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="caijanfeng/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/caijf-USTC/huggingface/runs/8zwymxds) 
+
+
+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: 4.49.0.dev0
+- Pytorch: 2.5.1
+- Datasets: 3.3.0
+- 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": 0.027384539475202127,
+    "train_runtime": 37232.0457,
+    "train_samples": 72441,
+    "train_samples_per_second": 1.946,
+    "train_steps_per_second": 0.008
+}

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": "4.49.0.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:cba7e93e3f26dd6f99f67bfe7fff8aeb6aa27c2ff82f6c70f984c3de8df95ee8
+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": 0.027384539475202127,
+    "train_runtime": 37232.0457,
+    "train_samples": 72441,
+    "train_samples_per_second": 1.946,
+    "train_steps_per_second": 0.008
+}

trainer_state.json

@@ -0,0 +1,804 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 100,
+  "global_step": 283,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "completion_length": 383.0911985397339,
+      "epoch": 0.0176678445229682,
+      "grad_norm": 0.9267308712005615,
+      "kl": 0.0008915185928344727,
+      "learning_rate": 3.448275862068966e-06,
+      "loss": 0.0,
+      "reward": 0.6448660973459482,
+      "reward_std": 0.43812552504241464,
+      "rewards/accuracy_reward": 0.1502232206054032,
+      "rewards/format_reward": 0.49464287869632245,
+      "step": 5
+    },
+    {
+      "completion_length": 224.4988929748535,
+      "epoch": 0.0353356890459364,
+      "grad_norm": 0.44213607907295227,
+      "kl": 0.31944580078125,
+      "learning_rate": 6.896551724137932e-06,
+      "loss": 0.0128,
+      "reward": 0.9520089730620385,
+      "reward_std": 0.26545318868011236,
+      "rewards/accuracy_reward": 0.07265625306172296,
+      "rewards/format_reward": 0.8793527193367481,
+      "step": 10
+    },
+    {
+      "completion_length": 157.38460483551026,
+      "epoch": 0.053003533568904596,
+      "grad_norm": 0.2661096155643463,
+      "kl": 0.05403594970703125,
+      "learning_rate": 1.0344827586206898e-05,
+      "loss": 0.0022,
+      "reward": 1.0422991596162319,
+      "reward_std": 0.21440245490521193,
+      "rewards/accuracy_reward": 0.09095982516882942,
+      "rewards/format_reward": 0.9513393238186836,
+      "step": 15
+    },
+    {
+      "completion_length": 96.58973650932312,
+      "epoch": 0.0706713780918728,
+      "grad_norm": 0.39351069927215576,
+      "kl": 0.15865478515625,
+      "learning_rate": 1.3793103448275863e-05,
+      "loss": 0.0063,
+      "reward": 1.1008929066359996,
+      "reward_std": 0.20822038110345603,
+      "rewards/accuracy_reward": 0.12767857781145722,
+      "rewards/format_reward": 0.9732143238186837,
+      "step": 20
+    },
+    {
+      "completion_length": 107.32266120910644,
+      "epoch": 0.08833922261484099,
+      "grad_norm": 0.661626398563385,
+      "kl": 0.35902099609375,
+      "learning_rate": 1.7241379310344828e-05,
+      "loss": 0.0144,
+      "reward": 1.128459869325161,
+      "reward_std": 0.2679577572271228,
+      "rewards/accuracy_reward": 0.17399554383009672,
+      "rewards/format_reward": 0.9544643260538578,
+      "step": 25
+    },
+    {
+      "completion_length": 47.275559878349306,
+      "epoch": 0.10600706713780919,
+      "grad_norm": 0.5010063052177429,
+      "kl": 0.5186767578125,
+      "learning_rate": 1.999923511388017e-05,
+      "loss": 0.0208,
+      "reward": 1.1618304058909417,
+      "reward_std": 0.210414170473814,
+      "rewards/accuracy_reward": 0.18805804364383222,
+      "rewards/format_reward": 0.9737723633646965,
+      "step": 30
+    },
+    {
+      "completion_length": 239.45213136672973,
+      "epoch": 0.12367491166077739,
+      "grad_norm": 0.5395922660827637,
+      "kl": 0.2320068359375,
+      "learning_rate": 1.9972476383747748e-05,
+      "loss": 0.0093,
+      "reward": 1.1856027349829674,
+      "reward_std": 0.2784878654405475,
+      "rewards/accuracy_reward": 0.2165178671013564,
+      "rewards/format_reward": 0.9690848588943481,
+      "step": 35
+    },
+    {
+      "completion_length": 170.19040813446045,
+      "epoch": 0.1413427561837456,
+      "grad_norm": 0.393367201089859,
+      "kl": 0.2654296875,
+      "learning_rate": 1.9907590277344582e-05,
+      "loss": 0.0106,
+      "reward": 1.1880580872297286,
+      "reward_std": 0.25068697994574907,
+      "rewards/accuracy_reward": 0.20792411724105478,
+      "rewards/format_reward": 0.9801339648663998,
+      "step": 40
+    },
+    {
+      "completion_length": 132.1880645751953,
+      "epoch": 0.15901060070671377,
+      "grad_norm": 0.2796829342842102,
+      "kl": 0.3013916015625,
+      "learning_rate": 1.9804824871166254e-05,
+      "loss": 0.0121,
+      "reward": 1.2225446954369545,
+      "reward_std": 0.24187104841694235,
+      "rewards/accuracy_reward": 0.24073661854490638,
+      "rewards/format_reward": 0.9818080730736256,
+      "step": 45
+    },
+    {
+      "completion_length": 166.0952097892761,
+      "epoch": 0.17667844522968199,
+      "grad_norm": 0.3038596510887146,
+      "kl": 0.27999267578125,
+      "learning_rate": 1.9664573064143604e-05,
+      "loss": 0.0112,
+      "reward": 1.243750052154064,
+      "reward_std": 0.2760961548425257,
+      "rewards/accuracy_reward": 0.27399554969742895,
+      "rewards/format_reward": 0.9697545059025288,
+      "step": 50
+    },
+    {
+      "completion_length": 146.12824268341063,
+      "epoch": 0.19434628975265017,
+      "grad_norm": 0.3787432312965393,
+      "kl": 0.40203857421875,
+      "learning_rate": 1.948737107548771e-05,
+      "loss": 0.0161,
+      "reward": 1.200334867835045,
+      "reward_std": 0.30924737490713594,
+      "rewards/accuracy_reward": 0.264620547182858,
+      "rewards/format_reward": 0.9357143260538578,
+      "step": 55
+    },
+    {
+      "completion_length": 63.74587321281433,
+      "epoch": 0.21201413427561838,
+      "grad_norm": 0.6207670569419861,
+      "kl": 0.74423828125,
+      "learning_rate": 1.9273896394584103e-05,
+      "loss": 0.0298,
+      "reward": 1.1473214782774448,
+      "reward_std": 0.2868926800787449,
+      "rewards/accuracy_reward": 0.22734376145526766,
+      "rewards/format_reward": 0.9199777208268642,
+      "step": 60
+    },
+    {
+      "completion_length": 110.70368802547455,
+      "epoch": 0.22968197879858657,
+      "grad_norm": 52.001220703125,
+      "kl": 1.921142578125,
+      "learning_rate": 1.9024965190774262e-05,
+      "loss": 0.0769,
+      "reward": 1.1512277290225028,
+      "reward_std": 0.27448821542784574,
+      "rewards/accuracy_reward": 0.23549108249135314,
+      "rewards/format_reward": 0.9157366491854191,
+      "step": 65
+    },
+    {
+      "completion_length": 226.51317901611327,
+      "epoch": 0.24734982332155478,
+      "grad_norm": 3.3317148685455322,
+      "kl": 1.24912109375,
+      "learning_rate": 1.8741529192927528e-05,
+      "loss": 0.05,
+      "reward": 1.0268973752856254,
+      "reward_std": 0.4129971068352461,
+      "rewards/accuracy_reward": 0.20267858097795396,
+      "rewards/format_reward": 0.824218786507845,
+      "step": 70
+    },
+    {
+      "completion_length": 128.79185905456544,
+      "epoch": 0.26501766784452296,
+      "grad_norm": 14.126457214355469,
+      "kl": 1.390869140625,
+      "learning_rate": 1.8424672050733577e-05,
+      "loss": 0.0556,
+      "reward": 1.1761161178350448,
+      "reward_std": 0.30296447649598124,
+      "rewards/accuracy_reward": 0.2532366174273193,
+      "rewards/format_reward": 0.9228795051574707,
+      "step": 75
+    },
+    {
+      "completion_length": 130.28717198371888,
+      "epoch": 0.2826855123674912,
+      "grad_norm": 12.251070976257324,
+      "kl": 4.2921875,
+      "learning_rate": 1.8075605191627242e-05,
+      "loss": 0.1717,
+      "reward": 1.15781254991889,
+      "reward_std": 0.31116877496242523,
+      "rewards/accuracy_reward": 0.2357142967171967,
+      "rewards/format_reward": 0.9220982573926448,
+      "step": 80
+    },
+    {
+      "completion_length": 166.793869304657,
+      "epoch": 0.3003533568904594,
+      "grad_norm": 8.950257301330566,
+      "kl": 3.652392578125,
+      "learning_rate": 1.7695663189185703e-05,
+      "loss": 0.1461,
+      "reward": 1.1366071909666062,
+      "reward_std": 0.37287568356841805,
+      "rewards/accuracy_reward": 0.24709822554141284,
+      "rewards/format_reward": 0.8895089700818062,
+      "step": 85
+    },
+    {
+      "completion_length": 127.73795194625855,
+      "epoch": 0.31802120141342755,
+      "grad_norm": 2.8447344303131104,
+      "kl": 2.64189453125,
+      "learning_rate": 1.7286298660705877e-05,
+      "loss": 0.1056,
+      "reward": 1.1922991588711738,
+      "reward_std": 0.31104280035942794,
+      "rewards/accuracy_reward": 0.25502233263105156,
+      "rewards/format_reward": 0.93727682903409,
+      "step": 90
+    },
+    {
+      "completion_length": 113.19252805709839,
+      "epoch": 0.33568904593639576,
+      "grad_norm": 1.4655897617340088,
+      "kl": 1.2039306640625,
+      "learning_rate": 1.6849076713469914e-05,
+      "loss": 0.0481,
+      "reward": 1.2165179058909417,
+      "reward_std": 0.29555373433977367,
+      "rewards/accuracy_reward": 0.2671875121071935,
+      "rewards/format_reward": 0.9493304006755352,
+      "step": 95
+    },
+    {
+      "completion_length": 86.61194581985474,
+      "epoch": 0.35335689045936397,
+      "grad_norm": 0.3733413517475128,
+      "kl": 0.6366943359375,
+      "learning_rate": 1.6385668960932143e-05,
+      "loss": 0.0255,
+      "reward": 1.1428571932017804,
+      "reward_std": 0.2961823304183781,
+      "rewards/accuracy_reward": 0.24252233356237413,
+      "rewards/format_reward": 0.9003348629921675,
+      "step": 100
+    },
+    {
+      "epoch": 0.35335689045936397,
+      "eval_completion_length": 124.61478996276855,
+      "eval_kl": 0.7275390625,
+      "eval_loss": 0.02880026213824749,
+      "eval_reward": 1.2354911267757416,
+      "eval_reward_std": 0.30917802825570107,
+      "eval_rewards/accuracy_reward": 0.2946428656578064,
+      "eval_rewards/format_reward": 0.940848246216774,
+      "eval_runtime": 36.0891,
+      "eval_samples_per_second": 2.743,
+      "eval_steps_per_second": 0.028,
+      "step": 100
+    },
+    {
+      "completion_length": 183.94967460632324,
+      "epoch": 0.3710247349823322,
+      "grad_norm": 0.29940587282180786,
+      "kl": 0.660888671875,
+      "learning_rate": 1.5897847131705194e-05,
+      "loss": 0.0264,
+      "reward": 1.2210938051342963,
+      "reward_std": 0.3049804212525487,
+      "rewards/accuracy_reward": 0.27265626164153217,
+      "rewards/format_reward": 0.9484375402331352,
+      "step": 105
+    },
+    {
+      "completion_length": 231.64610500335692,
+      "epoch": 0.38869257950530034,
+      "grad_norm": 1.9811211824417114,
+      "kl": 0.8390380859375,
+      "learning_rate": 1.5387476295779737e-05,
+      "loss": 0.0336,
+      "reward": 1.2466518417000771,
+      "reward_std": 0.3104502111673355,
+      "rewards/accuracy_reward": 0.3001116213388741,
+      "rewards/format_reward": 0.9465402208268643,
+      "step": 110
+    },
+    {
+      "completion_length": 193.47433891296387,
+      "epoch": 0.40636042402826855,
+      "grad_norm": 1.2147762775421143,
+      "kl": 1.2443359375,
+      "learning_rate": 1.4856507733875837e-05,
+      "loss": 0.0498,
+      "reward": 1.2349330991506577,
+      "reward_std": 0.30721699353307486,
+      "rewards/accuracy_reward": 0.2930803688708693,
+      "rewards/format_reward": 0.941852717846632,
+      "step": 115
+    },
+    {
+      "completion_length": 153.606591796875,
+      "epoch": 0.42402826855123676,
+      "grad_norm": 0.3809666931629181,
+      "kl": 0.4972412109375,
+      "learning_rate": 1.4306971477188223e-05,
+      "loss": 0.0199,
+      "reward": 1.2648438036441803,
+      "reward_std": 0.290036397613585,
+      "rewards/accuracy_reward": 0.3108259065076709,
+      "rewards/format_reward": 0.9540178947150707,
+      "step": 120
+    },
+    {
+      "completion_length": 145.02165756225585,
+      "epoch": 0.4416961130742049,
+      "grad_norm": 0.4991630017757416,
+      "kl": 0.4145751953125,
+      "learning_rate": 1.3740968546047935e-05,
+      "loss": 0.0166,
+      "reward": 1.2626116588711738,
+      "reward_std": 0.26841245898976923,
+      "rewards/accuracy_reward": 0.3021205497905612,
+      "rewards/format_reward": 0.9604911148548126,
+      "step": 125
+    },
+    {
+      "completion_length": 163.11552076339723,
+      "epoch": 0.45936395759717313,
+      "grad_norm": 0.31778135895729065,
+      "kl": 0.36231689453125,
+      "learning_rate": 1.3160662917174045e-05,
+      "loss": 0.0145,
+      "reward": 1.285379521548748,
+      "reward_std": 0.2959928734228015,
+      "rewards/accuracy_reward": 0.3406250147148967,
+      "rewards/format_reward": 0.9447545073926449,
+      "step": 130
+    },
+    {
+      "completion_length": 141.67891292572023,
+      "epoch": 0.47703180212014135,
+      "grad_norm": 0.34590408205986023,
+      "kl": 0.4644775390625,
+      "learning_rate": 1.2568273250226681e-05,
+      "loss": 0.0186,
+      "reward": 1.2870536252856255,
+      "reward_std": 0.26789645981043575,
+      "rewards/accuracy_reward": 0.33683037385344505,
+      "rewards/format_reward": 0.9502232536673546,
+      "step": 135
+    },
+    {
+      "completion_length": 144.7002299308777,
+      "epoch": 0.49469964664310956,
+      "grad_norm": 0.44773581624031067,
+      "kl": 0.4588134765625,
+      "learning_rate": 1.1966064405292887e-05,
+      "loss": 0.0184,
+      "reward": 1.2602679148316382,
+      "reward_std": 0.2673138614743948,
+      "rewards/accuracy_reward": 0.3029018001630902,
+      "rewards/format_reward": 0.9573661126196384,
+      "step": 140
+    },
+    {
+      "completion_length": 160.48873472213745,
+      "epoch": 0.5123674911660777,
+      "grad_norm": 0.3488762080669403,
+      "kl": 0.369140625,
+      "learning_rate": 1.1356338783736256e-05,
+      "loss": 0.0148,
+      "reward": 1.2993304133415222,
+      "reward_std": 0.27455081362277267,
+      "rewards/accuracy_reward": 0.3327009085565805,
+      "rewards/format_reward": 0.9666295036673546,
+      "step": 145
+    },
+    {
+      "completion_length": 171.54386882781984,
+      "epoch": 0.5300353356890459,
+      "grad_norm": 0.5754287242889404,
+      "kl": 0.3918701171875,
+      "learning_rate": 1.0741427525516463e-05,
+      "loss": 0.0157,
+      "reward": 1.3091518446803092,
+      "reward_std": 0.26976782493293283,
+      "rewards/accuracy_reward": 0.34665180258452893,
+      "rewards/format_reward": 0.9625000424683094,
+      "step": 150
+    },
+    {
+      "completion_length": 163.8817042350769,
+      "epoch": 0.5477031802120141,
+      "grad_norm": 0.24734921753406525,
+      "kl": 0.4376220703125,
+      "learning_rate": 1.012368159663363e-05,
+      "loss": 0.0175,
+      "reward": 1.3056920334696769,
+      "reward_std": 0.2626989148557186,
+      "rewards/accuracy_reward": 0.33872769456356766,
+      "rewards/format_reward": 0.9669643260538578,
+      "step": 155
+    },
+    {
+      "completion_length": 214.48315677642822,
+      "epoch": 0.5653710247349824,
+      "grad_norm": 0.39896252751350403,
+      "kl": 0.36552734375,
+      "learning_rate": 9.505462800772612e-06,
+      "loss": 0.0146,
+      "reward": 1.2992188043892383,
+      "reward_std": 0.3003792591392994,
+      "rewards/accuracy_reward": 0.34944198122248055,
+      "rewards/format_reward": 0.9497768260538578,
+      "step": 160
+    },
+    {
+      "completion_length": 212.60224132537843,
+      "epoch": 0.5830388692579506,
+      "grad_norm": 0.3877353072166443,
+      "kl": 0.3916259765625,
+      "learning_rate": 8.889134749511956e-06,
+      "loss": 0.0157,
+      "reward": 1.2678571984171867,
+      "reward_std": 0.33237045761197803,
+      "rewards/accuracy_reward": 0.329464303329587,
+      "rewards/format_reward": 0.9383928969502449,
+      "step": 165
+    },
+    {
+      "completion_length": 193.6290256500244,
+      "epoch": 0.6007067137809188,
+      "grad_norm": 0.2638029456138611,
+      "kl": 0.4411865234375,
+      "learning_rate": 8.277053825620836e-06,
+      "loss": 0.0176,
+      "reward": 1.2617187976837159,
+      "reward_std": 0.3131700936704874,
+      "rewards/accuracy_reward": 0.32912948057055474,
+      "rewards/format_reward": 0.9325893305242061,
+      "step": 170
+    },
+    {
+      "completion_length": 163.14465045928955,
+      "epoch": 0.6183745583038869,
+      "grad_norm": 0.3375968933105469,
+      "kl": 0.792431640625,
+      "learning_rate": 7.671560173993588e-06,
+      "loss": 0.0317,
+      "reward": 1.3142857745289802,
+      "reward_std": 0.2933783018961549,
+      "rewards/accuracy_reward": 0.3618303745985031,
+      "rewards/format_reward": 0.9524553991854191,
+      "step": 175
+    },
+    {
+      "completion_length": 168.05692710876465,
+      "epoch": 0.6360424028268551,
+      "grad_norm": 0.3316074311733246,
+      "kl": 0.4262451171875,
+      "learning_rate": 7.07496875466589e-06,
+      "loss": 0.0171,
+      "reward": 1.2792411267757415,
+      "reward_std": 0.32045796802267434,
+      "rewards/accuracy_reward": 0.3372768020257354,
+      "rewards/format_reward": 0.9419643253087997,
+      "step": 180
+    },
+    {
+      "completion_length": 162.42344417572022,
+      "epoch": 0.6537102473498233,
+      "grad_norm": 0.3786272704601288,
+      "kl": 0.5138427734375,
+      "learning_rate": 6.489560492119225e-06,
+      "loss": 0.0206,
+      "reward": 1.2774554193019867,
+      "reward_std": 0.31533648930490016,
+      "rewards/accuracy_reward": 0.33761162012815477,
+      "rewards/format_reward": 0.9398437954485417,
+      "step": 185
+    },
+    {
+      "completion_length": 138.84576539993287,
+      "epoch": 0.6713780918727915,
+      "grad_norm": 0.4081542491912842,
+      "kl": 0.4474365234375,
+      "learning_rate": 5.9175735547120975e-06,
+      "loss": 0.0179,
+      "reward": 1.3568080931901931,
+      "reward_std": 0.2501720578409731,
+      "rewards/accuracy_reward": 0.38180805277079344,
+      "rewards/format_reward": 0.9750000454485417,
+      "step": 190
+    },
+    {
+      "completion_length": 187.03762016296386,
+      "epoch": 0.6890459363957597,
+      "grad_norm": 0.4072677493095398,
+      "kl": 0.454052734375,
+      "learning_rate": 5.361194797579108e-06,
+      "loss": 0.0182,
+      "reward": 1.306808091700077,
+      "reward_std": 0.3165676988661289,
+      "rewards/accuracy_reward": 0.3659598369151354,
+      "rewards/format_reward": 0.9408482536673546,
+      "step": 195
+    },
+    {
+      "completion_length": 196.41183795928956,
+      "epoch": 0.7067137809187279,
+      "grad_norm": 0.2528441846370697,
+      "kl": 0.410107421875,
+      "learning_rate": 4.8225514017138205e-06,
+      "loss": 0.0164,
+      "reward": 1.3114955961704253,
+      "reward_std": 0.324984360858798,
+      "rewards/accuracy_reward": 0.36796876452863214,
+      "rewards/format_reward": 0.9435268275439739,
+      "step": 200
+    },
+    {
+      "epoch": 0.7067137809187279,
+      "eval_completion_length": 190.1671905517578,
+      "eval_kl": 0.3974609375,
+      "eval_loss": 0.013840513303875923,
+      "eval_reward": 1.3738839626312256,
+      "eval_reward_std": 0.3043372705578804,
+      "eval_rewards/accuracy_reward": 0.4107143059372902,
+      "eval_rewards/format_reward": 0.9631696939468384,
+      "eval_runtime": 41.1198,
+      "eval_samples_per_second": 2.408,
+      "eval_steps_per_second": 0.024,
+      "step": 200
+    },
+    {
+      "completion_length": 187.42009782791138,
+      "epoch": 0.7243816254416962,
+      "grad_norm": 0.5058914422988892,
+      "kl": 0.42244873046875,
+      "learning_rate": 4.303702741201431e-06,
+      "loss": 0.0169,
+      "reward": 1.3109375566244126,
+      "reward_std": 0.309923352021724,
+      "rewards/accuracy_reward": 0.35669644493609665,
+      "rewards/format_reward": 0.9542411133646965,
+      "step": 205
+    },
+    {
+      "completion_length": 196.0847183227539,
+      "epoch": 0.7420494699646644,
+      "grad_norm": 0.4528452157974243,
+      "kl": 0.459375,
+      "learning_rate": 3.8066325096949153e-06,
+      "loss": 0.0184,
+      "reward": 1.2954241663217545,
+      "reward_std": 0.3209618354216218,
+      "rewards/accuracy_reward": 0.354799123480916,
+      "rewards/format_reward": 0.9406250424683094,
+      "step": 210
+    },
+    {
+      "completion_length": 182.6784683227539,
+      "epoch": 0.7597173144876325,
+      "grad_norm": 0.5112515687942505,
+      "kl": 0.433740234375,
+      "learning_rate": 3.3332411362372063e-06,
+      "loss": 0.0174,
+      "reward": 1.31026791036129,
+      "reward_std": 0.32117727063596246,
+      "rewards/accuracy_reward": 0.3594866218045354,
+      "rewards/format_reward": 0.9507812909781933,
+      "step": 215
+    },
+    {
+      "completion_length": 200.36418285369874,
+      "epoch": 0.7773851590106007,
+      "grad_norm": 0.31226980686187744,
+      "kl": 0.4846923828125,
+      "learning_rate": 2.8853385194256677e-06,
+      "loss": 0.0194,
+      "reward": 1.273549161851406,
+      "reward_std": 0.334391163662076,
+      "rewards/accuracy_reward": 0.3455357301980257,
+      "rewards/format_reward": 0.9280134335160255,
+      "step": 220
+    },
+    {
+      "completion_length": 183.47835721969605,
+      "epoch": 0.7950530035335689,
+      "grad_norm": 0.437847375869751,
+      "kl": 0.418212890625,
+      "learning_rate": 2.464637107698046e-06,
+      "loss": 0.0167,
+      "reward": 1.3132812976837158,
+      "reward_std": 0.3066121691837907,
+      "rewards/accuracy_reward": 0.36618305146694186,
+      "rewards/format_reward": 0.9470982573926449,
+      "step": 225
+    },
+    {
+      "completion_length": 199.54431734085082,
+      "epoch": 0.8127208480565371,
+      "grad_norm": 0.35982805490493774,
+      "kl": 0.4347900390625,
+      "learning_rate": 2.072745352195794e-06,
+      "loss": 0.0174,
+      "reward": 1.286272370815277,
+      "reward_std": 0.333691646438092,
+      "rewards/accuracy_reward": 0.3571428755298257,
+      "rewards/format_reward": 0.9291295044124126,
+      "step": 230
+    },
+    {
+      "completion_length": 201.60123605728148,
+      "epoch": 0.8303886925795053,
+      "grad_norm": 0.4822141230106354,
+      "kl": 0.4885009765625,
+      "learning_rate": 1.7111615572361628e-06,
+      "loss": 0.0195,
+      "reward": 1.281696480512619,
+      "reward_std": 0.36514539532363416,
+      "rewards/accuracy_reward": 0.36071430388838055,
+      "rewards/format_reward": 0.9209821879863739,
+      "step": 235
+    },
+    {
+      "completion_length": 191.80257606506348,
+      "epoch": 0.8480565371024735,
+      "grad_norm": 0.3629634380340576,
+      "kl": 0.4655517578125,
+      "learning_rate": 1.381268151904298e-06,
+      "loss": 0.0186,
+      "reward": 1.3007813036441802,
+      "reward_std": 0.33475553449243306,
+      "rewards/accuracy_reward": 0.3681919816881418,
+      "rewards/format_reward": 0.9325893260538578,
+      "step": 240
+    },
+    {
+      "completion_length": 194.34130392074584,
+      "epoch": 0.8657243816254417,
+      "grad_norm": 0.42891520261764526,
+      "kl": 0.462939453125,
+      "learning_rate": 1.0843264046665558e-06,
+      "loss": 0.0185,
+      "reward": 1.302232201397419,
+      "reward_std": 0.3392474535852671,
+      "rewards/accuracy_reward": 0.3680803723633289,
+      "rewards/format_reward": 0.9341518305242061,
+      "step": 245
+    },
+    {
+      "completion_length": 200.85045433044434,
+      "epoch": 0.8833922261484098,
+      "grad_norm": 0.4686770737171173,
+      "kl": 0.4640380859375,
+      "learning_rate": 8.214716012124491e-07,
+      "loss": 0.0186,
+      "reward": 1.3125000596046448,
+      "reward_std": 0.3251642185263336,
+      "rewards/accuracy_reward": 0.3733259098604321,
+      "rewards/format_reward": 0.9391741506755352,
+      "step": 250
+    },
+    {
+      "completion_length": 194.45402736663817,
+      "epoch": 0.901060070671378,
+      "grad_norm": 0.854424774646759,
+      "kl": 0.4887451171875,
+      "learning_rate": 5.937087039615619e-07,
+      "loss": 0.0196,
+      "reward": 1.3106027349829674,
+      "reward_std": 0.32200891636312007,
+      "rewards/accuracy_reward": 0.3715401954948902,
+      "rewards/format_reward": 0.9390625439584255,
+      "step": 255
+    },
+    {
+      "completion_length": 201.23203868865966,
+      "epoch": 0.9187279151943463,
+      "grad_norm": 0.4338059723377228,
+      "kl": 0.49970703125,
+      "learning_rate": 4.019085098303077e-07,
+      "loss": 0.02,
+      "reward": 1.3261161267757415,
+      "reward_std": 0.33655967731028796,
+      "rewards/accuracy_reward": 0.39051341228187086,
+      "rewards/format_reward": 0.9356027200818062,
+      "step": 260
+    },
+    {
+      "completion_length": 204.1974422454834,
+      "epoch": 0.9363957597173145,
+      "grad_norm": 0.5758671164512634,
+      "kl": 0.49468994140625,
+      "learning_rate": 2.4680432094837394e-07,
+      "loss": 0.0198,
+      "reward": 1.300446479022503,
+      "reward_std": 0.3384882753714919,
+      "rewards/accuracy_reward": 0.3676339445635676,
+      "rewards/format_reward": 0.9328125424683094,
+      "step": 265
+    },
+    {
+      "completion_length": 199.08661632537843,
+      "epoch": 0.9540636042402827,
+      "grad_norm": 0.5108721852302551,
+      "kl": 0.45732421875,
+      "learning_rate": 1.289891410535593e-07,
+      "loss": 0.0183,
+      "reward": 1.3210938066244124,
+      "reward_std": 0.33634756207466127,
+      "rewards/accuracy_reward": 0.3814732311293483,
+      "rewards/format_reward": 0.939620578289032,
+      "step": 270
+    },
+    {
+      "completion_length": 200.83203945159912,
+      "epoch": 0.9717314487632509,
+      "grad_norm": 0.40297839045524597,
+      "kl": 0.4919921875,
+      "learning_rate": 4.8913408283934874e-08,
+      "loss": 0.0197,
+      "reward": 1.302455411851406,
+      "reward_std": 0.33796050902456043,
+      "rewards/accuracy_reward": 0.37109376695007085,
+      "rewards/format_reward": 0.9313616514205932,
+      "step": 275
+    },
+    {
+      "completion_length": 205.692866897583,
+      "epoch": 0.9893992932862191,
+      "grad_norm": 0.4729168713092804,
+      "kl": 0.4669677734375,
+      "learning_rate": 6.883273035447335e-09,
+      "loss": 0.0187,
+      "reward": 1.3194196969270706,
+      "reward_std": 0.3483607778325677,
+      "rewards/accuracy_reward": 0.38526787627488374,
+      "rewards/format_reward": 0.9341518260538578,
+      "step": 280
+    },
+    {
+      "completion_length": 200.4055379231771,
+      "epoch": 1.0,
+      "kl": 0.4570719401041667,
+      "reward": 1.3089658319950104,
+      "reward_std": 0.32773972923556965,
+      "rewards/accuracy_reward": 0.3723958491658171,
+      "rewards/format_reward": 0.9365699800352255,
+      "step": 283,
+      "total_flos": 0.0,
+      "train_loss": 0.027384539475202127,
+      "train_runtime": 37232.0457,
+      "train_samples_per_second": 1.946,
+      "train_steps_per_second": 0.008
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 283,
+  "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": 16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9bf478757144733b85efa7561d3d4fa41cf506de6534c1eb730a3131556c7e4
+size 7544