{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9999074845036544,
  "eval_steps": 100,
  "global_step": 1351,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.003700619853825516,
      "grad_norm": 6.406365314676644,
      "learning_rate": 7.352941176470589e-07,
      "loss": 1.3566,
      "mean_token_accuracy": 0.6660232350230217,
      "step": 5
    },
    {
      "epoch": 0.007401239707651032,
      "grad_norm": 6.123712399933804,
      "learning_rate": 1.4705882352941177e-06,
      "loss": 1.3913,
      "mean_token_accuracy": 0.6558953940868377,
      "step": 10
    },
    {
      "epoch": 0.011101859561476548,
      "grad_norm": 5.275448607027431,
      "learning_rate": 2.2058823529411767e-06,
      "loss": 1.3404,
      "mean_token_accuracy": 0.6668361470103263,
      "step": 15
    },
    {
      "epoch": 0.014802479415302064,
      "grad_norm": 3.222168252214971,
      "learning_rate": 2.9411764705882355e-06,
      "loss": 1.2979,
      "mean_token_accuracy": 0.6731562301516533,
      "step": 20
    },
    {
      "epoch": 0.01850309926912758,
      "grad_norm": 2.304245354229246,
      "learning_rate": 3.6764705882352946e-06,
      "loss": 1.2317,
      "mean_token_accuracy": 0.6820817664265633,
      "step": 25
    },
    {
      "epoch": 0.022203719122953096,
      "grad_norm": 2.2635687771667734,
      "learning_rate": 4.411764705882353e-06,
      "loss": 1.19,
      "mean_token_accuracy": 0.6874127045273781,
      "step": 30
    },
    {
      "epoch": 0.025904338976778612,
      "grad_norm": 1.6106011837583105,
      "learning_rate": 5.147058823529411e-06,
      "loss": 1.1346,
      "mean_token_accuracy": 0.6967306867241859,
      "step": 35
    },
    {
      "epoch": 0.029604958830604128,
      "grad_norm": 1.2780962221463628,
      "learning_rate": 5.882352941176471e-06,
      "loss": 1.0694,
      "mean_token_accuracy": 0.7095885664224625,
      "step": 40
    },
    {
      "epoch": 0.03330557868442964,
      "grad_norm": 1.637550942271707,
      "learning_rate": 6.61764705882353e-06,
      "loss": 1.056,
      "mean_token_accuracy": 0.7120530039072037,
      "step": 45
    },
    {
      "epoch": 0.03700619853825516,
      "grad_norm": 1.2301760518659035,
      "learning_rate": 7.352941176470589e-06,
      "loss": 1.03,
      "mean_token_accuracy": 0.7162090480327606,
      "step": 50
    },
    {
      "epoch": 0.04070681839208067,
      "grad_norm": 0.9170716667796568,
      "learning_rate": 8.088235294117648e-06,
      "loss": 1.0283,
      "mean_token_accuracy": 0.7158895432949066,
      "step": 55
    },
    {
      "epoch": 0.04440743824590619,
      "grad_norm": 1.0543932075765083,
      "learning_rate": 8.823529411764707e-06,
      "loss": 0.9877,
      "mean_token_accuracy": 0.7251855596899986,
      "step": 60
    },
    {
      "epoch": 0.048108058099731704,
      "grad_norm": 0.9843413417015394,
      "learning_rate": 9.558823529411766e-06,
      "loss": 0.9974,
      "mean_token_accuracy": 0.7208766683936119,
      "step": 65
    },
    {
      "epoch": 0.051808677953557224,
      "grad_norm": 0.9511465656581665,
      "learning_rate": 1.0294117647058823e-05,
      "loss": 0.9924,
      "mean_token_accuracy": 0.7231188163161277,
      "step": 70
    },
    {
      "epoch": 0.055509297807382736,
      "grad_norm": 2.825049696114851,
      "learning_rate": 1.1029411764705885e-05,
      "loss": 0.9782,
      "mean_token_accuracy": 0.7259459629654884,
      "step": 75
    },
    {
      "epoch": 0.059209917661208256,
      "grad_norm": 0.938362593121529,
      "learning_rate": 1.1764705882352942e-05,
      "loss": 0.9686,
      "mean_token_accuracy": 0.7275517180562019,
      "step": 80
    },
    {
      "epoch": 0.06291053751503377,
      "grad_norm": 0.9202800308554298,
      "learning_rate": 1.25e-05,
      "loss": 0.9794,
      "mean_token_accuracy": 0.7241677865386009,
      "step": 85
    },
    {
      "epoch": 0.06661115736885928,
      "grad_norm": 0.9332127502056355,
      "learning_rate": 1.323529411764706e-05,
      "loss": 0.9497,
      "mean_token_accuracy": 0.731817427277565,
      "step": 90
    },
    {
      "epoch": 0.0703117772226848,
      "grad_norm": 1.0837119699847717,
      "learning_rate": 1.3970588235294118e-05,
      "loss": 0.9852,
      "mean_token_accuracy": 0.7222246199846267,
      "step": 95
    },
    {
      "epoch": 0.07401239707651032,
      "grad_norm": 0.8948953129317817,
      "learning_rate": 1.4705882352941179e-05,
      "loss": 0.9463,
      "mean_token_accuracy": 0.7313005596399307,
      "step": 100
    },
    {
      "epoch": 0.07401239707651032,
      "eval_loss": 0.9788710474967957,
      "eval_mean_token_accuracy": 0.7260306134368434,
      "eval_runtime": 24.5669,
      "eval_samples_per_second": 20.882,
      "eval_steps_per_second": 1.343,
      "step": 100
    },
    {
      "epoch": 0.07771301693033583,
      "grad_norm": 0.9653033723452836,
      "learning_rate": 1.5441176470588237e-05,
      "loss": 0.9696,
      "mean_token_accuracy": 0.7255482524633408,
      "step": 105
    },
    {
      "epoch": 0.08141363678416134,
      "grad_norm": 0.9349548842470216,
      "learning_rate": 1.6176470588235296e-05,
      "loss": 0.9722,
      "mean_token_accuracy": 0.7248646602034569,
      "step": 110
    },
    {
      "epoch": 0.08511425663798686,
      "grad_norm": 0.8956374335067367,
      "learning_rate": 1.6911764705882355e-05,
      "loss": 0.9591,
      "mean_token_accuracy": 0.7272942900657654,
      "step": 115
    },
    {
      "epoch": 0.08881487649181238,
      "grad_norm": 0.9788467585120308,
      "learning_rate": 1.7647058823529414e-05,
      "loss": 0.9363,
      "mean_token_accuracy": 0.7335584476590157,
      "step": 120
    },
    {
      "epoch": 0.0925154963456379,
      "grad_norm": 0.9134626099519884,
      "learning_rate": 1.8382352941176472e-05,
      "loss": 0.9276,
      "mean_token_accuracy": 0.7360474243760109,
      "step": 125
    },
    {
      "epoch": 0.09621611619946341,
      "grad_norm": 0.9753561569488689,
      "learning_rate": 1.911764705882353e-05,
      "loss": 0.9263,
      "mean_token_accuracy": 0.7367036134004593,
      "step": 130
    },
    {
      "epoch": 0.09991673605328892,
      "grad_norm": 0.9447239853385726,
      "learning_rate": 1.985294117647059e-05,
      "loss": 0.9358,
      "mean_token_accuracy": 0.7343964487314224,
      "step": 135
    },
    {
      "epoch": 0.10361735590711445,
      "grad_norm": 0.9093049891627352,
      "learning_rate": 1.9999465148392906e-05,
      "loss": 0.9199,
      "mean_token_accuracy": 0.7370649874210358,
      "step": 140
    },
    {
      "epoch": 0.10731797576093996,
      "grad_norm": 0.9478337858776447,
      "learning_rate": 1.999729241179462e-05,
      "loss": 0.9378,
      "mean_token_accuracy": 0.7336751863360405,
      "step": 145
    },
    {
      "epoch": 0.11101859561476547,
      "grad_norm": 0.8569290211164649,
      "learning_rate": 1.999344872485215e-05,
      "loss": 0.9038,
      "mean_token_accuracy": 0.7422620430588722,
      "step": 150
    },
    {
      "epoch": 0.11471921546859098,
      "grad_norm": 1.036972518186118,
      "learning_rate": 1.9987934730000457e-05,
      "loss": 0.9031,
      "mean_token_accuracy": 0.741059948503971,
      "step": 155
    },
    {
      "epoch": 0.11841983532241651,
      "grad_norm": 0.951663289147615,
      "learning_rate": 1.998075134885022e-05,
      "loss": 0.9171,
      "mean_token_accuracy": 0.7384189561009407,
      "step": 160
    },
    {
      "epoch": 0.12212045517624202,
      "grad_norm": 0.8919119572868265,
      "learning_rate": 1.9971899782033853e-05,
      "loss": 0.9378,
      "mean_token_accuracy": 0.7339910715818405,
      "step": 165
    },
    {
      "epoch": 0.12582107503006754,
      "grad_norm": 0.9627820867919544,
      "learning_rate": 1.9961381509004785e-05,
      "loss": 0.9548,
      "mean_token_accuracy": 0.7276575401425361,
      "step": 170
    },
    {
      "epoch": 0.12952169488389306,
      "grad_norm": 1.0867703495663577,
      "learning_rate": 1.9949198287790215e-05,
      "loss": 0.8957,
      "mean_token_accuracy": 0.7430507227778435,
      "step": 175
    },
    {
      "epoch": 0.13322231473771856,
      "grad_norm": 0.9434565750923781,
      "learning_rate": 1.9935352154697257e-05,
      "loss": 0.9326,
      "mean_token_accuracy": 0.7323492169380188,
      "step": 180
    },
    {
      "epoch": 0.1369229345915441,
      "grad_norm": 0.9338703985952416,
      "learning_rate": 1.9919845423972603e-05,
      "loss": 0.9152,
      "mean_token_accuracy": 0.7383304908871651,
      "step": 185
    },
    {
      "epoch": 0.1406235544453696,
      "grad_norm": 0.8449751394948155,
      "learning_rate": 1.9902680687415704e-05,
      "loss": 0.9115,
      "mean_token_accuracy": 0.7386730879545211,
      "step": 190
    },
    {
      "epoch": 0.1443241742991951,
      "grad_norm": 0.9634477921697367,
      "learning_rate": 1.9883860813945596e-05,
      "loss": 0.9472,
      "mean_token_accuracy": 0.7278202414512634,
      "step": 195
    },
    {
      "epoch": 0.14802479415302064,
      "grad_norm": 0.8797771752554515,
      "learning_rate": 1.986338894912137e-05,
      "loss": 0.9133,
      "mean_token_accuracy": 0.7370981857180595,
      "step": 200
    },
    {
      "epoch": 0.14802479415302064,
      "eval_loss": 0.9444138407707214,
      "eval_mean_token_accuracy": 0.7318792343139648,
      "eval_runtime": 24.471,
      "eval_samples_per_second": 20.964,
      "eval_steps_per_second": 1.349,
      "step": 200
    },
    {
      "epoch": 0.15172541400684614,
      "grad_norm": 0.9730130557080766,
      "learning_rate": 1.9841268514616434e-05,
      "loss": 0.9072,
      "mean_token_accuracy": 0.7391636997461319,
      "step": 205
    },
    {
      "epoch": 0.15542603386067166,
      "grad_norm": 0.9307244607734099,
      "learning_rate": 1.9817503207646606e-05,
      "loss": 0.9083,
      "mean_token_accuracy": 0.7380757689476013,
      "step": 210
    },
    {
      "epoch": 0.1591266537144972,
      "grad_norm": 0.8802319735889644,
      "learning_rate": 1.979209700035216e-05,
      "loss": 0.9354,
      "mean_token_accuracy": 0.7329269647598267,
      "step": 215
    },
    {
      "epoch": 0.1628272735683227,
      "grad_norm": 0.8725363780112383,
      "learning_rate": 1.976505413913393e-05,
      "loss": 0.9128,
      "mean_token_accuracy": 0.7383672297000885,
      "step": 220
    },
    {
      "epoch": 0.16652789342214822,
      "grad_norm": 0.903336658421562,
      "learning_rate": 1.9736379143943565e-05,
      "loss": 0.9385,
      "mean_token_accuracy": 0.7308298230171204,
      "step": 225
    },
    {
      "epoch": 0.17022851327597371,
      "grad_norm": 0.8621415714119711,
      "learning_rate": 1.9706076807528044e-05,
      "loss": 0.9239,
      "mean_token_accuracy": 0.7339740186929703,
      "step": 230
    },
    {
      "epoch": 0.17392913312979924,
      "grad_norm": 0.9079088453730583,
      "learning_rate": 1.967415219462864e-05,
      "loss": 0.9217,
      "mean_token_accuracy": 0.7344419181346893,
      "step": 235
    },
    {
      "epoch": 0.17762975298362477,
      "grad_norm": 0.8253363960767214,
      "learning_rate": 1.9640610641134383e-05,
      "loss": 0.9103,
      "mean_token_accuracy": 0.7382244557142258,
      "step": 240
    },
    {
      "epoch": 0.18133037283745027,
      "grad_norm": 0.934084069747635,
      "learning_rate": 1.9605457753190224e-05,
      "loss": 0.9152,
      "mean_token_accuracy": 0.7362594842910767,
      "step": 245
    },
    {
      "epoch": 0.1850309926912758,
      "grad_norm": 0.8995132563871583,
      "learning_rate": 1.9568699406260016e-05,
      "loss": 0.931,
      "mean_token_accuracy": 0.7331584557890892,
      "step": 250
    },
    {
      "epoch": 0.18873161254510132,
      "grad_norm": 0.8420380152677727,
      "learning_rate": 1.953034174414449e-05,
      "loss": 0.8976,
      "mean_token_accuracy": 0.7409126415848732,
      "step": 255
    },
    {
      "epoch": 0.19243223239892682,
      "grad_norm": 0.8794392938382097,
      "learning_rate": 1.9490391177954383e-05,
      "loss": 0.8971,
      "mean_token_accuracy": 0.7426333785057068,
      "step": 260
    },
    {
      "epoch": 0.19613285225275234,
      "grad_norm": 0.871108034639782,
      "learning_rate": 1.944885438503888e-05,
      "loss": 0.9153,
      "mean_token_accuracy": 0.7366316050291062,
      "step": 265
    },
    {
      "epoch": 0.19983347210657784,
      "grad_norm": 0.8826333710088322,
      "learning_rate": 1.9405738307869565e-05,
      "loss": 0.9093,
      "mean_token_accuracy": 0.7374404519796371,
      "step": 270
    },
    {
      "epoch": 0.20353409196040337,
      "grad_norm": 0.9186513628105667,
      "learning_rate": 1.936105015288003e-05,
      "loss": 0.9206,
      "mean_token_accuracy": 0.7328536227345467,
      "step": 275
    },
    {
      "epoch": 0.2072347118142289,
      "grad_norm": 0.892096099101671,
      "learning_rate": 1.9314797389261426e-05,
      "loss": 0.9193,
      "mean_token_accuracy": 0.7348861888051033,
      "step": 280
    },
    {
      "epoch": 0.2109353316680544,
      "grad_norm": 0.8240507326046963,
      "learning_rate": 1.9266987747714036e-05,
      "loss": 0.897,
      "mean_token_accuracy": 0.7411905080080032,
      "step": 285
    },
    {
      "epoch": 0.21463595152187992,
      "grad_norm": 0.867466389731865,
      "learning_rate": 1.9217629219155172e-05,
      "loss": 0.9087,
      "mean_token_accuracy": 0.7377493485808373,
      "step": 290
    },
    {
      "epoch": 0.21833657137570542,
      "grad_norm": 0.8533932703930674,
      "learning_rate": 1.916673005338357e-05,
      "loss": 0.8899,
      "mean_token_accuracy": 0.7420720800757408,
      "step": 295
    },
    {
      "epoch": 0.22203719122953094,
      "grad_norm": 0.8385314803046641,
      "learning_rate": 1.9114298757700508e-05,
      "loss": 0.9295,
      "mean_token_accuracy": 0.7315363213419914,
      "step": 300
    },
    {
      "epoch": 0.22203719122953094,
      "eval_loss": 0.9261447191238403,
      "eval_mean_token_accuracy": 0.7356406341899525,
      "eval_runtime": 24.4594,
      "eval_samples_per_second": 20.974,
      "eval_steps_per_second": 1.349,
      "step": 300
    },
    {
      "epoch": 0.22573781108335647,
      "grad_norm": 0.8344167353514693,
      "learning_rate": 1.9060344095487916e-05,
      "loss": 0.9021,
      "mean_token_accuracy": 0.7388954728841781,
      "step": 305
    },
    {
      "epoch": 0.22943843093718197,
      "grad_norm": 0.8166938367608658,
      "learning_rate": 1.9004875084743624e-05,
      "loss": 0.8745,
      "mean_token_accuracy": 0.7464573740959167,
      "step": 310
    },
    {
      "epoch": 0.2331390507910075,
      "grad_norm": 0.8583610850262193,
      "learning_rate": 1.8947900996574133e-05,
      "loss": 0.8906,
      "mean_token_accuracy": 0.7421576008200645,
      "step": 315
    },
    {
      "epoch": 0.23683967064483302,
      "grad_norm": 0.85619616919622,
      "learning_rate": 1.8889431353645004e-05,
      "loss": 0.9083,
      "mean_token_accuracy": 0.7389416947960854,
      "step": 320
    },
    {
      "epoch": 0.24054029049865852,
      "grad_norm": 0.95321268460484,
      "learning_rate": 1.8829475928589272e-05,
      "loss": 0.9037,
      "mean_token_accuracy": 0.7408745899796486,
      "step": 325
    },
    {
      "epoch": 0.24424091035248405,
      "grad_norm": 0.9192097687099984,
      "learning_rate": 1.8768044742374008e-05,
      "loss": 0.8844,
      "mean_token_accuracy": 0.7437578573822975,
      "step": 330
    },
    {
      "epoch": 0.24794153020630955,
      "grad_norm": 0.8224814882376746,
      "learning_rate": 1.870514806262544e-05,
      "loss": 0.8793,
      "mean_token_accuracy": 0.7451973140239716,
      "step": 335
    },
    {
      "epoch": 0.2516421500601351,
      "grad_norm": 0.8277417287111005,
      "learning_rate": 1.8640796401912805e-05,
      "loss": 0.8937,
      "mean_token_accuracy": 0.7406307712197304,
      "step": 340
    },
    {
      "epoch": 0.2553427699139606,
      "grad_norm": 0.8459764358981483,
      "learning_rate": 1.8575000515991283e-05,
      "loss": 0.8979,
      "mean_token_accuracy": 0.7395652890205383,
      "step": 345
    },
    {
      "epoch": 0.2590433897677861,
      "grad_norm": 0.8033711981121063,
      "learning_rate": 1.850777140200427e-05,
      "loss": 0.8813,
      "mean_token_accuracy": 0.7449199944734574,
      "step": 350
    },
    {
      "epoch": 0.2627440096216116,
      "grad_norm": 0.7910864449390781,
      "learning_rate": 1.843912029664531e-05,
      "loss": 0.8965,
      "mean_token_accuracy": 0.7398232161998749,
      "step": 355
    },
    {
      "epoch": 0.2664446294754371,
      "grad_norm": 0.8451406694855826,
      "learning_rate": 1.8369058674280004e-05,
      "loss": 0.8943,
      "mean_token_accuracy": 0.7411910384893418,
      "step": 360
    },
    {
      "epoch": 0.27014524932926265,
      "grad_norm": 0.8333682523549819,
      "learning_rate": 1.8297598245028173e-05,
      "loss": 0.9132,
      "mean_token_accuracy": 0.7367936804890632,
      "step": 365
    },
    {
      "epoch": 0.2738458691830882,
      "grad_norm": 0.8175542933467003,
      "learning_rate": 1.8224750952806626e-05,
      "loss": 0.8769,
      "mean_token_accuracy": 0.7455912932753563,
      "step": 370
    },
    {
      "epoch": 0.2775464890369137,
      "grad_norm": 0.8434576136182329,
      "learning_rate": 1.815052897333284e-05,
      "loss": 0.8808,
      "mean_token_accuracy": 0.7459388568997383,
      "step": 375
    },
    {
      "epoch": 0.2812471088907392,
      "grad_norm": 0.8936045983724704,
      "learning_rate": 1.8074944712089925e-05,
      "loss": 0.9024,
      "mean_token_accuracy": 0.7386106207966805,
      "step": 380
    },
    {
      "epoch": 0.2849477287445647,
      "grad_norm": 0.8151663680460399,
      "learning_rate": 1.799801080225316e-05,
      "loss": 0.8912,
      "mean_token_accuracy": 0.7416763469576836,
      "step": 385
    },
    {
      "epoch": 0.2886483485983902,
      "grad_norm": 0.8482431323316972,
      "learning_rate": 1.7919740102578482e-05,
      "loss": 0.8887,
      "mean_token_accuracy": 0.7428235754370689,
      "step": 390
    },
    {
      "epoch": 0.29234896845221575,
      "grad_norm": 0.8424168395876551,
      "learning_rate": 1.7840145695253258e-05,
      "loss": 0.8769,
      "mean_token_accuracy": 0.744986218214035,
      "step": 395
    },
    {
      "epoch": 0.2960495883060413,
      "grad_norm": 0.7880509490552058,
      "learning_rate": 1.7759240883709745e-05,
      "loss": 0.8706,
      "mean_token_accuracy": 0.74755879342556,
      "step": 400
    },
    {
      "epoch": 0.2960495883060413,
      "eval_loss": 0.9140655398368835,
      "eval_mean_token_accuracy": 0.7376509543621179,
      "eval_runtime": 24.4423,
      "eval_samples_per_second": 20.988,
      "eval_steps_per_second": 1.35,
      "step": 400
    },
    {
      "epoch": 0.2997502081598668,
      "grad_norm": 0.9325819637971632,
      "learning_rate": 1.7677039190401538e-05,
      "loss": 0.8778,
      "mean_token_accuracy": 0.7454720690846444,
      "step": 405
    },
    {
      "epoch": 0.3034508280136923,
      "grad_norm": 0.855702135031522,
      "learning_rate": 1.759355435454342e-05,
      "loss": 0.8825,
      "mean_token_accuracy": 0.7436973512172699,
      "step": 410
    },
    {
      "epoch": 0.3071514478675178,
      "grad_norm": 0.8190150096476719,
      "learning_rate": 1.7508800329814993e-05,
      "loss": 0.8801,
      "mean_token_accuracy": 0.7447807624936104,
      "step": 415
    },
    {
      "epoch": 0.31085206772134333,
      "grad_norm": 0.7879901787308187,
      "learning_rate": 1.7422791282028457e-05,
      "loss": 0.8926,
      "mean_token_accuracy": 0.7424971207976341,
      "step": 420
    },
    {
      "epoch": 0.31455268757516885,
      "grad_norm": 0.8418635735026896,
      "learning_rate": 1.7335541586760928e-05,
      "loss": 0.8962,
      "mean_token_accuracy": 0.740935817360878,
      "step": 425
    },
    {
      "epoch": 0.3182533074289944,
      "grad_norm": 0.8832514235130526,
      "learning_rate": 1.7247065826951694e-05,
      "loss": 0.8843,
      "mean_token_accuracy": 0.7440172478556633,
      "step": 430
    },
    {
      "epoch": 0.32195392728281985,
      "grad_norm": 0.9664717868103249,
      "learning_rate": 1.715737879046483e-05,
      "loss": 0.8955,
      "mean_token_accuracy": 0.7396286860108375,
      "step": 435
    },
    {
      "epoch": 0.3256545471366454,
      "grad_norm": 0.8907608106819015,
      "learning_rate": 1.7066495467617552e-05,
      "loss": 0.891,
      "mean_token_accuracy": 0.7419100046157837,
      "step": 440
    },
    {
      "epoch": 0.3293551669904709,
      "grad_norm": 0.7921927517011862,
      "learning_rate": 1.6974431048674714e-05,
      "loss": 0.8823,
      "mean_token_accuracy": 0.7444690898060798,
      "step": 445
    },
    {
      "epoch": 0.33305578684429643,
      "grad_norm": 0.8845253263563585,
      "learning_rate": 1.6881200921309914e-05,
      "loss": 0.8908,
      "mean_token_accuracy": 0.7414721488952637,
      "step": 450
    },
    {
      "epoch": 0.33675640669812196,
      "grad_norm": 0.9038197553584879,
      "learning_rate": 1.6786820668033596e-05,
      "loss": 0.8902,
      "mean_token_accuracy": 0.7424716472625732,
      "step": 455
    },
    {
      "epoch": 0.34045702655194743,
      "grad_norm": 0.845478277089973,
      "learning_rate": 1.6691306063588583e-05,
      "loss": 0.8862,
      "mean_token_accuracy": 0.7440178290009498,
      "step": 460
    },
    {
      "epoch": 0.34415764640577295,
      "grad_norm": 0.7780737971349384,
      "learning_rate": 1.6594673072313478e-05,
      "loss": 0.8683,
      "mean_token_accuracy": 0.7474384486675263,
      "step": 465
    },
    {
      "epoch": 0.3478582662595985,
      "grad_norm": 0.7883315377481923,
      "learning_rate": 1.6496937845474375e-05,
      "loss": 0.852,
      "mean_token_accuracy": 0.7510827273130417,
      "step": 470
    },
    {
      "epoch": 0.351558886113424,
      "grad_norm": 0.8522289082869288,
      "learning_rate": 1.639811671856535e-05,
      "loss": 0.8999,
      "mean_token_accuracy": 0.7387788712978363,
      "step": 475
    },
    {
      "epoch": 0.35525950596724953,
      "grad_norm": 0.8084630790546208,
      "learning_rate": 1.6298226208578127e-05,
      "loss": 0.8818,
      "mean_token_accuracy": 0.7436941042542458,
      "step": 480
    },
    {
      "epoch": 0.358960125821075,
      "grad_norm": 0.9229148133761365,
      "learning_rate": 1.6197283011241423e-05,
      "loss": 0.8909,
      "mean_token_accuracy": 0.7398700997233391,
      "step": 485
    },
    {
      "epoch": 0.36266074567490053,
      "grad_norm": 0.7838660795136883,
      "learning_rate": 1.6095303998230432e-05,
      "loss": 0.8747,
      "mean_token_accuracy": 0.7453865185379982,
      "step": 490
    },
    {
      "epoch": 0.36636136552872606,
      "grad_norm": 0.8988778984216769,
      "learning_rate": 1.599230621434687e-05,
      "loss": 0.8798,
      "mean_token_accuracy": 0.7436122760176659,
      "step": 495
    },
    {
      "epoch": 0.3700619853825516,
      "grad_norm": 0.8288040124260576,
      "learning_rate": 1.5888306874670112e-05,
      "loss": 0.8839,
      "mean_token_accuracy": 0.7431044474244117,
      "step": 500
    },
    {
      "epoch": 0.3700619853825516,
      "eval_loss": 0.90340656042099,
      "eval_mean_token_accuracy": 0.7403560316923893,
      "eval_runtime": 24.4427,
      "eval_samples_per_second": 20.988,
      "eval_steps_per_second": 1.35,
      "step": 500
    },
    {
      "epoch": 0.3737626052363771,
      "grad_norm": 0.8197782426055139,
      "learning_rate": 1.5783323361679865e-05,
      "loss": 0.8799,
      "mean_token_accuracy": 0.7442870959639549,
      "step": 505
    },
    {
      "epoch": 0.37746322509020264,
      "grad_norm": 0.7912056501798941,
      "learning_rate": 1.567737322235084e-05,
      "loss": 0.8661,
      "mean_token_accuracy": 0.7487552657723426,
      "step": 510
    },
    {
      "epoch": 0.3811638449440281,
      "grad_norm": 0.8680442053546329,
      "learning_rate": 1.557047416521996e-05,
      "loss": 0.8446,
      "mean_token_accuracy": 0.7539586842060089,
      "step": 515
    },
    {
      "epoch": 0.38486446479785363,
      "grad_norm": 0.7966248473049342,
      "learning_rate": 1.546264405742654e-05,
      "loss": 0.8894,
      "mean_token_accuracy": 0.7421392247080802,
      "step": 520
    },
    {
      "epoch": 0.38856508465167916,
      "grad_norm": 0.7832801079547608,
      "learning_rate": 1.535390092172597e-05,
      "loss": 0.8401,
      "mean_token_accuracy": 0.7548262551426888,
      "step": 525
    },
    {
      "epoch": 0.3922657045055047,
      "grad_norm": 0.8617978523223222,
      "learning_rate": 1.5244262933477401e-05,
      "loss": 0.8676,
      "mean_token_accuracy": 0.7473892971873284,
      "step": 530
    },
    {
      "epoch": 0.3959663243593302,
      "grad_norm": 0.8638721458324728,
      "learning_rate": 1.5133748417605878e-05,
      "loss": 0.8734,
      "mean_token_accuracy": 0.7463245391845703,
      "step": 535
    },
    {
      "epoch": 0.3996669442131557,
      "grad_norm": 0.8732616764981507,
      "learning_rate": 1.5022375845539537e-05,
      "loss": 0.8742,
      "mean_token_accuracy": 0.74578056037426,
      "step": 540
    },
    {
      "epoch": 0.4033675640669812,
      "grad_norm": 0.8680491958485625,
      "learning_rate": 1.4910163832122278e-05,
      "loss": 0.9111,
      "mean_token_accuracy": 0.7350529983639718,
      "step": 545
    },
    {
      "epoch": 0.40706818392080674,
      "grad_norm": 0.845161443541315,
      "learning_rate": 1.4797131132502464e-05,
      "loss": 0.8543,
      "mean_token_accuracy": 0.7507802724838257,
      "step": 550
    },
    {
      "epoch": 0.41076880377463226,
      "grad_norm": 0.7949620736755997,
      "learning_rate": 1.4683296638998192e-05,
      "loss": 0.8959,
      "mean_token_accuracy": 0.7393335103988647,
      "step": 555
    },
    {
      "epoch": 0.4144694236284578,
      "grad_norm": 0.8137916315549841,
      "learning_rate": 1.4568679377939619e-05,
      "loss": 0.8599,
      "mean_token_accuracy": 0.7493755236268044,
      "step": 560
    },
    {
      "epoch": 0.41817004348228326,
      "grad_norm": 0.8553447099213015,
      "learning_rate": 1.4453298506488896e-05,
      "loss": 0.8562,
      "mean_token_accuracy": 0.7500303864479065,
      "step": 565
    },
    {
      "epoch": 0.4218706633361088,
      "grad_norm": 0.78741790468558,
      "learning_rate": 1.4337173309438236e-05,
      "loss": 0.8513,
      "mean_token_accuracy": 0.751435661315918,
      "step": 570
    },
    {
      "epoch": 0.4255712831899343,
      "grad_norm": 0.8198917515370167,
      "learning_rate": 1.4220323195986649e-05,
      "loss": 0.8892,
      "mean_token_accuracy": 0.740348969399929,
      "step": 575
    },
    {
      "epoch": 0.42927190304375984,
      "grad_norm": 0.8907770425857867,
      "learning_rate": 1.4102767696495885e-05,
      "loss": 0.87,
      "mean_token_accuracy": 0.7466455265879631,
      "step": 580
    },
    {
      "epoch": 0.43297252289758537,
      "grad_norm": 0.7746865966025234,
      "learning_rate": 1.398452645922611e-05,
      "loss": 0.8543,
      "mean_token_accuracy": 0.7507255434989929,
      "step": 585
    },
    {
      "epoch": 0.43667314275141084,
      "grad_norm": 0.8081107449852742,
      "learning_rate": 1.3865619247051916e-05,
      "loss": 0.8942,
      "mean_token_accuracy": 0.7397695079445838,
      "step": 590
    },
    {
      "epoch": 0.44037376260523636,
      "grad_norm": 0.7928994894049232,
      "learning_rate": 1.3746065934159123e-05,
      "loss": 0.8711,
      "mean_token_accuracy": 0.7459681868553162,
      "step": 595
    },
    {
      "epoch": 0.4440743824590619,
      "grad_norm": 0.7457840735614479,
      "learning_rate": 1.3625886502723008e-05,
      "loss": 0.8838,
      "mean_token_accuracy": 0.7431347534060478,
      "step": 600
    },
    {
      "epoch": 0.4440743824590619,
      "eval_loss": 0.8946976065635681,
      "eval_mean_token_accuracy": 0.7420943668394377,
      "eval_runtime": 24.4316,
      "eval_samples_per_second": 20.997,
      "eval_steps_per_second": 1.351,
      "step": 600
    },
    {
      "epoch": 0.4477750023128874,
      "grad_norm": 0.7926368650744795,
      "learning_rate": 1.3505101039568494e-05,
      "loss": 0.8728,
      "mean_token_accuracy": 0.7459641486406327,
      "step": 605
    },
    {
      "epoch": 0.45147562216671294,
      "grad_norm": 0.8000063636466884,
      "learning_rate": 1.3383729732812814e-05,
      "loss": 0.8669,
      "mean_token_accuracy": 0.7467781469225884,
      "step": 610
    },
    {
      "epoch": 0.4551762420205384,
      "grad_norm": 0.8077945425677387,
      "learning_rate": 1.3261792868491267e-05,
      "loss": 0.8805,
      "mean_token_accuracy": 0.7430326372385025,
      "step": 615
    },
    {
      "epoch": 0.45887686187436394,
      "grad_norm": 0.7941374586541651,
      "learning_rate": 1.3139310827166613e-05,
      "loss": 0.859,
      "mean_token_accuracy": 0.7493226900696754,
      "step": 620
    },
    {
      "epoch": 0.46257748172818947,
      "grad_norm": 0.8015163522302382,
      "learning_rate": 1.3016304080522657e-05,
      "loss": 0.8573,
      "mean_token_accuracy": 0.7498888701200486,
      "step": 625
    },
    {
      "epoch": 0.466278101582015,
      "grad_norm": 0.8987066171816293,
      "learning_rate": 1.2892793187942588e-05,
      "loss": 0.8589,
      "mean_token_accuracy": 0.7495828256011009,
      "step": 630
    },
    {
      "epoch": 0.4699787214358405,
      "grad_norm": 0.8210878729300097,
      "learning_rate": 1.2768798793072708e-05,
      "loss": 0.8397,
      "mean_token_accuracy": 0.7545604810118676,
      "step": 635
    },
    {
      "epoch": 0.47367934128966604,
      "grad_norm": 0.8134721151372588,
      "learning_rate": 1.2644341620372025e-05,
      "loss": 0.9003,
      "mean_token_accuracy": 0.7380212768912315,
      "step": 640
    },
    {
      "epoch": 0.4773799611434915,
      "grad_norm": 0.7617180215752571,
      "learning_rate": 1.2519442471648364e-05,
      "loss": 0.8451,
      "mean_token_accuracy": 0.7525444984436035,
      "step": 645
    },
    {
      "epoch": 0.48108058099731704,
      "grad_norm": 0.7862729854353447,
      "learning_rate": 1.2394122222581557e-05,
      "loss": 0.844,
      "mean_token_accuracy": 0.7545543164014816,
      "step": 650
    },
    {
      "epoch": 0.48478120085114257,
      "grad_norm": 0.7602815222283319,
      "learning_rate": 1.226840181923427e-05,
      "loss": 0.8628,
      "mean_token_accuracy": 0.7499380096793175,
      "step": 655
    },
    {
      "epoch": 0.4884818207049681,
      "grad_norm": 0.7942344985162466,
      "learning_rate": 1.214230227455106e-05,
      "loss": 0.8903,
      "mean_token_accuracy": 0.7406057506799698,
      "step": 660
    },
    {
      "epoch": 0.4921824405587936,
      "grad_norm": 0.827163943781085,
      "learning_rate": 1.201584466484629e-05,
      "loss": 0.8781,
      "mean_token_accuracy": 0.7433477059006691,
      "step": 665
    },
    {
      "epoch": 0.4958830604126191,
      "grad_norm": 0.8927178947512965,
      "learning_rate": 1.1889050126281405e-05,
      "loss": 0.8762,
      "mean_token_accuracy": 0.7436618626117706,
      "step": 670
    },
    {
      "epoch": 0.4995836802664446,
      "grad_norm": 0.893577761513805,
      "learning_rate": 1.1761939851332241e-05,
      "loss": 0.8583,
      "mean_token_accuracy": 0.7491728380322457,
      "step": 675
    },
    {
      "epoch": 0.5032843001202701,
      "grad_norm": 0.8911659991233541,
      "learning_rate": 1.1634535085246903e-05,
      "loss": 0.8347,
      "mean_token_accuracy": 0.7567671954631805,
      "step": 680
    },
    {
      "epoch": 0.5069849199740957,
      "grad_norm": 0.7822165331384006,
      "learning_rate": 1.1506857122494832e-05,
      "loss": 0.8404,
      "mean_token_accuracy": 0.754469695687294,
      "step": 685
    },
    {
      "epoch": 0.5106855398279212,
      "grad_norm": 0.7759748923454998,
      "learning_rate": 1.1378927303207637e-05,
      "loss": 0.8741,
      "mean_token_accuracy": 0.7437764629721642,
      "step": 690
    },
    {
      "epoch": 0.5143861596817467,
      "grad_norm": 0.8502606153370593,
      "learning_rate": 1.12507670096123e-05,
      "loss": 0.8582,
      "mean_token_accuracy": 0.7506067097187042,
      "step": 695
    },
    {
      "epoch": 0.5180867795355723,
      "grad_norm": 0.7749617835044338,
      "learning_rate": 1.1122397662457352e-05,
      "loss": 0.8538,
      "mean_token_accuracy": 0.7503936603665352,
      "step": 700
    },
    {
      "epoch": 0.5180867795355723,
      "eval_loss": 0.888458788394928,
      "eval_mean_token_accuracy": 0.7433922489484152,
      "eval_runtime": 24.5105,
      "eval_samples_per_second": 20.93,
      "eval_steps_per_second": 1.346,
      "step": 700
    },
    {
      "epoch": 0.5217873993893978,
      "grad_norm": 0.7706407219978753,
      "learning_rate": 1.0993840717432582e-05,
      "loss": 0.8553,
      "mean_token_accuracy": 0.7487444669008255,
      "step": 705
    },
    {
      "epoch": 0.5254880192432232,
      "grad_norm": 0.7705189840242923,
      "learning_rate": 1.0865117661582958e-05,
      "loss": 0.8858,
      "mean_token_accuracy": 0.7415965974330903,
      "step": 710
    },
    {
      "epoch": 0.5291886390970487,
      "grad_norm": 0.8411896195744224,
      "learning_rate": 1.0736250009717249e-05,
      "loss": 0.8884,
      "mean_token_accuracy": 0.7412499740719796,
      "step": 715
    },
    {
      "epoch": 0.5328892589508742,
      "grad_norm": 0.8230171145609085,
      "learning_rate": 1.0607259300812047e-05,
      "loss": 0.8705,
      "mean_token_accuracy": 0.745503506064415,
      "step": 720
    },
    {
      "epoch": 0.5365898788046998,
      "grad_norm": 0.8401376931438745,
      "learning_rate": 1.0478167094411733e-05,
      "loss": 0.8707,
      "mean_token_accuracy": 0.7456582605838775,
      "step": 725
    },
    {
      "epoch": 0.5402904986585253,
      "grad_norm": 0.7558836285060193,
      "learning_rate": 1.0348994967025012e-05,
      "loss": 0.8545,
      "mean_token_accuracy": 0.7487017199397087,
      "step": 730
    },
    {
      "epoch": 0.5439911185123508,
      "grad_norm": 0.7642061571005638,
      "learning_rate": 1.0219764508518595e-05,
      "loss": 0.8654,
      "mean_token_accuracy": 0.7462582185864448,
      "step": 735
    },
    {
      "epoch": 0.5476917383661764,
      "grad_norm": 0.8147049407956067,
      "learning_rate": 1.0090497318508687e-05,
      "loss": 0.8424,
      "mean_token_accuracy": 0.75324746966362,
      "step": 740
    },
    {
      "epoch": 0.5513923582200019,
      "grad_norm": 0.7571808100586339,
      "learning_rate": 9.961215002750799e-06,
      "loss": 0.8588,
      "mean_token_accuracy": 0.7491571202874183,
      "step": 745
    },
    {
      "epoch": 0.5550929780738274,
      "grad_norm": 0.8038741472628308,
      "learning_rate": 9.831939169528565e-06,
      "loss": 0.8582,
      "mean_token_accuracy": 0.7493733122944832,
      "step": 750
    },
    {
      "epoch": 0.5587935979276529,
      "grad_norm": 0.848259052134148,
      "learning_rate": 9.702691426042124e-06,
      "loss": 0.8526,
      "mean_token_accuracy": 0.7506043568253518,
      "step": 755
    },
    {
      "epoch": 0.5624942177814783,
      "grad_norm": 0.7842186156617634,
      "learning_rate": 9.573493374796694e-06,
      "loss": 0.8369,
      "mean_token_accuracy": 0.755862507224083,
      "step": 760
    },
    {
      "epoch": 0.5661948376353039,
      "grad_norm": 0.7485121651837016,
      "learning_rate": 9.444366609991916e-06,
      "loss": 0.8729,
      "mean_token_accuracy": 0.7438469439744949,
      "step": 765
    },
    {
      "epoch": 0.5698954574891294,
      "grad_norm": 0.9176291688747804,
      "learning_rate": 9.315332713912593e-06,
      "loss": 0.8815,
      "mean_token_accuracy": 0.7436545789241791,
      "step": 770
    },
    {
      "epoch": 0.5735960773429549,
      "grad_norm": 0.8416476293024356,
      "learning_rate": 9.18641325332142e-06,
      "loss": 0.8579,
      "mean_token_accuracy": 0.7489444330334664,
      "step": 775
    },
    {
      "epoch": 0.5772966971967805,
      "grad_norm": 0.8665126504141849,
      "learning_rate": 9.057629775854314e-06,
      "loss": 0.8691,
      "mean_token_accuracy": 0.7452979102730751,
      "step": 780
    },
    {
      "epoch": 0.580997317050606,
      "grad_norm": 0.7873121719067726,
      "learning_rate": 8.929003806418934e-06,
      "loss": 0.8609,
      "mean_token_accuracy": 0.7471121177077293,
      "step": 785
    },
    {
      "epoch": 0.5846979369044315,
      "grad_norm": 0.7603699249259593,
      "learning_rate": 8.800556843597002e-06,
      "loss": 0.8303,
      "mean_token_accuracy": 0.7559542834758759,
      "step": 790
    },
    {
      "epoch": 0.588398556758257,
      "grad_norm": 0.8208200680668915,
      "learning_rate": 8.672310356051023e-06,
      "loss": 0.8755,
      "mean_token_accuracy": 0.742991179227829,
      "step": 795
    },
    {
      "epoch": 0.5920991766120826,
      "grad_norm": 0.7539353236493428,
      "learning_rate": 8.544285778936004e-06,
      "loss": 0.8604,
      "mean_token_accuracy": 0.7469533935189248,
      "step": 800
    },
    {
      "epoch": 0.5920991766120826,
      "eval_loss": 0.882485568523407,
      "eval_mean_token_accuracy": 0.7446909926154397,
      "eval_runtime": 24.5244,
      "eval_samples_per_second": 20.918,
      "eval_steps_per_second": 1.346,
      "step": 800
    },
    {
      "epoch": 0.5957997964659081,
      "grad_norm": 0.7747168848469933,
      "learning_rate": 8.416504510316774e-06,
      "loss": 0.8687,
      "mean_token_accuracy": 0.7457119628787041,
      "step": 805
    },
    {
      "epoch": 0.5995004163197336,
      "grad_norm": 0.7965745611460714,
      "learning_rate": 8.28898790759152e-06,
      "loss": 0.8719,
      "mean_token_accuracy": 0.7440481051802635,
      "step": 810
    },
    {
      "epoch": 0.603201036173559,
      "grad_norm": 0.7287090635465163,
      "learning_rate": 8.161757283922084e-06,
      "loss": 0.8259,
      "mean_token_accuracy": 0.7569043532013893,
      "step": 815
    },
    {
      "epoch": 0.6069016560273846,
      "grad_norm": 0.7849161686448317,
      "learning_rate": 8.034833904671698e-06,
      "loss": 0.8725,
      "mean_token_accuracy": 0.7442520692944526,
      "step": 820
    },
    {
      "epoch": 0.6106022758812101,
      "grad_norm": 0.7891027028253116,
      "learning_rate": 7.908238983850666e-06,
      "loss": 0.8365,
      "mean_token_accuracy": 0.75490812510252,
      "step": 825
    },
    {
      "epoch": 0.6143028957350356,
      "grad_norm": 0.7870957116065473,
      "learning_rate": 7.781993680570656e-06,
      "loss": 0.8576,
      "mean_token_accuracy": 0.7477666437625885,
      "step": 830
    },
    {
      "epoch": 0.6180035155888611,
      "grad_norm": 0.8664448499276786,
      "learning_rate": 7.656119095508155e-06,
      "loss": 0.8332,
      "mean_token_accuracy": 0.7553980827331543,
      "step": 835
    },
    {
      "epoch": 0.6217041354426867,
      "grad_norm": 0.8003046870817029,
      "learning_rate": 7.530636267377706e-06,
      "loss": 0.8559,
      "mean_token_accuracy": 0.7494190320372581,
      "step": 840
    },
    {
      "epoch": 0.6254047552965122,
      "grad_norm": 0.8000920295001714,
      "learning_rate": 7.405566169415481e-06,
      "loss": 0.8869,
      "mean_token_accuracy": 0.7401010394096375,
      "step": 845
    },
    {
      "epoch": 0.6291053751503377,
      "grad_norm": 0.7895598832562634,
      "learning_rate": 7.280929705873818e-06,
      "loss": 0.8648,
      "mean_token_accuracy": 0.7470688298344612,
      "step": 850
    },
    {
      "epoch": 0.6328059950041632,
      "grad_norm": 0.7756015770435069,
      "learning_rate": 7.15674770852727e-06,
      "loss": 0.8294,
      "mean_token_accuracy": 0.7561267375946045,
      "step": 855
    },
    {
      "epoch": 0.6365066148579888,
      "grad_norm": 0.7508734070910927,
      "learning_rate": 7.033040933190776e-06,
      "loss": 0.8367,
      "mean_token_accuracy": 0.7520585179328918,
      "step": 860
    },
    {
      "epoch": 0.6402072347118142,
      "grad_norm": 0.7569428687497424,
      "learning_rate": 6.909830056250527e-06,
      "loss": 0.8498,
      "mean_token_accuracy": 0.7501048058271408,
      "step": 865
    },
    {
      "epoch": 0.6439078545656397,
      "grad_norm": 0.7984119233647183,
      "learning_rate": 6.787135671208126e-06,
      "loss": 0.8571,
      "mean_token_accuracy": 0.7490562096238136,
      "step": 870
    },
    {
      "epoch": 0.6476084744194652,
      "grad_norm": 0.8159593622616607,
      "learning_rate": 6.6649782852385554e-06,
      "loss": 0.8686,
      "mean_token_accuracy": 0.7469749033451081,
      "step": 875
    },
    {
      "epoch": 0.6513090942732908,
      "grad_norm": 0.7631802495882793,
      "learning_rate": 6.543378315762634e-06,
      "loss": 0.8657,
      "mean_token_accuracy": 0.7458467945456505,
      "step": 880
    },
    {
      "epoch": 0.6550097141271163,
      "grad_norm": 0.812760399325826,
      "learning_rate": 6.42235608703441e-06,
      "loss": 0.8635,
      "mean_token_accuracy": 0.7484910488128662,
      "step": 885
    },
    {
      "epoch": 0.6587103339809418,
      "grad_norm": 0.8219546330443505,
      "learning_rate": 6.301931826744189e-06,
      "loss": 0.8363,
      "mean_token_accuracy": 0.7542862921953202,
      "step": 890
    },
    {
      "epoch": 0.6624109538347673,
      "grad_norm": 0.7882602270918212,
      "learning_rate": 6.18212566263765e-06,
      "loss": 0.864,
      "mean_token_accuracy": 0.7470616161823272,
      "step": 895
    },
    {
      "epoch": 0.6661115736885929,
      "grad_norm": 0.8408472925032482,
      "learning_rate": 6.0629576191517035e-06,
      "loss": 0.8619,
      "mean_token_accuracy": 0.74820506721735,
      "step": 900
    },
    {
      "epoch": 0.6661115736885929,
      "eval_loss": 0.8769035935401917,
      "eval_mean_token_accuracy": 0.7459533539685336,
      "eval_runtime": 24.4702,
      "eval_samples_per_second": 20.964,
      "eval_steps_per_second": 1.349,
      "step": 900
    },
    {
      "epoch": 0.6698121935424184,
      "grad_norm": 0.7754763488387619,
      "learning_rate": 5.944447614067588e-06,
      "loss": 0.868,
      "mean_token_accuracy": 0.7466130316257477,
      "step": 905
    },
    {
      "epoch": 0.6735128133962439,
      "grad_norm": 0.7533565990898803,
      "learning_rate": 5.8266154551818225e-06,
      "loss": 0.8561,
      "mean_token_accuracy": 0.7474694743752479,
      "step": 910
    },
    {
      "epoch": 0.6772134332500694,
      "grad_norm": 0.848613374257325,
      "learning_rate": 5.709480836995509e-06,
      "loss": 0.867,
      "mean_token_accuracy": 0.7461282700300217,
      "step": 915
    },
    {
      "epoch": 0.6809140531038949,
      "grad_norm": 0.7588179424646748,
      "learning_rate": 5.593063337422595e-06,
      "loss": 0.8525,
      "mean_token_accuracy": 0.7497700050473213,
      "step": 920
    },
    {
      "epoch": 0.6846146729577204,
      "grad_norm": 0.7363445259820587,
      "learning_rate": 5.477382414517625e-06,
      "loss": 0.8616,
      "mean_token_accuracy": 0.7481978580355644,
      "step": 925
    },
    {
      "epoch": 0.6883152928115459,
      "grad_norm": 0.7698222134303955,
      "learning_rate": 5.362457403223495e-06,
      "loss": 0.8553,
      "mean_token_accuracy": 0.749049125611782,
      "step": 930
    },
    {
      "epoch": 0.6920159126653714,
      "grad_norm": 0.7718707701904592,
      "learning_rate": 5.248307512139818e-06,
      "loss": 0.8616,
      "mean_token_accuracy": 0.7483461976051331,
      "step": 935
    },
    {
      "epoch": 0.695716532519197,
      "grad_norm": 0.785721659768036,
      "learning_rate": 5.134951820312402e-06,
      "loss": 0.8557,
      "mean_token_accuracy": 0.7493070676922798,
      "step": 940
    },
    {
      "epoch": 0.6994171523730225,
      "grad_norm": 0.7715074813141185,
      "learning_rate": 5.022409274044346e-06,
      "loss": 0.8493,
      "mean_token_accuracy": 0.7499154567718506,
      "step": 945
    },
    {
      "epoch": 0.703117772226848,
      "grad_norm": 0.8335206414183822,
      "learning_rate": 4.910698683729371e-06,
      "loss": 0.8373,
      "mean_token_accuracy": 0.7533765882253647,
      "step": 950
    },
    {
      "epoch": 0.7068183920806735,
      "grad_norm": 0.7505821456653344,
      "learning_rate": 4.799838720707847e-06,
      "loss": 0.8675,
      "mean_token_accuracy": 0.745770500600338,
      "step": 955
    },
    {
      "epoch": 0.7105190119344991,
      "grad_norm": 0.7594914324330723,
      "learning_rate": 4.6898479141460415e-06,
      "loss": 0.8295,
      "mean_token_accuracy": 0.756369736790657,
      "step": 960
    },
    {
      "epoch": 0.7142196317883246,
      "grad_norm": 0.7204641559045669,
      "learning_rate": 4.580744647939163e-06,
      "loss": 0.8624,
      "mean_token_accuracy": 0.7481252133846283,
      "step": 965
    },
    {
      "epoch": 0.71792025164215,
      "grad_norm": 0.8174852819097401,
      "learning_rate": 4.472547157638674e-06,
      "loss": 0.8417,
      "mean_token_accuracy": 0.7523833066225052,
      "step": 970
    },
    {
      "epoch": 0.7216208714959755,
      "grad_norm": 0.862526259850387,
      "learning_rate": 4.365273527404384e-06,
      "loss": 0.8423,
      "mean_token_accuracy": 0.754362627863884,
      "step": 975
    },
    {
      "epoch": 0.7253214913498011,
      "grad_norm": 0.802739929512475,
      "learning_rate": 4.258941686981864e-06,
      "loss": 0.8469,
      "mean_token_accuracy": 0.7522266939282417,
      "step": 980
    },
    {
      "epoch": 0.7290221112036266,
      "grad_norm": 0.7931469636274437,
      "learning_rate": 4.15356940870567e-06,
      "loss": 0.8305,
      "mean_token_accuracy": 0.7555211216211319,
      "step": 985
    },
    {
      "epoch": 0.7327227310574521,
      "grad_norm": 0.7884974421740458,
      "learning_rate": 4.049174304528857e-06,
      "loss": 0.8327,
      "mean_token_accuracy": 0.7560925737023354,
      "step": 990
    },
    {
      "epoch": 0.7364233509112776,
      "grad_norm": 0.7868742780753382,
      "learning_rate": 3.945773823079315e-06,
      "loss": 0.8491,
      "mean_token_accuracy": 0.7494197428226471,
      "step": 995
    },
    {
      "epoch": 0.7401239707651032,
      "grad_norm": 0.7556900381775965,
      "learning_rate": 3.8433852467434175e-06,
      "loss": 0.8337,
      "mean_token_accuracy": 0.7557049483060837,
      "step": 1000
    },
    {
      "epoch": 0.7401239707651032,
      "eval_loss": 0.872385561466217,
      "eval_mean_token_accuracy": 0.7468928351546779,
      "eval_runtime": 24.5427,
      "eval_samples_per_second": 20.902,
      "eval_steps_per_second": 1.345,
      "step": 1000
    },
    {
      "epoch": 0.7438245906189287,
      "grad_norm": 0.7642811057752702,
      "learning_rate": 3.742025688777413e-06,
      "loss": 0.8082,
      "mean_token_accuracy": 0.7626117318868637,
      "step": 1005
    },
    {
      "epoch": 0.7475252104727542,
      "grad_norm": 0.7878330759901441,
      "learning_rate": 3.641712090447125e-06,
      "loss": 0.8584,
      "mean_token_accuracy": 0.7494427710771561,
      "step": 1010
    },
    {
      "epoch": 0.7512258303265797,
      "grad_norm": 0.7817001972699995,
      "learning_rate": 3.542461218196379e-06,
      "loss": 0.8596,
      "mean_token_accuracy": 0.7491939187049865,
      "step": 1015
    },
    {
      "epoch": 0.7549264501804053,
      "grad_norm": 0.7922908897094753,
      "learning_rate": 3.444289660844665e-06,
      "loss": 0.837,
      "mean_token_accuracy": 0.7543808072805405,
      "step": 1020
    },
    {
      "epoch": 0.7586270700342307,
      "grad_norm": 0.7945985767364506,
      "learning_rate": 3.347213826814456e-06,
      "loss": 0.8662,
      "mean_token_accuracy": 0.7463411048054696,
      "step": 1025
    },
    {
      "epoch": 0.7623276898880562,
      "grad_norm": 0.7549974443621008,
      "learning_rate": 3.2512499413887255e-06,
      "loss": 0.8303,
      "mean_token_accuracy": 0.7554700002074242,
      "step": 1030
    },
    {
      "epoch": 0.7660283097418817,
      "grad_norm": 0.8236761629232633,
      "learning_rate": 3.1564140439990256e-06,
      "loss": 0.8406,
      "mean_token_accuracy": 0.751871857047081,
      "step": 1035
    },
    {
      "epoch": 0.7697289295957073,
      "grad_norm": 0.7267400739764023,
      "learning_rate": 3.0627219855446667e-06,
      "loss": 0.8287,
      "mean_token_accuracy": 0.7558587804436684,
      "step": 1040
    },
    {
      "epoch": 0.7734295494495328,
      "grad_norm": 0.776879654823643,
      "learning_rate": 2.970189425743383e-06,
      "loss": 0.8361,
      "mean_token_accuracy": 0.7547884792089462,
      "step": 1045
    },
    {
      "epoch": 0.7771301693033583,
      "grad_norm": 0.7552506340173256,
      "learning_rate": 2.8788318305139808e-06,
      "loss": 0.8286,
      "mean_token_accuracy": 0.7568780824542045,
      "step": 1050
    },
    {
      "epoch": 0.7808307891571838,
      "grad_norm": 0.7580209757202828,
      "learning_rate": 2.7886644693913333e-06,
      "loss": 0.8459,
      "mean_token_accuracy": 0.750422203540802,
      "step": 1055
    },
    {
      "epoch": 0.7845314090110094,
      "grad_norm": 0.7719039642796498,
      "learning_rate": 2.6997024129742544e-06,
      "loss": 0.8404,
      "mean_token_accuracy": 0.7537851154804229,
      "step": 1060
    },
    {
      "epoch": 0.7882320288648349,
      "grad_norm": 0.7810831799894009,
      "learning_rate": 2.611960530406572e-06,
      "loss": 0.8273,
      "mean_token_accuracy": 0.7569017544388771,
      "step": 1065
    },
    {
      "epoch": 0.7919326487186604,
      "grad_norm": 0.8590214475734921,
      "learning_rate": 2.5254534868919077e-06,
      "loss": 0.8299,
      "mean_token_accuracy": 0.7555602207779885,
      "step": 1070
    },
    {
      "epoch": 0.7956332685724858,
      "grad_norm": 0.7451834304748566,
      "learning_rate": 2.4401957412425213e-06,
      "loss": 0.847,
      "mean_token_accuracy": 0.7509198769927025,
      "step": 1075
    },
    {
      "epoch": 0.7993338884263114,
      "grad_norm": 0.7394748846818531,
      "learning_rate": 2.3562015434626784e-06,
      "loss": 0.8478,
      "mean_token_accuracy": 0.7510868698358536,
      "step": 1080
    },
    {
      "epoch": 0.8030345082801369,
      "grad_norm": 0.7653067572984463,
      "learning_rate": 2.273484932366874e-06,
      "loss": 0.847,
      "mean_token_accuracy": 0.7503063544631005,
      "step": 1085
    },
    {
      "epoch": 0.8067351281339624,
      "grad_norm": 0.7743151365038076,
      "learning_rate": 2.192059733233408e-06,
      "loss": 0.8472,
      "mean_token_accuracy": 0.7531051859259605,
      "step": 1090
    },
    {
      "epoch": 0.810435747987788,
      "grad_norm": 0.7474033991103015,
      "learning_rate": 2.111939555493603e-06,
      "loss": 0.8459,
      "mean_token_accuracy": 0.7517727881669998,
      "step": 1095
    },
    {
      "epoch": 0.8141363678416135,
      "grad_norm": 0.779312232265544,
      "learning_rate": 2.0331377904571303e-06,
      "loss": 0.8276,
      "mean_token_accuracy": 0.7547013550996781,
      "step": 1100
    },
    {
      "epoch": 0.8141363678416135,
      "eval_loss": 0.8695138692855835,
      "eval_mean_token_accuracy": 0.7477564071163987,
      "eval_runtime": 24.4317,
      "eval_samples_per_second": 20.997,
      "eval_steps_per_second": 1.351,
      "step": 1100
    },
    {
      "epoch": 0.817836987695439,
      "grad_norm": 0.8034562413677603,
      "learning_rate": 1.9556676090737803e-06,
      "loss": 0.8317,
      "mean_token_accuracy": 0.7545135840773582,
      "step": 1105
    },
    {
      "epoch": 0.8215376075492645,
      "grad_norm": 0.7281049735394789,
      "learning_rate": 1.879541959732072e-06,
      "loss": 0.853,
      "mean_token_accuracy": 0.749828140437603,
      "step": 1110
    },
    {
      "epoch": 0.82523822740309,
      "grad_norm": 0.7645464251699309,
      "learning_rate": 1.8047735660950427e-06,
      "loss": 0.8095,
      "mean_token_accuracy": 0.7608827918767929,
      "step": 1115
    },
    {
      "epoch": 0.8289388472569156,
      "grad_norm": 0.7946888766498504,
      "learning_rate": 1.7313749249736266e-06,
      "loss": 0.8157,
      "mean_token_accuracy": 0.760844600200653,
      "step": 1120
    },
    {
      "epoch": 0.832639467110741,
      "grad_norm": 0.7522474191468512,
      "learning_rate": 1.6593583042379192e-06,
      "loss": 0.8347,
      "mean_token_accuracy": 0.7559689804911613,
      "step": 1125
    },
    {
      "epoch": 0.8363400869645665,
      "grad_norm": 0.7489873580556787,
      "learning_rate": 1.5887357407667314e-06,
      "loss": 0.8605,
      "mean_token_accuracy": 0.7474079817533493,
      "step": 1130
    },
    {
      "epoch": 0.840040706818392,
      "grad_norm": 0.7641134150797313,
      "learning_rate": 1.5195190384357405e-06,
      "loss": 0.8618,
      "mean_token_accuracy": 0.7480388507246971,
      "step": 1135
    },
    {
      "epoch": 0.8437413266722176,
      "grad_norm": 0.7204342341589023,
      "learning_rate": 1.4517197661445893e-06,
      "loss": 0.8529,
      "mean_token_accuracy": 0.7483357265591621,
      "step": 1140
    },
    {
      "epoch": 0.8474419465260431,
      "grad_norm": 0.7681754372477057,
      "learning_rate": 1.3853492558832472e-06,
      "loss": 0.8306,
      "mean_token_accuracy": 0.7552460536360741,
      "step": 1145
    },
    {
      "epoch": 0.8511425663798686,
      "grad_norm": 0.7429470084922081,
      "learning_rate": 1.3204186008379926e-06,
      "loss": 0.8389,
      "mean_token_accuracy": 0.7531604886054992,
      "step": 1150
    },
    {
      "epoch": 0.8548431862336942,
      "grad_norm": 0.6873318445062786,
      "learning_rate": 1.2569386535372807e-06,
      "loss": 0.8286,
      "mean_token_accuracy": 0.7552321195602417,
      "step": 1155
    },
    {
      "epoch": 0.8585438060875197,
      "grad_norm": 0.7640481385504215,
      "learning_rate": 1.1949200240378577e-06,
      "loss": 0.8255,
      "mean_token_accuracy": 0.7569879427552223,
      "step": 1160
    },
    {
      "epoch": 0.8622444259413452,
      "grad_norm": 0.7638091289655545,
      "learning_rate": 1.1343730781513896e-06,
      "loss": 0.8558,
      "mean_token_accuracy": 0.749303475022316,
      "step": 1165
    },
    {
      "epoch": 0.8659450457951707,
      "grad_norm": 0.7685982859289542,
      "learning_rate": 1.0753079357119134e-06,
      "loss": 0.8273,
      "mean_token_accuracy": 0.7565306261181831,
      "step": 1170
    },
    {
      "epoch": 0.8696456656489963,
      "grad_norm": 0.7114001929021425,
      "learning_rate": 1.017734468884417e-06,
      "loss": 0.8473,
      "mean_token_accuracy": 0.7502565905451775,
      "step": 1175
    },
    {
      "epoch": 0.8733462855028217,
      "grad_norm": 0.7348011760240892,
      "learning_rate": 9.616623005147952e-07,
      "loss": 0.8286,
      "mean_token_accuracy": 0.7560262143611908,
      "step": 1180
    },
    {
      "epoch": 0.8770469053566472,
      "grad_norm": 0.7429654371816976,
      "learning_rate": 9.071008025214767e-07,
      "loss": 0.8158,
      "mean_token_accuracy": 0.760336747765541,
      "step": 1185
    },
    {
      "epoch": 0.8807475252104727,
      "grad_norm": 0.8100937899257782,
      "learning_rate": 8.540590943290128e-07,
      "loss": 0.8328,
      "mean_token_accuracy": 0.7545322120189667,
      "step": 1190
    },
    {
      "epoch": 0.8844481450642983,
      "grad_norm": 0.7556302951610571,
      "learning_rate": 8.025460413438457e-07,
      "loss": 0.842,
      "mean_token_accuracy": 0.7518000155687332,
      "step": 1195
    },
    {
      "epoch": 0.8881487649181238,
      "grad_norm": 0.7214919634823889,
      "learning_rate": 7.525702534725443e-07,
      "loss": 0.8485,
      "mean_token_accuracy": 0.7501253366470337,
      "step": 1200
    },
    {
      "epoch": 0.8881487649181238,
      "eval_loss": 0.8681026101112366,
      "eval_mean_token_accuracy": 0.7479201501066034,
      "eval_runtime": 24.4445,
      "eval_samples_per_second": 20.986,
      "eval_steps_per_second": 1.35,
      "step": 1200
    },
    {
      "epoch": 0.8918493847719493,
      "grad_norm": 0.7363787692810877,
      "learning_rate": 7.041400836827439e-07,
      "loss": 0.8446,
      "mean_token_accuracy": 0.7505712598562241,
      "step": 1205
    },
    {
      "epoch": 0.8955500046257748,
      "grad_norm": 0.7760798741879421,
      "learning_rate": 6.572636266070265e-07,
      "loss": 0.8515,
      "mean_token_accuracy": 0.7495753020048141,
      "step": 1210
    },
    {
      "epoch": 0.8992506244796004,
      "grad_norm": 0.7375925847527678,
      "learning_rate": 6.119487171899807e-07,
      "loss": 0.8151,
      "mean_token_accuracy": 0.7605884402990342,
      "step": 1215
    },
    {
      "epoch": 0.9029512443334259,
      "grad_norm": 0.8483447358597658,
      "learning_rate": 5.682029293786673e-07,
      "loss": 0.8508,
      "mean_token_accuracy": 0.7498715907335282,
      "step": 1220
    },
    {
      "epoch": 0.9066518641872514,
      "grad_norm": 0.7427906597662893,
      "learning_rate": 5.26033574856708e-07,
      "loss": 0.8254,
      "mean_token_accuracy": 0.7563312321901321,
      "step": 1225
    },
    {
      "epoch": 0.9103524840410768,
      "grad_norm": 0.720458409183581,
      "learning_rate": 4.854477018222103e-07,
      "loss": 0.8374,
      "mean_token_accuracy": 0.7546701580286026,
      "step": 1230
    },
    {
      "epoch": 0.9140531038949024,
      "grad_norm": 0.7702820261310465,
      "learning_rate": 4.464520938097294e-07,
      "loss": 0.8397,
      "mean_token_accuracy": 0.7535283699631691,
      "step": 1235
    },
    {
      "epoch": 0.9177537237487279,
      "grad_norm": 0.8128830297797902,
      "learning_rate": 4.0905326855646186e-07,
      "loss": 0.855,
      "mean_token_accuracy": 0.7464580446481704,
      "step": 1240
    },
    {
      "epoch": 0.9214543436025534,
      "grad_norm": 0.7719201359932438,
      "learning_rate": 3.732574769128738e-07,
      "loss": 0.8585,
      "mean_token_accuracy": 0.7479515969753265,
      "step": 1245
    },
    {
      "epoch": 0.9251549634563789,
      "grad_norm": 0.767822410168291,
      "learning_rate": 3.390707017979311e-07,
      "loss": 0.8417,
      "mean_token_accuracy": 0.7507120683789253,
      "step": 1250
    },
    {
      "epoch": 0.9288555833102045,
      "grad_norm": 0.7931837924998084,
      "learning_rate": 3.06498657199108e-07,
      "loss": 0.8423,
      "mean_token_accuracy": 0.7526604071259498,
      "step": 1255
    },
    {
      "epoch": 0.93255620316403,
      "grad_norm": 0.7402614770859547,
      "learning_rate": 2.7554678721735675e-07,
      "loss": 0.831,
      "mean_token_accuracy": 0.7547580033540726,
      "step": 1260
    },
    {
      "epoch": 0.9362568230178555,
      "grad_norm": 0.7265080682796597,
      "learning_rate": 2.4622026515717654e-07,
      "loss": 0.8577,
      "mean_token_accuracy": 0.7485264018177986,
      "step": 1265
    },
    {
      "epoch": 0.939957442871681,
      "grad_norm": 0.7994816530184113,
      "learning_rate": 2.1852399266194312e-07,
      "loss": 0.8384,
      "mean_token_accuracy": 0.7530582755804062,
      "step": 1270
    },
    {
      "epoch": 0.9436580627255066,
      "grad_norm": 0.7408942586414351,
      "learning_rate": 1.9246259889464935e-07,
      "loss": 0.8395,
      "mean_token_accuracy": 0.7526318833231926,
      "step": 1275
    },
    {
      "epoch": 0.9473586825793321,
      "grad_norm": 0.7531498605702176,
      "learning_rate": 1.6804043976418438e-07,
      "loss": 0.8384,
      "mean_token_accuracy": 0.7517712652683258,
      "step": 1280
    },
    {
      "epoch": 0.9510593024331575,
      "grad_norm": 0.7640371632548638,
      "learning_rate": 1.4526159719728595e-07,
      "loss": 0.843,
      "mean_token_accuracy": 0.7513806536793709,
      "step": 1285
    },
    {
      "epoch": 0.954759922286983,
      "grad_norm": 0.7800995207503227,
      "learning_rate": 1.24129878456285e-07,
      "loss": 0.8482,
      "mean_token_accuracy": 0.750167365372181,
      "step": 1290
    },
    {
      "epoch": 0.9584605421408086,
      "grad_norm": 0.7781230447019286,
      "learning_rate": 1.0464881550276362e-07,
      "loss": 0.8571,
      "mean_token_accuracy": 0.7482250303030014,
      "step": 1295
    },
    {
      "epoch": 0.9621611619946341,
      "grad_norm": 0.7814221898628717,
      "learning_rate": 8.682166440721729e-08,
      "loss": 0.8554,
      "mean_token_accuracy": 0.7490108326077461,
      "step": 1300
    },
    {
      "epoch": 0.9621611619946341,
      "eval_loss": 0.8676031827926636,
      "eval_mean_token_accuracy": 0.7480372389157613,
      "eval_runtime": 24.4444,
      "eval_samples_per_second": 20.986,
      "eval_steps_per_second": 1.35,
      "step": 1300
    },
    {
      "epoch": 0.9658617818484596,
      "grad_norm": 0.7864179680038271,
      "learning_rate": 7.065140480483235e-08,
      "loss": 0.8612,
      "mean_token_accuracy": 0.7473657980561257,
      "step": 1305
    },
    {
      "epoch": 0.9695624017022851,
      "grad_norm": 0.7508526784486548,
      "learning_rate": 5.6140739397474445e-08,
      "loss": 0.8365,
      "mean_token_accuracy": 0.7531405627727509,
      "step": 1310
    },
    {
      "epoch": 0.9732630215561107,
      "grad_norm": 0.7164581752617882,
      "learning_rate": 4.329209350195651e-08,
      "loss": 0.8443,
      "mean_token_accuracy": 0.7516303405165672,
      "step": 1315
    },
    {
      "epoch": 0.9769636414099362,
      "grad_norm": 0.8498247341964699,
      "learning_rate": 3.210761464466639e-08,
      "loss": 0.8374,
      "mean_token_accuracy": 0.7550172284245491,
      "step": 1320
    },
    {
      "epoch": 0.9806642612637617,
      "grad_norm": 0.7153929981028929,
      "learning_rate": 2.2589172202635014e-08,
      "loss": 0.7977,
      "mean_token_accuracy": 0.7646528780460358,
      "step": 1325
    },
    {
      "epoch": 0.9843648811175872,
      "grad_norm": 0.7908203411439174,
      "learning_rate": 1.4738357091084177e-08,
      "loss": 0.8371,
      "mean_token_accuracy": 0.753947702050209,
      "step": 1330
    },
    {
      "epoch": 0.9880655009714127,
      "grad_norm": 0.7547450484335274,
      "learning_rate": 8.556481497521418e-09,
      "loss": 0.8498,
      "mean_token_accuracy": 0.7504909783601761,
      "step": 1335
    },
    {
      "epoch": 0.9917661208252382,
      "grad_norm": 0.7442071956341998,
      "learning_rate": 4.044578662419918e-09,
      "loss": 0.8265,
      "mean_token_accuracy": 0.7562764957547188,
      "step": 1340
    },
    {
      "epoch": 0.9954667406790637,
      "grad_norm": 0.7540578819730456,
      "learning_rate": 1.203402706525525e-09,
      "loss": 0.8332,
      "mean_token_accuracy": 0.7546069085597992,
      "step": 1345
    },
    {
      "epoch": 0.9991673605328892,
      "grad_norm": 0.7649196329052427,
      "learning_rate": 3.342850480869686e-11,
      "loss": 0.805,
      "mean_token_accuracy": 0.7623659715056419,
      "step": 1350
    },
    {
      "epoch": 0.9999074845036544,
      "mean_token_accuracy": 0.7686276957392693,
      "step": 1351,
      "total_flos": 76959195168768.0,
      "train_loss": 0.8854525579424984,
      "train_runtime": 18393.1053,
      "train_samples_per_second": 4.701,
      "train_steps_per_second": 0.073
    }
  ],
  "logging_steps": 5,
  "max_steps": 1351,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 76959195168768.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}