Instructions to use When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift")
model = AutoModelForCausalLM.from_pretrained("When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift

SGLang

How to use When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift with Docker Model Runner:
```
docker model run hf.co/When-Does-Reasoning-Matter/Qwen2.5-0.5B-ift
```

Qwen2.5-0.5B-ift / slurm.out

Nicolas-BZRD

Add files using upload-large-folder tool

be2e11f verified 7 months ago

raw

history blame

307 kB

	1: W0820 16:56:16.725000 1497662 torch/distributed/run.py:792]
	1: W0820 16:56:16.725000 1497662 torch/distributed/run.py:792] *****************************************
	1: W0820 16:56:16.725000 1497662 torch/distributed/run.py:792] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.
	1: W0820 16:56:16.725000 1497662 torch/distributed/run.py:792] *****************************************
	0: W0820 16:56:16.732000 880870 torch/distributed/run.py:792]
	0: W0820 16:56:16.732000 880870 torch/distributed/run.py:792] *****************************************
	0: W0820 16:56:16.732000 880870 torch/distributed/run.py:792] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.
	0: W0820 16:56:16.732000 880870 torch/distributed/run.py:792] *****************************************
	3: W0820 16:56:16.836000 271178 torch/distributed/run.py:792]
	3: W0820 16:56:16.836000 271178 torch/distributed/run.py:792] *****************************************
	3: W0820 16:56:16.836000 271178 torch/distributed/run.py:792] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.
	3: W0820 16:56:16.836000 271178 torch/distributed/run.py:792] *****************************************
	2: W0820 16:56:16.841000 1571176 torch/distributed/run.py:792]
	2: W0820 16:56:16.841000 1571176 torch/distributed/run.py:792] *****************************************
	2: W0820 16:56:16.841000 1571176 torch/distributed/run.py:792] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.
	2: W0820 16:56:16.841000 1571176 torch/distributed/run.py:792] *****************************************
	2: [2025-08-20 16:56:37,848] [INFO] [axolotl.utils.schemas.validation.check_eval_packing:118] [PID:1571252] [RANK:0] explicitly setting `eval_sample_packing` to match `sample_packing`[39m
	2: [2025-08-20 16:56:37,848] [INFO] [axolotl.utils.schemas.validation.hint_sample_packing_padding:217] [PID:1571252] [RANK:0] Setting `pad_to_sequence_len: true` to prevent memory leaks when sample_packing[39m
	1: [2025-08-20 16:56:38,039] [INFO] [axolotl.utils.schemas.validation.check_eval_packing:118] [PID:1497738] [RANK:0] explicitly setting `eval_sample_packing` to match `sample_packing`[39m
	1: [2025-08-20 16:56:38,039] [INFO] [axolotl.utils.schemas.validation.hint_sample_packing_padding:217] [PID:1497738] [RANK:0] Setting `pad_to_sequence_len: true` to prevent memory leaks when sample_packing[39m
	0: [2025-08-20 16:56:38,110] [INFO] [axolotl.utils.schemas.validation.check_eval_packing:118] [PID:880949] [RANK:0] explicitly setting `eval_sample_packing` to match `sample_packing`[39m
	0: [2025-08-20 16:56:38,111] [INFO] [axolotl.utils.schemas.validation.hint_sample_packing_padding:217] [PID:880949] [RANK:0] Setting `pad_to_sequence_len: true` to prevent memory leaks when sample_packing[39m
	3: [2025-08-20 16:56:38,340] [INFO] [axolotl.utils.schemas.validation.check_eval_packing:118] [PID:271254] [RANK:0] explicitly setting `eval_sample_packing` to match `sample_packing`[39m
	3: [2025-08-20 16:56:38,340] [INFO] [axolotl.utils.schemas.validation.hint_sample_packing_padding:217] [PID:271254] [RANK:0] Setting `pad_to_sequence_len: true` to prevent memory leaks when sample_packing[39m
	0: [2025-08-20 16:56:40,670] [INFO] [axolotl.utils.data.sft._load_raw_datasets:310] [PID:880950] [RANK:1] Loading raw datasets...[39m
	0: [2025-08-20 16:56:41,415] [INFO] [axolotl.utils.data.wrappers.get_dataset_wrapper:88] [PID:880950] [RANK:1] Loading dataset: /lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking with base_type: chat_template and prompt_style: None[39m
	0: [2025-08-20 16:56:41,708] [INFO] [axolotl.cli.config.load_cfg:244] [PID:880949] [RANK:0] config:
	0: {
	0: "activation_offloading": false,
	0: "auto_resume_from_checkpoints": true,
	0: "axolotl_config_path": "/lustre/fswork/projects/rech/dgo/udv55np/train/tmp/1755698958355263805.yaml",
	0: "base_model": "/lustre/fswork/projects/rech/qwv/udv55np/Qwen/Qwen2.5-0.5B",
	0: "base_model_config": "/lustre/fswork/projects/rech/qwv/udv55np/Qwen/Qwen2.5-0.5B",
	0: "batch_size": 16,
	0: "bf16": true,
	0: "capabilities": {
	0: "bf16": true,
	0: "compute_capability": "sm_90",
	0: "fp8": false,
	0: "n_gpu": 16,
	0: "n_node": 1
	0: },
	0: "chat_template": "qwen_25",
	0: "dataloader_num_workers": 16,
	0: "dataloader_pin_memory": true,
	0: "dataloader_prefetch_factor": 256,
	0: "dataset_prepared_path": "/lustre/fsn1/projects/rech/dgo/udv55np/dataset/Qwen3-235B-A22B/Qwen2.5-0.5B/mix_0",
	0: "dataset_processes": 192,
	0: "datasets": [
	0: {
	0: "chat_template": "tokenizer_default",
	0: "data_files": [
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0007.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0009.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0005.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0006.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0014.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0010.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0012.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0008.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0001.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0002.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0013.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0015.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0004.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0011.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0000.jsonl",
	0: "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking/0003.jsonl"
	0: ],
	0: "ds_type": "json",
	0: "field_messages": "conversations",
	0: "message_property_mappings": {
	0: "content": "content",
	0: "role": "role"
	0: },
	0: "path": "/lustre/fswork/projects/rech/qwv/udv55np/dataset/ift/Qwen3-235B-A22B/no_thinking",
	0: "trust_remote_code": false,
	0: "type": "chat_template"
	0: }
	0: ],
	0: "ddp": true,
	0: "deepspeed": {
	0: "bf16": {
	0: "enabled": true
	0: },
	0: "gradient_accumulation_steps": "auto",
	0: "gradient_clipping": "auto",
	0: "train_batch_size": "auto",
	0: "train_micro_batch_size_per_gpu": "auto",
	0: "wall_clock_breakdown": false,
	0: "zero_optimization": {
	0: "contiguous_gradients": true,
	0: "overlap_comm": true,
	0: "reduce_bucket_size": "auto",
	0: "stage": 3,
	0: "stage3_gather_16bit_weights_on_model_save": true,
	0: "stage3_param_persistence_threshold": "auto",
	0: "stage3_prefetch_bucket_size": "auto",
	0: "sub_group_size": 0
	0: }
	0: },
	0: "device": "cuda:0",
	0: "device_map": {
	0: "": 0
	0: },
	0: "env_capabilities": {
	0: "torch_version": "2.6.0"
	0: },
	0: "eval_batch_size": 1,
	0: "eval_causal_lm_metrics": [
	0: "sacrebleu",
	0: "comet",
	0: "ter",
	0: "chrf"
	0: ],
	0: "eval_max_new_tokens": 128,
	0: "eval_sample_packing": true,
	0: "eval_table_size": 0,
	0: "evals_per_epoch": 0,
	0: "flash_attention": true,
	0: "fp16": false,
	0: "gradient_accumulation_steps": 1,
	0: "gradient_checkpointing": true,
	0: "gradient_checkpointing_kwargs": {
	0: "use_reentrant": true
	0: },
	0: "learning_rate": 1e-05,
	0: "lisa_layers_attribute": "model.layers",
	0: "load_best_model_at_end": false,
	0: "load_in_4bit": false,
	0: "load_in_8bit": false,
	0: "local_rank": 0,
	0: "logging_steps": 10,
	0: "lora_dropout": 0.0,
	0: "loraplus_lr_embedding": 1e-06,
	0: "lr_scheduler": "warmup_stable_decay",
	0: "lr_scheduler_kwargs": {
	0: "min_lr_ratio": 0.1,
	0: "num_decay_steps": 300
	0: },
	0: "max_prompt_len": 512,
	0: "mean_resizing_embeddings": false,
	0: "micro_batch_size": 1,
	0: "model_config_type": "qwen2",
	0: "num_epochs": 1.0,
	0: "optimizer": "adamw_torch_fused",
	0: "output_dir": "/lustre/fswork/projects/rech/dgo/udv55np/ift/Qwen3-235B-A22B/Qwen2.5-0.5B/0",
	0: "pad_to_sequence_len": true,
	0: "pretrain_multipack_attn": true,
	0: "pretrain_multipack_buffer_size": 10000,
	0: "profiler_steps_start": 0,
	0: "qlora_sharded_model_loading": false,
	0: "ray_num_workers": 1,
	0: "resources_per_worker": {
	0: "GPU": 1
	0: },
	0: "sample_packing": true,
	0: "sample_packing_bin_size": 200,
	0: "sample_packing_group_size": 100000,
	0: "save_only_model": false,
	0: "save_safetensors": true,
	0: "save_steps": 0.2,
	0: "save_total_limit": 20,
	0: "sequence_len": 16384,
	0: "sequence_parallel_degree": 1,
	0: "shuffle_merged_datasets": true,
	0: "skip_prepare_dataset": false,
	0: "special_tokens": {
	0: "bos_token": "<\|im_start\|>",
	0: "eos_token": "<\|im_end\|>",
	0: "pad_token": "<\|endoftext\|>"
	0: },
	0: "strict": false,
	0: "tensor_parallel_size": 1,
	0: "tf32": false,
	0: "tiled_mlp_use_original_mlp": true,
	0: "tokenizer_config": "/lustre/fswork/projects/rech/qwv/udv55np/Qwen/Qwen2.5-0.5B",
	0: "torch_dtype": "torch.bfloat16",
	0: "train_on_inputs": false,
	0: "trl": {
	0: "log_completions": false,
	0: "mask_truncated_completions": false,
	0: "ref_model_mixup_alpha": 0.9,
	0: "ref_model_sync_steps": 64,
	0: "scale_rewards": true,
	0: "sync_ref_model": false,
	0: "use_vllm": false,
	0: "vllm_server_host": "0.0.0.0",
	0: "vllm_server_port": 8000
	0: },
	0: "use_ray": false,
	0: "use_tensorboard": true,
	0: "val_set_size": 0.0,
	0: "vllm": {
	0: "device": "auto",
	0: "dtype": "auto",
	0: "gpu_memory_utilization": 0.9,
	0: "host": "0.0.0.0",
	0: "port": 8000
	0: },
	0: "warmup_steps": 150,
	0: "weight_decay": 0.0,
	0: "world_size": 16
	0: }[39m
	0: [2025-08-20 16:56:41,710] [INFO] [axolotl.cli.checks.check_user_token:35] [PID:880949] [RANK:0] Skipping HuggingFace token verification because HF_HUB_OFFLINE is set to True. Only local files will be used.[39m
	0: Dropping Long Sequences (>16384) (num_proc=192): 0%\| \| 0/1393784 [00:00<?, ? examples/s] Dropping Long Sequences (>16384) (num_proc=192): 0%\| \| 1000/1393784 [00:01<35:49, 647.85 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 1%\|▏ \| 20000/1393784 [00:01<01:22, 16610.26 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 3%\|▎ \| 38000/1393784 [00:01<00:39, 34040.49 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 4%\|▍ \| 58000/1393784 [00:01<00:23, 55729.77 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 5%\|▌ \| 75000/1393784 [00:01<00:17, 73382.29 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 7%\|▋ \| 92000/1393784 [00:02<00:14, 88963.31 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 8%\|▊ \| 108000/1393784 [00:02<00:12, 100525.00 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 9%\|▉ \| 124000/1393784 [00:02<00:11, 112354.97
	0: examples/s] Dropping Long Sequences (>16384) (num_proc=192): 10%\|█ \| 140000/1393784 [00:02<00:10, 121259.64 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 11%\|█▏ \| 157000/1393784 [00:02<00:09, 132246.43 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 12%\|█▏ \| 174000/1393784 [00:02<00:08, 139321.12 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 14%\|█▎ \| 191000/1393784 [00:02<00:08, 146007.03 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 17%\|█▋ \| 242000/1393784 [00:02<00:04, 243494.26 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 21%\|██ \| 295000/1393784 [00:02<00:03, 322748.63 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 25%\|██▍ \| 342000/1393784 [00:03<00:02, 359443.89 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 33%\|███▎ \| 455780/1393784 [00:03<00:01, 582274.31 examples/s] Dropping Long Sequences (>16384) (num_pr
	0: oc=192): 40%\|███▉ \| 551820/1393784 [00:03<00:01, 691185.18 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 45%\|████▍ \| 623600/1393784 [00:03<00:01, 576400.08 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 49%\|████▉ \| 686420/1393784 [00:03<00:01, 487306.27 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 53%\|█████▎ \| 740460/1393784 [00:03<00:01, 492542.70 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 57%\|█████▋ \| 793580/1393784 [00:03<00:01, 456005.40 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 60%\|██████ \| 842620/1393784 [00:03<00:01, 435739.59 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 64%\|██████▍ \| 893920/1393784 [00:04<00:01, 448084.72 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 67%\|██████▋ \| 940480/1393784 [00:04<00:01, 440398.87 examples/s] Dropping Long Sequences (>16384) (nu
	0: m_proc=192): 71%\|███████ \| 985777/1393784 [00:04<00:00, 427743.44 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 74%\|███████▍ \| 1029331/1393784 [00:04<00:00, 405088.11 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 77%\|███████▋ \| 1078369/1393784 [00:04<00:00, 424379.65 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 81%\|████████ \| 1122516/1393784 [00:04<00:00, 422464.38 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 84%\|████████▎ \| 1166366/1393784 [00:04<00:00, 419730.36 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 87%\|████████▋ \| 1209510/1393784 [00:04<00:00, 414463.25 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 90%\|████████▉ \| 1252100/1393784 [00:04<00:00, 416483.55 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 93%\|█████████▎\| 1294021/1393784 [00:05<00:00, 395
	0: 480.06 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 96%\|█████████▌\| 1333942/1393784 [00:05<00:00, 365959.87 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 98%\|█████████▊\| 1371863/1393784 [00:05<00:00, 302760.23 examples/s] Dropping Long Sequences (>16384) (num_proc=192): 100%\|██████████\| 1393784/1393784 [00:05<00:00, 234154.83 examples/s]
	0: Drop Samples with Zero Trainable Tokens (num_proc=192): 0%\| \| 0/1393229 [00:00<?, ? examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 0%\| \| 1000/1393229 [00:02<50:25, 460.11 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 1%\| \| 10000/1393229 [00:02<03:53, 5936.20 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 1%\|▏ \| 18000/1393229 [00:02<01:57, 11688.24 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 2%\|▏ \| 24000/1393229 [00:02<01:23, 16404.24 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 2%\|▏ \| 32000/1393229 [00:02<00:57, 23801.82 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 3%\|▎ \| 43000/1393229 [00:02<00:37, 36428.10 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 4%\|▎ \| 51000/1393229 [00:02<00:30, 44018.44 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192):
	0: 4%\|▍ \| 60000/1393229 [00:02<00:25, 52964.31 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 5%\|▌ \| 73000/1393229 [00:03<00:19, 68476.84 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 6%\|▌ \| 86000/1393229 [00:03<00:15, 82477.73 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 7%\|▋ \| 102000/1393229 [00:03<00:12, 100557.74 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 8%\|▊ \| 115000/1393229 [00:03<00:11, 107583.80 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 9%\|▉ \| 129000/1393229 [00:03<00:11, 114110.88 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 10%\|█ \| 142000/1393229 [00:03<00:10, 116459.59 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 11%\|█ \| 155000/1393229 [00:03<00:10, 116347.96 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 12%\|█▏ \| 1
	0: 68000/1393229 [00:03<00:10, 118309.16 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 13%\|█▎ \| 181000/1393229 [00:03<00:09, 121411.68 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 14%\|█▍ \| 194000/1393229 [00:04<00:09, 120284.66 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 15%\|█▍ \| 207000/1393229 [00:04<00:10, 117740.48 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 16%\|█▌ \| 221000/1393229 [00:04<00:09, 120466.22 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 17%\|█▋ \| 234000/1393229 [00:04<00:09, 117526.73 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 18%\|█▊ \| 247000/1393229 [00:04<00:09, 120255.75 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 20%\|█▉ \| 275000/1393229 [00:04<00:06, 165223.39 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 27%\|██▋ \|
	0: 381000/1393229 [00:04<00:02, 420296.44 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 31%\|███ \| 425000/1393229 [00:04<00:02, 362665.16 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 33%\|███▎ \| 464000/1393229 [00:04<00:03, 303122.99 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 36%\|███▌ \| 498000/1393229 [00:05<00:02, 306727.35 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 38%\|███▊ \| 536000/1393229 [00:05<00:02, 319634.18 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 41%\|████ \| 570000/1393229 [00:05<00:02, 306649.64 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 43%\|████▎ \| 602000/1393229 [00:05<00:02, 287107.66 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 46%\|████▌ \| 636000/1393229 [00:05<00:02, 299362.86 examples/s] Drop Samples with Zero Trainable Tokens (num_p
	0: roc=192): 48%\|████▊ \| 667000/1393229 [00:05<00:02, 297606.28 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 50%\|█████ \| 700000/1393229 [00:05<00:02, 299908.66 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 53%\|█████▎ \| 741000/1393229 [00:05<00:01, 329982.47 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 56%\|█████▌ \| 777000/1393229 [00:05<00:01, 337493.04 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 58%\|█████▊ \| 812000/1393229 [00:06<00:01, 318276.65 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 61%\|██████ \| 850000/1393229 [00:06<00:01, 328251.90 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 63%\|██████▎ \| 883257/1393229 [00:06<00:01, 328594.35 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 67%\|██████▋ \| 927028/1393229 [00:06<00:01, 3579
	0: 51.22 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 69%\|██████▉ \| 964855/1393229 [00:06<00:01, 362652.76 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 72%\|███████▏ \| 1002196/1393229 [00:06<00:01, 351948.04 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 75%\|███████▌ \| 1045995/1393229 [00:06<00:00, 372957.80 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 78%\|███████▊ \| 1084107/1393229 [00:06<00:00, 374183.19 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 81%\|████████ \| 1126191/1393229 [00:06<00:00, 387411.68 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 84%\|████████▎ \| 1165273/1393229 [00:07<00:00, 375284.68 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 87%\|████████▋ \| 1207861/1393229 [00:07<00:00, 386771.76 examples/s] Drop Samples with
	0: Zero Trainable Tokens (num_proc=192): 90%\|████████▉ \| 1246981/1393229 [00:07<00:00, 358485.92 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 92%\|█████████▏\| 1286797/1393229 [00:07<00:00, 366308.44 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 95%\|█████████▌\| 1323917/1393229 [00:07<00:00, 367296.65 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 98%\|█████████▊\| 1361037/1393229 [00:07<00:00, 309401.49 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=192): 100%\|██████████\| 1393229/1393229 [00:08<00:00, 169350.94 examples/s]
	0: Add position_id column (Sample Packing) (num_proc=192): 0%\| \| 0/1393229 [00:00<?, ? examples/s] Add position_id column (Sample Packing) (num_proc=192): 0%\| \| 1000/1393229 [00:02<49:38, 467.35 examples/s] Add position_id column (Sample Packing) (num_proc=192): 1%\| \| 16000/1393229 [00:02<02:20, 9800.62 examples/s] Add position_id column (Sample Packing) (num_proc=192): 2%\|▏ \| 32000/1393229 [00:02<01:01, 22064.08 examples/s] Add position_id column (Sample Packing) (num_proc=192): 3%\|▎ \| 46000/1393229 [00:02<00:39, 34273.48 examples/s] Add position_id column (Sample Packing) (num_proc=192): 4%\|▍ \| 59000/1393229 [00:02<00:28, 46141.28 examples/s] Add position_id column (Sample Packing) (num_proc=192): 5%\|▌ \| 73000/1393229 [00:02<00:21, 60644.89 examples/s] Add position_id column (Sample Packing) (num_proc=192): 6%\|▌ \| 86000/1393229 [00:02<00:18, 69759.50 examples/s] Add position_id column (Sample Packing) (num_proc=192):
	0: 7%\|▋ \| 99000/1393229 [00:02<00:15, 81007.59 examples/s] Add position_id column (Sample Packing) (num_proc=192): 9%\|▊ \| 121000/1393229 [00:02<00:11, 111222.79 examples/s] Add position_id column (Sample Packing) (num_proc=192): 10%\|█ \| 143000/1393229 [00:03<00:09, 136046.10 examples/s] Add position_id column (Sample Packing) (num_proc=192): 12%\|█▏ \| 172000/1393229 [00:03<00:07, 174130.85 examples/s] Add position_id column (Sample Packing) (num_proc=192): 16%\|█▌ \| 220000/1393229 [00:03<00:04, 254733.96 examples/s] Add position_id column (Sample Packing) (num_proc=192): 18%\|█▊ \| 249000/1393229 [00:03<00:04, 236184.91 examples/s] Add position_id column (Sample Packing) (num_proc=192): 20%\|█▉ \| 276000/1393229 [00:03<00:04, 232808.84 examples/s] Add position_id column (Sample Packing) (num_proc=192): 22%\|██▏ \| 302000/1393229 [00:03<00:04, 228488.25 examples/s] Add position_id column (Sample Packing) (num_proc=192): 23%\|�
	0: ��█▎ \| 327000/1393229 [00:03<00:05, 203885.65 examples/s] Add position_id column (Sample Packing) (num_proc=192): 26%\|██▌ \| 358000/1393229 [00:03<00:04, 227576.28 examples/s] Add position_id column (Sample Packing) (num_proc=192): 28%\|██▊ \| 393000/1393229 [00:04<00:03, 259055.38 examples/s] Add position_id column (Sample Packing) (num_proc=192): 30%\|███ \| 421000/1393229 [00:04<00:03, 246904.86 examples/s] Add position_id column (Sample Packing) (num_proc=192): 32%\|███▏ \| 447000/1393229 [00:04<00:04, 235228.88 examples/s] Add position_id column (Sample Packing) (num_proc=192): 34%\|███▍ \| 472000/1393229 [00:04<00:03, 233225.59 examples/s] Add position_id column (Sample Packing) (num_proc=192): 36%\|███▌ \| 496000/1393229 [00:04<00:03, 230270.29 examples/s] Add position_id column (Sample Packing) (num_proc=192): 37%\|███▋ \| 520000/1393229 [00:04<00:04, 209309.78 examples/s] Add position_id column (Sample Packing
	0: ) (num_proc=192): 39%\|███▉ \| 548000/1393229 [00:04<00:03, 226028.22 examples/s] Add position_id column (Sample Packing) (num_proc=192): 42%\|████▏ \| 580000/1393229 [00:04<00:03, 246117.66 examples/s] Add position_id column (Sample Packing) (num_proc=192): 43%\|████▎ \| 606000/1393229 [00:04<00:03, 246544.10 examples/s] Add position_id column (Sample Packing) (num_proc=192): 45%\|████▌ \| 632000/1393229 [00:05<00:03, 250155.04 examples/s] Add position_id column (Sample Packing) (num_proc=192): 47%\|████▋ \| 658000/1393229 [00:05<00:02, 246068.08 examples/s] Add position_id column (Sample Packing) (num_proc=192): 49%\|████▉ \| 683000/1393229 [00:05<00:03, 226019.02 examples/s] Add position_id column (Sample Packing) (num_proc=192): 51%\|█████ \| 707000/1393229 [00:05<00:03, 207986.95 examples/s] Add position_id column (Sample Packing) (num_proc=192): 52%\|█████▏ \| 729000/1393229 [00:05<00:03, 207471.96 ex
	0: amples/s] Add position_id column (Sample Packing) (num_proc=192): 54%\|█████▍ \| 759000/1393229 [00:05<00:02, 230257.22 examples/s] Add position_id column (Sample Packing) (num_proc=192): 57%\|█████▋ \| 793000/1393229 [00:05<00:02, 255681.67 examples/s] Add position_id column (Sample Packing) (num_proc=192): 59%\|█████▉ \| 821000/1393229 [00:05<00:02, 262384.91 examples/s] Add position_id column (Sample Packing) (num_proc=192): 61%\|██████ \| 848000/1393229 [00:05<00:02, 247397.87 examples/s] Add position_id column (Sample Packing) (num_proc=192): 63%\|██████▎ \| 874000/1393229 [00:06<00:02, 236204.67 examples/s] Add position_id column (Sample Packing) (num_proc=192): 64%\|██████▍ \| 898000/1393229 [00:06<00:02, 218983.16 examples/s] Add position_id column (Sample Packing) (num_proc=192): 66%\|██████▌ \| 921000/1393229 [00:06<00:02, 214581.84 examples/s] Add position_id column (Sample Packing) (num_proc=192):
	0: 68%\|██████▊ \| 947000/1393229 [00:06<00:01, 226506.86 examples/s] Add position_id column (Sample Packing) (num_proc=192): 70%\|███████ \| 980000/1393229 [00:06<00:01, 252706.85 examples/s] Add position_id column (Sample Packing) (num_proc=192): 72%\|███████▏ \| 1006000/1393229 [00:06<00:01, 248273.15 examples/s] Add position_id column (Sample Packing) (num_proc=192): 74%\|███████▍ \| 1032000/1393229 [00:06<00:01, 240775.09 examples/s] Add position_id column (Sample Packing) (num_proc=192): 76%\|███████▌ \| 1059028/1393229 [00:06<00:01, 244445.80 examples/s] Add position_id column (Sample Packing) (num_proc=192): 78%\|███████▊ \| 1083855/1393229 [00:07<00:01, 230992.61 examples/s] Add position_id column (Sample Packing) (num_proc=192): 80%\|███████▉ \| 1107911/1393229 [00:07<00:01, 226200.76 examples/s] Add position_id column (Sample Packing) (num_proc=192): 82%\|████████▏ \| 1137453/
	0: 1393229 [00:07<00:01, 244710.67 examples/s] Add position_id column (Sample Packing) (num_proc=192): 84%\|████████▍ \| 1172251/1393229 [00:07<00:00, 272431.48 examples/s] Add position_id column (Sample Packing) (num_proc=192): 86%\|████████▋ \| 1204877/1393229 [00:07<00:00, 287615.86 examples/s] Add position_id column (Sample Packing) (num_proc=192): 89%\|████████▊ \| 1234268/1393229 [00:07<00:00, 280443.08 examples/s] Add position_id column (Sample Packing) (num_proc=192): 91%\|█████████ \| 1263344/1393229 [00:07<00:00, 280766.97 examples/s] Add position_id column (Sample Packing) (num_proc=192): 93%\|█████████▎\| 1292005/1393229 [00:07<00:00, 253167.65 examples/s] Add position_id column (Sample Packing) (num_proc=192): 95%\|█████████▍\| 1320589/1393229 [00:07<00:00, 260423.92 examples/s] Add position_id column (Sample Packing) (num_proc=192): 97%\|█████████▋\| 1347197/1393229 [00:08<00
	0: :00, 242331.70 examples/s] Add position_id column (Sample Packing) (num_proc=192): 98%\|█████████▊\| 1372085/1393229 [00:08<00:00, 208139.65 examples/s] Add position_id column (Sample Packing) (num_proc=192): 100%\|██████████\| 1393229/1393229 [00:08<00:00, 158552.36 examples/s]
	0: Saving the dataset (0/192 shards): 0%\| \| 0/1393229 [00:00<?, ? examples/s] Saving the dataset (0/192 shards): 0%\| \| 4000/1393229 [00:02<11:35, 1998.84 examples/s] Saving the dataset (1/192 shards): 7%\|▋ \| 95514/1393229 [00:02<10:49, 1998.84 examples/s] Saving the dataset (2/192 shards): 7%\|▋ \| 101514/1393229 [00:02<10:46, 1998.84 examples/s] Saving the dataset (3/192 shards): 9%\|▊ \| 120542/1393229 [00:02<10:36, 1998.84 examples/s] Saving the dataset (4/192 shards): 9%\|▉ \| 123542/1393229 [00:02<10:35, 1998.84 examples/s] Saving the dataset (5/192 shards): 9%\|▉ \| 123542/1393229 [00:02<10:35, 1998.84 examples/s] Saving the dataset (6/192 shards): 9%\|▉ \| 125542/1393229 [00:02<10:34, 1998.84 examples/s] Saving the dataset (7/192 shards): 11%\|█ \| 152056/1393229 [00:02<10:20, 1998.84 examples/s] Saving the dataset (8/192 shards): 11%\|█ \| 154313/1393229 [00:02<10:19, 1998.84 examples/s] Saving the data
	0: set (9/192 shards): 12%\|█▏ \| 163313/1393229 [00:02<10:15, 1998.84 examples/s] Saving the dataset (10/192 shards): 13%\|█▎ \| 180827/1393229 [00:02<10:06, 1998.84 examples/s] Saving the dataset (11/192 shards): 13%\|█▎ \| 181084/1393229 [00:02<10:06, 1998.84 examples/s] Saving the dataset (12/192 shards): 14%\|█▎ \| 189084/1393229 [00:02<10:02, 1998.84 examples/s] Saving the dataset (13/192 shards): 16%\|█▌ \| 220855/1393229 [00:02<09:46, 1998.84 examples/s] Saving the dataset (14/192 shards): 16%\|█▌ \| 220855/1393229 [00:02<09:46, 1998.84 examples/s] Saving the dataset (15/192 shards): 17%\|█▋ \| 238369/1393229 [00:02<09:37, 1998.84 examples/s] Saving the dataset (16/192 shards): 17%\|█▋ \| 238369/1393229 [00:02<09:37, 1998.84 examples/s] Saving the dataset (17/192 shards): 17%\|█▋ \| 238626/1393229 [00:02<09:37, 1998.84 examples/s] Saving the dataset (18/192 shards): 17%\|█▋ \| 240626/1393229 [00:02<09:36,
	0: 1998.84 examples/s] Saving the dataset (19/192 shards): 18%\|█▊ \| 246883/1393229 [00:02<09:33, 1998.84 examples/s] Saving the dataset (20/192 shards): 20%\|█▉ \| 272911/1393229 [00:02<09:20, 1998.84 examples/s] Saving the dataset (21/192 shards): 20%\|█▉ \| 272911/1393229 [00:02<09:20, 1998.84 examples/s] Saving the dataset (22/192 shards): 20%\|█▉ \| 272911/1393229 [00:02<09:20, 1998.84 examples/s] Saving the dataset (23/192 shards): 21%\|██ \| 286168/1393229 [00:02<09:13, 1998.84 examples/s] Saving the dataset (24/192 shards): 21%\|██ \| 288168/1393229 [00:02<09:12, 1998.84 examples/s] Saving the dataset (25/192 shards): 23%\|██▎ \| 315682/1393229 [00:02<08:59, 1998.84 examples/s] Saving the dataset (26/192 shards): 23%\|██▎ \| 323939/1393229 [00:02<08:54, 1998.84 examples/s] Saving the dataset (27/192 shards): 23%\|██▎ \| 326453/1393229 [00:02<08:53, 1998.84 examples/s] Saving the dataset (28/192 shards): 23%\|█
	0: █▎ \| 326453/1393229 [00:02<08:53, 1998.84 examples/s] Saving the dataset (29/192 shards): 24%\|██▍ \| 340967/1393229 [00:02<08:46, 1998.84 examples/s] Saving the dataset (30/192 shards): 24%\|██▍ \| 340967/1393229 [00:02<08:46, 1998.84 examples/s] Saving the dataset (31/192 shards): 25%\|██▍ \| 348224/1393229 [00:02<08:42, 1998.84 examples/s] Saving the dataset (32/192 shards): 25%\|██▍ \| 348224/1393229 [00:02<08:42, 1998.84 examples/s] Saving the dataset (33/192 shards): 27%\|██▋ \| 374509/1393229 [00:02<08:29, 1998.84 examples/s] Saving the dataset (34/192 shards): 27%\|██▋ \| 377509/1393229 [00:02<08:28, 1998.84 examples/s] Saving the dataset (35/192 shards): 27%\|██▋ \| 382509/1393229 [00:02<08:25, 1998.84 examples/s] Saving the dataset (36/192 shards): 28%\|██▊ \| 385509/1393229 [00:02<08:24, 1998.84 examples/s] Saving the dataset (37/192 shards): 28%\|██▊ \| 386766/1393229 [00:02<08:23, 1998.84
	0: examples/s] Saving the dataset (38/192 shards): 28%\|██▊ \| 394023/1393229 [00:02<08:19, 1998.84 examples/s] Saving the dataset (39/192 shards): 29%\|██▉ \| 404023/1393229 [00:02<08:14, 1998.84 examples/s] Saving the dataset (40/192 shards): 30%\|██▉ \| 413280/1393229 [00:02<08:10, 1998.84 examples/s] Saving the dataset (41/192 shards): 30%\|███ \| 422051/1393229 [00:02<08:05, 1998.84 examples/s] Saving the dataset (42/192 shards): 31%\|███ \| 426308/1393229 [00:02<08:03, 1998.84 examples/s] Saving the dataset (43/192 shards): 31%\|███ \| 426308/1393229 [00:02<08:03, 1998.84 examples/s] Saving the dataset (44/192 shards): 31%\|███ \| 426308/1393229 [00:02<08:03, 1998.84 examples/s] Saving the dataset (45/192 shards): 32%\|███▏ \| 442079/1393229 [00:02<07:55, 1998.84 examples/s] Saving the dataset (46/192 shards): 32%\|███▏ \| 442079/1393229 [00:02<07:55, 1998.84 examples/s] Saving the dataset (47/192 shards):
	0: 32%\|███▏ \| 450079/1393229 [00:02<07:51, 1998.84 examples/s] Saving the dataset (48/192 shards): 34%\|███▍ \| 471107/1393229 [00:02<07:41, 1998.84 examples/s] Saving the dataset (49/192 shards): 34%\|███▍ \| 476364/1393229 [00:02<07:38, 1998.84 examples/s] Saving the dataset (50/192 shards): 34%\|███▍ \| 478364/1393229 [00:02<07:37, 1998.84 examples/s] Saving the dataset (51/192 shards): 34%\|███▍ \| 478364/1393229 [00:02<07:37, 1998.84 examples/s] Saving the dataset (52/192 shards): 35%\|███▍ \| 481364/1393229 [00:02<07:36, 1998.84 examples/s] Saving the dataset (53/192 shards): 36%\|███▌ \| 502878/1393229 [00:02<07:25, 1998.84 examples/s] Saving the dataset (54/192 shards): 36%\|███▋ \| 505392/1393229 [00:02<07:24, 1998.84 examples/s] Saving the dataset (55/192 shards): 37%\|███▋ \| 517392/1393229 [00:02<07:18, 1998.84 examples/s] Saving the dataset (56/192 shards): 37%\|███▋ \| 517392/13
	0: 93229 [00:02<07:18, 1998.84 examples/s] Saving the dataset (57/192 shards): 38%\|███▊ \| 531649/1393229 [00:02<07:11, 1998.84 examples/s] Saving the dataset (58/192 shards): 40%\|███▉ \| 556420/1393229 [00:02<06:58, 1998.84 examples/s] Saving the dataset (59/192 shards): 40%\|███▉ \| 556420/1393229 [00:02<06:58, 1998.84 examples/s] Saving the dataset (60/192 shards): 40%\|████ \| 560420/1393229 [00:02<06:56, 1998.84 examples/s] Saving the dataset (61/192 shards): 42%\|████▏ \| 584448/1393229 [00:02<06:44, 1998.84 examples/s] Saving the dataset (62/192 shards): 42%\|████▏ \| 588448/1393229 [00:02<06:42, 1998.84 examples/s] Saving the dataset (63/192 shards): 42%\|████▏ \| 588705/1393229 [00:02<06:42, 1998.84 examples/s] Saving the dataset (64/192 shards): 42%\|████▏ \| 590705/1393229 [00:02<06:41, 1998.84 examples/s] Saving the dataset (65/192 shards): 43%\|████▎ \| 602962/1393229 [00:02<06:35, 1998
	0: .84 examples/s] Saving the dataset (66/192 shards): 44%\|████▍ \| 611219/1393229 [00:02<06:31, 1998.84 examples/s] Saving the dataset (67/192 shards): 44%\|████▍ \| 618219/1393229 [00:02<06:27, 1998.84 examples/s] Saving the dataset (68/192 shards): 46%\|████▋ \| 645247/1393229 [00:02<06:14, 1998.84 examples/s] Saving the dataset (69/192 shards): 46%\|████▋ \| 645247/1393229 [00:02<06:14, 1998.84 examples/s] Saving the dataset (70/192 shards): 47%\|████▋ \| 650504/1393229 [00:02<06:11, 1998.84 examples/s] Saving the dataset (71/192 shards): 47%\|████▋ \| 650504/1393229 [00:02<06:11, 1998.84 examples/s] Saving the dataset (72/192 shards): 47%\|████▋ \| 656761/1393229 [00:02<06:08, 1998.84 examples/s] Saving the dataset (73/192 shards): 48%\|████▊ \| 665761/1393229 [00:02<06:03, 1998.84 examples/s] Saving the dataset (74/192 shards): 50%\|█████ \| 697018/1393229 [00:02<05:48, 1998.84 examples/s]
	0: Saving the dataset (75/192 shards): 52%\|█████▏ \| 728275/1393229 [00:02<05:32, 1998.84 examples/s] Saving the dataset (76/192 shards): 53%\|█████▎ \| 743044/1393229 [00:02<05:25, 1998.84 examples/s] Saving the dataset (77/192 shards): 54%\|█████▎ \| 747557/1393229 [00:02<05:23, 1998.84 examples/s] Saving the dataset (78/192 shards): 55%\|█████▍ \| 761069/1393229 [00:02<05:16, 1998.84 examples/s] Saving the dataset (79/192 shards): 55%\|█████▍ \| 763325/1393229 [00:02<05:15, 1998.84 examples/s] Saving the dataset (80/192 shards): 55%\|█████▍ \| 763325/1393229 [00:02<05:15, 1998.84 examples/s] Saving the dataset (81/192 shards): 55%\|█████▌ \| 767581/1393229 [00:02<05:13, 1998.84 examples/s] Saving the dataset (82/192 shards): 55%\|█████▌ \| 769581/1393229 [00:02<05:12, 1998.84 examples/s] Saving the dataset (83/192 shards): 55%\|█████▌ \| 769581/1393229 [00:02<05:12, 1998.84 examples/s
	0: ] Saving the dataset (84/192 shards): 57%\|█████▋ \| 788093/1393229 [00:02<05:02, 1998.84 examples/s] Saving the dataset (85/192 shards): 58%\|█████▊ \| 802605/1393229 [00:02<04:55, 1998.84 examples/s] Saving the dataset (86/192 shards): 58%\|█████▊ \| 806605/1393229 [00:02<04:53, 1998.84 examples/s] Saving the dataset (87/192 shards): 58%\|█████▊ \| 811861/1393229 [00:02<04:50, 1998.84 examples/s] Saving the dataset (88/192 shards): 59%\|█████▊ \| 817117/1393229 [00:02<04:48, 1998.84 examples/s] Saving the dataset (89/192 shards): 59%\|█████▉ \| 825373/1393229 [00:02<04:44, 1998.84 examples/s] Saving the dataset (90/192 shards): 59%\|█████▉ \| 827629/1393229 [00:02<04:42, 1998.84 examples/s] Saving the dataset (91/192 shards): 60%\|█████▉ \| 829629/1393229 [00:02<04:41, 1998.84 examples/s] Saving the dataset (92/192 shards): 60%\|██████ \| 841885/1393229 [00:02<04:35, 1998.84 examples
	0: /s] Saving the dataset (93/192 shards): 61%\|██████▏ \| 853397/1393229 [00:02<04:30, 1998.84 examples/s] Saving the dataset (94/192 shards): 61%\|██████▏ \| 853397/1393229 [00:02<04:30, 1998.84 examples/s] Saving the dataset (95/192 shards): 61%\|██████▏ \| 853397/1393229 [00:02<04:30, 1998.84 examples/s] Saving the dataset (96/192 shards): 63%\|██████▎ \| 880165/1393229 [00:02<04:16, 1998.84 examples/s] Saving the dataset (97/192 shards): 65%\|██████▍ \| 901701/1393229 [00:02<04:05, 1998.84 examples/s] Saving the dataset (98/192 shards): 65%\|██████▍ \| 901701/1393229 [00:02<04:05, 1998.84 examples/s] Saving the dataset (99/192 shards): 65%\|██████▍ \| 901701/1393229 [00:02<04:05, 1998.84 examples/s] Saving the dataset (100/192 shards): 65%\|██████▌ \| 907701/1393229 [00:02<04:02, 1998.84 examples/s] Saving the dataset (101/192 shards): 65%\|██████▌ \| 907701/1393229 [00:02<0
	0: 4:02, 1998.84 examples/s] Saving the dataset (102/192 shards): 65%\|██████▌ \| 909701/1393229 [00:02<04:01, 1998.84 examples/s] Saving the dataset (103/192 shards): 66%\|██████▌ \| 913701/1393229 [00:02<03:59, 1998.84 examples/s] Saving the dataset (104/192 shards): 66%\|██████▌ \| 919701/1393229 [00:02<03:56, 1998.84 examples/s] Saving the dataset (105/192 shards): 67%\|██████▋ \| 933213/1393229 [00:02<03:50, 1998.84 examples/s] Saving the dataset (106/192 shards): 67%\|██████▋ \| 934469/1393229 [00:02<03:49, 1998.84 examples/s] Saving the dataset (107/192 shards): 68%\|██████▊ \| 946981/1393229 [00:02<03:43, 1998.84 examples/s] Saving the dataset (107/192 shards): 69%\|██████▊ \| 954493/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (108/192 shards): 69%\|██████▉ \| 965493/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (109/192 shards): 69%\|█████�
	0: ��▉ \| 965749/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (110/192 shards): 69%\|██████▉ \| 967517/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (111/192 shards): 70%\|██████▉ \| 969517/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (112/192 shards): 70%\|██████▉ \| 974517/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (113/192 shards): 70%\|███████ \| 978517/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (114/192 shards): 70%\|███████ \| 978517/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (115/192 shards): 70%\|███████ \| 978517/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (116/192 shards): 72%\|███████▏ \| 999029/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (117/192 shards): 72%\|███████▏ \| 1001029/1393229 [00:02<00:00, 635171.48 examples/s] Saving
	0: the dataset (118/192 shards): 73%\|███████▎ \| 1021053/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (119/192 shards): 74%\|███████▍ \| 1034053/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (120/192 shards): 74%\|███████▍ \| 1034053/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (121/192 shards): 75%\|███████▍ \| 1042821/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (122/192 shards): 76%\|███████▌ \| 1052589/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (123/192 shards): 76%\|███████▌ \| 1055589/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (124/192 shards): 76%\|███████▌ \| 1060845/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (125/192 shards): 76%\|███████▋ \| 1062845/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (126/192 shards): 76%\|████
	0: ███▋ \| 1065101/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (127/192 shards): 76%\|███████▋ \| 1065101/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (128/192 shards): 77%\|███████▋ \| 1067613/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (129/192 shards): 77%\|███████▋ \| 1069613/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (130/192 shards): 77%\|███████▋ \| 1074869/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (131/192 shards): 78%\|███████▊ \| 1079869/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (132/192 shards): 78%\|███████▊ \| 1083125/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (133/192 shards): 78%\|███████▊ \| 1093381/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (134/192 shards): 79%\|███████▉ \| 1104405/1393229 [00:02<00:00, 63
	0: 5171.48 examples/s] Saving the dataset (135/192 shards): 79%\|███████▉ \| 1105661/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (136/192 shards): 80%\|███████▉ \| 1109917/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (137/192 shards): 80%\|███████▉ \| 1109917/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (138/192 shards): 80%\|███████▉ \| 1109917/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (139/192 shards): 80%\|███████▉ \| 1109917/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (140/192 shards): 80%\|███████▉ \| 1113917/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (141/192 shards): 81%\|████████ \| 1125173/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (142/192 shards): 81%\|████████ \| 1128685/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (143/192
	0: shards): 82%\|████████▏ \| 1137941/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (144/192 shards): 82%\|████████▏ \| 1143941/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (145/192 shards): 83%\|████████▎ \| 1151453/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (146/192 shards): 83%\|████████▎ \| 1155709/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (147/192 shards): 84%\|████████▍ \| 1167221/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (148/192 shards): 84%\|████████▍ \| 1170221/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (149/192 shards): 84%\|████████▍ \| 1170221/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (150/192 shards): 85%\|████████▍ \| 1182733/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (151/192 shards): 86%\|██████
	0: ██▌ \| 1191733/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (152/192 shards): 87%\|████████▋ \| 1206245/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (153/192 shards): 88%\|████████▊ \| 1221013/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (154/192 shards): 88%\|████████▊ \| 1221013/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (155/192 shards): 88%\|████████▊ \| 1226269/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (156/192 shards): 88%\|████████▊ \| 1232269/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (157/192 shards): 89%\|████████▉ \| 1237525/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (158/192 shards): 89%\|████████▉ \| 1243525/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (159/192 shards): 90%\|█████████ \| 1254549/1393229 [00:
	0: 02<00:00, 635171.48 examples/s] Saving the dataset (160/192 shards): 90%\|█████████ \| 1254549/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (161/192 shards): 90%\|█████████ \| 1254549/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (162/192 shards): 90%\|█████████ \| 1257549/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (163/192 shards): 91%\|█████████ \| 1265805/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (164/192 shards): 92%\|█████████▏\| 1280061/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (165/192 shards): 93%\|█████████▎\| 1294317/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (166/192 shards): 94%\|█████████▍\| 1312085/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (167/192 shards): 94%\|█████████▍\| 1312085/1393229 [00:02<00:00, 635171.48 exam
	0: ples/s] Saving the dataset (168/192 shards): 94%\|█████████▍\| 1312085/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (169/192 shards): 95%\|█████████▍\| 1322109/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (170/192 shards): 95%\|█████████▍\| 1322109/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (171/192 shards): 95%\|█████████▌\| 1326109/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (172/192 shards): 95%\|█████████▌\| 1326109/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (173/192 shards): 95%\|█████████▌\| 1330133/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (174/192 shards): 95%\|█████████▌\| 1330133/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (175/192 shards): 95%\|█████████▌\| 1330133/1393229 [00:02<00:00, 635171.48 examples/s] Saving t
	0: he dataset (176/192 shards): 96%\|█████████▌\| 1335389/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (177/192 shards): 96%\|█████████▌\| 1340645/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (178/192 shards): 97%\|█████████▋\| 1344901/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (179/192 shards): 97%\|█████████▋\| 1344901/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (180/192 shards): 97%\|█████████▋\| 1349157/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (181/192 shards): 97%\|█████████▋\| 1353413/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (182/192 shards): 97%\|█████████▋\| 1354925/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (183/192 shards): 97%\|█████████▋\| 1354925/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (184/
	0: 192 shards): 98%\|█████████▊\| 1363181/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (185/192 shards): 99%\|█████████▉\| 1383693/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (186/192 shards): 99%\|█████████▉\| 1383949/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (187/192 shards): 100%\|█████████▉\| 1388717/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (188/192 shards): 100%\|█████████▉\| 1388717/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (189/192 shards): 100%\|█████████▉\| 1388717/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (190/192 shards): 100%\|█████████▉\| 1388717/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (191/192 shards): 100%\|██████████\| 1393229/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (192/192 shards): 100
	0: %\|██████████\| 1393229/1393229 [00:02<00:00, 635171.48 examples/s] Saving the dataset (192/192 shards): 100%\|██████████\| 1393229/1393229 [00:02<00:00, 629771.92 examples/s]
	0: [2025-08-20 16:57:32,211] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:471] [PID:880949] [RANK:0] Loading prepared dataset from disk at /lustre/fsn1/projects/rech/dgo/udv55np/dataset/Qwen3-235B-A22B/Qwen2.5-0.5B/mix_0/b92082c2117369c7b34fbc8edf04b9ae...[39m
	0: [2025-08-20 16:58:40,864] [INFO] [axolotl.utils.samplers.multipack.calc_min_len:435] [PID:880949] [RANK:0] gather_len_batches: [47617, 47617, 47617, 47615, 47616, 47616, 47617, 47617, 47616, 47616, 47617, 47617, 47617, 47616, 47617, 47617][39m
	0: [2025-08-20 16:58:40,875] [INFO] [axolotl.utils.trainer.calc_sample_packing_eff_est:496] [PID:880949] [RANK:0] sample_packing_eff_est across ranks: [0.9988111257553101, 0.9987481832504272, 0.9987691640853882, 0.9987901449203491, 0.9987901449203491, 0.9987481832504272, 0.9987901449203491, 0.9987901449203491, 0.9987901449203491, 0.9987901449203491, 0.9987901449203491, 0.9987901449203491, 0.9987481832504272, 0.9988111257553101, 0.9988111257553101, 0.9987691640853882][39m
	0: [2025-08-20 16:58:40,891] [INFO] [axolotl.utils.data.sft._prepare_standard_dataset:123] [PID:880949] [RANK:0] Maximum number of steps set at 2975[39m
	1: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	3: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	2: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	3: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	3: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	0: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	2: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	2: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	3: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	2: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	1: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	1: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	0: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	1: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	0: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	0: You are attempting to use Flash Attention 2.0 with a model not initialized on GPU. Make sure to move the model to GPU after initializing it on CPU with `model.to('cuda')`.
	0: [2025-08-20 16:58:42,585] [INFO] [axolotl.loaders.model._configure_embedding_dtypes:317] [PID:880949] [RANK:0] Converting modules to torch.bfloat16[39m
	0: [2025-08-20 16:58:45,139] [INFO] [axolotl.train.save_initial_configs:397] [PID:880949] [RANK:0] Pre-saving tokenizer to /lustre/fswork/projects/rech/dgo/udv55np/ift/Qwen3-235B-A22B/Qwen2.5-0.5B/0...[39m
	0: [2025-08-20 16:58:45,304] [INFO] [axolotl.train.save_initial_configs:400] [PID:880949] [RANK:0] Pre-saving model config to /lustre/fswork/projects/rech/dgo/udv55np/ift/Qwen3-235B-A22B/Qwen2.5-0.5B/0...[39m
	0: [2025-08-20 16:58:45,312] [INFO] [axolotl.train.execute_training:221] [PID:880949] [RANK:0] Starting trainer...[39m
	0: [2025-08-20 17:02:39,390] [INFO] [axolotl.utils.samplers.multipack.calc_min_len:435] [PID:880949] [RANK:0] gather_len_batches: [47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616, 47616][39m
	0: Parameter Offload - Persistent parameters statistics: param_count = 121, numel = 71552
	0: {'loss': 1.1147, 'grad_norm': 3.6220792776073254, 'learning_rate': 1.54e-06, 'epoch': 0.0}
	0: 0%\| \| 0/2975 [00:00<?, ?it/s] 0%\| \| 1/2975 [03:46<187:02:36, 226.41s/it] 0%\| \| 2/2975 [03:51<79:16:39, 96.00s/it] 0%\| \| 3/2975 [03:52<43:29:41, 52.69s/it] 0%\| \| 4/2975 [03:53<26:41:06, 32.33s/it] 0%\| \| 5/2975 [03:54<17:14:35, 20.90s/it] 0%\| \| 6/2975 [03:54<11:32:47, 14.00s/it] 0%\| \| 7/2975 [03:55<7:55:53, 9.62s/it] 0%\| \| 8/2975 [03:55<5:33:50, 6.75s/it] 0%\| \| 9/2975 [03:56<3:58:42, 4.83s/it] 0%\| \| 10/2975 [03:57<2:54:10, 3.52s/it] 0%\| \| 10/2975 [03:57<2:54:10, 3.52s/it] 0%\| \| 11/2975 [03:57<2:10:00, 2.63s/it] 0%\| \| 12/2975 [03:58<1:39:29, 2.01s/it] 0%\| \| 13/2975 [03:58<1:18:20, 1.59s/it] 0%\| \| 14/2975 [03:59<1:03:36, 1.29s/it] 1%\| \| 15/2975 [04:00<53:23, 1.08s/it] 1%\| \| 16/2975 [04:00<46:14, 1.07it/s] 1%\| \| 17/2975 [04:01<41:15, 1.19i
	0: {'loss': 1.0079, 'grad_norm': 2.185977392937427, 'learning_rate': 2.1400000000000003e-06, 'epoch': 0.01}
	0: {'loss': 0.9488, 'grad_norm': 1.279183640961094, 'learning_rate': 2.7400000000000004e-06, 'epoch': 0.01}
	0: t/s] 1%\| \| 18/2975 [04:01<37:47, 1.30it/s] 1%\| \| 19/2975 [04:02<35:23, 1.39it/s] 1%\| \| 20/2975 [04:03<33:38, 1.46it/s] 1%\| \| 20/2975 [04:03<33:38, 1.46it/s] 1%\| \| 21/2975 [04:03<32:31, 1.51it/s] 1%\| \| 22/2975 [04:04<31:44, 1.55it/s] 1%\| \| 23/2975 [04:04<31:08, 1.58it/s] 1%\| \| 24/2975 [04:05<30:43, 1.60it/s] 1%\| \| 25/2975 [04:06<30:25, 1.62it/s] 1%\| \| 26/2975 [04:06<30:12, 1.63it/s] 1%\| \| 27/2975 [04:07<30:06, 1.63it/s] 1%\| \| 28/2975 [04:07<30:05, 1.63it/s] 1%\| \| 29/2975 [04:08<30:07, 1.63it/s] 1%\| \| 30/2975 [04:09<30:09, 1.63it/s] 1%\| \| 30/2975 [04:09<30:09, 1.63it/s] 1%\| \| 31/2975 [04:09<30:07, 1.63it/s] 1%\| \| 32/2975 [04:10<30:02, 1.63it/s] 1%\| \| 33/2975 [04:10<29:54, 1.64it/s] 1%\| \|
	0: {'loss': 0.9124, 'grad_norm': 1.2547741524546454, 'learning_rate': 3.3400000000000006e-06, 'epoch': 0.01}
	0: {'loss': 0.8961, 'grad_norm': 1.0408630202611184, 'learning_rate': 3.94e-06, 'epoch': 0.02}
	0: 34/2975 [04:11<29:47, 1.65it/s] 1%\| \| 35/2975 [04:12<29:46, 1.65it/s] 1%\| \| 36/2975 [04:12<29:43, 1.65it/s] 1%\| \| 37/2975 [04:13<29:41, 1.65it/s] 1%\|▏ \| 38/2975 [04:14<29:41, 1.65it/s] 1%\|▏ \| 39/2975 [04:14<29:40, 1.65it/s] 1%\|▏ \| 40/2975 [04:15<29:38, 1.65it/s] 1%\|▏ \| 40/2975 [04:15<29:38, 1.65it/s] 1%\|▏ \| 41/2975 [04:15<29:40, 1.65it/s] 1%\|▏ \| 42/2975 [04:16<29:38, 1.65it/s] 1%\|▏ \| 43/2975 [04:17<29:40, 1.65it/s] 1%\|▏ \| 44/2975 [04:17<29:46, 1.64it/s] 2%\|▏ \| 45/2975 [04:18<29:54, 1.63it/s] 2%\|▏ \| 46/2975 [04:18<29:57, 1.63it/s] 2%\|▏ \| 47/2975 [04:19<29:56, 1.63it/s] 2%\|▏ \| 48/2975 [04:20<29:53, 1.63it/s] 2%\|▏ \| 49/2975 [04:20<29:51, 1.63it/s] 2%\|▏ \| 50/2975 [04:21<29:45, 1.64it/s] 2%\|▏
	0: {'loss': 0.8844, 'grad_norm': 0.9864649886106432, 'learning_rate': 4.54e-06, 'epoch': 0.02}
	0: \| 50/2975 [04:21<29:45, 1.64it/s] 2%\|▏ \| 51/2975 [04:21<29:42, 1.64it/s] 2%\|▏ \| 52/2975 [04:22<29:36, 1.65it/s] 2%\|▏ \| 53/2975 [04:23<29:35, 1.65it/s] 2%\|▏ \| 54/2975 [04:23<29:36, 1.64it/s] 2%\|▏ \| 55/2975 [04:24<29:39, 1.64it/s] 2%\|▏ \| 56/2975 [04:25<29:39, 1.64it/s] 2%\|▏ \| 57/2975 [04:25<29:42, 1.64it/s] 2%\|▏ \| 58/2975 [04:26<29:48, 1.63it/s] 2%\|▏ \| 59/2975 [04:26<29:51, 1.63it/s] 2%\|▏ \| 60/2975 [04:27<29:47, 1.63it/s] 2%\|▏ \| 60/2975 [04:27<29:47, 1.63it/s] 2%\|▏ \| 61/2975 [04:28<29:45, 1.63it/s] 2%\|▏ \| 62/2975 [04:28<29:39, 1.64it/s] 2%\|▏ \| 63/2975 [04:29<29:34, 1.64it/s] 2%\|▏ \| 64/2975 [04:29<29:32, 1.64it/s] 2%\|▏ \| 65/2975 [04:30<29:30, 1.64it/s] 2%\|▏ \| 66/2975 [04:31<29:27, 1.65it/s] 2%\|▏ \| 67/2975 [04:31<29:24, 1.65it
	0: {'loss': 0.8461, 'grad_norm': 1.0171978797051726, 'learning_rate': 5.140000000000001e-06, 'epoch': 0.02}
	0: {'loss': 0.8403, 'grad_norm': 0.9604348470003533, 'learning_rate': 5.74e-06, 'epoch': 0.03}
	0: /s] 2%\|▏ \| 68/2975 [04:32<29:28, 1.64it/s] 2%\|▏ \| 69/2975 [04:32<29:35, 1.64it/s] 2%\|▏ \| 70/2975 [04:33<29:41, 1.63it/s] 2%\|▏ \| 70/2975 [04:33<29:41, 1.63it/s] 2%\|▏ \| 71/2975 [04:34<29:48, 1.62it/s] 2%\|▏ \| 72/2975 [04:34<29:46, 1.62it/s] 2%\|▏ \| 73/2975 [04:35<29:43, 1.63it/s] 2%\|▏ \| 74/2975 [04:36<29:40, 1.63it/s] 3%\|▎ \| 75/2975 [04:36<29:37, 1.63it/s] 3%\|▎ \| 76/2975 [04:37<29:42, 1.63it/s] 3%\|▎ \| 77/2975 [04:37<29:43, 1.62it/s] 3%\|▎ \| 78/2975 [04:38<29:46, 1.62it/s] 3%\|▎ \| 79/2975 [04:39<29:43, 1.62it/s] 3%\|▎ \| 80/2975 [04:39<29:41, 1.63it/s] 3%\|▎ \| 80/2975 [04:39<29:41, 1.63it/s] 3%\|▎ \| 81/2975 [04:40<29:41, 1.62it/s] 3%\|▎ \| 82/2975 [04:40<29:37, 1.63it/s] 3%\|▎ \| 83/2975 [04:41<
	0: {'loss': 0.845, 'grad_norm': 0.9465021140599814, 'learning_rate': 6.34e-06, 'epoch': 0.03}
	0: 29:37, 1.63it/s] 3%\|▎ \| 84/2975 [04:42<29:38, 1.63it/s] 3%\|▎ \| 85/2975 [04:42<29:37, 1.63it/s] 3%\|▎ \| 86/2975 [04:43<29:46, 1.62it/s] 3%\|▎ \| 87/2975 [04:44<29:46, 1.62it/s] 3%\|▎ \| 88/2975 [04:44<29:42, 1.62it/s] 3%\|▎ \| 89/2975 [04:45<29:38, 1.62it/s] 3%\|▎ \| 90/2975 [04:45<29:36, 1.62it/s] 3%\|▎ \| 90/2975 [04:45<29:36, 1.62it/s] 3%\|▎ \| 91/2975 [04:46<29:38, 1.62it/s] 3%\|▎ \| 92/2975 [04:47<29:32, 1.63it/s] 3%\|▎ \| 93/2975 [04:47<29:32, 1.63it/s] 3%\|▎ \| 94/2975 [04:48<29:33, 1.62it/s] 3%\|▎ \| 95/2975 [04:48<29:32, 1.62it/s] 3%\|▎ \| 96/2975 [04:49<29:29, 1.63it/s] 3%\|▎ \| 97/2975 [04:50<29:31, 1.62it/s] 3%\|▎ \| 98/2975 [04:50<29:33, 1.62it/s] 3%\|▎ \| 99/2975 [04:51<29:34, 1.62it/s] 3%\|▎ \| 100/2975 [04:52<29:51, 1.60it/s]
	0: {'loss': 0.8482, 'grad_norm': 1.1297069466960947, 'learning_rate': 6.940000000000001e-06, 'epoch': 0.03}
	0: {'loss': 0.8297, 'grad_norm': 0.9648123456744243, 'learning_rate': 7.540000000000001e-06, 'epoch': 0.04}
	0: 3%\|▎ \| 100/2975 [04:52<29:51, 1.60it/s] 3%\|▎ \| 101/2975 [04:52<29:47, 1.61it/s] 3%\|▎ \| 102/2975 [04:53<29:42, 1.61it/s] 3%\|▎ \| 103/2975 [04:53<29:40, 1.61it/s] 3%\|▎ \| 104/2975 [04:54<29:35, 1.62it/s] 4%\|▎ \| 105/2975 [04:55<29:26, 1.63it/s] 4%\|▎ \| 106/2975 [04:55<29:20, 1.63it/s] 4%\|▎ \| 107/2975 [04:56<29:16, 1.63it/s] 4%\|▎ \| 108/2975 [04:56<29:14, 1.63it/s] 4%\|▎ \| 109/2975 [04:57<29:13, 1.63it/s] 4%\|▎ \| 110/2975 [04:58<29:11, 1.64it/s] 4%\|▎ \| 110/2975 [04:58<29:11, 1.64it/s] 4%\|▎ \| 111/2975 [04:58<29:11, 1.64it/s] 4%\|▍ \| 112/2975 [04:59<29:07, 1.64it/s] 4%\|▍ \| 113/2975 [05:00<29:10, 1.63it/s] 4%\|▍ \| 114/2975 [05:00<29:08, 1.64it/s] 4%\|▍ \| 115/2975 [05:01<29:09, 1.64it/s] 4%\|▍ \| 116/2975 [05:01<29
	0: {'loss': 0.8187, 'grad_norm': 0.9406876728676388, 'learning_rate': 8.14e-06, 'epoch': 0.04}
	0: {'loss': 0.8218, 'grad_norm': 0.925012843902631, 'learning_rate': 8.740000000000001e-06, 'epoch': 0.04}
	0: :07, 1.64it/s] 4%\|▍ \| 117/2975 [05:02<29:07, 1.64it/s] 4%\|▍ \| 118/2975 [05:03<29:06, 1.64it/s] 4%\|▍ \| 119/2975 [05:03<29:04, 1.64it/s] 4%\|▍ \| 120/2975 [05:04<29:12, 1.63it/s] 4%\|▍ \| 120/2975 [05:04<29:12, 1.63it/s] 4%\|▍ \| 121/2975 [05:04<29:13, 1.63it/s] 4%\|▍ \| 122/2975 [05:05<29:13, 1.63it/s] 4%\|▍ \| 123/2975 [05:06<29:13, 1.63it/s] 4%\|▍ \| 124/2975 [05:06<29:15, 1.62it/s] 4%\|▍ \| 125/2975 [05:07<29:14, 1.62it/s] 4%\|▍ \| 126/2975 [05:08<29:14, 1.62it/s] 4%\|▍ \| 127/2975 [05:08<29:13, 1.62it/s] 4%\|▍ \| 128/2975 [05:09<29:11, 1.63it/s] 4%\|▍ \| 129/2975 [05:09<29:11, 1.62it/s] 4%\|▍ \| 130/2975 [05:10<29:10, 1.63it/s] 4%\|▍ \| 130/2975 [05:10<29:10, 1.63it/s] 4%\|▍ \| 131/2975 [05:11<29:07, 1.63it/s] 4
	0: {'loss': 0.8419, 'grad_norm': 0.8502300171757229, 'learning_rate': 9.34e-06, 'epoch': 0.05}
	0: %\|▍ \| 132/2975 [05:11<29:01, 1.63it/s] 4%\|▍ \| 133/2975 [05:12<28:59, 1.63it/s] 5%\|▍ \| 134/2975 [05:12<28:52, 1.64it/s] 5%\|▍ \| 135/2975 [05:13<28:48, 1.64it/s] 5%\|▍ \| 136/2975 [05:14<28:52, 1.64it/s] 5%\|▍ \| 137/2975 [05:14<28:56, 1.63it/s] 5%\|▍ \| 138/2975 [05:15<29:00, 1.63it/s] 5%\|▍ \| 139/2975 [05:15<29:01, 1.63it/s] 5%\|▍ \| 140/2975 [05:16<29:05, 1.62it/s] 5%\|▍ \| 140/2975 [05:16<29:05, 1.62it/s] 5%\|▍ \| 141/2975 [05:17<29:09, 1.62it/s] 5%\|▍ \| 142/2975 [05:17<29:04, 1.62it/s] 5%\|▍ \| 143/2975 [05:18<29:00, 1.63it/s] 5%\|▍ \| 144/2975 [05:19<28:53, 1.63it/s] 5%\|▍ \| 145/2975 [05:19<31:39, 1.49it/s] 5%\|▍ \| 146/2975 [05:20<30:45, 1.53it/s] 5%\|▍ \| 147/2975 [05:21<30:07, 1.56it/s] 5%\|▍ \| 148/2975 [05:21<29:45, 1.58it/s] 5%\|▌ \| 14
	0: {'loss': 0.7937, 'grad_norm': 0.7785647727224466, 'learning_rate': 9.940000000000001e-06, 'epoch': 0.05}
	0: {'loss': 0.7883, 'grad_norm': 0.7493685913449163, 'learning_rate': 1e-05, 'epoch': 0.05}
	0: 9/2975 [05:22<29:27, 1.60it/s] 5%\|▌ \| 150/2975 [05:22<29:16, 1.61it/s] 5%\|▌ \| 150/2975 [05:22<29:16, 1.61it/s] 5%\|▌ \| 151/2975 [05:23<29:09, 1.61it/s] 5%\|▌ \| 152/2975 [05:24<28:58, 1.62it/s] 5%\|▌ \| 153/2975 [05:24<28:50, 1.63it/s] 5%\|▌ \| 154/2975 [05:25<28:42, 1.64it/s] 5%\|▌ \| 155/2975 [05:25<28:39, 1.64it/s] 5%\|▌ \| 156/2975 [05:26<28:35, 1.64it/s] 5%\|▌ \| 157/2975 [05:27<28:33, 1.64it/s] 5%\|▌ \| 158/2975 [05:27<28:32, 1.64it/s] 5%\|▌ \| 159/2975 [05:28<28:29, 1.65it/s] 5%\|▌ \| 160/2975 [05:28<28:25, 1.65it/s] 5%\|▌ \| 160/2975 [05:28<28:25, 1.65it/s] 5%\|▌ \| 161/2975 [05:29<28:27, 1.65it/s] 5%\|▌ \| 162/2975 [05:30<28:31, 1.64it/s] 5%\|▌ \| 163/2975 [05:30<28:32, 1.64it/s] 6%\|▌ \| 164/2975 [05:31<28:31
	0: {'loss': 0.79, 'grad_norm': 0.7957620935977624, 'learning_rate': 1e-05, 'epoch': 0.06}
	0: {'loss': 0.7847, 'grad_norm': 0.782598850636886, 'learning_rate': 1e-05, 'epoch': 0.06}
	0: , 1.64it/s] 6%\|▌ \| 165/2975 [05:32<28:29, 1.64it/s] 6%\|▌ \| 166/2975 [05:32<28:29, 1.64it/s] 6%\|▌ \| 167/2975 [05:33<28:31, 1.64it/s] 6%\|▌ \| 168/2975 [05:33<28:32, 1.64it/s] 6%\|▌ \| 169/2975 [05:34<28:32, 1.64it/s] 6%\|▌ \| 170/2975 [05:35<28:30, 1.64it/s] 6%\|▌ \| 170/2975 [05:35<28:30, 1.64it/s] 6%\|▌ \| 171/2975 [05:35<28:29, 1.64it/s] 6%\|▌ \| 172/2975 [05:36<28:29, 1.64it/s] 6%\|▌ \| 173/2975 [05:36<28:27, 1.64it/s] 6%\|▌ \| 174/2975 [05:37<28:23, 1.64it/s] 6%\|▌ \| 175/2975 [05:38<28:22, 1.64it/s] 6%\|▌ \| 176/2975 [05:38<28:23, 1.64it/s] 6%\|▌ \| 177/2975 [05:39<28:22, 1.64it/s] 6%\|▌ \| 178/2975 [05:39<28:25, 1.64it/s] 6%\|▌ \| 179/2975 [05:40<28:24, 1.64it/s] 6%\|▌ \| 180/2975 [05:41<28:26, 1.64it/s] 6%\|�
	0: {'loss': 0.8106, 'grad_norm': 0.8284804242874985, 'learning_rate': 1e-05, 'epoch': 0.06}
	0: �� \| 180/2975 [05:41<28:26, 1.64it/s] 6%\|▌ \| 181/2975 [05:41<28:25, 1.64it/s] 6%\|▌ \| 182/2975 [05:42<28:24, 1.64it/s] 6%\|▌ \| 183/2975 [05:42<28:22, 1.64it/s] 6%\|▌ \| 184/2975 [05:43<28:22, 1.64it/s] 6%\|▌ \| 185/2975 [05:44<28:27, 1.63it/s] 6%\|▋ \| 186/2975 [05:44<28:29, 1.63it/s] 6%\|▋ \| 187/2975 [05:45<28:33, 1.63it/s] 6%\|▋ \| 188/2975 [05:46<28:33, 1.63it/s] 6%\|▋ \| 189/2975 [05:46<28:35, 1.62it/s] 6%\|▋ \| 190/2975 [05:47<28:38, 1.62it/s] 6%\|▋ \| 190/2975 [05:47<28:38, 1.62it/s] 6%\|▋ \| 191/2975 [05:47<28:40, 1.62it/s] 6%\|▋ \| 192/2975 [05:48<28:37, 1.62it/s] 6%\|▋ \| 193/2975 [05:49<28:35, 1.62it/s] 7%\|▋ \| 194/2975 [05:49<28:32, 1.62it/s] 7%\|▋ \| 195/2975 [05:50<28:31, 1.62it/s] 7%\|▋ \| 196/2975 [05:51<28:29, 1.63it/s] 7%\|▋ \| 197/2
	0: {'loss': 0.7863, 'grad_norm': 0.7926107821646097, 'learning_rate': 1e-05, 'epoch': 0.07}
	0: {'loss': 0.7856, 'grad_norm': 0.8607046204689304, 'learning_rate': 1e-05, 'epoch': 0.07}
	0: 975 [05:51<28:27, 1.63it/s] 7%\|▋ \| 198/2975 [05:52<28:25, 1.63it/s] 7%\|▋ \| 199/2975 [05:52<28:27, 1.63it/s] 7%\|▋ \| 200/2975 [05:53<28:23, 1.63it/s] 7%\|▋ \| 200/2975 [05:53<28:23, 1.63it/s] 7%\|▋ \| 201/2975 [05:54<28:27, 1.62it/s] 7%\|▋ \| 202/2975 [05:54<28:26, 1.62it/s] 7%\|▋ \| 203/2975 [05:55<28:23, 1.63it/s] 7%\|▋ \| 204/2975 [05:55<28:26, 1.62it/s] 7%\|▋ \| 205/2975 [05:56<28:29, 1.62it/s] 7%\|▋ \| 206/2975 [05:57<28:26, 1.62it/s] 7%\|▋ \| 207/2975 [05:57<28:29, 1.62it/s] 7%\|▋ \| 208/2975 [05:58<28:29, 1.62it/s] 7%\|▋ \| 209/2975 [05:59<28:26, 1.62it/s] 7%\|▋ \| 210/2975 [05:59<28:23, 1.62it/s] 7%\|▋ \| 210/2975 [05:59<28:23, 1.62it/s] 7%\|▋ \| 211/2975 [06:00<28:27, 1.62it/s] 7%\|▋ \| 212/2975 [06:00<28:26,
	0: {'loss': 0.8051, 'grad_norm': 0.8692304162211446, 'learning_rate': 1e-05, 'epoch': 0.07}
	0: 1.62it/s] 7%\|▋ \| 213/2975 [06:01<28:25, 1.62it/s] 7%\|▋ \| 214/2975 [06:02<28:21, 1.62it/s] 7%\|▋ \| 215/2975 [06:02<28:19, 1.62it/s] 7%\|▋ \| 216/2975 [06:03<28:17, 1.63it/s] 7%\|▋ \| 217/2975 [06:03<28:15, 1.63it/s] 7%\|▋ \| 218/2975 [06:04<28:15, 1.63it/s] 7%\|▋ \| 219/2975 [06:05<28:12, 1.63it/s] 7%\|▋ \| 220/2975 [06:05<28:15, 1.63it/s] 7%\|▋ \| 220/2975 [06:05<28:15, 1.63it/s] 7%\|▋ \| 221/2975 [06:06<28:16, 1.62it/s] 7%\|▋ \| 222/2975 [06:07<28:14, 1.63it/s] 7%\|▋ \| 223/2975 [06:07<28:11, 1.63it/s] 8%\|▊ \| 224/2975 [06:08<28:11, 1.63it/s] 8%\|▊ \| 225/2975 [06:08<28:09, 1.63it/s] 8%\|▊ \| 226/2975 [06:09<28:06, 1.63it/s] 8%\|▊ \| 227/2975 [06:10<28:04, 1.63it/s] 8%\|▊ \| 228/2975 [06:10<28:07, 1.63it/s] 8%\|▊ \| 229/2975 [06:11<28:09, 1.63it/s] 8%\|▊
	0: {'loss': 0.7915, 'grad_norm': 0.8049170893159536, 'learning_rate': 1e-05, 'epoch': 0.08}
	0: {'loss': 0.781, 'grad_norm': 0.7687711775406885, 'learning_rate': 1e-05, 'epoch': 0.08}
	0: \| 230/2975 [06:11<28:06, 1.63it/s] 8%\|▊ \| 230/2975 [06:11<28:06, 1.63it/s] 8%\|▊ \| 231/2975 [06:12<28:09, 1.62it/s] 8%\|▊ \| 232/2975 [06:13<28:06, 1.63it/s] 8%\|▊ \| 233/2975 [06:13<28:05, 1.63it/s] 8%\|▊ \| 234/2975 [06:14<28:03, 1.63it/s] 8%\|▊ \| 235/2975 [06:15<28:00, 1.63it/s] 8%\|▊ \| 236/2975 [06:15<28:00, 1.63it/s] 8%\|▊ \| 237/2975 [06:16<28:00, 1.63it/s] 8%\|▊ \| 238/2975 [06:16<27:59, 1.63it/s] 8%\|▊ \| 239/2975 [06:17<28:03, 1.63it/s] 8%\|▊ \| 240/2975 [06:18<27:59, 1.63it/s] 8%\|▊ \| 240/2975 [06:18<27:59, 1.63it/s] 8%\|▊ \| 241/2975 [06:18<28:00, 1.63it/s] 8%\|▊ \| 242/2975 [06:19<27:59, 1.63it/s] 8%\|▊ \| 243/2975 [06:19<28:00, 1.63it/s] 8%\|▊ \| 244/2975 [06:20<28:00, 1.63it/s] 8%\|▊ \| 245/2975
	0: {'loss': 0.7829, 'grad_norm': 0.8128211126668727, 'learning_rate': 1e-05, 'epoch': 0.08}
	0: {'loss': 0.7748, 'grad_norm': 0.7604484943760149, 'learning_rate': 1e-05, 'epoch': 0.09}
	0: [06:21<28:00, 1.62it/s] 8%\|▊ \| 246/2975 [06:21<28:01, 1.62it/s] 8%\|▊ \| 247/2975 [06:22<28:01, 1.62it/s] 8%\|▊ \| 248/2975 [06:22<27:58, 1.63it/s] 8%\|▊ \| 249/2975 [06:23<27:56, 1.63it/s] 8%\|▊ \| 250/2975 [06:24<27:57, 1.62it/s] 8%\|▊ \| 250/2975 [06:24<27:57, 1.62it/s] 8%\|▊ \| 251/2975 [06:24<28:01, 1.62it/s] 8%\|▊ \| 252/2975 [06:25<27:58, 1.62it/s] 9%\|▊ \| 253/2975 [06:26<27:55, 1.62it/s] 9%\|▊ \| 254/2975 [06:26<27:56, 1.62it/s] 9%\|▊ \| 255/2975 [06:27<27:53, 1.63it/s] 9%\|▊ \| 256/2975 [06:27<27:49, 1.63it/s] 9%\|▊ \| 257/2975 [06:28<27:47, 1.63it/s] 9%\|▊ \| 258/2975 [06:29<27:47, 1.63it/s] 9%\|▊ \| 259/2975 [06:29<27:52, 1.62it/s] 9%\|▊ \| 260/2975 [06:30<27:51, 1.62it/s] 9%\|▊ \| 260/2975 [06:30<27:51, 1.6
	0: {'loss': 0.783, 'grad_norm': 0.7744995010014754, 'learning_rate': 1e-05, 'epoch': 0.09}
	0: 2it/s] 9%\|▉ \| 261/2975 [06:31<27:51, 1.62it/s] 9%\|▉ \| 262/2975 [06:31<27:50, 1.62it/s] 9%\|▉ \| 263/2975 [06:32<27:48, 1.63it/s] 9%\|▉ \| 264/2975 [06:32<27:48, 1.62it/s] 9%\|▉ \| 265/2975 [06:33<27:45, 1.63it/s] 9%\|▉ \| 266/2975 [06:34<27:45, 1.63it/s] 9%\|▉ \| 267/2975 [06:34<27:44, 1.63it/s] 9%\|▉ \| 268/2975 [06:35<27:43, 1.63it/s] 9%\|▉ \| 269/2975 [06:35<27:40, 1.63it/s] 9%\|▉ \| 270/2975 [06:36<27:38, 1.63it/s] 9%\|▉ \| 270/2975 [06:36<27:38, 1.63it/s] 9%\|▉ \| 271/2975 [06:37<27:40, 1.63it/s] 9%\|▉ \| 272/2975 [06:37<27:39, 1.63it/s] 9%\|▉ \| 273/2975 [06:38<27:39, 1.63it/s] 9%\|▉ \| 274/2975 [06:38<27:38, 1.63it/s] 9%\|▉ \| 275/2975 [06:39<27:39, 1.63it/s] 9%\|▉ \| 276/2975 [06:40<27:38, 1.63it/s] 9%\|▉ \| 277/2975 [06:40<27:39, 1.63it/s] 9%\|▉
	0: {'loss': 0.7854, 'grad_norm': 0.7643071268098258, 'learning_rate': 1e-05, 'epoch': 0.09}
	0: {'loss': 0.7895, 'grad_norm': 0.8158389807254296, 'learning_rate': 1e-05, 'epoch': 0.1}
	0: \| 278/2975 [06:41<27:37, 1.63it/s] 9%\|▉ \| 279/2975 [06:42<27:37, 1.63it/s] 9%\|▉ \| 280/2975 [06:42<27:35, 1.63it/s] 9%\|▉ \| 280/2975 [06:42<27:35, 1.63it/s] 9%\|▉ \| 281/2975 [06:43<27:35, 1.63it/s] 9%\|▉ \| 282/2975 [06:43<27:39, 1.62it/s] 10%\|▉ \| 283/2975 [06:44<27:39, 1.62it/s] 10%\|▉ \| 284/2975 [06:45<27:37, 1.62it/s] 10%\|▉ \| 285/2975 [06:45<27:33, 1.63it/s] 10%\|▉ \| 286/2975 [06:46<27:28, 1.63it/s] 10%\|▉ \| 287/2975 [06:46<27:25, 1.63it/s] 10%\|▉ \| 288/2975 [06:47<27:26, 1.63it/s] 10%\|▉ \| 289/2975 [06:48<27:25, 1.63it/s] 10%\|▉ \| 290/2975 [06:48<27:23, 1.63it/s] 10%\|▉ \| 290/2975 [06:48<27:23, 1.63it/s] 10%\|▉ \| 291/2975 [06:49<27:25, 1.63it/s] 10%\|▉ \| 292/2975 [06:50<27:21, 1.63it/s] 10%\|▉ \| 293/2975 [0
	0: {'loss': 0.7638, 'grad_norm': 0.747639615762065, 'learning_rate': 1e-05, 'epoch': 0.1}
	0: 6:50<27:19, 1.64it/s] 10%\|▉ \| 294/2975 [06:51<27:18, 1.64it/s] 10%\|▉ \| 295/2975 [06:51<27:19, 1.64it/s] 10%\|▉ \| 296/2975 [06:52<27:15, 1.64it/s] 10%\|▉ \| 297/2975 [06:53<27:10, 1.64it/s] 10%\|█ \| 298/2975 [06:53<27:07, 1.64it/s] 10%\|█ \| 299/2975 [06:54<27:07, 1.64it/s] 10%\|█ \| 300/2975 [06:54<27:11, 1.64it/s] 10%\|█ \| 300/2975 [06:54<27:11, 1.64it/s] 10%\|█ \| 301/2975 [06:55<27:10, 1.64it/s] 10%\|█ \| 302/2975 [06:56<27:08, 1.64it/s] 10%\|█ \| 303/2975 [06:56<27:05, 1.64it/s] 10%\|█ \| 304/2975 [06:57<27:03, 1.65it/s] 10%\|█ \| 305/2975 [06:57<27:00, 1.65it/s] 10%\|█ \| 306/2975 [06:58<26:56, 1.65it/s] 10%\|█ \| 307/2975 [06:59<26:54, 1.65it/s] 10%\|█ \| 308/2975 [06:59<26:54, 1.65it/s] 10%\|█ \| 309/2975 [07:00<26:58, 1.65it/s] 10%\|█ \| 310/2975 [07:00<26:58, 1.65i
	0: {'loss': 0.761, 'grad_norm': 0.7527151310109943, 'learning_rate': 1e-05, 'epoch': 0.1}
	0: {'loss': 0.7666, 'grad_norm': 0.7679616039639676, 'learning_rate': 1e-05, 'epoch': 0.11}
	0: t/s] 10%\|█ \| 310/2975 [07:00<26:58, 1.65it/s] 10%\|█ \| 311/2975 [07:01<27:00, 1.64it/s] 10%\|█ \| 312/2975 [07:02<26:59, 1.64it/s] 11%\|█ \| 313/2975 [07:02<26:59, 1.64it/s] 11%\|█ \| 314/2975 [07:03<26:57, 1.65it/s] 11%\|█ \| 315/2975 [07:04<26:55, 1.65it/s] 11%\|█ \| 316/2975 [07:04<26:56, 1.64it/s] 11%\|█ \| 317/2975 [07:05<26:59, 1.64it/s] 11%\|█ \| 318/2975 [07:05<26:55, 1.64it/s] 11%\|█ \| 319/2975 [07:06<26:52, 1.65it/s] 11%\|█ \| 320/2975 [07:07<26:52, 1.65it/s] 11%\|█ \| 320/2975 [07:07<26:52, 1.65it/s] 11%\|█ \| 321/2975 [07:07<26:53, 1.64it/s] 11%\|█ \| 322/2975 [07:08<26:51, 1.65it/s] 11%\|█ \| 323/2975 [07:08<26:51, 1.65it/s] 11%\|█ \| 324/2975 [07:09<26:51, 1.64it/s] 11%\|█ \| 325/2975 [07:10<26:54, 1.64it/s] 11%\|█
	0: {'loss': 0.748, 'grad_norm': 0.8669834737593406, 'learning_rate': 1e-05, 'epoch': 0.11}
	0: {'loss': 0.7805, 'grad_norm': 0.77301353399369, 'learning_rate': 1e-05, 'epoch': 0.11}
	0: \| 326/2975 [07:10<26:52, 1.64it/s] 11%\|█ \| 327/2975 [07:11<26:50, 1.64it/s] 11%\|█ \| 328/2975 [07:11<26:50, 1.64it/s] 11%\|█ \| 329/2975 [07:12<26:52, 1.64it/s] 11%\|█ \| 330/2975 [07:13<26:51, 1.64it/s] 11%\|█ \| 330/2975 [07:13<26:51, 1.64it/s] 11%\|█ \| 331/2975 [07:13<26:49, 1.64it/s] 11%\|█ \| 332/2975 [07:14<26:51, 1.64it/s] 11%\|█ \| 333/2975 [07:14<26:48, 1.64it/s] 11%\|█ \| 334/2975 [07:15<26:47, 1.64it/s] 11%\|█▏ \| 335/2975 [07:16<26:46, 1.64it/s] 11%\|█▏ \| 336/2975 [07:16<26:44, 1.64it/s] 11%\|█▏ \| 337/2975 [07:17<26:46, 1.64it/s] 11%\|█▏ \| 338/2975 [07:18<26:45, 1.64it/s] 11%\|█▏ \| 339/2975 [07:18<26:42, 1.64it/s] 11%\|█▏ \| 340/2975 [07:19<26:39, 1.65it/s] 11%\|█▏ \| 340/2975 [07:19<26:39, 1.65it/s] 11%\|█▏
	0: {'loss': 0.7913, 'grad_norm': 0.7979909975428289, 'learning_rate': 1e-05, 'epoch': 0.12}
	0: \| 341/2975 [07:19<26:43, 1.64it/s] 11%\|█▏ \| 342/2975 [07:20<26:41, 1.64it/s] 12%\|█▏ \| 343/2975 [07:21<26:41, 1.64it/s] 12%\|█▏ \| 344/2975 [07:21<26:38, 1.65it/s] 12%\|█▏ \| 345/2975 [07:22<26:35, 1.65it/s] 12%\|█▏ \| 346/2975 [07:22<26:31, 1.65it/s] 12%\|█▏ \| 347/2975 [07:23<26:29, 1.65it/s] 12%\|█▏ \| 348/2975 [07:24<26:27, 1.65it/s] 12%\|█▏ \| 349/2975 [07:24<26:26, 1.66it/s] 12%\|█▏ \| 350/2975 [07:25<26:25, 1.66it/s] 12%\|█▏ \| 350/2975 [07:25<26:25, 1.66it/s] 12%\|█▏ \| 351/2975 [07:25<26:27, 1.65it/s] 12%\|█▏ \| 352/2975 [07:26<26:26, 1.65it/s] 12%\|█▏ \| 353/2975 [07:27<26:28, 1.65it/s] 12%\|█▏ \| 354/2975 [07:27<26:28, 1.65it/s] 12%\|█▏ \| 355/2975 [07:28<26:28, 1.65it/s] 12%\|█▏ \| 356/2975 [07:28<26:28, 1.65it/s] 12%\|█▏ \| 357/2975 [07:29<26:30, 1.65it/s]
	0: {'loss': 0.7749, 'grad_norm': 0.752827606829707, 'learning_rate': 1e-05, 'epoch': 0.12}
	0: {'loss': 0.7592, 'grad_norm': 0.7632063835474084, 'learning_rate': 1e-05, 'epoch': 0.12}
	0: 12%\|█▏ \| 358/2975 [07:30<26:31, 1.64it/s] 12%\|█▏ \| 359/2975 [07:30<26:33, 1.64it/s] 12%\|█▏ \| 360/2975 [07:31<29:08, 1.50it/s] 12%\|█▏ \| 360/2975 [07:31<29:08, 1.50it/s] 12%\|█▏ \| 361/2975 [07:32<28:21, 1.54it/s] 12%\|█▏ \| 362/2975 [07:32<27:47, 1.57it/s] 12%\|█▏ \| 363/2975 [07:33<27:24, 1.59it/s] 12%\|█▏ \| 364/2975 [07:33<27:09, 1.60it/s] 12%\|█▏ \| 365/2975 [07:34<27:00, 1.61it/s] 12%\|█▏ \| 366/2975 [07:35<26:52, 1.62it/s] 12%\|█▏ \| 367/2975 [07:35<26:45, 1.62it/s] 12%\|█▏ \| 368/2975 [07:36<26:41, 1.63it/s] 12%\|█▏ \| 369/2975 [07:37<26:36, 1.63it/s] 12%\|█▏ \| 370/2975 [07:37<26:32, 1.64it/s] 12%\|█▏ \| 370/2975 [07:37<26:32, 1.64it/s] 12%\|█▏ \| 371/2975 [07:38<26:27, 1.64it/s] 13%\|█▎ \| 372/2975 [07:38<26:2
	0: {'loss': 0.7592, 'grad_norm': 0.741049028699946, 'learning_rate': 1e-05, 'epoch': 0.13}
	0: 4, 1.64it/s] 13%\|█▎ \| 373/2975 [07:39<26:20, 1.65it/s] 13%\|█▎ \| 374/2975 [07:40<26:19, 1.65it/s] 13%\|█▎ \| 375/2975 [07:40<26:18, 1.65it/s] 13%\|█▎ \| 376/2975 [07:41<26:18, 1.65it/s] 13%\|█▎ \| 377/2975 [07:41<26:15, 1.65it/s] 13%\|█▎ \| 378/2975 [07:42<26:12, 1.65it/s] 13%\|█▎ \| 379/2975 [07:43<26:08, 1.66it/s] 13%\|█▎ \| 380/2975 [07:43<26:07, 1.66it/s] 13%\|█▎ \| 380/2975 [07:43<26:07, 1.66it/s] 13%\|█▎ \| 381/2975 [07:44<26:09, 1.65it/s] 13%\|█▎ \| 382/2975 [07:44<26:09, 1.65it/s] 13%\|█▎ \| 383/2975 [07:45<26:07, 1.65it/s] 13%\|█▎ \| 384/2975 [07:46<26:05, 1.66it/s] 13%\|█▎ \| 385/2975 [07:46<26:05, 1.65it/s] 13%\|█▎ \| 386/2975 [07:47<26:07, 1.65it/s] 13%\|█▎ \| 387/2975 [07:47<26:07, 1.65it/s] 13%\|█▎ \| 388/2975 [07:48<26:10, 1.65it/s] 13%\|█▎ \| 38
	0: {'loss': 0.786, 'grad_norm': 0.808057993234383, 'learning_rate': 1e-05, 'epoch': 0.13}
	0: {'loss': 0.7761, 'grad_norm': 0.8398541202344658, 'learning_rate': 1e-05, 'epoch': 0.13}
	0: 9/2975 [07:49<26:11, 1.65it/s] 13%\|█▎ \| 390/2975 [07:49<26:13, 1.64it/s] 13%\|█▎ \| 390/2975 [07:49<26:13, 1.64it/s] 13%\|█▎ \| 391/2975 [07:50<26:13, 1.64it/s] 13%\|█▎ \| 392/2975 [07:50<26:13, 1.64it/s] 13%\|█▎ \| 393/2975 [07:51<26:13, 1.64it/s] 13%\|█▎ \| 394/2975 [07:52<26:12, 1.64it/s] 13%\|█▎ \| 395/2975 [07:52<26:08, 1.65it/s] 13%\|█▎ \| 396/2975 [07:53<26:03, 1.65it/s] 13%\|█▎ \| 397/2975 [07:54<26:00, 1.65it/s] 13%\|█▎ \| 398/2975 [07:54<26:00, 1.65it/s] 13%\|█▎ \| 399/2975 [07:55<25:59, 1.65it/s] 13%\|█▎ \| 400/2975 [07:55<25:59, 1.65it/s] 13%\|█▎ \| 400/2975 [07:55<25:59, 1.65it/s] 13%\|█▎ \| 401/2975 [07:56<26:00, 1.65it/s] 14%\|█▎ \| 402/2975 [07:57<25:57, 1.65it/s] 14%\|█▎ \| 403/2975 [07:57<25:55, 1.65it/s] 14%\|█
	0: {'loss': 0.7596, 'grad_norm': 0.8017339744514405, 'learning_rate': 1e-05, 'epoch': 0.14}
	0: ▎ \| 404/2975 [07:58<25:52, 1.66it/s] 14%\|█▎ \| 405/2975 [07:58<25:50, 1.66it/s] 14%\|█▎ \| 406/2975 [07:59<25:48, 1.66it/s] 14%\|█▎ \| 407/2975 [08:00<25:48, 1.66it/s] 14%\|█▎ \| 408/2975 [08:00<25:46, 1.66it/s] 14%\|█▎ \| 409/2975 [08:01<25:47, 1.66it/s] 14%\|█▍ \| 410/2975 [08:01<25:47, 1.66it/s] 14%\|█▍ \| 410/2975 [08:01<25:47, 1.66it/s] 14%\|█▍ \| 411/2975 [08:02<25:47, 1.66it/s] 14%\|█▍ \| 412/2975 [08:03<25:46, 1.66it/s] 14%\|█▍ \| 413/2975 [08:03<25:46, 1.66it/s] 14%\|█▍ \| 414/2975 [08:04<25:48, 1.65it/s] 14%\|█▍ \| 415/2975 [08:04<25:48, 1.65it/s] 14%\|█▍ \| 416/2975 [08:05<25:49, 1.65it/s] 14%\|█▍ \| 417/2975 [08:06<25:48, 1.65it/s] 14%\|█▍ \| 418/2975 [08:06<25:48, 1.65it/s] 14%\|█▍ \| 419/2975 [08:07<25:50, 1.65it/s] 14%\|█▍ \| 420/2975 [08:07<25:53,
	0: {'loss': 0.765, 'grad_norm': 0.7680169702718432, 'learning_rate': 1e-05, 'epoch': 0.14}
	0: {'loss': 0.7588, 'grad_norm': 0.7973598542191226, 'learning_rate': 1e-05, 'epoch': 0.14}
	0: 1.64it/s] 14%\|█▍ \| 420/2975 [08:07<25:53, 1.64it/s] 14%\|█▍ \| 421/2975 [08:08<25:58, 1.64it/s] 14%\|█▍ \| 422/2975 [08:09<25:56, 1.64it/s] 14%\|█▍ \| 423/2975 [08:09<25:55, 1.64it/s] 14%\|█▍ \| 424/2975 [08:10<25:53, 1.64it/s] 14%\|█▍ \| 425/2975 [08:10<25:51, 1.64it/s] 14%\|█▍ \| 426/2975 [08:11<25:48, 1.65it/s] 14%\|█▍ \| 427/2975 [08:12<25:55, 1.64it/s] 14%\|█▍ \| 428/2975 [08:12<25:49, 1.64it/s] 14%\|█▍ \| 429/2975 [08:13<25:45, 1.65it/s] 14%\|█▍ \| 430/2975 [08:14<25:42, 1.65it/s] 14%\|█▍ \| 430/2975 [08:14<25:42, 1.65it/s] 14%\|█▍ \| 431/2975 [08:14<25:42, 1.65it/s] 15%\|█▍ \| 432/2975 [08:15<25:41, 1.65it/s] 15%\|█▍ \| 433/2975 [08:15<25:38, 1.65it/s] 15%\|█▍ \| 434/2975 [08:16<25:38, 1.65it/s] 15%\|█▍ \| 435/2975
	0: {'loss': 0.757, 'grad_norm': 0.7879157045491778, 'learning_rate': 1e-05, 'epoch': 0.15}
	0: {'loss': 0.7522, 'grad_norm': 0.7679264682899368, 'learning_rate': 1e-05, 'epoch': 0.15}
	0: [08:17<25:38, 1.65it/s] 15%\|█▍ \| 436/2975 [08:17<25:36, 1.65it/s] 15%\|█▍ \| 437/2975 [08:18<25:35, 1.65it/s] 15%\|█▍ \| 438/2975 [08:18<25:51, 1.64it/s] 15%\|█▍ \| 439/2975 [08:19<25:44, 1.64it/s] 15%\|█▍ \| 440/2975 [08:20<25:40, 1.65it/s] 15%\|█▍ \| 440/2975 [08:20<25:40, 1.65it/s] 15%\|█▍ \| 441/2975 [08:20<25:39, 1.65it/s] 15%\|█▍ \| 442/2975 [08:21<25:37, 1.65it/s] 15%\|█▍ \| 443/2975 [08:21<25:38, 1.65it/s] 15%\|█▍ \| 444/2975 [08:22<25:37, 1.65it/s] 15%\|█▍ \| 445/2975 [08:23<25:36, 1.65it/s] 15%\|█▍ \| 446/2975 [08:23<25:35, 1.65it/s] 15%\|█▌ \| 447/2975 [08:24<25:35, 1.65it/s] 15%\|█▌ \| 448/2975 [08:24<25:35, 1.65it/s] 15%\|█▌ \| 449/2975 [08:25<25:32, 1.65it/s] 15%\|█▌ \| 450/2975 [08:26<25:30, 1.65it/s] 15%\|█▌
	0: {'loss': 0.7731, 'grad_norm': 0.7671720022789881, 'learning_rate': 1e-05, 'epoch': 0.15}
	0: \| 450/2975 [08:26<25:30, 1.65it/s] 15%\|█▌ \| 451/2975 [08:26<25:29, 1.65it/s] 15%\|█▌ \| 452/2975 [08:27<25:31, 1.65it/s] 15%\|█▌ \| 453/2975 [08:27<25:29, 1.65it/s] 15%\|█▌ \| 454/2975 [08:28<25:26, 1.65it/s] 15%\|█▌ \| 455/2975 [08:29<25:26, 1.65it/s] 15%\|█▌ \| 456/2975 [08:29<25:28, 1.65it/s] 15%\|█▌ \| 457/2975 [08:30<25:29, 1.65it/s] 15%\|█▌ \| 458/2975 [08:31<25:28, 1.65it/s] 15%\|█▌ \| 459/2975 [08:31<25:29, 1.65it/s] 15%\|█▌ \| 460/2975 [08:32<25:30, 1.64it/s] 15%\|█▌ \| 460/2975 [08:32<25:30, 1.64it/s] 15%\|█▌ \| 461/2975 [08:32<25:30, 1.64it/s] 16%\|█▌ \| 462/2975 [08:33<25:26, 1.65it/s] 16%\|█▌ \| 463/2975 [08:34<25:23, 1.65it/s] 16%\|█▌ \| 464/2975 [08:34<25:21, 1.65it/s] 16%\|█▌ \| 465/2975 [08:35<25:18, 1.65it/s] 16%\|█▌ \| 466/2975 [08:35<25:15, 1.66it/
	0: {'loss': 0.7697, 'grad_norm': 0.7672119462561652, 'learning_rate': 1e-05, 'epoch': 0.16}
	0: {'loss': 0.7506, 'grad_norm': 0.7215765122797166, 'learning_rate': 1e-05, 'epoch': 0.16}
	0: s] 16%\|█▌ \| 467/2975 [08:36<25:14, 1.66it/s] 16%\|█▌ \| 468/2975 [08:37<25:24, 1.64it/s] 16%\|█▌ \| 469/2975 [08:37<25:20, 1.65it/s] 16%\|█▌ \| 470/2975 [08:38<25:17, 1.65it/s] 16%\|█▌ \| 470/2975 [08:38<25:17, 1.65it/s] 16%\|█▌ \| 471/2975 [08:38<25:16, 1.65it/s] 16%\|█▌ \| 472/2975 [08:39<25:15, 1.65it/s] 16%\|█▌ \| 473/2975 [08:40<25:15, 1.65it/s] 16%\|█▌ \| 474/2975 [08:40<25:16, 1.65it/s] 16%\|█▌ \| 475/2975 [08:41<25:16, 1.65it/s] 16%\|█▌ \| 476/2975 [08:41<25:17, 1.65it/s] 16%\|█▌ \| 477/2975 [08:42<25:16, 1.65it/s] 16%\|█▌ \| 478/2975 [08:43<25:13, 1.65it/s] 16%\|█▌ \| 479/2975 [08:43<25:16, 1.65it/s] 16%\|█▌ \| 480/2975 [08:44<25:18, 1.64it/s] 16%\|█▌ \| 480/2975 [08:44<25:18, 1.64it/s] 16%\|█▌ \| 481/2975 [08:44<
	0: {'loss': 0.773, 'grad_norm': 0.775884303997141, 'learning_rate': 1e-05, 'epoch': 0.16}
	0: 25:18, 1.64it/s] 16%\|█▌ \| 482/2975 [08:45<25:18, 1.64it/s] 16%\|█▌ \| 483/2975 [08:46<25:16, 1.64it/s] 16%\|█▋ \| 484/2975 [08:46<25:12, 1.65it/s] 16%\|█▋ \| 485/2975 [08:47<25:08, 1.65it/s] 16%\|█▋ \| 486/2975 [08:48<25:05, 1.65it/s] 16%\|█▋ \| 487/2975 [08:48<25:04, 1.65it/s] 16%\|█▋ \| 488/2975 [08:49<25:02, 1.66it/s] 16%\|█▋ \| 489/2975 [08:49<25:02, 1.65it/s] 16%\|█▋ \| 490/2975 [08:50<25:01, 1.65it/s] 16%\|█▋ \| 490/2975 [08:50<25:01, 1.65it/s] 17%\|█▋ \| 491/2975 [08:51<25:00, 1.66it/s] 17%\|█▋ \| 492/2975 [08:51<24:59, 1.66it/s] 17%\|█▋ \| 493/2975 [08:52<24:59, 1.66it/s] 17%\|█▋ \| 494/2975 [08:52<24:59, 1.65it/s] 17%\|█▋ \| 495/2975 [08:53<24:58, 1.65it/s] 17%\|█▋ \| 496/2975 [08:54<24:59, 1.65it/s] 17%\|█▋ \| 497/2975 [08:54<24:58, 1.65it/s] 17%\|█▋
	0: {'loss': 0.7643, 'grad_norm': 0.754784845884012, 'learning_rate': 1e-05, 'epoch': 0.17}
	0: {'loss': 0.7816, 'grad_norm': 0.7816116256031539, 'learning_rate': 1e-05, 'epoch': 0.17}
	0: \| 498/2975 [08:55<24:56, 1.66it/s] 17%\|█▋ \| 499/2975 [08:55<25:28, 1.62it/s] 17%\|█▋ \| 500/2975 [08:56<25:23, 1.62it/s] 17%\|█▋ \| 500/2975 [08:56<25:23, 1.62it/s] 17%\|█▋ \| 501/2975 [08:57<25:44, 1.60it/s] 17%\|█▋ \| 502/2975 [08:57<25:29, 1.62it/s] 17%\|█▋ \| 503/2975 [08:58<25:18, 1.63it/s] 17%\|█▋ \| 504/2975 [08:58<25:09, 1.64it/s] 17%\|█▋ \| 505/2975 [08:59<25:28, 1.62it/s] 17%\|█▋ \| 506/2975 [09:00<25:23, 1.62it/s] 17%\|█▋ \| 507/2975 [09:00<25:18, 1.63it/s] 17%\|█▋ \| 508/2975 [09:01<25:12, 1.63it/s] 17%\|█▋ \| 509/2975 [09:02<25:06, 1.64it/s] 17%\|█▋ \| 510/2975 [09:02<25:02, 1.64it/s] 17%\|█▋ \| 510/2975 [09:02<25:02, 1.64it/s] 17%\|█▋ \| 511/2975 [09:03<25:00, 1.64it/s] 17%\|█▋ \| 512/2975 [09:03<24:59, 1.64it/s] 17%
	0: {'loss': 0.7499, 'grad_norm': 0.7626195725481466, 'learning_rate': 1e-05, 'epoch': 0.17}
	0: \|█▋ \| 513/2975 [09:04<24:55, 1.65it/s] 17%\|█▋ \| 514/2975 [09:05<24:52, 1.65it/s] 17%\|█▋ \| 515/2975 [09:05<24:50, 1.65it/s] 17%\|█▋ \| 516/2975 [09:06<24:47, 1.65it/s] 17%\|█▋ \| 517/2975 [09:06<24:45, 1.65it/s] 17%\|█▋ \| 518/2975 [09:07<24:44, 1.66it/s] 17%\|█▋ \| 519/2975 [09:08<24:42, 1.66it/s] 17%\|█▋ \| 520/2975 [09:08<24:41, 1.66it/s] 17%\|█▋ \| 520/2975 [09:08<24:41, 1.66it/s] 18%\|█▊ \| 521/2975 [09:09<24:41, 1.66it/s] 18%\|█▊ \| 522/2975 [09:09<24:41, 1.66it/s] 18%\|█▊ \| 523/2975 [09:10<24:41, 1.65it/s] 18%\|█▊ \| 524/2975 [09:11<24:40, 1.66it/s] 18%\|█▊ \| 525/2975 [09:11<24:38, 1.66it/s] 18%\|█▊ \| 526/2975 [09:12<24:37, 1.66it/s] 18%\|█▊ \| 527/2975 [09:12<24:36, 1.66it/s] 18%\|█▊ \| 528/2975 [09:13<24:34, 1.66it/s] 18%\|█▊ \| 529/2975 [09:14<24:3
	0: {'loss': 0.7507, 'grad_norm': 0.7542918658420218, 'learning_rate': 1e-05, 'epoch': 0.18}
	0: {'loss': 0.767, 'grad_norm': 0.7877507030671485, 'learning_rate': 1e-05, 'epoch': 0.18}
	0: 5, 1.66it/s] 18%\|█▊ \| 530/2975 [09:14<24:33, 1.66it/s] 18%\|█▊ \| 530/2975 [09:14<24:33, 1.66it/s] 18%\|█▊ \| 531/2975 [09:15<24:35, 1.66it/s] 18%\|█▊ \| 532/2975 [09:15<24:33, 1.66it/s] 18%\|█▊ \| 533/2975 [09:16<24:32, 1.66it/s] 18%\|█▊ \| 534/2975 [09:17<25:02, 1.62it/s] 18%\|█▊ \| 535/2975 [09:17<24:56, 1.63it/s] 18%\|█▊ \| 536/2975 [09:18<24:52, 1.63it/s] 18%\|█▊ \| 537/2975 [09:19<24:48, 1.64it/s] 18%\|█▊ \| 538/2975 [09:19<24:45, 1.64it/s] 18%\|█▊ \| 539/2975 [09:20<24:42, 1.64it/s] 18%\|█▊ \| 540/2975 [09:20<24:42, 1.64it/s] 18%\|█▊ \| 540/2975 [09:20<24:42, 1.64it/s] 18%\|█▊ \| 541/2975 [09:21<24:42, 1.64it/s] 18%\|█▊ \| 542/2975 [09:22<24:38, 1.65it/s] 18%\|█▊ \| 543/2975 [09:22<24:35, 1.65it/s] 18%\|█▊ \| 544/2
	0: {'loss': 0.7464, 'grad_norm': 0.7620889079841904, 'learning_rate': 1e-05, 'epoch': 0.18}
	0: 975 [09:23<24:34, 1.65it/s] 18%\|█▊ \| 545/2975 [09:23<24:34, 1.65it/s] 18%\|█▊ \| 546/2975 [09:24<24:33, 1.65it/s] 18%\|█▊ \| 547/2975 [09:25<24:31, 1.65it/s] 18%\|█▊ \| 548/2975 [09:25<24:28, 1.65it/s] 18%\|█▊ \| 549/2975 [09:26<24:51, 1.63it/s] 18%\|█▊ \| 550/2975 [09:26<24:47, 1.63it/s] 18%\|█▊ \| 550/2975 [09:26<24:47, 1.63it/s] 19%\|█▊ \| 551/2975 [09:27<24:42, 1.63it/s] 19%\|█▊ \| 552/2975 [09:28<24:39, 1.64it/s] 19%\|█▊ \| 553/2975 [09:28<24:52, 1.62it/s] 19%\|█▊ \| 554/2975 [09:29<25:12, 1.60it/s] 19%\|█▊ \| 555/2975 [09:30<24:57, 1.62it/s] 19%\|█▊ \| 556/2975 [09:30<24:50, 1.62it/s] 19%\|█▊ \| 557/2975 [09:31<24:45, 1.63it/s] 19%\|█▉ \| 558/2975 [09:31<24:41, 1.63it/s] 19%\|█▉ \| 559/2975 [09:32<24:36, 1.64it/s] 19%\|█▉ \| 560/2975 [09:33<24:34, 1.64it/s]
	0: {'loss': 0.7649, 'grad_norm': 0.8003034932416693, 'learning_rate': 1e-05, 'epoch': 0.19}
	0: {'loss': 0.7572, 'grad_norm': 0.7129834167859199, 'learning_rate': 1e-05, 'epoch': 0.19}
	0: 19%\|█▉ \| 560/2975 [09:33<24:34, 1.64it/s] 19%\|█▉ \| 561/2975 [09:33<24:37, 1.63it/s] 19%\|█▉ \| 562/2975 [09:34<24:35, 1.63it/s] 19%\|█▉ \| 563/2975 [09:34<24:30, 1.64it/s] 19%\|█▉ \| 564/2975 [09:35<24:25, 1.64it/s] 19%\|█▉ \| 565/2975 [09:36<24:23, 1.65it/s] 19%\|█▉ \| 566/2975 [09:36<24:31, 1.64it/s] 19%\|█▉ \| 567/2975 [09:37<24:27, 1.64it/s] 19%\|█▉ \| 568/2975 [09:37<24:23, 1.64it/s] 19%\|█▉ \| 569/2975 [09:38<24:20, 1.65it/s] 19%\|█▉ \| 570/2975 [09:39<24:15, 1.65it/s] 19%\|█▉ \| 570/2975 [09:39<24:15, 1.65it/s] 19%\|█▉ \| 571/2975 [09:39<24:15, 1.65it/s] 19%\|█▉ \| 572/2975 [09:40<24:14, 1.65it/s] 19%\|█▉ \| 573/2975 [09:40<24:12, 1.65it/s] 19%\|█▉ \| 574/2975 [09:41<24:09, 1.66it/s] 19%\|█▉ \| 575/2975 [09:42<24:07, 1.6
	0: {'loss': 0.7671, 'grad_norm': 0.7883813838017206, 'learning_rate': 1e-05, 'epoch': 0.19}
	0: {'loss': 0.7548, 'grad_norm': 0.8388848664403168, 'learning_rate': 1e-05, 'epoch': 0.2}
	0: 6it/s] 19%\|█▉ \| 576/2975 [09:42<24:07, 1.66it/s] 19%\|█▉ \| 577/2975 [09:43<24:08, 1.66it/s] 19%\|█▉ \| 578/2975 [09:43<24:08, 1.65it/s] 19%\|█▉ \| 579/2975 [09:44<24:11, 1.65it/s] 19%\|█▉ \| 580/2975 [09:45<24:11, 1.65it/s] 19%\|█▉ \| 580/2975 [09:45<24:11, 1.65it/s] 20%\|█▉ \| 581/2975 [09:45<24:11, 1.65it/s] 20%\|█▉ \| 582/2975 [09:46<24:13, 1.65it/s] 20%\|█▉ \| 583/2975 [09:47<24:13, 1.65it/s] 20%\|█▉ \| 584/2975 [09:47<24:12, 1.65it/s] 20%\|█▉ \| 585/2975 [09:48<24:10, 1.65it/s] 20%\|█▉ \| 586/2975 [09:48<24:09, 1.65it/s] 20%\|█▉ \| 587/2975 [09:49<24:09, 1.65it/s] 20%\|█▉ \| 588/2975 [09:50<24:07, 1.65it/s] 20%\|█▉ \| 589/2975 [09:50<24:20, 1.63it/s] 20%\|█▉ \| 590/2975 [09:51<24:15, 1.64it/s] 20%\|█▉ \| 590/2975 [09
	0: {'loss': 0.755, 'grad_norm': 0.7507527819754194, 'learning_rate': 1e-05, 'epoch': 0.2}
	0: :51<24:15, 1.64it/s] 20%\|█▉ \| 591/2975 [09:51<24:14, 1.64it/s] 20%\|█▉ \| 592/2975 [09:52<24:09, 1.64it/s] 20%\|█▉ \| 593/2975 [09:53<24:06, 1.65it/s] 20%\|█▉ \| 594/2975 [09:53<24:04, 1.65it/s] 20%\|██ \| 595/2975 [09:54<24:05, 1.65it/s] 20%\|██ \| 596/2975 [09:56<47:29, 1.20s/it] 20%\|██ \| 597/2975 [09:57<40:28, 1.02s/it] 20%\|██ \| 598/2975 [09:58<35:33, 1.11it/s] 20%\|██ \| 599/2975 [09:58<32:05, 1.23it/s] 20%\|██ \| 600/2975 [09:59<29:40, 1.33it/s] 20%\|██ \| 600/2975 [09:59<29:40, 1.33it/s] 20%\|██ \| 601/2975 [09:59<28:00, 1.41it/s] 20%\|██ \| 602/2975 [10:00<26:49, 1.47it/s] 20%\|██ \| 603/2975 [10:01<25:59, 1.52it/s] 20%\|██ \| 604/2975 [10:01<25:23, 1.56it/s] 20%\|██ \| 605/2975 [10:02<24:59, 1.58it/s] 20%\|██ \| 606/2975 [10:02<24:39, 1.60it/s] 20%\|██
	0: {'loss': 0.7547, 'grad_norm': 0.8164223892765191, 'learning_rate': 1e-05, 'epoch': 0.2}
	0: {'loss': 0.7495, 'grad_norm': 0.7622680122421158, 'learning_rate': 1e-05, 'epoch': 0.21}
	0: \| 607/2975 [10:03<24:27, 1.61it/s] 20%\|██ \| 608/2975 [10:04<24:18, 1.62it/s] 20%\|██ \| 609/2975 [10:04<24:10, 1.63it/s] 21%\|██ \| 610/2975 [10:05<24:05, 1.64it/s] 21%\|██ \| 610/2975 [10:05<24:05, 1.64it/s] 21%\|██ \| 611/2975 [10:06<24:02, 1.64it/s] 21%\|██ \| 612/2975 [10:06<23:59, 1.64it/s] 21%\|██ \| 613/2975 [10:07<23:56, 1.64it/s] 21%\|██ \| 614/2975 [10:07<23:54, 1.65it/s] 21%\|██ \| 615/2975 [10:08<23:51, 1.65it/s] 21%\|██ \| 616/2975 [10:09<23:48, 1.65it/s] 21%\|██ \| 617/2975 [10:09<23:48, 1.65it/s] 21%\|██ \| 618/2975 [10:10<23:45, 1.65it/s] 21%\|██ \| 619/2975 [10:10<23:44, 1.65it/s] 21%\|██ \| 620/2975 [10:11<23:45, 1.65it/s] 21%\|██ \| 620/2975 [10:11<23:45, 1.65it/s] 21%\|██ \| 621/2975 [10:12<23:48, 1.65it/s]
	0: {'loss': 0.7315, 'grad_norm': 0.7531375977537063, 'learning_rate': 1e-05, 'epoch': 0.21}
	0: 21%\|██ \| 622/2975 [10:12<23:50, 1.65it/s] 21%\|██ \| 623/2975 [10:13<24:07, 1.62it/s] 21%\|██ \| 624/2975 [10:13<24:03, 1.63it/s] 21%\|██ \| 625/2975 [10:14<24:14, 1.62it/s] 21%\|██ \| 626/2975 [10:15<24:03, 1.63it/s] 21%\|██ \| 627/2975 [10:15<23:55, 1.64it/s] 21%\|██ \| 628/2975 [10:16<23:50, 1.64it/s] 21%\|██ \| 629/2975 [10:16<23:46, 1.64it/s] 21%\|██ \| 630/2975 [10:17<23:42, 1.65it/s] 21%\|██ \| 630/2975 [10:17<23:42, 1.65it/s] 21%\|██ \| 631/2975 [10:18<23:43, 1.65it/s] 21%\|██ \| 632/2975 [10:18<23:42, 1.65it/s] 21%\|██▏ \| 633/2975 [10:19<23:40, 1.65it/s] 21%\|██▏ \| 634/2975 [10:19<23:41, 1.65it/s] 21%\|██▏ \| 635/2975 [10:20<23:40, 1.65it/s] 21%\|██▏ \| 636/2975 [10:21<23:39, 1.65it/s] 21%\|██▏ \| 637/2975 [10:21<23:39, 1.65it/s] 21%\|██▏ \| 638/
	0: {'loss': 0.7475, 'grad_norm': 0.744626566048428, 'learning_rate': 1e-05, 'epoch': 0.22}
	0: {'loss': 0.7474, 'grad_norm': 0.842353042268259, 'learning_rate': 1e-05, 'epoch': 0.22}
	0: 2975 [10:22<23:40, 1.65it/s] 21%\|██▏ \| 639/2975 [10:23<23:43, 1.64it/s] 22%\|██▏ \| 640/2975 [10:23<23:38, 1.65it/s] 22%\|██▏ \| 640/2975 [10:23<23:38, 1.65it/s] 22%\|██▏ \| 641/2975 [10:24<23:39, 1.64it/s] 22%\|██▏ \| 642/2975 [10:24<23:37, 1.65it/s] 22%\|██▏ \| 643/2975 [10:25<23:50, 1.63it/s] 22%\|██▏ \| 644/2975 [10:26<23:42, 1.64it/s] 22%\|██▏ \| 645/2975 [10:26<23:38, 1.64it/s] 22%\|██▏ \| 646/2975 [10:27<23:35, 1.65it/s] 22%\|██▏ \| 647/2975 [10:27<23:32, 1.65it/s] 22%\|██▏ \| 648/2975 [10:28<23:28, 1.65it/s] 22%\|██▏ \| 649/2975 [10:29<23:27, 1.65it/s] 22%\|██▏ \| 650/2975 [10:29<23:44, 1.63it/s] 22%\|██▏ \| 650/2975 [10:29<23:44, 1.63it/s] 22%\|██▏ \| 651/2975 [10:30<23:37, 1.64it/s] 22%\|██▏ \| 652/2975 [10
	0: {'loss': 0.7385, 'grad_norm': 0.7784520565914156, 'learning_rate': 1e-05, 'epoch': 0.22}
	0: :30<23:31, 1.65it/s] 22%\|██▏ \| 653/2975 [10:31<23:29, 1.65it/s] 22%\|██▏ \| 654/2975 [10:32<23:30, 1.65it/s] 22%\|██▏ \| 655/2975 [10:32<23:31, 1.64it/s] 22%\|██▏ \| 656/2975 [10:33<23:35, 1.64it/s] 22%\|██▏ \| 657/2975 [10:34<23:36, 1.64it/s] 22%\|██▏ \| 658/2975 [10:34<23:35, 1.64it/s] 22%\|██▏ \| 659/2975 [10:35<23:32, 1.64it/s] 22%\|██▏ \| 660/2975 [10:35<23:30, 1.64it/s] 22%\|██▏ \| 660/2975 [10:35<23:30, 1.64it/s] 22%\|██▏ \| 661/2975 [10:36<23:27, 1.64it/s] 22%\|██▏ \| 662/2975 [10:37<23:22, 1.65it/s] 22%\|██▏ \| 663/2975 [10:37<23:19, 1.65it/s] 22%\|██▏ \| 664/2975 [10:38<23:17, 1.65it/s] 22%\|██▏ \| 665/2975 [10:38<23:16, 1.65it/s] 22%\|██▏ \| 666/2975 [10:39<23:20, 1.65it/s] 22%\|██▏ \| 667/2975 [10:40<23:20, 1.65it/s] 22%\|██▏ \| 668/2975 [10:40
	0: {'loss': 0.7611, 'grad_norm': 0.7492523467278513, 'learning_rate': 1e-05, 'epoch': 0.23}
	0: {'loss': 0.7608, 'grad_norm': 0.7700016384605249, 'learning_rate': 1e-05, 'epoch': 0.23}
	0: <23:20, 1.65it/s] 22%\|██▏ \| 669/2975 [10:41<23:19, 1.65it/s] 23%\|██▎ \| 670/2975 [10:41<23:20, 1.65it/s] 23%\|██▎ \| 670/2975 [10:41<23:20, 1.65it/s] 23%\|██▎ \| 671/2975 [10:42<23:21, 1.64it/s] 23%\|██▎ \| 672/2975 [10:43<23:21, 1.64it/s] 23%\|██▎ \| 673/2975 [10:43<23:23, 1.64it/s] 23%\|██▎ \| 674/2975 [10:44<23:22, 1.64it/s] 23%\|██▎ \| 675/2975 [10:44<23:21, 1.64it/s] 23%\|██▎ \| 676/2975 [10:45<23:19, 1.64it/s] 23%\|██▎ \| 677/2975 [10:46<23:23, 1.64it/s] 23%\|██▎ \| 678/2975 [10:46<23:41, 1.62it/s] 23%\|██▎ \| 679/2975 [10:47<23:35, 1.62it/s] 23%\|██▎ \| 680/2975 [10:48<23:30, 1.63it/s] 23%\|██▎ \| 680/2975 [10:48<23:30, 1.63it/s] 23%\|██▎ \| 681/2975 [10:48<23:28, 1.63it/s] 23%\|██▎ \| 682/2975 [10:49<23:26,
	0: {'loss': 0.7482, 'grad_norm': 0.7657815124970616, 'learning_rate': 1e-05, 'epoch': 0.23}
	0: 1.63it/s] 23%\|██▎ \| 683/2975 [10:49<23:23, 1.63it/s] 23%\|██▎ \| 684/2975 [10:50<23:19, 1.64it/s] 23%\|██▎ \| 685/2975 [10:51<23:17, 1.64it/s] 23%\|██▎ \| 686/2975 [10:51<23:14, 1.64it/s] 23%\|██▎ \| 687/2975 [10:52<23:13, 1.64it/s] 23%\|██▎ \| 688/2975 [10:52<23:13, 1.64it/s] 23%\|██▎ \| 689/2975 [10:53<23:14, 1.64it/s] 23%\|██▎ \| 690/2975 [10:54<23:17, 1.64it/s] 23%\|██▎ \| 690/2975 [10:54<23:17, 1.64it/s] 23%\|██▎ \| 691/2975 [10:54<23:17, 1.63it/s] 23%\|██▎ \| 692/2975 [10:55<23:32, 1.62it/s] 23%\|██▎ \| 693/2975 [10:55<23:24, 1.62it/s] 23%\|██▎ \| 694/2975 [10:56<23:22, 1.63it/s] 23%\|██▎ \| 695/2975 [10:57<23:15, 1.63it/s] 23%\|██▎ \| 696/2975 [10:57<23:11, 1.64it/s] 23%\|██▎ \| 697/2975 [10:58<23:10, 1.64it/s] 23%\|██▎ \| 698/2975 [10:59<23:10, 1.
	0: {'loss': 0.7855, 'grad_norm': 0.7712451026387313, 'learning_rate': 1e-05, 'epoch': 0.24}
	0: {'loss': 0.7508, 'grad_norm': 0.7790654777806169, 'learning_rate': 1e-05, 'epoch': 0.24}
	0: 64it/s] 23%\|██▎ \| 699/2975 [10:59<23:12, 1.63it/s] 24%\|██▎ \| 700/2975 [11:00<23:13, 1.63it/s] 24%\|██▎ \| 700/2975 [11:00<23:13, 1.63it/s] 24%\|██▎ \| 701/2975 [11:00<23:13, 1.63it/s] 24%\|██▎ \| 702/2975 [11:01<23:09, 1.64it/s] 24%\|██▎ \| 703/2975 [11:02<23:05, 1.64it/s] 24%\|██▎ \| 704/2975 [11:02<23:02, 1.64it/s] 24%\|██▎ \| 705/2975 [11:03<23:01, 1.64it/s] 24%\|██▎ \| 706/2975 [11:03<23:02, 1.64it/s] 24%\|██▍ \| 707/2975 [11:04<23:01, 1.64it/s] 24%\|██▍ \| 708/2975 [11:05<23:01, 1.64it/s] 24%\|██▍ \| 709/2975 [11:05<23:00, 1.64it/s] 24%\|██▍ \| 710/2975 [11:06<23:01, 1.64it/s] 24%\|██▍ \| 710/2975 [11:06<23:01, 1.64it/s] 24%\|██▍ \| 711/2975 [11:06<23:04, 1.64it/s] 24%\|██▍ \| 712/2975 [11:07<23:05, 1.63it/s]
	0: {'loss': 0.742, 'grad_norm': 0.7958963482387478, 'learning_rate': 1e-05, 'epoch': 0.24}
	0: 24%\|██▍ \| 713/2975 [11:08<23:05, 1.63it/s] 24%\|██▍ \| 714/2975 [11:08<23:02, 1.64it/s] 24%\|██▍ \| 715/2975 [11:09<23:00, 1.64it/s] 24%\|██▍ \| 716/2975 [11:10<22:58, 1.64it/s] 24%\|██▍ \| 717/2975 [11:10<22:57, 1.64it/s] 24%\|██▍ \| 718/2975 [11:11<22:54, 1.64it/s] 24%\|██▍ \| 719/2975 [11:11<22:53, 1.64it/s] 24%\|██▍ \| 720/2975 [11:12<22:49, 1.65it/s] 24%\|██▍ \| 720/2975 [11:12<22:49, 1.65it/s] 24%\|██▍ \| 721/2975 [11:13<22:51, 1.64it/s] 24%\|██▍ \| 722/2975 [11:13<22:50, 1.64it/s] 24%\|██▍ \| 723/2975 [11:14<22:47, 1.65it/s] 24%\|██▍ \| 724/2975 [11:14<22:46, 1.65it/s] 24%\|██▍ \| 725/2975 [11:15<22:44, 1.65it/s] 24%\|██▍ \| 726/2975 [11:16<22:48, 1.64it/s] 24%\|██▍ \| 727/2975 [11:16<23:01, 1.63it/s] 24%\|██▍ \| 728/2975 [11:17<22:56, 1.63it/s] 25
	0: {'loss': 0.7683, 'grad_norm': 0.7584229493647255, 'learning_rate': 1e-05, 'epoch': 0.25}
	0: {'loss': 0.7302, 'grad_norm': 0.7917073972184041, 'learning_rate': 1e-05, 'epoch': 0.25}
	0: %\|██▍ \| 729/2975 [11:17<22:51, 1.64it/s] 25%\|██▍ \| 730/2975 [11:18<22:48, 1.64it/s] 25%\|██▍ \| 730/2975 [11:18<22:48, 1.64it/s] 25%\|██▍ \| 731/2975 [11:19<22:45, 1.64it/s] 25%\|██▍ \| 732/2975 [11:19<22:43, 1.65it/s] 25%\|██▍ \| 733/2975 [11:20<22:40, 1.65it/s] 25%\|██▍ \| 734/2975 [11:20<22:38, 1.65it/s] 25%\|██▍ \| 735/2975 [11:21<22:38, 1.65it/s] 25%\|██▍ \| 736/2975 [11:22<22:36, 1.65it/s] 25%\|██▍ \| 737/2975 [11:22<22:37, 1.65it/s] 25%\|██▍ \| 738/2975 [11:23<22:36, 1.65it/s] 25%\|██▍ \| 739/2975 [11:23<22:35, 1.65it/s] 25%\|██▍ \| 740/2975 [11:24<22:35, 1.65it/s] 25%\|██▍ \| 740/2975 [11:24<22:35, 1.65it/s] 25%\|██▍ \| 741/2975 [11:25<22:38, 1.64it/s] 25%\|██▍ \| 742/2975 [11:25<22:43, 1.64it/s] 25%\|██
	0: {'loss': 0.7357, 'grad_norm': 0.7683530057363408, 'learning_rate': 1e-05, 'epoch': 0.25}
	0: ▍ \| 743/2975 [11:26<22:43, 1.64it/s] 25%\|██▌ \| 744/2975 [11:27<22:43, 1.64it/s] 25%\|██▌ \| 745/2975 [11:27<22:43, 1.64it/s] 25%\|██▌ \| 746/2975 [11:28<22:40, 1.64it/s] 25%\|██▌ \| 747/2975 [11:28<22:38, 1.64it/s] 25%\|██▌ \| 748/2975 [11:29<22:35, 1.64it/s] 25%\|██▌ \| 749/2975 [11:30<22:35, 1.64it/s] 25%\|██▌ \| 750/2975 [11:30<22:34, 1.64it/s] 25%\|██▌ \| 750/2975 [11:30<22:34, 1.64it/s] 25%\|██▌ \| 751/2975 [11:31<22:33, 1.64it/s] 25%\|██▌ \| 752/2975 [11:31<22:29, 1.65it/s] 25%\|██▌ \| 753/2975 [11:32<22:26, 1.65it/s] 25%\|██▌ \| 754/2975 [11:33<23:07, 1.60it/s] 25%\|██▌ \| 755/2975 [11:33<22:58, 1.61it/s] 25%\|██▌ \| 756/2975 [11:34<22:50, 1.62it/s] 25%\|██▌ \| 757/2975 [11:35<22:43, 1.63it/s] 25%\|██▌ \| 758/2975 [11:35<22:38, 1.63it/s] 26%\|██▌
	0: {'loss': 0.7429, 'grad_norm': 0.732943318594699, 'learning_rate': 1e-05, 'epoch': 0.26}
	0: {'loss': 0.745, 'grad_norm': 0.7598610714362597, 'learning_rate': 1e-05, 'epoch': 0.26}
	0: \| 759/2975 [11:36<22:34, 1.64it/s] 26%\|██▌ \| 760/2975 [11:36<22:31, 1.64it/s] 26%\|██▌ \| 760/2975 [11:36<22:31, 1.64it/s] 26%\|██▌ \| 761/2975 [11:37<22:30, 1.64it/s] 26%\|██▌ \| 762/2975 [11:38<22:28, 1.64it/s] 26%\|██▌ \| 763/2975 [11:38<22:28, 1.64it/s] 26%\|██▌ \| 764/2975 [11:39<22:29, 1.64it/s] 26%\|██▌ \| 765/2975 [11:39<22:31, 1.64it/s] 26%\|██▌ \| 766/2975 [11:40<22:32, 1.63it/s] 26%\|██▌ \| 767/2975 [11:41<22:31, 1.63it/s] 26%\|██▌ \| 768/2975 [11:41<22:29, 1.64it/s] 26%\|██▌ \| 769/2975 [11:42<22:27, 1.64it/s] 26%\|██▌ \| 770/2975 [11:42<22:24, 1.64it/s] 26%\|██▌ \| 770/2975 [11:42<22:24, 1.64it/s] 26%\|██▌ \| 771/2975 [11:43<22:23, 1.64it/s] 26%\|██▌ \| 772/2975 [11:44<22:21, 1.64it/s] 26%\|██▌ \|
	0: {'loss': 0.7358, 'grad_norm': 0.7804717351373442, 'learning_rate': 1e-05, 'epoch': 0.26}
	0: 773/2975 [11:44<22:17, 1.65it/s] 26%\|██▌ \| 774/2975 [11:45<22:13, 1.65it/s] 26%\|██▌ \| 775/2975 [11:45<22:10, 1.65it/s] 26%\|██▌ \| 776/2975 [11:46<22:10, 1.65it/s] 26%\|██▌ \| 777/2975 [11:47<22:10, 1.65it/s] 26%\|██▌ \| 778/2975 [11:47<22:09, 1.65it/s] 26%\|██▌ \| 779/2975 [11:48<22:09, 1.65it/s] 26%\|██▌ \| 780/2975 [11:49<22:09, 1.65it/s] 26%\|██▌ \| 780/2975 [11:49<22:09, 1.65it/s] 26%\|██▋ \| 781/2975 [11:49<22:12, 1.65it/s] 26%\|██▋ \| 782/2975 [11:50<23:18, 1.57it/s] 26%\|██▋ \| 783/2975 [11:50<23:02, 1.59it/s] 26%\|██▋ \| 784/2975 [11:51<22:48, 1.60it/s] 26%\|██▋ \| 785/2975 [11:52<22:39, 1.61it/s] 26%\|██▋ \| 786/2975 [11:52<22:31, 1.62it/s] 26%\|██▋ \| 787/2975 [11:53<22:25, 1.63it/s] 26%\|██▋ \| 788/2975 [11:53<22:22, 1.63it/s] 27%\|██▋ \| 78
	0: {'loss': 0.7658, 'grad_norm': 0.749539480965823, 'learning_rate': 1e-05, 'epoch': 0.27}
	0: {'loss': 0.7477, 'grad_norm': 0.7999449395678967, 'learning_rate': 1e-05, 'epoch': 0.27}
	0: 9/2975 [11:54<22:17, 1.63it/s] 27%\|██▋ \| 790/2975 [11:55<22:14, 1.64it/s] 27%\|██▋ \| 790/2975 [11:55<22:14, 1.64it/s] 27%\|██▋ \| 791/2975 [11:55<22:11, 1.64it/s] 27%\|██▋ \| 792/2975 [11:56<22:07, 1.64it/s] 27%\|██▋ \| 793/2975 [11:57<22:02, 1.65it/s] 27%\|██▋ \| 794/2975 [11:57<21:59, 1.65it/s] 27%\|██▋ \| 795/2975 [11:58<21:56, 1.66it/s] 27%\|██▋ \| 796/2975 [11:58<21:55, 1.66it/s] 27%\|██▋ \| 797/2975 [11:59<21:56, 1.65it/s] 27%\|██▋ \| 798/2975 [12:00<21:58, 1.65it/s] 27%\|██▋ \| 799/2975 [12:00<22:01, 1.65it/s] 27%\|██▋ \| 800/2975 [12:01<22:04, 1.64it/s] 27%\|██▋ \| 800/2975 [12:01<22:04, 1.64it/s] 27%\|██▋ \| 801/2975 [12:01<22:09, 1.64it/s] 27%\|██▋ \| 802/2975 [12:02<22:14, 1.63it/s] 27%\|██▋ \| 803/2975 [
	0: {'loss': 0.7294, 'grad_norm': 0.7329522590368147, 'learning_rate': 1e-05, 'epoch': 0.27}
	0: 12:03<22:12, 1.63it/s] 27%\|██▋ \| 804/2975 [12:03<22:10, 1.63it/s] 27%\|██▋ \| 805/2975 [12:04<22:07, 1.63it/s] 27%\|██▋ \| 806/2975 [12:04<22:04, 1.64it/s] 27%\|██▋ \| 807/2975 [12:05<22:01, 1.64it/s] 27%\|██▋ \| 808/2975 [12:06<22:00, 1.64it/s] 27%\|██▋ \| 809/2975 [12:06<21:58, 1.64it/s] 27%\|██▋ \| 810/2975 [12:07<21:57, 1.64it/s] 27%\|██▋ \| 810/2975 [12:07<21:57, 1.64it/s] 27%\|██▋ \| 811/2975 [12:07<21:57, 1.64it/s] 27%\|██▋ \| 812/2975 [12:08<21:58, 1.64it/s] 27%\|██▋ \| 813/2975 [12:09<22:00, 1.64it/s] 27%\|██▋ \| 814/2975 [12:09<22:00, 1.64it/s] 27%\|██▋ \| 815/2975 [12:10<22:01, 1.63it/s] 27%\|██▋ \| 816/2975 [12:11<21:59, 1.64it/s] 27%\|██▋ \| 817/2975 [12:11<21:56, 1.64it/s] 27%\|██▋ \| 818/2975 [12:12<21:54, 1.64it/s] 28%\|██▊ \| 819/2975 [12:
	0: {'loss': 0.7372, 'grad_norm': 0.7607255267529802, 'learning_rate': 1e-05, 'epoch': 0.28}
	0: {'loss': 0.7557, 'grad_norm': 0.7377892901126538, 'learning_rate': 1e-05, 'epoch': 0.28}
	0: 12<21:53, 1.64it/s] 28%\|██▊ \| 820/2975 [12:13<21:49, 1.65it/s] 28%\|██▊ \| 820/2975 [12:13<21:49, 1.65it/s] 28%\|██▊ \| 821/2975 [12:14<21:49, 1.64it/s] 28%\|██▊ \| 822/2975 [12:14<21:49, 1.64it/s] 28%\|██▊ \| 823/2975 [12:15<21:48, 1.64it/s] 28%\|██▊ \| 824/2975 [12:15<21:49, 1.64it/s] 28%\|██▊ \| 825/2975 [12:16<21:50, 1.64it/s] 28%\|██▊ \| 826/2975 [12:17<21:52, 1.64it/s] 28%\|██▊ \| 827/2975 [12:17<21:51, 1.64it/s] 28%\|██▊ \| 828/2975 [12:18<21:51, 1.64it/s] 28%\|██▊ \| 829/2975 [12:18<21:50, 1.64it/s] 28%\|██▊ \| 830/2975 [12:19<21:48, 1.64it/s] 28%\|██▊ \| 830/2975 [12:19<21:48, 1.64it/s] 28%\|██▊ \| 831/2975 [12:20<21:47, 1.64it/s] 28%\|██▊ \| 832/2975 [12:20<21:48, 1.64it/s] 28%\|██▊ \| 833/2975 [12:21<21:48
	0: {'loss': 0.7693, 'grad_norm': 0.7531349345953301, 'learning_rate': 1e-05, 'epoch': 0.28}
	0: , 1.64it/s] 28%\|██▊ \| 834/2975 [12:22<21:45, 1.64it/s] 28%\|██▊ \| 835/2975 [12:22<21:42, 1.64it/s] 28%\|██▊ \| 836/2975 [12:23<21:40, 1.64it/s] 28%\|██▊ \| 837/2975 [12:23<21:41, 1.64it/s] 28%\|██▊ \| 838/2975 [12:24<21:40, 1.64it/s] 28%\|██▊ \| 839/2975 [12:25<21:37, 1.65it/s] 28%\|██▊ \| 840/2975 [12:25<21:35, 1.65it/s] 28%\|██▊ \| 840/2975 [12:25<21:35, 1.65it/s] 28%\|██▊ \| 841/2975 [12:26<21:38, 1.64it/s] 28%\|██▊ \| 842/2975 [12:26<21:38, 1.64it/s] 28%\|██▊ \| 843/2975 [12:27<21:38, 1.64it/s] 28%\|██▊ \| 844/2975 [12:28<21:36, 1.64it/s] 28%\|██▊ \| 845/2975 [12:28<21:36, 1.64it/s] 28%\|██▊ \| 846/2975 [12:29<21:35, 1.64it/s] 28%\|██▊ \| 847/2975 [12:29<21:34, 1.64it/s] 29%\|██▊ \| 848/2975 [12:30<21:33, 1.64it/s] 29%\|██▊ \| 849/2975 [12:31<21:31,
	0: {'loss': 0.7344, 'grad_norm': 0.7340430827625565, 'learning_rate': 1e-05, 'epoch': 0.29}
	0: {'loss': 0.7447, 'grad_norm': 0.7838462726074723, 'learning_rate': 1e-05, 'epoch': 0.29}
	0: 1.65it/s] 29%\|██▊ \| 850/2975 [12:31<21:30, 1.65it/s] 29%\|██▊ \| 850/2975 [12:31<21:30, 1.65it/s] 29%\|██▊ \| 851/2975 [12:32<21:30, 1.65it/s] 29%\|██▊ \| 852/2975 [12:33<23:24, 1.51it/s] 29%\|██▊ \| 853/2975 [12:33<23:23, 1.51it/s] 29%\|██▊ \| 854/2975 [12:34<22:51, 1.55it/s] 29%\|██▊ \| 855/2975 [12:35<22:26, 1.57it/s] 29%\|██▉ \| 856/2975 [12:35<22:09, 1.59it/s] 29%\|██▉ \| 857/2975 [12:36<21:58, 1.61it/s] 29%\|██▉ \| 858/2975 [12:36<21:49, 1.62it/s] 29%\|██▉ \| 859/2975 [12:37<21:41, 1.63it/s] 29%\|██▉ \| 860/2975 [12:38<21:38, 1.63it/s] 29%\|██▉ \| 860/2975 [12:38<21:38, 1.63it/s] 29%\|██▉ \| 861/2975 [12:38<21:35, 1.63it/s] 29%\|██▉ \| 862/2975 [12:39<21:33, 1.63it/s] 29%\|██▉ \| 863/2975 [12:39<22:02, 1.60it/s
	0: {'loss': 0.739, 'grad_norm': 0.8091788262795472, 'learning_rate': 1e-05, 'epoch': 0.29}
	0: ] 29%\|██▉ \| 864/2975 [12:40<21:55, 1.61it/s] 29%\|██▉ \| 865/2975 [12:41<21:45, 1.62it/s] 29%\|██▉ \| 866/2975 [12:41<21:39, 1.62it/s] 29%\|██▉ \| 867/2975 [12:42<21:57, 1.60it/s] 29%\|██▉ \| 868/2975 [12:43<21:48, 1.61it/s] 29%\|██▉ \| 869/2975 [12:43<22:06, 1.59it/s] 29%\|██▉ \| 870/2975 [12:44<21:58, 1.60it/s] 29%\|██▉ \| 870/2975 [12:44<21:58, 1.60it/s] 29%\|██▉ \| 871/2975 [12:44<21:49, 1.61it/s] 29%\|██▉ \| 872/2975 [12:45<21:38, 1.62it/s] 29%\|██▉ \| 873/2975 [12:46<21:30, 1.63it/s] 29%\|██▉ \| 874/2975 [12:46<21:26, 1.63it/s] 29%\|██▉ \| 875/2975 [12:47<21:26, 1.63it/s] 29%\|██▉ \| 876/2975 [12:47<21:24, 1.63it/s] 29%\|██▉ \| 877/2975 [12:48<21:21, 1.64it/s] 30%\|██▉ \| 878/2975 [12:49<21:19, 1.64it/s] 30%\|██▉ \| 879/2975 [12:49<21:17, 1.64it/s]
	0: {'loss': 0.7427, 'grad_norm': 0.7583526965895802, 'learning_rate': 1e-05, 'epoch': 0.3}
	0: {'loss': 0.7427, 'grad_norm': 0.7926390640915221, 'learning_rate': 1e-05, 'epoch': 0.3}
	0: 30%\|██▉ \| 880/2975 [12:50<21:17, 1.64it/s] 30%\|██▉ \| 880/2975 [12:50<21:17, 1.64it/s] 30%\|██▉ \| 881/2975 [12:51<21:16, 1.64it/s] 30%\|██▉ \| 882/2975 [12:51<21:13, 1.64it/s] 30%\|██▉ \| 883/2975 [12:52<21:16, 1.64it/s] 30%\|██▉ \| 884/2975 [12:52<22:18, 1.56it/s] 30%\|██▉ \| 885/2975 [12:53<21:56, 1.59it/s] 30%\|██▉ \| 886/2975 [12:54<21:46, 1.60it/s] 30%\|██▉ \| 887/2975 [12:54<21:37, 1.61it/s] 30%\|██▉ \| 888/2975 [12:55<21:30, 1.62it/s] 30%\|██▉ \| 889/2975 [12:55<21:24, 1.62it/s] 30%\|██▉ \| 890/2975 [12:56<21:19, 1.63it/s] 30%\|██▉ \| 890/2975 [12:56<21:19, 1.63it/s] 30%\|██▉ \| 891/2975 [12:57<21:15, 1.63it/s] 30%\|██▉ \| 892/2975 [12:57<21:13, 1.64it/s] 30%\|███ \| 893/2975 [12:58<21:11, 1.64it/s] 30%\|█�
	0: {'loss': 0.762, 'grad_norm': 0.7419955033128212, 'learning_rate': 1e-05, 'epoch': 0.3}
	0: ��█ \| 894/2975 [12:59<21:10, 1.64it/s] 30%\|███ \| 895/2975 [12:59<21:09, 1.64it/s] 30%\|███ \| 896/2975 [13:00<21:07, 1.64it/s] 30%\|███ \| 897/2975 [13:00<21:07, 1.64it/s] 30%\|███ \| 898/2975 [13:01<21:07, 1.64it/s] 30%\|███ \| 899/2975 [13:02<21:05, 1.64it/s] 30%\|███ \| 900/2975 [13:02<21:41, 1.59it/s] 30%\|███ \| 900/2975 [13:02<21:41, 1.59it/s] 30%\|███ \| 901/2975 [13:03<21:29, 1.61it/s] 30%\|███ \| 902/2975 [13:03<21:20, 1.62it/s] 30%\|███ \| 903/2975 [13:04<21:14, 1.63it/s] 30%\|███ \| 904/2975 [13:05<21:10, 1.63it/s] 30%\|███ \| 905/2975 [13:05<21:05, 1.64it/s] 30%\|███ \| 906/2975 [13:06<21:05, 1.64it/s] 30%\|███ \| 907/2975 [13:07<21:02, 1.64it/s] 31%\|███ \| 908/2975 [13:07<21:01, 1.64it/s] 31%\|███ \| 909/2975 [13:08<20:59, 1.64it/s] 31%\|██�
	0: {'loss': 0.7173, 'grad_norm': 0.7312540923740953, 'learning_rate': 1e-05, 'epoch': 0.31}
	0: {'loss': 0.7546, 'grad_norm': 0.8241647849928874, 'learning_rate': 1e-05, 'epoch': 0.31}
	0: �� \| 910/2975 [13:08<20:58, 1.64it/s] 31%\|███ \| 910/2975 [13:08<20:58, 1.64it/s] 31%\|███ \| 911/2975 [13:09<20:57, 1.64it/s] 31%\|███ \| 912/2975 [13:10<20:55, 1.64it/s] 31%\|███ \| 913/2975 [13:10<20:55, 1.64it/s] 31%\|███ \| 914/2975 [13:11<20:55, 1.64it/s] 31%\|███ \| 915/2975 [13:11<20:54, 1.64it/s] 31%\|███ \| 916/2975 [13:12<20:51, 1.64it/s] 31%\|███ \| 917/2975 [13:13<20:48, 1.65it/s] 31%\|███ \| 918/2975 [13:13<20:48, 1.65it/s] 31%\|███ \| 919/2975 [13:14<20:48, 1.65it/s] 31%\|███ \| 920/2975 [13:14<20:47, 1.65it/s] 31%\|███ \| 920/2975 [13:14<20:47, 1.65it/s] 31%\|███ \| 921/2975 [13:15<20:45, 1.65it/s] 31%\|███ \| 922/2975 [13:16<20:43, 1.65it/s] 31%\|███ \| 923/2975 [13:16<20:42, 1.65it/s] 31%\|███
	0: {'loss': 0.7578, 'grad_norm': 0.742815794399239, 'learning_rate': 1e-05, 'epoch': 0.31}
	0: \| 924/2975 [13:17<20:43, 1.65it/s] 31%\|███ \| 925/2975 [13:17<20:43, 1.65it/s] 31%\|███ \| 926/2975 [13:18<20:45, 1.64it/s] 31%\|███ \| 927/2975 [13:19<20:46, 1.64it/s] 31%\|███ \| 928/2975 [13:19<20:47, 1.64it/s] 31%\|███ \| 929/2975 [13:20<20:48, 1.64it/s] 31%\|███▏ \| 930/2975 [13:21<20:47, 1.64it/s] 31%\|███▏ \| 930/2975 [13:21<20:47, 1.64it/s] 31%\|███▏ \| 931/2975 [13:21<20:47, 1.64it/s] 31%\|███▏ \| 932/2975 [13:22<20:43, 1.64it/s] 31%\|███▏ \| 933/2975 [13:22<21:10, 1.61it/s] 31%\|███▏ \| 934/2975 [13:23<21:01, 1.62it/s] 31%\|███▏ \| 935/2975 [13:24<20:56, 1.62it/s] 31%\|███▏ \| 936/2975 [13:24<20:53, 1.63it/s] 31%\|███▏ \| 937/2975 [13:25<20:50, 1.63it/s] 32%\|███▏ \| 938/2975 [13:25<20:49, 1.63it/s] 32%\|███▏ \| 939/2975 [13:26<20:45, 1.64it/s]
	0: {'loss': 0.7538, 'grad_norm': 0.8110139545092059, 'learning_rate': 1e-05, 'epoch': 0.32}
	0: {'loss': 0.742, 'grad_norm': 0.7619714082364365, 'learning_rate': 1e-05, 'epoch': 0.32}
	0: 32%\|███▏ \| 940/2975 [13:27<20:41, 1.64it/s] 32%\|███▏ \| 940/2975 [13:27<20:41, 1.64it/s] 32%\|███▏ \| 941/2975 [13:27<20:40, 1.64it/s] 32%\|███▏ \| 942/2975 [13:28<20:38, 1.64it/s] 32%\|███▏ \| 943/2975 [13:28<20:37, 1.64it/s] 32%\|███▏ \| 944/2975 [13:29<20:38, 1.64it/s] 32%\|███▏ \| 945/2975 [13:30<20:38, 1.64it/s] 32%\|███▏ \| 946/2975 [13:30<20:39, 1.64it/s] 32%\|███▏ \| 947/2975 [13:31<20:38, 1.64it/s] 32%\|███▏ \| 948/2975 [13:32<20:39, 1.64it/s] 32%\|███▏ \| 949/2975 [13:32<20:39, 1.63it/s] 32%\|███▏ \| 950/2975 [13:33<20:39, 1.63it/s] 32%\|███▏ \| 950/2975 [13:33<20:39, 1.63it/s] 32%\|███▏ \| 951/2975 [13:33<20:39, 1.63it/s] 32%\|███▏ \| 952/2975 [13:34<20:36, 1.64it/s] 32%\|███▏ \| 953/2975 [1
	0: {'loss': 0.7647, 'grad_norm': 0.7569373817314637, 'learning_rate': 1e-05, 'epoch': 0.32}
	0: 3:35<20:36, 1.64it/s] 32%\|███▏ \| 954/2975 [13:35<20:32, 1.64it/s] 32%\|███▏ \| 955/2975 [13:36<20:29, 1.64it/s] 32%\|███▏ \| 956/2975 [13:36<20:27, 1.65it/s] 32%\|███▏ \| 957/2975 [13:37<20:27, 1.64it/s] 32%\|███▏ \| 958/2975 [13:38<20:26, 1.65it/s] 32%\|███▏ \| 959/2975 [13:38<20:25, 1.64it/s] 32%\|███▏ \| 960/2975 [13:39<20:26, 1.64it/s] 32%\|███▏ \| 960/2975 [13:39<20:26, 1.64it/s] 32%\|███▏ \| 961/2975 [13:39<20:28, 1.64it/s] 32%\|███▏ \| 962/2975 [13:40<20:30, 1.64it/s] 32%\|███▏ \| 963/2975 [13:41<20:32, 1.63it/s] 32%\|███▏ \| 964/2975 [13:41<20:30, 1.63it/s] 32%\|███▏ \| 965/2975 [13:42<20:27, 1.64it/s] 32%\|███▏ \| 966/2975 [13:43<20:24, 1.64it/s] 33%\|███▎ \| 967/2975 [13:43<20:21, 1.64it/s] 33%\|███▎ \| 968/2975 [13:44<20:19, 1.65it/s] 33%\|
	0: {'loss': 0.7504, 'grad_norm': 0.7201689920498835, 'learning_rate': 1e-05, 'epoch': 0.33}
	0: {'loss': 0.7495, 'grad_norm': 0.7858981946322333, 'learning_rate': 1e-05, 'epoch': 0.33}
	0: ███▎ \| 969/2975 [13:44<20:19, 1.65it/s] 33%\|███▎ \| 970/2975 [13:45<20:18, 1.65it/s] 33%\|███▎ \| 970/2975 [13:45<20:18, 1.65it/s] 33%\|███▎ \| 971/2975 [13:46<20:18, 1.64it/s] 33%\|███▎ \| 972/2975 [13:46<20:18, 1.64it/s] 33%\|███▎ \| 973/2975 [13:47<20:18, 1.64it/s] 33%\|███▎ \| 974/2975 [13:47<20:19, 1.64it/s] 33%\|███▎ \| 975/2975 [13:48<20:20, 1.64it/s] 33%\|███▎ \| 976/2975 [13:49<20:19, 1.64it/s] 33%\|███▎ \| 977/2975 [13:49<20:20, 1.64it/s] 33%\|███▎ \| 978/2975 [13:50<20:22, 1.63it/s] 33%\|███▎ \| 979/2975 [13:50<20:20, 1.64it/s] 33%\|███▎ \| 980/2975 [13:51<20:16, 1.64it/s] 33%\|███▎ \| 980/2975 [13:51<20:16, 1.64it/s] 33%\|███▎ \| 981/2975 [13:52<20:17, 1.64it/s] 33%\|███▎ \| 982/2975 [13:52
	0: {'loss': 0.7351, 'grad_norm': 0.8083503181543557, 'learning_rate': 1e-05, 'epoch': 0.33}
	0: <20:13, 1.64it/s] 33%\|███▎ \| 983/2975 [13:53<20:11, 1.64it/s] 33%\|███▎ \| 984/2975 [13:53<20:11, 1.64it/s] 33%\|███▎ \| 985/2975 [13:54<20:10, 1.64it/s] 33%\|███▎ \| 986/2975 [13:55<20:07, 1.65it/s] 33%\|███▎ \| 987/2975 [13:55<20:06, 1.65it/s] 33%\|███▎ \| 988/2975 [13:56<20:06, 1.65it/s] 33%\|███▎ \| 989/2975 [13:57<20:07, 1.65it/s] 33%\|███▎ \| 990/2975 [13:57<20:09, 1.64it/s] 33%\|███▎ \| 990/2975 [13:57<20:09, 1.64it/s] 33%\|███▎ \| 991/2975 [13:58<20:11, 1.64it/s] 33%\|███▎ \| 992/2975 [13:58<20:11, 1.64it/s] 33%\|███▎ \| 993/2975 [13:59<20:10, 1.64it/s] 33%\|███▎ \| 994/2975 [14:00<20:08, 1.64it/s] 33%\|███▎ \| 995/2975 [14:00<20:06, 1.64it/s] 33%\|███▎ \| 996/2975 [14:01<20:03, 1.64it/s] 34%\|███▎ \| 997/2975 [14:01<20:00, 1.65it/s] 34%\|█�
	0: {'loss': 0.7259, 'grad_norm': 0.7450275326316524, 'learning_rate': 1e-05, 'epoch': 0.34}
	0: {'loss': 0.7221, 'grad_norm': 0.7648787973503811, 'learning_rate': 1e-05, 'epoch': 0.34}
	0: ��█▎ \| 998/2975 [14:02<19:58, 1.65it/s] 34%\|███▎ \| 999/2975 [14:03<19:57, 1.65it/s] 34%\|███▎ \| 1000/2975 [14:03<19:56, 1.65it/s] 34%\|███▎ \| 1000/2975 [14:03<19:56, 1.65it/s] 34%\|███▎ \| 1001/2975 [14:04<19:59, 1.65it/s] 34%\|███▎ \| 1002/2975 [14:04<20:00, 1.64it/s] 34%\|███▎ \| 1003/2975 [14:05<19:58, 1.65it/s] 34%\|███▎ \| 1004/2975 [14:06<19:55, 1.65it/s] 34%\|███▍ \| 1005/2975 [14:06<19:54, 1.65it/s] 34%\|███▍ \| 1006/2975 [14:07<19:53, 1.65it/s] 34%\|███▍ \| 1007/2975 [14:07<19:53, 1.65it/s] 34%\|███▍ \| 1008/2975 [14:08<19:52, 1.65it/s] 34%\|███▍ \| 1009/2975 [14:09<19:51, 1.65it/s] 34%\|███▍ \| 1010/2975 [14:09<19:53, 1.65it/s] 34%\|███▍ \| 1010/2975 [14:09<19:53, 1.65it/s] 34%\|███▍ \| 1011
	0: {'loss': 0.7375, 'grad_norm': 0.8533932684671076, 'learning_rate': 1e-05, 'epoch': 0.34}
	0: /2975 [14:10<19:56, 1.64it/s] 34%\|███▍ \| 1012/2975 [14:10<19:54, 1.64it/s] 34%\|███▍ \| 1013/2975 [14:11<19:53, 1.64it/s] 34%\|███▍ \| 1014/2975 [14:12<19:54, 1.64it/s] 34%\|███▍ \| 1015/2975 [14:12<19:55, 1.64it/s] 34%\|███▍ \| 1016/2975 [14:13<19:54, 1.64it/s] 34%\|███▍ \| 1017/2975 [14:14<19:54, 1.64it/s] 34%\|███▍ \| 1018/2975 [14:14<19:55, 1.64it/s] 34%\|███▍ \| 1019/2975 [14:15<19:54, 1.64it/s] 34%\|███▍ \| 1020/2975 [14:15<19:54, 1.64it/s] 34%\|███▍ \| 1020/2975 [14:15<19:54, 1.64it/s] 34%\|███▍ \| 1021/2975 [14:16<19:56, 1.63it/s] 34%\|███▍ \| 1022/2975 [14:17<19:55, 1.63it/s] 34%\|███▍ \| 1023/2975 [14:17<19:54, 1.63it/s] 34%\|███▍ \| 1024/2975 [14:18<19:53, 1.64it/s] 34%\|███▍ \| 1025/2975 [14:18<19:51, 1.64it/s] 34%\|███▍ \| 1026/2975 [14:1
	0: {'loss': 0.7551, 'grad_norm': 0.7901510519026402, 'learning_rate': 1e-05, 'epoch': 0.35}
	0: 9<19:48, 1.64it/s] 35%\|███▍ \| 1027/2975 [14:20<19:46, 1.64it/s] 35%\|███▍ \| 1028/2975 [14:20<19:44, 1.64it/s] 35%\|███▍ \| 1029/2975 [14:21<19:45, 1.64it/s] 35%\|███▍ \| 1030/2975 [14:21<19:44, 1.64it/s] 35%\|███▍ \| 1030/2975 [14:21<19:44, 1.64it/s] 35%\|███▍ \| 1031/2975 [14:22<19:44, 1.64it/s] 35%\|███▍ \| 1032/2975 [14:23<19:44, 1.64it/s] 35%\|███▍ \| 1033/2975 [14:23<19:45, 1.64it/s] 35%\|███▍ \| 1034/2975 [14:24<19:45, 1.64it/s] 35%\|███▍ \| 1035/2975 [14:25<19:43, 1.64it/s] 35%\|███▍ \| 1036/2975 [14:25<19:41, 1.64it/s] 35%\|███▍ \| 1037/2975 [14:26<19:41, 1.64it/s] 35%\|███▍ \| 1038/2975 [14:26<19:40, 1.64it/s] 35%\|███▍ \| 1039/2975 [14:27<19:38, 1.64it/s] 35%\|███▍ \| 1040/2975 [14:28<19:36, 1.65it/s]
	0: {'loss': 0.7455, 'grad_norm': 0.7253433698188633, 'learning_rate': 1e-05, 'epoch': 0.35}
	0: {'loss': 0.7301, 'grad_norm': 0.7578554379488228, 'learning_rate': 1e-05, 'epoch': 0.35}
	0: 35%\|███▍ \| 1040/2975 [14:28<19:36, 1.65it/s] 35%\|███▍ \| 1041/2975 [14:28<19:34, 1.65it/s] 35%\|███▌ \| 1042/2975 [14:29<19:33, 1.65it/s] 35%\|███▌ \| 1043/2975 [14:29<19:33, 1.65it/s] 35%\|███▌ \| 1044/2975 [14:30<19:31, 1.65it/s] 35%\|███▌ \| 1045/2975 [14:31<19:40, 1.64it/s] 35%\|███▌ \| 1046/2975 [14:31<19:36, 1.64it/s] 35%\|███▌ \| 1047/2975 [14:32<19:34, 1.64it/s] 35%\|███▌ \| 1048/2975 [14:32<19:34, 1.64it/s] 35%\|███▌ \| 1049/2975 [14:33<19:34, 1.64it/s] 35%\|███▌ \| 1050/2975 [14:34<19:32, 1.64it/s] 35%\|███▌ \| 1050/2975 [14:34<19:32, 1.64it/s] 35%\|███▌ \| 1051/2975 [14:34<19:31, 1.64it/s] 35%\|███▌ \| 1052/2975 [14:35<19:29, 1.64it/s] 35%\|███▌ \| 1053/2975 [14:35<19:26, 1.65it/s] 35%\|███▌ \| 1054/2975 [14:36<19:25, 1.65it/s] 35%\|█�
	0: {'loss': 0.7492, 'grad_norm': 0.7766259324764115, 'learning_rate': 1e-05, 'epoch': 0.36}
	0: ��█▌ \| 1055/2975 [14:37<19:24, 1.65it/s] 35%\|███▌ \| 1056/2975 [14:37<21:22, 1.50it/s] 36%\|███▌ \| 1057/2975 [14:38<20:45, 1.54it/s] 36%\|███▌ \| 1058/2975 [14:39<20:18, 1.57it/s] 36%\|███▌ \| 1059/2975 [14:39<19:59, 1.60it/s] 36%\|███▌ \| 1060/2975 [14:40<19:47, 1.61it/s] 36%\|███▌ \| 1060/2975 [14:40<19:47, 1.61it/s] 36%\|███▌ \| 1061/2975 [14:41<19:39, 1.62it/s] 36%\|███▌ \| 1062/2975 [14:41<19:35, 1.63it/s] 36%\|███▌ \| 1063/2975 [14:42<19:33, 1.63it/s] 36%\|███▌ \| 1064/2975 [14:42<19:29, 1.63it/s] 36%\|███▌ \| 1065/2975 [14:43<19:28, 1.63it/s] 36%\|███▌ \| 1066/2975 [14:44<19:26, 1.64it/s] 36%\|███▌ \| 1067/2975 [14:44<19:25, 1.64it/s] 36%\|███▌ \| 1068/2975 [14:45<19:22, 1.64it/s] 36%\|███▌ \| 1069/2975 [14:45<19:20, 1.64it/s] 36%\|███▌
	0: {'loss': 0.7338, 'grad_norm': 0.8065374744219792, 'learning_rate': 1e-05, 'epoch': 0.36}
	0: {'loss': 0.7436, 'grad_norm': 0.7526874737814758, 'learning_rate': 1e-05, 'epoch': 0.36}
	0: \| 1070/2975 [14:46<19:17, 1.65it/s] 36%\|███▌ \| 1070/2975 [14:46<19:17, 1.65it/s] 36%\|███▌ \| 1071/2975 [14:47<19:17, 1.65it/s] 36%\|███▌ \| 1072/2975 [14:47<19:15, 1.65it/s] 36%\|███▌ \| 1073/2975 [14:48<19:15, 1.65it/s] 36%\|███▌ \| 1074/2975 [14:48<19:14, 1.65it/s] 36%\|███▌ \| 1075/2975 [14:49<19:12, 1.65it/s] 36%\|███▌ \| 1076/2975 [14:50<19:10, 1.65it/s] 36%\|███▌ \| 1077/2975 [14:50<19:10, 1.65it/s] 36%\|███▌ \| 1078/2975 [14:51<19:09, 1.65it/s] 36%\|███▋ \| 1079/2975 [14:51<19:12, 1.65it/s] 36%\|███▋ \| 1080/2975 [14:52<19:14, 1.64it/s] 36%\|███▋ \| 1080/2975 [14:52<19:14, 1.64it/s] 36%\|███▋ \| 1081/2975 [14:53<19:14, 1.64it/s] 36%\|███▋ \| 1082/2975 [14:53<19:15, 1.64it/s] 36%\|███▋ \| 1083/2975 [14
	0: {'loss': 0.7384, 'grad_norm': 0.8016191676712313, 'learning_rate': 1e-05, 'epoch': 0.37}
	0: :54<19:15, 1.64it/s] 36%\|███▋ \| 1084/2975 [14:55<19:11, 1.64it/s] 36%\|███▋ \| 1085/2975 [14:55<19:09, 1.64it/s] 37%\|███▋ \| 1086/2975 [14:56<19:06, 1.65it/s] 37%\|███▋ \| 1087/2975 [14:56<19:04, 1.65it/s] 37%\|███▋ \| 1088/2975 [14:57<19:02, 1.65it/s] 37%\|███▋ \| 1089/2975 [14:58<19:00, 1.65it/s] 37%\|███▋ \| 1090/2975 [14:58<18:58, 1.66it/s] 37%\|███▋ \| 1090/2975 [14:58<18:58, 1.66it/s] 37%\|███▋ \| 1091/2975 [14:59<18:57, 1.66it/s] 37%\|███▋ \| 1092/2975 [14:59<18:58, 1.65it/s] 37%\|███▋ \| 1093/2975 [15:00<18:57, 1.65it/s] 37%\|███▋ \| 1094/2975 [15:01<18:55, 1.66it/s] 37%\|███▋ \| 1095/2975 [15:01<18:54, 1.66it/s] 37%\|███▋ \| 1096/2975 [15:02<18:53, 1.66it/s] 37%\|███▋ \| 1097/2975 [15:02<18:53, 1.66it/s] 37%\|███▋ \| 1098/2975 [15:03<18:53,
	0: {'loss': 0.7236, 'grad_norm': 0.7541213322859073, 'learning_rate': 1e-05, 'epoch': 0.37}
	0: {'loss': 0.7439, 'grad_norm': 0.767711525718183, 'learning_rate': 1e-05, 'epoch': 0.37}
	0: 1.66it/s] 37%\|███▋ \| 1099/2975 [15:04<18:54, 1.65it/s] 37%\|███▋ \| 1100/2975 [15:04<18:55, 1.65it/s] 37%\|███▋ \| 1100/2975 [15:04<18:55, 1.65it/s] 37%\|███▋ \| 1101/2975 [15:05<18:56, 1.65it/s] 37%\|███▋ \| 1102/2975 [15:05<18:56, 1.65it/s] 37%\|███▋ \| 1103/2975 [15:06<18:57, 1.65it/s] 37%\|███▋ \| 1104/2975 [15:07<18:57, 1.65it/s] 37%\|███▋ \| 1105/2975 [15:07<18:59, 1.64it/s] 37%\|███▋ \| 1106/2975 [15:08<18:59, 1.64it/s] 37%\|███▋ \| 1107/2975 [15:08<18:57, 1.64it/s] 37%\|███▋ \| 1108/2975 [15:09<18:55, 1.64it/s] 37%\|███▋ \| 1109/2975 [15:10<18:53, 1.65it/s] 37%\|███▋ \| 1110/2975 [15:10<18:53, 1.65it/s] 37%\|███▋ \| 1110/2975 [15:10<18:53, 1.65it/s] 37%\|███▋ \| 1111/2975 [15:11<18:54, 1.64it/s] 37%\|�
	0: {'loss': 0.7662, 'grad_norm': 0.7476627932421003, 'learning_rate': 1e-05, 'epoch': 0.38}
	0: �██▋ \| 1112/2975 [15:11<18:52, 1.65it/s] 37%\|███▋ \| 1113/2975 [15:12<18:50, 1.65it/s] 37%\|███▋ \| 1114/2975 [15:13<18:47, 1.65it/s] 37%\|███▋ \| 1115/2975 [15:13<18:46, 1.65it/s] 38%\|███▊ \| 1116/2975 [15:14<18:45, 1.65it/s] 38%\|███▊ \| 1117/2975 [15:15<18:43, 1.65it/s] 38%\|███▊ \| 1118/2975 [15:15<18:42, 1.65it/s] 38%\|███▊ \| 1119/2975 [15:16<18:40, 1.66it/s] 38%\|███▊ \| 1120/2975 [15:16<18:39, 1.66it/s] 38%\|███▊ \| 1120/2975 [15:16<18:39, 1.66it/s] 38%\|███▊ \| 1121/2975 [15:17<18:39, 1.66it/s] 38%\|███▊ \| 1122/2975 [15:18<18:37, 1.66it/s] 38%\|███▊ \| 1123/2975 [15:18<18:37, 1.66it/s] 38%\|███▊ \| 1124/2975 [15:19<18:36, 1.66it/s] 38%\|███▊ \| 1125/2975 [15:19<18:35, 1.66it/s] 38%\|███▊ \| 1126/2975 [15:20<18:35, 1.66it/s] 38%\|███▊
	0: {'loss': 0.7154, 'grad_norm': 0.7440026914213119, 'learning_rate': 1e-05, 'epoch': 0.38}
	0: {'loss': 0.7241, 'grad_norm': 0.7597698137132717, 'learning_rate': 1e-05, 'epoch': 0.38}
	0: \| 1127/2975 [15:21<18:34, 1.66it/s] 38%\|███▊ \| 1128/2975 [15:21<18:34, 1.66it/s] 38%\|███▊ \| 1129/2975 [15:22<18:32, 1.66it/s] 38%\|███▊ \| 1130/2975 [15:22<18:32, 1.66it/s] 38%\|███▊ \| 1130/2975 [15:22<18:32, 1.66it/s] 38%\|███▊ \| 1131/2975 [15:23<18:33, 1.66it/s] 38%\|███▊ \| 1132/2975 [15:24<18:33, 1.65it/s] 38%\|███▊ \| 1133/2975 [15:24<18:33, 1.65it/s] 38%\|███▊ \| 1134/2975 [15:25<18:34, 1.65it/s] 38%\|███▊ \| 1135/2975 [15:25<18:56, 1.62it/s] 38%\|███▊ \| 1136/2975 [15:26<18:50, 1.63it/s] 38%\|███▊ \| 1137/2975 [15:27<18:44, 1.63it/s] 38%\|███▊ \| 1138/2975 [15:27<18:40, 1.64it/s] 38%\|███▊ \| 1139/2975 [15:28<18:38, 1.64it/s] 38%\|███▊ \| 1140/2975 [15:28<18:35, 1.65it/s] 38%\|███▊ \| 1140/2975 [
	0: {'loss': 0.7401, 'grad_norm': 0.7943983699535594, 'learning_rate': 1e-05, 'epoch': 0.39}
	0: 15:28<18:35, 1.65it/s] 38%\|███▊ \| 1141/2975 [15:29<18:36, 1.64it/s] 38%\|███▊ \| 1142/2975 [15:30<18:33, 1.65it/s] 38%\|███▊ \| 1143/2975 [15:30<18:31, 1.65it/s] 38%\|███▊ \| 1144/2975 [15:31<18:29, 1.65it/s] 38%\|███▊ \| 1145/2975 [15:31<18:27, 1.65it/s] 39%\|███▊ \| 1146/2975 [15:32<18:26, 1.65it/s] 39%\|███▊ \| 1147/2975 [15:33<18:25, 1.65it/s] 39%\|███▊ \| 1148/2975 [15:33<18:25, 1.65it/s] 39%\|███▊ \| 1149/2975 [15:34<18:26, 1.65it/s] 39%\|███▊ \| 1150/2975 [15:35<18:28, 1.65it/s] 39%\|███▊ \| 1150/2975 [15:35<18:28, 1.65it/s] 39%\|███▊ \| 1151/2975 [15:35<18:31, 1.64it/s] 39%\|███▊ \| 1152/2975 [15:36<18:38, 1.63it/s] 39%\|███▉ \| 1153/2975 [15:36<18:35, 1.63it/s] 39%\|███▉ \| 1154/2975 [15:37<19:00, 1.60it/s] 39%\|███▉ \| 1155/2975 [15:38<18:49
	0: {'loss': 0.741, 'grad_norm': 0.755713448663527, 'learning_rate': 1e-05, 'epoch': 0.39}
	0: , 1.61it/s] 39%\|███▉ \| 1156/2975 [15:38<18:41, 1.62it/s] 39%\|███▉ \| 1157/2975 [15:39<18:37, 1.63it/s] 39%\|███▉ \| 1158/2975 [15:39<18:34, 1.63it/s] 39%\|███▉ \| 1159/2975 [15:40<18:32, 1.63it/s] 39%\|███▉ \| 1160/2975 [15:41<18:29, 1.64it/s] 39%\|███▉ \| 1160/2975 [15:41<18:29, 1.64it/s] 39%\|███▉ \| 1161/2975 [15:41<18:27, 1.64it/s] 39%\|███▉ \| 1162/2975 [15:42<18:24, 1.64it/s] 39%\|███▉ \| 1163/2975 [15:42<18:23, 1.64it/s] 39%\|███▉ \| 1164/2975 [15:43<18:22, 1.64it/s] 39%\|███▉ \| 1165/2975 [15:44<18:20, 1.64it/s] 39%\|███▉ \| 1166/2975 [15:44<18:19, 1.65it/s] 39%\|███▉ \| 1167/2975 [15:45<18:17, 1.65it/s] 39%\|███▉ \| 1168/2975 [15:46<18:15, 1.65it/s] 39%\|███▉ \| 1169/2975 [15:46<18:16, 1.65it/s] 39%\|███▉ \| 1170/2975 [15:47<18:15, 1.65it/s
	0: {'loss': 0.7348, 'grad_norm': 0.763836705857911, 'learning_rate': 1e-05, 'epoch': 0.39}
	0: {'loss': 0.7486, 'grad_norm': 0.7260225673474382, 'learning_rate': 1e-05, 'epoch': 0.4}
	0: ] 39%\|███▉ \| 1170/2975 [15:47<18:15, 1.65it/s] 39%\|███▉ \| 1171/2975 [15:47<18:17, 1.64it/s] 39%\|███▉ \| 1172/2975 [15:48<18:17, 1.64it/s] 39%\|███▉ \| 1173/2975 [15:49<18:15, 1.65it/s] 39%\|███▉ \| 1174/2975 [15:49<18:16, 1.64it/s] 39%\|███▉ \| 1175/2975 [15:50<18:16, 1.64it/s] 40%\|███▉ \| 1176/2975 [15:50<18:16, 1.64it/s] 40%\|███▉ \| 1177/2975 [15:51<18:14, 1.64it/s] 40%\|███▉ \| 1178/2975 [15:52<18:13, 1.64it/s] 40%\|███▉ \| 1179/2975 [15:52<18:12, 1.64it/s] 40%\|███▉ \| 1180/2975 [15:53<18:11, 1.64it/s] 40%\|███▉ \| 1180/2975 [15:53<18:11, 1.64it/s] 40%\|███▉ \| 1181/2975 [15:53<18:10, 1.64it/s] 40%\|███▉ \| 1182/2975 [15:54<18:11, 1.64it/s] 40%\|███▉ \| 1183/2975 [15:55<18:09, 1.64it/s] 40%\|███�
	0: {'loss': 0.7411, 'grad_norm': 0.7526510518592818, 'learning_rate': 1e-05, 'epoch': 0.4}
	0: � \| 1184/2975 [15:55<18:07, 1.65it/s] 40%\|███▉ \| 1185/2975 [15:56<18:04, 1.65it/s] 40%\|███▉ \| 1186/2975 [15:56<18:13, 1.64it/s] 40%\|███▉ \| 1187/2975 [15:57<18:25, 1.62it/s] 40%\|███▉ \| 1188/2975 [15:58<18:19, 1.63it/s] 40%\|███▉ \| 1189/2975 [15:58<18:15, 1.63it/s] 40%\|████ \| 1190/2975 [15:59<18:12, 1.63it/s] 40%\|████ \| 1190/2975 [15:59<18:12, 1.63it/s] 40%\|████ \| 1191/2975 [16:02<38:34, 1.30s/it] 40%\|████ \| 1192/2975 [16:02<32:25, 1.09s/it] 40%\|████ \| 1193/2975 [16:03<28:07, 1.06it/s] 40%\|████ \| 1194/2975 [16:04<25:06, 1.18it/s] 40%\|████ \| 1195/2975 [16:04<22:59, 1.29it/s] 40%\|████ \| 1196/2975 [16:05<21:29, 1.38it/s] 40%\|████ \| 1197/2975 [16:06<20:25, 1.45it/s] 40%\|████ \| 1198/2975 [16:06<19:57, 1.48it/s] 40%\|████ \| 11
	0: {'loss': 0.708, 'grad_norm': 0.763605139895749, 'learning_rate': 1e-05, 'epoch': 0.4}
	0: {'loss': 0.7189, 'grad_norm': 0.743714237682584, 'learning_rate': 1e-05, 'epoch': 0.41}
	0: 99/2975 [16:07<19:23, 1.53it/s] 40%\|████ \| 1200/2975 [16:07<18:58, 1.56it/s] 40%\|████ \| 1200/2975 [16:07<18:58, 1.56it/s] 40%\|████ \| 1201/2975 [16:08<18:41, 1.58it/s] 40%\|████ \| 1202/2975 [16:09<18:26, 1.60it/s] 40%\|████ \| 1203/2975 [16:09<18:18, 1.61it/s] 40%\|████ \| 1204/2975 [16:10<18:12, 1.62it/s] 41%\|████ \| 1205/2975 [16:10<18:07, 1.63it/s] 41%\|████ \| 1206/2975 [16:11<18:01, 1.64it/s] 41%\|████ \| 1207/2975 [16:12<17:56, 1.64it/s] 41%\|████ \| 1208/2975 [16:12<17:52, 1.65it/s] 41%\|████ \| 1209/2975 [16:13<17:50, 1.65it/s] 41%\|████ \| 1210/2975 [16:13<17:50, 1.65it/s] 41%\|████ \| 1210/2975 [16:13<17:50, 1.65it/s] 41%\|████ \| 1211/2975 [16:14<17:49, 1.65it/s] 41%\|████ \| 1212/2975 [16:15<17:
	0: {'loss': 0.7215, 'grad_norm': 0.7671818270062326, 'learning_rate': 1e-05, 'epoch': 0.41}
	0: 46, 1.65it/s] 41%\|████ \| 1213/2975 [16:15<17:45, 1.65it/s] 41%\|████ \| 1214/2975 [16:16<17:45, 1.65it/s] 41%\|████ \| 1215/2975 [16:16<17:46, 1.65it/s] 41%\|████ \| 1216/2975 [16:17<17:46, 1.65it/s] 41%\|████ \| 1217/2975 [16:18<17:44, 1.65it/s] 41%\|████ \| 1218/2975 [16:18<17:44, 1.65it/s] 41%\|████ \| 1219/2975 [16:19<17:43, 1.65it/s] 41%\|████ \| 1220/2975 [16:19<17:43, 1.65it/s] 41%\|████ \| 1220/2975 [16:19<17:43, 1.65it/s] 41%\|████ \| 1221/2975 [16:20<17:44, 1.65it/s] 41%\|████ \| 1222/2975 [16:21<17:43, 1.65it/s] 41%\|████ \| 1223/2975 [16:21<17:42, 1.65it/s] 41%\|████ \| 1224/2975 [16:22<17:41, 1.65it/s] 41%\|████ \| 1225/2975 [16:23<17:40, 1.65it/s] 41%\|████ \| 1226/2975 [16:23<17:39, 1.65it/s] 41%\|████ \| 1227/2975 [16:24<17:43, 1.64it
	0: {'loss': 0.7156, 'grad_norm': 0.7977023629995006, 'learning_rate': 1e-05, 'epoch': 0.41}
	0: {'loss': 0.733, 'grad_norm': 0.7843304726970324, 'learning_rate': 1e-05, 'epoch': 0.42}
	0: /s] 41%\|████▏ \| 1228/2975 [16:24<17:43, 1.64it/s] 41%\|████▏ \| 1229/2975 [16:25<18:03, 1.61it/s] 41%\|████▏ \| 1230/2975 [16:26<17:52, 1.63it/s] 41%\|████▏ \| 1230/2975 [16:26<17:52, 1.63it/s] 41%\|████▏ \| 1231/2975 [16:26<17:46, 1.64it/s] 41%\|████▏ \| 1232/2975 [16:27<17:41, 1.64it/s] 41%\|████▏ \| 1233/2975 [16:27<17:38, 1.65it/s] 41%\|████▏ \| 1234/2975 [16:28<17:35, 1.65it/s] 42%\|████▏ \| 1235/2975 [16:29<17:32, 1.65it/s] 42%\|████▏ \| 1236/2975 [16:29<17:41, 1.64it/s] 42%\|████▏ \| 1237/2975 [16:30<17:38, 1.64it/s] 42%\|████▏ \| 1238/2975 [16:30<17:36, 1.64it/s] 42%\|████▏ \| 1239/2975 [16:31<17:35, 1.65it/s] 42%\|████▏ \| 1240/2975 [16:32<17:35, 1.64it/s] 42%\|████▏ \| 1240/2975 [16:32<17
	0: {'loss': 0.7065, 'grad_norm': 0.7143588421452666, 'learning_rate': 1e-05, 'epoch': 0.42}
	0: :35, 1.64it/s] 42%\|████▏ \| 1241/2975 [16:32<17:37, 1.64it/s] 42%\|████▏ \| 1242/2975 [16:33<17:36, 1.64it/s] 42%\|████▏ \| 1243/2975 [16:34<18:26, 1.57it/s] 42%\|████▏ \| 1244/2975 [16:34<18:22, 1.57it/s] 42%\|████▏ \| 1245/2975 [16:35<18:05, 1.59it/s] 42%\|████▏ \| 1246/2975 [16:36<18:33, 1.55it/s] 42%\|████▏ \| 1247/2975 [16:36<18:27, 1.56it/s] 42%\|████▏ \| 1248/2975 [16:37<18:07, 1.59it/s] 42%\|████▏ \| 1249/2975 [16:37<17:53, 1.61it/s] 42%\|████▏ \| 1250/2975 [16:38<17:43, 1.62it/s] 42%\|████▏ \| 1250/2975 [16:38<17:43, 1.62it/s] 42%\|████▏ \| 1251/2975 [16:39<17:37, 1.63it/s] 42%\|████▏ \| 1252/2975 [16:39<17:32, 1.64it/s] 42%\|████▏ \| 1253/2975 [16:40<17:28, 1.64it/s] 42%\|████▏ \| 1254/2975 [16:40<17:25, 1.65it/s] 42%\|████▏
	0: {'loss': 0.726, 'grad_norm': 0.7248263242252937, 'learning_rate': 1e-05, 'epoch': 0.42}
	0: \| 1255/2975 [16:41<17:23, 1.65it/s] 42%\|████▏ \| 1256/2975 [16:42<18:16, 1.57it/s] 42%\|████▏ \| 1257/2975 [16:42<18:32, 1.54it/s] 42%\|████▏ \| 1258/2975 [16:43<18:10, 1.57it/s] 42%\|████▏ \| 1259/2975 [16:44<17:56, 1.59it/s] 42%\|████▏ \| 1260/2975 [16:44<17:43, 1.61it/s] 42%\|████▏ \| 1260/2975 [16:44<17:43, 1.61it/s] 42%\|████▏ \| 1261/2975 [16:45<17:37, 1.62it/s] 42%\|████▏ \| 1262/2975 [16:45<17:30, 1.63it/s] 42%\|████▏ \| 1263/2975 [16:46<17:25, 1.64it/s] 42%\|████▏ \| 1264/2975 [16:47<17:20, 1.64it/s] 43%\|████▎ \| 1265/2975 [16:47<17:17, 1.65it/s] 43%\|████▎ \| 1266/2975 [16:48<17:15, 1.65it/s] 43%\|████▎ \| 1267/2975 [16:48<17:14, 1.65it/s] 43%\|████▎ \| 1268/2975 [16:49<17:16, 1.65it/s] 43%\|████▎ \| 1269/2975 [16:50<17:15, 1.65it/s] 43%
	0: {'loss': 0.719, 'grad_norm': 0.7889701499095937, 'learning_rate': 1e-05, 'epoch': 0.43}
	0: {'loss': 0.7321, 'grad_norm': 0.7621716701405825, 'learning_rate': 1e-05, 'epoch': 0.43}
	0: \|████▎ \| 1270/2975 [16:50<17:14, 1.65it/s] 43%\|████▎ \| 1270/2975 [16:50<17:14, 1.65it/s] 43%\|████▎ \| 1271/2975 [16:51<17:16, 1.64it/s] 43%\|████▎ \| 1272/2975 [16:51<17:15, 1.64it/s] 43%\|████▎ \| 1273/2975 [16:52<17:13, 1.65it/s] 43%\|████▎ \| 1274/2975 [16:53<17:12, 1.65it/s] 43%\|████▎ \| 1275/2975 [16:53<17:12, 1.65it/s] 43%\|████▎ \| 1276/2975 [16:54<17:12, 1.65it/s] 43%\|████▎ \| 1277/2975 [16:54<17:11, 1.65it/s] 43%\|████▎ \| 1278/2975 [16:55<17:10, 1.65it/s] 43%\|████▎ \| 1279/2975 [16:56<17:08, 1.65it/s] 43%\|████▎ \| 1280/2975 [16:56<17:07, 1.65it/s] 43%\|████▎ \| 1280/2975 [16:56<17:07, 1.65it/s] 43%\|████▎ \| 1281/2975 [16:57<17:06, 1.65it/s] 43%\|████▎ \| 1282/2975 [16:58<17:05, 1.
	0: {'loss': 0.74, 'grad_norm': 0.7827637856545197, 'learning_rate': 1e-05, 'epoch': 0.43}
	0: 65it/s] 43%\|████▎ \| 1283/2975 [16:58<17:04, 1.65it/s] 43%\|████▎ \| 1284/2975 [16:59<17:04, 1.65it/s] 43%\|████▎ \| 1285/2975 [16:59<17:07, 1.64it/s] 43%\|████▎ \| 1286/2975 [17:00<17:07, 1.64it/s] 43%\|████▎ \| 1287/2975 [17:01<17:05, 1.65it/s] 43%\|████▎ \| 1288/2975 [17:01<17:04, 1.65it/s] 43%\|████▎ \| 1289/2975 [17:02<17:03, 1.65it/s] 43%\|████▎ \| 1290/2975 [17:02<17:03, 1.65it/s] 43%\|████▎ \| 1290/2975 [17:02<17:03, 1.65it/s] 43%\|████▎ \| 1291/2975 [17:03<17:02, 1.65it/s] 43%\|████▎ \| 1292/2975 [17:04<17:01, 1.65it/s] 43%\|████▎ \| 1293/2975 [17:04<17:00, 1.65it/s] 43%\|████▎ \| 1294/2975 [17:05<16:58, 1.65it/s] 44%\|████▎ \| 1295/2975 [17:05<16:55, 1.65it/s] 44%\|████▎ \| 1296/2975 [17:06<16:54, 1.66it/s] 44%\|████▎ \| 1297/2
	0: {'loss': 0.7228, 'grad_norm': 0.7484708286727815, 'learning_rate': 1e-05, 'epoch': 0.44}
	0: {'loss': 0.7199, 'grad_norm': 0.743631669253817, 'learning_rate': 1e-05, 'epoch': 0.44}
	0: 975 [17:07<16:52, 1.66it/s] 44%\|████▎ \| 1298/2975 [17:07<16:52, 1.66it/s] 44%\|████▎ \| 1299/2975 [17:08<16:51, 1.66it/s] 44%\|████▎ \| 1300/2975 [17:08<16:50, 1.66it/s] 44%\|████▎ \| 1300/2975 [17:08<16:50, 1.66it/s] 44%\|████▎ \| 1301/2975 [17:09<16:50, 1.66it/s] 44%\|████▍ \| 1302/2975 [17:10<16:48, 1.66it/s] 44%\|████▍ \| 1303/2975 [17:10<16:49, 1.66it/s] 44%\|████▍ \| 1304/2975 [17:11<16:50, 1.65it/s] 44%\|████▍ \| 1305/2975 [17:11<16:49, 1.65it/s] 44%\|████▍ \| 1306/2975 [17:12<16:48, 1.65it/s] 44%\|████▍ \| 1307/2975 [17:13<16:47, 1.66it/s] 44%\|████▍ \| 1308/2975 [17:13<16:48, 1.65it/s] 44%\|████▍ \| 1309/2975 [17:14<17:10, 1.62it/s] 44%\|████▍ \| 1310/2975 [17:15<17:09, 1.62it/s] 44%\|████▍
	0: {'loss': 0.7227, 'grad_norm': 0.7640136035361755, 'learning_rate': 1e-05, 'epoch': 0.44}
	0: \| 1310/2975 [17:15<17:09, 1.62it/s] 44%\|████▍ \| 1311/2975 [17:15<17:04, 1.62it/s] 44%\|████▍ \| 1312/2975 [17:16<16:58, 1.63it/s] 44%\|████▍ \| 1313/2975 [17:16<16:54, 1.64it/s] 44%\|████▍ \| 1314/2975 [17:17<16:52, 1.64it/s] 44%\|████▍ \| 1315/2975 [17:18<16:48, 1.65it/s] 44%\|████▍ \| 1316/2975 [17:18<16:45, 1.65it/s] 44%\|████▍ \| 1317/2975 [17:19<16:43, 1.65it/s] 44%\|████▍ \| 1318/2975 [17:19<16:42, 1.65it/s] 44%\|████▍ \| 1319/2975 [17:20<16:42, 1.65it/s] 44%\|████▍ \| 1320/2975 [17:21<16:41, 1.65it/s] 44%\|████▍ \| 1320/2975 [17:21<16:41, 1.65it/s] 44%\|████▍ \| 1321/2975 [17:21<17:41, 1.56it/s] 44%\|████▍ \| 1322/2975 [17:22<17:51, 1.54it/s] 44%\|████▍ \| 1323/2975 [17:23<17:29, 1.57it/s] 45%\|████▍ \| 1324/2975 [17:23<17:14, 1.60it/s]
	0: {'loss': 0.7302, 'grad_norm': 0.7597085554218037, 'learning_rate': 1e-05, 'epoch': 0.45}
	0: 45%\|████▍ \| 1325/2975 [17:24<17:02, 1.61it/s] 45%\|████▍ \| 1326/2975 [17:24<16:54, 1.63it/s] 45%\|████▍ \| 1327/2975 [17:25<16:47, 1.64it/s] 45%\|████▍ \| 1328/2975 [17:26<16:43, 1.64it/s] 45%\|████▍ \| 1329/2975 [17:26<16:39, 1.65it/s] 45%\|████▍ \| 1330/2975 [17:27<16:35, 1.65it/s] 45%\|████▍ \| 1330/2975 [17:27<16:35, 1.65it/s] 45%\|████▍ \| 1331/2975 [17:27<16:34, 1.65it/s] 45%\|████▍ \| 1332/2975 [17:28<16:32, 1.66it/s] 45%\|████▍ \| 1333/2975 [17:29<16:30, 1.66it/s] 45%\|████▍ \| 1334/2975 [17:29<16:31, 1.66it/s] 45%\|████▍ \| 1335/2975 [17:30<16:31, 1.65it/s] 45%\|████▍ \| 1336/2975 [17:30<16:30, 1.65it/s] 45%\|████▍ \| 1337/2975 [17:31<16:29, 1.66it/s] 45%\|████▍ \| 1338/2975 [17:32<16:28, 1.66it/s] 45%\|████▌ \| 1339/2975 [17:
	0: {'loss': 0.728, 'grad_norm': 0.7845822733538016, 'learning_rate': 1e-05, 'epoch': 0.45}
	0: {'loss': 0.7165, 'grad_norm': 0.7776081763736031, 'learning_rate': 1e-05, 'epoch': 0.45}
	0: 32<16:28, 1.65it/s] 45%\|████▌ \| 1340/2975 [17:33<16:29, 1.65it/s] 45%\|████▌ \| 1340/2975 [17:33<16:29, 1.65it/s] 45%\|████▌ \| 1341/2975 [17:33<16:29, 1.65it/s] 45%\|████▌ \| 1342/2975 [17:34<16:29, 1.65it/s] 45%\|████▌ \| 1343/2975 [17:35<16:29, 1.65it/s] 45%\|████▌ \| 1344/2975 [17:35<16:28, 1.65it/s] 45%\|████▌ \| 1345/2975 [17:36<16:27, 1.65it/s] 45%\|████▌ \| 1346/2975 [17:36<16:26, 1.65it/s] 45%\|████▌ \| 1347/2975 [17:37<16:27, 1.65it/s] 45%\|████▌ \| 1348/2975 [17:38<16:26, 1.65it/s] 45%\|████▌ \| 1349/2975 [17:38<16:26, 1.65it/s] 45%\|████▌ \| 1350/2975 [17:39<16:26, 1.65it/s] 45%\|████▌ \| 1350/2975 [17:39<16:26, 1.65it/s] 45%\|████▌ \| 1351/2975 [17:40<16:26, 1.65it/s] 45%\|████▌ \| 13
	0: {'loss': 0.7278, 'grad_norm': 0.7524412952090146, 'learning_rate': 1e-05, 'epoch': 0.46}
	0: 52/2975 [17:40<16:25, 1.65it/s] 45%\|████▌ \| 1353/2975 [17:41<16:26, 1.64it/s] 46%\|████▌ \| 1354/2975 [17:41<16:32, 1.63it/s] 46%\|████▌ \| 1355/2975 [17:42<16:30, 1.64it/s] 46%\|████▌ \| 1356/2975 [17:43<16:25, 1.64it/s] 46%\|████▌ \| 1357/2975 [17:43<16:22, 1.65it/s] 46%\|████▌ \| 1358/2975 [17:44<16:21, 1.65it/s] 46%\|████▌ \| 1359/2975 [17:44<16:19, 1.65it/s] 46%\|████▌ \| 1360/2975 [17:45<16:16, 1.65it/s] 46%\|████▌ \| 1360/2975 [17:45<16:16, 1.65it/s] 46%\|████▌ \| 1361/2975 [17:46<16:15, 1.65it/s] 46%\|████▌ \| 1362/2975 [17:46<16:14, 1.66it/s] 46%\|████▌ \| 1363/2975 [17:47<16:13, 1.66it/s] 46%\|████▌ \| 1364/2975 [17:47<16:12, 1.66it/s] 46%\|████▌ \| 1365/2975 [17:48<16:12, 1.65it/s] 46%\|████▌ \| 1366/2975 [17:49<16:11, 1.66it/s] 46%\|█
	0: {'loss': 0.7409, 'grad_norm': 0.7305820316850615, 'learning_rate': 1e-05, 'epoch': 0.46}
	0: ███▌ \| 1367/2975 [17:49<16:11, 1.66it/s] 46%\|████▌ \| 1368/2975 [17:50<16:12, 1.65it/s] 46%\|████▌ \| 1369/2975 [17:50<16:12, 1.65it/s] 46%\|████▌ \| 1370/2975 [17:51<16:13, 1.65it/s] 46%\|████▌ \| 1370/2975 [17:51<16:13, 1.65it/s] 46%\|████▌ \| 1371/2975 [17:52<22:42, 1.18it/s] 46%\|████▌ \| 1372/2975 [17:53<21:46, 1.23it/s] 46%\|████▌ \| 1373/2975 [17:54<20:07, 1.33it/s] 46%\|████▌ \| 1374/2975 [17:54<18:57, 1.41it/s] 46%\|████▌ \| 1375/2975 [17:55<18:08, 1.47it/s] 46%\|████▋ \| 1376/2975 [17:56<17:31, 1.52it/s] 46%\|████▋ \| 1377/2975 [17:56<17:05, 1.56it/s] 46%\|████▋ \| 1378/2975 [17:57<16:47, 1.58it/s] 46%\|████▋ \| 1379/2975 [17:57<16:35, 1.60it/s] 46%\|████▋ \| 1380/2975 [17:58<16:26, 1.62it/s]
	0: {'loss': 0.7224, 'grad_norm': 0.7595749245205473, 'learning_rate': 1e-05, 'epoch': 0.46}
	0: {'loss': 0.7343, 'grad_norm': 0.737663184076695, 'learning_rate': 1e-05, 'epoch': 0.47}
	0: 46%\|████▋ \| 1380/2975 [17:58<16:26, 1.62it/s] 46%\|████▋ \| 1381/2975 [17:59<16:20, 1.63it/s] 46%\|████▋ \| 1382/2975 [17:59<16:18, 1.63it/s] 46%\|████▋ \| 1383/2975 [18:00<16:14, 1.63it/s] 47%\|████▋ \| 1384/2975 [18:00<16:11, 1.64it/s] 47%\|████▋ \| 1385/2975 [18:01<16:10, 1.64it/s] 47%\|████▋ \| 1386/2975 [18:02<17:46, 1.49it/s] 47%\|████▋ \| 1387/2975 [18:03<19:14, 1.38it/s] 47%\|████▋ \| 1388/2975 [18:03<18:18, 1.44it/s] 47%\|████▋ \| 1389/2975 [18:04<17:37, 1.50it/s] 47%\|████▋ \| 1390/2975 [18:05<19:22, 1.36it/s] 47%\|████▋ \| 1390/2975 [18:05<19:22, 1.36it/s] 47%\|████▋ \| 1391/2975 [18:05<18:26, 1.43it/s] 47%\|████▋ \| 1392/2975 [18:06<17:58, 1.47it/s] 47%\|████▋ \| 1393/2975 [18:07<19:11, 1.37it/s] 47%\|████▋ \| 1394/2975
	0: {'loss': 0.7235, 'grad_norm': 0.7518853779242005, 'learning_rate': 1e-05, 'epoch': 0.47}
	0: [18:08<18:40, 1.41it/s] 47%\|████▋ \| 1395/2975 [18:08<17:50, 1.48it/s] 47%\|████▋ \| 1396/2975 [18:09<17:16, 1.52it/s] 47%\|████▋ \| 1397/2975 [18:09<16:51, 1.56it/s] 47%\|████▋ \| 1398/2975 [18:10<16:34, 1.58it/s] 47%\|████▋ \| 1399/2975 [18:11<16:23, 1.60it/s] 47%\|████▋ \| 1400/2975 [18:11<16:15, 1.62it/s] 47%\|████▋ \| 1400/2975 [18:11<16:15, 1.62it/s] 47%\|████▋ \| 1401/2975 [18:12<18:11, 1.44it/s] 47%\|████▋ \| 1402/2975 [18:13<17:42, 1.48it/s] 47%\|████▋ \| 1403/2975 [18:13<17:12, 1.52it/s] 47%\|████▋ \| 1404/2975 [18:14<16:48, 1.56it/s] 47%\|████▋ \| 1405/2975 [18:15<16:38, 1.57it/s] 47%\|████▋ \| 1406/2975 [18:15<16:25, 1.59it/s] 47%\|████▋ \| 1407/2975 [18:16<16:16, 1.61it/s] 47%\|████▋ \| 1408/2975 [18:16<16:15, 1.61it/s] 47%\|███�
	0: {'loss': 0.7288, 'grad_norm': 0.7594287015442452, 'learning_rate': 1e-05, 'epoch': 0.47}
	0: {'loss': 0.7071, 'grad_norm': 0.7398135145941482, 'learning_rate': 1e-05, 'epoch': 0.48}
	0: �▋ \| 1409/2975 [18:17<16:06, 1.62it/s] 47%\|████▋ \| 1410/2975 [18:18<15:59, 1.63it/s] 47%\|████▋ \| 1410/2975 [18:18<15:59, 1.63it/s] 47%\|████▋ \| 1411/2975 [18:18<15:58, 1.63it/s] 47%\|████▋ \| 1412/2975 [18:19<15:55, 1.64it/s] 47%\|████▋ \| 1413/2975 [18:19<15:53, 1.64it/s] 48%\|████▊ \| 1414/2975 [18:20<15:52, 1.64it/s] 48%\|████▊ \| 1415/2975 [18:21<15:50, 1.64it/s] 48%\|████▊ \| 1416/2975 [18:21<15:49, 1.64it/s] 48%\|████▊ \| 1417/2975 [18:22<16:31, 1.57it/s] 48%\|████▊ \| 1418/2975 [18:23<16:32, 1.57it/s] 48%\|████▊ \| 1419/2975 [18:23<16:17, 1.59it/s] 48%\|████▊ \| 1420/2975 [18:24<16:25, 1.58it/s] 48%\|████▊ \| 1420/2975 [18:24<16:25, 1.58it/s] 48%\|████▊ \| 1421/2975 [18:25<16:13, 1.60it/s] 48%
	0: {'loss': 0.7393, 'grad_norm': 0.769642418005638, 'learning_rate': 1e-05, 'epoch': 0.48}
	0: \|████▊ \| 1422/2975 [18:25<16:02, 1.61it/s] 48%\|████▊ \| 1423/2975 [18:26<16:10, 1.60it/s] 48%\|████▊ \| 1424/2975 [18:26<15:59, 1.62it/s] 48%\|████▊ \| 1425/2975 [18:27<15:51, 1.63it/s] 48%\|████▊ \| 1426/2975 [18:28<15:45, 1.64it/s] 48%\|████▊ \| 1427/2975 [18:28<15:41, 1.64it/s] 48%\|████▊ \| 1428/2975 [18:29<15:37, 1.65it/s] 48%\|████▊ \| 1429/2975 [18:29<15:36, 1.65it/s] 48%\|████▊ \| 1430/2975 [18:30<15:36, 1.65it/s] 48%\|████▊ \| 1430/2975 [18:30<15:36, 1.65it/s] 48%\|████▊ \| 1431/2975 [18:31<15:38, 1.64it/s] 48%\|████▊ \| 1432/2975 [18:31<15:39, 1.64it/s] 48%\|████▊ \| 1433/2975 [18:32<15:40, 1.64it/s] 48%\|████▊ \| 1434/2975 [18:32<15:40, 1.64it/s] 48%\|████▊ \| 1435/2975 [18:33<15:40, 1.64it/s] 48%\|████▊ \| 1436/2975 [18:34<1
	0: {'loss': 0.722, 'grad_norm': 0.7687981704462479, 'learning_rate': 1e-05, 'epoch': 0.48}
	0: 5:37, 1.64it/s] 48%\|████▊ \| 1437/2975 [18:34<15:36, 1.64it/s] 48%\|████▊ \| 1438/2975 [18:35<15:47, 1.62it/s] 48%\|████▊ \| 1439/2975 [18:35<15:41, 1.63it/s] 48%\|████▊ \| 1440/2975 [18:36<15:37, 1.64it/s] 48%\|████▊ \| 1440/2975 [18:36<15:37, 1.64it/s] 48%\|████▊ \| 1441/2975 [18:37<15:34, 1.64it/s] 48%\|████▊ \| 1442/2975 [18:37<15:33, 1.64it/s] 49%\|████▊ \| 1443/2975 [18:38<15:33, 1.64it/s] 49%\|████▊ \| 1444/2975 [18:39<15:34, 1.64it/s] 49%\|████▊ \| 1445/2975 [18:39<15:32, 1.64it/s] 49%\|████▊ \| 1446/2975 [18:40<15:31, 1.64it/s] 49%\|████▊ \| 1447/2975 [18:40<15:30, 1.64it/s] 49%\|████▊ \| 1448/2975 [18:41<15:28, 1.64it/s] 49%\|████▊ \| 1449/2975 [18:42<15:26, 1.65it/s] 49%\|████▊ \| 1450/2975 [18:42<15:25, 1.65it/s]
	0: {'loss': 0.7337, 'grad_norm': 0.737611485833095, 'learning_rate': 1e-05, 'epoch': 0.49}
	0: {'loss': 0.7532, 'grad_norm': 0.753601200505015, 'learning_rate': 1e-05, 'epoch': 0.49}
	0: 49%\|████▊ \| 1450/2975 [18:42<15:25, 1.65it/s] 49%\|████▉ \| 1451/2975 [18:43<15:24, 1.65it/s] 49%\|████▉ \| 1452/2975 [18:43<15:22, 1.65it/s] 49%\|████▉ \| 1453/2975 [18:44<15:22, 1.65it/s] 49%\|████▉ \| 1454/2975 [18:45<15:20, 1.65it/s] 49%\|████▉ \| 1455/2975 [18:45<15:19, 1.65it/s] 49%\|████▉ \| 1456/2975 [18:46<15:19, 1.65it/s] 49%\|████▉ \| 1457/2975 [18:46<15:19, 1.65it/s] 49%\|████▉ \| 1458/2975 [18:47<15:17, 1.65it/s] 49%\|████▉ \| 1459/2975 [18:48<15:16, 1.65it/s] 49%\|████▉ \| 1460/2975 [18:48<15:16, 1.65it/s] 49%\|████▉ \| 1460/2975 [18:48<15:16, 1.65it/s] 49%\|████▉ \| 1461/2975 [18:49<15:18, 1.65it/s] 49%\|████▉ \| 1462/2975 [18:49<15:18, 1.65it/s] 49%\|████▉ \| 1463/2975 [18:50<15:17, 1.65it/s] 49%\|██�
	0: {'loss': 0.7061, 'grad_norm': 0.7431195611428175, 'learning_rate': 1e-05, 'epoch': 0.49}
	0: ��█▉ \| 1464/2975 [18:51<15:16, 1.65it/s] 49%\|████▉ \| 1465/2975 [18:51<15:17, 1.65it/s] 49%\|████▉ \| 1466/2975 [18:52<15:18, 1.64it/s] 49%\|████▉ \| 1467/2975 [18:52<15:17, 1.64it/s] 49%\|████▉ \| 1468/2975 [18:53<15:16, 1.64it/s] 49%\|████▉ \| 1469/2975 [18:54<15:15, 1.65it/s] 49%\|████▉ \| 1470/2975 [18:54<15:14, 1.65it/s] 49%\|████▉ \| 1470/2975 [18:54<15:14, 1.65it/s] 49%\|████▉ \| 1471/2975 [18:55<15:14, 1.65it/s] 49%\|████▉ \| 1472/2975 [18:56<15:12, 1.65it/s] 50%\|████▉ \| 1473/2975 [18:56<15:11, 1.65it/s] 50%\|████▉ \| 1474/2975 [18:57<15:10, 1.65it/s] 50%\|████▉ \| 1475/2975 [18:57<15:10, 1.65it/s] 50%\|████▉ \| 1476/2975 [18:58<15:10, 1.65it/s] 50%\|████▉ \| 1477/2975 [18:59<15:09, 1.65it/s] 50%\|████▉ \| 1478/2975 [18:59<15:09, 1
	0: {'loss': 0.7238, 'grad_norm': 0.7227391902402309, 'learning_rate': 1e-05, 'epoch': 0.5}
	0: {'loss': 0.7362, 'grad_norm': 0.7838094626607218, 'learning_rate': 1e-05, 'epoch': 0.5}
	0: .65it/s] 50%\|████▉ \| 1479/2975 [19:00<15:10, 1.64it/s] 50%\|████▉ \| 1480/2975 [19:00<15:09, 1.64it/s] 50%\|████▉ \| 1480/2975 [19:00<15:09, 1.64it/s] 50%\|████▉ \| 1481/2975 [19:01<15:07, 1.65it/s] 50%\|████▉ \| 1482/2975 [19:02<15:05, 1.65it/s] 50%\|████▉ \| 1483/2975 [19:02<15:05, 1.65it/s] 50%\|████▉ \| 1484/2975 [19:03<15:05, 1.65it/s] 50%\|████▉ \| 1485/2975 [19:03<15:06, 1.64it/s] 50%\|████▉ \| 1486/2975 [19:04<15:05, 1.64it/s] 50%\|████▉ \| 1487/2975 [19:05<15:04, 1.64it/s] 50%\|█████ \| 1488/2975 [19:05<15:04, 1.64it/s] 50%\|█████ \| 1489/2975 [19:06<15:05, 1.64it/s] 50%\|█████ \| 1490/2975 [19:06<15:04, 1.64it/s] 50%\|█████ \| 1490/2975 [19:06<15:04, 1.64it/s] 50%\|█████ \| 1491/2975 [19:
	0: {'loss': 0.7061, 'grad_norm': 0.7235667413143033, 'learning_rate': 1e-05, 'epoch': 0.5}
	0: 07<15:03, 1.64it/s] 50%\|█████ \| 1492/2975 [19:08<15:02, 1.64it/s] 50%\|█████ \| 1493/2975 [19:08<15:14, 1.62it/s] 50%\|█████ \| 1494/2975 [19:09<15:12, 1.62it/s] 50%\|█████ \| 1495/2975 [19:10<15:09, 1.63it/s] 50%\|█████ \| 1496/2975 [19:10<15:09, 1.63it/s] 50%\|█████ \| 1497/2975 [19:11<15:06, 1.63it/s] 50%\|█████ \| 1498/2975 [19:11<15:04, 1.63it/s] 50%\|█████ \| 1499/2975 [19:12<15:01, 1.64it/s] 50%\|█████ \| 1500/2975 [19:13<14:59, 1.64it/s] 50%\|█████ \| 1500/2975 [19:13<14:59, 1.64it/s] 50%\|█████ \| 1501/2975 [19:13<15:00, 1.64it/s] 50%\|█████ \| 1502/2975 [19:14<15:02, 1.63it/s] 51%\|█████ \| 1503/2975 [19:14<15:00, 1.63it/s] 51%\|█████ \| 1504/2975 [19:15<15:00, 1.63it/s] 51%\|█████ \| 1505/2975 [19:16<14:59, 1.63it/s] 51%\|█████
	0: {'loss': 0.7325, 'grad_norm': 0.7931800443283757, 'learning_rate': 1e-05, 'epoch': 0.51}
	0: \| 1506/2975 [19:16<14:57, 1.64it/s] 51%\|█████ \| 1507/2975 [19:17<14:57, 1.63it/s] 51%\|█████ \| 1508/2975 [19:17<14:57, 1.63it/s] 51%\|█████ \| 1509/2975 [19:18<14:56, 1.64it/s] 51%\|█████ \| 1510/2975 [19:19<14:53, 1.64it/s] 51%\|█████ \| 1510/2975 [19:19<14:53, 1.64it/s] 51%\|█████ \| 1511/2975 [19:19<14:55, 1.63it/s] 51%\|█████ \| 1512/2975 [19:20<14:55, 1.63it/s] 51%\|█████ \| 1513/2975 [19:21<14:54, 1.63it/s] 51%\|█████ \| 1514/2975 [19:21<14:53, 1.64it/s] 51%\|█████ \| 1515/2975 [19:22<15:04, 1.61it/s] 51%\|█████ \| 1516/2975 [19:22<15:00, 1.62it/s] 51%\|█████ \| 1517/2975 [19:23<14:56, 1.63it/s] 51%\|█████ \| 1518/2975 [19:24<14:53, 1.63it/s] 51%\|█████ \| 1519/2975 [19:24<15:06, 1.61it/s] 51%\|█████ \| 1520/2975 [19:25<15:00, 1.62it/s]
	0: {'loss': 0.7227, 'grad_norm': 0.7509840361931992, 'learning_rate': 1e-05, 'epoch': 0.51}
	0: {'loss': 0.7477, 'grad_norm': 0.7622116209089915, 'learning_rate': 1e-05, 'epoch': 0.51}
	0: 51%\|█████ \| 1520/2975 [19:25<15:00, 1.62it/s] 51%\|█████ \| 1521/2975 [19:25<14:57, 1.62it/s] 51%\|█████ \| 1522/2975 [19:26<14:52, 1.63it/s] 51%\|█████ \| 1523/2975 [19:27<14:49, 1.63it/s] 51%\|█████ \| 1524/2975 [19:27<14:48, 1.63it/s] 51%\|█████▏ \| 1525/2975 [19:28<14:46, 1.64it/s] 51%\|█████▏ \| 1526/2975 [19:29<14:46, 1.64it/s] 51%\|█████▏ \| 1527/2975 [19:29<14:45, 1.64it/s] 51%\|█████▏ \| 1528/2975 [19:30<14:45, 1.63it/s] 51%\|█████▏ \| 1529/2975 [19:30<14:44, 1.63it/s] 51%\|█████▏ \| 1530/2975 [19:31<14:44, 1.63it/s] 51%\|█████▏ \| 1530/2975 [19:31<14:44, 1.63it/s] 51%\|█████▏ \| 1531/2975 [19:32<14:44, 1.63it/s] 51%\|█████▏ \| 1532/2975 [19:32<14:43, 1.63it/s] 52%\|█████▏ \| 15
	0: {'loss': 0.7109, 'grad_norm': 0.7106310730304262, 'learning_rate': 1e-05, 'epoch': 0.52}
	0: 33/2975 [19:33<14:42, 1.63it/s] 52%\|█████▏ \| 1534/2975 [19:33<14:42, 1.63it/s] 52%\|█████▏ \| 1535/2975 [19:34<14:52, 1.61it/s] 52%\|█████▏ \| 1536/2975 [19:35<14:47, 1.62it/s] 52%\|█████▏ \| 1537/2975 [19:35<14:45, 1.62it/s] 52%\|█████▏ \| 1538/2975 [19:36<14:42, 1.63it/s] 52%\|█████▏ \| 1539/2975 [19:37<15:03, 1.59it/s] 52%\|█████▏ \| 1540/2975 [19:37<14:55, 1.60it/s] 52%\|█████▏ \| 1540/2975 [19:37<14:55, 1.60it/s] 52%\|█████▏ \| 1541/2975 [19:38<14:49, 1.61it/s] 52%\|█████▏ \| 1542/2975 [19:38<14:44, 1.62it/s] 52%\|█████▏ \| 1543/2975 [19:39<14:39, 1.63it/s] 52%\|█████▏ \| 1544/2975 [19:40<14:38, 1.63it/s] 52%\|█████▏ \| 1545/2975 [19:40<14:36, 1.63it/s] 52%\|█████▏ \| 1546/2975 [19:41<14:46, 1.61it/s] 52%\|█████▏ \| 1547/2975 [19
	0: {'loss': 0.7287, 'grad_norm': 0.9284595641558513, 'learning_rate': 1e-05, 'epoch': 0.52}
	0: {'loss': 0.7064, 'grad_norm': 0.7437149827305557, 'learning_rate': 1e-05, 'epoch': 0.52}
	0: :41<14:41, 1.62it/s] 52%\|█████▏ \| 1548/2975 [19:42<14:38, 1.62it/s] 52%\|█████▏ \| 1549/2975 [19:43<14:35, 1.63it/s] 52%\|█████▏ \| 1550/2975 [19:43<14:32, 1.63it/s] 52%\|█████▏ \| 1550/2975 [19:43<14:32, 1.63it/s] 52%\|█████▏ \| 1551/2975 [19:44<14:31, 1.63it/s] 52%\|█████▏ \| 1552/2975 [19:45<14:30, 1.63it/s] 52%\|█████▏ \| 1553/2975 [19:45<14:29, 1.64it/s] 52%\|█████▏ \| 1554/2975 [19:46<14:34, 1.62it/s] 52%\|█████▏ \| 1555/2975 [19:46<14:33, 1.63it/s] 52%\|█████▏ \| 1556/2975 [19:47<14:30, 1.63it/s] 52%\|█████▏ \| 1557/2975 [19:48<14:27, 1.63it/s] 52%\|█████▏ \| 1558/2975 [19:48<14:26, 1.64it/s] 52%\|█████▏ \| 1559/2975 [19:49<14:25, 1.64it/s] 52%\|█████▏ \| 1560/2975 [19:49<14:26, 1.63it/s]
	0: {'loss': 0.7335, 'grad_norm': 0.7134687197364542, 'learning_rate': 1e-05, 'epoch': 0.53}
	0: 52%\|█████▏ \| 1560/2975 [19:49<14:26, 1.63it/s] 52%\|█████▏ \| 1561/2975 [19:50<14:25, 1.63it/s] 53%\|█████▎ \| 1562/2975 [19:51<14:24, 1.63it/s] 53%\|█████▎ \| 1563/2975 [19:51<14:24, 1.63it/s] 53%\|█████▎ \| 1564/2975 [19:52<14:23, 1.63it/s] 53%\|█████▎ \| 1565/2975 [19:52<14:22, 1.64it/s] 53%\|█████▎ \| 1566/2975 [19:53<14:18, 1.64it/s] 53%\|█████▎ \| 1567/2975 [19:54<14:17, 1.64it/s] 53%\|█████▎ \| 1568/2975 [19:54<14:19, 1.64it/s] 53%\|█████▎ \| 1569/2975 [19:55<14:19, 1.64it/s] 53%\|█████▎ \| 1570/2975 [19:56<14:18, 1.64it/s] 53%\|█████▎ \| 1570/2975 [19:56<14:18, 1.64it/s] 53%\|█████▎ \| 1571/2975 [19:56<14:19, 1.63it/s] 53%\|█████▎ \| 1572/2975 [19:57<14:19, 1.63it/s] 53%\|█████▎ \| 1573/2975 [19:57<14:18, 1.63it/s] 53%\|██
	0: {'loss': 0.7279, 'grad_norm': 0.7685737145218493, 'learning_rate': 1e-05, 'epoch': 0.53}
	0: ███▎ \| 1574/2975 [19:58<14:17, 1.63it/s] 53%\|█████▎ \| 1575/2975 [19:59<14:15, 1.64it/s] 53%\|█████▎ \| 1576/2975 [19:59<14:16, 1.63it/s] 53%\|█████▎ \| 1577/2975 [20:00<14:15, 1.63it/s] 53%\|█████▎ \| 1578/2975 [20:00<14:14, 1.64it/s] 53%\|█████▎ \| 1579/2975 [20:01<14:13, 1.64it/s] 53%\|█████▎ \| 1580/2975 [20:02<14:14, 1.63it/s] 53%\|█████▎ \| 1580/2975 [20:02<14:14, 1.63it/s] 53%\|█████▎ \| 1581/2975 [20:02<14:15, 1.63it/s] 53%\|█████▎ \| 1582/2975 [20:03<14:14, 1.63it/s] 53%\|█████▎ \| 1583/2975 [20:04<14:15, 1.63it/s] 53%\|█████▎ \| 1584/2975 [20:04<14:14, 1.63it/s] 53%\|█████▎ \| 1585/2975 [20:05<14:13, 1.63it/s] 53%\|█████▎ \| 1586/2975 [20:05<14:12, 1.63it/s] 53%\|█████▎ \| 1587/2975 [20:06<14:11, 1.63it/s] 53%\|█████�
	0: {'loss': 0.7372, 'grad_norm': 0.7335726270641828, 'learning_rate': 1e-05, 'epoch': 0.53}
	0: {'loss': 0.7061, 'grad_norm': 0.7425514182122959, 'learning_rate': 1e-05, 'epoch': 0.54}
	0: � \| 1588/2975 [20:07<14:10, 1.63it/s] 53%\|█████▎ \| 1589/2975 [20:07<14:08, 1.63it/s] 53%\|█████▎ \| 1590/2975 [20:08<14:08, 1.63it/s] 53%\|█████▎ \| 1590/2975 [20:08<14:08, 1.63it/s] 53%\|█████▎ \| 1591/2975 [20:08<14:07, 1.63it/s] 54%\|█████▎ \| 1592/2975 [20:09<14:05, 1.64it/s] 54%\|█████▎ \| 1593/2975 [20:10<14:04, 1.64it/s] 54%\|█████▎ \| 1594/2975 [20:10<14:03, 1.64it/s] 54%\|█████▎ \| 1595/2975 [20:11<14:02, 1.64it/s] 54%\|█████▎ \| 1596/2975 [20:11<14:01, 1.64it/s] 54%\|█████▎ \| 1597/2975 [20:12<14:00, 1.64it/s] 54%\|█████▎ \| 1598/2975 [20:13<13:59, 1.64it/s] 54%\|█████▎ \| 1599/2975 [20:13<13:59, 1.64it/s] 54%\|█████▍ \| 1600/2975 [20:14<14:00, 1.64it/s] 54%\|█████▍ \| 1600/2975 [20:1
	0: {'loss': 0.7128, 'grad_norm': 0.756614439567717, 'learning_rate': 1e-05, 'epoch': 0.54}
	0: 4<14:00, 1.64it/s] 54%\|█████▍ \| 1601/2975 [20:15<14:00, 1.64it/s] 54%\|█████▍ \| 1602/2975 [20:15<13:59, 1.64it/s] 54%\|█████▍ \| 1603/2975 [20:16<13:57, 1.64it/s] 54%\|█████▍ \| 1604/2975 [20:16<13:57, 1.64it/s] 54%\|█████▍ \| 1605/2975 [20:17<13:55, 1.64it/s] 54%\|█████▍ \| 1606/2975 [20:18<13:53, 1.64it/s] 54%\|█████▍ \| 1607/2975 [20:18<13:53, 1.64it/s] 54%\|█████▍ \| 1608/2975 [20:19<13:52, 1.64it/s] 54%\|█████▍ \| 1609/2975 [20:19<13:53, 1.64it/s] 54%\|█████▍ \| 1610/2975 [20:20<13:52, 1.64it/s] 54%\|█████▍ \| 1610/2975 [20:20<13:52, 1.64it/s] 54%\|█████▍ \| 1611/2975 [20:21<13:52, 1.64it/s] 54%\|█████▍ \| 1612/2975 [20:21<13:53, 1.64it/s] 54%\|█████▍ \| 1613/2975 [20:22<13:52, 1.64it/s] 54%\|█████▍ \| 1614/2975 [20:22<13:52, 1
	0: {'loss': 0.7238, 'grad_norm': 0.7183712888084608, 'learning_rate': 1e-05, 'epoch': 0.54}
	0: .64it/s] 54%\|█████▍ \| 1615/2975 [20:23<13:51, 1.64it/s] 54%\|█████▍ \| 1616/2975 [20:24<13:50, 1.64it/s] 54%\|█████▍ \| 1617/2975 [20:24<13:48, 1.64it/s] 54%\|█████▍ \| 1618/2975 [20:25<13:46, 1.64it/s] 54%\|█████▍ \| 1619/2975 [20:26<13:44, 1.64it/s] 54%\|█████▍ \| 1620/2975 [20:26<13:43, 1.65it/s] 54%\|█████▍ \| 1620/2975 [20:26<13:43, 1.65it/s] 54%\|█████▍ \| 1621/2975 [20:27<13:43, 1.64it/s] 55%\|█████▍ \| 1622/2975 [20:27<13:44, 1.64it/s] 55%\|█████▍ \| 1623/2975 [20:28<13:44, 1.64it/s] 55%\|█████▍ \| 1624/2975 [20:29<13:44, 1.64it/s] 55%\|█████▍ \| 1625/2975 [20:29<13:46, 1.63it/s] 55%\|█████▍ \| 1626/2975 [20:30<13:47, 1.63it/s] 55%\|█████▍ \| 1627/2975 [20:30<13:47, 1.63it/s] 55%\|█████▍ \| 1628/2975 [20:31<13:44, 1.63it/s] 5
	0: {'loss': 0.7151, 'grad_norm': 0.7421727883946161, 'learning_rate': 1e-05, 'epoch': 0.55}
	0: {'loss': 0.7215, 'grad_norm': 0.7387996700462125, 'learning_rate': 1e-05, 'epoch': 0.55}
	0: 5%\|█████▍ \| 1629/2975 [20:32<13:42, 1.64it/s] 55%\|█████▍ \| 1630/2975 [20:32<13:41, 1.64it/s] 55%\|█████▍ \| 1630/2975 [20:32<13:41, 1.64it/s] 55%\|█████▍ \| 1631/2975 [20:33<13:40, 1.64it/s] 55%\|█████▍ \| 1632/2975 [20:33<13:39, 1.64it/s] 55%\|█████▍ \| 1633/2975 [20:34<13:38, 1.64it/s] 55%\|█████▍ \| 1634/2975 [20:35<13:37, 1.64it/s] 55%\|█████▍ \| 1635/2975 [20:35<13:38, 1.64it/s] 55%\|█████▍ \| 1636/2975 [20:36<13:38, 1.64it/s] 55%\|█████▌ \| 1637/2975 [20:37<13:38, 1.63it/s] 55%\|█████▌ \| 1638/2975 [20:37<13:39, 1.63it/s] 55%\|█████▌ \| 1639/2975 [20:38<13:37, 1.63it/s] 55%\|█████▌ \| 1640/2975 [20:38<13:35, 1.64it/s] 55%\|█████▌ \| 1640/2975 [20:38<13:35, 1.64it/s] 55%\|█████▌
	0: {'loss': 0.7069, 'grad_norm': 0.7347750601522001, 'learning_rate': 1e-05, 'epoch': 0.55}
	0: \| 1641/2975 [20:39<13:34, 1.64it/s] 55%\|█████▌ \| 1642/2975 [20:40<13:32, 1.64it/s] 55%\|█████▌ \| 1643/2975 [20:40<13:31, 1.64it/s] 55%\|█████▌ \| 1644/2975 [20:41<13:29, 1.65it/s] 55%\|█████▌ \| 1645/2975 [20:41<13:28, 1.65it/s] 55%\|█████▌ \| 1646/2975 [20:42<13:27, 1.65it/s] 55%\|█████▌ \| 1647/2975 [20:43<13:27, 1.65it/s] 55%\|█████▌ \| 1648/2975 [20:43<13:27, 1.64it/s] 55%\|█████▌ \| 1649/2975 [20:44<13:28, 1.64it/s] 55%\|█████▌ \| 1650/2975 [20:44<13:31, 1.63it/s] 55%\|█████▌ \| 1650/2975 [20:44<13:31, 1.63it/s] 55%\|█████▌ \| 1651/2975 [20:45<13:32, 1.63it/s] 56%\|█████▌ \| 1652/2975 [20:46<13:30, 1.63it/s] 56%\|█████▌ \| 1653/2975 [20:46<13:28, 1.64it/s] 56%\|█████▌ \| 1654/2975 [20:47<13:27, 1.64it/s] 56%\|█████▌ \| 1655/2
	0: {'loss': 0.7276, 'grad_norm': 0.7590718949791706, 'learning_rate': 1e-05, 'epoch': 0.56}
	0: 975 [20:47<13:25, 1.64it/s] 56%\|█████▌ \| 1656/2975 [20:48<13:24, 1.64it/s] 56%\|█████▌ \| 1657/2975 [20:49<13:21, 1.64it/s] 56%\|█████▌ \| 1658/2975 [20:49<13:20, 1.64it/s] 56%\|█████▌ \| 1659/2975 [20:50<13:20, 1.64it/s] 56%\|█████▌ \| 1660/2975 [20:51<13:19, 1.64it/s] 56%\|█████▌ \| 1660/2975 [20:51<13:19, 1.64it/s] 56%\|█████▌ \| 1661/2975 [20:51<13:21, 1.64it/s] 56%\|█████▌ \| 1662/2975 [20:52<13:22, 1.64it/s] 56%\|█████▌ \| 1663/2975 [20:52<13:23, 1.63it/s] 56%\|█████▌ \| 1664/2975 [20:53<13:22, 1.63it/s] 56%\|█████▌ \| 1665/2975 [20:54<13:20, 1.64it/s] 56%\|█████▌ \| 1666/2975 [20:54<13:19, 1.64it/s] 56%\|█████▌ \| 1667/2975 [20:55<13:18, 1.64it/s] 56%\|█████▌ \| 1668/2975 [20:55<13:16, 1.64it/s] 56%\|█████▌ \| 1669/2975 [20:56<
	0: {'loss': 0.7199, 'grad_norm': 0.7649717152658091, 'learning_rate': 1e-05, 'epoch': 0.56}
	0: {'loss': 0.7277, 'grad_norm': 0.7437977542383272, 'learning_rate': 1e-05, 'epoch': 0.56}
	0: 13:16, 1.64it/s] 56%\|█████▌ \| 1670/2975 [20:57<13:15, 1.64it/s] 56%\|█████▌ \| 1670/2975 [20:57<13:15, 1.64it/s] 56%\|█████▌ \| 1671/2975 [20:57<13:14, 1.64it/s] 56%\|█████▌ \| 1672/2975 [20:58<13:13, 1.64it/s] 56%\|█████▌ \| 1673/2975 [20:58<13:12, 1.64it/s] 56%\|█████▋ \| 1674/2975 [20:59<13:14, 1.64it/s] 56%\|█████▋ \| 1675/2975 [21:00<13:15, 1.63it/s] 56%\|█████▋ \| 1676/2975 [21:00<13:16, 1.63it/s] 56%\|█████▋ \| 1677/2975 [21:01<13:14, 1.63it/s] 56%\|█████▋ \| 1678/2975 [21:02<13:12, 1.64it/s] 56%\|█████▋ \| 1679/2975 [21:02<13:09, 1.64it/s] 56%\|█████▋ \| 1680/2975 [21:03<13:08, 1.64it/s] 56%\|█████▋ \| 1680/2975 [21:03<13:08, 1.64it/s] 57%\|█████▋ \| 1681/2975 [21:03<13:08, 1.64it/s] 57%
	0: {'loss': 0.7005, 'grad_norm': 0.7046944932910209, 'learning_rate': 1e-05, 'epoch': 0.57}
	0: \|█████▋ \| 1682/2975 [21:04<13:07, 1.64it/s] 57%\|█████▋ \| 1683/2975 [21:05<13:07, 1.64it/s] 57%\|█████▋ \| 1684/2975 [21:05<13:06, 1.64it/s] 57%\|█████▋ \| 1685/2975 [21:06<13:05, 1.64it/s] 57%\|█████▋ \| 1686/2975 [21:06<13:05, 1.64it/s] 57%\|█████▋ \| 1687/2975 [21:07<13:06, 1.64it/s] 57%\|█████▋ \| 1688/2975 [21:08<13:05, 1.64it/s] 57%\|█████▋ \| 1689/2975 [21:08<13:05, 1.64it/s] 57%\|█████▋ \| 1690/2975 [21:09<13:04, 1.64it/s] 57%\|█████▋ \| 1690/2975 [21:09<13:04, 1.64it/s] 57%\|█████▋ \| 1691/2975 [21:09<13:04, 1.64it/s] 57%\|█████▋ \| 1692/2975 [21:10<13:03, 1.64it/s] 57%\|█████▋ \| 1693/2975 [21:11<13:03, 1.64it/s] 57%\|█████▋ \| 1694/2975 [21:11<13:02, 1.64it/s] 57%\|█████▋ \| 1695/2975 [21:12<13:01, 1.64it/s] 57%\|███�
	0: {'loss': 0.7045, 'grad_norm': 0.7092249753117572, 'learning_rate': 1e-05, 'epoch': 0.57}
	0: ��█▋ \| 1696/2975 [21:12<12:58, 1.64it/s] 57%\|█████▋ \| 1697/2975 [21:13<12:57, 1.64it/s] 57%\|█████▋ \| 1698/2975 [21:14<12:56, 1.64it/s] 57%\|█████▋ \| 1699/2975 [21:14<12:57, 1.64it/s] 57%\|█████▋ \| 1700/2975 [21:15<13:08, 1.62it/s] 57%\|█████▋ \| 1700/2975 [21:15<13:08, 1.62it/s] 57%\|█████▋ \| 1701/2975 [21:16<13:04, 1.62it/s] 57%\|█████▋ \| 1702/2975 [21:16<13:00, 1.63it/s] 57%\|█████▋ \| 1703/2975 [21:17<12:59, 1.63it/s] 57%\|█████▋ \| 1704/2975 [21:17<13:07, 1.61it/s] 57%\|█████▋ \| 1705/2975 [21:18<13:03, 1.62it/s] 57%\|█████▋ \| 1706/2975 [21:19<13:01, 1.62it/s] 57%\|█████▋ \| 1707/2975 [21:19<13:00, 1.63it/s] 57%\|█████▋ \| 1708/2975 [21:20<12:58, 1.63it/s] 57%\|█████▋ \| 1709/2975 [21:20<12:57, 1.63it/s] 57%\|█████▋
	0: {'loss': 0.7211, 'grad_norm': 0.908931769106596, 'learning_rate': 1e-05, 'epoch': 0.57}
	0: {'loss': 0.734, 'grad_norm': 0.7677796479365407, 'learning_rate': 1e-05, 'epoch': 0.58}
	0: \| 1710/2975 [21:21<12:56, 1.63it/s] 57%\|█████▋ \| 1710/2975 [21:21<12:56, 1.63it/s] 58%\|█████▊ \| 1711/2975 [21:22<12:55, 1.63it/s] 58%\|█████▊ \| 1712/2975 [21:22<12:52, 1.63it/s] 58%\|█████▊ \| 1713/2975 [21:23<12:50, 1.64it/s] 58%\|█████▊ \| 1714/2975 [21:24<12:48, 1.64it/s] 58%\|█████▊ \| 1715/2975 [21:24<12:46, 1.64it/s] 58%\|█████▊ \| 1716/2975 [21:25<12:46, 1.64it/s] 58%\|█████▊ \| 1717/2975 [21:25<12:46, 1.64it/s] 58%\|█████▊ \| 1718/2975 [21:26<12:45, 1.64it/s] 58%\|█████▊ \| 1719/2975 [21:27<12:43, 1.64it/s] 58%\|█████▊ \| 1720/2975 [21:27<12:43, 1.64it/s] 58%\|█████▊ \| 1720/2975 [21:27<12:43, 1.64it/s] 58%\|█████▊ \| 1721/2975 [21:28<12:43, 1.64it/s] 58%\|█████▊ \| 1722/2975 [21:28<12
	0: {'loss': 0.7207, 'grad_norm': 0.8264399873360426, 'learning_rate': 1e-05, 'epoch': 0.58}
	0: :43, 1.64it/s] 58%\|█████▊ \| 1723/2975 [21:29<12:44, 1.64it/s] 58%\|█████▊ \| 1724/2975 [21:30<12:44, 1.64it/s] 58%\|█████▊ \| 1725/2975 [21:30<12:43, 1.64it/s] 58%\|█████▊ \| 1726/2975 [21:31<12:42, 1.64it/s] 58%\|█████▊ \| 1727/2975 [21:31<12:41, 1.64it/s] 58%\|█████▊ \| 1728/2975 [21:32<12:40, 1.64it/s] 58%\|█████▊ \| 1729/2975 [21:33<12:39, 1.64it/s] 58%\|█████▊ \| 1730/2975 [21:33<12:40, 1.64it/s] 58%\|█████▊ \| 1730/2975 [21:33<12:40, 1.64it/s] 58%\|█████▊ \| 1731/2975 [21:34<13:54, 1.49it/s] 58%\|█████▊ \| 1732/2975 [21:35<13:29, 1.54it/s] 58%\|█████▊ \| 1733/2975 [21:35<13:13, 1.57it/s] 58%\|█████▊ \| 1734/2975 [21:36<13:02, 1.59it/s] 58%\|█████▊ \| 1735/2975 [21:37<12:53, 1.60it/s] 58%\|█████▊ \| 1736/2975 [21:37<12:48, 1.61i
	0: {'loss': 0.711, 'grad_norm': 0.8011410157530903, 'learning_rate': 1e-05, 'epoch': 0.58}
	0: t/s] 58%\|█████▊ \| 1737/2975 [21:38<12:43, 1.62it/s] 58%\|█████▊ \| 1738/2975 [21:38<12:41, 1.62it/s] 58%\|█████▊ \| 1739/2975 [21:39<12:38, 1.63it/s] 58%\|█████▊ \| 1740/2975 [21:40<12:35, 1.63it/s] 58%\|█████▊ \| 1740/2975 [21:40<12:35, 1.63it/s] 59%\|█████▊ \| 1741/2975 [21:40<12:34, 1.64it/s] 59%\|█████▊ \| 1742/2975 [21:41<12:31, 1.64it/s] 59%\|█████▊ \| 1743/2975 [21:41<12:29, 1.64it/s] 59%\|█████▊ \| 1744/2975 [21:42<12:27, 1.65it/s] 59%\|█████▊ \| 1745/2975 [21:43<12:25, 1.65it/s] 59%\|█████▊ \| 1746/2975 [21:43<12:25, 1.65it/s] 59%\|█████▊ \| 1747/2975 [21:44<12:25, 1.65it/s] 59%\|█████▉ \| 1748/2975 [21:44<12:24, 1.65it/s] 59%\|█████▉ \| 1749/2975 [21:45<12:22, 1.65it/s] 59%\|█████▉ \| 1750/2975 [21:46<12:21, 1.65it/s]
	0: {'loss': 0.7079, 'grad_norm': 0.7102845427696028, 'learning_rate': 1e-05, 'epoch': 0.59}
	0: {'loss': 0.7247, 'grad_norm': 0.7090127036567279, 'learning_rate': 1e-05, 'epoch': 0.59}
	0: 59%\|█████▉ \| 1750/2975 [21:46<12:21, 1.65it/s] 59%\|█████▉ \| 1751/2975 [21:46<12:33, 1.63it/s] 59%\|█████▉ \| 1752/2975 [21:47<12:31, 1.63it/s] 59%\|█████▉ \| 1753/2975 [21:48<12:29, 1.63it/s] 59%\|█████▉ \| 1754/2975 [21:48<12:27, 1.63it/s] 59%\|█████▉ \| 1755/2975 [21:49<12:24, 1.64it/s] 59%\|█████▉ \| 1756/2975 [21:49<12:23, 1.64it/s] 59%\|█████▉ \| 1757/2975 [21:50<12:21, 1.64it/s] 59%\|█████▉ \| 1758/2975 [21:51<12:19, 1.65it/s] 59%\|█████▉ \| 1759/2975 [21:51<12:19, 1.64it/s] 59%\|█████▉ \| 1760/2975 [21:52<12:17, 1.65it/s] 59%\|█████▉ \| 1760/2975 [21:52<12:17, 1.65it/s] 59%\|█████▉ \| 1761/2975 [21:52<12:16, 1.65it/s] 59%\|█████▉ \| 1762/2975 [21:53<12:16, 1.65it/s] 59%\|█████▉ \|
	0: {'loss': 0.7057, 'grad_norm': 0.724586519482281, 'learning_rate': 1e-05, 'epoch': 0.59}
	0: 1763/2975 [21:54<12:15, 1.65it/s] 59%\|█████▉ \| 1764/2975 [21:54<12:14, 1.65it/s] 59%\|█████▉ \| 1765/2975 [21:55<12:13, 1.65it/s] 59%\|█████▉ \| 1766/2975 [21:55<12:11, 1.65it/s] 59%\|█████▉ \| 1767/2975 [21:56<12:11, 1.65it/s] 59%\|█████▉ \| 1768/2975 [21:57<12:12, 1.65it/s] 59%\|█████▉ \| 1769/2975 [21:57<12:12, 1.65it/s] 59%\|█████▉ \| 1770/2975 [21:58<12:14, 1.64it/s] 59%\|█████▉ \| 1770/2975 [21:58<12:14, 1.64it/s] 60%\|█████▉ \| 1771/2975 [21:58<12:14, 1.64it/s] 60%\|█████▉ \| 1772/2975 [21:59<12:13, 1.64it/s] 60%\|█████▉ \| 1773/2975 [22:00<12:12, 1.64it/s] 60%\|█████▉ \| 1774/2975 [22:00<12:11, 1.64it/s] 60%\|█████▉ \| 1775/2975 [22:01<12:10, 1.64it/s] 60%\|█████▉ \| 1776/2975 [22:01<12:09, 1.64it/s] 60%\|█████▉ \| 1777/2975
	0: {'loss': 0.7076, 'grad_norm': 0.738490278103292, 'learning_rate': 1e-05, 'epoch': 0.6}
	0: [22:02<12:07, 1.65it/s] 60%\|█████▉ \| 1778/2975 [22:03<12:06, 1.65it/s] 60%\|█████▉ \| 1779/2975 [22:03<12:06, 1.65it/s] 60%\|█████▉ \| 1780/2975 [22:04<12:05, 1.65it/s] 60%\|█████▉ \| 1780/2975 [22:04<12:05, 1.65it/s] 60%\|█████▉ \| 1781/2975 [22:05<12:05, 1.65it/s] 60%\|█████▉ \| 1782/2975 [22:05<12:05, 1.64it/s] 60%\|█████▉ \| 1783/2975 [22:06<12:04, 1.64it/s] 60%\|█████▉ \| 1784/2975 [22:06<12:04, 1.64it/s] 60%\|██████ \| 1785/2975 [22:07<12:04, 1.64it/s] 60%\|██████ \| 1786/2975 [22:10<24:05, 1.22s/it] 60%\|██████ \| 1787/2975 [22:10<20:28, 1.03s/it] 60%\|██████ \| 1788/2975 [22:11<17:56, 1.10it/s] 60%\|██████ \| 1789/2975 [22:11<16:09, 1.22it/s] 60%\|██████ \| 1790/2975 [22:12<14:53, 1.33it/s]
	0: {'loss': 0.7015, 'grad_norm': 0.7563688162947536, 'learning_rate': 1e-05, 'epoch': 0.6}
	0: {'loss': 0.7154, 'grad_norm': 0.7767154655838906, 'learning_rate': 1e-05, 'epoch': 0.6}
	0: 60%\|██████ \| 1790/2975 [22:12<14:53, 1.33it/s] 60%\|██████ \| 1791/2975 [22:13<14:00, 1.41it/s] 60%\|██████ \| 1792/2975 [22:13<13:23, 1.47it/s] 60%\|██████ \| 1793/2975 [22:14<12:57, 1.52it/s] 60%\|██████ \| 1794/2975 [22:14<12:39, 1.56it/s] 60%\|██████ \| 1795/2975 [22:15<12:25, 1.58it/s] 60%\|██████ \| 1796/2975 [22:16<12:15, 1.60it/s] 60%\|██████ \| 1797/2975 [22:16<12:09, 1.61it/s] 60%\|██████ \| 1798/2975 [22:17<12:04, 1.62it/s] 60%\|██████ \| 1799/2975 [22:17<12:00, 1.63it/s] 61%\|██████ \| 1800/2975 [22:18<11:57, 1.64it/s] 61%\|██████ \| 1800/2975 [22:18<11:57, 1.64it/s] 61%\|██████ \| 1801/2975 [22:19<11:58, 1.63it/s] 61%\|██████ \| 1802/2975 [22:19<11:59, 1.63it/s] 61%\|██████ \| 1803/2975 [22:20<11:57, 1.63it/s] 61%\|█
	0: {'loss': 0.7147, 'grad_norm': 0.7588920449284606, 'learning_rate': 1e-05, 'epoch': 0.61}
	0: █████ \| 1804/2975 [22:21<11:56, 1.63it/s] 61%\|██████ \| 1805/2975 [22:21<11:55, 1.64it/s] 61%\|██████ \| 1806/2975 [22:22<11:53, 1.64it/s] 61%\|██████ \| 1807/2975 [22:22<11:52, 1.64it/s] 61%\|██████ \| 1808/2975 [22:23<11:50, 1.64it/s] 61%\|██████ \| 1809/2975 [22:24<11:48, 1.65it/s] 61%\|██████ \| 1810/2975 [22:24<11:47, 1.65it/s] 61%\|██████ \| 1810/2975 [22:24<11:47, 1.65it/s] 61%\|██████ \| 1811/2975 [22:25<11:48, 1.64it/s] 61%\|██████ \| 1812/2975 [22:25<11:47, 1.64it/s] 61%\|██████ \| 1813/2975 [22:26<11:46, 1.65it/s] 61%\|██████ \| 1814/2975 [22:27<11:45, 1.65it/s] 61%\|██████ \| 1815/2975 [22:27<11:45, 1.64it/s] 61%\|██████ \| 1816/2975 [22:28<11:46, 1.64it/s] 61%\|██████ \| 1817/2975 [22:28<11:46, 1.64it/s] 61%\|████�
	0: {'loss': 0.7259, 'grad_norm': 0.7499492826943769, 'learning_rate': 1e-05, 'epoch': 0.61}
	0: {'loss': 0.7127, 'grad_norm': 0.7463183831563831, 'learning_rate': 1e-05, 'epoch': 0.61}
	0: �█ \| 1818/2975 [22:29<11:47, 1.64it/s] 61%\|██████ \| 1819/2975 [22:30<11:46, 1.64it/s] 61%\|██████ \| 1820/2975 [22:30<11:45, 1.64it/s] 61%\|██████ \| 1820/2975 [22:30<11:45, 1.64it/s] 61%\|██████ \| 1821/2975 [22:31<11:44, 1.64it/s] 61%\|██████ \| 1822/2975 [22:32<11:42, 1.64it/s] 61%\|██████▏ \| 1823/2975 [22:32<11:41, 1.64it/s] 61%\|██████▏ \| 1824/2975 [22:33<11:40, 1.64it/s] 61%\|██████▏ \| 1825/2975 [22:33<11:40, 1.64it/s] 61%\|██████▏ \| 1826/2975 [22:34<11:39, 1.64it/s] 61%\|██████▏ \| 1827/2975 [22:35<11:38, 1.64it/s] 61%\|██████▏ \| 1828/2975 [22:35<11:37, 1.64it/s] 61%\|██████▏ \| 1829/2975 [22:36<11:37, 1.64it/s] 62%\|██████▏ \| 1830/2975 [22:36<11:37, 1.64it/s] 62%\|██████�
	0: {'loss': 0.7244, 'grad_norm': 0.7877095775584843, 'learning_rate': 1e-05, 'epoch': 0.62}
	0: � \| 1830/2975 [22:36<11:37, 1.64it/s] 62%\|██████▏ \| 1831/2975 [22:37<11:38, 1.64it/s] 62%\|██████▏ \| 1832/2975 [22:38<11:42, 1.63it/s] 62%\|██████▏ \| 1833/2975 [22:38<11:40, 1.63it/s] 62%\|██████▏ \| 1834/2975 [22:39<11:38, 1.63it/s] 62%\|██████▏ \| 1835/2975 [22:39<11:37, 1.64it/s] 62%\|██████▏ \| 1836/2975 [22:40<11:33, 1.64it/s] 62%\|██████▏ \| 1837/2975 [22:41<11:32, 1.64it/s] 62%\|██████▏ \| 1838/2975 [22:41<11:32, 1.64it/s] 62%\|██████▏ \| 1839/2975 [22:42<11:31, 1.64it/s] 62%\|██████▏ \| 1840/2975 [22:42<11:29, 1.65it/s] 62%\|██████▏ \| 1840/2975 [22:42<11:29, 1.65it/s] 62%\|██████▏ \| 1841/2975 [22:43<11:29, 1.65it/s] 62%\|██████▏ \| 1842/2975 [22:44<11:29, 1.64it/s] 62%\|██████▏ \| 1843/2975 [22:44<11:28, 1.64it/s] 62%\|�
	0: {'loss': 0.7075, 'grad_norm': 0.7279988466982945, 'learning_rate': 1e-05, 'epoch': 0.62}
	0: ��█████▏ \| 1844/2975 [22:45<11:26, 1.65it/s] 62%\|██████▏ \| 1845/2975 [22:46<11:28, 1.64it/s] 62%\|██████▏ \| 1846/2975 [22:46<11:27, 1.64it/s] 62%\|██████▏ \| 1847/2975 [22:47<11:26, 1.64it/s] 62%\|██████▏ \| 1848/2975 [22:47<11:25, 1.64it/s] 62%\|██████▏ \| 1849/2975 [22:48<11:25, 1.64it/s] 62%\|██████▏ \| 1850/2975 [22:49<11:24, 1.64it/s] 62%\|██████▏ \| 1850/2975 [22:49<11:24, 1.64it/s] 62%\|██████▏ \| 1851/2975 [22:49<11:24, 1.64it/s] 62%\|██████▏ \| 1852/2975 [22:50<11:23, 1.64it/s] 62%\|██████▏ \| 1853/2975 [22:50<11:21, 1.65it/s] 62%\|██████▏ \| 1854/2975 [22:51<11:21, 1.65it/s] 62%\|██████▏ \| 1855/2975 [22:52<11:20, 1.65it/s] 62%\|██████▏ \| 1856/2975 [22:52<11:20, 1.64it/s] 62%\|██████▏ \| 1857/2975 [22:53<11:19
	0: {'loss': 0.7137, 'grad_norm': 0.7349254032326754, 'learning_rate': 1e-05, 'epoch': 0.62}
	0: , 1.65it/s] 62%\|██████▏ \| 1858/2975 [22:53<11:19, 1.64it/s] 62%\|██████▏ \| 1859/2975 [22:54<11:18, 1.65it/s] 63%\|██████▎ \| 1860/2975 [22:55<11:17, 1.64it/s] 63%\|██████▎ \| 1860/2975 [22:55<11:17, 1.64it/s] 63%\|██████▎ \| 1861/2975 [22:55<11:18, 1.64it/s] 63%\|██████▎ \| 1862/2975 [22:56<11:18, 1.64it/s] 63%\|██████▎ \| 1863/2975 [22:56<11:18, 1.64it/s] 63%\|██████▎ \| 1864/2975 [22:57<11:18, 1.64it/s] 63%\|██████▎ \| 1865/2975 [22:58<11:18, 1.64it/s] 63%\|██████▎ \| 1866/2975 [22:58<11:18, 1.63it/s] 63%\|██████▎ \| 1867/2975 [22:59<11:17, 1.64it/s] 63%\|██████▎ \| 1868/2975 [23:00<11:17, 1.63it/s] 63%\|██████▎ \| 1869/2975 [23:00<11:16, 1.64it/s] 63%\|██████▎ \| 1870/2975 [23:01<11:14, 1.64it/s]
	0: {'loss': 0.704, 'grad_norm': 0.7241088554589961, 'learning_rate': 1e-05, 'epoch': 0.63}
	0: {'loss': 0.7115, 'grad_norm': 0.7283852481874094, 'learning_rate': 1e-05, 'epoch': 0.63}
	0: 63%\|██████▎ \| 1870/2975 [23:01<11:14, 1.64it/s] 63%\|██████▎ \| 1871/2975 [23:01<11:13, 1.64it/s] 63%\|██████▎ \| 1872/2975 [23:02<11:11, 1.64it/s] 63%\|██████▎ \| 1873/2975 [23:03<11:11, 1.64it/s] 63%\|██████▎ \| 1874/2975 [23:03<11:09, 1.64it/s] 63%\|██████▎ \| 1875/2975 [23:04<11:11, 1.64it/s] 63%\|██████▎ \| 1876/2975 [23:04<11:10, 1.64it/s] 63%\|██████▎ \| 1877/2975 [23:05<11:09, 1.64it/s] 63%\|██████▎ \| 1878/2975 [23:06<11:09, 1.64it/s] 63%\|██████▎ \| 1879/2975 [23:06<11:08, 1.64it/s] 63%\|██████▎ \| 1880/2975 [23:07<11:08, 1.64it/s] 63%\|██████▎ \| 1880/2975 [23:07<11:08, 1.64it/s] 63%\|██████▎ \| 1881/2975 [23:07<11:07, 1.64it/s] 63%\|██████▎ \| 1882/2975 [23:08<11:05, 1.64it/s] 63%\|██████▎
	0: {'loss': 0.7055, 'grad_norm': 0.7733755864857219, 'learning_rate': 1e-05, 'epoch': 0.64}
	0: \| 1883/2975 [23:09<11:04, 1.64it/s] 63%\|██████▎ \| 1884/2975 [23:09<11:03, 1.64it/s] 63%\|██████▎ \| 1885/2975 [23:10<11:02, 1.64it/s] 63%\|██████▎ \| 1886/2975 [23:11<11:01, 1.65it/s] 63%\|██████▎ \| 1887/2975 [23:11<11:00, 1.65it/s] 63%\|██████▎ \| 1888/2975 [23:12<11:01, 1.64it/s] 63%\|██████▎ \| 1889/2975 [23:12<11:00, 1.64it/s] 64%\|██████▎ \| 1890/2975 [23:13<11:00, 1.64it/s] 64%\|██████▎ \| 1890/2975 [23:13<11:00, 1.64it/s] 64%\|██████▎ \| 1891/2975 [23:14<11:01, 1.64it/s] 64%\|██████▎ \| 1892/2975 [23:14<11:01, 1.64it/s] 64%\|██████▎ \| 1893/2975 [23:15<11:00, 1.64it/s] 64%\|██████▎ \| 1894/2975 [23:15<10:59, 1.64it/s] 64%\|██████▎ \| 1895/2975 [23:16<10:57, 1.64it/s] 64%\|██████▎ \| 1896/2975 [23:17<10:56, 1.64it/s] 64%\|█�
	0: {'loss': 0.701, 'grad_norm': 0.8224231396862601, 'learning_rate': 1e-05, 'epoch': 0.64}
	0: ��████▍ \| 1897/2975 [23:17<10:55, 1.65it/s] 64%\|██████▍ \| 1898/2975 [23:18<10:54, 1.65it/s] 64%\|██████▍ \| 1899/2975 [23:18<10:53, 1.65it/s] 64%\|██████▍ \| 1900/2975 [23:19<10:51, 1.65it/s] 64%\|██████▍ \| 1900/2975 [23:19<10:51, 1.65it/s] 64%\|██████▍ \| 1901/2975 [23:20<10:51, 1.65it/s] 64%\|██████▍ \| 1902/2975 [23:20<10:50, 1.65it/s] 64%\|██████▍ \| 1903/2975 [23:21<10:51, 1.65it/s] 64%\|██████▍ \| 1904/2975 [23:21<10:51, 1.64it/s] 64%\|██████▍ \| 1905/2975 [23:22<10:50, 1.64it/s] 64%\|██████▍ \| 1906/2975 [23:23<10:50, 1.64it/s] 64%\|██████▍ \| 1907/2975 [23:23<10:49, 1.64it/s] 64%\|██████▍ \| 1908/2975 [23:24<10:48, 1.64it/s] 64%\|██████▍ \| 1909/2975 [23:25<10:48, 1.64it/s] 64%\|██████▍ \| 1910/2975 [23:25<10:47,
	0: {'loss': 0.7237, 'grad_norm': 0.7734811371365439, 'learning_rate': 1e-05, 'epoch': 0.64}
	0: {'loss': 0.695, 'grad_norm': 0.7196135161346667, 'learning_rate': 1e-05, 'epoch': 0.65}
	0: 1.65it/s] 64%\|██████▍ \| 1910/2975 [23:25<10:47, 1.65it/s] 64%\|██████▍ \| 1911/2975 [23:26<10:46, 1.65it/s] 64%\|██████▍ \| 1912/2975 [23:26<10:46, 1.64it/s] 64%\|██████▍ \| 1913/2975 [23:27<10:46, 1.64it/s] 64%\|██████▍ \| 1914/2975 [23:28<10:44, 1.65it/s] 64%\|██████▍ \| 1915/2975 [23:28<10:43, 1.65it/s] 64%\|██████▍ \| 1916/2975 [23:29<10:43, 1.64it/s] 64%\|██████▍ \| 1917/2975 [23:29<10:45, 1.64it/s] 64%\|██████▍ \| 1918/2975 [23:30<10:43, 1.64it/s] 65%\|██████▍ \| 1919/2975 [23:31<10:42, 1.64it/s] 65%\|██████▍ \| 1920/2975 [23:31<10:42, 1.64it/s] 65%\|██████▍ \| 1920/2975 [23:31<10:42, 1.64it/s] 65%\|██████▍ \| 1921/2975 [23:32<10:43, 1.64it/s] 65%\|██████▍ \| 1922/2975 [23:32<10
	0: {'loss': 0.7186, 'grad_norm': 0.7760509033852916, 'learning_rate': 1e-05, 'epoch': 0.65}
	0: :42, 1.64it/s] 65%\|██████▍ \| 1923/2975 [23:33<10:40, 1.64it/s] 65%\|██████▍ \| 1924/2975 [23:34<10:40, 1.64it/s] 65%\|██████▍ \| 1925/2975 [23:34<10:40, 1.64it/s] 65%\|██████▍ \| 1926/2975 [23:35<10:41, 1.64it/s] 65%\|██████▍ \| 1927/2975 [23:35<10:40, 1.64it/s] 65%\|██████▍ \| 1928/2975 [23:36<11:41, 1.49it/s] 65%\|██████▍ \| 1929/2975 [23:37<11:21, 1.53it/s] 65%\|██████▍ \| 1930/2975 [23:37<11:06, 1.57it/s] 65%\|██████▍ \| 1930/2975 [23:37<11:06, 1.57it/s] 65%\|██████▍ \| 1931/2975 [23:38<10:55, 1.59it/s] 65%\|██████▍ \| 1932/2975 [23:39<10:47, 1.61it/s] 65%\|██████▍ \| 1933/2975 [23:39<10:42, 1.62it/s] 65%\|██████▌ \| 1934/2975 [23:40<10:39, 1.63it/s] 65%\|██████▌ \| 1935/2975 [23:41<10:36, 1.64it/s] 65%\|██████▌ \|
	0: {'loss': 0.6914, 'grad_norm': 0.7313402259756968, 'learning_rate': 1e-05, 'epoch': 0.65}
	0: 1936/2975 [23:41<10:33, 1.64it/s] 65%\|██████▌ \| 1937/2975 [23:42<10:31, 1.64it/s] 65%\|██████▌ \| 1938/2975 [23:42<10:30, 1.64it/s] 65%\|██████▌ \| 1939/2975 [23:43<10:29, 1.65it/s] 65%\|██████▌ \| 1940/2975 [23:44<10:28, 1.65it/s] 65%\|██████▌ \| 1940/2975 [23:44<10:28, 1.65it/s] 65%\|██████▌ \| 1941/2975 [23:44<10:28, 1.65it/s] 65%\|██████▌ \| 1942/2975 [23:45<10:26, 1.65it/s] 65%\|██████▌ \| 1943/2975 [23:45<10:25, 1.65it/s] 65%\|██████▌ \| 1944/2975 [23:46<10:24, 1.65it/s] 65%\|██████▌ \| 1945/2975 [23:47<10:25, 1.65it/s] 65%\|██████▌ \| 1946/2975 [23:47<10:26, 1.64it/s] 65%\|██████▌ \| 1947/2975 [23:48<10:26, 1.64it/s] 65%\|██████▌ \| 1948/2975 [23:48<10:26, 1.64it/s] 66%\|██████▌ \| 1949/2975 [23:49<10:26, 1.64it/s] 66%\|██�
	0: {'loss': 0.7146, 'grad_norm': 0.734114065969648, 'learning_rate': 1e-05, 'epoch': 0.66}
	0: {'loss': 0.7279, 'grad_norm': 0.7403305578088101, 'learning_rate': 1e-05, 'epoch': 0.66}
	0: ��███▌ \| 1950/2975 [23:50<10:24, 1.64it/s] 66%\|██████▌ \| 1950/2975 [23:50<10:24, 1.64it/s] 66%\|██████▌ \| 1951/2975 [23:50<10:24, 1.64it/s] 66%\|██████▌ \| 1952/2975 [23:51<10:22, 1.64it/s] 66%\|██████▌ \| 1953/2975 [23:51<10:21, 1.64it/s] 66%\|██████▌ \| 1954/2975 [23:52<10:20, 1.65it/s] 66%\|██████▌ \| 1955/2975 [23:53<10:18, 1.65it/s] 66%\|██████▌ \| 1956/2975 [23:53<10:18, 1.65it/s] 66%\|██████▌ \| 1957/2975 [23:54<10:17, 1.65it/s] 66%\|██████▌ \| 1958/2975 [23:54<10:16, 1.65it/s] 66%\|██████▌ \| 1959/2975 [23:55<10:14, 1.65it/s] 66%\|██████▌ \| 1960/2975 [23:56<10:13, 1.65it/s] 66%\|██████▌ \| 1960/2975 [23:56<10:13, 1.65it/s] 66%\|██████▌ \| 1961/2975 [23:56<10:14, 1.65it/s] 66%\|�
	0: {'loss': 0.7024, 'grad_norm': 0.7210850729422128, 'learning_rate': 1e-05, 'epoch': 0.66}
	0: ��█████▌ \| 1962/2975 [23:57<10:14, 1.65it/s] 66%\|██████▌ \| 1963/2975 [23:58<10:16, 1.64it/s] 66%\|██████▌ \| 1964/2975 [23:58<10:17, 1.64it/s] 66%\|██████▌ \| 1965/2975 [23:59<10:17, 1.63it/s] 66%\|██████▌ \| 1966/2975 [23:59<10:18, 1.63it/s] 66%\|██████▌ \| 1967/2975 [24:00<10:17, 1.63it/s] 66%\|██████▌ \| 1968/2975 [24:01<10:15, 1.64it/s] 66%\|██████▌ \| 1969/2975 [24:01<10:13, 1.64it/s] 66%\|██████▌ \| 1970/2975 [24:02<10:11, 1.64it/s] 66%\|██████▌ \| 1970/2975 [24:02<10:11, 1.64it/s] 66%\|██████▋ \| 1971/2975 [24:02<10:10, 1.64it/s] 66%\|██████▋ \| 1972/2975 [24:03<10:09, 1.65it/s] 66%\|██████▋ \| 1973/2975 [24:04<10:08, 1.65it/s] 66%\|██████▋ \| 1974/2975 [24:04<10:07, 1.65it/s] 66%\|██████▋ \| 1975/2975 [24:05<10:06
	0: {'loss': 0.7115, 'grad_norm': 0.7123479453497946, 'learning_rate': 1e-05, 'epoch': 0.67}
	0: , 1.65it/s] 66%\|██████▋ \| 1976/2975 [24:05<10:06, 1.65it/s] 66%\|██████▋ \| 1977/2975 [24:06<10:05, 1.65it/s] 66%\|██████▋ \| 1978/2975 [24:07<10:05, 1.65it/s] 67%\|██████▋ \| 1979/2975 [24:07<10:06, 1.64it/s] 67%\|██████▋ \| 1980/2975 [24:08<10:06, 1.64it/s] 67%\|██████▋ \| 1980/2975 [24:08<10:06, 1.64it/s] 67%\|██████▋ \| 1981/2975 [24:09<10:07, 1.64it/s] 67%\|██████▋ \| 1982/2975 [24:09<10:06, 1.64it/s] 67%\|██████▋ \| 1983/2975 [24:10<10:06, 1.64it/s] 67%\|██████▋ \| 1984/2975 [24:10<10:05, 1.64it/s] 67%\|██████▋ \| 1985/2975 [24:11<10:04, 1.64it/s] 67%\|██████▋ \| 1986/2975 [24:12<10:02, 1.64it/s] 67%\|██████▋ \| 1987/2975 [24:12<10:00, 1.64it/s] 67%\|██████▋ \| 1988/2975 [24:13<09:59, 1.65it/s] 67%\|██████▋ \| 198
	0: {'loss': 0.7311, 'grad_norm': 0.7600061639464267, 'learning_rate': 1e-05, 'epoch': 0.67}
	0: {'loss': 0.7109, 'grad_norm': 0.7994233915758029, 'learning_rate': 1e-05, 'epoch': 0.67}
	0: 9/2975 [24:13<09:58, 1.65it/s] 67%\|██████▋ \| 1990/2975 [24:14<09:57, 1.65it/s] 67%\|██████▋ \| 1990/2975 [24:14<09:57, 1.65it/s] 67%\|██████▋ \| 1991/2975 [24:15<09:57, 1.65it/s] 67%\|██████▋ \| 1992/2975 [24:15<09:57, 1.65it/s] 67%\|██████▋ \| 1993/2975 [24:16<09:56, 1.65it/s] 67%\|██████▋ \| 1994/2975 [24:16<09:55, 1.65it/s] 67%\|██████▋ \| 1995/2975 [24:17<09:54, 1.65it/s] 67%\|██████▋ \| 1996/2975 [24:18<09:54, 1.65it/s] 67%\|██████▋ \| 1997/2975 [24:18<09:54, 1.65it/s] 67%\|██████▋ \| 1998/2975 [24:19<09:55, 1.64it/s] 67%\|██████▋ \| 1999/2975 [24:19<09:54, 1.64it/s] 67%\|██████▋ \| 2000/2975 [24:20<09:53, 1.64it/s] 67%\|██████▋ \| 2000/2975 [24:20<09:53, 1.64it/s] 67%\|██████▋
	0: {'loss': 0.7085, 'grad_norm': 0.7520626757283446, 'learning_rate': 1e-05, 'epoch': 0.68}
	0: \| 2001/2975 [24:21<09:53, 1.64it/s] 67%\|██████▋ \| 2002/2975 [24:21<09:52, 1.64it/s] 67%\|██████▋ \| 2003/2975 [24:22<09:51, 1.64it/s] 67%\|██████▋ \| 2004/2975 [24:22<09:50, 1.65it/s] 67%\|██████▋ \| 2005/2975 [24:23<09:48, 1.65it/s] 67%\|██████▋ \| 2006/2975 [24:24<09:47, 1.65it/s] 67%\|██████▋ \| 2007/2975 [24:24<09:47, 1.65it/s] 67%\|██████▋ \| 2008/2975 [24:25<09:46, 1.65it/s] 68%\|██████▊ \| 2009/2975 [24:26<09:45, 1.65it/s] 68%\|██████▊ \| 2010/2975 [24:26<09:43, 1.65it/s] 68%\|██████▊ \| 2010/2975 [24:26<09:43, 1.65it/s] 68%\|██████▊ \| 2011/2975 [24:27<09:43, 1.65it/s] 68%\|██████▊ \| 2012/2975 [24:27<09:42, 1.65it/s] 68%\|██████▊ \| 2013/2975 [24:28<09:42, 1.65it/s] 68%\|██████▊ \| 2014/2975 [24:29<09:42, 1.65it/s] 68%\|█�
	0: {'loss': 0.7145, 'grad_norm': 0.7895646915881295, 'learning_rate': 1e-05, 'epoch': 0.68}
	0: ��████▊ \| 2015/2975 [24:29<09:42, 1.65it/s] 68%\|██████▊ \| 2016/2975 [24:30<09:42, 1.65it/s] 68%\|██████▊ \| 2017/2975 [24:30<09:43, 1.64it/s] 68%\|██████▊ \| 2018/2975 [24:31<09:43, 1.64it/s] 68%\|██████▊ \| 2019/2975 [24:32<09:43, 1.64it/s] 68%\|██████▊ \| 2020/2975 [24:32<09:42, 1.64it/s] 68%\|██████▊ \| 2020/2975 [24:32<09:42, 1.64it/s] 68%\|██████▊ \| 2021/2975 [24:33<09:42, 1.64it/s] 68%\|██████▊ \| 2022/2975 [24:33<09:41, 1.64it/s] 68%\|██████▊ \| 2023/2975 [24:34<09:40, 1.64it/s] 68%\|██████▊ \| 2024/2975 [24:35<09:38, 1.64it/s] 68%\|██████▊ \| 2025/2975 [24:35<09:37, 1.65it/s] 68%\|██████▊ \| 2026/2975 [24:36<09:36, 1.65it/s] 68%\|██████▊ \| 2027/2975 [24:36<09:35, 1.65it/s] 68%\|██████▊ \| 2028/2975 [24:37<09:34,
	0: {'loss': 0.7168, 'grad_norm': 0.7015561317834488, 'learning_rate': 1e-05, 'epoch': 0.68}
	0: {'loss': 0.7098, 'grad_norm': 0.7445727424355767, 'learning_rate': 1e-05, 'epoch': 0.69}
	0: 1.65it/s] 68%\|██████▊ \| 2029/2975 [24:38<09:32, 1.65it/s] 68%\|██████▊ \| 2030/2975 [24:38<09:32, 1.65it/s] 68%\|██████▊ \| 2030/2975 [24:38<09:32, 1.65it/s] 68%\|██████▊ \| 2031/2975 [24:39<09:32, 1.65it/s] 68%\|██████▊ \| 2032/2975 [24:39<09:32, 1.65it/s] 68%\|██████▊ \| 2033/2975 [24:40<09:32, 1.65it/s] 68%\|██████▊ \| 2034/2975 [24:41<09:31, 1.65it/s] 68%\|██████▊ \| 2035/2975 [24:41<09:30, 1.65it/s] 68%\|██████▊ \| 2036/2975 [24:42<09:30, 1.65it/s] 68%\|██████▊ \| 2037/2975 [24:43<09:30, 1.64it/s] 69%\|██████▊ \| 2038/2975 [24:43<09:29, 1.64it/s] 69%\|██████▊ \| 2039/2975 [24:44<09:30, 1.64it/s] 69%\|██████▊ \| 2040/2975 [24:44<09:30, 1.64it/s] 69%\|██████▊ \| 2040/2975 [24:44<09
	0: {'loss': 0.6964, 'grad_norm': 0.7581584631465635, 'learning_rate': 1e-05, 'epoch': 0.69}
	0: :30, 1.64it/s] 69%\|██████▊ \| 2041/2975 [24:45<09:29, 1.64it/s] 69%\|██████▊ \| 2042/2975 [24:46<09:28, 1.64it/s] 69%\|██████▊ \| 2043/2975 [24:46<09:27, 1.64it/s] 69%\|██████▊ \| 2044/2975 [24:47<09:25, 1.65it/s] 69%\|██████▊ \| 2045/2975 [24:47<09:24, 1.65it/s] 69%\|██████▉ \| 2046/2975 [24:48<09:23, 1.65it/s] 69%\|██████▉ \| 2047/2975 [24:49<09:22, 1.65it/s] 69%\|██████▉ \| 2048/2975 [24:49<09:23, 1.65it/s] 69%\|██████▉ \| 2049/2975 [24:50<09:23, 1.64it/s] 69%\|██████▉ \| 2050/2975 [24:50<09:23, 1.64it/s] 69%\|██████▉ \| 2050/2975 [24:50<09:23, 1.64it/s] 69%\|██████▉ \| 2051/2975 [24:51<09:22, 1.64it/s] 69%\|██████▉ \| 2052/2975 [24:52<09:22, 1.64it/s] 69%\|██████▉ \| 2053/2975 [24:52<09:22, 1.64it/s] 69%\|██████▉ \|
	0: {'loss': 0.7142, 'grad_norm': 0.7617928015384502, 'learning_rate': 1e-05, 'epoch': 0.69}
	0: 2054/2975 [24:53<09:21, 1.64it/s] 69%\|██████▉ \| 2055/2975 [24:53<09:20, 1.64it/s] 69%\|██████▉ \| 2056/2975 [24:54<09:20, 1.64it/s] 69%\|██████▉ \| 2057/2975 [24:55<09:19, 1.64it/s] 69%\|██████▉ \| 2058/2975 [24:55<09:18, 1.64it/s] 69%\|██████▉ \| 2059/2975 [24:56<09:17, 1.64it/s] 69%\|██████▉ \| 2060/2975 [24:57<09:16, 1.64it/s] 69%\|██████▉ \| 2060/2975 [24:57<09:16, 1.64it/s] 69%\|██████▉ \| 2061/2975 [24:57<09:16, 1.64it/s] 69%\|██████▉ \| 2062/2975 [24:58<09:16, 1.64it/s] 69%\|██████▉ \| 2063/2975 [24:58<09:24, 1.61it/s] 69%\|██████▉ \| 2064/2975 [24:59<09:22, 1.62it/s] 69%\|██████▉ \| 2065/2975 [25:00<09:19, 1.63it/s] 69%\|██████▉ \| 2066/2975 [25:00<09:17, 1.63it/s] 69%\|██████▉ \| 2067/2975 [25:01<09:15, 1.64it/s] 70%\|██�
	0: {'loss': 0.6963, 'grad_norm': 0.8090055655264392, 'learning_rate': 1e-05, 'epoch': 0.7}
	0: {'loss': 0.7027, 'grad_norm': 0.721421977318063, 'learning_rate': 1e-05, 'epoch': 0.7}
	0: ��███▉ \| 2068/2975 [25:01<09:13, 1.64it/s] 70%\|██████▉ \| 2069/2975 [25:02<09:10, 1.65it/s] 70%\|██████▉ \| 2070/2975 [25:03<09:09, 1.65it/s] 70%\|██████▉ \| 2070/2975 [25:03<09:09, 1.65it/s] 70%\|██████▉ \| 2071/2975 [25:03<09:09, 1.65it/s] 70%\|██████▉ \| 2072/2975 [25:04<09:08, 1.64it/s] 70%\|██████▉ \| 2073/2975 [25:04<09:09, 1.64it/s] 70%\|██████▉ \| 2074/2975 [25:05<09:08, 1.64it/s] 70%\|██████▉ \| 2075/2975 [25:06<09:06, 1.65it/s] 70%\|██████▉ \| 2076/2975 [25:06<09:06, 1.64it/s] 70%\|██████▉ \| 2077/2975 [25:07<09:06, 1.64it/s] 70%\|██████▉ \| 2078/2975 [25:08<09:06, 1.64it/s] 70%\|██████▉ \| 2079/2975 [25:08<09:16, 1.61it/s] 70%\|██████▉ \| 2080/2975 [25:09<09:12, 1.62it/s] 70%\|�
	0: {'loss': 0.7205, 'grad_norm': 0.7883438921335163, 'learning_rate': 1e-05, 'epoch': 0.7}
	0: ��█████▉ \| 2080/2975 [25:09<09:12, 1.62it/s] 70%\|██████▉ \| 2081/2975 [25:09<09:10, 1.62it/s] 70%\|██████▉ \| 2082/2975 [25:10<09:09, 1.63it/s] 70%\|███████ \| 2083/2975 [25:11<09:07, 1.63it/s] 70%\|███████ \| 2084/2975 [25:11<09:05, 1.63it/s] 70%\|███████ \| 2085/2975 [25:12<09:04, 1.64it/s] 70%\|███████ \| 2086/2975 [25:12<09:02, 1.64it/s] 70%\|███████ \| 2087/2975 [25:13<09:00, 1.64it/s] 70%\|███████ \| 2088/2975 [25:14<09:00, 1.64it/s] 70%\|███████ \| 2089/2975 [25:14<08:59, 1.64it/s] 70%\|███████ \| 2090/2975 [25:15<08:58, 1.64it/s] 70%\|███████ \| 2090/2975 [25:15<08:58, 1.64it/s] 70%\|███████ \| 2091/2975 [25:15<08:58, 1.64it/s] 70%\|███████ \| 2092/2975 [25:16<08:57, 1.64it/s] 70%\|███████ \| 2093/2975 [25:17<08:56
	0: {'loss': 0.713, 'grad_norm': 0.7444611156598626, 'learning_rate': 1e-05, 'epoch': 0.71}
	0: , 1.64it/s] 70%\|███████ \| 2094/2975 [25:17<08:55, 1.64it/s] 70%\|███████ \| 2095/2975 [25:18<08:55, 1.64it/s] 70%\|███████ \| 2096/2975 [25:19<08:55, 1.64it/s] 70%\|███████ \| 2097/2975 [25:19<08:54, 1.64it/s] 71%\|███████ \| 2098/2975 [25:20<08:54, 1.64it/s] 71%\|███████ \| 2099/2975 [25:20<08:54, 1.64it/s] 71%\|███████ \| 2100/2975 [25:21<08:53, 1.64it/s] 71%\|███████ \| 2100/2975 [25:21<08:53, 1.64it/s] 71%\|███████ \| 2101/2975 [25:22<08:52, 1.64it/s] 71%\|███████ \| 2102/2975 [25:22<08:51, 1.64it/s] 71%\|███████ \| 2103/2975 [25:23<08:49, 1.65it/s] 71%\|███████ \| 2104/2975 [25:23<08:48, 1.65it/s] 71%\|███████ \| 2105/2975 [25:24<08:48, 1.65it/s] 71%\|███████ \| 2106/2975 [25:25<08:47, 1.65it/s] 71%\|███████ \| 210
	0: {'loss': 0.6955, 'grad_norm': 0.7425154567734555, 'learning_rate': 1e-05, 'epoch': 0.71}
	0: 7/2975 [25:25<08:46, 1.65it/s] 71%\|███████ \| 2108/2975 [25:26<08:45, 1.65it/s] 71%\|███████ \| 2109/2975 [25:26<08:45, 1.65it/s] 71%\|███████ \| 2110/2975 [25:27<08:45, 1.65it/s] 71%\|███████ \| 2110/2975 [25:27<08:45, 1.65it/s] 71%\|███████ \| 2111/2975 [25:28<08:45, 1.65it/s] 71%\|███████ \| 2112/2975 [25:28<08:44, 1.64it/s] 71%\|███████ \| 2113/2975 [25:29<08:45, 1.64it/s] 71%\|███████ \| 2114/2975 [25:29<08:45, 1.64it/s] 71%\|███████ \| 2115/2975 [25:30<08:45, 1.64it/s] 71%\|███████ \| 2116/2975 [25:31<08:44, 1.64it/s] 71%\|███████ \| 2117/2975 [25:31<08:42, 1.64it/s] 71%\|███████ \| 2118/2975 [25:32<08:41, 1.64it/s] 71%\|███████ \| 2119/2975 [25:33<08:40, 1.64it/s] 71%\|███████▏ \| 2120/2975 [25:33<08:38, 1.65it/s]
	0: {'loss': 0.7219, 'grad_norm': 0.7459426729679086, 'learning_rate': 1e-05, 'epoch': 0.71}
	0: {'loss': 0.7215, 'grad_norm': 0.7310305789819423, 'learning_rate': 1e-05, 'epoch': 0.72}
	0: 71%\|███████▏ \| 2120/2975 [25:33<08:38, 1.65it/s] 71%\|███████▏ \| 2121/2975 [25:34<08:37, 1.65it/s] 71%\|███████▏ \| 2122/2975 [25:34<08:36, 1.65it/s] 71%\|███████▏ \| 2123/2975 [25:35<08:35, 1.65it/s] 71%\|███████▏ \| 2124/2975 [25:36<08:34, 1.66it/s] 71%\|███████▏ \| 2125/2975 [25:36<08:33, 1.66it/s] 71%\|███████▏ \| 2126/2975 [25:37<08:32, 1.66it/s] 71%\|███████▏ \| 2127/2975 [25:37<08:31, 1.66it/s] 72%\|███████▏ \| 2128/2975 [25:38<08:30, 1.66it/s] 72%\|███████▏ \| 2129/2975 [25:39<08:29, 1.66it/s] 72%\|███████▏ \| 2130/2975 [25:39<08:28, 1.66it/s] 72%\|███████▏ \| 2130/2975 [25:39<08:28, 1.66it/s] 72%\|███████▏ \| 2131/2975 [25:40<08:28, 1.66it/s] 72%\|███████▏ \| 2132/2975 [25:
	0: {'loss': 0.7142, 'grad_norm': 0.7238767511429106, 'learning_rate': 1e-05, 'epoch': 0.72}
	0: 40<08:28, 1.66it/s] 72%\|███████▏ \| 2133/2975 [25:41<08:27, 1.66it/s] 72%\|███████▏ \| 2134/2975 [25:42<08:26, 1.66it/s] 72%\|███████▏ \| 2135/2975 [25:42<08:25, 1.66it/s] 72%\|███████▏ \| 2136/2975 [25:43<08:25, 1.66it/s] 72%\|███████▏ \| 2137/2975 [25:43<08:25, 1.66it/s] 72%\|███████▏ \| 2138/2975 [25:44<08:24, 1.66it/s] 72%\|███████▏ \| 2139/2975 [25:45<08:24, 1.66it/s] 72%\|███████▏ \| 2140/2975 [25:45<08:24, 1.66it/s] 72%\|███████▏ \| 2140/2975 [25:45<08:24, 1.66it/s] 72%\|███████▏ \| 2141/2975 [25:46<08:23, 1.66it/s] 72%\|███████▏ \| 2142/2975 [25:46<08:23, 1.65it/s] 72%\|███████▏ \| 2143/2975 [25:47<08:25, 1.65it/s] 72%\|███████▏ \| 2144/2975 [25:48<08:24, 1.65it/s] 72%\|███████▏ \| 2145/2975 [25:48<08:23, 1.65it/s
	0: {'loss': 0.7035, 'grad_norm': 0.7743032216599863, 'learning_rate': 1e-05, 'epoch': 0.72}
	0: ] 72%\|███████▏ \| 2146/2975 [25:49<08:23, 1.65it/s] 72%\|███████▏ \| 2147/2975 [25:49<08:22, 1.65it/s] 72%\|███████▏ \| 2148/2975 [25:50<08:23, 1.64it/s] 72%\|███████▏ \| 2149/2975 [25:51<08:23, 1.64it/s] 72%\|███████▏ \| 2150/2975 [25:51<08:23, 1.64it/s] 72%\|███████▏ \| 2150/2975 [25:51<08:23, 1.64it/s] 72%\|███████▏ \| 2151/2975 [25:52<08:22, 1.64it/s] 72%\|███████▏ \| 2152/2975 [25:52<08:21, 1.64it/s] 72%\|███████▏ \| 2153/2975 [25:53<08:20, 1.64it/s] 72%\|███████▏ \| 2154/2975 [25:54<08:19, 1.65it/s] 72%\|███████▏ \| 2155/2975 [25:54<08:17, 1.65it/s] 72%\|███████▏ \| 2156/2975 [25:55<08:16, 1.65it/s] 73%\|███████▎ \| 2157/2975 [25:56<08:15, 1.65it/s] 73%\|███████▎ \| 2158/2975 [25:56<08:14, 1.65it/s] 73%\|████
	0: {'loss': 0.7082, 'grad_norm': 0.7386252426659861, 'learning_rate': 1e-05, 'epoch': 0.73}
	0: {'loss': 0.6979, 'grad_norm': 0.7193218822348061, 'learning_rate': 1e-05, 'epoch': 0.73}
	0: ███▎ \| 2159/2975 [25:57<08:13, 1.65it/s] 73%\|███████▎ \| 2160/2975 [25:57<08:12, 1.65it/s] 73%\|███████▎ \| 2160/2975 [25:57<08:12, 1.65it/s] 73%\|███████▎ \| 2161/2975 [25:58<08:11, 1.66it/s] 73%\|███████▎ \| 2162/2975 [25:59<08:10, 1.66it/s] 73%\|███████▎ \| 2163/2975 [25:59<08:10, 1.66it/s] 73%\|███████▎ \| 2164/2975 [26:00<08:09, 1.66it/s] 73%\|███████▎ \| 2165/2975 [26:00<08:10, 1.65it/s] 73%\|███████▎ \| 2166/2975 [26:01<08:09, 1.65it/s] 73%\|███████▎ \| 2167/2975 [26:02<08:08, 1.65it/s] 73%\|███████▎ \| 2168/2975 [26:02<08:08, 1.65it/s] 73%\|███████▎ \| 2169/2975 [26:03<08:08, 1.65it/s] 73%\|███████▎ \| 2170/2975 [26:03<08:09, 1.65it/s] 73%\|███████▎ \| 2170/2975 [26:03<0
	0: {'loss': 0.7047, 'grad_norm': 0.7075505087205055, 'learning_rate': 1e-05, 'epoch': 0.73}
	0: 8:09, 1.65it/s] 73%\|███████▎ \| 2171/2975 [26:04<08:09, 1.64it/s] 73%\|███████▎ \| 2172/2975 [26:05<08:08, 1.64it/s] 73%\|███████▎ \| 2173/2975 [26:05<08:08, 1.64it/s] 73%\|███████▎ \| 2174/2975 [26:06<08:08, 1.64it/s] 73%\|███████▎ \| 2175/2975 [26:06<08:07, 1.64it/s] 73%\|███████▎ \| 2176/2975 [26:07<08:05, 1.64it/s] 73%\|███████▎ \| 2177/2975 [26:08<08:04, 1.65it/s] 73%\|███████▎ \| 2178/2975 [26:08<08:03, 1.65it/s] 73%\|███████▎ \| 2179/2975 [26:09<08:02, 1.65it/s] 73%\|███████▎ \| 2180/2975 [26:09<08:02, 1.65it/s] 73%\|███████▎ \| 2180/2975 [26:09<08:02, 1.65it/s] 73%\|███████▎ \| 2181/2975 [26:10<08:01, 1.65it/s] 73%\|███████▎ \| 2182/2975 [26:11<08:00, 1.65it/s] 73%\|███████▎ \| 2183/2975 [26:11<07:59, 1.65it/s] 7
	0: {'loss': 0.716, 'grad_norm': 0.7458277455926593, 'learning_rate': 1e-05, 'epoch': 0.74}
	0: 3%\|███████▎ \| 2184/2975 [26:12<07:58, 1.65it/s] 73%\|███████▎ \| 2185/2975 [26:12<07:58, 1.65it/s] 73%\|███████▎ \| 2186/2975 [26:13<07:57, 1.65it/s] 74%\|███████▎ \| 2187/2975 [26:14<07:57, 1.65it/s] 74%\|███████▎ \| 2188/2975 [26:14<07:57, 1.65it/s] 74%\|███████▎ \| 2189/2975 [26:15<07:56, 1.65it/s] 74%\|███████▎ \| 2190/2975 [26:16<07:55, 1.65it/s] 74%\|███████▎ \| 2190/2975 [26:16<07:55, 1.65it/s] 74%\|███████▎ \| 2191/2975 [26:16<07:56, 1.65it/s] 74%\|███████▎ \| 2192/2975 [26:17<07:55, 1.65it/s] 74%\|███████▎ \| 2193/2975 [26:17<07:54, 1.65it/s] 74%\|███████▎ \| 2194/2975 [26:18<07:53, 1.65it/s] 74%\|███████▍ \| 2195/2975 [26:19<07:54, 1.64it/s] 74%\|███████▍ \| 2196/2975 [26:19<07:53, 1.65it/s] 74%\|█████�
	0: {'loss': 0.7161, 'grad_norm': 0.7587663874878008, 'learning_rate': 1e-05, 'epoch': 0.74}
	0: ��█▍ \| 2197/2975 [26:20<07:53, 1.64it/s] 74%\|███████▍ \| 2198/2975 [26:20<07:53, 1.64it/s] 74%\|███████▍ \| 2199/2975 [26:21<07:52, 1.64it/s] 74%\|███████▍ \| 2200/2975 [26:22<07:51, 1.64it/s] 74%\|███████▍ \| 2200/2975 [26:22<07:51, 1.64it/s] 74%\|███████▍ \| 2201/2975 [26:22<07:51, 1.64it/s] 74%\|███████▍ \| 2202/2975 [26:23<07:50, 1.64it/s] 74%\|███████▍ \| 2203/2975 [26:23<07:50, 1.64it/s] 74%\|███████▍ \| 2204/2975 [26:24<07:49, 1.64it/s] 74%\|███████▍ \| 2205/2975 [26:25<07:48, 1.64it/s] 74%\|███████▍ \| 2206/2975 [26:25<07:47, 1.64it/s] 74%\|███████▍ \| 2207/2975 [26:26<07:47, 1.64it/s] 74%\|███████▍ \| 2208/2975 [26:26<07:46, 1.64it/s] 74%\|███████▍ \| 2209/2975 [26:27<07:46, 1.64it/s] 74%\|███████▍ \| 2210/29
	0: {'loss': 0.7097, 'grad_norm': 0.7377475407946548, 'learning_rate': 1e-05, 'epoch': 0.74}
	0: {'loss': 0.7072, 'grad_norm': 0.77022569766666, 'learning_rate': 1e-05, 'epoch': 0.75}
	0: 75 [26:28<07:46, 1.64it/s] 74%\|███████▍ \| 2210/2975 [26:28<07:46, 1.64it/s] 74%\|███████▍ \| 2211/2975 [26:28<07:46, 1.64it/s] 74%\|███████▍ \| 2212/2975 [26:29<07:45, 1.64it/s] 74%\|███████▍ \| 2213/2975 [26:30<07:45, 1.64it/s] 74%\|███████▍ \| 2214/2975 [26:30<07:44, 1.64it/s] 74%\|███████▍ \| 2215/2975 [26:31<07:42, 1.64it/s] 74%\|███████▍ \| 2216/2975 [26:31<07:41, 1.65it/s] 75%\|███████▍ \| 2217/2975 [26:32<07:40, 1.65it/s] 75%\|███████▍ \| 2218/2975 [26:33<07:39, 1.65it/s] 75%\|███████▍ \| 2219/2975 [26:33<07:38, 1.65it/s] 75%\|███████▍ \| 2220/2975 [26:34<07:38, 1.65it/s] 75%\|███████▍ \| 2220/2975 [26:34<07:38, 1.65it/s] 75%\|███████▍ \| 2221/2975 [26:34<07:38, 1.65it/s] 75%\|�
	0: {'loss': 0.7296, 'grad_norm': 0.8032957399887488, 'learning_rate': 1e-05, 'epoch': 0.75}
	0: ��██████▍ \| 2222/2975 [26:35<07:36, 1.65it/s] 75%\|███████▍ \| 2223/2975 [26:36<07:36, 1.65it/s] 75%\|███████▍ \| 2224/2975 [26:36<07:35, 1.65it/s] 75%\|███████▍ \| 2225/2975 [26:37<07:41, 1.63it/s] 75%\|███████▍ \| 2226/2975 [26:37<07:40, 1.63it/s] 75%\|███████▍ \| 2227/2975 [26:38<07:37, 1.63it/s] 75%\|███████▍ \| 2228/2975 [26:39<07:36, 1.64it/s] 75%\|███████▍ \| 2229/2975 [26:39<07:35, 1.64it/s] 75%\|███████▍ \| 2230/2975 [26:40<07:34, 1.64it/s] 75%\|███████▍ \| 2230/2975 [26:40<07:34, 1.64it/s] 75%\|███████▍ \| 2231/2975 [26:40<07:33, 1.64it/s] 75%\|███████▌ \| 2232/2975 [26:41<07:31, 1.64it/s] 75%\|███████▌ \| 2233/2975 [26:42<07:31, 1.64it/s] 75%\|███████▌ \| 2234/2975 [26:42<07:29, 1.65it/s] 75%\|██████�
	0: {'loss': 0.7217, 'grad_norm': 0.8944500828928701, 'learning_rate': 1e-05, 'epoch': 0.75}
	0: �▌ \| 2235/2975 [26:43<07:28, 1.65it/s] 75%\|███████▌ \| 2236/2975 [26:44<07:27, 1.65it/s] 75%\|███████▌ \| 2237/2975 [26:44<07:27, 1.65it/s] 75%\|███████▌ \| 2238/2975 [26:45<07:26, 1.65it/s] 75%\|███████▌ \| 2239/2975 [26:45<07:26, 1.65it/s] 75%\|███████▌ \| 2240/2975 [26:46<07:26, 1.65it/s] 75%\|███████▌ \| 2240/2975 [26:46<07:26, 1.65it/s] 75%\|███████▌ \| 2241/2975 [26:47<07:27, 1.64it/s] 75%\|███████▌ \| 2242/2975 [26:47<07:26, 1.64it/s] 75%\|███████▌ \| 2243/2975 [26:48<07:26, 1.64it/s] 75%\|███████▌ \| 2244/2975 [26:48<07:30, 1.62it/s] 75%\|███████▌ \| 2245/2975 [26:49<07:27, 1.63it/s] 75%\|███████▌ \| 2246/2975 [26:50<07:25, 1.64it/s] 76%\|███████▌ \| 2247/2975 [26:50<07:24, 1.64it/s] 76%\|███████▌ \| 2248/2975 [
	0: {'loss': 0.7271, 'grad_norm': 0.7628192737409064, 'learning_rate': 1e-05, 'epoch': 0.76}
	0: {'loss': 0.7152, 'grad_norm': 0.7357763726593973, 'learning_rate': 1e-05, 'epoch': 0.76}
	0: 26:51<07:23, 1.64it/s] 76%\|███████▌ \| 2249/2975 [26:51<07:22, 1.64it/s] 76%\|███████▌ \| 2250/2975 [26:52<07:21, 1.64it/s] 76%\|███████▌ \| 2250/2975 [26:52<07:21, 1.64it/s] 76%\|███████▌ \| 2251/2975 [26:53<07:21, 1.64it/s] 76%\|███████▌ \| 2252/2975 [26:53<07:21, 1.64it/s] 76%\|███████▌ \| 2253/2975 [26:54<07:20, 1.64it/s] 76%\|███████▌ \| 2254/2975 [26:55<07:20, 1.64it/s] 76%\|███████▌ \| 2255/2975 [26:55<07:19, 1.64it/s] 76%\|███████▌ \| 2256/2975 [26:56<07:19, 1.64it/s] 76%\|███████▌ \| 2257/2975 [26:56<07:18, 1.64it/s] 76%\|███████▌ \| 2258/2975 [26:57<07:16, 1.64it/s] 76%\|███████▌ \| 2259/2975 [26:58<07:16, 1.64it/s] 76%\|███████▌ \| 2260/2975 [26:58<07:14, 1.65it/s] 76%\|█�
	0: {'loss': 0.6946, 'grad_norm': 0.7573853557180044, 'learning_rate': 1e-05, 'epoch': 0.76}
	0: �█████▌ \| 2260/2975 [26:58<07:14, 1.65it/s] 76%\|███████▌ \| 2261/2975 [26:59<07:13, 1.65it/s] 76%\|███████▌ \| 2262/2975 [26:59<07:13, 1.64it/s] 76%\|███████▌ \| 2263/2975 [27:00<07:13, 1.64it/s] 76%\|███████▌ \| 2264/2975 [27:01<07:13, 1.64it/s] 76%\|███████▌ \| 2265/2975 [27:01<07:19, 1.62it/s] 76%\|███████▌ \| 2266/2975 [27:02<07:16, 1.62it/s] 76%\|███████▌ \| 2267/2975 [27:02<07:14, 1.63it/s] 76%\|███████▌ \| 2268/2975 [27:03<07:12, 1.63it/s] 76%\|███████▋ \| 2269/2975 [27:04<07:11, 1.64it/s] 76%\|███████▋ \| 2270/2975 [27:04<07:09, 1.64it/s] 76%\|███████▋ \| 2270/2975 [27:04<07:09, 1.64it/s] 76%\|███████▋ \| 2271/2975 [27:05<07:08, 1.64it/s] 76%\|███████▋ \| 2272/2975 [27:05<07:07, 1.64it/s] 76%\|███████▋
	0: {'loss': 0.6989, 'grad_norm': 0.7584677946820176, 'learning_rate': 1e-05, 'epoch': 0.77}
	0: \| 2273/2975 [27:06<07:05, 1.65it/s] 76%\|███████▋ \| 2274/2975 [27:07<07:05, 1.65it/s] 76%\|███████▋ \| 2275/2975 [27:07<07:04, 1.65it/s] 77%\|███████▋ \| 2276/2975 [27:08<07:03, 1.65it/s] 77%\|███████▋ \| 2277/2975 [27:09<07:01, 1.65it/s] 77%\|███████▋ \| 2278/2975 [27:09<07:01, 1.66it/s] 77%\|███████▋ \| 2279/2975 [27:10<07:01, 1.65it/s] 77%\|███████▋ \| 2280/2975 [27:10<07:01, 1.65it/s] 77%\|███████▋ \| 2280/2975 [27:10<07:01, 1.65it/s] 77%\|███████▋ \| 2281/2975 [27:11<07:01, 1.65it/s] 77%\|███████▋ \| 2282/2975 [27:12<07:01, 1.65it/s] 77%\|███████▋ \| 2283/2975 [27:12<07:00, 1.64it/s] 77%\|███████▋ \| 2284/2975 [27:13<07:00, 1.64it/s] 77%\|███████▋ \| 2285/2975 [27:13<07:00, 1.64it/s] 77%\|███████▋ \| 2286/2975 [27:1
	0: {'loss': 0.6925, 'grad_norm': 0.7468414910667224, 'learning_rate': 1e-05, 'epoch': 0.77}
	0: 4<06:59, 1.64it/s] 77%\|███████▋ \| 2287/2975 [27:15<06:59, 1.64it/s] 77%\|███████▋ \| 2288/2975 [27:15<06:58, 1.64it/s] 77%\|███████▋ \| 2289/2975 [27:16<06:57, 1.64it/s] 77%\|███████▋ \| 2290/2975 [27:16<06:55, 1.65it/s] 77%\|███████▋ \| 2290/2975 [27:16<06:55, 1.65it/s] 77%\|███████▋ \| 2291/2975 [27:17<06:54, 1.65it/s] 77%\|███████▋ \| 2292/2975 [27:18<06:53, 1.65it/s] 77%\|███████▋ \| 2293/2975 [27:18<06:52, 1.65it/s] 77%\|███████▋ \| 2294/2975 [27:19<06:51, 1.65it/s] 77%\|███████▋ \| 2295/2975 [27:19<06:51, 1.65it/s] 77%\|███████▋ \| 2296/2975 [27:20<06:50, 1.66it/s] 77%\|███████▋ \| 2297/2975 [27:21<06:49, 1.66it/s] 77%\|███████▋ \| 2298/2975 [27:21<06:49, 1.65it/s] 77%\|███████▋ \| 2299/2975 [27:22<06:48, 1.65it/s]
	0: {'loss': 0.7231, 'grad_norm': 0.7731197620627839, 'learning_rate': 1e-05, 'epoch': 0.77}
	0: {'loss': 0.7266, 'grad_norm': 0.7237676476481548, 'learning_rate': 1e-05, 'epoch': 0.78}
	0: 77%\|███████▋ \| 2300/2975 [27:22<06:47, 1.66it/s] 77%\|███████▋ \| 2300/2975 [27:22<06:47, 1.66it/s] 77%\|███████▋ \| 2301/2975 [27:23<06:47, 1.66it/s] 77%\|███████▋ \| 2302/2975 [27:24<06:47, 1.65it/s] 77%\|███████▋ \| 2303/2975 [27:24<06:47, 1.65it/s] 77%\|███████▋ \| 2304/2975 [27:25<06:47, 1.65it/s] 77%\|███████▋ \| 2305/2975 [27:25<06:47, 1.64it/s] 78%\|███████▊ \| 2306/2975 [27:26<06:47, 1.64it/s] 78%\|███████▊ \| 2307/2975 [27:27<06:46, 1.64it/s] 78%\|███████▊ \| 2308/2975 [27:27<06:45, 1.64it/s] 78%\|███████▊ \| 2309/2975 [27:28<06:45, 1.64it/s] 78%\|███████▊ \| 2310/2975 [27:31<13:27, 1.21s/it] 78%\|███████▊ \| 2310/2975 [27:31<13:27, 1.21s/it] 78%\|███████▊ \|
	0: {'loss': 0.7102, 'grad_norm': 0.7563453931672067, 'learning_rate': 1e-05, 'epoch': 0.78}
	0: 2311/2975 [27:31<11:27, 1.04s/it] 78%\|███████▊ \| 2312/2975 [27:32<10:06, 1.09it/s] 78%\|███████▊ \| 2313/2975 [27:34<15:12, 1.38s/it] 78%\|███████▊ \| 2314/2975 [27:35<12:52, 1.17s/it] 78%\|███████▊ \| 2315/2975 [27:36<11:00, 1.00s/it] 78%\|███████▊ \| 2316/2975 [27:36<09:41, 1.13it/s] 78%\|███████▊ \| 2317/2975 [27:37<08:46, 1.25it/s] 78%\|███████▊ \| 2318/2975 [27:37<08:07, 1.35it/s] 78%\|███████▊ \| 2319/2975 [27:38<07:40, 1.42it/s] 78%\|███████▊ \| 2320/2975 [27:39<07:21, 1.49it/s] 78%\|███████▊ \| 2320/2975 [27:39<07:21, 1.49it/s] 78%\|███████▊ \| 2321/2975 [27:39<07:07, 1.53it/s] 78%\|███████▊ \| 2322/2975 [27:40<06:57, 1.56it/s] 78%\|███████▊ \| 2323/2975 [27:40<06:54, 1.57it/s] 78%\|███████▊ \| 2324/2975 [27:41<06
	0: {'loss': 0.6981, 'grad_norm': 0.7137144777380231, 'learning_rate': 1e-05, 'epoch': 0.78}
	0: :53, 1.57it/s] 78%\|███████▊ \| 2325/2975 [27:42<07:07, 1.52it/s] 78%\|███████▊ \| 2326/2975 [27:42<07:01, 1.54it/s] 78%\|███████▊ \| 2327/2975 [27:43<06:54, 1.56it/s] 78%\|███████▊ \| 2328/2975 [27:44<08:55, 1.21it/s] 78%\|███████▊ \| 2329/2975 [27:45<08:13, 1.31it/s] 78%\|███████▊ \| 2330/2975 [27:46<07:42, 1.39it/s] 78%\|███████▊ \| 2330/2975 [27:46<07:42, 1.39it/s] 78%\|███████▊ \| 2331/2975 [27:46<07:21, 1.46it/s] 78%\|███████▊ \| 2332/2975 [27:49<13:33, 1.26s/it] 78%\|███████▊ \| 2333/2975 [27:49<11:34, 1.08s/it] 78%\|███████▊ \| 2334/2975 [27:50<10:03, 1.06it/s] 78%\|███████▊ \| 2335/2975 [27:51<08:59, 1.19it/s] 79%\|███████▊ \| 2336/2975 [27:51<08:12, 1.30it/s] 79%\|███████▊ \| 2337/2975 [27:52<07:40, 1.38it/s] 79
	0: {'loss': 0.7058, 'grad_norm': 0.7535662693178167, 'learning_rate': 1e-05, 'epoch': 0.79}
	0: %\|███████▊ \| 2338/2975 [27:52<07:17, 1.46it/s] 79%\|███████▊ \| 2339/2975 [27:53<07:00, 1.51it/s] 79%\|███████▊ \| 2340/2975 [27:54<06:49, 1.55it/s] 79%\|███████▊ \| 2340/2975 [27:54<06:49, 1.55it/s] 79%\|███████▊ \| 2341/2975 [27:54<06:40, 1.58it/s] 79%\|███████▊ \| 2342/2975 [27:55<06:34, 1.60it/s] 79%\|███████▉ \| 2343/2975 [27:55<06:29, 1.62it/s] 79%\|███████▉ \| 2344/2975 [27:56<06:26, 1.63it/s] 79%\|███████▉ \| 2345/2975 [27:57<06:23, 1.64it/s] 79%\|███████▉ \| 2346/2975 [27:57<06:21, 1.65it/s] 79%\|███████▉ \| 2347/2975 [27:58<06:20, 1.65it/s] 79%\|███████▉ \| 2348/2975 [27:58<06:19, 1.65it/s] 79%\|███████▉ \| 2349/2975 [27:59<06:18, 1.65it/s] 79%\|███████▉ \| 2350/2975 [28:00<06:17, 1.65it/s]
	0: {'loss': 0.7054, 'grad_norm': 0.7380719530155119, 'learning_rate': 1e-05, 'epoch': 0.79}
	0: {'loss': 0.7059, 'grad_norm': 0.7417744754153898, 'learning_rate': 1e-05, 'epoch': 0.79}
	0: 79%\|███████▉ \| 2350/2975 [28:00<06:17, 1.65it/s] 79%\|███████▉ \| 2351/2975 [28:00<06:17, 1.65it/s] 79%\|███████▉ \| 2352/2975 [28:01<06:16, 1.66it/s] 79%\|███████▉ \| 2353/2975 [28:01<06:15, 1.66it/s] 79%\|███████▉ \| 2354/2975 [28:02<06:14, 1.66it/s] 79%\|███████▉ \| 2355/2975 [28:03<06:13, 1.66it/s] 79%\|███████▉ \| 2356/2975 [28:03<06:12, 1.66it/s] 79%\|███████▉ \| 2357/2975 [28:04<06:11, 1.66it/s] 79%\|███████▉ \| 2358/2975 [28:04<06:11, 1.66it/s] 79%\|███████▉ \| 2359/2975 [28:05<06:11, 1.66it/s] 79%\|███████▉ \| 2360/2975 [28:06<06:10, 1.66it/s] 79%\|███████▉ \| 2360/2975 [28:06<06:10, 1.66it/s] 79%\|███████▉ \| 2361/2975 [28:06<06:10, 1.66it/s] 79%\|███████▉ \| 2362/2975 [28:07<06:09,
	0: {'loss': 0.693, 'grad_norm': 0.7211124714468495, 'learning_rate': 1e-05, 'epoch': 0.8}
	0: 1.66it/s] 79%\|███████▉ \| 2363/2975 [28:08<06:09, 1.66it/s] 79%\|███████▉ \| 2364/2975 [28:08<06:08, 1.66it/s] 79%\|███████▉ \| 2365/2975 [28:09<06:07, 1.66it/s] 80%\|███████▉ \| 2366/2975 [28:09<06:07, 1.66it/s] 80%\|███████▉ \| 2367/2975 [28:10<06:06, 1.66it/s] 80%\|███████▉ \| 2368/2975 [28:11<06:05, 1.66it/s] 80%\|███████▉ \| 2369/2975 [28:11<06:05, 1.66it/s] 80%\|███████▉ \| 2370/2975 [28:12<06:04, 1.66it/s] 80%\|███████▉ \| 2370/2975 [28:12<06:04, 1.66it/s] 80%\|███████▉ \| 2371/2975 [28:12<06:04, 1.66it/s] 80%\|███████▉ \| 2372/2975 [28:13<06:03, 1.66it/s] 80%\|███████▉ \| 2373/2975 [28:14<06:02, 1.66it/s] 80%\|███████▉ \| 2374/2975 [28:14<06:02, 1.66it/s] 80%\|███████▉ \| 2375/2975 [28:15<06:01, 1.66it/s] 80%\|�
	0: {'loss': 0.724, 'grad_norm': 0.770275334209149, 'learning_rate': 1e-05, 'epoch': 0.8}
	0: �██████▉ \| 2376/2975 [28:15<06:00, 1.66it/s] 80%\|███████▉ \| 2377/2975 [28:16<05:59, 1.66it/s] 80%\|███████▉ \| 2378/2975 [28:17<05:59, 1.66it/s] 80%\|███████▉ \| 2379/2975 [28:17<05:58, 1.66it/s] 80%\|████████ \| 2380/2975 [28:18<05:58, 1.66it/s] 80%\|████████ \| 2380/2975 [28:18<05:58, 1.66it/s] 80%\|████████ \| 2381/2975 [28:20<11:43, 1.18s/it] 80%\|████████ \| 2382/2975 [28:21<09:59, 1.01s/it] 80%\|████████ \| 2383/2975 [28:22<08:47, 1.12it/s] 80%\|████████ \| 2384/2975 [28:22<07:56, 1.24it/s] 80%\|████████ \| 2385/2975 [28:23<07:21, 1.34it/s] 80%\|████████ \| 2386/2975 [28:23<06:56, 1.41it/s] 80%\|████████ \| 2387/2975 [28:24<06:38, 1.48it/s] 80%\|████████ \| 2388/2975 [28:25<06:24, 1.53it/s] 80%\|███████
	0: {'loss': 0.6791, 'grad_norm': 0.7226620658914946, 'learning_rate': 1e-05, 'epoch': 0.8}
	0: {'loss': 0.7061, 'grad_norm': 0.8579084355726249, 'learning_rate': 1e-05, 'epoch': 0.81}
	0: █ \| 2389/2975 [28:25<06:15, 1.56it/s] 80%\|████████ \| 2390/2975 [28:26<06:08, 1.59it/s] 80%\|████████ \| 2390/2975 [28:26<06:08, 1.59it/s] 80%\|████████ \| 2391/2975 [28:26<06:03, 1.61it/s] 80%\|████████ \| 2392/2975 [28:27<06:01, 1.61it/s] 80%\|████████ \| 2393/2975 [28:28<05:57, 1.63it/s] 80%\|████████ \| 2394/2975 [28:28<05:54, 1.64it/s] 81%\|████████ \| 2395/2975 [28:29<05:52, 1.65it/s] 81%\|████████ \| 2396/2975 [28:29<05:51, 1.65it/s] 81%\|████████ \| 2397/2975 [28:30<05:50, 1.65it/s] 81%\|████████ \| 2398/2975 [28:31<05:49, 1.65it/s] 81%\|████████ \| 2399/2975 [28:31<05:48, 1.65it/s] 81%\|████████ \| 2400/2975 [28:32<05:48, 1.65it/s] 81%\|████████ \| 2400/2975 [28:32<05:48, 1.
	0: {'loss': 0.6831, 'grad_norm': 0.7141326078218623, 'learning_rate': 1e-05, 'epoch': 0.81}
	0: 65it/s] 81%\|████████ \| 2401/2975 [28:32<05:47, 1.65it/s] 81%\|████████ \| 2402/2975 [28:33<05:46, 1.65it/s] 81%\|████████ \| 2403/2975 [28:34<05:45, 1.65it/s] 81%\|████████ \| 2404/2975 [28:34<05:45, 1.65it/s] 81%\|████████ \| 2405/2975 [28:35<05:44, 1.65it/s] 81%\|████████ \| 2406/2975 [28:35<05:44, 1.65it/s] 81%\|████████ \| 2407/2975 [28:36<05:44, 1.65it/s] 81%\|████████ \| 2408/2975 [28:37<05:43, 1.65it/s] 81%\|████████ \| 2409/2975 [28:37<05:43, 1.65it/s] 81%\|████████ \| 2410/2975 [28:38<05:48, 1.62it/s] 81%\|████████ \| 2410/2975 [28:38<05:48, 1.62it/s] 81%\|████████ \| 2411/2975 [28:39<05:46, 1.63it/s] 81%\|████████ \| 2412/2975 [28:39<05:44, 1.64it/s] 81%\|████████ \| 2413/2975 [28:40<05:43, 1.64it/s] 81%\|██
	0: {'loss': 0.7209, 'grad_norm': 0.7866118483290073, 'learning_rate': 1e-05, 'epoch': 0.81}
	0: ██████ \| 2414/2975 [28:40<05:41, 1.64it/s] 81%\|████████ \| 2415/2975 [28:41<05:40, 1.65it/s] 81%\|████████ \| 2416/2975 [28:42<05:38, 1.65it/s] 81%\|████████ \| 2417/2975 [28:42<05:37, 1.65it/s] 81%\|████████▏ \| 2418/2975 [28:43<05:36, 1.65it/s] 81%\|████████▏ \| 2419/2975 [28:43<05:35, 1.66it/s] 81%\|████████▏ \| 2420/2975 [28:44<05:36, 1.65it/s] 81%\|████████▏ \| 2420/2975 [28:44<05:36, 1.65it/s] 81%\|████████▏ \| 2421/2975 [28:45<05:36, 1.65it/s] 81%\|████████▏ \| 2422/2975 [28:45<05:35, 1.65it/s] 81%\|████████▏ \| 2423/2975 [28:46<05:33, 1.65it/s] 81%\|████████▏ \| 2424/2975 [28:46<05:33, 1.65it/s] 82%\|████████▏ \| 2425/2975 [28:47<05:32, 1.65it/s] 82%\|████████▏ \| 2426/2975 [28:48<05:32, 1.65it/s] 82%\|█�
	0: {'loss': 0.699, 'grad_norm': 0.7373297652878902, 'learning_rate': 1e-05, 'epoch': 0.82}
	0: �██████▏ \| 2427/2975 [28:48<05:31, 1.65it/s] 82%\|████████▏ \| 2428/2975 [28:49<05:30, 1.65it/s] 82%\|████████▏ \| 2429/2975 [28:49<05:30, 1.65it/s] 82%\|████████▏ \| 2430/2975 [28:50<05:30, 1.65it/s] 82%\|████████▏ \| 2430/2975 [28:50<05:30, 1.65it/s] 82%\|████████▏ \| 2431/2975 [28:51<05:30, 1.65it/s] 82%\|████████▏ \| 2432/2975 [28:51<05:29, 1.65it/s] 82%\|████████▏ \| 2433/2975 [28:52<05:29, 1.65it/s] 82%\|████████▏ \| 2434/2975 [28:52<05:28, 1.65it/s] 82%\|████████▏ \| 2435/2975 [28:53<05:27, 1.65it/s] 82%\|████████▏ \| 2436/2975 [28:54<05:26, 1.65it/s] 82%\|████████▏ \| 2437/2975 [28:54<05:26, 1.65it/s] 82%\|████████▏ \| 2438/2975 [28:55<05:25, 1.65it/s] 82%\|████████▏ \| 2439/2975 [28:55<05:25, 1.65it/s]
	0: {'loss': 0.7002, 'grad_norm': 0.7990406058141861, 'learning_rate': 1e-05, 'epoch': 0.82}
	0: {'loss': 0.7136, 'grad_norm': 0.7467991232331265, 'learning_rate': 1e-05, 'epoch': 0.82}
	0: 82%\|████████▏ \| 2440/2975 [28:56<05:24, 1.65it/s] 82%\|████████▏ \| 2440/2975 [28:56<05:24, 1.65it/s] 82%\|████████▏ \| 2441/2975 [28:57<05:59, 1.48it/s] 82%\|████████▏ \| 2442/2975 [28:58<05:48, 1.53it/s] 82%\|████████▏ \| 2443/2975 [28:58<05:39, 1.57it/s] 82%\|████████▏ \| 2444/2975 [28:59<05:32, 1.60it/s] 82%\|████████▏ \| 2445/2975 [28:59<05:28, 1.61it/s] 82%\|████████▏ \| 2446/2975 [29:00<05:25, 1.63it/s] 82%\|████████▏ \| 2447/2975 [29:01<05:22, 1.64it/s] 82%\|████████▏ \| 2448/2975 [29:01<05:20, 1.64it/s] 82%\|████████▏ \| 2449/2975 [29:02<05:19, 1.65it/s] 82%\|████████▏ \| 2450/2975 [29:02<05:18, 1.65it/s] 82%\|████████▏ \| 2450/2975 [29:02<05:18, 1.65it/s] 82%\|█�
	0: {'loss': 0.7189, 'grad_norm': 0.7557954002797767, 'learning_rate': 1e-05, 'epoch': 0.83}
	0: ��██████▏ \| 2451/2975 [29:03<05:17, 1.65it/s] 82%\|████████▏ \| 2452/2975 [29:04<05:16, 1.65it/s] 82%\|████████▏ \| 2453/2975 [29:04<05:15, 1.66it/s] 82%\|████████▏ \| 2454/2975 [29:05<05:14, 1.66it/s] 83%\|████████▎ \| 2455/2975 [29:05<05:13, 1.66it/s] 83%\|████████▎ \| 2456/2975 [29:06<05:12, 1.66it/s] 83%\|████████▎ \| 2457/2975 [29:07<05:12, 1.66it/s] 83%\|████████▎ \| 2458/2975 [29:07<05:11, 1.66it/s] 83%\|████████▎ \| 2459/2975 [29:08<05:11, 1.66it/s] 83%\|████████▎ \| 2460/2975 [29:08<05:10, 1.66it/s] 83%\|████████▎ \| 2460/2975 [29:08<05:10, 1.66it/s] 83%\|████████▎ \| 2461/2975 [29:09<05:10, 1.66it/s] 83%\|████████▎ \| 2462/2975 [29:10<05:09, 1.66it/s] 83%\|████████▎ \| 2463/2975 [29:10<05:08, 1.66it/s]
	0: {'loss': 0.7223, 'grad_norm': 0.7298360516103363, 'learning_rate': 1e-05, 'epoch': 0.83}
	0: 83%\|████████▎ \| 2464/2975 [29:11<05:07, 1.66it/s] 83%\|████████▎ \| 2465/2975 [29:11<05:07, 1.66it/s] 83%\|████████▎ \| 2466/2975 [29:12<05:06, 1.66it/s] 83%\|████████▎ \| 2467/2975 [29:13<05:05, 1.66it/s] 83%\|████████▎ \| 2468/2975 [29:13<05:05, 1.66it/s] 83%\|████████▎ \| 2469/2975 [29:14<05:04, 1.66it/s] 83%\|████████▎ \| 2470/2975 [29:14<05:04, 1.66it/s] 83%\|████████▎ \| 2470/2975 [29:14<05:04, 1.66it/s] 83%\|████████▎ \| 2471/2975 [29:15<05:03, 1.66it/s] 83%\|████████▎ \| 2472/2975 [29:16<05:03, 1.66it/s] 83%\|████████▎ \| 2473/2975 [29:16<05:03, 1.66it/s] 83%\|████████▎ \| 2474/2975 [29:17<05:02, 1.66it/s] 83%\|████████▎ \| 2475/2975 [29:17<05:01, 1.66it/s] 83%\|████████▎ \| 2476/2975 [29:18<05:00, 1
	0: {'loss': 0.674, 'grad_norm': 0.7417790471996113, 'learning_rate': 1e-05, 'epoch': 0.83}
	0: .66it/s] 83%\|████████▎ \| 2477/2975 [29:19<04:59, 1.66it/s] 83%\|████████▎ \| 2478/2975 [29:19<04:59, 1.66it/s] 83%\|████████▎ \| 2479/2975 [29:20<04:58, 1.66it/s] 83%\|████████▎ \| 2480/2975 [29:20<04:57, 1.66it/s] 83%\|████████▎ \| 2480/2975 [29:20<04:57, 1.66it/s] 83%\|████████▎ \| 2481/2975 [29:21<04:57, 1.66it/s] 83%\|████████▎ \| 2482/2975 [29:22<04:56, 1.66it/s] 83%\|████████▎ \| 2483/2975 [29:22<04:55, 1.66it/s] 83%\|████████▎ \| 2484/2975 [29:23<04:54, 1.66it/s] 84%\|████████▎ \| 2485/2975 [29:23<04:54, 1.66it/s] 84%\|████████▎ \| 2486/2975 [29:24<04:54, 1.66it/s] 84%\|████████▎ \| 2487/2975 [29:25<05:22, 1.51it/s] 84%\|████████▎ \| 2488/2975 [29:25<05:13, 1.55it/s] 84%\|████████▎ \| 2489/2975 [29:26<
	0: {'loss': 0.7084, 'grad_norm': 0.7594059959897228, 'learning_rate': 1e-05, 'epoch': 0.84}
	0: {'loss': 0.7075, 'grad_norm': 0.7080358402783196, 'learning_rate': 1e-05, 'epoch': 0.84}
	0: 05:06, 1.59it/s] 84%\|████████▎ \| 2490/2975 [29:27<05:01, 1.61it/s] 84%\|████████▎ \| 2490/2975 [29:27<05:01, 1.61it/s] 84%\|████████▎ \| 2491/2975 [29:27<04:58, 1.62it/s] 84%\|████████▍ \| 2492/2975 [29:28<04:55, 1.63it/s] 84%\|████████▍ \| 2493/2975 [29:28<04:53, 1.64it/s] 84%\|████████▍ \| 2494/2975 [29:29<04:58, 1.61it/s] 84%\|████████▍ \| 2495/2975 [29:30<04:56, 1.62it/s] 84%\|████████▍ \| 2496/2975 [29:30<04:56, 1.62it/s] 84%\|████████▍ \| 2497/2975 [29:31<04:54, 1.62it/s] 84%\|████████▍ \| 2498/2975 [29:32<04:52, 1.63it/s] 84%\|████████▍ \| 2499/2975 [29:32<04:51, 1.63it/s] 84%\|████████▍ \| 2500/2975 [29:33<04:49, 1.64it/s] 84%\|████████▍ \| 2500/2975 [29:33<04:49,
	0: {'loss': 0.6857, 'grad_norm': 0.7324494757206906, 'learning_rate': 1e-05, 'epoch': 0.84}
	0: 1.64it/s] 84%\|████████▍ \| 2501/2975 [29:33<04:51, 1.63it/s] 84%\|████████▍ \| 2502/2975 [29:34<04:48, 1.64it/s] 84%\|████████▍ \| 2503/2975 [29:35<04:46, 1.65it/s] 84%\|████████▍ \| 2504/2975 [29:35<04:44, 1.66it/s] 84%\|████████▍ \| 2505/2975 [29:36<04:43, 1.66it/s] 84%\|████████▍ \| 2506/2975 [29:36<04:42, 1.66it/s] 84%\|████████▍ \| 2507/2975 [29:37<04:41, 1.66it/s] 84%\|████████▍ \| 2508/2975 [29:38<04:41, 1.66it/s] 84%\|████████▍ \| 2509/2975 [29:38<04:40, 1.66it/s] 84%\|████████▍ \| 2510/2975 [29:39<04:39, 1.66it/s] 84%\|████████▍ \| 2510/2975 [29:39<04:39, 1.66it/s] 84%\|████████▍ \| 2511/2975 [29:39<04:39, 1.66it/s] 84%\|████████▍ \| 2512/2975 [29:40<04:38, 1.66it/s] 84%\|████████▍ \| 2513/2975 [29:41
	0: {'loss': 0.6943, 'grad_norm': 0.7764822865310044, 'learning_rate': 1e-05, 'epoch': 0.85}
	0: <04:38, 1.66it/s] 85%\|████████▍ \| 2514/2975 [29:41<04:37, 1.66it/s] 85%\|████████▍ \| 2515/2975 [29:42<04:36, 1.66it/s] 85%\|████████▍ \| 2516/2975 [29:42<04:36, 1.66it/s] 85%\|████████▍ \| 2517/2975 [29:43<04:35, 1.66it/s] 85%\|████████▍ \| 2518/2975 [29:44<04:35, 1.66it/s] 85%\|████████▍ \| 2519/2975 [29:44<04:34, 1.66it/s] 85%\|████████▍ \| 2520/2975 [29:45<04:33, 1.66it/s] 85%\|████████▍ \| 2520/2975 [29:45<04:33, 1.66it/s] 85%\|████████▍ \| 2521/2975 [29:45<04:33, 1.66it/s] 85%\|████████▍ \| 2522/2975 [29:46<04:32, 1.66it/s] 85%\|████████▍ \| 2523/2975 [29:47<04:32, 1.66it/s] 85%\|████████▍ \| 2524/2975 [29:47<04:31, 1.66it/s] 85%\|████████▍ \| 2525/2975 [29:48<04:31, 1.66it/s] 85%\|████████▍ \| 2526/29
	0: {'loss': 0.6938, 'grad_norm': 0.7442177955212868, 'learning_rate': 1e-05, 'epoch': 0.85}
	0: 75 [29:48<04:30, 1.66it/s] 85%\|████████▍ \| 2527/2975 [29:49<04:29, 1.66it/s] 85%\|████████▍ \| 2528/2975 [29:50<04:29, 1.66it/s] 85%\|████████▌ \| 2529/2975 [29:50<04:28, 1.66it/s] 85%\|████████▌ \| 2530/2975 [29:51<04:28, 1.66it/s] 85%\|████████▌ \| 2530/2975 [29:51<04:28, 1.66it/s] 85%\|████████▌ \| 2531/2975 [29:51<04:27, 1.66it/s] 85%\|████████▌ \| 2532/2975 [29:52<04:26, 1.66it/s] 85%\|████████▌ \| 2533/2975 [29:53<04:26, 1.66it/s] 85%\|████████▌ \| 2534/2975 [29:53<04:25, 1.66it/s] 85%\|████████▌ \| 2535/2975 [29:54<04:25, 1.66it/s] 85%\|████████▌ \| 2536/2975 [29:54<04:24, 1.66it/s] 85%\|████████▌ \| 2537/2975 [29:55<04:23, 1.66it/s] 85%\|████████▌ \| 2538/2975 [29:56<04:22, 1.66it/s] 85%\|████████▌
	0: {'loss': 0.6942, 'grad_norm': 0.7121672642088279, 'learning_rate': 1e-05, 'epoch': 0.85}
	0: {'loss': 0.6817, 'grad_norm': 0.7192337762381629, 'learning_rate': 1e-05, 'epoch': 0.86}
	0: \| 2539/2975 [29:56<04:22, 1.66it/s] 85%\|████████▌ \| 2540/2975 [29:57<04:22, 1.66it/s] 85%\|████████▌ \| 2540/2975 [29:57<04:22, 1.66it/s] 85%\|████████▌ \| 2541/2975 [29:57<04:21, 1.66it/s] 85%\|████████▌ \| 2542/2975 [29:58<04:21, 1.66it/s] 85%\|████████▌ \| 2543/2975 [29:59<04:20, 1.66it/s] 86%\|████████▌ \| 2544/2975 [29:59<04:20, 1.66it/s] 86%\|████████▌ \| 2545/2975 [30:00<04:19, 1.66it/s] 86%\|████████▌ \| 2546/2975 [30:00<04:18, 1.66it/s] 86%\|████████▌ \| 2547/2975 [30:01<04:17, 1.66it/s] 86%\|████████▌ \| 2548/2975 [30:02<04:20, 1.64it/s] 86%\|████████▌ \| 2549/2975 [30:02<04:19, 1.64it/s] 86%\|████████▌ \| 2550/2975 [30:03<04:17, 1.65it/s] 86%\|████████▌ \| 2550/2
	0: {'loss': 0.7133, 'grad_norm': 0.7485050272309117, 'learning_rate': 1e-05, 'epoch': 0.86}
	0: 975 [30:03<04:17, 1.65it/s] 86%\|████████▌ \| 2551/2975 [30:04<04:16, 1.65it/s] 86%\|████████▌ \| 2552/2975 [30:04<04:15, 1.66it/s] 86%\|████████▌ \| 2553/2975 [30:05<04:14, 1.66it/s] 86%\|████████▌ \| 2554/2975 [30:05<04:13, 1.66it/s] 86%\|████████▌ \| 2555/2975 [30:06<04:12, 1.66it/s] 86%\|████████▌ \| 2556/2975 [30:07<04:12, 1.66it/s] 86%\|████████▌ \| 2557/2975 [30:07<04:11, 1.66it/s] 86%\|████████▌ \| 2558/2975 [30:08<04:11, 1.66it/s] 86%\|████████▌ \| 2559/2975 [30:08<04:10, 1.66it/s] 86%\|████████▌ \| 2560/2975 [30:09<04:09, 1.66it/s] 86%\|████████▌ \| 2560/2975 [30:09<04:09, 1.66it/s] 86%\|████████▌ \| 2561/2975 [30:10<04:39, 1.48it/s] 86%\|████████▌ \| 2562/2975 [30:10<04:30, 1.52it/s] 86%\|████████▌
	0: {'loss': 0.699, 'grad_norm': 0.7334239029144919, 'learning_rate': 1e-05, 'epoch': 0.86}
	0: \| 2563/2975 [30:11<04:24, 1.56it/s] 86%\|████████▌ \| 2564/2975 [30:12<04:19, 1.59it/s] 86%\|████████▌ \| 2565/2975 [30:12<04:15, 1.60it/s] 86%\|████████▋ \| 2566/2975 [30:13<04:12, 1.62it/s] 86%\|████████▋ \| 2567/2975 [30:13<04:10, 1.63it/s] 86%\|████████▋ \| 2568/2975 [30:14<04:08, 1.64it/s] 86%\|████████▋ \| 2569/2975 [30:15<04:07, 1.64it/s] 86%\|████████▋ \| 2570/2975 [30:15<04:05, 1.65it/s] 86%\|████████▋ \| 2570/2975 [30:15<04:05, 1.65it/s] 86%\|████████▋ \| 2571/2975 [30:16<04:05, 1.65it/s] 86%\|████████▋ \| 2572/2975 [30:16<04:04, 1.65it/s] 86%\|████████▋ \| 2573/2975 [30:17<04:03, 1.65it/s] 87%\|████████▋ \| 2574/2975 [30:18<04:02, 1.65it/s] 87%\|████████▋ \| 2575/2975 [30:18<04:02, 1.65it/s] 87%\|██████
	0: {'loss': 0.7182, 'grad_norm': 0.76424968884289, 'learning_rate': 1e-05, 'epoch': 0.87}
	0: ██▋ \| 2576/2975 [30:19<04:00, 1.66it/s] 87%\|████████▋ \| 2577/2975 [30:19<04:00, 1.66it/s] 87%\|████████▋ \| 2578/2975 [30:20<03:59, 1.66it/s] 87%\|████████▋ \| 2579/2975 [30:21<03:58, 1.66it/s] 87%\|████████▋ \| 2580/2975 [30:21<03:58, 1.66it/s] 87%\|████████▋ \| 2580/2975 [30:21<03:58, 1.66it/s] 87%\|████████▋ \| 2581/2975 [30:22<03:58, 1.65it/s] 87%\|████████▋ \| 2582/2975 [30:22<03:57, 1.65it/s] 87%\|████████▋ \| 2583/2975 [30:23<03:56, 1.66it/s] 87%\|████████▋ \| 2584/2975 [30:24<03:55, 1.66it/s] 87%\|████████▋ \| 2585/2975 [30:24<03:55, 1.66it/s] 87%\|████████▋ \| 2586/2975 [30:25<03:54, 1.66it/s] 87%\|████████▋ \| 2587/2975 [30:25<03:53, 1.66it/s] 87%\|████████▋ \| 2588/2975 [30:26<03:53, 1.66it/s] 87%\|███
	0: {'loss': 0.7038, 'grad_norm': 0.7340157897457464, 'learning_rate': 1e-05, 'epoch': 0.87}
	0: {'loss': 0.7181, 'grad_norm': 0.7260402815882088, 'learning_rate': 1e-05, 'epoch': 0.87}
	0: █████▋ \| 2589/2975 [30:27<03:52, 1.66it/s] 87%\|████████▋ \| 2590/2975 [30:27<03:52, 1.66it/s] 87%\|████████▋ \| 2590/2975 [30:27<03:52, 1.66it/s] 87%\|████████▋ \| 2591/2975 [30:28<03:51, 1.66it/s] 87%\|████████▋ \| 2592/2975 [30:29<03:50, 1.66it/s] 87%\|████████▋ \| 2593/2975 [30:29<03:50, 1.66it/s] 87%\|████████▋ \| 2594/2975 [30:30<03:49, 1.66it/s] 87%\|████████▋ \| 2595/2975 [30:30<03:49, 1.66it/s] 87%\|████████▋ \| 2596/2975 [30:31<03:48, 1.66it/s] 87%\|████████▋ \| 2597/2975 [30:32<03:48, 1.66it/s] 87%\|████████▋ \| 2598/2975 [30:32<03:47, 1.65it/s] 87%\|████████▋ \| 2599/2975 [30:33<03:47, 1.65it/s] 87%\|████████▋ \| 2600/2975 [30:33<03:47, 1.65it/s] 87%\|█████�
	0: {'loss': 0.6887, 'grad_norm': 0.7334851286396781, 'learning_rate': 1e-05, 'epoch': 0.88}
	0: �██▋ \| 2600/2975 [30:33<03:47, 1.65it/s] 87%\|████████▋ \| 2601/2975 [30:34<03:46, 1.65it/s] 87%\|████████▋ \| 2602/2975 [30:35<03:45, 1.65it/s] 87%\|████████▋ \| 2603/2975 [30:35<03:45, 1.65it/s] 88%\|████████▊ \| 2604/2975 [30:36<03:44, 1.65it/s] 88%\|████████▊ \| 2605/2975 [30:36<03:44, 1.65it/s] 88%\|████████▊ \| 2606/2975 [30:37<03:43, 1.65it/s] 88%\|████████▊ \| 2607/2975 [30:38<03:42, 1.65it/s] 88%\|████████▊ \| 2608/2975 [30:38<03:41, 1.66it/s] 88%\|████████▊ \| 2609/2975 [30:39<03:41, 1.66it/s] 88%\|████████▊ \| 2610/2975 [30:39<03:41, 1.65it/s] 88%\|████████▊ \| 2610/2975 [30:39<03:41, 1.65it/s] 88%\|████████▊ \| 2611/2975 [30:40<03:40, 1.65it/s] 88%\|████████▊ \| 2612/2975 [30:41<03:39, 1.65it/s] 88%\|██�
	0: {'loss': 0.6814, 'grad_norm': 0.726296902846085, 'learning_rate': 1e-05, 'epoch': 0.88}
	0: �█████▊ \| 2613/2975 [30:41<03:39, 1.65it/s] 88%\|████████▊ \| 2614/2975 [30:42<03:38, 1.65it/s] 88%\|████████▊ \| 2615/2975 [30:42<03:37, 1.65it/s] 88%\|████████▊ \| 2616/2975 [30:43<03:36, 1.65it/s] 88%\|████████▊ \| 2617/2975 [30:44<03:36, 1.66it/s] 88%\|████████▊ \| 2618/2975 [30:44<03:35, 1.66it/s] 88%\|████████▊ \| 2619/2975 [30:45<03:34, 1.66it/s] 88%\|████████▊ \| 2620/2975 [30:46<04:04, 1.45it/s] 88%\|████████▊ \| 2620/2975 [30:46<04:04, 1.45it/s] 88%\|████████▊ \| 2621/2975 [30:46<03:55, 1.50it/s] 88%\|████████▊ \| 2622/2975 [30:47<03:48, 1.54it/s] 88%\|████████▊ \| 2623/2975 [30:48<03:43, 1.57it/s] 88%\|████████▊ \| 2624/2975 [30:48<03:40, 1.59it/s] 88%\|████████▊ \| 2625/2975 [30:49<03:37, 1.61it/s] 88%
	0: {'loss': 0.709, 'grad_norm': 0.7449056931497603, 'learning_rate': 1e-05, 'epoch': 0.88}
	0: \|████████▊ \| 2626/2975 [30:49<03:35, 1.62it/s] 88%\|████████▊ \| 2627/2975 [30:50<03:33, 1.63it/s] 88%\|████████▊ \| 2628/2975 [30:51<03:32, 1.64it/s] 88%\|████████▊ \| 2629/2975 [30:51<03:31, 1.64it/s] 88%\|████████▊ \| 2630/2975 [30:52<03:29, 1.64it/s] 88%\|████████▊ \| 2630/2975 [30:52<03:29, 1.64it/s] 88%\|████████▊ \| 2631/2975 [30:52<03:29, 1.65it/s] 88%\|████████▊ \| 2632/2975 [30:53<03:27, 1.65it/s] 89%\|████████▊ \| 2633/2975 [30:54<03:27, 1.65it/s] 89%\|████████▊ \| 2634/2975 [30:54<03:26, 1.65it/s] 89%\|████████▊ \| 2635/2975 [30:55<03:25, 1.66it/s] 89%\|████████▊ \| 2636/2975 [30:55<03:25, 1.65it/s] 89%\|████████▊ \| 2637/2975 [30:56<03:24, 1.65it/s] 89%\|████████▊ \| 2638/2975 [30:57<03:23, 1.65i
	0: {'loss': 0.6999, 'grad_norm': 0.7202233468170463, 'learning_rate': 1e-05, 'epoch': 0.89}
	0: {'loss': 0.7101, 'grad_norm': 0.7424194875658202, 'learning_rate': 1e-05, 'epoch': 0.89}
	0: t/s] 89%\|████████▊ \| 2639/2975 [30:57<03:23, 1.65it/s] 89%\|████████▊ \| 2640/2975 [30:58<03:22, 1.65it/s] 89%\|████████▊ \| 2640/2975 [30:58<03:22, 1.65it/s] 89%\|████████▉ \| 2641/2975 [30:58<03:22, 1.65it/s] 89%\|████████▉ \| 2642/2975 [30:59<03:21, 1.65it/s] 89%\|████████▉ \| 2643/2975 [31:00<03:21, 1.65it/s] 89%\|████████▉ \| 2644/2975 [31:00<03:20, 1.65it/s] 89%\|████████▉ \| 2645/2975 [31:01<03:19, 1.65it/s] 89%\|████████▉ \| 2646/2975 [31:01<03:18, 1.66it/s] 89%\|████████▉ \| 2647/2975 [31:02<03:17, 1.66it/s] 89%\|████████▉ \| 2648/2975 [31:03<03:17, 1.66it/s] 89%\|████████▉ \| 2649/2975 [31:03<03:16, 1.66it/s] 89%\|████████▉ \| 2650/2975 [31:04<03:15, 1.66it/s] 89
	0: {'loss': 0.6882, 'grad_norm': 0.7092015768110003, 'learning_rate': 1e-05, 'epoch': 0.89}
	0: %\|████████▉ \| 2650/2975 [31:04<03:15, 1.66it/s] 89%\|████████▉ \| 2651/2975 [31:04<03:15, 1.66it/s] 89%\|████████▉ \| 2652/2975 [31:05<03:14, 1.66it/s] 89%\|████████▉ \| 2653/2975 [31:06<03:14, 1.66it/s] 89%\|████████▉ \| 2654/2975 [31:06<03:13, 1.66it/s] 89%\|████████▉ \| 2655/2975 [31:07<03:14, 1.64it/s] 89%\|████████▉ \| 2656/2975 [31:08<03:42, 1.43it/s] 89%\|████████▉ \| 2657/2975 [31:08<03:33, 1.49it/s] 89%\|████████▉ \| 2658/2975 [31:09<03:26, 1.54it/s] 89%\|████████▉ \| 2659/2975 [31:10<03:21, 1.57it/s] 89%\|████████▉ \| 2660/2975 [31:10<03:18, 1.59it/s] 89%\|████████▉ \| 2660/2975 [31:10<03:18, 1.59it/s] 89%\|████████▉ \| 2661/2975 [31:11<03:16, 1.60it/s] 89%\|████████▉ \| 2662/2975 [31:11<03:14, 1.61
	0: {'loss': 0.6977, 'grad_norm': 0.7068212020146698, 'learning_rate': 1e-05, 'epoch': 0.9}
	0: it/s] 90%\|████████▉ \| 2663/2975 [31:12<03:12, 1.62it/s] 90%\|████████▉ \| 2664/2975 [31:13<03:11, 1.63it/s] 90%\|████████▉ \| 2665/2975 [31:13<03:09, 1.63it/s] 90%\|████████▉ \| 2666/2975 [31:14<03:08, 1.64it/s] 90%\|████████▉ \| 2667/2975 [31:15<03:07, 1.65it/s] 90%\|████████▉ \| 2668/2975 [31:15<03:29, 1.47it/s] 90%\|████████▉ \| 2669/2975 [31:16<03:22, 1.51it/s] 90%\|████████▉ \| 2670/2975 [31:17<03:17, 1.55it/s] 90%\|████████▉ \| 2670/2975 [31:17<03:17, 1.55it/s] 90%\|████████▉ \| 2671/2975 [31:17<03:13, 1.57it/s] 90%\|████████▉ \| 2672/2975 [31:18<03:10, 1.59it/s] 90%\|████████▉ \| 2673/2975 [31:18<03:08, 1.61it/s] 90%\|████████▉ \| 2674/2975 [31:19<03:05, 1.62it/s] 90%\|████████▉ \| 2675/2975 [31:20<03:
	0: {'loss': 0.7009, 'grad_norm': 0.7066569735462942, 'learning_rate': 9.996052735444863e-06, 'epoch': 0.9}
	0: 03, 1.63it/s] 90%\|████████▉ \| 2676/2975 [31:20<03:02, 1.64it/s] 90%\|████████▉ \| 2677/2975 [31:21<03:01, 1.65it/s] 90%\|█████████ \| 2678/2975 [31:21<03:00, 1.65it/s] 90%\|█████████ \| 2679/2975 [31:22<02:58, 1.65it/s] 90%\|█████████ \| 2680/2975 [31:23<03:23, 1.45it/s] 90%\|█████████ \| 2680/2975 [31:23<03:23, 1.45it/s] 90%\|█████████ \| 2681/2975 [31:24<03:15, 1.50it/s] 90%\|█████████ \| 2682/2975 [31:24<03:09, 1.55it/s] 90%\|█████████ \| 2683/2975 [31:25<03:05, 1.58it/s] 90%\|█████████ \| 2684/2975 [31:25<03:01, 1.60it/s] 90%\|█████████ \| 2685/2975 [31:26<02:59, 1.62it/s] 90%\|█████████ \| 2686/2975 [31:27<02:57, 1.63it/s] 90%\|█████████ \| 2687/2975 [31:27<02:55, 1.64it/s] 90%\|█████████ \| 2688/2975 [
	0: {'loss': 0.707, 'grad_norm': 0.7598696013918776, 'learning_rate': 9.951725498333449e-06, 'epoch': 0.9}
	0: 31:28<02:54, 1.64it/s] 90%\|█████████ \| 2689/2975 [31:28<02:53, 1.65it/s] 90%\|█████████ \| 2690/2975 [31:29<02:52, 1.65it/s] 90%\|█████████ \| 2690/2975 [31:29<02:52, 1.65it/s] 90%\|█████████ \| 2691/2975 [31:30<02:52, 1.65it/s] 90%\|█████████ \| 2692/2975 [31:30<02:51, 1.65it/s] 91%\|█████████ \| 2693/2975 [31:31<02:50, 1.66it/s] 91%\|█████████ \| 2694/2975 [31:31<02:49, 1.66it/s] 91%\|█████████ \| 2695/2975 [31:32<02:48, 1.66it/s] 91%\|█████████ \| 2696/2975 [31:33<02:47, 1.66it/s] 91%\|█████████ \| 2697/2975 [31:33<02:47, 1.66it/s] 91%\|█████████ \| 2698/2975 [31:34<02:46, 1.66it/s] 91%\|█████████ \| 2699/2975 [31:34<02:46, 1.66it/s] 91%\|█████████ \| 2700/2975 [31:35<02:45, 1.66it/s]
	0: {'loss': 0.6899, 'grad_norm': 0.7517116214718325, 'learning_rate': 9.858624225078841e-06, 'epoch': 0.91}
	0: {'loss': 0.708, 'grad_norm': 0.7320599916190836, 'learning_rate': 9.717768952713514e-06, 'epoch': 0.91}
	0: 91%\|█████████ \| 2700/2975 [31:35<02:45, 1.66it/s] 91%\|█████████ \| 2701/2975 [31:36<02:44, 1.66it/s] 91%\|█████████ \| 2702/2975 [31:36<02:44, 1.66it/s] 91%\|█████████ \| 2703/2975 [31:37<02:44, 1.66it/s] 91%\|█████████ \| 2704/2975 [31:37<02:43, 1.65it/s] 91%\|█████████ \| 2705/2975 [31:38<02:42, 1.66it/s] 91%\|█████████ \| 2706/2975 [31:39<02:42, 1.66it/s] 91%\|█████████ \| 2707/2975 [31:39<02:41, 1.66it/s] 91%\|█████████ \| 2708/2975 [31:40<02:40, 1.66it/s] 91%\|█████████ \| 2709/2975 [31:40<02:40, 1.66it/s] 91%\|█████████ \| 2710/2975 [31:41<02:39, 1.66it/s] 91%\|█████████ \| 2710/2975 [31:41<02:39, 1.66it/s] 91%\|█████████ \| 2711/2975 [31:42<02:39, 1.66it/s] 91%\|█████████ \| 2712/2975
	0: {'loss': 0.696, 'grad_norm': 0.7651357941181524, 'learning_rate': 9.530702921077358e-06, 'epoch': 0.91}
	0: [31:42<02:38, 1.66it/s] 91%\|█████████ \| 2713/2975 [31:43<02:37, 1.66it/s] 91%\|█████████ \| 2714/2975 [31:43<02:37, 1.66it/s] 91%\|█████████▏\| 2715/2975 [31:44<02:59, 1.45it/s] 91%\|█████████▏\| 2716/2975 [31:45<02:52, 1.50it/s] 91%\|█████████▏\| 2717/2975 [31:46<02:46, 1.55it/s] 91%\|█████████▏\| 2718/2975 [31:46<02:42, 1.58it/s] 91%\|█████████▏\| 2719/2975 [31:47<02:39, 1.60it/s] 91%\|█████████▏\| 2720/2975 [31:47<02:37, 1.62it/s] 91%\|█████████▏\| 2720/2975 [31:47<02:37, 1.62it/s] 91%\|█████████▏\| 2721/2975 [31:48<02:35, 1.63it/s] 91%\|█████████▏\| 2722/2975 [31:49<02:34, 1.64it/s] 92%\|█████████▏\| 2723/2975 [31:49<02:32, 1.65it/s] 92%\|█████████▏\| 2724/2975 [31:50<02:32, 1.65it/s] 92%\|███
	0: {'loss': 0.7053, 'grad_norm': 0.7197188710960789, 'learning_rate': 9.29947566475907e-06, 'epoch': 0.92}
	0: ██████▏\| 2725/2975 [31:50<02:31, 1.65it/s] 92%\|█████████▏\| 2726/2975 [31:51<02:30, 1.66it/s] 92%\|█████████▏\| 2727/2975 [31:52<02:29, 1.66it/s] 92%\|█████████▏\| 2728/2975 [31:52<02:29, 1.66it/s] 92%\|█████████▏\| 2729/2975 [31:53<02:28, 1.66it/s] 92%\|█████████▏\| 2730/2975 [31:53<02:27, 1.66it/s] 92%\|█████████▏\| 2730/2975 [31:53<02:27, 1.66it/s] 92%\|█████████▏\| 2731/2975 [31:54<02:27, 1.66it/s] 92%\|█████████▏\| 2732/2975 [31:55<02:26, 1.66it/s] 92%\|█████████▏\| 2733/2975 [31:55<02:25, 1.66it/s] 92%\|█████████▏\| 2734/2975 [31:56<02:25, 1.66it/s] 92%\|█████████▏\| 2735/2975 [31:56<02:24, 1.66it/s] 92%\|█████████▏\| 2736/2975 [31:57<02:24, 1.66it/s] 92%\|█████████▏\| 2737/2975 [3
	0: {'loss': 0.6905, 'grad_norm': 0.7172619922394774, 'learning_rate': 9.02662055796628e-06, 'epoch': 0.92}
	0: 1:58<02:23, 1.66it/s] 92%\|█████████▏\| 2738/2975 [31:58<02:44, 1.44it/s] 92%\|█████████▏\| 2739/2975 [31:59<02:58, 1.32it/s] 92%\|█████████▏\| 2740/2975 [32:00<02:46, 1.41it/s] 92%\|█████████▏\| 2740/2975 [32:00<02:46, 1.41it/s] 92%\|█████████▏\| 2741/2975 [32:01<02:38, 1.47it/s] 92%\|█████████▏\| 2742/2975 [32:01<02:32, 1.53it/s] 92%\|█████████▏\| 2743/2975 [32:02<02:28, 1.56it/s] 92%\|█████████▏\| 2744/2975 [32:02<02:25, 1.59it/s] 92%\|█████████▏\| 2745/2975 [32:03<02:22, 1.61it/s] 92%\|█████████▏\| 2746/2975 [32:04<02:20, 1.63it/s] 92%\|█████████▏\| 2747/2975 [32:04<02:19, 1.64it/s] 92%\|█████████▏\| 2748/2975 [32:05<02:18, 1.64it/s] 92%\|█████████▏\| 2749/2975 [32:05<02:17, 1.64it/s] 92%\|██�
	0: {'loss': 0.6869, 'grad_norm': 0.7121173622621597, 'learning_rate': 8.715127058347615e-06, 'epoch': 0.92}
	0: {'loss': 0.7213, 'grad_norm': 0.7341579336149818, 'learning_rate': 8.368407953869105e-06, 'epoch': 0.93}
	0: ��██████▏\| 2750/2975 [32:06<02:16, 1.65it/s] 92%\|█████████▏\| 2750/2975 [32:06<02:16, 1.65it/s] 92%\|█████████▏\| 2751/2975 [32:07<02:15, 1.65it/s] 93%\|█████████▎\| 2752/2975 [32:07<02:15, 1.65it/s] 93%\|█████████▎\| 2753/2975 [32:08<02:14, 1.66it/s] 93%\|█████████▎\| 2754/2975 [32:08<02:13, 1.66it/s] 93%\|█████████▎\| 2755/2975 [32:09<02:12, 1.66it/s] 93%\|█████████▎\| 2756/2975 [32:10<02:32, 1.44it/s] 93%\|█████████▎\| 2757/2975 [32:11<02:25, 1.50it/s] 93%\|█████████▎\| 2758/2975 [32:11<02:20, 1.54it/s] 93%\|█████████▎\| 2759/2975 [32:12<02:17, 1.58it/s] 93%\|█████████▎\| 2760/2975 [32:12<02:14, 1.60it/s] 93%\|█████████▎\| 2760/2975 [32:12<02:14, 1.60
	0: {'loss': 0.7156, 'grad_norm': 0.7549621234139194, 'learning_rate': 7.99026197159505e-06, 'epoch': 0.93}
	0: it/s] 93%\|█████████▎\| 2761/2975 [32:13<02:12, 1.62it/s] 93%\|█████████▎\| 2762/2975 [32:14<02:10, 1.63it/s] 93%\|█████████▎\| 2763/2975 [32:14<02:09, 1.64it/s] 93%\|█████████▎\| 2764/2975 [32:15<02:07, 1.65it/s] 93%\|█████████▎\| 2765/2975 [32:15<02:06, 1.66it/s] 93%\|█████████▎\| 2766/2975 [32:16<02:06, 1.66it/s] 93%\|█████████▎\| 2767/2975 [32:17<02:05, 1.66it/s] 93%\|█████████▎\| 2768/2975 [32:17<02:04, 1.66it/s] 93%\|█████████▎\| 2769/2975 [32:18<02:04, 1.66it/s] 93%\|█████████▎\| 2770/2975 [32:18<02:03, 1.66it/s] 93%\|█████████▎\| 2770/2975 [32:18<02:03, 1.66it/s] 93%\|█████████▎\| 2771/2975 [32:19<02:02, 1.66it/s] 93%\|█████████▎\| 2772/2975 [32:20<02:02, 1.66it/s] 93%\|████████
	0: {'loss': 0.6948, 'grad_norm': 0.7215693093672365, 'learning_rate': 7.584832158039379e-06, 'epoch': 0.93}
	0: █▎\| 2773/2975 [32:20<02:01, 1.66it/s] 93%\|█████████▎\| 2774/2975 [32:21<02:01, 1.66it/s] 93%\|█████████▎\| 2775/2975 [32:21<02:00, 1.66it/s] 93%\|█████████▎\| 2776/2975 [32:22<01:59, 1.66it/s] 93%\|█████████▎\| 2777/2975 [32:23<01:59, 1.66it/s] 93%\|█████████▎\| 2778/2975 [32:23<01:58, 1.66it/s] 93%\|█████████▎\| 2779/2975 [32:24<01:57, 1.66it/s] 93%\|█████████▎\| 2780/2975 [32:24<01:57, 1.66it/s] 93%\|█████████▎\| 2780/2975 [32:24<01:57, 1.66it/s] 93%\|█████████▎\| 2781/2975 [32:25<01:56, 1.66it/s] 94%\|█████████▎\| 2782/2975 [32:26<01:56, 1.66it/s] 94%\|█████████▎\| 2783/2975 [32:26<01:55, 1.66it/s] 94%\|█████████▎\| 2784/2975 [32:27<01:55, 1.66it/s] 94%\|█████████▎\| 2785/2975 [32:27<01:54, 1.
	0: {'loss': 0.6965, 'grad_norm': 0.7326912491360482, 'learning_rate': 7.156560487081052e-06, 'epoch': 0.94}
	0: 66it/s] 94%\|█████████▎\| 2786/2975 [32:28<01:53, 1.66it/s] 94%\|█████████▎\| 2787/2975 [32:29<01:53, 1.66it/s] 94%\|█████████▎\| 2788/2975 [32:29<01:52, 1.66it/s] 94%\|█████████▎\| 2789/2975 [32:30<01:52, 1.66it/s] 94%\|█████████▍\| 2790/2975 [32:30<01:51, 1.66it/s] 94%\|█████████▍\| 2790/2975 [32:30<01:51, 1.66it/s] 94%\|█████████▍\| 2791/2975 [32:31<01:51, 1.66it/s] 94%\|█████████▍\| 2792/2975 [32:32<01:50, 1.66it/s] 94%\|█████████▍\| 2793/2975 [32:32<01:49, 1.66it/s] 94%\|█████████▍\| 2794/2975 [32:33<01:49, 1.66it/s] 94%\|█████████▍\| 2795/2975 [32:33<01:48, 1.66it/s] 94%\|█████████▍\| 2796/2975 [32:34<01:47, 1.66it/s] 94%\|█████████▍\| 2797/2975 [32:35<01:47, 1.66it/s] 94%\|███████�
	0: {'loss': 0.6906, 'grad_norm': 0.72139445955845, 'learning_rate': 6.710139192768695e-06, 'epoch': 0.94}
	0: ��█▍\| 2798/2975 [32:35<01:46, 1.66it/s] 94%\|█████████▍\| 2799/2975 [32:36<01:45, 1.66it/s] 94%\|█████████▍\| 2800/2975 [32:36<01:45, 1.66it/s] 94%\|█████████▍\| 2800/2975 [32:36<01:45, 1.66it/s] 94%\|█████████▍\| 2801/2975 [32:37<01:45, 1.66it/s] 94%\|█████████▍\| 2802/2975 [32:38<01:57, 1.47it/s] 94%\|█████████▍\| 2803/2975 [32:39<01:53, 1.52it/s] 94%\|█████████▍\| 2804/2975 [32:39<01:49, 1.56it/s] 94%\|█████████▍\| 2805/2975 [32:40<01:47, 1.58it/s] 94%\|█████████▍\| 2806/2975 [32:40<01:45, 1.60it/s] 94%\|█████████▍\| 2807/2975 [32:41<01:43, 1.62it/s] 94%\|█████████▍\| 2808/2975 [32:42<01:42, 1.63it/s] 94%\|█████████▍\| 2809/2975 [32:42<01:41, 1.63it/s] 94%\|█████████▍\| 2810/2975 [32:43<01:40,
	0: {'loss': 0.6742, 'grad_norm': 0.6937605745619208, 'learning_rate': 6.25045936022246e-06, 'epoch': 0.94}
	0: {'loss': 0.698, 'grad_norm': 0.7320316503097014, 'learning_rate': 5.782557337881911e-06, 'epoch': 0.95}
	0: 1.64it/s] 94%\|█████████▍\| 2810/2975 [32:43<01:40, 1.64it/s] 94%\|█████████▍\| 2811/2975 [32:43<01:39, 1.64it/s] 95%\|█████████▍\| 2812/2975 [32:44<01:39, 1.63it/s] 95%\|█████████▍\| 2813/2975 [32:45<01:38, 1.64it/s] 95%\|█████████▍\| 2814/2975 [32:45<01:37, 1.65it/s] 95%\|█████████▍\| 2815/2975 [32:46<01:36, 1.65it/s] 95%\|█████████▍\| 2816/2975 [32:46<01:36, 1.65it/s] 95%\|█████████▍\| 2817/2975 [32:47<01:35, 1.65it/s] 95%\|█████████▍\| 2818/2975 [32:48<01:34, 1.66it/s] 95%\|█████████▍\| 2819/2975 [32:48<01:34, 1.66it/s] 95%\|█████████▍\| 2820/2975 [32:49<01:33, 1.66it/s] 95%\|█████████▍\| 2820/2975 [32:49<01:33, 1.66it/s] 95%\|█████████▍\| 2821/29
	0: {'loss': 0.695, 'grad_norm': 0.7076182410530591, 'learning_rate': 5.311559558218603e-06, 'epoch': 0.95}
	0: 75 [32:49<01:33, 1.66it/s] 95%\|█████████▍\| 2822/2975 [32:50<01:32, 1.66it/s] 95%\|█████████▍\| 2823/2975 [32:51<01:31, 1.66it/s] 95%\|█████████▍\| 2824/2975 [32:51<01:30, 1.66it/s] 95%\|█████████▍\| 2825/2975 [32:52<01:30, 1.66it/s] 95%\|█████████▍\| 2826/2975 [32:52<01:29, 1.66it/s] 95%\|█████████▌\| 2827/2975 [32:53<01:29, 1.66it/s] 95%\|█████████▌\| 2828/2975 [32:54<01:28, 1.66it/s] 95%\|█████████▌\| 2829/2975 [32:54<01:27, 1.66it/s] 95%\|█████████▌\| 2830/2975 [32:55<01:27, 1.66it/s] 95%\|█████████▌\| 2830/2975 [32:55<01:27, 1.66it/s] 95%\|█████████▌\| 2831/2975 [32:55<01:26, 1.66it/s] 95%\|█████████▌\| 2832/2975 [32:56<01:26, 1.66it/s] 95%\|█████████▌\| 2833/2975 [32:57<01:25, 1.66it/s] 95%\|�
	0: {'loss': 0.6936, 'grad_norm': 0.7227746574659001, 'learning_rate': 4.842626371469149e-06, 'epoch': 0.95}
	0: �████████▌\| 2834/2975 [32:57<01:24, 1.66it/s] 95%\|█████████▌\| 2835/2975 [32:58<01:24, 1.66it/s] 95%\|█████████▌\| 2836/2975 [32:58<01:23, 1.66it/s] 95%\|█████████▌\| 2837/2975 [32:59<01:23, 1.66it/s] 95%\|█████████▌\| 2838/2975 [33:00<01:22, 1.66it/s] 95%\|█████████▌\| 2839/2975 [33:00<01:21, 1.66it/s] 95%\|█████████▌\| 2840/2975 [33:01<01:21, 1.66it/s] 95%\|█████████▌\| 2840/2975 [33:01<01:21, 1.66it/s] 95%\|█████████▌\| 2841/2975 [33:01<01:20, 1.66it/s] 96%\|█████████▌\| 2842/2975 [33:02<01:20, 1.66it/s] 96%\|█████████▌\| 2843/2975 [33:03<01:19, 1.66it/s] 96%\|█████████▌\| 2844/2975 [33:03<01:18, 1.66it/s] 96%\|█████████▌\| 2845/2975 [33:04<01:18, 1.66it/s] 96%\|█████████▌\| 2846/
	0: {'loss': 0.6863, 'grad_norm': 0.7091047859832311, 'learning_rate': 4.380895507758155e-06, 'epoch': 0.96}
	0: 2975 [33:04<01:17, 1.66it/s] 96%\|█████████▌\| 2847/2975 [33:05<01:16, 1.66it/s] 96%\|█████████▌\| 2848/2975 [33:06<01:16, 1.66it/s] 96%\|█████████▌\| 2849/2975 [33:06<01:15, 1.66it/s] 96%\|█████████▌\| 2850/2975 [33:07<01:15, 1.66it/s] 96%\|█████████▌\| 2850/2975 [33:07<01:15, 1.66it/s] 96%\|█████████▌\| 2851/2975 [33:07<01:14, 1.66it/s] 96%\|█████████▌\| 2852/2975 [33:08<01:14, 1.66it/s] 96%\|█████████▌\| 2853/2975 [33:09<01:13, 1.66it/s] 96%\|█████████▌\| 2854/2975 [33:09<01:12, 1.66it/s] 96%\|█████████▌\| 2855/2975 [33:10<01:12, 1.66it/s] 96%\|█████████▌\| 2856/2975 [33:10<01:11, 1.66it/s] 96%\|█████████▌\| 2857/2975 [33:11<01:10, 1.66it/s] 96%\|█████████▌\| 2858/2975 [33:12<01:10, 1.66it/s] 96%\|
	0: {'loss': 0.6856, 'grad_norm': 0.7247132123386185, 'learning_rate': 3.931425787051832e-06, 'epoch': 0.96}
	0: █████████▌\| 2859/2975 [33:12<01:09, 1.66it/s] 96%\|█████████▌\| 2860/2975 [33:13<01:09, 1.66it/s] 96%\|█████████▌\| 2860/2975 [33:13<01:09, 1.66it/s] 96%\|█████████▌\| 2861/2975 [33:13<01:08, 1.66it/s] 96%\|█████████▌\| 2862/2975 [33:14<01:08, 1.66it/s] 96%\|█████████▌\| 2863/2975 [33:15<01:07, 1.66it/s] 96%\|█████████▋\| 2864/2975 [33:15<01:06, 1.66it/s] 96%\|█████████▋\| 2865/2975 [33:16<01:06, 1.66it/s] 96%\|█████████▋\| 2866/2975 [33:17<01:05, 1.66it/s] 96%\|█████████▋\| 2867/2975 [33:17<01:05, 1.66it/s] 96%\|█████████▋\| 2868/2975 [33:18<01:04, 1.66it/s] 96%\|█████████▋\| 2869/2975 [33:18<01:03, 1.66it/s] 96%\|█████████▋\| 2870/2975 [33:19<01:03, 1.66it/s]
	0: {'loss': 0.7047, 'grad_norm': 0.6897697496973473, 'learning_rate': 3.499141693667828e-06, 'epoch': 0.96}
	0: {'loss': 0.6937, 'grad_norm': 0.7064837688525987, 'learning_rate': 3.0887794225945143e-06, 'epoch': 0.97}
	0: 96%\|█████████▋\| 2870/2975 [33:19<01:03, 1.66it/s] 97%\|█████████▋\| 2871/2975 [33:20<01:02, 1.66it/s] 97%\|█████████▋\| 2872/2975 [33:20<01:02, 1.66it/s] 97%\|█████████▋\| 2873/2975 [33:21<01:01, 1.66it/s] 97%\|█████████▋\| 2874/2975 [33:21<01:00, 1.66it/s] 97%\|█████████▋\| 2875/2975 [33:22<01:00, 1.66it/s] 97%\|█████████▋\| 2876/2975 [33:23<00:59, 1.66it/s] 97%\|█████████▋\| 2877/2975 [33:23<00:58, 1.66it/s] 97%\|█████████▋\| 2878/2975 [33:24<00:58, 1.66it/s] 97%\|█████████▋\| 2879/2975 [33:24<00:57, 1.66it/s] 97%\|█████████▋\| 2880/2975 [33:25<00:57, 1.66it/s] 97%\|█████████▋\| 2880/2975 [33:25<00:57, 1.66it/s] 97%\|█████████▋\| 2881/2975 [33:26<00:56, 1.66it/s] 97%\|█████�
	0: {'loss': 0.6974, 'grad_norm': 0.7154357858752877, 'learning_rate': 2.7048349887476038e-06, 'epoch': 0.97}
	0: �███▋\| 2882/2975 [33:26<00:55, 1.66it/s] 97%\|█████████▋\| 2883/2975 [33:27<00:55, 1.66it/s] 97%\|█████████▋\| 2884/2975 [33:27<00:54, 1.66it/s] 97%\|█████████▋\| 2885/2975 [33:28<00:54, 1.66it/s] 97%\|█████████▋\| 2886/2975 [33:29<00:53, 1.66it/s] 97%\|█████████▋\| 2887/2975 [33:29<00:53, 1.66it/s] 97%\|█████████▋\| 2888/2975 [33:30<00:52, 1.66it/s] 97%\|█████████▋\| 2889/2975 [33:30<00:51, 1.66it/s] 97%\|█████████▋\| 2890/2975 [33:31<00:51, 1.66it/s] 97%\|█████████▋\| 2890/2975 [33:31<00:51, 1.66it/s] 97%\|█████████▋\| 2891/2975 [33:32<00:50, 1.66it/s] 97%\|█████████▋\| 2892/2975 [33:32<00:50, 1.66it/s] 97%\|█████████▋\| 2893/2975 [33:33<00:49, 1.66it/s] 97%\|█████████▋\| 2894/2975 [33:33<00:
	0: {'loss': 0.7042, 'grad_norm': 0.6873844650072911, 'learning_rate': 2.3515149676898554e-06, 'epoch': 0.97}
	0: 48, 1.66it/s] 97%\|█████████▋\| 2895/2975 [33:34<00:48, 1.66it/s] 97%\|█████████▋\| 2896/2975 [33:35<00:47, 1.66it/s] 97%\|█████████▋\| 2897/2975 [33:35<00:46, 1.66it/s] 97%\|█████████▋\| 2898/2975 [33:36<00:46, 1.66it/s] 97%\|█████████▋\| 2899/2975 [33:36<00:45, 1.66it/s] 97%\|█████████▋\| 2900/2975 [33:37<00:45, 1.66it/s] 97%\|█████████▋\| 2900/2975 [33:37<00:45, 1.66it/s] 98%\|█████████▊\| 2901/2975 [33:38<00:44, 1.66it/s] 98%\|█████████▊\| 2902/2975 [33:38<00:43, 1.66it/s] 98%\|█████████▊\| 2903/2975 [33:39<00:43, 1.66it/s] 98%\|█████████▊\| 2904/2975 [33:39<00:42, 1.66it/s] 98%\|█████████▊\| 2905/2975 [33:40<00:42, 1.66it/s] 98%\|█████████▊\| 2906/2975 [33:41<00:41, 1.66it/s] 98%\|█████
	0: {'loss': 0.6907, 'grad_norm': 0.6885606635440618, 'learning_rate': 2.032690407508949e-06, 'epoch': 0.98}
	0: ████▊\| 2907/2975 [33:41<00:40, 1.66it/s] 98%\|█████████▊\| 2908/2975 [33:42<00:40, 1.66it/s] 98%\|█████████▊\| 2909/2975 [33:42<00:39, 1.66it/s] 98%\|█████████▊\| 2910/2975 [33:43<00:39, 1.66it/s] 98%\|█████████▊\| 2910/2975 [33:43<00:39, 1.66it/s] 98%\|█████████▊\| 2911/2975 [33:44<00:38, 1.66it/s] 98%\|█████████▊\| 2912/2975 [33:44<00:37, 1.66it/s] 98%\|█████████▊\| 2913/2975 [33:45<00:37, 1.66it/s] 98%\|█████████▊\| 2914/2975 [33:45<00:36, 1.66it/s] 98%\|█████████▊\| 2915/2975 [33:46<00:36, 1.66it/s] 98%\|█████████▊\| 2916/2975 [33:47<00:35, 1.66it/s] 98%\|█████████▊\| 2917/2975 [33:47<00:35, 1.63it/s] 98%\|█████████▊\| 2918/2975 [33:48<00:34, 1.64it/s] 98%\|█████████▊\| 2919/2975 [33:48<0
	0: {'loss': 0.6971, 'grad_norm': 0.6687159290240817, 'learning_rate': 1.7518544168045527e-06, 'epoch': 0.98}
	0: {'loss': 0.6822, 'grad_norm': 0.6974668873540893, 'learning_rate': 1.5120838934595338e-06, 'epoch': 0.98}
	0: 0:34, 1.65it/s] 98%\|█████████▊\| 2920/2975 [33:49<00:33, 1.65it/s] 98%\|█████████▊\| 2920/2975 [33:49<00:33, 1.65it/s] 98%\|█████████▊\| 2921/2975 [33:50<00:32, 1.65it/s] 98%\|█████████▊\| 2922/2975 [33:50<00:32, 1.65it/s] 98%\|█████████▊\| 2923/2975 [33:51<00:31, 1.65it/s] 98%\|█████████▊\| 2924/2975 [33:52<00:30, 1.65it/s] 98%\|█████████▊\| 2925/2975 [33:52<00:30, 1.65it/s] 98%\|█████████▊\| 2926/2975 [33:53<00:29, 1.65it/s] 98%\|█████████▊\| 2927/2975 [33:53<00:29, 1.64it/s] 98%\|█████████▊\| 2928/2975 [33:54<00:28, 1.65it/s] 98%\|█████████▊\| 2929/2975 [33:55<00:27, 1.65it/s] 98%\|█████████▊\| 2930/2975 [33:55<00:27, 1.65it/s] 98%\|█████████▊\|
	0: {'loss': 0.6865, 'grad_norm': 0.6900700149375906, 'learning_rate': 1.316005813502869e-06, 'epoch': 0.99}
	0: 2930/2975 [33:55<00:27, 1.65it/s] 99%\|█████████▊\| 2931/2975 [33:56<00:26, 1.65it/s] 99%\|█████████▊\| 2932/2975 [33:56<00:26, 1.65it/s] 99%\|█████████▊\| 2933/2975 [33:57<00:25, 1.65it/s] 99%\|█████████▊\| 2934/2975 [33:58<00:24, 1.65it/s] 99%\|█████████▊\| 2935/2975 [33:58<00:24, 1.65it/s] 99%\|█████████▊\| 2936/2975 [33:59<00:23, 1.65it/s] 99%\|█████████▊\| 2937/2975 [33:59<00:23, 1.65it/s] 99%\|█████████▉\| 2938/2975 [34:00<00:22, 1.65it/s] 99%\|█████████▉\| 2939/2975 [34:01<00:21, 1.65it/s] 99%\|█████████▉\| 2940/2975 [34:01<00:21, 1.66it/s] 99%\|█████████▉\| 2940/2975 [34:01<00:21, 1.66it/s] 99%\|█████████▉\| 2941/2975 [34:02<00:20, 1.66it/s] 99%\|█████████▉\| 2942/2975 [34:02<00:19, 1.66it/s]
	0: {'loss': 0.7135, 'grad_norm': 0.6716948809212612, 'learning_rate': 1.1657684494105386e-06, 'epoch': 0.99}
	0: 99%\|█████████▉\| 2943/2975 [34:03<00:19, 1.66it/s] 99%\|█████████▉\| 2944/2975 [34:04<00:18, 1.66it/s] 99%\|█████████▉\| 2945/2975 [34:04<00:18, 1.66it/s] 99%\|█████████▉\| 2946/2975 [34:05<00:17, 1.66it/s] 99%\|█████████▉\| 2947/2975 [34:05<00:16, 1.66it/s] 99%\|█████████▉\| 2948/2975 [34:06<00:16, 1.66it/s] 99%\|█████████▉\| 2949/2975 [34:07<00:15, 1.66it/s] 99%\|█████████▉\| 2950/2975 [34:07<00:15, 1.66it/s] 99%\|█████████▉\| 2950/2975 [34:07<00:15, 1.66it/s] 99%\|█████████▉\| 2951/2975 [34:08<00:14, 1.66it/s] 99%\|█████████▉\| 2952/2975 [34:08<00:13, 1.66it/s] 99%\|█████████▉\| 2953/2975 [34:09<00:13, 1.66it/s] 99%\|█████████▉\| 2954/2975 [34:10<00:12, 1.66it/s] 99%\|█████████▉
	0: {'loss': 0.6859, 'grad_norm': 0.6712348667758804, 'learning_rate': 1.0630178331827281e-06, 'epoch': 0.99}
	0: \| 2955/2975 [34:10<00:12, 1.66it/s] 99%\|█████████▉\| 2956/2975 [34:11<00:11, 1.66it/s] 99%\|█████████▉\| 2957/2975 [34:11<00:10, 1.66it/s] 99%\|█████████▉\| 2958/2975 [34:12<00:10, 1.66it/s] 99%\|█████████▉\| 2959/2975 [34:13<00:09, 1.66it/s] 99%\|█████████▉\| 2960/2975 [34:13<00:09, 1.66it/s] 99%\|█████████▉\| 2960/2975 [34:13<00:09, 1.66it/s] 100%\|█████████▉\| 2961/2975 [34:14<00:08, 1.66it/s] 100%\|█████████▉\| 2962/2975 [34:14<00:07, 1.66it/s] 100%\|█████████▉\| 2963/2975 [34:15<00:07, 1.66it/s] 100%\|█████████▉\| 2964/2975 [34:16<00:06, 1.66it/s] 100%\|█████████▉\| 2965/2975 [34:16<00:06, 1.66it/s] 100%\|█████████▉\| 2966/2975 [34:17<00:05, 1.66it/s] 100%\|█████████▉\| 2967/2975 [34:17<00:04, 1.65it/s
	0: {'loss': 0.6809, 'grad_norm': 0.6768243143096558, 'learning_rate': 1.008879722072778e-06, 'epoch': 1.0}
	0: {'train_runtime': 2064.8117, 'train_samples_per_second': 23.053, 'train_steps_per_second': 1.441, 'train_loss': 0.7352094054021755, 'epoch': 1.0}
	0: ] 100%\|█████████▉\| 2968/2975 [34:18<00:04, 1.66it/s] 100%\|█████████▉\| 2969/2975 [34:19<00:03, 1.66it/s] 100%\|█████████▉\| 2970/2975 [34:19<00:03, 1.66it/s] 100%\|█████████▉\| 2970/2975 [34:19<00:03, 1.66it/s] 100%\|█████████▉\| 2971/2975 [34:20<00:02, 1.66it/s] 100%\|█████████▉\| 2972/2975 [34:20<00:01, 1.66it/s] 100%\|█████████▉\| 2973/2975 [34:21<00:01, 1.66it/s] 100%\|█████████▉\| 2974/2975 [34:22<00:00, 1.66it/s] 100%\|██████████\| 2975/2975 [34:22<00:00, 1.66it/s] 100%\|██████████\| 2975/2975 [34:24<00:00, 1.66it/s] 100%\|██████████\| 2975/2975 [34:24<00:00, 1.44it/s]
	0: [2025-08-20 17:37:09,165] [INFO] [axolotl.train.save_trained_model:246] [PID:880949] [RANK:0] Training completed! Saving trained model to /lustre/fswork/projects/rech/dgo/udv55np/ift/Qwen3-235B-A22B/Qwen2.5-0.5B/0.[39m
	0: [2025-08-20 17:37:10,570] [INFO] [axolotl.train.save_trained_model:331] [PID:880949] [RANK:0] Model successfully saved to /lustre/fswork/projects/rech/dgo/udv55np/ift/Qwen3-235B-A22B/Qwen2.5-0.5B/0[39m