Model save

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,58 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: transformers
+model_name: Qwen2.5-1.5B-Open-R1-Distill
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for Qwen2.5-1.5B-Open-R1-Distill
+
+This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="bluuluu/Qwen2.5-1.5B-Open-R1-Distill", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/2741919970-hustvl/huggingface/runs/8v0s3ur8) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.15.0.dev0
+- Transformers: 4.49.0.dev0
+- Pytorch: 2.5.1
+- Datasets: 3.2.0
+- Tokenizers: 0.21.0
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

all_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 76959556042752.0,
+    "train_loss": 0.8068137678457665,
+    "train_runtime": 18078.9101,
+    "train_samples": 16610,
+    "train_samples_per_second": 2.392,
+    "train_steps_per_second": 0.075
+}

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.49.0.dev0"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:954f58fa168bc0d18b848661a47f2b087e94736c27a2cd4e959343518ab78bd0
+size 3087467144

special_tokens_map.json

@@ -0,0 +1,25 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|im_end|>"
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 76959556042752.0,
+    "train_loss": 0.8068137678457665,
+    "train_runtime": 18078.9101,
+    "train_samples": 16610,
+    "train_samples_per_second": 2.392,
+    "train_steps_per_second": 0.075
+}

trainer_state.json

@@ -0,0 +1,2320 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9998149861239592,
+  "eval_steps": 100,
+  "global_step": 1351,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0037002775208140612,
+      "grad_norm": 2.8957433700561523,
+      "learning_rate": 6.5522123978507875e-06,
+      "loss": 1.2055,
+      "mean_token_accuracy": 0.6906560870025514,
+      "step": 5
+    },
+    {
+      "epoch": 0.0074005550416281225,
+      "grad_norm": 1.8598837852478027,
+      "learning_rate": 9.37409668112298e-06,
+      "loss": 1.1409,
+      "mean_token_accuracy": 0.698853009887776,
+      "step": 10
+    },
+    {
+      "epoch": 0.011100832562442183,
+      "grad_norm": 1.537341594696045,
+      "learning_rate": 1.1024793168211608e-05,
+      "loss": 0.9966,
+      "mean_token_accuracy": 0.7270418325735069,
+      "step": 15
+    },
+    {
+      "epoch": 0.014801110083256245,
+      "grad_norm": 1.109471082687378,
+      "learning_rate": 1.219598096439517e-05,
+      "loss": 0.9761,
+      "mean_token_accuracy": 0.7272552272192254,
+      "step": 20
+    },
+    {
+      "epoch": 0.018501387604070305,
+      "grad_norm": 1.0324221849441528,
+      "learning_rate": 1.3104424795701575e-05,
+      "loss": 0.9551,
+      "mean_token_accuracy": 0.7313410652722199,
+      "step": 25
+    },
+    {
+      "epoch": 0.022201665124884366,
+      "grad_norm": 0.8868595957756042,
+      "learning_rate": 1.3846677451483799e-05,
+      "loss": 0.9201,
+      "mean_token_accuracy": 0.7379118125275668,
+      "step": 30
+    },
+    {
+      "epoch": 0.025901942645698426,
+      "grad_norm": 0.9547779560089111,
+      "learning_rate": 1.4474243129971968e-05,
+      "loss": 0.9167,
+      "mean_token_accuracy": 0.7379963784805907,
+      "step": 35
+    },
+    {
+      "epoch": 0.02960222016651249,
+      "grad_norm": 1.0785847902297974,
+      "learning_rate": 1.501786524766736e-05,
+      "loss": 0.8836,
+      "mean_token_accuracy": 0.746341194692062,
+      "step": 40
+    },
+    {
+      "epoch": 0.03330249768732655,
+      "grad_norm": 0.8535160422325134,
+      "learning_rate": 1.5497373938572426e-05,
+      "loss": 0.91,
+      "mean_token_accuracy": 0.7387234984522606,
+      "step": 45
+    },
+    {
+      "epoch": 0.03700277520814061,
+      "grad_norm": 0.9231573939323425,
+      "learning_rate": 1.592630907897377e-05,
+      "loss": 0.8773,
+      "mean_token_accuracy": 0.74560689911493,
+      "step": 50
+    },
+    {
+      "epoch": 0.04070305272895467,
+      "grad_norm": 0.9026180505752563,
+      "learning_rate": 1.6314328111538255e-05,
+      "loss": 0.8634,
+      "mean_token_accuracy": 0.7512907472411612,
+      "step": 55
+    },
+    {
+      "epoch": 0.04440333024976873,
+      "grad_norm": 1.0450937747955322,
+      "learning_rate": 1.6668561734755987e-05,
+      "loss": 0.8587,
+      "mean_token_accuracy": 0.7512712527442093,
+      "step": 60
+    },
+    {
+      "epoch": 0.04810360777058279,
+      "grad_norm": 1.0268136262893677,
+      "learning_rate": 1.6994425079669313e-05,
+      "loss": 0.8858,
+      "mean_token_accuracy": 0.7436999012547456,
+      "step": 65
+    },
+    {
+      "epoch": 0.05180388529139685,
+      "grad_norm": 1.0223065614700317,
+      "learning_rate": 1.7296127413244162e-05,
+      "loss": 0.8555,
+      "mean_token_accuracy": 0.7516197988052247,
+      "step": 70
+    },
+    {
+      "epoch": 0.05550416281221091,
+      "grad_norm": 0.9143498539924622,
+      "learning_rate": 1.7577005566062396e-05,
+      "loss": 0.914,
+      "mean_token_accuracy": 0.7367069906630946,
+      "step": 75
+    },
+    {
+      "epoch": 0.05920444033302498,
+      "grad_norm": 0.9483162760734558,
+      "learning_rate": 1.783974953093955e-05,
+      "loss": 0.872,
+      "mean_token_accuracy": 0.7465734032042433,
+      "step": 80
+    },
+    {
+      "epoch": 0.06290471785383904,
+      "grad_norm": 1.0431548357009888,
+      "learning_rate": 1.808655954803422e-05,
+      "loss": 0.8958,
+      "mean_token_accuracy": 0.739948661671668,
+      "step": 85
+    },
+    {
+      "epoch": 0.0666049953746531,
+      "grad_norm": 0.9597674012184143,
+      "learning_rate": 1.8319258221844618e-05,
+      "loss": 0.8504,
+      "mean_token_accuracy": 0.7529073109309405,
+      "step": 90
+    },
+    {
+      "epoch": 0.07030527289546716,
+      "grad_norm": 0.8986235857009888,
+      "learning_rate": 1.8539372284516762e-05,
+      "loss": 0.8689,
+      "mean_token_accuracy": 0.7483927127084614,
+      "step": 95
+    },
+    {
+      "epoch": 0.07400555041628122,
+      "grad_norm": 1.116733431816101,
+      "learning_rate": 1.874819336224596e-05,
+      "loss": 0.8829,
+      "mean_token_accuracy": 0.7434675110209781,
+      "step": 100
+    },
+    {
+      "epoch": 0.07400555041628122,
+      "eval_loss": 0.8832501769065857,
+      "eval_mean_token_accuracy": 0.7398939720176742,
+      "eval_runtime": 28.2477,
+      "eval_samples_per_second": 9.098,
+      "eval_steps_per_second": 1.168,
+      "step": 100
+    },
+    {
+      "epoch": 0.07770582793709528,
+      "grad_norm": 1.0423495769500732,
+      "learning_rate": 1.8946823900332786e-05,
+      "loss": 0.9029,
+      "mean_token_accuracy": 0.7380944747487156,
+      "step": 105
+    },
+    {
+      "epoch": 0.08140610545790934,
+      "grad_norm": 0.8770654797554016,
+      "learning_rate": 1.9136212394810443e-05,
+      "loss": 0.8512,
+      "mean_token_accuracy": 0.752658273084104,
+      "step": 110
+    },
+    {
+      "epoch": 0.0851063829787234,
+      "grad_norm": 0.9110291600227356,
+      "learning_rate": 1.9317180786056083e-05,
+      "loss": 0.8404,
+      "mean_token_accuracy": 0.7541489019069838,
+      "step": 115
+    },
+    {
+      "epoch": 0.08880666049953746,
+      "grad_norm": 1.0504673719406128,
+      "learning_rate": 1.949044601802818e-05,
+      "loss": 0.7908,
+      "mean_token_accuracy": 0.7690538308961964,
+      "step": 120
+    },
+    {
+      "epoch": 0.09250693802035152,
+      "grad_norm": 0.9688966870307922,
+      "learning_rate": 1.9656637193552367e-05,
+      "loss": 0.8283,
+      "mean_token_accuracy": 0.7583009069462628,
+      "step": 125
+    },
+    {
+      "epoch": 0.09620721554116558,
+      "grad_norm": 1.0260767936706543,
+      "learning_rate": 1.9816309362941505e-05,
+      "loss": 0.8764,
+      "mean_token_accuracy": 0.744777112966561,
+      "step": 130
+    },
+    {
+      "epoch": 0.09990749306197964,
+      "grad_norm": 0.9919949769973755,
+      "learning_rate": 1.996995470893325e-05,
+      "loss": 0.8696,
+      "mean_token_accuracy": 0.7461172694007374,
+      "step": 135
+    },
+    {
+      "epoch": 0.1036077705827937,
+      "grad_norm": 1.040398120880127,
+      "learning_rate": 2e-05,
+      "loss": 0.8734,
+      "mean_token_accuracy": 0.7447351537110054,
+      "step": 140
+    },
+    {
+      "epoch": 0.10730804810360776,
+      "grad_norm": 0.8929896354675293,
+      "learning_rate": 2e-05,
+      "loss": 0.8464,
+      "mean_token_accuracy": 0.7520211094548359,
+      "step": 145
+    },
+    {
+      "epoch": 0.11100832562442182,
+      "grad_norm": 1.0192770957946777,
+      "learning_rate": 2e-05,
+      "loss": 0.8769,
+      "mean_token_accuracy": 0.744968839209957,
+      "step": 150
+    },
+    {
+      "epoch": 0.1147086031452359,
+      "grad_norm": 1.0098509788513184,
+      "learning_rate": 2e-05,
+      "loss": 0.8323,
+      "mean_token_accuracy": 0.755749875133377,
+      "step": 155
+    },
+    {
+      "epoch": 0.11840888066604996,
+      "grad_norm": 1.056443691253662,
+      "learning_rate": 2e-05,
+      "loss": 0.8581,
+      "mean_token_accuracy": 0.7490507941249067,
+      "step": 160
+    },
+    {
+      "epoch": 0.12210915818686402,
+      "grad_norm": 0.876708447933197,
+      "learning_rate": 2e-05,
+      "loss": 0.8488,
+      "mean_token_accuracy": 0.7515548475709357,
+      "step": 165
+    },
+    {
+      "epoch": 0.12580943570767808,
+      "grad_norm": 0.9257837533950806,
+      "learning_rate": 2e-05,
+      "loss": 0.858,
+      "mean_token_accuracy": 0.7489278577453964,
+      "step": 170
+    },
+    {
+      "epoch": 0.12950971322849214,
+      "grad_norm": 1.0198274850845337,
+      "learning_rate": 2e-05,
+      "loss": 0.8361,
+      "mean_token_accuracy": 0.7551833607296417,
+      "step": 175
+    },
+    {
+      "epoch": 0.1332099907493062,
+      "grad_norm": 1.0134333372116089,
+      "learning_rate": 2e-05,
+      "loss": 0.8212,
+      "mean_token_accuracy": 0.7584818009818037,
+      "step": 180
+    },
+    {
+      "epoch": 0.13691026827012026,
+      "grad_norm": 1.0174723863601685,
+      "learning_rate": 2e-05,
+      "loss": 0.8628,
+      "mean_token_accuracy": 0.7476798968375172,
+      "step": 185
+    },
+    {
+      "epoch": 0.14061054579093432,
+      "grad_norm": 0.9869109988212585,
+      "learning_rate": 2e-05,
+      "loss": 0.8443,
+      "mean_token_accuracy": 0.7523148595161331,
+      "step": 190
+    },
+    {
+      "epoch": 0.14431082331174838,
+      "grad_norm": 0.9373570084571838,
+      "learning_rate": 2e-05,
+      "loss": 0.8582,
+      "mean_token_accuracy": 0.7481680463659037,
+      "step": 195
+    },
+    {
+      "epoch": 0.14801110083256244,
+      "grad_norm": 0.9523370862007141,
+      "learning_rate": 2e-05,
+      "loss": 0.839,
+      "mean_token_accuracy": 0.7536938574339362,
+      "step": 200
+    },
+    {
+      "epoch": 0.14801110083256244,
+      "eval_loss": 0.8592866659164429,
+      "eval_mean_token_accuracy": 0.7450125681976898,
+      "eval_runtime": 28.1828,
+      "eval_samples_per_second": 9.119,
+      "eval_steps_per_second": 1.171,
+      "step": 200
+    },
+    {
+      "epoch": 0.1517113783533765,
+      "grad_norm": 1.0168991088867188,
+      "learning_rate": 2e-05,
+      "loss": 0.8189,
+      "mean_token_accuracy": 0.7584566375595763,
+      "step": 205
+    },
+    {
+      "epoch": 0.15541165587419056,
+      "grad_norm": 0.9532285928726196,
+      "learning_rate": 2e-05,
+      "loss": 0.8488,
+      "mean_token_accuracy": 0.7494834024795103,
+      "step": 210
+    },
+    {
+      "epoch": 0.15911193339500462,
+      "grad_norm": 0.8827746510505676,
+      "learning_rate": 2e-05,
+      "loss": 0.8187,
+      "mean_token_accuracy": 0.7595932905014217,
+      "step": 215
+    },
+    {
+      "epoch": 0.16281221091581868,
+      "grad_norm": 0.8878557682037354,
+      "learning_rate": 2e-05,
+      "loss": 0.8276,
+      "mean_token_accuracy": 0.7566971009274142,
+      "step": 220
+    },
+    {
+      "epoch": 0.16651248843663274,
+      "grad_norm": 0.8196055293083191,
+      "learning_rate": 2e-05,
+      "loss": 0.8143,
+      "mean_token_accuracy": 0.7594526820177664,
+      "step": 225
+    },
+    {
+      "epoch": 0.1702127659574468,
+      "grad_norm": 0.8733744025230408,
+      "learning_rate": 2e-05,
+      "loss": 0.8433,
+      "mean_token_accuracy": 0.7518748700823068,
+      "step": 230
+    },
+    {
+      "epoch": 0.17391304347826086,
+      "grad_norm": 0.851927638053894,
+      "learning_rate": 2e-05,
+      "loss": 0.8615,
+      "mean_token_accuracy": 0.7447105825612464,
+      "step": 235
+    },
+    {
+      "epoch": 0.17761332099907493,
+      "grad_norm": 0.9009714126586914,
+      "learning_rate": 2e-05,
+      "loss": 0.8498,
+      "mean_token_accuracy": 0.7484718849440476,
+      "step": 240
+    },
+    {
+      "epoch": 0.18131359851988899,
+      "grad_norm": 0.9394606351852417,
+      "learning_rate": 2e-05,
+      "loss": 0.8294,
+      "mean_token_accuracy": 0.7557432874020297,
+      "step": 245
+    },
+    {
+      "epoch": 0.18501387604070305,
+      "grad_norm": 0.9246826171875,
+      "learning_rate": 2e-05,
+      "loss": 0.8004,
+      "mean_token_accuracy": 0.7644653627399571,
+      "step": 250
+    },
+    {
+      "epoch": 0.1887141535615171,
+      "grad_norm": 0.8407145142555237,
+      "learning_rate": 2e-05,
+      "loss": 0.8377,
+      "mean_token_accuracy": 0.7536744187275825,
+      "step": 255
+    },
+    {
+      "epoch": 0.19241443108233117,
+      "grad_norm": 0.8852797746658325,
+      "learning_rate": 2e-05,
+      "loss": 0.8338,
+      "mean_token_accuracy": 0.7559093371950457,
+      "step": 260
+    },
+    {
+      "epoch": 0.19611470860314523,
+      "grad_norm": 1.0910946130752563,
+      "learning_rate": 2e-05,
+      "loss": 0.8022,
+      "mean_token_accuracy": 0.7614075804389667,
+      "step": 265
+    },
+    {
+      "epoch": 0.1998149861239593,
+      "grad_norm": 0.9022583961486816,
+      "learning_rate": 2e-05,
+      "loss": 0.8226,
+      "mean_token_accuracy": 0.7575589018517326,
+      "step": 270
+    },
+    {
+      "epoch": 0.20351526364477335,
+      "grad_norm": 0.8302850723266602,
+      "learning_rate": 2e-05,
+      "loss": 0.8237,
+      "mean_token_accuracy": 0.757602458948465,
+      "step": 275
+    },
+    {
+      "epoch": 0.2072155411655874,
+      "grad_norm": 0.9338579773902893,
+      "learning_rate": 2e-05,
+      "loss": 0.8081,
+      "mean_token_accuracy": 0.7613954278187921,
+      "step": 280
+    },
+    {
+      "epoch": 0.21091581868640147,
+      "grad_norm": 0.9413399696350098,
+      "learning_rate": 2e-05,
+      "loss": 0.8522,
+      "mean_token_accuracy": 0.7486690000063428,
+      "step": 285
+    },
+    {
+      "epoch": 0.21461609620721553,
+      "grad_norm": 0.850651741027832,
+      "learning_rate": 2e-05,
+      "loss": 0.8085,
+      "mean_token_accuracy": 0.7597455023324446,
+      "step": 290
+    },
+    {
+      "epoch": 0.2183163737280296,
+      "grad_norm": 0.8325985074043274,
+      "learning_rate": 2e-05,
+      "loss": 0.8015,
+      "mean_token_accuracy": 0.7625035134700381,
+      "step": 295
+    },
+    {
+      "epoch": 0.22201665124884365,
+      "grad_norm": 0.8833345770835876,
+      "learning_rate": 2e-05,
+      "loss": 0.8442,
+      "mean_token_accuracy": 0.7495757818615388,
+      "step": 300
+    },
+    {
+      "epoch": 0.22201665124884365,
+      "eval_loss": 0.8446890115737915,
+      "eval_mean_token_accuracy": 0.7479223023166885,
+      "eval_runtime": 28.2235,
+      "eval_samples_per_second": 9.106,
+      "eval_steps_per_second": 1.169,
+      "step": 300
+    },
+    {
+      "epoch": 0.22571692876965774,
+      "grad_norm": 1.0072492361068726,
+      "learning_rate": 2e-05,
+      "loss": 0.8126,
+      "mean_token_accuracy": 0.7591269186975097,
+      "step": 305
+    },
+    {
+      "epoch": 0.2294172062904718,
+      "grad_norm": 0.9031651616096497,
+      "learning_rate": 2e-05,
+      "loss": 0.8281,
+      "mean_token_accuracy": 0.7545658190081542,
+      "step": 310
+    },
+    {
+      "epoch": 0.23311748381128586,
+      "grad_norm": 0.8532033562660217,
+      "learning_rate": 2e-05,
+      "loss": 0.8178,
+      "mean_token_accuracy": 0.7587916660812325,
+      "step": 315
+    },
+    {
+      "epoch": 0.23681776133209992,
+      "grad_norm": 0.9483544826507568,
+      "learning_rate": 2e-05,
+      "loss": 0.8451,
+      "mean_token_accuracy": 0.7510137117833572,
+      "step": 320
+    },
+    {
+      "epoch": 0.24051803885291398,
+      "grad_norm": 0.9190371036529541,
+      "learning_rate": 2e-05,
+      "loss": 0.8236,
+      "mean_token_accuracy": 0.7560184304098765,
+      "step": 325
+    },
+    {
+      "epoch": 0.24421831637372804,
+      "grad_norm": 0.8880885243415833,
+      "learning_rate": 2e-05,
+      "loss": 0.8116,
+      "mean_token_accuracy": 0.7590534226852792,
+      "step": 330
+    },
+    {
+      "epoch": 0.2479185938945421,
+      "grad_norm": 0.9330101609230042,
+      "learning_rate": 2e-05,
+      "loss": 0.8302,
+      "mean_token_accuracy": 0.7547550656313512,
+      "step": 335
+    },
+    {
+      "epoch": 0.25161887141535616,
+      "grad_norm": 0.8674152493476868,
+      "learning_rate": 2e-05,
+      "loss": 0.8351,
+      "mean_token_accuracy": 0.7541465255919348,
+      "step": 340
+    },
+    {
+      "epoch": 0.2553191489361702,
+      "grad_norm": 0.9083060026168823,
+      "learning_rate": 2e-05,
+      "loss": 0.8345,
+      "mean_token_accuracy": 0.7528752633371677,
+      "step": 345
+    },
+    {
+      "epoch": 0.2590194264569843,
+      "grad_norm": 0.9060890078544617,
+      "learning_rate": 2e-05,
+      "loss": 0.8216,
+      "mean_token_accuracy": 0.7575168041307812,
+      "step": 350
+    },
+    {
+      "epoch": 0.2627197039777983,
+      "grad_norm": 0.8555945158004761,
+      "learning_rate": 2e-05,
+      "loss": 0.783,
+      "mean_token_accuracy": 0.7672426588048891,
+      "step": 355
+    },
+    {
+      "epoch": 0.2664199814986124,
+      "grad_norm": 0.8146703839302063,
+      "learning_rate": 2e-05,
+      "loss": 0.8284,
+      "mean_token_accuracy": 0.7543931949875237,
+      "step": 360
+    },
+    {
+      "epoch": 0.27012025901942643,
+      "grad_norm": 0.9369416832923889,
+      "learning_rate": 2e-05,
+      "loss": 0.8016,
+      "mean_token_accuracy": 0.7621930585328915,
+      "step": 365
+    },
+    {
+      "epoch": 0.2738205365402405,
+      "grad_norm": 0.7886361479759216,
+      "learning_rate": 2e-05,
+      "loss": 0.8076,
+      "mean_token_accuracy": 0.7606381698886774,
+      "step": 370
+    },
+    {
+      "epoch": 0.27752081406105455,
+      "grad_norm": 0.9856371879577637,
+      "learning_rate": 2e-05,
+      "loss": 0.8025,
+      "mean_token_accuracy": 0.7619788416802726,
+      "step": 375
+    },
+    {
+      "epoch": 0.28122109158186864,
+      "grad_norm": 0.9179413318634033,
+      "learning_rate": 2e-05,
+      "loss": 0.8102,
+      "mean_token_accuracy": 0.7585128919069817,
+      "step": 380
+    },
+    {
+      "epoch": 0.2849213691026827,
+      "grad_norm": 0.945956289768219,
+      "learning_rate": 2e-05,
+      "loss": 0.7964,
+      "mean_token_accuracy": 0.7646142601688602,
+      "step": 385
+    },
+    {
+      "epoch": 0.28862164662349676,
+      "grad_norm": 0.9356254935264587,
+      "learning_rate": 2e-05,
+      "loss": 0.8281,
+      "mean_token_accuracy": 0.7552207106091118,
+      "step": 390
+    },
+    {
+      "epoch": 0.2923219241443108,
+      "grad_norm": 0.8899054527282715,
+      "learning_rate": 2e-05,
+      "loss": 0.8269,
+      "mean_token_accuracy": 0.7552909439013166,
+      "step": 395
+    },
+    {
+      "epoch": 0.2960222016651249,
+      "grad_norm": 0.9562066793441772,
+      "learning_rate": 2e-05,
+      "loss": 0.8067,
+      "mean_token_accuracy": 0.7602417180096708,
+      "step": 400
+    },
+    {
+      "epoch": 0.2960222016651249,
+      "eval_loss": 0.8349066376686096,
+      "eval_mean_token_accuracy": 0.7500406710334881,
+      "eval_runtime": 28.1787,
+      "eval_samples_per_second": 9.12,
+      "eval_steps_per_second": 1.171,
+      "step": 400
+    },
+    {
+      "epoch": 0.299722479185939,
+      "grad_norm": 1.0249305963516235,
+      "learning_rate": 2e-05,
+      "loss": 0.8002,
+      "mean_token_accuracy": 0.7625656294392676,
+      "step": 405
+    },
+    {
+      "epoch": 0.303422756706753,
+      "grad_norm": 0.8880975842475891,
+      "learning_rate": 2e-05,
+      "loss": 0.8045,
+      "mean_token_accuracy": 0.7602248209978305,
+      "step": 410
+    },
+    {
+      "epoch": 0.3071230342275671,
+      "grad_norm": 0.8439946174621582,
+      "learning_rate": 2e-05,
+      "loss": 0.8164,
+      "mean_token_accuracy": 0.75832583094361,
+      "step": 415
+    },
+    {
+      "epoch": 0.3108233117483811,
+      "grad_norm": 0.9006332159042358,
+      "learning_rate": 2e-05,
+      "loss": 0.8231,
+      "mean_token_accuracy": 0.7574048409584961,
+      "step": 420
+    },
+    {
+      "epoch": 0.3145235892691952,
+      "grad_norm": 0.8789312243461609,
+      "learning_rate": 2e-05,
+      "loss": 0.8076,
+      "mean_token_accuracy": 0.7590419888879876,
+      "step": 425
+    },
+    {
+      "epoch": 0.31822386679000925,
+      "grad_norm": 0.8401655554771423,
+      "learning_rate": 2e-05,
+      "loss": 0.8405,
+      "mean_token_accuracy": 0.7504104069436911,
+      "step": 430
+    },
+    {
+      "epoch": 0.32192414431082333,
+      "grad_norm": 0.8797054290771484,
+      "learning_rate": 2e-05,
+      "loss": 0.8261,
+      "mean_token_accuracy": 0.7548758435645662,
+      "step": 435
+    },
+    {
+      "epoch": 0.32562442183163737,
+      "grad_norm": 0.8514414429664612,
+      "learning_rate": 2e-05,
+      "loss": 0.7757,
+      "mean_token_accuracy": 0.7683710895600167,
+      "step": 440
+    },
+    {
+      "epoch": 0.32932469935245146,
+      "grad_norm": 0.8692080974578857,
+      "learning_rate": 2e-05,
+      "loss": 0.8089,
+      "mean_token_accuracy": 0.7593312312489597,
+      "step": 445
+    },
+    {
+      "epoch": 0.3330249768732655,
+      "grad_norm": 0.8913412690162659,
+      "learning_rate": 2e-05,
+      "loss": 0.8234,
+      "mean_token_accuracy": 0.7565642409607426,
+      "step": 450
+    },
+    {
+      "epoch": 0.3367252543940796,
+      "grad_norm": 0.8350377678871155,
+      "learning_rate": 2e-05,
+      "loss": 0.7822,
+      "mean_token_accuracy": 0.7671309817518462,
+      "step": 455
+    },
+    {
+      "epoch": 0.3404255319148936,
+      "grad_norm": 0.8672026991844177,
+      "learning_rate": 2e-05,
+      "loss": 0.8021,
+      "mean_token_accuracy": 0.7619013855929817,
+      "step": 460
+    },
+    {
+      "epoch": 0.3441258094357077,
+      "grad_norm": 0.8754116892814636,
+      "learning_rate": 2e-05,
+      "loss": 0.8281,
+      "mean_token_accuracy": 0.7542948754676929,
+      "step": 465
+    },
+    {
+      "epoch": 0.34782608695652173,
+      "grad_norm": 0.8823506236076355,
+      "learning_rate": 2e-05,
+      "loss": 0.8224,
+      "mean_token_accuracy": 0.7557379489990506,
+      "step": 470
+    },
+    {
+      "epoch": 0.3515263644773358,
+      "grad_norm": 0.8378307819366455,
+      "learning_rate": 2e-05,
+      "loss": 0.8235,
+      "mean_token_accuracy": 0.7557773294176604,
+      "step": 475
+    },
+    {
+      "epoch": 0.35522664199814985,
+      "grad_norm": 0.764640748500824,
+      "learning_rate": 2e-05,
+      "loss": 0.7811,
+      "mean_token_accuracy": 0.7674092078356458,
+      "step": 480
+    },
+    {
+      "epoch": 0.35892691951896394,
+      "grad_norm": 0.8759243488311768,
+      "learning_rate": 2e-05,
+      "loss": 0.8143,
+      "mean_token_accuracy": 0.7586287599334172,
+      "step": 485
+    },
+    {
+      "epoch": 0.36262719703977797,
+      "grad_norm": 0.8526579141616821,
+      "learning_rate": 2e-05,
+      "loss": 0.7947,
+      "mean_token_accuracy": 0.7634890650281875,
+      "step": 490
+    },
+    {
+      "epoch": 0.36632747456059206,
+      "grad_norm": 0.8783523440361023,
+      "learning_rate": 2e-05,
+      "loss": 0.8113,
+      "mean_token_accuracy": 0.758303520901475,
+      "step": 495
+    },
+    {
+      "epoch": 0.3700277520814061,
+      "grad_norm": 0.8912361264228821,
+      "learning_rate": 2e-05,
+      "loss": 0.8004,
+      "mean_token_accuracy": 0.7601969218920052,
+      "step": 500
+    },
+    {
+      "epoch": 0.3700277520814061,
+      "eval_loss": 0.8273892402648926,
+      "eval_mean_token_accuracy": 0.7517986458893432,
+      "eval_runtime": 28.2564,
+      "eval_samples_per_second": 9.095,
+      "eval_steps_per_second": 1.168,
+      "step": 500
+    },
+    {
+      "epoch": 0.3737280296022202,
+      "grad_norm": 0.8783790469169617,
+      "learning_rate": 2e-05,
+      "loss": 0.8166,
+      "mean_token_accuracy": 0.7571132549186148,
+      "step": 505
+    },
+    {
+      "epoch": 0.3774283071230342,
+      "grad_norm": 0.8384273648262024,
+      "learning_rate": 2e-05,
+      "loss": 0.7976,
+      "mean_token_accuracy": 0.7625027165364132,
+      "step": 510
+    },
+    {
+      "epoch": 0.3811285846438483,
+      "grad_norm": 0.8020018339157104,
+      "learning_rate": 2e-05,
+      "loss": 0.8005,
+      "mean_token_accuracy": 0.7613467266492346,
+      "step": 515
+    },
+    {
+      "epoch": 0.38482886216466233,
+      "grad_norm": 0.864335298538208,
+      "learning_rate": 2e-05,
+      "loss": 0.8166,
+      "mean_token_accuracy": 0.7572026062300129,
+      "step": 520
+    },
+    {
+      "epoch": 0.3885291396854764,
+      "grad_norm": 0.8236207365989685,
+      "learning_rate": 2e-05,
+      "loss": 0.8345,
+      "mean_token_accuracy": 0.7512810127712488,
+      "step": 525
+    },
+    {
+      "epoch": 0.39222941720629045,
+      "grad_norm": 0.922486424446106,
+      "learning_rate": 2e-05,
+      "loss": 0.8128,
+      "mean_token_accuracy": 0.7574585249731419,
+      "step": 530
+    },
+    {
+      "epoch": 0.39592969472710454,
+      "grad_norm": 0.8133776187896729,
+      "learning_rate": 2e-05,
+      "loss": 0.8045,
+      "mean_token_accuracy": 0.7600427585053521,
+      "step": 535
+    },
+    {
+      "epoch": 0.3996299722479186,
+      "grad_norm": 0.9678435921669006,
+      "learning_rate": 2e-05,
+      "loss": 0.8434,
+      "mean_token_accuracy": 0.748454559157412,
+      "step": 540
+    },
+    {
+      "epoch": 0.40333024976873266,
+      "grad_norm": 0.8359988927841187,
+      "learning_rate": 2e-05,
+      "loss": 0.8218,
+      "mean_token_accuracy": 0.7548506933279708,
+      "step": 545
+    },
+    {
+      "epoch": 0.4070305272895467,
+      "grad_norm": 0.9044314622879028,
+      "learning_rate": 2e-05,
+      "loss": 0.773,
+      "mean_token_accuracy": 0.7704903178347341,
+      "step": 550
+    },
+    {
+      "epoch": 0.4107308048103608,
+      "grad_norm": 0.8256170153617859,
+      "learning_rate": 2e-05,
+      "loss": 0.7998,
+      "mean_token_accuracy": 0.7627984143152889,
+      "step": 555
+    },
+    {
+      "epoch": 0.4144310823311748,
+      "grad_norm": 0.8337195515632629,
+      "learning_rate": 2e-05,
+      "loss": 0.7815,
+      "mean_token_accuracy": 0.7671558285449628,
+      "step": 560
+    },
+    {
+      "epoch": 0.4181313598519889,
+      "grad_norm": 0.824918270111084,
+      "learning_rate": 2e-05,
+      "loss": 0.7931,
+      "mean_token_accuracy": 0.7636928037001625,
+      "step": 565
+    },
+    {
+      "epoch": 0.42183163737280294,
+      "grad_norm": 0.8702638149261475,
+      "learning_rate": 2e-05,
+      "loss": 0.7799,
+      "mean_token_accuracy": 0.766696658125059,
+      "step": 570
+    },
+    {
+      "epoch": 0.425531914893617,
+      "grad_norm": 0.9347019195556641,
+      "learning_rate": 2e-05,
+      "loss": 0.8027,
+      "mean_token_accuracy": 0.759642292609576,
+      "step": 575
+    },
+    {
+      "epoch": 0.42923219241443106,
+      "grad_norm": 0.8226842284202576,
+      "learning_rate": 2e-05,
+      "loss": 0.8318,
+      "mean_token_accuracy": 0.7513268577092702,
+      "step": 580
+    },
+    {
+      "epoch": 0.43293246993524515,
+      "grad_norm": 0.8303760886192322,
+      "learning_rate": 2e-05,
+      "loss": 0.7847,
+      "mean_token_accuracy": 0.7666960565561379,
+      "step": 585
+    },
+    {
+      "epoch": 0.4366327474560592,
+      "grad_norm": 0.864280641078949,
+      "learning_rate": 2e-05,
+      "loss": 0.7967,
+      "mean_token_accuracy": 0.7617017729986968,
+      "step": 590
+    },
+    {
+      "epoch": 0.44033302497687327,
+      "grad_norm": 0.8039905428886414,
+      "learning_rate": 2e-05,
+      "loss": 0.7952,
+      "mean_token_accuracy": 0.7631796585579562,
+      "step": 595
+    },
+    {
+      "epoch": 0.4440333024976873,
+      "grad_norm": 0.8144698143005371,
+      "learning_rate": 2e-05,
+      "loss": 0.7595,
+      "mean_token_accuracy": 0.7731878196083116,
+      "step": 600
+    },
+    {
+      "epoch": 0.4440333024976873,
+      "eval_loss": 0.8222786784172058,
+      "eval_mean_token_accuracy": 0.7522364520321178,
+      "eval_runtime": 28.1935,
+      "eval_samples_per_second": 9.116,
+      "eval_steps_per_second": 1.17,
+      "step": 600
+    },
+    {
+      "epoch": 0.4477335800185014,
+      "grad_norm": 0.898853600025177,
+      "learning_rate": 2e-05,
+      "loss": 0.8311,
+      "mean_token_accuracy": 0.7522999408717911,
+      "step": 605
+    },
+    {
+      "epoch": 0.4514338575393155,
+      "grad_norm": 0.7734112739562988,
+      "learning_rate": 2e-05,
+      "loss": 0.7892,
+      "mean_token_accuracy": 0.7639044411576632,
+      "step": 610
+    },
+    {
+      "epoch": 0.4551341350601295,
+      "grad_norm": 0.8656638860702515,
+      "learning_rate": 2e-05,
+      "loss": 0.7942,
+      "mean_token_accuracy": 0.7623657197010026,
+      "step": 615
+    },
+    {
+      "epoch": 0.4588344125809436,
+      "grad_norm": 0.9421690106391907,
+      "learning_rate": 2e-05,
+      "loss": 0.7726,
+      "mean_token_accuracy": 0.7677914333370994,
+      "step": 620
+    },
+    {
+      "epoch": 0.46253469010175763,
+      "grad_norm": 0.810966968536377,
+      "learning_rate": 2e-05,
+      "loss": 0.7738,
+      "mean_token_accuracy": 0.7682241466373756,
+      "step": 625
+    },
+    {
+      "epoch": 0.4662349676225717,
+      "grad_norm": 0.8764069080352783,
+      "learning_rate": 2e-05,
+      "loss": 0.764,
+      "mean_token_accuracy": 0.7716749303779987,
+      "step": 630
+    },
+    {
+      "epoch": 0.46993524514338575,
+      "grad_norm": 0.8801226615905762,
+      "learning_rate": 2e-05,
+      "loss": 0.7868,
+      "mean_token_accuracy": 0.7653086539883042,
+      "step": 635
+    },
+    {
+      "epoch": 0.47363552266419984,
+      "grad_norm": 0.789652943611145,
+      "learning_rate": 2e-05,
+      "loss": 0.7996,
+      "mean_token_accuracy": 0.7604761749526181,
+      "step": 640
+    },
+    {
+      "epoch": 0.47733580018501387,
+      "grad_norm": 0.8484530448913574,
+      "learning_rate": 2e-05,
+      "loss": 0.7966,
+      "mean_token_accuracy": 0.7630161460550556,
+      "step": 645
+    },
+    {
+      "epoch": 0.48103607770582796,
+      "grad_norm": 0.8077908754348755,
+      "learning_rate": 2e-05,
+      "loss": 0.7931,
+      "mean_token_accuracy": 0.763731608290881,
+      "step": 650
+    },
+    {
+      "epoch": 0.484736355226642,
+      "grad_norm": 0.8435987830162048,
+      "learning_rate": 2e-05,
+      "loss": 0.7695,
+      "mean_token_accuracy": 0.7693811651575391,
+      "step": 655
+    },
+    {
+      "epoch": 0.4884366327474561,
+      "grad_norm": 0.7886292934417725,
+      "learning_rate": 2e-05,
+      "loss": 0.794,
+      "mean_token_accuracy": 0.7628652561375162,
+      "step": 660
+    },
+    {
+      "epoch": 0.4921369102682701,
+      "grad_norm": 0.9320831894874573,
+      "learning_rate": 2e-05,
+      "loss": 0.8054,
+      "mean_token_accuracy": 0.7584387909891809,
+      "step": 665
+    },
+    {
+      "epoch": 0.4958371877890842,
+      "grad_norm": 0.8228521943092346,
+      "learning_rate": 2e-05,
+      "loss": 0.7931,
+      "mean_token_accuracy": 0.762219202960193,
+      "step": 670
+    },
+    {
+      "epoch": 0.49953746530989823,
+      "grad_norm": 0.9157373905181885,
+      "learning_rate": 2e-05,
+      "loss": 0.7726,
+      "mean_token_accuracy": 0.766911525229679,
+      "step": 675
+    },
+    {
+      "epoch": 0.5032377428307123,
+      "grad_norm": 0.8757467865943909,
+      "learning_rate": 2e-05,
+      "loss": 0.7925,
+      "mean_token_accuracy": 0.7622793847521553,
+      "step": 680
+    },
+    {
+      "epoch": 0.5069380203515264,
+      "grad_norm": 0.9015477895736694,
+      "learning_rate": 2e-05,
+      "loss": 0.808,
+      "mean_token_accuracy": 0.7594514579501054,
+      "step": 685
+    },
+    {
+      "epoch": 0.5106382978723404,
+      "grad_norm": 0.8006423711776733,
+      "learning_rate": 2e-05,
+      "loss": 0.7802,
+      "mean_token_accuracy": 0.7662706193752193,
+      "step": 690
+    },
+    {
+      "epoch": 0.5143385753931545,
+      "grad_norm": 0.9912610650062561,
+      "learning_rate": 2e-05,
+      "loss": 0.7897,
+      "mean_token_accuracy": 0.761338202651713,
+      "step": 695
+    },
+    {
+      "epoch": 0.5180388529139686,
+      "grad_norm": 0.8785173296928406,
+      "learning_rate": 2e-05,
+      "loss": 0.7866,
+      "mean_token_accuracy": 0.7645523463170034,
+      "step": 700
+    },
+    {
+      "epoch": 0.5180388529139686,
+      "eval_loss": 0.816895604133606,
+      "eval_mean_token_accuracy": 0.7541631101443679,
+      "eval_runtime": 28.2109,
+      "eval_samples_per_second": 9.11,
+      "eval_steps_per_second": 1.17,
+      "step": 700
+    },
+    {
+      "epoch": 0.5217391304347826,
+      "grad_norm": 0.9194395542144775,
+      "learning_rate": 2e-05,
+      "loss": 0.7962,
+      "mean_token_accuracy": 0.7632879072410094,
+      "step": 705
+    },
+    {
+      "epoch": 0.5254394079555966,
+      "grad_norm": 0.8533481359481812,
+      "learning_rate": 2e-05,
+      "loss": 0.7945,
+      "mean_token_accuracy": 0.7621978633882525,
+      "step": 710
+    },
+    {
+      "epoch": 0.5291396854764108,
+      "grad_norm": 0.7783718705177307,
+      "learning_rate": 2e-05,
+      "loss": 0.7771,
+      "mean_token_accuracy": 0.7678239959564587,
+      "step": 715
+    },
+    {
+      "epoch": 0.5328399629972248,
+      "grad_norm": 0.9274916052818298,
+      "learning_rate": 2e-05,
+      "loss": 0.7996,
+      "mean_token_accuracy": 0.7604699077057977,
+      "step": 720
+    },
+    {
+      "epoch": 0.5365402405180388,
+      "grad_norm": 0.8770431280136108,
+      "learning_rate": 2e-05,
+      "loss": 0.7845,
+      "mean_token_accuracy": 0.7654583712388715,
+      "step": 725
+    },
+    {
+      "epoch": 0.5402405180388529,
+      "grad_norm": 0.8453335762023926,
+      "learning_rate": 2e-05,
+      "loss": 0.7938,
+      "mean_token_accuracy": 0.761018252526195,
+      "step": 730
+    },
+    {
+      "epoch": 0.543940795559667,
+      "grad_norm": 0.7490786910057068,
+      "learning_rate": 2e-05,
+      "loss": 0.8167,
+      "mean_token_accuracy": 0.7549528979795714,
+      "step": 735
+    },
+    {
+      "epoch": 0.547641073080481,
+      "grad_norm": 0.9425523281097412,
+      "learning_rate": 2e-05,
+      "loss": 0.8059,
+      "mean_token_accuracy": 0.7600661300171085,
+      "step": 740
+    },
+    {
+      "epoch": 0.5513413506012951,
+      "grad_norm": 0.8289265036582947,
+      "learning_rate": 2e-05,
+      "loss": 0.7797,
+      "mean_token_accuracy": 0.7658190325778043,
+      "step": 745
+    },
+    {
+      "epoch": 0.5550416281221091,
+      "grad_norm": 0.808990478515625,
+      "learning_rate": 2e-05,
+      "loss": 0.7919,
+      "mean_token_accuracy": 0.7639361658364197,
+      "step": 750
+    },
+    {
+      "epoch": 0.5587419056429233,
+      "grad_norm": 0.8714848756790161,
+      "learning_rate": 2e-05,
+      "loss": 0.8133,
+      "mean_token_accuracy": 0.7559352930609992,
+      "step": 755
+    },
+    {
+      "epoch": 0.5624421831637373,
+      "grad_norm": 0.8202362060546875,
+      "learning_rate": 2e-05,
+      "loss": 0.7967,
+      "mean_token_accuracy": 0.7620283378489665,
+      "step": 760
+    },
+    {
+      "epoch": 0.5661424606845513,
+      "grad_norm": 0.8363901376724243,
+      "learning_rate": 2e-05,
+      "loss": 0.7856,
+      "mean_token_accuracy": 0.7647013079094134,
+      "step": 765
+    },
+    {
+      "epoch": 0.5698427382053654,
+      "grad_norm": 0.7312168478965759,
+      "learning_rate": 2e-05,
+      "loss": 0.7807,
+      "mean_token_accuracy": 0.7657156813892808,
+      "step": 770
+    },
+    {
+      "epoch": 0.5735430157261795,
+      "grad_norm": 0.8106317520141602,
+      "learning_rate": 2e-05,
+      "loss": 0.7833,
+      "mean_token_accuracy": 0.7646042951530488,
+      "step": 775
+    },
+    {
+      "epoch": 0.5772432932469935,
+      "grad_norm": 0.879208505153656,
+      "learning_rate": 2e-05,
+      "loss": 0.7906,
+      "mean_token_accuracy": 0.7629727484373736,
+      "step": 780
+    },
+    {
+      "epoch": 0.5809435707678076,
+      "grad_norm": 0.8280200958251953,
+      "learning_rate": 2e-05,
+      "loss": 0.7931,
+      "mean_token_accuracy": 0.7624752387623053,
+      "step": 785
+    },
+    {
+      "epoch": 0.5846438482886216,
+      "grad_norm": 0.8161231279373169,
+      "learning_rate": 2e-05,
+      "loss": 0.7979,
+      "mean_token_accuracy": 0.7606783559652616,
+      "step": 790
+    },
+    {
+      "epoch": 0.5883441258094357,
+      "grad_norm": 0.8787475228309631,
+      "learning_rate": 2e-05,
+      "loss": 0.8031,
+      "mean_token_accuracy": 0.7610423345421407,
+      "step": 795
+    },
+    {
+      "epoch": 0.5920444033302498,
+      "grad_norm": 0.8519960641860962,
+      "learning_rate": 2e-05,
+      "loss": 0.8219,
+      "mean_token_accuracy": 0.7543098449152789,
+      "step": 800
+    },
+    {
+      "epoch": 0.5920444033302498,
+      "eval_loss": 0.8132702708244324,
+      "eval_mean_token_accuracy": 0.7547294497721886,
+      "eval_runtime": 28.2905,
+      "eval_samples_per_second": 9.084,
+      "eval_steps_per_second": 1.166,
+      "step": 800
+    },
+    {
+      "epoch": 0.5957446808510638,
+      "grad_norm": 0.8904110193252563,
+      "learning_rate": 2e-05,
+      "loss": 0.7317,
+      "mean_token_accuracy": 0.7804520508941676,
+      "step": 805
+    },
+    {
+      "epoch": 0.599444958371878,
+      "grad_norm": 0.8179667592048645,
+      "learning_rate": 2e-05,
+      "loss": 0.7845,
+      "mean_token_accuracy": 0.7655715065591602,
+      "step": 810
+    },
+    {
+      "epoch": 0.603145235892692,
+      "grad_norm": 0.7966557145118713,
+      "learning_rate": 2e-05,
+      "loss": 0.7778,
+      "mean_token_accuracy": 0.7667068888225074,
+      "step": 815
+    },
+    {
+      "epoch": 0.606845513413506,
+      "grad_norm": 0.8500820994377136,
+      "learning_rate": 2e-05,
+      "loss": 0.7909,
+      "mean_token_accuracy": 0.7624461584090862,
+      "step": 820
+    },
+    {
+      "epoch": 0.61054579093432,
+      "grad_norm": 0.876806378364563,
+      "learning_rate": 2e-05,
+      "loss": 0.7986,
+      "mean_token_accuracy": 0.7598412421319061,
+      "step": 825
+    },
+    {
+      "epoch": 0.6142460684551342,
+      "grad_norm": 0.8352357745170593,
+      "learning_rate": 2e-05,
+      "loss": 0.8045,
+      "mean_token_accuracy": 0.7595125698366062,
+      "step": 830
+    },
+    {
+      "epoch": 0.6179463459759482,
+      "grad_norm": 0.8609703779220581,
+      "learning_rate": 2e-05,
+      "loss": 0.7653,
+      "mean_token_accuracy": 0.7705634200292611,
+      "step": 835
+    },
+    {
+      "epoch": 0.6216466234967623,
+      "grad_norm": 0.8633068203926086,
+      "learning_rate": 2e-05,
+      "loss": 0.7696,
+      "mean_token_accuracy": 0.7682237681411259,
+      "step": 840
+    },
+    {
+      "epoch": 0.6253469010175763,
+      "grad_norm": 0.8756479620933533,
+      "learning_rate": 2e-05,
+      "loss": 0.7786,
+      "mean_token_accuracy": 0.7672769183942045,
+      "step": 845
+    },
+    {
+      "epoch": 0.6290471785383904,
+      "grad_norm": 0.792131781578064,
+      "learning_rate": 2e-05,
+      "loss": 0.8053,
+      "mean_token_accuracy": 0.7596922873144693,
+      "step": 850
+    },
+    {
+      "epoch": 0.6327474560592045,
+      "grad_norm": 0.7844769954681396,
+      "learning_rate": 2e-05,
+      "loss": 0.7427,
+      "mean_token_accuracy": 0.7763440537258867,
+      "step": 855
+    },
+    {
+      "epoch": 0.6364477335800185,
+      "grad_norm": 0.770456850528717,
+      "learning_rate": 2e-05,
+      "loss": 0.8082,
+      "mean_token_accuracy": 0.7579604529340764,
+      "step": 860
+    },
+    {
+      "epoch": 0.6401480111008325,
+      "grad_norm": 0.8686452507972717,
+      "learning_rate": 2e-05,
+      "loss": 0.7895,
+      "mean_token_accuracy": 0.7622386697563842,
+      "step": 865
+    },
+    {
+      "epoch": 0.6438482886216467,
+      "grad_norm": 0.7865906357765198,
+      "learning_rate": 2e-05,
+      "loss": 0.7622,
+      "mean_token_accuracy": 0.7711946538689294,
+      "step": 870
+    },
+    {
+      "epoch": 0.6475485661424607,
+      "grad_norm": 0.8619543313980103,
+      "learning_rate": 2e-05,
+      "loss": 0.7776,
+      "mean_token_accuracy": 0.7680711539071554,
+      "step": 875
+    },
+    {
+      "epoch": 0.6512488436632747,
+      "grad_norm": 0.8202876448631287,
+      "learning_rate": 2e-05,
+      "loss": 0.7633,
+      "mean_token_accuracy": 0.7725941263025453,
+      "step": 880
+    },
+    {
+      "epoch": 0.6549491211840888,
+      "grad_norm": 0.7937086820602417,
+      "learning_rate": 2e-05,
+      "loss": 0.803,
+      "mean_token_accuracy": 0.7600367861507152,
+      "step": 885
+    },
+    {
+      "epoch": 0.6586493987049029,
+      "grad_norm": 0.889613687992096,
+      "learning_rate": 2e-05,
+      "loss": 0.7807,
+      "mean_token_accuracy": 0.7660721298978383,
+      "step": 890
+    },
+    {
+      "epoch": 0.6623496762257169,
+      "grad_norm": 0.9099228382110596,
+      "learning_rate": 2e-05,
+      "loss": 0.7911,
+      "mean_token_accuracy": 0.7620904760689602,
+      "step": 895
+    },
+    {
+      "epoch": 0.666049953746531,
+      "grad_norm": 0.8228977918624878,
+      "learning_rate": 2e-05,
+      "loss": 0.7695,
+      "mean_token_accuracy": 0.7692359508353532,
+      "step": 900
+    },
+    {
+      "epoch": 0.666049953746531,
+      "eval_loss": 0.8099245429039001,
+      "eval_mean_token_accuracy": 0.7551735807763004,
+      "eval_runtime": 28.1688,
+      "eval_samples_per_second": 9.124,
+      "eval_steps_per_second": 1.172,
+      "step": 900
+    },
+    {
+      "epoch": 0.669750231267345,
+      "grad_norm": 0.8004016280174255,
+      "learning_rate": 2e-05,
+      "loss": 0.7658,
+      "mean_token_accuracy": 0.7700296924370021,
+      "step": 905
+    },
+    {
+      "epoch": 0.6734505087881592,
+      "grad_norm": 0.8088281154632568,
+      "learning_rate": 2e-05,
+      "loss": 0.7449,
+      "mean_token_accuracy": 0.7756549136654808,
+      "step": 910
+    },
+    {
+      "epoch": 0.6771507863089732,
+      "grad_norm": 0.8305289149284363,
+      "learning_rate": 2e-05,
+      "loss": 0.7748,
+      "mean_token_accuracy": 0.7683322851408529,
+      "step": 915
+    },
+    {
+      "epoch": 0.6808510638297872,
+      "grad_norm": 0.8163439631462097,
+      "learning_rate": 2e-05,
+      "loss": 0.7852,
+      "mean_token_accuracy": 0.7640544732853833,
+      "step": 920
+    },
+    {
+      "epoch": 0.6845513413506013,
+      "grad_norm": 0.8671082854270935,
+      "learning_rate": 2e-05,
+      "loss": 0.8049,
+      "mean_token_accuracy": 0.7582682713568054,
+      "step": 925
+    },
+    {
+      "epoch": 0.6882516188714154,
+      "grad_norm": 0.8410265445709229,
+      "learning_rate": 2e-05,
+      "loss": 0.7935,
+      "mean_token_accuracy": 0.7615948024373346,
+      "step": 930
+    },
+    {
+      "epoch": 0.6919518963922294,
+      "grad_norm": 0.8609654307365417,
+      "learning_rate": 2e-05,
+      "loss": 0.793,
+      "mean_token_accuracy": 0.7631270531301052,
+      "step": 935
+    },
+    {
+      "epoch": 0.6956521739130435,
+      "grad_norm": 0.829707145690918,
+      "learning_rate": 2e-05,
+      "loss": 0.7669,
+      "mean_token_accuracy": 0.7696065254786408,
+      "step": 940
+    },
+    {
+      "epoch": 0.6993524514338575,
+      "grad_norm": 0.8150595426559448,
+      "learning_rate": 2e-05,
+      "loss": 0.7984,
+      "mean_token_accuracy": 0.7615524716113254,
+      "step": 945
+    },
+    {
+      "epoch": 0.7030527289546716,
+      "grad_norm": 0.8580748438835144,
+      "learning_rate": 2e-05,
+      "loss": 0.7689,
+      "mean_token_accuracy": 0.7683636381982694,
+      "step": 950
+    },
+    {
+      "epoch": 0.7067530064754857,
+      "grad_norm": 0.774782121181488,
+      "learning_rate": 2e-05,
+      "loss": 0.7895,
+      "mean_token_accuracy": 0.7641451735925326,
+      "step": 955
+    },
+    {
+      "epoch": 0.7104532839962997,
+      "grad_norm": 0.8466439843177795,
+      "learning_rate": 2e-05,
+      "loss": 0.7651,
+      "mean_token_accuracy": 0.7699041692846329,
+      "step": 960
+    },
+    {
+      "epoch": 0.7141535615171137,
+      "grad_norm": 0.8498072028160095,
+      "learning_rate": 2e-05,
+      "loss": 0.7708,
+      "mean_token_accuracy": 0.7683624308125098,
+      "step": 965
+    },
+    {
+      "epoch": 0.7178538390379279,
+      "grad_norm": 0.8351811170578003,
+      "learning_rate": 2e-05,
+      "loss": 0.7811,
+      "mean_token_accuracy": 0.7650397336041128,
+      "step": 970
+    },
+    {
+      "epoch": 0.7215541165587419,
+      "grad_norm": 0.9042247533798218,
+      "learning_rate": 2e-05,
+      "loss": 0.7654,
+      "mean_token_accuracy": 0.7692840039456211,
+      "step": 975
+    },
+    {
+      "epoch": 0.7252543940795559,
+      "grad_norm": 0.7749535441398621,
+      "learning_rate": 2e-05,
+      "loss": 0.7597,
+      "mean_token_accuracy": 0.7701962410947761,
+      "step": 980
+    },
+    {
+      "epoch": 0.72895467160037,
+      "grad_norm": 0.8533946871757507,
+      "learning_rate": 2e-05,
+      "loss": 0.7784,
+      "mean_token_accuracy": 0.7647845122361627,
+      "step": 985
+    },
+    {
+      "epoch": 0.7326549491211841,
+      "grad_norm": 0.8168575763702393,
+      "learning_rate": 2e-05,
+      "loss": 0.8086,
+      "mean_token_accuracy": 0.757385044925169,
+      "step": 990
+    },
+    {
+      "epoch": 0.7363552266419982,
+      "grad_norm": 0.7941954135894775,
+      "learning_rate": 2e-05,
+      "loss": 0.765,
+      "mean_token_accuracy": 0.7690882272352915,
+      "step": 995
+    },
+    {
+      "epoch": 0.7400555041628122,
+      "grad_norm": 0.8234696984291077,
+      "learning_rate": 2e-05,
+      "loss": 0.7771,
+      "mean_token_accuracy": 0.7658752902385906,
+      "step": 1000
+    },
+    {
+      "epoch": 0.7400555041628122,
+      "eval_loss": 0.8039809465408325,
+      "eval_mean_token_accuracy": 0.7569865406645019,
+      "eval_runtime": 28.2336,
+      "eval_samples_per_second": 9.103,
+      "eval_steps_per_second": 1.169,
+      "step": 1000
+    },
+    {
+      "epoch": 0.7437557816836263,
+      "grad_norm": 0.8295962810516357,
+      "learning_rate": 2e-05,
+      "loss": 0.8089,
+      "mean_token_accuracy": 0.7569489958101997,
+      "step": 1005
+    },
+    {
+      "epoch": 0.7474560592044404,
+      "grad_norm": 0.7815067768096924,
+      "learning_rate": 2e-05,
+      "loss": 0.7758,
+      "mean_token_accuracy": 0.7671153767210239,
+      "step": 1010
+    },
+    {
+      "epoch": 0.7511563367252544,
+      "grad_norm": 0.8126241564750671,
+      "learning_rate": 2e-05,
+      "loss": 0.7379,
+      "mean_token_accuracy": 0.7785501124249352,
+      "step": 1015
+    },
+    {
+      "epoch": 0.7548566142460684,
+      "grad_norm": 0.8162810802459717,
+      "learning_rate": 2e-05,
+      "loss": 0.8102,
+      "mean_token_accuracy": 0.7573220349482289,
+      "step": 1020
+    },
+    {
+      "epoch": 0.7585568917668826,
+      "grad_norm": 0.7573408484458923,
+      "learning_rate": 2e-05,
+      "loss": 0.7662,
+      "mean_token_accuracy": 0.7698151470774073,
+      "step": 1025
+    },
+    {
+      "epoch": 0.7622571692876966,
+      "grad_norm": 0.7962117791175842,
+      "learning_rate": 2e-05,
+      "loss": 0.7612,
+      "mean_token_accuracy": 0.7716137946938584,
+      "step": 1030
+    },
+    {
+      "epoch": 0.7659574468085106,
+      "grad_norm": 0.7962700724601746,
+      "learning_rate": 2e-05,
+      "loss": 0.7647,
+      "mean_token_accuracy": 0.7699719865120008,
+      "step": 1035
+    },
+    {
+      "epoch": 0.7696577243293247,
+      "grad_norm": 0.8017210960388184,
+      "learning_rate": 2e-05,
+      "loss": 0.7729,
+      "mean_token_accuracy": 0.7656573176756548,
+      "step": 1040
+    },
+    {
+      "epoch": 0.7733580018501388,
+      "grad_norm": 0.8346449136734009,
+      "learning_rate": 2e-05,
+      "loss": 0.7636,
+      "mean_token_accuracy": 0.7693838985258703,
+      "step": 1045
+    },
+    {
+      "epoch": 0.7770582793709528,
+      "grad_norm": 0.8304547667503357,
+      "learning_rate": 2e-05,
+      "loss": 0.7722,
+      "mean_token_accuracy": 0.7666433599313455,
+      "step": 1050
+    },
+    {
+      "epoch": 0.7807585568917669,
+      "grad_norm": 0.8001357913017273,
+      "learning_rate": 2e-05,
+      "loss": 0.7786,
+      "mean_token_accuracy": 0.7662119403935679,
+      "step": 1055
+    },
+    {
+      "epoch": 0.7844588344125809,
+      "grad_norm": 0.8312594294548035,
+      "learning_rate": 2e-05,
+      "loss": 0.7863,
+      "mean_token_accuracy": 0.7634551853341814,
+      "step": 1060
+    },
+    {
+      "epoch": 0.788159111933395,
+      "grad_norm": 0.8669179677963257,
+      "learning_rate": 2e-05,
+      "loss": 0.7696,
+      "mean_token_accuracy": 0.768203040960474,
+      "step": 1065
+    },
+    {
+      "epoch": 0.7918593894542091,
+      "grad_norm": 0.9093945026397705,
+      "learning_rate": 2e-05,
+      "loss": 0.7506,
+      "mean_token_accuracy": 0.7740881651757804,
+      "step": 1070
+    },
+    {
+      "epoch": 0.7955596669750231,
+      "grad_norm": 0.9110556244850159,
+      "learning_rate": 2e-05,
+      "loss": 0.7723,
+      "mean_token_accuracy": 0.768156867902089,
+      "step": 1075
+    },
+    {
+      "epoch": 0.7992599444958371,
+      "grad_norm": 0.8719515204429626,
+      "learning_rate": 2e-05,
+      "loss": 0.7487,
+      "mean_token_accuracy": 0.7738584815308094,
+      "step": 1080
+    },
+    {
+      "epoch": 0.8029602220166513,
+      "grad_norm": 0.855759859085083,
+      "learning_rate": 2e-05,
+      "loss": 0.8085,
+      "mean_token_accuracy": 0.7569202538917256,
+      "step": 1085
+    },
+    {
+      "epoch": 0.8066604995374653,
+      "grad_norm": 0.7778683304786682,
+      "learning_rate": 2e-05,
+      "loss": 0.8039,
+      "mean_token_accuracy": 0.7584921428765562,
+      "step": 1090
+    },
+    {
+      "epoch": 0.8103607770582794,
+      "grad_norm": 0.7796655893325806,
+      "learning_rate": 2e-05,
+      "loss": 0.7847,
+      "mean_token_accuracy": 0.7638983818899377,
+      "step": 1095
+    },
+    {
+      "epoch": 0.8140610545790934,
+      "grad_norm": 0.8250465989112854,
+      "learning_rate": 2e-05,
+      "loss": 0.7639,
+      "mean_token_accuracy": 0.7687881349384388,
+      "step": 1100
+    },
+    {
+      "epoch": 0.8140610545790934,
+      "eval_loss": 0.800523042678833,
+      "eval_mean_token_accuracy": 0.7575268288806666,
+      "eval_runtime": 28.2503,
+      "eval_samples_per_second": 9.097,
+      "eval_steps_per_second": 1.168,
+      "step": 1100
+    },
+    {
+      "epoch": 0.8177613320999075,
+      "grad_norm": 0.8301143646240234,
+      "learning_rate": 2e-05,
+      "loss": 0.7764,
+      "mean_token_accuracy": 0.7657188242099898,
+      "step": 1105
+    },
+    {
+      "epoch": 0.8214616096207216,
+      "grad_norm": 0.8132140040397644,
+      "learning_rate": 2e-05,
+      "loss": 0.7783,
+      "mean_token_accuracy": 0.7655952486721203,
+      "step": 1110
+    },
+    {
+      "epoch": 0.8251618871415356,
+      "grad_norm": 0.8092007637023926,
+      "learning_rate": 2e-05,
+      "loss": 0.7873,
+      "mean_token_accuracy": 0.7621894900380742,
+      "step": 1115
+    },
+    {
+      "epoch": 0.8288621646623496,
+      "grad_norm": 0.8420717120170593,
+      "learning_rate": 2e-05,
+      "loss": 0.7788,
+      "mean_token_accuracy": 0.7661002615281637,
+      "step": 1120
+    },
+    {
+      "epoch": 0.8325624421831638,
+      "grad_norm": 0.819154679775238,
+      "learning_rate": 2e-05,
+      "loss": 0.7834,
+      "mean_token_accuracy": 0.7640260223222127,
+      "step": 1125
+    },
+    {
+      "epoch": 0.8362627197039778,
+      "grad_norm": 0.7833020091056824,
+      "learning_rate": 2e-05,
+      "loss": 0.7673,
+      "mean_token_accuracy": 0.7671381666306092,
+      "step": 1130
+    },
+    {
+      "epoch": 0.8399629972247918,
+      "grad_norm": 0.8282708525657654,
+      "learning_rate": 2e-05,
+      "loss": 0.7663,
+      "mean_token_accuracy": 0.76869465856567,
+      "step": 1135
+    },
+    {
+      "epoch": 0.8436632747456059,
+      "grad_norm": 0.8674660325050354,
+      "learning_rate": 2e-05,
+      "loss": 0.8198,
+      "mean_token_accuracy": 0.7535738698791236,
+      "step": 1140
+    },
+    {
+      "epoch": 0.84736355226642,
+      "grad_norm": 0.8629141449928284,
+      "learning_rate": 2e-05,
+      "loss": 0.8062,
+      "mean_token_accuracy": 0.7576205181138813,
+      "step": 1145
+    },
+    {
+      "epoch": 0.851063829787234,
+      "grad_norm": 0.8082600235939026,
+      "learning_rate": 2e-05,
+      "loss": 0.7517,
+      "mean_token_accuracy": 0.7717635510287453,
+      "step": 1150
+    },
+    {
+      "epoch": 0.8547641073080481,
+      "grad_norm": 0.7617512941360474,
+      "learning_rate": 2e-05,
+      "loss": 0.7643,
+      "mean_token_accuracy": 0.7704166399970936,
+      "step": 1155
+    },
+    {
+      "epoch": 0.8584643848288621,
+      "grad_norm": 0.7678114771842957,
+      "learning_rate": 2e-05,
+      "loss": 0.7729,
+      "mean_token_accuracy": 0.7668813717759277,
+      "step": 1160
+    },
+    {
+      "epoch": 0.8621646623496763,
+      "grad_norm": 0.7924060821533203,
+      "learning_rate": 2e-05,
+      "loss": 0.7521,
+      "mean_token_accuracy": 0.7735818962967718,
+      "step": 1165
+    },
+    {
+      "epoch": 0.8658649398704903,
+      "grad_norm": 0.881769061088562,
+      "learning_rate": 2e-05,
+      "loss": 0.7859,
+      "mean_token_accuracy": 0.7643183082912849,
+      "step": 1170
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "grad_norm": 0.7947380542755127,
+      "learning_rate": 2e-05,
+      "loss": 0.7305,
+      "mean_token_accuracy": 0.7786316210966694,
+      "step": 1175
+    },
+    {
+      "epoch": 0.8732654949121184,
+      "grad_norm": 0.8072516918182373,
+      "learning_rate": 2e-05,
+      "loss": 0.7676,
+      "mean_token_accuracy": 0.7692856950594738,
+      "step": 1180
+    },
+    {
+      "epoch": 0.8769657724329325,
+      "grad_norm": 0.8792457580566406,
+      "learning_rate": 2e-05,
+      "loss": 0.7984,
+      "mean_token_accuracy": 0.7623190074905516,
+      "step": 1185
+    },
+    {
+      "epoch": 0.8806660499537465,
+      "grad_norm": 0.8246925473213196,
+      "learning_rate": 2e-05,
+      "loss": 0.7604,
+      "mean_token_accuracy": 0.7710242769603142,
+      "step": 1190
+    },
+    {
+      "epoch": 0.8843663274745606,
+      "grad_norm": 0.8517448902130127,
+      "learning_rate": 2e-05,
+      "loss": 0.8127,
+      "mean_token_accuracy": 0.7556866127163658,
+      "step": 1195
+    },
+    {
+      "epoch": 0.8880666049953746,
+      "grad_norm": 0.8204156160354614,
+      "learning_rate": 2e-05,
+      "loss": 0.7559,
+      "mean_token_accuracy": 0.7712058797842442,
+      "step": 1200
+    },
+    {
+      "epoch": 0.8880666049953746,
+      "eval_loss": 0.797502875328064,
+      "eval_mean_token_accuracy": 0.7579622750028506,
+      "eval_runtime": 28.4023,
+      "eval_samples_per_second": 9.049,
+      "eval_steps_per_second": 1.162,
+      "step": 1200
+    },
+    {
+      "epoch": 0.8917668825161887,
+      "grad_norm": 0.8131943345069885,
+      "learning_rate": 2e-05,
+      "loss": 0.7571,
+      "mean_token_accuracy": 0.7733840665520189,
+      "step": 1205
+    },
+    {
+      "epoch": 0.8954671600370028,
+      "grad_norm": 0.73580002784729,
+      "learning_rate": 2e-05,
+      "loss": 0.7727,
+      "mean_token_accuracy": 0.7674258088788803,
+      "step": 1210
+    },
+    {
+      "epoch": 0.8991674375578168,
+      "grad_norm": 0.7590300440788269,
+      "learning_rate": 2e-05,
+      "loss": 0.7637,
+      "mean_token_accuracy": 0.7691722572737629,
+      "step": 1215
+    },
+    {
+      "epoch": 0.902867715078631,
+      "grad_norm": 0.7864495515823364,
+      "learning_rate": 2e-05,
+      "loss": 0.794,
+      "mean_token_accuracy": 0.7610983305690475,
+      "step": 1220
+    },
+    {
+      "epoch": 0.906567992599445,
+      "grad_norm": 0.7783846259117126,
+      "learning_rate": 2e-05,
+      "loss": 0.7698,
+      "mean_token_accuracy": 0.7677923415686079,
+      "step": 1225
+    },
+    {
+      "epoch": 0.910268270120259,
+      "grad_norm": 0.8439836502075195,
+      "learning_rate": 2e-05,
+      "loss": 0.8021,
+      "mean_token_accuracy": 0.7596970516172707,
+      "step": 1230
+    },
+    {
+      "epoch": 0.913968547641073,
+      "grad_norm": 0.875133752822876,
+      "learning_rate": 2e-05,
+      "loss": 0.7796,
+      "mean_token_accuracy": 0.7652020678902339,
+      "step": 1235
+    },
+    {
+      "epoch": 0.9176688251618872,
+      "grad_norm": 0.7731521725654602,
+      "learning_rate": 2e-05,
+      "loss": 0.7595,
+      "mean_token_accuracy": 0.7703574952610036,
+      "step": 1240
+    },
+    {
+      "epoch": 0.9213691026827012,
+      "grad_norm": 0.7407891750335693,
+      "learning_rate": 2e-05,
+      "loss": 0.7665,
+      "mean_token_accuracy": 0.7682582461530055,
+      "step": 1245
+    },
+    {
+      "epoch": 0.9250693802035153,
+      "grad_norm": 0.790274977684021,
+      "learning_rate": 2e-05,
+      "loss": 0.7489,
+      "mean_token_accuracy": 0.7736910322129121,
+      "step": 1250
+    },
+    {
+      "epoch": 0.9287696577243293,
+      "grad_norm": 0.81461501121521,
+      "learning_rate": 2e-05,
+      "loss": 0.7741,
+      "mean_token_accuracy": 0.7667225211713106,
+      "step": 1255
+    },
+    {
+      "epoch": 0.9324699352451434,
+      "grad_norm": 0.7998383045196533,
+      "learning_rate": 2e-05,
+      "loss": 0.7807,
+      "mean_token_accuracy": 0.7625351777097331,
+      "step": 1260
+    },
+    {
+      "epoch": 0.9361702127659575,
+      "grad_norm": 0.8241930603981018,
+      "learning_rate": 2e-05,
+      "loss": 0.8072,
+      "mean_token_accuracy": 0.7569770056683507,
+      "step": 1265
+    },
+    {
+      "epoch": 0.9398704902867715,
+      "grad_norm": 0.7618784308433533,
+      "learning_rate": 2e-05,
+      "loss": 0.7772,
+      "mean_token_accuracy": 0.7643117414360313,
+      "step": 1270
+    },
+    {
+      "epoch": 0.9435707678075855,
+      "grad_norm": 0.7704834938049316,
+      "learning_rate": 2e-05,
+      "loss": 0.7789,
+      "mean_token_accuracy": 0.7653304006081735,
+      "step": 1275
+    },
+    {
+      "epoch": 0.9472710453283997,
+      "grad_norm": 0.7337019443511963,
+      "learning_rate": 2e-05,
+      "loss": 0.772,
+      "mean_token_accuracy": 0.7669845879245384,
+      "step": 1280
+    },
+    {
+      "epoch": 0.9509713228492137,
+      "grad_norm": 0.8250733017921448,
+      "learning_rate": 2e-05,
+      "loss": 0.7785,
+      "mean_token_accuracy": 0.7657440636307564,
+      "step": 1285
+    },
+    {
+      "epoch": 0.9546716003700277,
+      "grad_norm": 0.839878499507904,
+      "learning_rate": 2e-05,
+      "loss": 0.8066,
+      "mean_token_accuracy": 0.7591835950929369,
+      "step": 1290
+    },
+    {
+      "epoch": 0.9583718778908418,
+      "grad_norm": 0.8156315088272095,
+      "learning_rate": 2e-05,
+      "loss": 0.7704,
+      "mean_token_accuracy": 0.7668827112901107,
+      "step": 1295
+    },
+    {
+      "epoch": 0.9620721554116559,
+      "grad_norm": 0.8538619875907898,
+      "learning_rate": 2e-05,
+      "loss": 0.7589,
+      "mean_token_accuracy": 0.7709486799753794,
+      "step": 1300
+    },
+    {
+      "epoch": 0.9620721554116559,
+      "eval_loss": 0.7957349419593811,
+      "eval_mean_token_accuracy": 0.7585383150980717,
+      "eval_runtime": 28.3961,
+      "eval_samples_per_second": 9.051,
+      "eval_steps_per_second": 1.162,
+      "step": 1300
+    },
+    {
+      "epoch": 0.96577243293247,
+      "grad_norm": 0.8585022687911987,
+      "learning_rate": 2e-05,
+      "loss": 0.7955,
+      "mean_token_accuracy": 0.7603014430216614,
+      "step": 1305
+    },
+    {
+      "epoch": 0.969472710453284,
+      "grad_norm": 0.8347756862640381,
+      "learning_rate": 2e-05,
+      "loss": 0.8013,
+      "mean_token_accuracy": 0.7593268261742192,
+      "step": 1310
+    },
+    {
+      "epoch": 0.973172987974098,
+      "grad_norm": 0.7872015833854675,
+      "learning_rate": 2e-05,
+      "loss": 0.7683,
+      "mean_token_accuracy": 0.7696177228132353,
+      "step": 1315
+    },
+    {
+      "epoch": 0.9768732654949122,
+      "grad_norm": 0.8233494162559509,
+      "learning_rate": 2e-05,
+      "loss": 0.7706,
+      "mean_token_accuracy": 0.7678653349992682,
+      "step": 1320
+    },
+    {
+      "epoch": 0.9805735430157262,
+      "grad_norm": 0.8023855090141296,
+      "learning_rate": 2e-05,
+      "loss": 0.7783,
+      "mean_token_accuracy": 0.7656580825758337,
+      "step": 1325
+    },
+    {
+      "epoch": 0.9842738205365402,
+      "grad_norm": 0.8429985642433167,
+      "learning_rate": 2e-05,
+      "loss": 0.7669,
+      "mean_token_accuracy": 0.7683449900233187,
+      "step": 1330
+    },
+    {
+      "epoch": 0.9879740980573543,
+      "grad_norm": 0.792027473449707,
+      "learning_rate": 2e-05,
+      "loss": 0.767,
+      "mean_token_accuracy": 0.7692297252303795,
+      "step": 1335
+    },
+    {
+      "epoch": 0.9916743755781684,
+      "grad_norm": 0.8096156120300293,
+      "learning_rate": 2e-05,
+      "loss": 0.7841,
+      "mean_token_accuracy": 0.7638673896787455,
+      "step": 1340
+    },
+    {
+      "epoch": 0.9953746530989824,
+      "grad_norm": 0.9265767931938171,
+      "learning_rate": 2e-05,
+      "loss": 0.771,
+      "mean_token_accuracy": 0.7675258668736931,
+      "step": 1345
+    },
+    {
+      "epoch": 0.9990749306197965,
+      "grad_norm": 0.8164911866188049,
+      "learning_rate": 2e-05,
+      "loss": 0.7838,
+      "mean_token_accuracy": 0.7645547380257296,
+      "step": 1350
+    },
+    {
+      "epoch": 0.9998149861239592,
+      "mean_token_accuracy": 0.7587965060316426,
+      "step": 1351,
+      "total_flos": 76959556042752.0,
+      "train_loss": 0.8068137678457665,
+      "train_runtime": 18078.9101,
+      "train_samples_per_second": 2.392,
+      "train_steps_per_second": 0.075
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1351,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 76959556042752.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d60e2086dd88927d46a3ffa1d5ec1ab404d991d69e1288db51a9e380ef723a6
+size 8632