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 +1,68 @@
+---
+base_model: Qwen/Qwen2.5-1.5B-Instruct
+library_name: transformers
+model_name: Qwen2.5-1.5B-Open-R1-GRPO
+tags:
+- generated_from_trainer
+- trl
+- grpo
+licence: license
+---
+
+# Model Card for Qwen2.5-1.5B-Open-R1-GRPO
+
+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="Kurt232/Qwen2.5-1.5B-Open-R1-GRPO", 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/wenjie-du-hong-kong-university-of-science-and-technology/huggingface/runs/1acyst2a) 
+
+
+This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
+
+### Framework versions
+
+- TRL: 0.16.0.dev0
+- Transformers: 4.50.0.dev0
+- Pytorch: 2.5.1
+- Datasets: 3.2.0
+- Tokenizers: 0.21.0
+
+## Citations
+
+Cite GRPO as:
+
+```bibtex
+@article{zhihong2024deepseekmath,
+    title        = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
+    author       = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
+    year         = 2024,
+    eprint       = {arXiv:2402.03300},
+}
+
+```
+
+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": 0.0,
+    "train_loss": 2.1605342640346974,
+    "train_runtime": 168649.3171,
+    "train_samples": 72441,
+    "train_samples_per_second": 0.43,
+    "train_steps_per_second": 0.003
+}

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": 32768,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.50.0.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": "4.50.0.dev0"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c78eb42c4a1b1efdea95ef058faa54c8756553e3229f2dd200e41af683ecfd03
+size 3554214752

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "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": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5eee858c5123a4279c3e1f7b81247343f356ac767940b2692a928ad929543214
+size 11422063

tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "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": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 0.0,
+    "train_loss": 2.1605342640346974,
+    "train_runtime": 168649.3171,
+    "train_samples": 72441,
+    "train_samples_per_second": 0.43,
+    "train_steps_per_second": 0.003
+}

trainer_state.json

@@ -0,0 +1,1581 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9982883330572581,
+  "eval_steps": 100,
+  "global_step": 565,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "completion_length": 395.45969400405886,
+      "epoch": 0.008834410027055381,
+      "grad_norm": 0.7056356072425842,
+      "kl": 0.00015916824340820313,
+      "learning_rate": 1.7543859649122807e-06,
+      "loss": 0.0,
+      "reward": 0.6100000126287342,
+      "reward_std": 0.42495151571929457,
+      "rewards/accuracy_reward": 0.14562500298488884,
+      "rewards/format_reward": 0.4643750095041469,
+      "step": 5
+    },
+    {
+      "completion_length": 288.1106297492981,
+      "epoch": 0.017668820054110763,
+      "grad_norm": 10.0836763381958,
+      "kl": 0.0852059006690979,
+      "learning_rate": 3.5087719298245615e-06,
+      "loss": 0.0034,
+      "reward": 0.8221875139512121,
+      "reward_std": 0.35551070063374934,
+      "rewards/accuracy_reward": 0.11125000189058483,
+      "rewards/format_reward": 0.7109375111758709,
+      "step": 10
+    },
+    {
+      "completion_length": 220.40937814712524,
+      "epoch": 0.026503230081166142,
+      "grad_norm": 0.4689132273197174,
+      "kl": 0.04801025390625,
+      "learning_rate": 5.263157894736842e-06,
+      "loss": 0.0019,
+      "reward": 0.988125017657876,
+      "reward_std": 0.2385118653997779,
+      "rewards/accuracy_reward": 0.07843750123865903,
+      "rewards/format_reward": 0.9096875101327896,
+      "step": 15
+    },
+    {
+      "completion_length": 203.07937822341918,
+      "epoch": 0.035337640108221526,
+      "grad_norm": 0.4062890112400055,
+      "kl": 0.03817596435546875,
+      "learning_rate": 7.017543859649123e-06,
+      "loss": 0.0015,
+      "reward": 1.0215625200420617,
+      "reward_std": 0.2450850388035178,
+      "rewards/accuracy_reward": 0.10656250198371708,
+      "rewards/format_reward": 0.9150000106543302,
+      "step": 20
+    },
+    {
+      "completion_length": 164.8200032711029,
+      "epoch": 0.044172050135276905,
+      "grad_norm": 1.4767407178878784,
+      "kl": 0.05220718383789062,
+      "learning_rate": 8.771929824561405e-06,
+      "loss": 0.0021,
+      "reward": 1.058125016838312,
+      "reward_std": 0.2080313797108829,
+      "rewards/accuracy_reward": 0.10625000225845724,
+      "rewards/format_reward": 0.9518750045448542,
+      "step": 25
+    },
+    {
+      "completion_length": 205.17625331878662,
+      "epoch": 0.053006460162332285,
+      "grad_norm": 0.3233347535133362,
+      "kl": 0.06426353454589843,
+      "learning_rate": 1.0526315789473684e-05,
+      "loss": 0.0026,
+      "reward": 1.1231250219047069,
+      "reward_std": 0.24855279684998094,
+      "rewards/accuracy_reward": 0.16875000279396773,
+      "rewards/format_reward": 0.9543750032782554,
+      "step": 30
+    },
+    {
+      "completion_length": 254.10000500679016,
+      "epoch": 0.061840870189387664,
+      "grad_norm": 0.3409097194671631,
+      "kl": 0.07096328735351562,
+      "learning_rate": 1.2280701754385966e-05,
+      "loss": 0.0028,
+      "reward": 1.103750016912818,
+      "reward_std": 0.33230566498823466,
+      "rewards/accuracy_reward": 0.20093750341329725,
+      "rewards/format_reward": 0.9028125114738941,
+      "step": 35
+    },
+    {
+      "completion_length": 285.9718804359436,
+      "epoch": 0.07067528021644305,
+      "grad_norm": 0.3942728042602539,
+      "kl": 0.08272323608398438,
+      "learning_rate": 1.4035087719298246e-05,
+      "loss": 0.0033,
+      "reward": 1.1353125210851431,
+      "reward_std": 0.38983988179825246,
+      "rewards/accuracy_reward": 0.24750000431668012,
+      "rewards/format_reward": 0.8878125142306089,
+      "step": 40
+    },
+    {
+      "completion_length": 232.73781714439392,
+      "epoch": 0.07950969024349842,
+      "grad_norm": 0.3255186080932617,
+      "kl": 0.07653045654296875,
+      "learning_rate": 1.578947368421053e-05,
+      "loss": 0.0031,
+      "reward": 1.1790625225752591,
+      "reward_std": 0.28494237596169114,
+      "rewards/accuracy_reward": 0.22125000345986337,
+      "rewards/format_reward": 0.9578125022351742,
+      "step": 45
+    },
+    {
+      "completion_length": 247.60344228744506,
+      "epoch": 0.08834410027055381,
+      "grad_norm": 0.2690398395061493,
+      "kl": 0.087286376953125,
+      "learning_rate": 1.754385964912281e-05,
+      "loss": 0.0035,
+      "reward": 1.229687524214387,
+      "reward_std": 0.2820313207805157,
+      "rewards/accuracy_reward": 0.2587500046938658,
+      "rewards/format_reward": 0.9709374994039536,
+      "step": 50
+    },
+    {
+      "completion_length": 333.2628190994263,
+      "epoch": 0.09717851029760918,
+      "grad_norm": 0.2579941153526306,
+      "kl": 0.108929443359375,
+      "learning_rate": 1.929824561403509e-05,
+      "loss": 0.0044,
+      "reward": 1.2840625256299973,
+      "reward_std": 0.30264704609289766,
+      "rewards/accuracy_reward": 0.3265625053783879,
+      "rewards/format_reward": 0.9575000021606683,
+      "step": 55
+    },
+    {
+      "completion_length": 417.37250623703005,
+      "epoch": 0.10601292032466457,
+      "grad_norm": 3.9840168952941895,
+      "kl": 8.018402099609375,
+      "learning_rate": 1.9998279033654883e-05,
+      "loss": 0.3202,
+      "reward": 1.128125017695129,
+      "reward_std": 0.4620271602179855,
+      "rewards/accuracy_reward": 0.34187500597909093,
+      "rewards/format_reward": 0.786250009573996,
+      "step": 60
+    },
+    {
+      "completion_length": 461.0903203010559,
+      "epoch": 0.11484733035171994,
+      "grad_norm": 1054268.875,
+      "kl": 5414.609799194336,
+      "learning_rate": 1.9987764162142615e-05,
+      "loss": 216.3807,
+      "reward": 0.991250017285347,
+      "reward_std": 0.5101292508654296,
+      "rewards/accuracy_reward": 0.3278125050244853,
+      "rewards/format_reward": 0.6634375125169754,
+      "step": 65
+    },
+    {
+      "completion_length": 311.6025058746338,
+      "epoch": 0.12368174037877533,
+      "grad_norm": 2.1096460819244385,
+      "kl": 0.6299407958984375,
+      "learning_rate": 1.996770055201962e-05,
+      "loss": 0.0252,
+      "reward": 1.098750017210841,
+      "reward_std": 0.4555227609351277,
+      "rewards/accuracy_reward": 0.3271875062957406,
+      "rewards/format_reward": 0.771562509983778,
+      "step": 70
+    },
+    {
+      "completion_length": 235.89250383377075,
+      "epoch": 0.1325161504058307,
+      "grad_norm": 6.618839263916016,
+      "kl": 0.4949859619140625,
+      "learning_rate": 1.9938107384973165e-05,
+      "loss": 0.0198,
+      "reward": 1.249375019967556,
+      "reward_std": 0.3586165915243328,
+      "rewards/accuracy_reward": 0.34125000652857124,
+      "rewards/format_reward": 0.9081250090152025,
+      "step": 75
+    },
+    {
+      "completion_length": 250.77844104766845,
+      "epoch": 0.1413505604328861,
+      "grad_norm": 1.8069761991500854,
+      "kl": 0.939288330078125,
+      "learning_rate": 1.9899012953363002e-05,
+      "loss": 0.0376,
+      "reward": 1.1634375214576722,
+      "reward_std": 0.43803766365163027,
+      "rewards/accuracy_reward": 0.3309375068871304,
+      "rewards/format_reward": 0.8325000129640102,
+      "step": 80
+    },
+    {
+      "completion_length": 245.42875361442566,
+      "epoch": 0.15018497045994147,
+      "grad_norm": 0.7974065542221069,
+      "kl": 1.3499786376953125,
+      "learning_rate": 1.9850454633172632e-05,
+      "loss": 0.054,
+      "reward": 1.121250021085143,
+      "reward_std": 0.41360275098122656,
+      "rewards/accuracy_reward": 0.2906250062864274,
+      "rewards/format_reward": 0.8306250110268593,
+      "step": 85
+    },
+    {
+      "completion_length": 213.91344094276428,
+      "epoch": 0.15901938048699685,
+      "grad_norm": 0.6790722012519836,
+      "kl": 0.401104736328125,
+      "learning_rate": 1.979247884827625e-05,
+      "loss": 0.016,
+      "reward": 1.261875021830201,
+      "reward_std": 0.2916932702064514,
+      "rewards/accuracy_reward": 0.31187500653322786,
+      "rewards/format_reward": 0.9500000022351742,
+      "step": 90
+    },
+    {
+      "completion_length": 273.06938033103944,
+      "epoch": 0.16785379051405222,
+      "grad_norm": 19.757688522338867,
+      "kl": 7.057659912109375,
+      "learning_rate": 1.9725141026055473e-05,
+      "loss": 0.2824,
+      "reward": 1.1515625223517418,
+      "reward_std": 0.37032804545015097,
+      "rewards/accuracy_reward": 0.27531250528991225,
+      "rewards/format_reward": 0.876250010728836,
+      "step": 95
+    },
+    {
+      "completion_length": 242.15219221115112,
+      "epoch": 0.17668820054110762,
+      "grad_norm": 1.281561017036438,
+      "kl": 1.34007568359375,
+      "learning_rate": 1.9648505544408343e-05,
+      "loss": 0.0536,
+      "reward": 0.9990625161677599,
+      "reward_std": 0.4486269262619317,
+      "rewards/accuracy_reward": 0.19250000277534127,
+      "rewards/format_reward": 0.8065625119954347,
+      "step": 100
+    },
+    {
+      "epoch": 0.17668820054110762,
+      "eval_completion_length": 160.48298971993583,
+      "eval_kl": 1.2287946428571428,
+      "eval_loss": 0.05083000287413597,
+      "eval_reward": 1.112500011920929,
+      "eval_reward_std": 0.34407747643334524,
+      "eval_rewards/accuracy_reward": 0.21428571960755757,
+      "eval_rewards/format_reward": 0.8982142976352147,
+      "eval_runtime": 36.089,
+      "eval_samples_per_second": 2.743,
+      "eval_steps_per_second": 0.055,
+      "step": 100
+    },
+    {
+      "completion_length": 155.20687766075133,
+      "epoch": 0.185522610568163,
+      "grad_norm": 0.5366517901420593,
+      "kl": 0.703033447265625,
+      "learning_rate": 1.9562645670201278e-05,
+      "loss": 0.0281,
+      "reward": 1.1596875220537186,
+      "reward_std": 0.2591133718378842,
+      "rewards/accuracy_reward": 0.20656250272877513,
+      "rewards/format_reward": 0.9531250018626451,
+      "step": 105
+    },
+    {
+      "completion_length": 219.61187934875488,
+      "epoch": 0.19435702059521837,
+      "grad_norm": 6.434405326843262,
+      "kl": 37.474554443359374,
+      "learning_rate": 1.9467643489222704e-05,
+      "loss": 1.5045,
+      "reward": 1.1209375146776437,
+      "reward_std": 0.32721211602911354,
+      "rewards/accuracy_reward": 0.21062500362750142,
+      "rewards/format_reward": 0.9103125061839819,
+      "step": 110
+    },
+    {
+      "completion_length": 297.05250535011294,
+      "epoch": 0.20319143062227377,
+      "grad_norm": 0.3138395845890045,
+      "kl": 51.60116882324219,
+      "learning_rate": 1.9363589827705494e-05,
+      "loss": 2.0679,
+      "reward": 0.9950000163167715,
+      "reward_std": 0.4325931371189654,
+      "rewards/accuracy_reward": 0.17125000276137142,
+      "rewards/format_reward": 0.823750014975667,
+      "step": 115
+    },
+    {
+      "completion_length": 214.70187797546387,
+      "epoch": 0.21202584064932914,
+      "grad_norm": 0.2661781311035156,
+      "kl": 0.1655609130859375,
+      "learning_rate": 1.9250584165493102e-05,
+      "loss": 0.0066,
+      "reward": 1.2171875227242708,
+      "reward_std": 0.2870821573305875,
+      "rewards/accuracy_reward": 0.27031250488944353,
+      "rewards/format_reward": 0.9468750044703483,
+      "step": 120
+    },
+    {
+      "completion_length": 250.79937930107116,
+      "epoch": 0.2208602506763845,
+      "grad_norm": 0.2946310043334961,
+      "kl": 0.1904541015625,
+      "learning_rate": 1.9128734540932494e-05,
+      "loss": 0.0076,
+      "reward": 1.2265625208616258,
+      "reward_std": 0.321342809824273,
+      "rewards/accuracy_reward": 0.29343750413972886,
+      "rewards/format_reward": 0.9331250052899123,
+      "step": 125
+    },
+    {
+      "completion_length": 325.37844333648684,
+      "epoch": 0.22969466070343988,
+      "grad_norm": 0.2663358747959137,
+      "kl": 0.13707275390625,
+      "learning_rate": 1.899815744758478e-05,
+      "loss": 0.0055,
+      "reward": 1.1243750214576722,
+      "reward_std": 0.43622313924133777,
+      "rewards/accuracy_reward": 0.28187500485219064,
+      "rewards/format_reward": 0.8425000134855509,
+      "step": 130
+    },
+    {
+      "completion_length": 245.49687938690187,
+      "epoch": 0.23852907073049529,
+      "grad_norm": 0.27225080132484436,
+      "kl": 0.1351104736328125,
+      "learning_rate": 1.8858977722852273e-05,
+      "loss": 0.0054,
+      "reward": 1.230937520414591,
+      "reward_std": 0.2990158367436379,
+      "rewards/accuracy_reward": 0.286875005136244,
+      "rewards/format_reward": 0.9440625056624412,
+      "step": 135
+    },
+    {
+      "completion_length": 335.2887560844421,
+      "epoch": 0.24736348075755066,
+      "grad_norm": 0.27616289258003235,
+      "kl": 0.1483642578125,
+      "learning_rate": 1.8711328428628492e-05,
+      "loss": 0.0059,
+      "reward": 1.179062517732382,
+      "reward_std": 0.39579583145678043,
+      "rewards/accuracy_reward": 0.2965625042794272,
+      "rewards/format_reward": 0.8825000107288361,
+      "step": 140
+    },
+    {
+      "completion_length": 297.70031728744505,
+      "epoch": 0.25619789078460603,
+      "grad_norm": 0.2805606722831726,
+      "kl": 0.1417388916015625,
+      "learning_rate": 1.855535072408516e-05,
+      "loss": 0.0057,
+      "reward": 1.2018750231713056,
+      "reward_std": 0.3620848820544779,
+      "rewards/accuracy_reward": 0.2940625047776848,
+      "rewards/format_reward": 0.9078125089406968,
+      "step": 145
+    },
+    {
+      "completion_length": 324.3646910190582,
+      "epoch": 0.2650323008116614,
+      "grad_norm": 2.052229404449463,
+      "kl": 0.1634857177734375,
+      "learning_rate": 1.839119373071791e-05,
+      "loss": 0.0065,
+      "reward": 1.1421875178813934,
+      "reward_std": 0.4344090681057423,
+      "rewards/accuracy_reward": 0.276250005629845,
+      "rewards/format_reward": 0.8659375131130218,
+      "step": 150
+    },
+    {
+      "completion_length": 390.25656843185425,
+      "epoch": 0.2738667108387168,
+      "grad_norm": 17.338764190673828,
+      "kl": 0.8824493408203125,
+      "learning_rate": 1.8219014389779586e-05,
+      "loss": 0.0353,
+      "reward": 1.0956250216811896,
+      "reward_std": 0.4632612488232553,
+      "rewards/accuracy_reward": 0.2831250050337985,
+      "rewards/format_reward": 0.812500012293458,
+      "step": 155
+    },
+    {
+      "completion_length": 339.919068813324,
+      "epoch": 0.2827011208657722,
+      "grad_norm": 0.5709920525550842,
+      "kl": 0.8438385009765625,
+      "learning_rate": 1.8038977312237583e-05,
+      "loss": 0.0339,
+      "reward": 1.1243750188499688,
+      "reward_std": 0.44248684560880064,
+      "rewards/accuracy_reward": 0.2725000043865293,
+      "rewards/format_reward": 0.8518750129267574,
+      "step": 160
+    },
+    {
+      "completion_length": 225.35000352859498,
+      "epoch": 0.2915355308928276,
+      "grad_norm": 1.9149246215820312,
+      "kl": 1.168243408203125,
+      "learning_rate": 1.785125462139855e-05,
+      "loss": 0.0467,
+      "reward": 1.1393750187009573,
+      "reward_std": 0.3354927910026163,
+      "rewards/accuracy_reward": 0.23656250403728335,
+      "rewards/format_reward": 0.9028125092387199,
+      "step": 165
+    },
+    {
+      "completion_length": 265.2453175067902,
+      "epoch": 0.30036994091988295,
+      "grad_norm": 1.2403807640075684,
+      "kl": 2.22581787109375,
+      "learning_rate": 1.765602578835102e-05,
+      "loss": 0.0891,
+      "reward": 1.0812500189989804,
+      "reward_std": 0.398513315198943,
+      "rewards/accuracy_reward": 0.21593750319443644,
+      "rewards/format_reward": 0.8653125107288361,
+      "step": 170
+    },
+    {
+      "completion_length": 162.558753156662,
+      "epoch": 0.3092043509469383,
+      "grad_norm": 0.8905084729194641,
+      "kl": 1.0909912109375,
+      "learning_rate": 1.745347746038319e-05,
+      "loss": 0.0437,
+      "reward": 1.1509375225752592,
+      "reward_std": 0.2991133214440197,
+      "rewards/accuracy_reward": 0.2137500040233135,
+      "rewards/format_reward": 0.9371875058859587,
+      "step": 175
+    },
+    {
+      "completion_length": 151.561878490448,
+      "epoch": 0.3180387609739937,
+      "grad_norm": 0.7335211634635925,
+      "kl": 1.7427215576171875,
+      "learning_rate": 1.724380328253998e-05,
+      "loss": 0.0698,
+      "reward": 1.2050000198185444,
+      "reward_std": 0.27624500943347813,
+      "rewards/accuracy_reward": 0.24375000514555722,
+      "rewards/format_reward": 0.9612500034272671,
+      "step": 180
+    },
+    {
+      "completion_length": 214.4893784046173,
+      "epoch": 0.32687317100104907,
+      "grad_norm": 1.8747870922088623,
+      "kl": 0.5605438232421875,
+      "learning_rate": 1.7027203712489902e-05,
+      "loss": 0.0224,
+      "reward": 1.1746875189244748,
+      "reward_std": 0.3283426255453378,
+      "rewards/accuracy_reward": 0.2471875047776848,
+      "rewards/format_reward": 0.92750000692904,
+      "step": 185
+    },
+    {
+      "completion_length": 246.84844212532045,
+      "epoch": 0.33570758102810444,
+      "grad_norm": 0.9456897974014282,
+      "kl": 1.3440460205078124,
+      "learning_rate": 1.6803885828878798e-05,
+      "loss": 0.0538,
+      "reward": 1.1721875224262477,
+      "reward_std": 0.35966599816456435,
+      "rewards/accuracy_reward": 0.2609375044237822,
+      "rewards/format_reward": 0.9112500131130219,
+      "step": 190
+    },
+    {
+      "completion_length": 256.84125423431396,
+      "epoch": 0.34454199105515987,
+      "grad_norm": 7.396022796630859,
+      "kl": 2.101116943359375,
+      "learning_rate": 1.657406313335358e-05,
+      "loss": 0.0841,
+      "reward": 1.1690625209361314,
+      "reward_std": 0.3737951037939638,
+      "rewards/accuracy_reward": 0.26375000465195625,
+      "rewards/format_reward": 0.9053125131875277,
+      "step": 195
+    },
+    {
+      "completion_length": 293.7046928405762,
+      "epoch": 0.35337640108221524,
+      "grad_norm": 5.65695858001709,
+      "kl": 312.5613220214844,
+      "learning_rate": 1.633795534644538e-05,
+      "loss": 12.4926,
+      "reward": 1.0831250209361314,
+      "reward_std": 0.44232793282717464,
+      "rewards/accuracy_reward": 0.24312500432133674,
+      "rewards/format_reward": 0.8400000127032399,
+      "step": 200
+    },
+    {
+      "epoch": 0.35337640108221524,
+      "eval_completion_length": 322.5714329310826,
+      "eval_kl": 1.8136160714285714,
+      "eval_loss": 0.07279199361801147,
+      "eval_reward": 1.0625000085149492,
+      "eval_reward_std": 0.4705401488712856,
+      "eval_rewards/accuracy_reward": 0.25535714839185986,
+      "eval_rewards/format_reward": 0.8071428707667759,
+      "eval_runtime": 61.7397,
+      "eval_samples_per_second": 1.604,
+      "eval_steps_per_second": 0.032,
+      "step": 200
+    },
+    {
+      "completion_length": 341.8271938323975,
+      "epoch": 0.3622108111092706,
+      "grad_norm": 3.2196381092071533,
+      "kl": 3.080609130859375,
+      "learning_rate": 1.609578819750708e-05,
+      "loss": 0.1232,
+      "reward": 1.0606250196695328,
+      "reward_std": 0.47473005009815095,
+      "rewards/accuracy_reward": 0.2515625037252903,
+      "rewards/format_reward": 0.8090625129640102,
+      "step": 205
+    },
+    {
+      "completion_length": 320.10562963485717,
+      "epoch": 0.371045221136326,
+      "grad_norm": 0.5747339725494385,
+      "kl": 0.7265289306640625,
+      "learning_rate": 1.5847793208906228e-05,
+      "loss": 0.0291,
+      "reward": 1.2040625192224979,
+      "reward_std": 0.41384371900931,
+      "rewards/accuracy_reward": 0.3168750056996942,
+      "rewards/format_reward": 0.8871875137090683,
+      "step": 210
+    },
+    {
+      "completion_length": 269.2675045013428,
+      "epoch": 0.37987963116338136,
+      "grad_norm": 0.7235865592956543,
+      "kl": 0.3179962158203125,
+      "learning_rate": 1.5594207474679533e-05,
+      "loss": 0.0127,
+      "reward": 1.2943750232458116,
+      "reward_std": 0.33030168660916387,
+      "rewards/accuracy_reward": 0.3456250076182187,
+      "rewards/format_reward": 0.9487500034272671,
+      "step": 215
+    },
+    {
+      "completion_length": 261.3346913814545,
+      "epoch": 0.38871404119043673,
+      "grad_norm": 1.222240924835205,
+      "kl": 0.407818603515625,
+      "learning_rate": 1.533527343386062e-05,
+      "loss": 0.0163,
+      "reward": 1.2256250228732823,
+      "reward_std": 0.325825364748016,
+      "rewards/accuracy_reward": 0.2937500055413693,
+      "rewards/format_reward": 0.931875005736947,
+      "step": 220
+    },
+    {
+      "completion_length": 305.0543807029724,
+      "epoch": 0.3975484512174921,
+      "grad_norm": 0.40610432624816895,
+      "kl": 0.1676788330078125,
+      "learning_rate": 1.5071238638697731e-05,
+      "loss": 0.0067,
+      "reward": 1.26843752078712,
+      "reward_std": 0.3284678528551012,
+      "rewards/accuracy_reward": 0.3296875049360096,
+      "rewards/format_reward": 0.9387500051409006,
+      "step": 225
+    },
+    {
+      "completion_length": 406.37750749588014,
+      "epoch": 0.40638286124454753,
+      "grad_norm": 30.780855178833008,
+      "kl": 0.97349853515625,
+      "learning_rate": 1.4802355517982956e-05,
+      "loss": 0.0389,
+      "reward": 1.1640625223517418,
+      "reward_std": 0.4788378866389394,
+      "rewards/accuracy_reward": 0.33156250438187274,
+      "rewards/format_reward": 0.8325000109151006,
+      "step": 230
+    },
+    {
+      "completion_length": 430.0178194999695,
+      "epoch": 0.4152172712716029,
+      "grad_norm": 0.38293465971946716,
+      "kl": 0.36756591796875,
+      "learning_rate": 1.452888113571929e-05,
+      "loss": 0.0147,
+      "reward": 1.1506250152364372,
+      "reward_std": 0.4521612086333334,
+      "rewards/accuracy_reward": 0.3231250060489401,
+      "rewards/format_reward": 0.8275000104680658,
+      "step": 235
+    },
+    {
+      "completion_length": 295.6890670776367,
+      "epoch": 0.4240516812986583,
+      "grad_norm": 0.8215426802635193,
+      "kl": 0.2094635009765625,
+      "learning_rate": 1.4251076945356233e-05,
+      "loss": 0.0084,
+      "reward": 1.27562501989305,
+      "reward_std": 0.3160497205797583,
+      "rewards/accuracy_reward": 0.3431250059278682,
+      "rewards/format_reward": 0.9325000073760747,
+      "step": 240
+    },
+    {
+      "completion_length": 253.41000318527222,
+      "epoch": 0.43288609132571365,
+      "grad_norm": 7.472289085388184,
+      "kl": 6.22340087890625,
+      "learning_rate": 1.3969208539828873e-05,
+      "loss": 0.249,
+      "reward": 0.9746875168755651,
+      "reward_std": 0.49195458251051605,
+      "rewards/accuracy_reward": 0.21687500386033207,
+      "rewards/format_reward": 0.7578125147148966,
+      "step": 245
+    },
+    {
+      "completion_length": 297.43594336509705,
+      "epoch": 0.441720501352769,
+      "grad_norm": 0.7884982228279114,
+      "kl": 0.66375732421875,
+      "learning_rate": 1.3683545397639433e-05,
+      "loss": 0.0265,
+      "reward": 0.943125014565885,
+      "reward_std": 0.47752332291565835,
+      "rewards/accuracy_reward": 0.21625000413041562,
+      "rewards/format_reward": 0.7268750105053187,
+      "step": 250
+    },
+    {
+      "completion_length": 245.30719223022462,
+      "epoch": 0.4505549113798244,
+      "grad_norm": 0.726028323173523,
+      "kl": 0.21495361328125,
+      "learning_rate": 1.3394360625224067e-05,
+      "loss": 0.0086,
+      "reward": 1.3168750215321778,
+      "reward_std": 0.26591075989417734,
+      "rewards/accuracy_reward": 0.35125000493135305,
+      "rewards/format_reward": 0.9656250003725291,
+      "step": 255
+    },
+    {
+      "completion_length": 284.3690680503845,
+      "epoch": 0.45938932140687977,
+      "grad_norm": 1.1150134801864624,
+      "kl": 0.8627227783203125,
+      "learning_rate": 1.3101930695851186e-05,
+      "loss": 0.0345,
+      "reward": 1.3371875263750552,
+      "reward_std": 0.23761457363143562,
+      "rewards/accuracy_reward": 0.36343750746455045,
+      "rewards/format_reward": 0.9737499997019767,
+      "step": 260
+    },
+    {
+      "completion_length": 325.9565673828125,
+      "epoch": 0.4682237314339352,
+      "grad_norm": 0.41818270087242126,
+      "kl": 2.020770263671875,
+      "learning_rate": 1.2806535185300931e-05,
+      "loss": 0.0808,
+      "reward": 1.2912500217556953,
+      "reward_std": 0.3179181146901101,
+      "rewards/accuracy_reward": 0.35875000681262464,
+      "rewards/format_reward": 0.9325000055134296,
+      "step": 265
+    },
+    {
+      "completion_length": 318.7375056743622,
+      "epoch": 0.47705814146099057,
+      "grad_norm": 0.2984035313129425,
+      "kl": 0.5583831787109375,
+      "learning_rate": 1.2508456504578538e-05,
+      "loss": 0.0223,
+      "reward": 1.2543750200420618,
+      "reward_std": 0.3499622279778123,
+      "rewards/accuracy_reward": 0.3296875057742,
+      "rewards/format_reward": 0.924687511101365,
+      "step": 270
+    },
+    {
+      "completion_length": 287.0968791007996,
+      "epoch": 0.48589255148804594,
+      "grad_norm": 0.33989837765693665,
+      "kl": 0.371783447265625,
+      "learning_rate": 1.2207979629917061e-05,
+      "loss": 0.0149,
+      "reward": 1.2256250225007534,
+      "reward_std": 0.3394947111140937,
+      "rewards/accuracy_reward": 0.29718750540632755,
+      "rewards/format_reward": 0.9284375078976155,
+      "step": 275
+    },
+    {
+      "completion_length": 287.8387555122375,
+      "epoch": 0.4947269615151013,
+      "grad_norm": 0.268026739358902,
+      "kl": 0.1927764892578125,
+      "learning_rate": 1.1905391830327685e-05,
+      "loss": 0.0077,
+      "reward": 1.202500019222498,
+      "reward_std": 0.35878146667964755,
+      "rewards/accuracy_reward": 0.2943750049686059,
+      "rewards/format_reward": 0.9081250105053187,
+      "step": 280
+    },
+    {
+      "completion_length": 284.0490675926209,
+      "epoch": 0.5035613715421567,
+      "grad_norm": 1.2430206537246704,
+      "kl": 0.2335784912109375,
+      "learning_rate": 1.1600982392957978e-05,
+      "loss": 0.0094,
+      "reward": 1.20906252078712,
+      "reward_std": 0.3613354285713285,
+      "rewards/accuracy_reward": 0.30906250688713044,
+      "rewards/format_reward": 0.9000000122934579,
+      "step": 285
+    },
+    {
+      "completion_length": 264.8518793106079,
+      "epoch": 0.5123957815692121,
+      "grad_norm": 1.2759640216827393,
+      "kl": 0.343011474609375,
+      "learning_rate": 1.1295042346520755e-05,
+      "loss": 0.0137,
+      "reward": 1.268750026449561,
+      "reward_std": 0.3411198881920427,
+      "rewards/accuracy_reward": 0.3359375048661605,
+      "rewards/format_reward": 0.9328125044703484,
+      "step": 290
+    },
+    {
+      "completion_length": 276.43125615119936,
+      "epoch": 0.5212301915962675,
+      "grad_norm": 0.371543824672699,
+      "kl": 0.31484375,
+      "learning_rate": 1.0987864183057943e-05,
+      "loss": 0.0126,
+      "reward": 1.2600000239908695,
+      "reward_std": 0.35433362079784275,
+      "rewards/accuracy_reward": 0.3359375072643161,
+      "rewards/format_reward": 0.9240625072270632,
+      "step": 295
+    },
+    {
+      "completion_length": 285.17188062667844,
+      "epoch": 0.5300646016233228,
+      "grad_norm": 0.30253562331199646,
+      "kl": 0.2199676513671875,
+      "learning_rate": 1.067974157830539e-05,
+      "loss": 0.0088,
+      "reward": 1.2521875225007535,
+      "reward_std": 0.3534000650513917,
+      "rewards/accuracy_reward": 0.3321875044610351,
+      "rewards/format_reward": 0.9200000088661909,
+      "step": 300
+    },
+    {
+      "epoch": 0.5300646016233228,
+      "eval_completion_length": 251.39116995675224,
+      "eval_kl": 0.18108258928571427,
+      "eval_loss": 0.006980654317885637,
+      "eval_reward": 1.3089285918644495,
+      "eval_reward_std": 0.33055316124643597,
+      "eval_rewards/accuracy_reward": 0.3642857202461788,
+      "eval_rewards/format_reward": 0.9446428673607963,
+      "eval_runtime": 59.9414,
+      "eval_samples_per_second": 1.652,
+      "eval_steps_per_second": 0.033,
+      "step": 300
+    },
+    {
+      "completion_length": 284.31687960624697,
+      "epoch": 0.5388990116503782,
+      "grad_norm": 0.3302873969078064,
+      "kl": 0.2164642333984375,
+      "learning_rate": 1.0370969110926052e-05,
+      "loss": 0.0087,
+      "reward": 1.2406250204890967,
+      "reward_std": 0.35780384358949957,
+      "rewards/accuracy_reward": 0.3165625065099448,
+      "rewards/format_reward": 0.9240625072270632,
+      "step": 305
+    },
+    {
+      "completion_length": 327.0396939754486,
+      "epoch": 0.5477334216774336,
+      "grad_norm": 0.4746917188167572,
+      "kl": 0.3468963623046875,
+      "learning_rate": 1.0061841980879941e-05,
+      "loss": 0.0139,
+      "reward": 1.1775000210851432,
+      "reward_std": 0.41827466948889197,
+      "rewards/accuracy_reward": 0.30406250536907464,
+      "rewards/format_reward": 0.873437513411045,
+      "step": 310
+    },
+    {
+      "completion_length": 348.985320186615,
+      "epoch": 0.556567831704489,
+      "grad_norm": 0.31228217482566833,
+      "kl": 0.30557861328125,
+      "learning_rate": 9.752655727200051e-06,
+      "loss": 0.0122,
+      "reward": 1.2109375238418578,
+      "reward_std": 0.4383382148575038,
+      "rewards/accuracy_reward": 0.3440625049173832,
+      "rewards/format_reward": 0.8668750118464231,
+      "step": 315
+    },
+    {
+      "completion_length": 343.36594438552856,
+      "epoch": 0.5654022417315444,
+      "grad_norm": 0.47009924054145813,
+      "kl": 0.3079986572265625,
+      "learning_rate": 9.443705945444158e-06,
+      "loss": 0.0123,
+      "reward": 1.1912500225007534,
+      "reward_std": 0.4135822183918208,
+      "rewards/accuracy_reward": 0.31593750645406543,
+      "rewards/format_reward": 0.8753125127404928,
+      "step": 320
+    },
+    {
+      "completion_length": 347.44625635147094,
+      "epoch": 0.5742366517585997,
+      "grad_norm": 0.35182997584342957,
+      "kl": 0.278741455078125,
+      "learning_rate": 9.135288005092546e-06,
+      "loss": 0.0112,
+      "reward": 1.194062525779009,
+      "reward_std": 0.42705982849001883,
+      "rewards/accuracy_reward": 0.3281250067055225,
+      "rewards/format_reward": 0.8659375134855509,
+      "step": 325
+    },
+    {
+      "completion_length": 280.76656765937804,
+      "epoch": 0.5830710617856552,
+      "grad_norm": 0.27393487095832825,
+      "kl": 0.212225341796875,
+      "learning_rate": 8.827696767161902e-06,
+      "loss": 0.0085,
+      "reward": 1.2568750269711018,
+      "reward_std": 0.287053491268307,
+      "rewards/accuracy_reward": 0.2962500054156408,
+      "rewards/format_reward": 0.9606250025331974,
+      "step": 330
+    },
+    {
+      "completion_length": 288.0281304359436,
+      "epoch": 0.5919054718127105,
+      "grad_norm": 0.22333523631095886,
+      "kl": 0.1795684814453125,
+      "learning_rate": 8.52122630230531e-06,
+      "loss": 0.0072,
+      "reward": 1.2878125224262476,
+      "reward_std": 0.2826185919344425,
+      "rewards/accuracy_reward": 0.3234375049592927,
+      "rewards/format_reward": 0.964375001564622,
+      "step": 335
+    },
+    {
+      "completion_length": 316.2690674781799,
+      "epoch": 0.6007398818397659,
+      "grad_norm": 0.35810768604278564,
+      "kl": 0.192633056640625,
+      "learning_rate": 8.216169609667854e-06,
+      "loss": 0.0077,
+      "reward": 1.2306250248104333,
+      "reward_std": 0.35166344842873515,
+      "rewards/accuracy_reward": 0.3065625059651211,
+      "rewards/format_reward": 0.9240625072270632,
+      "step": 340
+    },
+    {
+      "completion_length": 359.2031311035156,
+      "epoch": 0.6095742918668212,
+      "grad_norm": 0.33678558468818665,
+      "kl": 0.2140533447265625,
+      "learning_rate": 7.91281833676665e-06,
+      "loss": 0.0086,
+      "reward": 1.1700000200420617,
+      "reward_std": 0.43951611914671956,
+      "rewards/accuracy_reward": 0.305937506374903,
+      "rewards/format_reward": 0.8640625115484,
+      "step": 345
+    },
+    {
+      "completion_length": 343.9143795967102,
+      "epoch": 0.6184087018938766,
+      "grad_norm": 0.3433804512023926,
+      "kl": 0.2072998046875,
+      "learning_rate": 7.6114625006630885e-06,
+      "loss": 0.0083,
+      "reward": 1.2115625184029342,
+      "reward_std": 0.39576993524096904,
+      "rewards/accuracy_reward": 0.3228125046007335,
+      "rewards/format_reward": 0.8887500122189522,
+      "step": 350
+    },
+    {
+      "completion_length": 303.42125554084777,
+      "epoch": 0.6272431119209321,
+      "grad_norm": 0.296512633562088,
+      "kl": 0.209527587890625,
+      "learning_rate": 7.312390210693863e-06,
+      "loss": 0.0084,
+      "reward": 1.245625025779009,
+      "reward_std": 0.3566357252653688,
+      "rewards/accuracy_reward": 0.32593750569503754,
+      "rewards/format_reward": 0.919687508419156,
+      "step": 355
+    },
+    {
+      "completion_length": 272.6603165149689,
+      "epoch": 0.6360775219479874,
+      "grad_norm": 0.4612409174442291,
+      "kl": 0.25294189453125,
+      "learning_rate": 7.015887393025847e-06,
+      "loss": 0.0101,
+      "reward": 1.2715625170618297,
+      "reward_std": 0.35159414759837093,
+      "rewards/accuracy_reward": 0.3450000061420724,
+      "rewards/format_reward": 0.9265625070780515,
+      "step": 360
+    },
+    {
+      "completion_length": 306.7465679168701,
+      "epoch": 0.6449119319750428,
+      "grad_norm": 0.5136480927467346,
+      "kl": 0.351116943359375,
+      "learning_rate": 6.722237517298232e-06,
+      "loss": 0.014,
+      "reward": 1.2143750261515378,
+      "reward_std": 0.40818659723736345,
+      "rewards/accuracy_reward": 0.32218750594183804,
+      "rewards/format_reward": 0.8921875134110451,
+      "step": 365
+    },
+    {
+      "completion_length": 293.43188071250916,
+      "epoch": 0.6537463420020981,
+      "grad_norm": 1.3061071634292603,
+      "kl": 0.31224365234375,
+      "learning_rate": 6.431721325613138e-06,
+      "loss": 0.0125,
+      "reward": 1.236562515050173,
+      "reward_std": 0.3789135571103543,
+      "rewards/accuracy_reward": 0.32000000490806996,
+      "rewards/format_reward": 0.9165625084191561,
+      "step": 370
+    },
+    {
+      "completion_length": 278.6196911811829,
+      "epoch": 0.6625807520291536,
+      "grad_norm": 0.5205467939376831,
+      "kl": 0.27296142578125,
+      "learning_rate": 6.144616564133927e-06,
+      "loss": 0.0109,
+      "reward": 1.2200000256299972,
+      "reward_std": 0.36232984592206774,
+      "rewards/accuracy_reward": 0.3078125056345016,
+      "rewards/format_reward": 0.912187509611249,
+      "step": 375
+    },
+    {
+      "completion_length": 256.6856292724609,
+      "epoch": 0.6714151620562089,
+      "grad_norm": 0.4971334934234619,
+      "kl": 1.381793212890625,
+      "learning_rate": 5.8611977175477355e-06,
+      "loss": 0.0552,
+      "reward": 1.2637500211596489,
+      "reward_std": 0.3477331514935941,
+      "rewards/accuracy_reward": 0.3281250061001629,
+      "rewards/format_reward": 0.9356250062584877,
+      "step": 380
+    },
+    {
+      "completion_length": 292.6896943569183,
+      "epoch": 0.6802495720832643,
+      "grad_norm": 0.4106786251068115,
+      "kl": 0.2263092041015625,
+      "learning_rate": 5.581735746646134e-06,
+      "loss": 0.0091,
+      "reward": 1.2271875157952308,
+      "reward_std": 0.3678107146173716,
+      "rewards/accuracy_reward": 0.3059375059790909,
+      "rewards/format_reward": 0.921250008046627,
+      "step": 385
+    },
+    {
+      "completion_length": 338.68781661987305,
+      "epoch": 0.6890839821103197,
+      "grad_norm": 0.31216996908187866,
+      "kl": 0.23704833984375,
+      "learning_rate": 5.306497829274785e-06,
+      "loss": 0.0095,
+      "reward": 1.2384375259280205,
+      "reward_std": 0.3888756018597633,
+      "rewards/accuracy_reward": 0.33750000621657816,
+      "rewards/format_reward": 0.9009375102818012,
+      "step": 390
+    },
+    {
+      "completion_length": 350.2625060081482,
+      "epoch": 0.697918392137375,
+      "grad_norm": 0.38919156789779663,
+      "kl": 0.2536376953125,
+      "learning_rate": 5.035747104899738e-06,
+      "loss": 0.0101,
+      "reward": 1.2350000202655793,
+      "reward_std": 0.3974794507026672,
+      "rewards/accuracy_reward": 0.3350000062491745,
+      "rewards/format_reward": 0.9000000089406968,
+      "step": 395
+    },
+    {
+      "completion_length": 333.70625457763674,
+      "epoch": 0.7067528021644305,
+      "grad_norm": 0.2801852524280548,
+      "kl": 0.2380279541015625,
+      "learning_rate": 4.76974242303464e-06,
+      "loss": 0.0095,
+      "reward": 1.2937500216066837,
+      "reward_std": 0.32844443861395123,
+      "rewards/accuracy_reward": 0.3625000061234459,
+      "rewards/format_reward": 0.9312500074505806,
+      "step": 400
+    },
+    {
+      "epoch": 0.7067528021644305,
+      "eval_completion_length": 332.7068655831473,
+      "eval_kl": 0.251953125,
+      "eval_loss": 0.009230787865817547,
+      "eval_reward": 1.2875000068119593,
+      "eval_reward_std": 0.3499350760664259,
+      "eval_rewards/accuracy_reward": 0.35714286140033175,
+      "eval_rewards/format_reward": 0.9303571496691022,
+      "eval_runtime": 61.6271,
+      "eval_samples_per_second": 1.606,
+      "eval_steps_per_second": 0.032,
+      "step": 400
+    },
+    {
+      "completion_length": 348.7931314468384,
+      "epoch": 0.7155872121914858,
+      "grad_norm": 0.6133540868759155,
+      "kl": 0.256671142578125,
+      "learning_rate": 4.508738095769278e-06,
+      "loss": 0.0103,
+      "reward": 1.2540625236928462,
+      "reward_std": 0.3829473619814962,
+      "rewards/accuracy_reward": 0.34531250735744834,
+      "rewards/format_reward": 0.9087500102818012,
+      "step": 405
+    },
+    {
+      "completion_length": 355.2525053024292,
+      "epoch": 0.7244216222185412,
+      "grad_norm": 0.3552949130535126,
+      "kl": 0.2611785888671875,
+      "learning_rate": 4.252983654636115e-06,
+      "loss": 0.0104,
+      "reward": 1.2318750265985727,
+      "reward_std": 0.371259527374059,
+      "rewards/accuracy_reward": 0.32937500602565706,
+      "rewards/format_reward": 0.9025000102818013,
+      "step": 410
+    },
+    {
+      "completion_length": 341.5478183746338,
+      "epoch": 0.7332560322455965,
+      "grad_norm": 0.326225221157074,
+      "kl": 0.23895263671875,
+      "learning_rate": 4.002723612047272e-06,
+      "loss": 0.0096,
+      "reward": 1.2762500245124102,
+      "reward_std": 0.39916003863327204,
+      "rewards/accuracy_reward": 0.37593750609084964,
+      "rewards/format_reward": 0.9003125108778477,
+      "step": 415
+    },
+    {
+      "completion_length": 352.67813138961793,
+      "epoch": 0.742090442272652,
+      "grad_norm": 1.1807901859283447,
+      "kl": 4.40692138671875,
+      "learning_rate": 3.7581972275299606e-06,
+      "loss": 0.1764,
+      "reward": 1.2221875179558992,
+      "reward_std": 0.3770625921431929,
+      "rewards/accuracy_reward": 0.3328125061932951,
+      "rewards/format_reward": 0.8893750138580799,
+      "step": 420
+    },
+    {
+      "completion_length": 344.5912554740906,
+      "epoch": 0.7509248522997074,
+      "grad_norm": 0.6603362560272217,
+      "kl": 0.2891998291015625,
+      "learning_rate": 3.5196382789839477e-06,
+      "loss": 0.0116,
+      "reward": 1.2221875227987766,
+      "reward_std": 0.4097146320156753,
+      "rewards/accuracy_reward": 0.34687500637955965,
+      "rewards/format_reward": 0.8753125123679638,
+      "step": 425
+    },
+    {
+      "completion_length": 340.8268808364868,
+      "epoch": 0.7597592623267627,
+      "grad_norm": 0.3733387291431427,
+      "kl": 0.2832794189453125,
+      "learning_rate": 3.2872748391796736e-06,
+      "loss": 0.0113,
+      "reward": 1.2106250237673521,
+      "reward_std": 0.39458047463558615,
+      "rewards/accuracy_reward": 0.32937500667758285,
+      "rewards/format_reward": 0.8812500111758709,
+      "step": 430
+    },
+    {
+      "completion_length": 320.6984436035156,
+      "epoch": 0.7685936723538181,
+      "grad_norm": 0.40383297204971313,
+      "kl": 0.24058837890625,
+      "learning_rate": 3.061329057710711e-06,
+      "loss": 0.0096,
+      "reward": 1.2590625163167715,
+      "reward_std": 0.3429459627252072,
+      "rewards/accuracy_reward": 0.3390625057509169,
+      "rewards/format_reward": 0.9200000084936619,
+      "step": 435
+    },
+    {
+      "completion_length": 311.40313014984133,
+      "epoch": 0.7774280823808735,
+      "grad_norm": 0.5515756607055664,
+      "kl": 0.209765625,
+      "learning_rate": 2.8420169486090765e-06,
+      "loss": 0.0084,
+      "reward": 1.2771875217556954,
+      "reward_std": 0.34724566363729537,
+      "rewards/accuracy_reward": 0.34937500532250854,
+      "rewards/format_reward": 0.9278125081211328,
+      "step": 440
+    },
+    {
+      "completion_length": 329.884068107605,
+      "epoch": 0.7862624924079289,
+      "grad_norm": 0.3992791473865509,
+      "kl": 0.232708740234375,
+      "learning_rate": 2.6295481838263628e-06,
+      "loss": 0.0093,
+      "reward": 1.2671875212341548,
+      "reward_std": 0.3570181305985898,
+      "rewards/accuracy_reward": 0.34968750572297724,
+      "rewards/format_reward": 0.9175000086426734,
+      "step": 445
+    },
+    {
+      "completion_length": 347.53969345092776,
+      "epoch": 0.7950969024349842,
+      "grad_norm": 0.45689699053764343,
+      "kl": 0.2296356201171875,
+      "learning_rate": 2.4241258927782197e-06,
+      "loss": 0.0092,
+      "reward": 1.2487500190734864,
+      "reward_std": 0.39153953776694833,
+      "rewards/accuracy_reward": 0.34531250717118384,
+      "rewards/format_reward": 0.9034375082701445,
+      "step": 450
+    },
+    {
+      "completion_length": 348.45406847000123,
+      "epoch": 0.8039313124620396,
+      "grad_norm": 0.5398163199424744,
+      "kl": 0.3041656494140625,
+      "learning_rate": 2.2259464681437404e-06,
+      "loss": 0.0122,
+      "reward": 1.238437521085143,
+      "reward_std": 0.37067002258263526,
+      "rewards/accuracy_reward": 0.33906250675208865,
+      "rewards/format_reward": 0.8993750099092722,
+      "step": 455
+    },
+    {
+      "completion_length": 361.00125675201417,
+      "epoch": 0.8127657224890951,
+      "grad_norm": 0.5371863842010498,
+      "kl": 0.225506591796875,
+      "learning_rate": 2.0351993781055192e-06,
+      "loss": 0.009,
+      "reward": 1.253125024214387,
+      "reward_std": 0.39060982735827565,
+      "rewards/accuracy_reward": 0.3550000062212348,
+      "rewards/format_reward": 0.8981250114738941,
+      "step": 460
+    },
+    {
+      "completion_length": 347.938130569458,
+      "epoch": 0.8216001325161504,
+      "grad_norm": 0.5616302490234375,
+      "kl": 219.42605285644532,
+      "learning_rate": 1.8520669852097573e-06,
+      "loss": 8.7857,
+      "reward": 1.2534375235438346,
+      "reward_std": 0.3921773022972047,
+      "rewards/accuracy_reward": 0.35656250605825335,
+      "rewards/format_reward": 0.8968750108033419,
+      "step": 465
+    },
+    {
+      "completion_length": 355.52563123703004,
+      "epoch": 0.8304345425432058,
+      "grad_norm": 0.5742087960243225,
+      "kl": 0.2433929443359375,
+      "learning_rate": 1.6767243720197302e-06,
+      "loss": 0.0097,
+      "reward": 1.2378125201910735,
+      "reward_std": 0.40007057855837047,
+      "rewards/accuracy_reward": 0.34625000662636013,
+      "rewards/format_reward": 0.8915625076740981,
+      "step": 470
+    },
+    {
+      "completion_length": 346.7978181838989,
+      "epoch": 0.8392689525702611,
+      "grad_norm": 3.535346269607544,
+      "kl": 0.69002685546875,
+      "learning_rate": 1.509339173729214e-06,
+      "loss": 0.0277,
+      "reward": 1.23781252913177,
+      "reward_std": 0.401255727186799,
+      "rewards/accuracy_reward": 0.34593750624917446,
+      "rewards/format_reward": 0.8918750081211329,
+      "step": 475
+    },
+    {
+      "completion_length": 340.69406814575194,
+      "epoch": 0.8481033625973166,
+      "grad_norm": 0.2953658401966095,
+      "kl": 0.262261962890625,
+      "learning_rate": 1.3500714178959084e-06,
+      "loss": 0.0105,
+      "reward": 1.243437521159649,
+      "reward_std": 0.3896955798845738,
+      "rewards/accuracy_reward": 0.34843750614672897,
+      "rewards/format_reward": 0.8950000096112489,
+      "step": 480
+    },
+    {
+      "completion_length": 354.4678184509277,
+      "epoch": 0.8569377726243719,
+      "grad_norm": 0.4164278507232666,
+      "kl": 0.227349853515625,
+      "learning_rate": 1.1990733714481185e-06,
+      "loss": 0.0091,
+      "reward": 1.2381250221282243,
+      "reward_std": 0.3951717680785805,
+      "rewards/accuracy_reward": 0.34937500648666175,
+      "rewards/format_reward": 0.8887500129640102,
+      "step": 485
+    },
+    {
+      "completion_length": 352.11563148498533,
+      "epoch": 0.8657721826514273,
+      "grad_norm": 0.5129674077033997,
+      "kl": 0.233819580078125,
+      "learning_rate": 1.0564893951109068e-06,
+      "loss": 0.0094,
+      "reward": 1.2071875222027302,
+      "reward_std": 0.37804696392267945,
+      "rewards/accuracy_reward": 0.3178125045960769,
+      "rewards/format_reward": 0.8893750116229058,
+      "step": 490
+    },
+    {
+      "completion_length": 354.6234440803528,
+      "epoch": 0.8746065926784827,
+      "grad_norm": 0.4102705419063568,
+      "kl": 0.276885986328125,
+      "learning_rate": 9.224558053909615e-07,
+      "loss": 0.0111,
+      "reward": 1.2128125235438347,
+      "reward_std": 0.38704944350756704,
+      "rewards/accuracy_reward": 0.32593750627711415,
+      "rewards/format_reward": 0.8868750117719173,
+      "step": 495
+    },
+    {
+      "completion_length": 346.5503177642822,
+      "epoch": 0.883441002705538,
+      "grad_norm": 0.70513516664505,
+      "kl": 0.772918701171875,
+      "learning_rate": 7.971007442520363e-07,
+      "loss": 0.031,
+      "reward": 1.2296875193715096,
+      "reward_std": 0.38830175981856885,
+      "rewards/accuracy_reward": 0.34218750533182174,
+      "rewards/format_reward": 0.8875000096857548,
+      "step": 500
+    },
+    {
+      "epoch": 0.883441002705538,
+      "eval_completion_length": 339.28299822126115,
+      "eval_kl": 0.25753348214285715,
+      "eval_loss": 0.00743135716766119,
+      "eval_reward": 1.2214285816465105,
+      "eval_reward_std": 0.41976045284952435,
+      "eval_rewards/accuracy_reward": 0.3339285786662783,
+      "eval_rewards/format_reward": 0.8875000136239188,
+      "eval_runtime": 61.4511,
+      "eval_samples_per_second": 1.611,
+      "eval_steps_per_second": 0.033,
+      "step": 500
+    },
+    {
+      "completion_length": 339.661568069458,
+      "epoch": 0.8922754127325935,
+      "grad_norm": 0.539588212966919,
+      "kl": 0.24498291015625,
+      "learning_rate": 6.805440566056554e-07,
+      "loss": 0.0098,
+      "reward": 1.2340625192970038,
+      "reward_std": 0.3814273901283741,
+      "rewards/accuracy_reward": 0.33062500660307703,
+      "rewards/format_reward": 0.9034375101327896,
+      "step": 505
+    },
+    {
+      "completion_length": 322.16969327926637,
+      "epoch": 0.9011098227596488,
+      "grad_norm": 1.271132469177246,
+      "kl": 0.26392822265625,
+      "learning_rate": 5.72897175734145e-07,
+      "loss": 0.0106,
+      "reward": 1.2703125223517417,
+      "reward_std": 0.3676451860461384,
+      "rewards/accuracy_reward": 0.3693750074598938,
+      "rewards/format_reward": 0.9009375102818012,
+      "step": 510
+    },
+    {
+      "completion_length": 312.22781734466554,
+      "epoch": 0.9099442327867042,
+      "grad_norm": 0.48748400807380676,
+      "kl": 0.2677764892578125,
+      "learning_rate": 4.7426301675554285e-07,
+      "loss": 0.0107,
+      "reward": 1.2534375242888927,
+      "reward_std": 0.3893388253170997,
+      "rewards/accuracy_reward": 0.3481250064447522,
+      "rewards/format_reward": 0.9053125087171793,
+      "step": 515
+    },
+    {
+      "completion_length": 329.6862547874451,
+      "epoch": 0.9187786428137595,
+      "grad_norm": 1.5978107452392578,
+      "kl": 0.26893310546875,
+      "learning_rate": 3.847358782322608e-07,
+      "loss": 0.0108,
+      "reward": 1.2581250235438346,
+      "reward_std": 0.390413482952863,
+      "rewards/accuracy_reward": 0.35781250612344595,
+      "rewards/format_reward": 0.9003125101327896,
+      "step": 520
+    },
+    {
+      "completion_length": 328.39188108444216,
+      "epoch": 0.927613052840815,
+      "grad_norm": 0.5838448405265808,
+      "kl": 0.31304931640625,
+      "learning_rate": 3.044013520175337e-07,
+      "loss": 0.0125,
+      "reward": 1.2337500229477882,
+      "reward_std": 0.3816523430403322,
+      "rewards/accuracy_reward": 0.33281250556465236,
+      "rewards/format_reward": 0.9009375125169754,
+      "step": 525
+    },
+    {
+      "completion_length": 337.2209428787231,
+      "epoch": 0.9364474628678704,
+      "grad_norm": 0.6874144673347473,
+      "kl": 0.27744140625,
+      "learning_rate": 2.3333624142587884e-07,
+      "loss": 0.0111,
+      "reward": 1.230000025033951,
+      "reward_std": 0.4204695597756654,
+      "rewards/accuracy_reward": 0.337500005797483,
+      "rewards/format_reward": 0.8925000090152025,
+      "step": 530
+    },
+    {
+      "completion_length": 331.3159426689148,
+      "epoch": 0.9452818728949257,
+      "grad_norm": 0.5779227614402771,
+      "kl": 0.255389404296875,
+      "learning_rate": 1.7160848780576334e-07,
+      "loss": 0.0102,
+      "reward": 1.2512500252574683,
+      "reward_std": 0.39169925842434167,
+      "rewards/accuracy_reward": 0.34906250566709784,
+      "rewards/format_reward": 0.9021875113248825,
+      "step": 535
+    },
+    {
+      "completion_length": 324.4328186035156,
+      "epoch": 0.9541162829219811,
+      "grad_norm": 0.9910773634910583,
+      "kl": 0.2505706787109375,
+      "learning_rate": 1.192771055847197e-07,
+      "loss": 0.01,
+      "reward": 1.2465625245124101,
+      "reward_std": 0.38332765088416637,
+      "rewards/accuracy_reward": 0.33718750630505384,
+      "rewards/format_reward": 0.9093750111758709,
+      "step": 540
+    },
+    {
+      "completion_length": 325.0950054168701,
+      "epoch": 0.9629506929490365,
+      "grad_norm": 0.4231570065021515,
+      "kl": 0.2594268798828125,
+      "learning_rate": 7.639212584897082e-08,
+      "loss": 0.0104,
+      "reward": 1.2393750201910734,
+      "reward_std": 0.3615853681694716,
+      "rewards/accuracy_reward": 0.33531250508967786,
+      "rewards/format_reward": 0.9040625128895045,
+      "step": 545
+    },
+    {
+      "completion_length": 336.5759434700012,
+      "epoch": 0.9717851029760919,
+      "grad_norm": 0.4935801029205322,
+      "kl": 0.25858154296875,
+      "learning_rate": 4.2994548511525735e-08,
+      "loss": 0.0103,
+      "reward": 1.2403125245124103,
+      "reward_std": 0.4053161561954767,
+      "rewards/accuracy_reward": 0.3540625063236803,
+      "rewards/format_reward": 0.8862500093877316,
+      "step": 550
+    },
+    {
+      "completion_length": 348.5090696334839,
+      "epoch": 0.9806195130031473,
+      "grad_norm": 0.47521111369132996,
+      "kl": 0.26097412109375,
+      "learning_rate": 1.9116303114480316e-08,
+      "loss": 0.0104,
+      "reward": 1.2050000194460153,
+      "reward_std": 0.4146237867884338,
+      "rewards/accuracy_reward": 0.32812500463332983,
+      "rewards/format_reward": 0.8768750112503767,
+      "step": 555
+    },
+    {
+      "completion_length": 330.13719339370726,
+      "epoch": 0.9894539230302026,
+      "grad_norm": 0.6407186985015869,
+      "kl": 0.2815765380859375,
+      "learning_rate": 4.7802183029710406e-09,
+      "loss": 0.0113,
+      "reward": 1.2556250225752592,
+      "reward_std": 0.3957557766698301,
+      "rewards/accuracy_reward": 0.36156250678468493,
+      "rewards/format_reward": 0.8940625116229057,
+      "step": 560
+    },
+    {
+      "completion_length": 322.6503179550171,
+      "epoch": 0.9982883330572581,
+      "grad_norm": 0.38322028517723083,
+      "kl": 0.26986083984375,
+      "learning_rate": 0.0,
+      "loss": 0.0108,
+      "reward": 1.2721875175833701,
+      "reward_std": 0.39191843890585004,
+      "rewards/accuracy_reward": 0.3693750073434785,
+      "rewards/format_reward": 0.9028125129640102,
+      "step": 565
+    },
+    {
+      "epoch": 0.9982883330572581,
+      "step": 565,
+      "total_flos": 0.0,
+      "train_loss": 2.1605342640346974,
+      "train_runtime": 168649.3171,
+      "train_samples_per_second": 0.43,
+      "train_steps_per_second": 0.003
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 565,
+  "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": 0.0,
+  "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 sha256:2113c0bd1649fffaf27f97e48a3aceaad6568b360e3850557966beff04d57531
+size 7608