second": 56.113, + "eval_steps_per_second": 28.057, + "step": 60300 + }, + { + "acc": 0.65479336, + "epoch": 1.5298072044647388, + "grad_norm": 4.65625, + "learning_rate": 1.4364756465618167e-06, + "loss": 1.63274002, + "memory(GiB)": 111.15, + "step": 60305, + "train_speed(iter/s)": 0.447148 + }, + { + "acc": 0.67495222, + "epoch": 1.5299340436326738, + "grad_norm": 5.21875, + "learning_rate": 1.4357401531584792e-06, + "loss": 1.55048943, + "memory(GiB)": 111.15, + "step": 60310, + "train_speed(iter/s)": 0.447168 + }, + { + "acc": 0.6499495, + "epoch": 1.5300608828006088, + "grad_norm": 5.90625, + "learning_rate": 1.4350048165272006e-06, + "loss": 1.59010515, + "memory(GiB)": 111.15, + "step": 60315, + "train_speed(iter/s)": 0.447187 + }, + { + "acc": 0.64584255, + "epoch": 1.5301877219685438, + "grad_norm": 6.0625, + "learning_rate": 1.4342696367003272e-06, + "loss": 1.66916962, + "memory(GiB)": 111.15, + "step": 60320, + "train_speed(iter/s)": 0.447207 + }, + { + "acc": 0.64300461, + "epoch": 1.5303145611364788, + "grad_norm": 6.21875, + "learning_rate": 1.433534613710193e-06, + "loss": 1.57774324, + "memory(GiB)": 111.15, + "step": 60325, + "train_speed(iter/s)": 0.447226 + }, + { + "acc": 0.65697756, + "epoch": 1.530441400304414, + "grad_norm": 5.28125, + "learning_rate": 1.4327997475891331e-06, + "loss": 1.61551876, + "memory(GiB)": 111.15, + "step": 60330, + "train_speed(iter/s)": 0.447246 + }, + { + "acc": 0.65353632, + "epoch": 1.5305682394723492, + "grad_norm": 5.34375, + "learning_rate": 1.432065038369465e-06, + "loss": 1.60021133, + "memory(GiB)": 111.15, + "step": 60335, + "train_speed(iter/s)": 0.447266 + }, + { + "acc": 0.65272646, + "epoch": 1.5306950786402842, + "grad_norm": 5.1875, + "learning_rate": 1.4313304860835048e-06, + "loss": 1.5747921, + "memory(GiB)": 111.15, + "step": 60340, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.64689336, + "epoch": 1.5308219178082192, + "grad_norm": 4.9375, + "learning_rate": 1.4305960907635641e-06, + "loss": 1.65024338, + "memory(GiB)": 111.15, + "step": 60345, + "train_speed(iter/s)": 0.447305 + }, + { + "acc": 0.66598849, + "epoch": 1.5309487569761542, + "grad_norm": 5.03125, + "learning_rate": 1.4298618524419455e-06, + "loss": 1.60760021, + "memory(GiB)": 111.15, + "step": 60350, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.64598293, + "epoch": 1.5310755961440892, + "grad_norm": 4.96875, + "learning_rate": 1.4291277711509388e-06, + "loss": 1.65229912, + "memory(GiB)": 111.15, + "step": 60355, + "train_speed(iter/s)": 0.447344 + }, + { + "acc": 0.66168156, + "epoch": 1.5312024353120244, + "grad_norm": 5.125, + "learning_rate": 1.428393846922837e-06, + "loss": 1.50968323, + "memory(GiB)": 111.15, + "step": 60360, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.6686934, + "epoch": 1.5313292744799594, + "grad_norm": 4.65625, + "learning_rate": 1.4276600797899199e-06, + "loss": 1.5272604, + "memory(GiB)": 111.15, + "step": 60365, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.65748577, + "epoch": 1.5314561136478946, + "grad_norm": 5.28125, + "learning_rate": 1.426926469784463e-06, + "loss": 1.56713152, + "memory(GiB)": 111.15, + "step": 60370, + "train_speed(iter/s)": 0.447404 + }, + { + "acc": 0.65747113, + "epoch": 1.5315829528158296, + "grad_norm": 5.9375, + "learning_rate": 1.42619301693873e-06, + "loss": 1.68293991, + "memory(GiB)": 111.15, + "step": 60375, + "train_speed(iter/s)": 0.447423 + }, + { + "acc": 0.65968018, + "epoch": 1.5317097919837646, + "grad_norm": 5.21875, + "learning_rate": 1.4254597212849858e-06, + "loss": 1.59811478, + "memory(GiB)": 111.15, + "step": 60380, + "train_speed(iter/s)": 0.447443 + }, + { + "acc": 0.64730406, + "epoch": 1.5318366311516995, + "grad_norm": 5.875, + "learning_rate": 1.4247265828554819e-06, + "loss": 1.61895466, + "memory(GiB)": 111.15, + "step": 60385, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.64390965, + "epoch": 1.5319634703196348, + "grad_norm": 6.1875, + "learning_rate": 1.423993601682465e-06, + "loss": 1.67728577, + "memory(GiB)": 111.15, + "step": 60390, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.64851685, + "epoch": 1.5320903094875697, + "grad_norm": 4.65625, + "learning_rate": 1.423260777798176e-06, + "loss": 1.60320778, + "memory(GiB)": 111.15, + "step": 60395, + "train_speed(iter/s)": 0.447502 + }, + { + "acc": 0.64834175, + "epoch": 1.532217148655505, + "grad_norm": 6.3125, + "learning_rate": 1.4225281112348466e-06, + "loss": 1.63140259, + "memory(GiB)": 111.15, + "step": 60400, + "train_speed(iter/s)": 0.447522 + }, + { + "epoch": 1.532217148655505, + "eval_acc": 0.6434736182032914, + "eval_loss": 1.6076481342315674, + "eval_runtime": 114.0919, + "eval_samples_per_second": 55.832, + "eval_steps_per_second": 27.916, + "step": 60400 + }, + { + "acc": 0.62324972, + "epoch": 1.53234398782344, + "grad_norm": 6.03125, + "learning_rate": 1.421795602024703e-06, + "loss": 1.72411613, + "memory(GiB)": 111.15, + "step": 60405, + "train_speed(iter/s)": 0.447148 + }, + { + "acc": 0.6538218, + "epoch": 1.532470826991375, + "grad_norm": 5.4375, + "learning_rate": 1.4210632501999643e-06, + "loss": 1.63780956, + "memory(GiB)": 111.15, + "step": 60410, + "train_speed(iter/s)": 0.447167 + }, + { + "acc": 0.65091381, + "epoch": 1.53259766615931, + "grad_norm": 5.1875, + "learning_rate": 1.4203310557928428e-06, + "loss": 1.60922546, + "memory(GiB)": 111.15, + "step": 60415, + "train_speed(iter/s)": 0.447187 + }, + { + "acc": 0.63339977, + "epoch": 1.532724505327245, + "grad_norm": 5.46875, + "learning_rate": 1.4195990188355435e-06, + "loss": 1.72450371, + "memory(GiB)": 111.15, + "step": 60420, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.64613514, + "epoch": 1.5328513444951801, + "grad_norm": 6.625, + "learning_rate": 1.418867139360265e-06, + "loss": 1.62459164, + "memory(GiB)": 111.15, + "step": 60425, + "train_speed(iter/s)": 0.447226 + }, + { + "acc": 0.65572944, + "epoch": 1.5329781836631153, + "grad_norm": 6.25, + "learning_rate": 1.418135417399198e-06, + "loss": 1.62145004, + "memory(GiB)": 111.15, + "step": 60430, + "train_speed(iter/s)": 0.447245 + }, + { + "acc": 0.64348392, + "epoch": 1.5331050228310503, + "grad_norm": 6.5625, + "learning_rate": 1.4174038529845273e-06, + "loss": 1.70151253, + "memory(GiB)": 111.15, + "step": 60435, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.64090037, + "epoch": 1.5332318619989853, + "grad_norm": 5.84375, + "learning_rate": 1.4166724461484304e-06, + "loss": 1.6722662, + "memory(GiB)": 111.15, + "step": 60440, + "train_speed(iter/s)": 0.447285 + }, + { + "acc": 0.63684158, + "epoch": 1.5333587011669203, + "grad_norm": 4.4375, + "learning_rate": 1.4159411969230758e-06, + "loss": 1.67982483, + "memory(GiB)": 111.15, + "step": 60445, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.65052872, + "epoch": 1.5334855403348553, + "grad_norm": 4.9375, + "learning_rate": 1.4152101053406325e-06, + "loss": 1.57772598, + "memory(GiB)": 111.15, + "step": 60450, + "train_speed(iter/s)": 0.447324 + }, + { + "acc": 0.6526248, + "epoch": 1.5336123795027905, + "grad_norm": 5.3125, + "learning_rate": 1.4144791714332517e-06, + "loss": 1.5826829, + "memory(GiB)": 111.15, + "step": 60455, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.6502265, + "epoch": 1.5337392186707255, + "grad_norm": 5.0, + "learning_rate": 1.4137483952330855e-06, + "loss": 1.65462093, + "memory(GiB)": 111.15, + "step": 60460, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.65932989, + "epoch": 1.5338660578386607, + "grad_norm": 6.4375, + "learning_rate": 1.4130177767722753e-06, + "loss": 1.61532593, + "memory(GiB)": 111.15, + "step": 60465, + "train_speed(iter/s)": 0.447383 + }, + { + "acc": 0.65794768, + "epoch": 1.5339928970065957, + "grad_norm": 5.15625, + "learning_rate": 1.4122873160829603e-06, + "loss": 1.59271402, + "memory(GiB)": 111.15, + "step": 60470, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.63623314, + "epoch": 1.5341197361745307, + "grad_norm": 4.3125, + "learning_rate": 1.4115570131972655e-06, + "loss": 1.65817299, + "memory(GiB)": 111.15, + "step": 60475, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.65987453, + "epoch": 1.5342465753424657, + "grad_norm": 5.03125, + "learning_rate": 1.4108268681473136e-06, + "loss": 1.61983643, + "memory(GiB)": 111.15, + "step": 60480, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.66050558, + "epoch": 1.5343734145104007, + "grad_norm": 8.25, + "learning_rate": 1.4100968809652215e-06, + "loss": 1.63374863, + "memory(GiB)": 111.15, + "step": 60485, + "train_speed(iter/s)": 0.44746 + }, + { + "acc": 0.64792719, + "epoch": 1.5345002536783359, + "grad_norm": 6.5, + "learning_rate": 1.4093670516830982e-06, + "loss": 1.63369961, + "memory(GiB)": 111.15, + "step": 60490, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.6424355, + "epoch": 1.534627092846271, + "grad_norm": 5.65625, + "learning_rate": 1.4086373803330417e-06, + "loss": 1.64856873, + "memory(GiB)": 111.15, + "step": 60495, + "train_speed(iter/s)": 0.4475 + }, + { + "acc": 0.64706125, + "epoch": 1.534753932014206, + "grad_norm": 5.84375, + "learning_rate": 1.4079078669471457e-06, + "loss": 1.6626915, + "memory(GiB)": 111.15, + "step": 60500, + "train_speed(iter/s)": 0.447519 + }, + { + "epoch": 1.534753932014206, + "eval_acc": 0.6433766961874973, + "eval_loss": 1.6075080633163452, + "eval_runtime": 113.5733, + "eval_samples_per_second": 56.087, + "eval_steps_per_second": 28.044, + "step": 60500 + }, + { + "acc": 0.64826808, + "epoch": 1.534880771182141, + "grad_norm": 5.0625, + "learning_rate": 1.4071785115575005e-06, + "loss": 1.57810593, + "memory(GiB)": 111.15, + "step": 60505, + "train_speed(iter/s)": 0.447148 + }, + { + "acc": 0.67026925, + "epoch": 1.535007610350076, + "grad_norm": 5.5, + "learning_rate": 1.4064493141961872e-06, + "loss": 1.55655556, + "memory(GiB)": 111.15, + "step": 60510, + "train_speed(iter/s)": 0.447167 + }, + { + "acc": 0.66556005, + "epoch": 1.535134449518011, + "grad_norm": 6.59375, + "learning_rate": 1.4057202748952736e-06, + "loss": 1.60155907, + "memory(GiB)": 111.15, + "step": 60515, + "train_speed(iter/s)": 0.447187 + }, + { + "acc": 0.65573401, + "epoch": 1.5352612886859462, + "grad_norm": 6.5, + "learning_rate": 1.4049913936868314e-06, + "loss": 1.58641186, + "memory(GiB)": 111.15, + "step": 60520, + "train_speed(iter/s)": 0.447206 + }, + { + "acc": 0.63558912, + "epoch": 1.5353881278538812, + "grad_norm": 5.75, + "learning_rate": 1.4042626706029184e-06, + "loss": 1.66035233, + "memory(GiB)": 111.15, + "step": 60525, + "train_speed(iter/s)": 0.447226 + }, + { + "acc": 0.63600302, + "epoch": 1.5355149670218164, + "grad_norm": 5.5625, + "learning_rate": 1.4035341056755864e-06, + "loss": 1.65681152, + "memory(GiB)": 111.15, + "step": 60530, + "train_speed(iter/s)": 0.447246 + }, + { + "acc": 0.66763563, + "epoch": 1.5356418061897514, + "grad_norm": 6.375, + "learning_rate": 1.402805698936882e-06, + "loss": 1.57165413, + "memory(GiB)": 111.15, + "step": 60535, + "train_speed(iter/s)": 0.447265 + }, + { + "acc": 0.65661802, + "epoch": 1.5357686453576864, + "grad_norm": 5.0, + "learning_rate": 1.4020774504188428e-06, + "loss": 1.64245491, + "memory(GiB)": 111.15, + "step": 60540, + "train_speed(iter/s)": 0.447284 + }, + { + "acc": 0.65500813, + "epoch": 1.5358954845256214, + "grad_norm": 5.40625, + "learning_rate": 1.4013493601535016e-06, + "loss": 1.63940773, + "memory(GiB)": 111.15, + "step": 60545, + "train_speed(iter/s)": 0.447304 + }, + { + "acc": 0.63921537, + "epoch": 1.5360223236935566, + "grad_norm": 7.0625, + "learning_rate": 1.400621428172882e-06, + "loss": 1.70576019, + "memory(GiB)": 111.15, + "step": 60550, + "train_speed(iter/s)": 0.447323 + }, + { + "acc": 0.64921303, + "epoch": 1.5361491628614916, + "grad_norm": 6.46875, + "learning_rate": 1.399893654509002e-06, + "loss": 1.66003113, + "memory(GiB)": 111.15, + "step": 60555, + "train_speed(iter/s)": 0.447343 + }, + { + "acc": 0.64164772, + "epoch": 1.5362760020294268, + "grad_norm": 5.03125, + "learning_rate": 1.3991660391938721e-06, + "loss": 1.63210545, + "memory(GiB)": 111.15, + "step": 60560, + "train_speed(iter/s)": 0.447363 + }, + { + "acc": 0.65728035, + "epoch": 1.5364028411973618, + "grad_norm": 6.03125, + "learning_rate": 1.398438582259497e-06, + "loss": 1.64674511, + "memory(GiB)": 111.15, + "step": 60565, + "train_speed(iter/s)": 0.447382 + }, + { + "acc": 0.64097548, + "epoch": 1.5365296803652968, + "grad_norm": 5.59375, + "learning_rate": 1.3977112837378726e-06, + "loss": 1.62449322, + "memory(GiB)": 111.15, + "step": 60570, + "train_speed(iter/s)": 0.447402 + }, + { + "acc": 0.64287138, + "epoch": 1.5366565195332318, + "grad_norm": 5.4375, + "learning_rate": 1.3969841436609888e-06, + "loss": 1.68213959, + "memory(GiB)": 111.15, + "step": 60575, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.65072603, + "epoch": 1.5367833587011668, + "grad_norm": 4.71875, + "learning_rate": 1.396257162060829e-06, + "loss": 1.57509623, + "memory(GiB)": 111.15, + "step": 60580, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.63273506, + "epoch": 1.536910197869102, + "grad_norm": 4.40625, + "learning_rate": 1.395530338969367e-06, + "loss": 1.68754463, + "memory(GiB)": 111.15, + "step": 60585, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.64341364, + "epoch": 1.5370370370370372, + "grad_norm": 5.28125, + "learning_rate": 1.3948036744185767e-06, + "loss": 1.60501328, + "memory(GiB)": 111.15, + "step": 60590, + "train_speed(iter/s)": 0.447481 + }, + { + "acc": 0.64636669, + "epoch": 1.5371638762049722, + "grad_norm": 6.125, + "learning_rate": 1.3940771684404153e-06, + "loss": 1.57733593, + "memory(GiB)": 111.15, + "step": 60595, + "train_speed(iter/s)": 0.4475 + }, + { + "acc": 0.66116476, + "epoch": 1.5372907153729072, + "grad_norm": 4.6875, + "learning_rate": 1.393350821066839e-06, + "loss": 1.53389454, + "memory(GiB)": 111.15, + "step": 60600, + "train_speed(iter/s)": 0.44752 + }, + { + "epoch": 1.5372907153729072, + "eval_acc": 0.6433666697720704, + "eval_loss": 1.6075342893600464, + "eval_runtime": 112.7921, + "eval_samples_per_second": 56.476, + "eval_steps_per_second": 28.238, + "step": 60600 + }, + { + "acc": 0.67261925, + "epoch": 1.5374175545408422, + "grad_norm": 5.1875, + "learning_rate": 1.3926246323297948e-06, + "loss": 1.58707895, + "memory(GiB)": 111.15, + "step": 60605, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.6555295, + "epoch": 1.5375443937087772, + "grad_norm": 5.40625, + "learning_rate": 1.3918986022612285e-06, + "loss": 1.57955837, + "memory(GiB)": 111.15, + "step": 60610, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.65749207, + "epoch": 1.5376712328767124, + "grad_norm": 5.46875, + "learning_rate": 1.3911727308930684e-06, + "loss": 1.61847801, + "memory(GiB)": 111.15, + "step": 60615, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.64135647, + "epoch": 1.5377980720446474, + "grad_norm": 5.0625, + "learning_rate": 1.3904470182572428e-06, + "loss": 1.61763554, + "memory(GiB)": 111.15, + "step": 60620, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.64865985, + "epoch": 1.5379249112125826, + "grad_norm": 5.6875, + "learning_rate": 1.3897214643856744e-06, + "loss": 1.68862057, + "memory(GiB)": 111.15, + "step": 60625, + "train_speed(iter/s)": 0.447229 + }, + { + "acc": 0.66992974, + "epoch": 1.5380517503805176, + "grad_norm": 5.625, + "learning_rate": 1.388996069310276e-06, + "loss": 1.56425686, + "memory(GiB)": 111.15, + "step": 60630, + "train_speed(iter/s)": 0.447248 + }, + { + "acc": 0.65079689, + "epoch": 1.5381785895484525, + "grad_norm": 5.1875, + "learning_rate": 1.3882708330629514e-06, + "loss": 1.66791401, + "memory(GiB)": 111.15, + "step": 60635, + "train_speed(iter/s)": 0.447268 + }, + { + "acc": 0.65144224, + "epoch": 1.5383054287163875, + "grad_norm": 4.9375, + "learning_rate": 1.3875457556755989e-06, + "loss": 1.64638042, + "memory(GiB)": 111.15, + "step": 60640, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.65259495, + "epoch": 1.5384322678843225, + "grad_norm": 5.75, + "learning_rate": 1.386820837180114e-06, + "loss": 1.58751564, + "memory(GiB)": 111.15, + "step": 60645, + "train_speed(iter/s)": 0.447307 + }, + { + "acc": 0.66133375, + "epoch": 1.5385591070522577, + "grad_norm": 5.53125, + "learning_rate": 1.386096077608382e-06, + "loss": 1.5624774, + "memory(GiB)": 111.15, + "step": 60650, + "train_speed(iter/s)": 0.447326 + }, + { + "acc": 0.6543314, + "epoch": 1.538685946220193, + "grad_norm": 5.09375, + "learning_rate": 1.385371476992276e-06, + "loss": 1.5872963, + "memory(GiB)": 111.15, + "step": 60655, + "train_speed(iter/s)": 0.447346 + }, + { + "acc": 0.64273057, + "epoch": 1.538812785388128, + "grad_norm": 5.6875, + "learning_rate": 1.3846470353636726e-06, + "loss": 1.65926266, + "memory(GiB)": 111.15, + "step": 60660, + "train_speed(iter/s)": 0.447365 + }, + { + "acc": 0.64185829, + "epoch": 1.538939624556063, + "grad_norm": 4.625, + "learning_rate": 1.3839227527544336e-06, + "loss": 1.62867298, + "memory(GiB)": 111.15, + "step": 60665, + "train_speed(iter/s)": 0.447385 + }, + { + "acc": 0.65462084, + "epoch": 1.539066463723998, + "grad_norm": 4.78125, + "learning_rate": 1.3831986291964184e-06, + "loss": 1.60534401, + "memory(GiB)": 111.15, + "step": 60670, + "train_speed(iter/s)": 0.447404 + }, + { + "acc": 0.65237169, + "epoch": 1.539193302891933, + "grad_norm": 5.71875, + "learning_rate": 1.382474664721472e-06, + "loss": 1.64823837, + "memory(GiB)": 111.15, + "step": 60675, + "train_speed(iter/s)": 0.447424 + }, + { + "acc": 0.65648985, + "epoch": 1.5393201420598681, + "grad_norm": 5.59375, + "learning_rate": 1.3817508593614425e-06, + "loss": 1.5509881, + "memory(GiB)": 111.15, + "step": 60680, + "train_speed(iter/s)": 0.447443 + }, + { + "acc": 0.65039139, + "epoch": 1.539446981227803, + "grad_norm": 5.15625, + "learning_rate": 1.381027213148165e-06, + "loss": 1.66168499, + "memory(GiB)": 111.15, + "step": 60685, + "train_speed(iter/s)": 0.447462 + }, + { + "acc": 0.64376497, + "epoch": 1.5395738203957383, + "grad_norm": 5.9375, + "learning_rate": 1.3803037261134678e-06, + "loss": 1.6164196, + "memory(GiB)": 111.15, + "step": 60690, + "train_speed(iter/s)": 0.447482 + }, + { + "acc": 0.65428486, + "epoch": 1.5397006595636733, + "grad_norm": 5.6875, + "learning_rate": 1.3795803982891736e-06, + "loss": 1.62890549, + "memory(GiB)": 111.15, + "step": 60695, + "train_speed(iter/s)": 0.447501 + }, + { + "acc": 0.65184288, + "epoch": 1.5398274987316083, + "grad_norm": 5.25, + "learning_rate": 1.3788572297070974e-06, + "loss": 1.60314903, + "memory(GiB)": 111.15, + "step": 60700, + "train_speed(iter/s)": 0.447521 + }, + { + "epoch": 1.5398274987316083, + "eval_acc": 0.6433787850240446, + "eval_loss": 1.6075439453125, + "eval_runtime": 113.3662, + "eval_samples_per_second": 56.19, + "eval_steps_per_second": 28.095, + "step": 60700 + }, + { + "acc": 0.65966654, + "epoch": 1.5399543378995433, + "grad_norm": 4.5625, + "learning_rate": 1.3781342203990478e-06, + "loss": 1.58435192, + "memory(GiB)": 111.15, + "step": 60705, + "train_speed(iter/s)": 0.447151 + }, + { + "acc": 0.65473027, + "epoch": 1.5400811770674785, + "grad_norm": 5.5625, + "learning_rate": 1.3774113703968255e-06, + "loss": 1.65098686, + "memory(GiB)": 111.15, + "step": 60710, + "train_speed(iter/s)": 0.447171 + }, + { + "acc": 0.64713755, + "epoch": 1.5402080162354135, + "grad_norm": 5.34375, + "learning_rate": 1.3766886797322248e-06, + "loss": 1.58670111, + "memory(GiB)": 111.15, + "step": 60715, + "train_speed(iter/s)": 0.44719 + }, + { + "acc": 0.67034826, + "epoch": 1.5403348554033487, + "grad_norm": 4.125, + "learning_rate": 1.3759661484370324e-06, + "loss": 1.50350189, + "memory(GiB)": 111.15, + "step": 60720, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.66067238, + "epoch": 1.5404616945712837, + "grad_norm": 6.25, + "learning_rate": 1.3752437765430294e-06, + "loss": 1.58406754, + "memory(GiB)": 111.15, + "step": 60725, + "train_speed(iter/s)": 0.44723 + }, + { + "acc": 0.65374212, + "epoch": 1.5405885337392187, + "grad_norm": 5.3125, + "learning_rate": 1.3745215640819886e-06, + "loss": 1.62547417, + "memory(GiB)": 111.15, + "step": 60730, + "train_speed(iter/s)": 0.447249 + }, + { + "acc": 0.6654377, + "epoch": 1.5407153729071537, + "grad_norm": 5.65625, + "learning_rate": 1.373799511085676e-06, + "loss": 1.55596266, + "memory(GiB)": 111.15, + "step": 60735, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.64887209, + "epoch": 1.5408422120750886, + "grad_norm": 5.15625, + "learning_rate": 1.3730776175858506e-06, + "loss": 1.66182766, + "memory(GiB)": 111.15, + "step": 60740, + "train_speed(iter/s)": 0.447288 + }, + { + "acc": 0.66101518, + "epoch": 1.5409690512430239, + "grad_norm": 5.375, + "learning_rate": 1.3723558836142631e-06, + "loss": 1.61735401, + "memory(GiB)": 111.15, + "step": 60745, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.66229515, + "epoch": 1.541095890410959, + "grad_norm": 4.53125, + "learning_rate": 1.371634309202663e-06, + "loss": 1.57234831, + "memory(GiB)": 111.15, + "step": 60750, + "train_speed(iter/s)": 0.447327 + }, + { + "acc": 0.64774632, + "epoch": 1.541222729578894, + "grad_norm": 5.375, + "learning_rate": 1.3709128943827842e-06, + "loss": 1.61919231, + "memory(GiB)": 111.15, + "step": 60755, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.63367605, + "epoch": 1.541349568746829, + "grad_norm": 5.78125, + "learning_rate": 1.3701916391863573e-06, + "loss": 1.67353516, + "memory(GiB)": 111.15, + "step": 60760, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.63736086, + "epoch": 1.541476407914764, + "grad_norm": 5.84375, + "learning_rate": 1.3694705436451093e-06, + "loss": 1.63860474, + "memory(GiB)": 111.15, + "step": 60765, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.66485586, + "epoch": 1.541603247082699, + "grad_norm": 5.5, + "learning_rate": 1.368749607790758e-06, + "loss": 1.52697983, + "memory(GiB)": 111.15, + "step": 60770, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.65197663, + "epoch": 1.5417300862506342, + "grad_norm": 5.40625, + "learning_rate": 1.3680288316550095e-06, + "loss": 1.55682659, + "memory(GiB)": 111.15, + "step": 60775, + "train_speed(iter/s)": 0.447425 + }, + { + "acc": 0.64534974, + "epoch": 1.5418569254185692, + "grad_norm": 6.53125, + "learning_rate": 1.3673082152695672e-06, + "loss": 1.6965704, + "memory(GiB)": 111.15, + "step": 60780, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.62665434, + "epoch": 1.5419837645865044, + "grad_norm": 4.6875, + "learning_rate": 1.3665877586661296e-06, + "loss": 1.70153236, + "memory(GiB)": 111.15, + "step": 60785, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.65638051, + "epoch": 1.5421106037544394, + "grad_norm": 4.9375, + "learning_rate": 1.3658674618763862e-06, + "loss": 1.57305479, + "memory(GiB)": 111.15, + "step": 60790, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.64082646, + "epoch": 1.5422374429223744, + "grad_norm": 5.8125, + "learning_rate": 1.365147324932014e-06, + "loss": 1.63223419, + "memory(GiB)": 111.15, + "step": 60795, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.63855052, + "epoch": 1.5423642820903094, + "grad_norm": 6.03125, + "learning_rate": 1.3644273478646925e-06, + "loss": 1.67751484, + "memory(GiB)": 111.15, + "step": 60800, + "train_speed(iter/s)": 0.447523 + }, + { + "epoch": 1.5423642820903094, + "eval_acc": 0.6434381079819875, + "eval_loss": 1.6075670719146729, + "eval_runtime": 113.7525, + "eval_samples_per_second": 55.999, + "eval_steps_per_second": 27.999, + "step": 60800 + }, + { + "acc": 0.64639711, + "epoch": 1.5424911212582444, + "grad_norm": 5.5625, + "learning_rate": 1.3637075307060877e-06, + "loss": 1.66062603, + "memory(GiB)": 111.15, + "step": 60805, + "train_speed(iter/s)": 0.447153 + }, + { + "acc": 0.65552855, + "epoch": 1.5426179604261796, + "grad_norm": 4.65625, + "learning_rate": 1.362987873487862e-06, + "loss": 1.6144783, + "memory(GiB)": 111.15, + "step": 60810, + "train_speed(iter/s)": 0.447172 + }, + { + "acc": 0.64176884, + "epoch": 1.5427447995941148, + "grad_norm": 5.3125, + "learning_rate": 1.362268376241665e-06, + "loss": 1.63098412, + "memory(GiB)": 111.15, + "step": 60815, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.62565832, + "epoch": 1.5428716387620498, + "grad_norm": 6.8125, + "learning_rate": 1.3615490389991476e-06, + "loss": 1.72352066, + "memory(GiB)": 111.15, + "step": 60820, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.65074596, + "epoch": 1.5429984779299848, + "grad_norm": 5.21875, + "learning_rate": 1.360829861791948e-06, + "loss": 1.61862831, + "memory(GiB)": 111.15, + "step": 60825, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.64856572, + "epoch": 1.5431253170979198, + "grad_norm": 5.46875, + "learning_rate": 1.3601108446516985e-06, + "loss": 1.60735035, + "memory(GiB)": 111.15, + "step": 60830, + "train_speed(iter/s)": 0.44725 + }, + { + "acc": 0.66562719, + "epoch": 1.5432521562658548, + "grad_norm": 5.90625, + "learning_rate": 1.3593919876100254e-06, + "loss": 1.57543087, + "memory(GiB)": 111.15, + "step": 60835, + "train_speed(iter/s)": 0.44727 + }, + { + "acc": 0.64553766, + "epoch": 1.54337899543379, + "grad_norm": 5.71875, + "learning_rate": 1.3586732906985467e-06, + "loss": 1.65457153, + "memory(GiB)": 111.15, + "step": 60840, + "train_speed(iter/s)": 0.447289 + }, + { + "acc": 0.65302753, + "epoch": 1.543505834601725, + "grad_norm": 5.84375, + "learning_rate": 1.357954753948874e-06, + "loss": 1.5954071, + "memory(GiB)": 111.15, + "step": 60845, + "train_speed(iter/s)": 0.447309 + }, + { + "acc": 0.67651176, + "epoch": 1.5436326737696602, + "grad_norm": 5.09375, + "learning_rate": 1.3572363773926117e-06, + "loss": 1.55727234, + "memory(GiB)": 111.15, + "step": 60850, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.66119065, + "epoch": 1.5437595129375952, + "grad_norm": 6.1875, + "learning_rate": 1.3565181610613571e-06, + "loss": 1.61036644, + "memory(GiB)": 111.15, + "step": 60855, + "train_speed(iter/s)": 0.447348 + }, + { + "acc": 0.64773784, + "epoch": 1.5438863521055302, + "grad_norm": 5.90625, + "learning_rate": 1.3558001049867008e-06, + "loss": 1.63665276, + "memory(GiB)": 111.15, + "step": 60860, + "train_speed(iter/s)": 0.447368 + }, + { + "acc": 0.65569735, + "epoch": 1.5440131912734651, + "grad_norm": 5.125, + "learning_rate": 1.3550822092002264e-06, + "loss": 1.59874134, + "memory(GiB)": 111.15, + "step": 60865, + "train_speed(iter/s)": 0.447387 + }, + { + "acc": 0.65851564, + "epoch": 1.5441400304414004, + "grad_norm": 4.53125, + "learning_rate": 1.3543644737335099e-06, + "loss": 1.54289837, + "memory(GiB)": 111.15, + "step": 60870, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.63209233, + "epoch": 1.5442668696093353, + "grad_norm": 5.75, + "learning_rate": 1.35364689861812e-06, + "loss": 1.6784441, + "memory(GiB)": 111.15, + "step": 60875, + "train_speed(iter/s)": 0.447426 + }, + { + "acc": 0.66247635, + "epoch": 1.5443937087772706, + "grad_norm": 5.59375, + "learning_rate": 1.3529294838856194e-06, + "loss": 1.59007797, + "memory(GiB)": 111.15, + "step": 60880, + "train_speed(iter/s)": 0.447445 + }, + { + "acc": 0.63497677, + "epoch": 1.5445205479452055, + "grad_norm": 5.8125, + "learning_rate": 1.3522122295675616e-06, + "loss": 1.62615585, + "memory(GiB)": 111.15, + "step": 60885, + "train_speed(iter/s)": 0.447465 + }, + { + "acc": 0.64899282, + "epoch": 1.5446473871131405, + "grad_norm": 6.1875, + "learning_rate": 1.351495135695499e-06, + "loss": 1.63584118, + "memory(GiB)": 111.15, + "step": 60890, + "train_speed(iter/s)": 0.447484 + }, + { + "acc": 0.65946703, + "epoch": 1.5447742262810755, + "grad_norm": 4.75, + "learning_rate": 1.3507782023009692e-06, + "loss": 1.59263887, + "memory(GiB)": 111.15, + "step": 60895, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.65517211, + "epoch": 1.5449010654490105, + "grad_norm": 5.21875, + "learning_rate": 1.3500614294155056e-06, + "loss": 1.60624809, + "memory(GiB)": 111.15, + "step": 60900, + "train_speed(iter/s)": 0.447523 + }, + { + "epoch": 1.5449010654490105, + "eval_acc": 0.6433980023202797, + "eval_loss": 1.6075862646102905, + "eval_runtime": 113.5432, + "eval_samples_per_second": 56.102, + "eval_steps_per_second": 28.051, + "step": 60900 + }, + { + "acc": 0.66381407, + "epoch": 1.5450279046169457, + "grad_norm": 5.25, + "learning_rate": 1.3493448170706347e-06, + "loss": 1.61134605, + "memory(GiB)": 111.15, + "step": 60905, + "train_speed(iter/s)": 0.447154 + }, + { + "acc": 0.65100985, + "epoch": 1.545154743784881, + "grad_norm": 5.125, + "learning_rate": 1.348628365297881e-06, + "loss": 1.64735832, + "memory(GiB)": 111.15, + "step": 60910, + "train_speed(iter/s)": 0.447173 + }, + { + "acc": 0.64934096, + "epoch": 1.545281582952816, + "grad_norm": 5.3125, + "learning_rate": 1.3479120741287526e-06, + "loss": 1.60158768, + "memory(GiB)": 111.15, + "step": 60915, + "train_speed(iter/s)": 0.447192 + }, + { + "acc": 0.64882383, + "epoch": 1.545408422120751, + "grad_norm": 4.90625, + "learning_rate": 1.3471959435947552e-06, + "loss": 1.70731888, + "memory(GiB)": 111.15, + "step": 60920, + "train_speed(iter/s)": 0.447211 + }, + { + "acc": 0.63997927, + "epoch": 1.545535261288686, + "grad_norm": 4.96875, + "learning_rate": 1.3464799737273898e-06, + "loss": 1.65356178, + "memory(GiB)": 111.15, + "step": 60925, + "train_speed(iter/s)": 0.447231 + }, + { + "acc": 0.64967322, + "epoch": 1.545662100456621, + "grad_norm": 5.9375, + "learning_rate": 1.3457641645581487e-06, + "loss": 1.62131824, + "memory(GiB)": 111.15, + "step": 60930, + "train_speed(iter/s)": 0.44725 + }, + { + "acc": 0.6425518, + "epoch": 1.545788939624556, + "grad_norm": 5.5, + "learning_rate": 1.3450485161185133e-06, + "loss": 1.66434174, + "memory(GiB)": 111.15, + "step": 60935, + "train_speed(iter/s)": 0.447269 + }, + { + "acc": 0.64992213, + "epoch": 1.545915778792491, + "grad_norm": 6.59375, + "learning_rate": 1.344333028439961e-06, + "loss": 1.58736763, + "memory(GiB)": 111.15, + "step": 60940, + "train_speed(iter/s)": 0.447289 + }, + { + "acc": 0.66314464, + "epoch": 1.5460426179604263, + "grad_norm": 5.65625, + "learning_rate": 1.3436177015539647e-06, + "loss": 1.57411308, + "memory(GiB)": 111.15, + "step": 60945, + "train_speed(iter/s)": 0.447308 + }, + { + "acc": 0.67148051, + "epoch": 1.5461694571283613, + "grad_norm": 5.59375, + "learning_rate": 1.3429025354919877e-06, + "loss": 1.57624998, + "memory(GiB)": 111.15, + "step": 60950, + "train_speed(iter/s)": 0.447328 + }, + { + "acc": 0.63875656, + "epoch": 1.5462962962962963, + "grad_norm": 5.53125, + "learning_rate": 1.3421875302854826e-06, + "loss": 1.6338501, + "memory(GiB)": 111.15, + "step": 60955, + "train_speed(iter/s)": 0.447347 + }, + { + "acc": 0.63833761, + "epoch": 1.5464231354642313, + "grad_norm": 5.09375, + "learning_rate": 1.3414726859659016e-06, + "loss": 1.63572445, + "memory(GiB)": 111.15, + "step": 60960, + "train_speed(iter/s)": 0.447367 + }, + { + "acc": 0.63177052, + "epoch": 1.5465499746321663, + "grad_norm": 5.25, + "learning_rate": 1.3407580025646866e-06, + "loss": 1.67703743, + "memory(GiB)": 111.15, + "step": 60965, + "train_speed(iter/s)": 0.447386 + }, + { + "acc": 0.65079594, + "epoch": 1.5466768138001015, + "grad_norm": 6.6875, + "learning_rate": 1.3400434801132716e-06, + "loss": 1.65920067, + "memory(GiB)": 111.15, + "step": 60970, + "train_speed(iter/s)": 0.447406 + }, + { + "acc": 0.64886904, + "epoch": 1.5468036529680367, + "grad_norm": 5.21875, + "learning_rate": 1.3393291186430852e-06, + "loss": 1.6103117, + "memory(GiB)": 111.15, + "step": 60975, + "train_speed(iter/s)": 0.447425 + }, + { + "acc": 0.63858614, + "epoch": 1.5469304921359717, + "grad_norm": 7.84375, + "learning_rate": 1.338614918185548e-06, + "loss": 1.66281147, + "memory(GiB)": 111.15, + "step": 60980, + "train_speed(iter/s)": 0.447444 + }, + { + "acc": 0.65356932, + "epoch": 1.5470573313039067, + "grad_norm": 4.59375, + "learning_rate": 1.3379008787720732e-06, + "loss": 1.6343811, + "memory(GiB)": 111.15, + "step": 60985, + "train_speed(iter/s)": 0.447464 + }, + { + "acc": 0.6623704, + "epoch": 1.5471841704718416, + "grad_norm": 5.59375, + "learning_rate": 1.3371870004340681e-06, + "loss": 1.64885101, + "memory(GiB)": 111.15, + "step": 60990, + "train_speed(iter/s)": 0.447483 + }, + { + "acc": 0.66089325, + "epoch": 1.5473110096397766, + "grad_norm": 5.15625, + "learning_rate": 1.3364732832029315e-06, + "loss": 1.59869976, + "memory(GiB)": 111.15, + "step": 60995, + "train_speed(iter/s)": 0.447503 + }, + { + "acc": 0.65752225, + "epoch": 1.5474378488077118, + "grad_norm": 5.25, + "learning_rate": 1.335759727110057e-06, + "loss": 1.59758549, + "memory(GiB)": 111.15, + "step": 61000, + "train_speed(iter/s)": 0.447522 + }, + { + "epoch": 1.5474378488077118, + "eval_acc": 0.6434193084530619, + "eval_loss": 1.607518196105957, + "eval_runtime": 114.048, + "eval_samples_per_second": 55.854, + "eval_steps_per_second": 27.927, + "step": 61000 + }, + { + "acc": 0.64996653, + "epoch": 1.5475646879756468, + "grad_norm": 6.3125, + "learning_rate": 1.335046332186829e-06, + "loss": 1.63765888, + "memory(GiB)": 111.15, + "step": 61005, + "train_speed(iter/s)": 0.447153 + }, + { + "acc": 0.6601757, + "epoch": 1.547691527143582, + "grad_norm": 5.5625, + "learning_rate": 1.3343330984646262e-06, + "loss": 1.55656118, + "memory(GiB)": 111.15, + "step": 61010, + "train_speed(iter/s)": 0.447172 + }, + { + "acc": 0.66854343, + "epoch": 1.547818366311517, + "grad_norm": 5.0, + "learning_rate": 1.33362002597482e-06, + "loss": 1.62073479, + "memory(GiB)": 111.15, + "step": 61015, + "train_speed(iter/s)": 0.447191 + }, + { + "acc": 0.64274397, + "epoch": 1.547945205479452, + "grad_norm": 4.9375, + "learning_rate": 1.3329071147487743e-06, + "loss": 1.63980618, + "memory(GiB)": 111.15, + "step": 61020, + "train_speed(iter/s)": 0.44721 + }, + { + "acc": 0.6483264, + "epoch": 1.548072044647387, + "grad_norm": 5.5625, + "learning_rate": 1.3321943648178442e-06, + "loss": 1.5995472, + "memory(GiB)": 111.15, + "step": 61025, + "train_speed(iter/s)": 0.447229 + }, + { + "acc": 0.64403276, + "epoch": 1.5481988838153222, + "grad_norm": 4.96875, + "learning_rate": 1.3314817762133848e-06, + "loss": 1.63764954, + "memory(GiB)": 111.15, + "step": 61030, + "train_speed(iter/s)": 0.447249 + }, + { + "acc": 0.6641202, + "epoch": 1.5483257229832572, + "grad_norm": 5.59375, + "learning_rate": 1.330769348966734e-06, + "loss": 1.64345703, + "memory(GiB)": 111.15, + "step": 61035, + "train_speed(iter/s)": 0.447268 + }, + { + "acc": 0.64847636, + "epoch": 1.5484525621511924, + "grad_norm": 5.5625, + "learning_rate": 1.3300570831092292e-06, + "loss": 1.62792549, + "memory(GiB)": 111.15, + "step": 61040, + "train_speed(iter/s)": 0.447287 + }, + { + "acc": 0.6679575, + "epoch": 1.5485794013191274, + "grad_norm": 4.8125, + "learning_rate": 1.3293449786721973e-06, + "loss": 1.53033714, + "memory(GiB)": 111.15, + "step": 61045, + "train_speed(iter/s)": 0.447306 + }, + { + "acc": 0.6463676, + "epoch": 1.5487062404870624, + "grad_norm": 5.59375, + "learning_rate": 1.3286330356869648e-06, + "loss": 1.65582237, + "memory(GiB)": 111.15, + "step": 61050, + "train_speed(iter/s)": 0.447325 + }, + { + "acc": 0.63849764, + "epoch": 1.5488330796549974, + "grad_norm": 5.875, + "learning_rate": 1.3279212541848413e-06, + "loss": 1.67342072, + "memory(GiB)": 111.15, + "step": 61055, + "train_speed(iter/s)": 0.447345 + }, + { + "acc": 0.6395184, + "epoch": 1.5489599188229324, + "grad_norm": 6.53125, + "learning_rate": 1.3272096341971342e-06, + "loss": 1.66531467, + "memory(GiB)": 111.15, + "step": 61060, + "train_speed(iter/s)": 0.447364 + }, + { + "acc": 0.65087419, + "epoch": 1.5490867579908676, + "grad_norm": 6.0, + "learning_rate": 1.326498175755147e-06, + "loss": 1.63271427, + "memory(GiB)": 111.15, + "step": 61065, + "train_speed(iter/s)": 0.447384 + }, + { + "acc": 0.66073451, + "epoch": 1.5492135971588028, + "grad_norm": 7.1875, + "learning_rate": 1.3257868788901722e-06, + "loss": 1.64948101, + "memory(GiB)": 111.15, + "step": 61070, + "train_speed(iter/s)": 0.447403 + }, + { + "acc": 0.65232401, + "epoch": 1.5493404363267378, + "grad_norm": 5.125, + "learning_rate": 1.3250757436334932e-06, + "loss": 1.5962863, + "memory(GiB)": 111.15, + "step": 61075, + "train_speed(iter/s)": 0.447422 + }, + { + "acc": 0.6545404, + "epoch": 1.5494672754946728, + "grad_norm": 6.03125, + "learning_rate": 1.3243647700163887e-06, + "loss": 1.58335609, + "memory(GiB)": 111.15, + "step": 61080, + "train_speed(iter/s)": 0.447441 + }, + { + "acc": 0.6480979, + "epoch": 1.5495941146626078, + "grad_norm": 4.78125, + "learning_rate": 1.323653958070134e-06, + "loss": 1.63703709, + "memory(GiB)": 111.15, + "step": 61085, + "train_speed(iter/s)": 0.447461 + }, + { + "acc": 0.64916668, + "epoch": 1.5497209538305428, + "grad_norm": 6.3125, + "learning_rate": 1.3229433078259928e-06, + "loss": 1.60829544, + "memory(GiB)": 111.15, + "step": 61090, + "train_speed(iter/s)": 0.44748 + }, + { + "acc": 0.64160552, + "epoch": 1.549847792998478, + "grad_norm": 4.8125, + "learning_rate": 1.3222328193152195e-06, + "loss": 1.57272158, + "memory(GiB)": 111.15, + "step": 61095, + "train_speed(iter/s)": 0.447499 + }, + { + "acc": 0.66049376, + "epoch": 1.549974632166413, + "grad_norm": 5.5, + "learning_rate": 1.3215224925690683e-06, + "loss": 1.64803162, + "memory(GiB)": 111.15, + "step": 61100, + "train_speed(iter/s)": 0.447519 + }, + { + "epoch": 1.549974632166413, + "eval_acc": 0.6433670875393798, + "eval_loss": 1.607504963874817, + "eval_runtime": 114.8701, + "eval_samples_per_second": 55.454, + "eval_steps_per_second": 27.727, + "step": 61100 + }, + { + "acc": 0.66711845, + "epoch": 1.5501014713343482, + "grad_norm": 4.75, + "learning_rate": 1.3208123276187807e-06, + "loss": 1.55229549, + "memory(GiB)": 111.15, + "step": 61105, + "train_speed(iter/s)": 0.447147 + }, + { + "acc": 0.64694872, + "epoch": 1.5502283105022832, + "grad_norm": 4.71875, + "learning_rate": 1.3201023244955952e-06, + "loss": 1.62568359, + "memory(GiB)": 111.15, + "step": 61110, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.66077862, + "epoch": 1.5503551496702181, + "grad_norm": 4.8125, + "learning_rate": 1.319392483230736e-06, + "loss": 1.60535698, + "memory(GiB)": 111.15, + "step": 61115, + "train_speed(iter/s)": 0.447185 + }, + { + "acc": 0.64251194, + "epoch": 1.5504819888381531, + "grad_norm": 5.25, + "learning_rate": 1.318682803855429e-06, + "loss": 1.70156612, + "memory(GiB)": 111.15, + "step": 61120, + "train_speed(iter/s)": 0.447204 + }, + { + "acc": 0.65304813, + "epoch": 1.5506088280060881, + "grad_norm": 6.21875, + "learning_rate": 1.3179732864008888e-06, + "loss": 1.65208435, + "memory(GiB)": 111.15, + "step": 61125, + "train_speed(iter/s)": 0.447224 + }, + { + "acc": 0.65565825, + "epoch": 1.5507356671740233, + "grad_norm": 5.65625, + "learning_rate": 1.3172639308983226e-06, + "loss": 1.6057724, + "memory(GiB)": 111.15, + "step": 61130, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.66161776, + "epoch": 1.5508625063419585, + "grad_norm": 4.9375, + "learning_rate": 1.3165547373789306e-06, + "loss": 1.56553326, + "memory(GiB)": 111.15, + "step": 61135, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.64531598, + "epoch": 1.5509893455098935, + "grad_norm": 4.90625, + "learning_rate": 1.3158457058739066e-06, + "loss": 1.6227932, + "memory(GiB)": 111.15, + "step": 61140, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.65768933, + "epoch": 1.5511161846778285, + "grad_norm": 5.84375, + "learning_rate": 1.3151368364144373e-06, + "loss": 1.64881725, + "memory(GiB)": 111.15, + "step": 61145, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.64592133, + "epoch": 1.5512430238457635, + "grad_norm": 5.4375, + "learning_rate": 1.3144281290317012e-06, + "loss": 1.63901348, + "memory(GiB)": 111.15, + "step": 61150, + "train_speed(iter/s)": 0.447321 + }, + { + "acc": 0.65983334, + "epoch": 1.5513698630136985, + "grad_norm": 5.90625, + "learning_rate": 1.3137195837568716e-06, + "loss": 1.53631496, + "memory(GiB)": 111.15, + "step": 61155, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.65735245, + "epoch": 1.5514967021816337, + "grad_norm": 5.4375, + "learning_rate": 1.313011200621112e-06, + "loss": 1.6238739, + "memory(GiB)": 111.15, + "step": 61160, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.64902515, + "epoch": 1.5516235413495687, + "grad_norm": 4.25, + "learning_rate": 1.312302979655582e-06, + "loss": 1.60589142, + "memory(GiB)": 111.15, + "step": 61165, + "train_speed(iter/s)": 0.447379 + }, + { + "acc": 0.65051441, + "epoch": 1.551750380517504, + "grad_norm": 5.96875, + "learning_rate": 1.3115949208914302e-06, + "loss": 1.62225189, + "memory(GiB)": 111.15, + "step": 61170, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.64366775, + "epoch": 1.551877219685439, + "grad_norm": 4.625, + "learning_rate": 1.3108870243598022e-06, + "loss": 1.65092125, + "memory(GiB)": 111.15, + "step": 61175, + "train_speed(iter/s)": 0.447418 + }, + { + "acc": 0.64787331, + "epoch": 1.552004058853374, + "grad_norm": 5.59375, + "learning_rate": 1.310179290091833e-06, + "loss": 1.66537113, + "memory(GiB)": 111.15, + "step": 61180, + "train_speed(iter/s)": 0.447437 + }, + { + "acc": 0.66157169, + "epoch": 1.5521308980213089, + "grad_norm": 5.5625, + "learning_rate": 1.3094717181186518e-06, + "loss": 1.56579065, + "memory(GiB)": 111.15, + "step": 61185, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64603529, + "epoch": 1.552257737189244, + "grad_norm": 6.65625, + "learning_rate": 1.3087643084713836e-06, + "loss": 1.60785942, + "memory(GiB)": 111.15, + "step": 61190, + "train_speed(iter/s)": 0.447476 + }, + { + "acc": 0.65732479, + "epoch": 1.552384576357179, + "grad_norm": 5.84375, + "learning_rate": 1.30805706118114e-06, + "loss": 1.59543486, + "memory(GiB)": 111.15, + "step": 61195, + "train_speed(iter/s)": 0.447495 + }, + { + "acc": 0.64583817, + "epoch": 1.5525114155251143, + "grad_norm": 6.0625, + "learning_rate": 1.3073499762790287e-06, + "loss": 1.63877754, + "memory(GiB)": 111.15, + "step": 61200, + "train_speed(iter/s)": 0.447514 + }, + { + "epoch": 1.5525114155251143, + "eval_acc": 0.6433833804644487, + "eval_loss": 1.6075305938720703, + "eval_runtime": 113.8115, + "eval_samples_per_second": 55.97, + "eval_steps_per_second": 27.985, + "step": 61200 + }, + { + "acc": 0.65100346, + "epoch": 1.5526382546930493, + "grad_norm": 4.6875, + "learning_rate": 1.306643053796154e-06, + "loss": 1.6063448, + "memory(GiB)": 111.15, + "step": 61205, + "train_speed(iter/s)": 0.447146 + }, + { + "acc": 0.65683699, + "epoch": 1.5527650938609843, + "grad_norm": 6.8125, + "learning_rate": 1.3059362937636084e-06, + "loss": 1.62712421, + "memory(GiB)": 111.15, + "step": 61210, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.66392717, + "epoch": 1.5528919330289193, + "grad_norm": 4.8125, + "learning_rate": 1.3052296962124756e-06, + "loss": 1.53788662, + "memory(GiB)": 111.15, + "step": 61215, + "train_speed(iter/s)": 0.447185 + }, + { + "acc": 0.63626208, + "epoch": 1.5530187721968542, + "grad_norm": 5.75, + "learning_rate": 1.3045232611738357e-06, + "loss": 1.6668684, + "memory(GiB)": 111.15, + "step": 61220, + "train_speed(iter/s)": 0.447204 + }, + { + "acc": 0.64396672, + "epoch": 1.5531456113647895, + "grad_norm": 5.375, + "learning_rate": 1.3038169886787632e-06, + "loss": 1.61758366, + "memory(GiB)": 111.15, + "step": 61225, + "train_speed(iter/s)": 0.447224 + }, + { + "acc": 0.67758198, + "epoch": 1.5532724505327247, + "grad_norm": 7.40625, + "learning_rate": 1.3031108787583235e-06, + "loss": 1.57856312, + "memory(GiB)": 111.15, + "step": 61230, + "train_speed(iter/s)": 0.447243 + }, + { + "acc": 0.64081879, + "epoch": 1.5533992897006597, + "grad_norm": 5.15625, + "learning_rate": 1.3024049314435694e-06, + "loss": 1.60206642, + "memory(GiB)": 111.15, + "step": 61235, + "train_speed(iter/s)": 0.447262 + }, + { + "acc": 0.6437161, + "epoch": 1.5535261288685946, + "grad_norm": 5.4375, + "learning_rate": 1.301699146765557e-06, + "loss": 1.62397499, + "memory(GiB)": 111.15, + "step": 61240, + "train_speed(iter/s)": 0.447281 + }, + { + "acc": 0.66659822, + "epoch": 1.5536529680365296, + "grad_norm": 4.875, + "learning_rate": 1.3009935247553274e-06, + "loss": 1.62088966, + "memory(GiB)": 111.15, + "step": 61245, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.63716702, + "epoch": 1.5537798072044646, + "grad_norm": 5.3125, + "learning_rate": 1.3002880654439192e-06, + "loss": 1.69327583, + "memory(GiB)": 111.15, + "step": 61250, + "train_speed(iter/s)": 0.44732 + }, + { + "acc": 0.6428359, + "epoch": 1.5539066463723998, + "grad_norm": 4.8125, + "learning_rate": 1.2995827688623568e-06, + "loss": 1.67391853, + "memory(GiB)": 111.15, + "step": 61255, + "train_speed(iter/s)": 0.447339 + }, + { + "acc": 0.66446075, + "epoch": 1.5540334855403348, + "grad_norm": 4.90625, + "learning_rate": 1.298877635041667e-06, + "loss": 1.51591797, + "memory(GiB)": 111.15, + "step": 61260, + "train_speed(iter/s)": 0.447358 + }, + { + "acc": 0.64247451, + "epoch": 1.55416032470827, + "grad_norm": 4.96875, + "learning_rate": 1.2981726640128633e-06, + "loss": 1.65702934, + "memory(GiB)": 111.15, + "step": 61265, + "train_speed(iter/s)": 0.447377 + }, + { + "acc": 0.64806128, + "epoch": 1.554287163876205, + "grad_norm": 4.5625, + "learning_rate": 1.297467855806953e-06, + "loss": 1.55410271, + "memory(GiB)": 111.15, + "step": 61270, + "train_speed(iter/s)": 0.447397 + }, + { + "acc": 0.65592566, + "epoch": 1.55441400304414, + "grad_norm": 5.53125, + "learning_rate": 1.2967632104549371e-06, + "loss": 1.63307343, + "memory(GiB)": 111.15, + "step": 61275, + "train_speed(iter/s)": 0.447416 + }, + { + "acc": 0.64620557, + "epoch": 1.554540842212075, + "grad_norm": 4.84375, + "learning_rate": 1.296058727987809e-06, + "loss": 1.61457481, + "memory(GiB)": 111.15, + "step": 61280, + "train_speed(iter/s)": 0.447435 + }, + { + "acc": 0.65431786, + "epoch": 1.55466768138001, + "grad_norm": 4.8125, + "learning_rate": 1.295354408436555e-06, + "loss": 1.67000694, + "memory(GiB)": 111.15, + "step": 61285, + "train_speed(iter/s)": 0.447454 + }, + { + "acc": 0.65833549, + "epoch": 1.5547945205479452, + "grad_norm": 6.3125, + "learning_rate": 1.294650251832154e-06, + "loss": 1.57710438, + "memory(GiB)": 111.15, + "step": 61290, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.64726534, + "epoch": 1.5549213597158804, + "grad_norm": 5.96875, + "learning_rate": 1.2939462582055784e-06, + "loss": 1.58634167, + "memory(GiB)": 111.15, + "step": 61295, + "train_speed(iter/s)": 0.447493 + }, + { + "acc": 0.66010089, + "epoch": 1.5550481988838154, + "grad_norm": 5.0625, + "learning_rate": 1.2932424275877926e-06, + "loss": 1.56881275, + "memory(GiB)": 111.15, + "step": 61300, + "train_speed(iter/s)": 0.447512 + }, + { + "epoch": 1.5550481988838154, + "eval_acc": 0.643369176375927, + "eval_loss": 1.6075748205184937, + "eval_runtime": 113.3154, + "eval_samples_per_second": 56.215, + "eval_steps_per_second": 28.107, + "step": 61300 + }, + { + "acc": 0.65704412, + "epoch": 1.5551750380517504, + "grad_norm": 5.5625, + "learning_rate": 1.2925387600097543e-06, + "loss": 1.61461449, + "memory(GiB)": 111.15, + "step": 61305, + "train_speed(iter/s)": 0.447146 + }, + { + "acc": 0.64017768, + "epoch": 1.5553018772196854, + "grad_norm": 4.1875, + "learning_rate": 1.291835255502414e-06, + "loss": 1.58383713, + "memory(GiB)": 111.15, + "step": 61310, + "train_speed(iter/s)": 0.447165 + }, + { + "acc": 0.63994064, + "epoch": 1.5554287163876204, + "grad_norm": 4.53125, + "learning_rate": 1.2911319140967148e-06, + "loss": 1.62919044, + "memory(GiB)": 111.15, + "step": 61315, + "train_speed(iter/s)": 0.447184 + }, + { + "acc": 0.64251842, + "epoch": 1.5555555555555556, + "grad_norm": 5.09375, + "learning_rate": 1.290428735823593e-06, + "loss": 1.64685726, + "memory(GiB)": 111.15, + "step": 61320, + "train_speed(iter/s)": 0.447203 + }, + { + "acc": 0.63787889, + "epoch": 1.5556823947234906, + "grad_norm": 5.3125, + "learning_rate": 1.2897257207139758e-06, + "loss": 1.70393219, + "memory(GiB)": 111.15, + "step": 61325, + "train_speed(iter/s)": 0.447223 + }, + { + "acc": 0.64722834, + "epoch": 1.5558092338914258, + "grad_norm": 5.28125, + "learning_rate": 1.28902286879879e-06, + "loss": 1.60652599, + "memory(GiB)": 111.15, + "step": 61330, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.64182625, + "epoch": 1.5559360730593608, + "grad_norm": 5.5625, + "learning_rate": 1.2883201801089445e-06, + "loss": 1.65496826, + "memory(GiB)": 111.15, + "step": 61335, + "train_speed(iter/s)": 0.447261 + }, + { + "acc": 0.63590565, + "epoch": 1.5560629122272958, + "grad_norm": 5.0, + "learning_rate": 1.2876176546753494e-06, + "loss": 1.68897133, + "memory(GiB)": 111.15, + "step": 61340, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.64597483, + "epoch": 1.5561897513952307, + "grad_norm": 5.40625, + "learning_rate": 1.286915292528903e-06, + "loss": 1.68069782, + "memory(GiB)": 111.15, + "step": 61345, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.64783001, + "epoch": 1.556316590563166, + "grad_norm": 5.8125, + "learning_rate": 1.286213093700503e-06, + "loss": 1.60983677, + "memory(GiB)": 111.15, + "step": 61350, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.65176487, + "epoch": 1.556443429731101, + "grad_norm": 5.78125, + "learning_rate": 1.28551105822103e-06, + "loss": 1.57596111, + "memory(GiB)": 111.15, + "step": 61355, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.64773107, + "epoch": 1.5565702688990362, + "grad_norm": 5.59375, + "learning_rate": 1.2848091861213636e-06, + "loss": 1.59041634, + "memory(GiB)": 111.15, + "step": 61360, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.63490582, + "epoch": 1.5566971080669711, + "grad_norm": 5.5, + "learning_rate": 1.2841074774323775e-06, + "loss": 1.71108589, + "memory(GiB)": 111.15, + "step": 61365, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.63696947, + "epoch": 1.5568239472349061, + "grad_norm": 4.8125, + "learning_rate": 1.2834059321849363e-06, + "loss": 1.65976295, + "memory(GiB)": 111.15, + "step": 61370, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.6757596, + "epoch": 1.5569507864028411, + "grad_norm": 5.75, + "learning_rate": 1.2827045504098928e-06, + "loss": 1.62564716, + "memory(GiB)": 111.15, + "step": 61375, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.65840054, + "epoch": 1.5570776255707761, + "grad_norm": 5.8125, + "learning_rate": 1.2820033321381009e-06, + "loss": 1.59154978, + "memory(GiB)": 111.15, + "step": 61380, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.66883268, + "epoch": 1.5572044647387113, + "grad_norm": 5.53125, + "learning_rate": 1.2813022774004024e-06, + "loss": 1.56321373, + "memory(GiB)": 111.15, + "step": 61385, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.65005627, + "epoch": 1.5573313039066465, + "grad_norm": 5.40625, + "learning_rate": 1.280601386227634e-06, + "loss": 1.64493694, + "memory(GiB)": 111.15, + "step": 61390, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.64819651, + "epoch": 1.5574581430745815, + "grad_norm": 5.25, + "learning_rate": 1.279900658650619e-06, + "loss": 1.66435699, + "memory(GiB)": 111.15, + "step": 61395, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.63696933, + "epoch": 1.5575849822425165, + "grad_norm": 6.09375, + "learning_rate": 1.2792000947001842e-06, + "loss": 1.66545067, + "memory(GiB)": 111.15, + "step": 61400, + "train_speed(iter/s)": 0.447511 + }, + { + "epoch": 1.5575849822425165, + "eval_acc": 0.643394660181804, + "eval_loss": 1.6075458526611328, + "eval_runtime": 113.9246, + "eval_samples_per_second": 55.914, + "eval_steps_per_second": 27.957, + "step": 61400 + }, + { + "acc": 0.65034833, + "epoch": 1.5577118214104515, + "grad_norm": 4.6875, + "learning_rate": 1.2784996944071415e-06, + "loss": 1.65339413, + "memory(GiB)": 111.15, + "step": 61405, + "train_speed(iter/s)": 0.447144 + }, + { + "acc": 0.65096321, + "epoch": 1.5578386605783865, + "grad_norm": 5.3125, + "learning_rate": 1.2777994578022972e-06, + "loss": 1.67180061, + "memory(GiB)": 111.15, + "step": 61410, + "train_speed(iter/s)": 0.447164 + }, + { + "acc": 0.65301952, + "epoch": 1.5579654997463217, + "grad_norm": 6.0, + "learning_rate": 1.2770993849164514e-06, + "loss": 1.65988407, + "memory(GiB)": 111.15, + "step": 61415, + "train_speed(iter/s)": 0.447183 + }, + { + "acc": 0.63351002, + "epoch": 1.5580923389142567, + "grad_norm": 5.375, + "learning_rate": 1.276399475780396e-06, + "loss": 1.72411804, + "memory(GiB)": 111.15, + "step": 61420, + "train_speed(iter/s)": 0.447202 + }, + { + "acc": 0.65545011, + "epoch": 1.558219178082192, + "grad_norm": 4.9375, + "learning_rate": 1.2756997304249164e-06, + "loss": 1.6246727, + "memory(GiB)": 111.15, + "step": 61425, + "train_speed(iter/s)": 0.447222 + }, + { + "acc": 0.64902658, + "epoch": 1.558346017250127, + "grad_norm": 5.5625, + "learning_rate": 1.2750001488807906e-06, + "loss": 1.60857716, + "memory(GiB)": 111.15, + "step": 61430, + "train_speed(iter/s)": 0.447241 + }, + { + "acc": 0.66258368, + "epoch": 1.5584728564180619, + "grad_norm": 6.21875, + "learning_rate": 1.2743007311787892e-06, + "loss": 1.64253368, + "memory(GiB)": 111.15, + "step": 61435, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.65929813, + "epoch": 1.5585996955859969, + "grad_norm": 5.5625, + "learning_rate": 1.2736014773496757e-06, + "loss": 1.57473793, + "memory(GiB)": 111.15, + "step": 61440, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.6619729, + "epoch": 1.5587265347539319, + "grad_norm": 5.6875, + "learning_rate": 1.2729023874242064e-06, + "loss": 1.55340233, + "memory(GiB)": 111.15, + "step": 61445, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.63537884, + "epoch": 1.558853373921867, + "grad_norm": 6.125, + "learning_rate": 1.2722034614331303e-06, + "loss": 1.73287506, + "memory(GiB)": 111.15, + "step": 61450, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.63352337, + "epoch": 1.5589802130898023, + "grad_norm": 5.46875, + "learning_rate": 1.2715046994071889e-06, + "loss": 1.68081436, + "memory(GiB)": 111.15, + "step": 61455, + "train_speed(iter/s)": 0.447338 + }, + { + "acc": 0.64626141, + "epoch": 1.5591070522577373, + "grad_norm": 5.46875, + "learning_rate": 1.2708061013771179e-06, + "loss": 1.62925262, + "memory(GiB)": 111.15, + "step": 61460, + "train_speed(iter/s)": 0.447357 + }, + { + "acc": 0.65293136, + "epoch": 1.5592338914256723, + "grad_norm": 4.5625, + "learning_rate": 1.2701076673736428e-06, + "loss": 1.56802216, + "memory(GiB)": 111.15, + "step": 61465, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.65000052, + "epoch": 1.5593607305936072, + "grad_norm": 5.5625, + "learning_rate": 1.269409397427488e-06, + "loss": 1.63789253, + "memory(GiB)": 111.15, + "step": 61470, + "train_speed(iter/s)": 0.447396 + }, + { + "acc": 0.6595891, + "epoch": 1.5594875697615422, + "grad_norm": 5.625, + "learning_rate": 1.2687112915693622e-06, + "loss": 1.52816992, + "memory(GiB)": 111.15, + "step": 61475, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.639323, + "epoch": 1.5596144089294774, + "grad_norm": 5.09375, + "learning_rate": 1.2680133498299729e-06, + "loss": 1.70628777, + "memory(GiB)": 111.15, + "step": 61480, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.64578795, + "epoch": 1.5597412480974124, + "grad_norm": 5.03125, + "learning_rate": 1.2673155722400177e-06, + "loss": 1.63455505, + "memory(GiB)": 111.15, + "step": 61485, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.64178705, + "epoch": 1.5598680872653476, + "grad_norm": 5.5625, + "learning_rate": 1.2666179588301908e-06, + "loss": 1.68312836, + "memory(GiB)": 111.15, + "step": 61490, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.6560174, + "epoch": 1.5599949264332826, + "grad_norm": 5.0625, + "learning_rate": 1.2659205096311738e-06, + "loss": 1.60937672, + "memory(GiB)": 111.15, + "step": 61495, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.64065237, + "epoch": 1.5601217656012176, + "grad_norm": 5.3125, + "learning_rate": 1.2652232246736423e-06, + "loss": 1.61399994, + "memory(GiB)": 111.15, + "step": 61500, + "train_speed(iter/s)": 0.447511 + }, + { + "epoch": 1.5601217656012176, + "eval_acc": 0.6433934068798756, + "eval_loss": 1.6075328588485718, + "eval_runtime": 113.7234, + "eval_samples_per_second": 56.013, + "eval_steps_per_second": 28.007, + "step": 61500 + }, + { + "acc": 0.65963111, + "epoch": 1.5602486047691526, + "grad_norm": 5.53125, + "learning_rate": 1.2645261039882694e-06, + "loss": 1.68677311, + "memory(GiB)": 111.15, + "step": 61505, + "train_speed(iter/s)": 0.447145 + }, + { + "acc": 0.64536543, + "epoch": 1.5603754439370878, + "grad_norm": 5.5625, + "learning_rate": 1.263829147605718e-06, + "loss": 1.65197525, + "memory(GiB)": 111.15, + "step": 61510, + "train_speed(iter/s)": 0.447164 + }, + { + "acc": 0.65322604, + "epoch": 1.5605022831050228, + "grad_norm": 5.59375, + "learning_rate": 1.26313235555664e-06, + "loss": 1.61498413, + "memory(GiB)": 111.15, + "step": 61515, + "train_speed(iter/s)": 0.447183 + }, + { + "acc": 0.66757555, + "epoch": 1.560629122272958, + "grad_norm": 5.5625, + "learning_rate": 1.2624357278716832e-06, + "loss": 1.60602112, + "memory(GiB)": 111.15, + "step": 61520, + "train_speed(iter/s)": 0.447202 + }, + { + "acc": 0.63282995, + "epoch": 1.560755961440893, + "grad_norm": 5.1875, + "learning_rate": 1.2617392645814913e-06, + "loss": 1.63219109, + "memory(GiB)": 111.15, + "step": 61525, + "train_speed(iter/s)": 0.447221 + }, + { + "acc": 0.64947481, + "epoch": 1.560882800608828, + "grad_norm": 5.0625, + "learning_rate": 1.2610429657166983e-06, + "loss": 1.68866959, + "memory(GiB)": 111.15, + "step": 61530, + "train_speed(iter/s)": 0.44724 + }, + { + "acc": 0.66347675, + "epoch": 1.561009639776763, + "grad_norm": 4.96875, + "learning_rate": 1.2603468313079265e-06, + "loss": 1.54928513, + "memory(GiB)": 111.15, + "step": 61535, + "train_speed(iter/s)": 0.44726 + }, + { + "acc": 0.65699701, + "epoch": 1.561136478944698, + "grad_norm": 5.03125, + "learning_rate": 1.2596508613857982e-06, + "loss": 1.56519012, + "memory(GiB)": 111.15, + "step": 61540, + "train_speed(iter/s)": 0.447279 + }, + { + "acc": 0.642169, + "epoch": 1.5612633181126332, + "grad_norm": 5.1875, + "learning_rate": 1.258955055980925e-06, + "loss": 1.63950119, + "memory(GiB)": 111.15, + "step": 61545, + "train_speed(iter/s)": 0.447298 + }, + { + "acc": 0.64145336, + "epoch": 1.5613901572805684, + "grad_norm": 4.90625, + "learning_rate": 1.258259415123911e-06, + "loss": 1.6263176, + "memory(GiB)": 111.15, + "step": 61550, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.6615572, + "epoch": 1.5615169964485034, + "grad_norm": 4.90625, + "learning_rate": 1.2575639388453532e-06, + "loss": 1.53849754, + "memory(GiB)": 111.15, + "step": 61555, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.64867358, + "epoch": 1.5616438356164384, + "grad_norm": 5.5, + "learning_rate": 1.2568686271758423e-06, + "loss": 1.60444965, + "memory(GiB)": 111.15, + "step": 61560, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.64127812, + "epoch": 1.5617706747843734, + "grad_norm": 4.9375, + "learning_rate": 1.2561734801459612e-06, + "loss": 1.65938988, + "memory(GiB)": 111.15, + "step": 61565, + "train_speed(iter/s)": 0.447376 + }, + { + "acc": 0.65793314, + "epoch": 1.5618975139523084, + "grad_norm": 5.5625, + "learning_rate": 1.2554784977862856e-06, + "loss": 1.60109558, + "memory(GiB)": 111.15, + "step": 61570, + "train_speed(iter/s)": 0.447395 + }, + { + "acc": 0.65105429, + "epoch": 1.5620243531202436, + "grad_norm": 5.4375, + "learning_rate": 1.2547836801273833e-06, + "loss": 1.62095833, + "memory(GiB)": 111.15, + "step": 61575, + "train_speed(iter/s)": 0.447415 + }, + { + "acc": 0.64547191, + "epoch": 1.5621511922881786, + "grad_norm": 5.8125, + "learning_rate": 1.2540890271998162e-06, + "loss": 1.63362293, + "memory(GiB)": 111.15, + "step": 61580, + "train_speed(iter/s)": 0.447434 + }, + { + "acc": 0.63778062, + "epoch": 1.5622780314561138, + "grad_norm": 4.625, + "learning_rate": 1.2533945390341379e-06, + "loss": 1.71499786, + "memory(GiB)": 111.15, + "step": 61585, + "train_speed(iter/s)": 0.447453 + }, + { + "acc": 0.64397125, + "epoch": 1.5624048706240488, + "grad_norm": 5.0625, + "learning_rate": 1.2527002156608946e-06, + "loss": 1.63156395, + "memory(GiB)": 111.15, + "step": 61590, + "train_speed(iter/s)": 0.447473 + }, + { + "acc": 0.65627794, + "epoch": 1.5625317097919837, + "grad_norm": 5.90625, + "learning_rate": 1.2520060571106275e-06, + "loss": 1.61884251, + "memory(GiB)": 111.15, + "step": 61595, + "train_speed(iter/s)": 0.447492 + }, + { + "acc": 0.639114, + "epoch": 1.5626585489599187, + "grad_norm": 5.40625, + "learning_rate": 1.2513120634138665e-06, + "loss": 1.68505249, + "memory(GiB)": 111.15, + "step": 61600, + "train_speed(iter/s)": 0.447512 + }, + { + "epoch": 1.5626585489599187, + "eval_acc": 0.6434510587685808, + "eval_loss": 1.6074637174606323, + "eval_runtime": 113.4336, + "eval_samples_per_second": 56.156, + "eval_steps_per_second": 28.078, + "step": 61600 + }, + { + "acc": 0.6551476, + "epoch": 1.5627853881278537, + "grad_norm": 5.625, + "learning_rate": 1.250618234601138e-06, + "loss": 1.55927391, + "memory(GiB)": 111.15, + "step": 61605, + "train_speed(iter/s)": 0.447147 + }, + { + "acc": 0.66035366, + "epoch": 1.562912227295789, + "grad_norm": 5.53125, + "learning_rate": 1.2499245707029595e-06, + "loss": 1.55254002, + "memory(GiB)": 111.15, + "step": 61610, + "train_speed(iter/s)": 0.447166 + }, + { + "acc": 0.64660215, + "epoch": 1.5630390664637241, + "grad_norm": 5.65625, + "learning_rate": 1.2492310717498412e-06, + "loss": 1.62770233, + "memory(GiB)": 111.15, + "step": 61615, + "train_speed(iter/s)": 0.447186 + }, + { + "acc": 0.66949577, + "epoch": 1.5631659056316591, + "grad_norm": 5.90625, + "learning_rate": 1.2485377377722863e-06, + "loss": 1.62398911, + "memory(GiB)": 111.15, + "step": 61620, + "train_speed(iter/s)": 0.447205 + }, + { + "acc": 0.66608438, + "epoch": 1.5632927447995941, + "grad_norm": 8.625, + "learning_rate": 1.2478445688007894e-06, + "loss": 1.60231686, + "memory(GiB)": 111.15, + "step": 61625, + "train_speed(iter/s)": 0.447224 + }, + { + "acc": 0.65691652, + "epoch": 1.5634195839675291, + "grad_norm": 5.53125, + "learning_rate": 1.2471515648658434e-06, + "loss": 1.58413467, + "memory(GiB)": 111.15, + "step": 61630, + "train_speed(iter/s)": 0.447244 + }, + { + "acc": 0.66044369, + "epoch": 1.563546423135464, + "grad_norm": 5.5, + "learning_rate": 1.2464587259979254e-06, + "loss": 1.59290543, + "memory(GiB)": 111.15, + "step": 61635, + "train_speed(iter/s)": 0.447263 + }, + { + "acc": 0.64670906, + "epoch": 1.5636732623033993, + "grad_norm": 7.625, + "learning_rate": 1.2457660522275095e-06, + "loss": 1.65377197, + "memory(GiB)": 111.15, + "step": 61640, + "train_speed(iter/s)": 0.447282 + }, + { + "acc": 0.65072718, + "epoch": 1.5638001014713343, + "grad_norm": 6.9375, + "learning_rate": 1.2450735435850654e-06, + "loss": 1.63144951, + "memory(GiB)": 111.15, + "step": 61645, + "train_speed(iter/s)": 0.447301 + }, + { + "acc": 0.64775829, + "epoch": 1.5639269406392695, + "grad_norm": 5.9375, + "learning_rate": 1.244381200101053e-06, + "loss": 1.64949837, + "memory(GiB)": 111.15, + "step": 61650, + "train_speed(iter/s)": 0.447321 + }, + { + "acc": 0.64952374, + "epoch": 1.5640537798072045, + "grad_norm": 5.53125, + "learning_rate": 1.2436890218059217e-06, + "loss": 1.67823315, + "memory(GiB)": 111.15, + "step": 61655, + "train_speed(iter/s)": 0.44734 + }, + { + "acc": 0.65062633, + "epoch": 1.5641806189751395, + "grad_norm": 5.0625, + "learning_rate": 1.2429970087301163e-06, + "loss": 1.60019112, + "memory(GiB)": 111.15, + "step": 61660, + "train_speed(iter/s)": 0.447359 + }, + { + "acc": 0.63922529, + "epoch": 1.5643074581430745, + "grad_norm": 5.5625, + "learning_rate": 1.2423051609040777e-06, + "loss": 1.67870598, + "memory(GiB)": 111.15, + "step": 61665, + "train_speed(iter/s)": 0.447379 + }, + { + "acc": 0.6406601, + "epoch": 1.5644342973110097, + "grad_norm": 5.46875, + "learning_rate": 1.2416134783582368e-06, + "loss": 1.69136963, + "memory(GiB)": 111.15, + "step": 61670, + "train_speed(iter/s)": 0.447398 + }, + { + "acc": 0.65114212, + "epoch": 1.5645611364789447, + "grad_norm": 5.09375, + "learning_rate": 1.2409219611230116e-06, + "loss": 1.63624611, + "memory(GiB)": 111.15, + "step": 61675, + "train_speed(iter/s)": 0.447417 + }, + { + "acc": 0.65818195, + "epoch": 1.56468797564688, + "grad_norm": 5.0625, + "learning_rate": 1.2402306092288236e-06, + "loss": 1.59679136, + "memory(GiB)": 111.15, + "step": 61680, + "train_speed(iter/s)": 0.447436 + }, + { + "acc": 0.66412153, + "epoch": 1.5648148148148149, + "grad_norm": 5.9375, + "learning_rate": 1.2395394227060793e-06, + "loss": 1.66023216, + "memory(GiB)": 111.15, + "step": 61685, + "train_speed(iter/s)": 0.447456 + }, + { + "acc": 0.64547167, + "epoch": 1.5649416539827499, + "grad_norm": 5.625, + "learning_rate": 1.238848401585182e-06, + "loss": 1.60134354, + "memory(GiB)": 111.15, + "step": 61690, + "train_speed(iter/s)": 0.447475 + }, + { + "acc": 0.6417614, + "epoch": 1.5650684931506849, + "grad_norm": 6.0625, + "learning_rate": 1.2381575458965218e-06, + "loss": 1.59035425, + "memory(GiB)": 111.15, + "step": 61695, + "train_speed(iter/s)": 0.447494 + }, + { + "acc": 0.65285177, + "epoch": 1.5651953323186198, + "grad_norm": 5.78125, + "learning_rate": 1.2374668556704888e-06, + "loss": 1.6354454, + "memory(GiB)": 111.15, + "step": 61700, + "train_speed(iter/s)": 0.447513 + }, + { + "epoch": 1.5651953323186198, + "eval_acc": 0.6434740359706009, + "eval_loss": 1.6074992418289185, + "eval_runtime": 114.5349, + "eval_samples_per_second": 55.616, + "eval_steps_per_second": 27.808, + "step": 61700 + }, + { + "acc": 0.65718641, + "epoch": 1.565322171486555, + "grad_norm": 6.125, + "learning_rate": 1.2367763309374625e-06, + "loss": 1.56777267, + "memory(GiB)": 111.15, + "step": 61705, + "train_speed(iter/s)": 0.447146 + }, + { + "acc": 0.64926414, + "epoch": 1.5654490106544903, + "grad_norm": 4.6875, + "learning_rate": 1.2360859717278145e-06, + "loss": 1.61757069, + "memory(GiB)": 111.15, + "step": 61710, + "train_speed(iter/s)": 0.447165 + }, + { + "acc": 0.6372757, + "epoch": 1.5655758498224253, + "grad_norm": 5.40625, + "learning_rate": 1.2353957780719106e-06, + "loss": 1.63773212, + "memory(GiB)": 111.15, + "step": 61715, + "train_speed(iter/s)": 0.447184 + }, + { + "acc": 0.63815513, + "epoch": 1.5657026889903602, + "grad_norm": 5.625, + "learning_rate": 1.2347057500001075e-06, + "loss": 1.7205162, + "memory(GiB)": 111.15, + "step": 61720, + "train_speed(iter/s)": 0.447203 + }, + { + "acc": 0.65624876, + "epoch": 1.5658295281582952, + "grad_norm": 5.0625, + "learning_rate": 1.2340158875427566e-06, + "loss": 1.60262699, + "memory(GiB)": 111.15, + "step": 61725, + "train_speed(iter/s)": 0.447222 + }, + { + "acc": 0.64388313, + "epoch": 1.5659563673262302, + "grad_norm": 5.59375, + "learning_rate": 1.2333261907302013e-06, + "loss": 1.61945648, + "memory(GiB)": 111.15, + "step": 61730, + "train_speed(iter/s)": 0.447242 + }, + { + "acc": 0.66546459, + "epoch": 1.5660832064941654, + "grad_norm": 4.9375, + "learning_rate": 1.2326366595927763e-06, + "loss": 1.59049377, + "memory(GiB)": 111.15, + "step": 61735, + "train_speed(iter/s)": 0.447261 + }, + { + "acc": 0.64993162, + "epoch": 1.5662100456621004, + "grad_norm": 4.875, + "learning_rate": 1.2319472941608118e-06, + "loss": 1.62627602, + "memory(GiB)": 111.15, + "step": 61740, + "train_speed(iter/s)": 0.44728 + }, + { + "acc": 0.65627518, + "epoch": 1.5663368848300356, + "grad_norm": 5.8125, + "learning_rate": 1.231258094464628e-06, + "loss": 1.6339304, + "memory(GiB)": 111.15, + "step": 61745, + "train_speed(iter/s)": 0.447299 + }, + { + "acc": 0.65322571, + "epoch": 1.5664637239979706, + "grad_norm": 6.125, + "learning_rate": 1.23056906053454e-06, + "loss": 1.6280899, + "memory(GiB)": 111.15, + "step": 61750, + "train_speed(iter/s)": 0.447318 + }, + { + "acc": 0.63951859, + "epoch": 1.5665905631659056, + "grad_norm": 5.46875, + "learning_rate": 1.2298801924008535e-06, + "loss": 1.67523937, + "memory(GiB)": 111.15, + "step": 61755, + "train_speed(iter/s)": 0.447337 + }, + { + "acc": 0.64100933, + "epoch": 1.5667174023338406, + "grad_norm": 4.875, + "learning_rate": 1.2291914900938685e-06, + "loss": 1.62723808, + "memory(GiB)": 111.15, + "step": 61760, + "train_speed(iter/s)": 0.447356 + }, + { + "acc": 0.64722714, + "epoch": 1.5668442415017756, + "grad_norm": 6.125, + "learning_rate": 1.2285029536438759e-06, + "loss": 1.63608131, + "memory(GiB)": 111.15, + "step": 61765, + "train_speed(iter/s)": 0.447375 + }, + { + "acc": 0.63908167, + "epoch": 1.5669710806697108, + "grad_norm": 4.28125, + "learning_rate": 1.227814583081165e-06, + "loss": 1.64320145, + "memory(GiB)": 111.15, + "step": 61770, + "train_speed(iter/s)": 0.447394 + }, + { + "acc": 0.65528545, + "epoch": 1.567097919837646, + "grad_norm": 7.25, + "learning_rate": 1.2271263784360088e-06, + "loss": 1.62080231, + "memory(GiB)": 111.15, + "step": 61775, + "train_speed(iter/s)": 0.447414 + }, + { + "acc": 0.65400348, + "epoch": 1.567224759005581, + "grad_norm": 5.40625, + "learning_rate": 1.2264383397386787e-06, + "loss": 1.66877098, + "memory(GiB)": 111.15, + "step": 61780, + "train_speed(iter/s)": 0.447433 + }, + { + "acc": 0.65651884, + "epoch": 1.567351598173516, + "grad_norm": 6.3125, + "learning_rate": 1.225750467019437e-06, + "loss": 1.6654871, + "memory(GiB)": 111.15, + "step": 61785, + "train_speed(iter/s)": 0.447452 + }, + { + "acc": 0.64565377, + "epoch": 1.567478437341451, + "grad_norm": 5.34375, + "learning_rate": 1.2250627603085435e-06, + "loss": 1.6973526, + "memory(GiB)": 111.15, + "step": 61790, + "train_speed(iter/s)": 0.44747 + }, + { + "acc": 0.65449662, + "epoch": 1.567605276509386, + "grad_norm": 6.53125, + "learning_rate": 1.2243752196362423e-06, + "loss": 1.63603992, + "memory(GiB)": 111.15, + "step": 61795, + "train_speed(iter/s)": 0.447489 + }, + { + "acc": 0.64652414, + "epoch": 1.5677321156773212, + "grad_norm": 5.125, + "learning_rate": 1.2236878450327743e-06, + "loss": 1.65817223, + "memory(GiB)": 111.15, + "step": 61800, + "train_speed(iter/s)": 0.447508 + }, + { + "epoch": 1.5677321156773212, + "eval_acc": 0.6434209795222998, + "eval_loss": 1.607554316520691, + "eval_runtime": 113.8144, + "eval_samples_per_second": 55.968, + "eval_steps_per_second": 27.984, + "step": 61800 + }, + { + "acc": 0.64618378, + "epoch": 1.5678589548452562, + "grad_norm": 6.21875, + "learning_rate": 1.223000636528377e-06, + "loss": 1.66398659, + "memory(GiB)": 111.15, + "step": 61805, + "train_speed(iter/s)": 0.447144 + }, + { + "acc": 0.66865191, + "epoch": 1.5679857940131914, + "grad_norm": 4.75, + "learning_rate": 1.2223135941532754e-06, + "loss": 1.54046526, + "memory(GiB)": 111.15, + "step": 61810, + "train_speed(iter/s)": 0.447163 + }, + { + "acc": 0.65558996, + "epoch": 1.5681126331811264, + "grad_norm": 6.0625, + "learning_rate": 1.2216267179376857e-06, + "loss": 1.61166077, + "memory(GiB)": 111.15, + "step": 61815, + "train_speed(iter/s)": 0.447182 + }, + { + "acc": 0.6555738, + "epoch": 1.5682394723490614, + "grad_norm": 5.65625, + "learning_rate": 1.2209400079118233e-06, + "loss": 1.63383522, + "memory(GiB)": 111.15, + "step": 61820, + "train_speed(iter/s)": 0.447201 + }, + { + "acc": 0.6585166, + "epoch": 1.5683663115169963, + "grad_norm": 4.5, + "learning_rate": 1.2202534641058916e-06, + "loss": 1.59268799, + "memory(GiB)": 111.15, + "step": 61825, + "train_speed(iter/s)": 0.44722 + }, + { + "acc": 0.67606874, + "epoch": 1.5684931506849316, + "grad_norm": 4.90625, + "learning_rate": 1.2195670865500896e-06, + "loss": 1.4574728, + "memory(GiB)": 111.15, + "step": 61830, + "train_speed(iter/s)": 0.447238 + }, + { + "acc": 0.65137987, + "epoch": 1.5686199898528665, + "grad_norm": 5.21875, + "learning_rate": 1.2188808752746022e-06, + "loss": 1.66996269, + "memory(GiB)": 111.15, + "step": 61835, + "train_speed(iter/s)": 0.447258 + }, + { + "acc": 0.66961575, + "epoch": 1.5687468290208018, + "grad_norm": 6.3125, + "learning_rate": 1.2181948303096176e-06, + "loss": 1.59006548, + "memory(GiB)": 111.15, + "step": 61840, + "train_speed(iter/s)": 0.447277 + }, + { + "acc": 0.65125914, + "epoch": 1.5688736681887367, + "grad_norm": 5.21875, + "learning_rate": 1.2175089516853083e-06, + "loss": 1.55854263, + "memory(GiB)": 111.15, + "step": 61845, + "train_speed(iter/s)": 0.447296 + }, + { + "acc": 0.66302423, + "epoch": 1.5690005073566717, + "grad_norm": 4.8125, + "learning_rate": 1.216823239431843e-06, + "loss": 1.58232918, + "memory(GiB)": 111.15, + "step": 61850, + "train_speed(iter/s)": 0.447315 + }, + { + "acc": 0.65608358, + "epoch": 1.5691273465246067, + "grad_norm": 6.8125, + "learning_rate": 1.2161376935793827e-06, + "loss": 1.59891691, + "memory(GiB)": 111.15, + "step": 61855, + "train_speed(iter/s)": 0.447334 + }, + { + "acc": 0.64724455, + "epoch": 1.5692541856925417, + "grad_norm": 5.71875, + "learning_rate": 1.21545231415808e-06, + "loss": 1.64509583, + "memory(GiB)": 111.15, + "step": 61860, + "train_speed(iter/s)": 0.447353 + }, + { + "acc": 0.63872747, + "epoch": 1.569381024860477, + "grad_norm": 5.40625, + "learning_rate": 1.2147671011980816e-06, + "loss": 1.68681488, + "memory(GiB)": 111.15, + "step": 61865, + "train_speed(iter/s)": 0.447372 + }, + { + "acc": 0.66064386, + "epoch": 1.5695078640284121, + "grad_norm": 5.59375, + "learning_rate": 1.2140820547295256e-06, + "loss": 1.57880478, + "memory(GiB)": 111.15, + "step": 61870, + "train_speed(iter/s)": 0.447391 + }, + { + "acc": 0.64495306, + "epoch": 1.5696347031963471, + "grad_norm": 5.0, + "learning_rate": 1.2133971747825435e-06, + "loss": 1.56886768, + "memory(GiB)": 111.15, + "step": 61875, + "train_speed(iter/s)": 0.44741 + }, + { + "acc": 0.65349288, + "epoch": 1.5697615423642821, + "grad_norm": 5.875, + "learning_rate": 1.2127124613872603e-06, + "loss": 1.64039383, + "memory(GiB)": 111.15, + "step": 61880, + "train_speed(iter/s)": 0.447429 + }, + { + "acc": 0.64967198, + "epoch": 1.569888381532217, + "grad_norm": 6.0625, + "learning_rate": 1.2120279145737918e-06, + "loss": 1.61907654, + "memory(GiB)": 111.15, + "step": 61885, + "train_speed(iter/s)": 0.447448 + }, + { + "acc": 0.66700048, + "epoch": 1.570015220700152, + "grad_norm": 5.28125, + "learning_rate": 1.2113435343722474e-06, + "loss": 1.55923624, + "memory(GiB)": 111.15, + "step": 61890, + "train_speed(iter/s)": 0.447467 + }, + { + "acc": 0.64505396, + "epoch": 1.5701420598680873, + "grad_norm": 5.03125, + "learning_rate": 1.21065932081273e-06, + "loss": 1.59934921, + "memory(GiB)": 111.15, + "step": 61895, + "train_speed(iter/s)": 0.447486 + }, + { + "acc": 0.65376916, + "epoch": 1.5702688990360223, + "grad_norm": 4.65625, + "learning_rate": 1.2099752739253334e-06, + "loss": 1.62111588, + "memory(GiB)": 111.15, + "step": 61900, + "train_speed(iter/s)": 0.447505 + }, + { + "epoch": 1.5702688990360223, + "eval_acc": 0.6434289171011794, + "eval_loss": 1.6074942350387573, + "eval_runtime": 113.2304, + "eval_samples_per_second": 56.257, + "eval_steps_per_second": 28.128, + "step": 61900 + }, + { + "acc": 0.66481819, + "epoch": 1.5703957382039575, + "grad_norm": 5.625, + "learning_rate": 1.209291393740144e-06, + "loss": 1.61883698, + "memory(GiB)": 111.15, + "step": 61905, + "train_speed(iter/s)": 0.447143 + }, + { + "acc": 0.64166856, + "epoch": 1.5705225773718925, + "grad_norm": 5.5625, + "learning_rate": 1.2086076802872472e-06, + "loss": 1.64801559, + "memory(GiB)": 111.15, + "step": 61910, + "train_speed(iter/s)": 0.447162 + }, + { + "acc": 0.64591441, + "epoch": 1.5706494165398275, + "grad_norm": 5.1875, + "learning_rate": 1.2079241335967096e-06, + "loss": 1.68066788, + "memory(GiB)": 111.15, + "step": 61915, + "train_speed(iter/s)": 0.447181 + }, + { + "acc": 0.64614167, + "epoch": 1.5707762557077625, + "grad_norm": 5.4375, + "learning_rate": 1.207240753698599e-06, + "loss": 1.63730164, + "memory(GiB)": 111.15, + "step": 61920, + "train_speed(iter/s)": 0.4472 + }, + { + "acc": 0.68177509, + "epoch": 1.5709030948756975, + "grad_norm": 6.1875, + "learning_rate": 1.2065575406229723e-06, + "loss": 1.48169909, + "memory(GiB)": 111.15, + "step": 61925, + "train_speed(iter/s)": 0.447219 + }, + { + "acc": 0.66844225, + "epoch": 1.5710299340436327, + "grad_norm": 5.375, + "learning_rate": 1.2058744943998847e-06, + "loss": 1.57001219, + "memory(GiB)": 111.15, + "step": 61930, + "train_speed(iter/s)": 0.447238 + }, + { + "acc": 0.65815463, + "epoch": 1.5711567732115679, + "grad_norm": 6.5625, + "learning_rate": 1.2051916150593746e-06, + "loss": 1.63338013, + "memory(GiB)": 111.15, + "step": 61935, + "train_speed(iter/s)": 0.447257 + }, + { + "acc": 0.64229174, + "epoch": 1.5712836123795029, + "grad_norm": 5.6875, + "learning_rate": 1.2045089026314783e-06, + "loss": 1.62774906, + "memory(GiB)": 111.15, + "step": 61940, + "train_speed(iter/s)": 0.447276 + }, + { + "acc": 0.64065957, + "epoch": 1.5714104515474379, + "grad_norm": 5.28125, + "learning_rate": 1.2038263571462278e-06, + "loss": 1.61140385, + "memory(GiB)": 111.15, + "step": 61945, + "train_speed(iter/s)": 0.447295 + }, + { + "acc": 0.64558754, + "epoch": 1.5715372907153728, + "grad_norm": 7.71875, + "learning_rate": 1.203143978633644e-06, + "loss": 1.64823112, + "memory(GiB)": 111.15, + "step": 61950, + "train_speed(iter/s)": 0.447314 + }, + { + "acc": 0.64385748, + "epoch": 1.5716641298833078, + "grad_norm": 6.03125, + "learning_rate": 1.2024617671237388e-06, + "loss": 1.59452457, + "memory(GiB)": 111.15, + "step": 61955, + "train_speed(iter/s)": 0.447333 + }, + { + "acc": 0.63561811, + "epoch": 1.571790969051243, + "grad_norm": 3.9375, + "learning_rate": 1.2017797226465178e-06, + "loss": 1.63195915, + "memory(GiB)": 111.15, + "step": 61960, + "train_speed(iter/s)": 0.447352 + }, + { + "acc": 0.6284585, + "epoch": 1.571917808219178, + "grad_norm": 4.84375, + "learning_rate": 1.2010978452319843e-06, + "loss": 1.66161652, + "memory(GiB)": 111.15, + "step": 61965, + "train_speed(iter/s)": 0.447371 + }, + { + "acc": 0.64103985, + "epoch": 1.5720446473871132, + "grad_norm": 5.03125, + "learning_rate": 1.2004161349101295e-06, + "loss": 1.62771263, + "memory(GiB)": 111.15, + "step": 61970, + "train_speed(iter/s)": 0.44739 + }, + { + "acc": 0.66199942, + "epoch": 1.5721714865550482, + "grad_norm": 7.0, + "learning_rate": 1.1997345917109348e-06, + "loss": 1.61285133, + "memory(GiB)": 111.15, + "step": 61975, + "train_speed(iter/s)": 0.447409 + }, + { + "acc": 0.65300317, + "epoch": 1.5722983257229832, + "grad_norm": 5.15625, + "learning_rate": 1.1990532156643808e-06, + "loss": 1.64526367, + "memory(GiB)": 111.15, + "step": 61980, + "train_speed(iter/s)": 0.447428 + }, + { + "acc": 0.65439763, + "epoch": 1.5724251648909182, + "grad_norm": 5.375, + "learning_rate": 1.198372006800436e-06, + "loss": 1.62791061, + "memory(GiB)": 111.15, + "step": 61985, + "train_speed(iter/s)": 0.447447 + }, + { + "acc": 0.63873773, + "epoch": 1.5725520040588534, + "grad_norm": 5.34375, + "learning_rate": 1.1976909651490637e-06, + "loss": 1.64584446, + "memory(GiB)": 111.15, + "step": 61990, + "train_speed(iter/s)": 0.447466 + }, + { + "acc": 0.66477251, + "epoch": 1.5726788432267884, + "grad_norm": 4.28125, + "learning_rate": 1.1970100907402188e-06, + "loss": 1.59328194, + "memory(GiB)": 111.15, + "step": 61995, + "train_speed(iter/s)": 0.447485 + }, + { + "acc": 0.65711193, + "epoch": 1.5728056823947236, + "grad_norm": 6.875, + "learning_rate": 1.1963293836038492e-06, + "loss": 1.63846951, + "memory(GiB)": 111.15, + "step": 62000, + "train_speed(iter/s)": 0.447504 + }, + { + "epoch": 1.5728056823947236, + "eval_acc": 0.6434197262203715, + "eval_loss": 1.6074186563491821, + "eval_runtime": 113.09, + "eval_samples_per_second": 56.327, + "eval_steps_per_second": 28.163, + "step": 62000 + } + ], + "logging_steps": 5, + "max_steps": 78840, + "num_input_tokens_seen": 0, + "num_train_epochs": 2, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 1.1594429930783375e+18, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +}