commit 23e0010cda080690815f345f50fe63434d8a9422 Author: ModelHub XC Date: Tue Jun 16 05:36:17 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: harsha070/expfinal-qwen-island-s42-lambda-0p0 Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..52373fe --- /dev/null +++ b/.gitattributes @@ -0,0 +1,36 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz 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 diff --git a/README.md b/README.md new file mode 100644 index 0000000..d57f07b --- /dev/null +++ b/README.md @@ -0,0 +1,67 @@ +--- +base_model: Qwen/Qwen2.5-3B-Instruct +library_name: transformers +model_name: expfinal-qwen-island-s42-lambda-0p0 +tags: +- generated_from_trainer +- grpo +- trl +licence: license +--- + +# Model Card for expfinal-qwen-island-s42-lambda-0p0 + +This model is a fine-tuned version of [Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-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="harsha070/expfinal-qwen-island-s42-lambda-0p0", device="cuda") +output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0] +print(output["generated_text"]) +``` + +## Training procedure + +[Visualize in Weights & Biases](https://wandb.ai/models-self5933/obfuscation-early-warning/runs/w7r1ivm1) + + + +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: 1.3.0 +- Transformers: 5.7.0 +- Pytorch: 2.11.0 +- Datasets: 4.8.5 +- Tokenizers: 0.22.2 + +## Citations + +Cite GRPO as: + +```bibtex +@article{shao2024deepseekmath, + 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 +@software{vonwerra2020trl, + title = {{TRL: Transformers Reinforcement Learning}}, + author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin}, + license = {Apache-2.0}, + url = {https://github.com/huggingface/trl}, + year = {2020} +} +``` \ No newline at end of file diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\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 XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/completions/completions_00010.parquet b/completions/completions_00010.parquet new file mode 100644 index 0000000..7509291 --- /dev/null +++ b/completions/completions_00010.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:709d5f1748fb55b0462c6b61c3eb2f9ec74ea16c8373b0d32ee0a618c6311bd3 +size 18500 diff --git a/completions/completions_00020.parquet b/completions/completions_00020.parquet new file mode 100644 index 0000000..66c104e --- /dev/null +++ b/completions/completions_00020.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a9b70299be669593867088632af6e2ac9332993402b6d23582cdcb161b161245 +size 20219 diff --git a/completions/completions_00030.parquet b/completions/completions_00030.parquet new file mode 100644 index 0000000..1e63307 --- /dev/null +++ b/completions/completions_00030.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19274048def0678aaae235dccfee33e64691ef8225a4ae5b873f643b0c47eb8b +size 24563 diff --git a/completions/completions_00040.parquet b/completions/completions_00040.parquet new file mode 100644 index 0000000..f81cbe1 --- /dev/null +++ b/completions/completions_00040.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1c2b9f7c4d5742a82eada06a275d0b18e30c6a6e2692f129977cf33eb9d08fd3 +size 21514 diff --git a/completions/completions_00050.parquet b/completions/completions_00050.parquet new file mode 100644 index 0000000..75ff766 --- /dev/null +++ b/completions/completions_00050.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3f45b4f7cff64bbd18c32d94b096828c57f4f3193c6c5d7f9a9545d9af050fc +size 27209 diff --git a/completions/completions_00060.parquet b/completions/completions_00060.parquet new file mode 100644 index 0000000..f727530 --- /dev/null +++ b/completions/completions_00060.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5071d1f07f49831a7cc85dc41bcd5ad3ae4aa7b14562b82932039fc747e85d0d +size 22894 diff --git a/completions/completions_00070.parquet b/completions/completions_00070.parquet new file mode 100644 index 0000000..01fdc1c --- /dev/null +++ b/completions/completions_00070.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3792f5cb2075ce7b444c54e237529a0e0a0161920941e25f7b92b89ff062db38 +size 21338 diff --git a/completions/completions_00080.parquet b/completions/completions_00080.parquet new file mode 100644 index 0000000..e065101 --- /dev/null +++ b/completions/completions_00080.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:00825a02117e4ca226ae94bf40d21eb6086a3d101f72e69131af83cb1141d9aa +size 21057 diff --git a/completions/completions_00090.parquet b/completions/completions_00090.parquet new file mode 100644 index 0000000..e31b0e8 --- /dev/null +++ b/completions/completions_00090.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ed71a7755dfa973f55a514a2d2cf7015b669ae63e8e0960e3cd9ce513dead6ac +size 22697 diff --git a/completions/completions_00100.parquet b/completions/completions_00100.parquet new file mode 100644 index 0000000..e19779b --- /dev/null +++ b/completions/completions_00100.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6c0d4224c99a311dd866c863871ca09f9a168096ee669c8a66b19ba67240f7f0 +size 24068 diff --git a/completions/completions_00110.parquet b/completions/completions_00110.parquet new file mode 100644 index 0000000..20c89a5 --- /dev/null +++ b/completions/completions_00110.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b41e1b8759b1b1000e36d3e74fa8da3067eb4b2274eb1b827863de954ab091c +size 24804 diff --git a/completions/completions_00120.parquet b/completions/completions_00120.parquet new file mode 100644 index 0000000..a804eb4 --- /dev/null +++ b/completions/completions_00120.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:671507b63f1e5f22f963785f15d17e6decca4359f123c5e0cc50ad6d43c56104 +size 22330 diff --git a/completions/completions_00130.parquet b/completions/completions_00130.parquet new file mode 100644 index 0000000..e4e9721 --- /dev/null +++ b/completions/completions_00130.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c943d189936de8af274d38829e2286c2d6fad1e4d26d253db1153912abfb6f6 +size 23072 diff --git a/completions/completions_00140.parquet b/completions/completions_00140.parquet new file mode 100644 index 0000000..9d5ac40 --- /dev/null +++ b/completions/completions_00140.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c8d8430dcfd67ff33bd0570cb856d84a45f7f1298dc933af80602870837d49f2 +size 24019 diff --git a/completions/completions_00150.parquet b/completions/completions_00150.parquet new file mode 100644 index 0000000..a3a0b8e --- /dev/null +++ b/completions/completions_00150.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4d937515c8720c05d349bab053159685dfae3616027ac87867ea9e90dc0ff5bb +size 21606 diff --git a/completions/completions_00160.parquet b/completions/completions_00160.parquet new file mode 100644 index 0000000..a050d70 --- /dev/null +++ b/completions/completions_00160.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:79e27f7525ffcf770056b275a4fe9791033657edd2cd2c86d1bb1a11bfd6ff20 +size 24367 diff --git a/config.json b/config.json new file mode 100644 index 0000000..64c100d --- /dev/null +++ b/config.json @@ -0,0 +1,69 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 2048, + "initializer_range": 0.02, + "intermediate_size": 11008, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 70, + "model_type": "qwen2", + "num_attention_heads": 16, + "num_hidden_layers": 36, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.7.0", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..aaf8639 --- /dev/null +++ b/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.05, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.7.0" +} diff --git a/last-checkpoint/chat_template.jinja b/last-checkpoint/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/last-checkpoint/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\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 XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/last-checkpoint/config.json b/last-checkpoint/config.json new file mode 100644 index 0000000..64c100d --- /dev/null +++ b/last-checkpoint/config.json @@ -0,0 +1,69 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "bfloat16", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 2048, + "initializer_range": 0.02, + "intermediate_size": 11008, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 70, + "model_type": "qwen2", + "num_attention_heads": 16, + "num_hidden_layers": 36, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.7.0", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/last-checkpoint/generation_config.json b/last-checkpoint/generation_config.json new file mode 100644 index 0000000..aaf8639 --- /dev/null +++ b/last-checkpoint/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.05, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.7.0" +} diff --git a/last-checkpoint/model.safetensors b/last-checkpoint/model.safetensors new file mode 100644 index 0000000..c39a6d8 --- /dev/null +++ b/last-checkpoint/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19ec7b8100fdbb732326b9938aaf5de706ac0cd9a0aa6fbb12e56ff42f02fc49 +size 6171927112 diff --git a/last-checkpoint/tokenizer.json b/last-checkpoint/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/last-checkpoint/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/last-checkpoint/tokenizer_config.json b/last-checkpoint/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/last-checkpoint/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_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|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/last-checkpoint/trainer_state.json b/last-checkpoint/trainer_state.json new file mode 100644 index 0000000..1a75489 --- /dev/null +++ b/last-checkpoint/trainer_state.json @@ -0,0 +1,482 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.28, + "eval_steps": 500, + "global_step": 160, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.05, + "completions/max_length": 894.8, + "completions/max_terminated_length": 821.2, + "completions/mean_length": 560.2, + "completions/mean_terminated_length": 540.6345306396485, + "completions/min_length": 279.5, + "completions/min_terminated_length": 279.5, + "entropy": 0.38268125932663677, + "epoch": 0.08, + "frac_reward_zero_std": 0.75, + "grad_norm": 1.4375, + "kl": 0.07789193278222228, + "learning_rate": 9.4375e-06, + "loss": -0.007836591452360153, + "num_tokens": 65460.0, + "reward": 0.65, + "reward_std": 0.46797851026058196, + "rewards/JointRewardFunction/mean": 0.65, + "rewards/JointRewardFunction/std": 0.4679785281419754, + "step": 10, + "step_time": 36.23508502000004 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.075, + "completions/max_length": 900.1, + "completions/max_terminated_length": 823.7, + "completions/mean_length": 578.525, + "completions/mean_terminated_length": 543.8116729736328, + "completions/min_length": 254.6, + "completions/min_terminated_length": 254.6, + "entropy": 0.23411482032388448, + "epoch": 0.16, + "frac_reward_zero_std": 0.775, + "grad_norm": 1.1484375, + "kl": 0.1648747116792947, + "learning_rate": 8.8125e-06, + "loss": 0.0052785202860832214, + "num_tokens": 132386.0, + "reward": 0.6625, + "reward_std": 0.46628117859363555, + "rewards/JointRewardFunction/mean": 0.6625, + "rewards/JointRewardFunction/std": 0.4662812024354935, + "step": 20, + "step_time": 37.938748809599566 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025, + "completions/max_length": 876.9, + "completions/max_terminated_length": 829.5, + "completions/mean_length": 622.2, + "completions/mean_terminated_length": 611.7642944335937, + "completions/min_length": 477.6, + "completions/min_terminated_length": 477.6, + "entropy": 0.22628286899998784, + "epoch": 0.24, + "frac_reward_zero_std": 0.875, + "grad_norm": 0.0218505859375, + "kl": 0.12199527090415359, + "learning_rate": 8.1875e-06, + "loss": 0.007959160953760147, + "num_tokens": 202606.0, + "reward": 0.9375, + "reward_std": 0.12246559262275696, + "rewards/JointRewardFunction/mean": 0.9375, + "rewards/JointRewardFunction/std": 0.12246559858322144, + "step": 30, + "step_time": 37.057444848399975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025, + "completions/max_length": 810.6, + "completions/max_terminated_length": 731.4, + "completions/mean_length": 596.8625, + "completions/mean_terminated_length": 584.707144165039, + "completions/min_length": 438.4, + "completions/min_terminated_length": 438.4, + "entropy": 0.42070485297590493, + "epoch": 0.32, + "frac_reward_zero_std": 0.85, + "grad_norm": 1.1015625, + "kl": 0.10627949037589132, + "learning_rate": 7.5625e-06, + "loss": -0.00913204848766327, + "num_tokens": 270929.0, + "reward": 0.9, + "reward_std": 0.20411193668842315, + "rewards/JointRewardFunction/mean": 0.9, + "rewards/JointRewardFunction/std": 0.20411194264888763, + "step": 40, + "step_time": 34.16536340820039 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.05, + "completions/max_length": 916.6, + "completions/max_terminated_length": 800.7, + "completions/mean_length": 642.275, + "completions/mean_terminated_length": 620.6982238769531, + "completions/min_length": 483.4, + "completions/min_terminated_length": 483.4, + "entropy": 0.545534435659647, + "epoch": 0.4, + "frac_reward_zero_std": 0.825, + "grad_norm": 0.9921875, + "kl": 0.09743389897048474, + "learning_rate": 6.9375e-06, + "loss": 0.019783291220664977, + "num_tokens": 343141.0, + "reward": 0.9125, + "reward_std": 0.19864802658557892, + "rewards/JointRewardFunction/mean": 0.9125, + "rewards/JointRewardFunction/std": 0.19864802658557892, + "step": 50, + "step_time": 38.55176728389906 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 774.9, + "completions/max_terminated_length": 774.9, + "completions/mean_length": 599.525, + "completions/mean_terminated_length": 599.525, + "completions/min_length": 438.8, + "completions/min_terminated_length": 438.8, + "entropy": 0.5808290097862482, + "epoch": 0.48, + "frac_reward_zero_std": 0.925, + "grad_norm": 0.035400390625, + "kl": 0.10702053690329194, + "learning_rate": 6.3125e-06, + "loss": -0.009540864825248718, + "num_tokens": 411443.0, + "reward": 0.9625, + "reward_std": 0.10606601536273956, + "rewards/JointRewardFunction/mean": 0.9625, + "rewards/JointRewardFunction/std": 0.10606601536273956, + "step": 60, + "step_time": 33.111361246699744 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0375, + "completions/max_length": 848.1, + "completions/max_terminated_length": 757.4, + "completions/mean_length": 626.325, + "completions/mean_terminated_length": 609.8857238769531, + "completions/min_length": 511.5, + "completions/min_terminated_length": 511.5, + "entropy": 0.4436331996694207, + "epoch": 0.56, + "frac_reward_zero_std": 0.9, + "grad_norm": 0.060791015625, + "kl": 0.08952742610126734, + "learning_rate": 5.6875e-06, + "loss": 0.014259077608585358, + "num_tokens": 482005.0, + "reward": 0.95, + "reward_std": 0.11700168251991272, + "rewards/JointRewardFunction/mean": 0.95, + "rewards/JointRewardFunction/std": 0.11700168251991272, + "step": 70, + "step_time": 35.991650615099935 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0125, + "completions/max_length": 787.7, + "completions/max_terminated_length": 747.8, + "completions/mean_length": 593.25, + "completions/mean_terminated_length": 587.5642883300782, + "completions/min_length": 493.6, + "completions/min_terminated_length": 493.6, + "entropy": 0.328166064620018, + "epoch": 0.64, + "frac_reward_zero_std": 0.95, + "grad_norm": 0.017822265625, + "kl": 0.1307119549252093, + "learning_rate": 5.0625e-06, + "loss": 0.0031396135687828063, + "num_tokens": 550079.0, + "reward": 0.975, + "reward_std": 0.07071067690849304, + "rewards/JointRewardFunction/mean": 0.975, + "rewards/JointRewardFunction/std": 0.07071067690849304, + "step": 80, + "step_time": 33.55307105900029 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 788.4, + "completions/max_terminated_length": 788.4, + "completions/mean_length": 619.825, + "completions/mean_terminated_length": 619.825, + "completions/min_length": 510.4, + "completions/min_terminated_length": 510.4, + "entropy": 0.4017201948910952, + "epoch": 0.72, + "frac_reward_zero_std": 0.925, + "grad_norm": 0.0172119140625, + "kl": 0.08959094756282866, + "learning_rate": 4.4375e-06, + "loss": 0.001467562187463045, + "num_tokens": 620285.0, + "reward": 0.9625, + "reward_std": 0.10606601536273956, + "rewards/JointRewardFunction/mean": 0.9625, + "rewards/JointRewardFunction/std": 0.10606601536273956, + "step": 90, + "step_time": 33.41816141909967 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.05, + "completions/max_length": 911.0, + "completions/max_terminated_length": 831.5, + "completions/mean_length": 671.025, + "completions/mean_terminated_length": 653.7214416503906, + "completions/min_length": 537.5, + "completions/min_terminated_length": 537.5, + "entropy": 0.4100338226184249, + "epoch": 0.8, + "frac_reward_zero_std": 0.875, + "grad_norm": 0.021240234375, + "kl": 0.07856191159226, + "learning_rate": 3.8125e-06, + "loss": 0.017280958592891693, + "num_tokens": 694633.0, + "reward": 0.9375, + "reward_std": 0.1767766922712326, + "rewards/JointRewardFunction/mean": 0.9375, + "rewards/JointRewardFunction/std": 0.1767766922712326, + "step": 100, + "step_time": 38.310592102500415 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 867.4, + "completions/max_terminated_length": 779.6, + "completions/mean_length": 642.0875, + "completions/mean_terminated_length": 616.8726257324219, + "completions/min_length": 501.5, + "completions/min_terminated_length": 501.5, + "entropy": 0.40164962466806176, + "epoch": 0.88, + "frac_reward_zero_std": 0.925, + "grad_norm": 1.1328125, + "kl": 0.0828359558712691, + "learning_rate": 3.1875e-06, + "loss": 0.009746464341878891, + "num_tokens": 766544.0, + "reward": 0.9625, + "reward_std": 0.0816463440656662, + "rewards/JointRewardFunction/mean": 0.9625, + "rewards/JointRewardFunction/std": 0.0816463440656662, + "step": 110, + "step_time": 36.619024862399236 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0375, + "completions/max_length": 807.3, + "completions/max_terminated_length": 713.2, + "completions/mean_length": 611.7875, + "completions/mean_terminated_length": 595.7928649902344, + "completions/min_length": 508.7, + "completions/min_terminated_length": 508.7, + "entropy": 0.40618473663926125, + "epoch": 0.96, + "frac_reward_zero_std": 0.925, + "grad_norm": 1.03125, + "kl": 0.08407443668693304, + "learning_rate": 2.5625e-06, + "loss": 0.01084473505616188, + "num_tokens": 835859.0, + "reward": 0.9625, + "reward_std": 0.10606601536273956, + "rewards/JointRewardFunction/mean": 0.9625, + "rewards/JointRewardFunction/std": 0.10606601536273956, + "step": 120, + "step_time": 34.16594214020042 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0125, + "completions/max_length": 751.9, + "completions/max_terminated_length": 748.3, + "completions/mean_length": 631.1, + "completions/mean_terminated_length": 626.9517883300781, + "completions/min_length": 544.1, + "completions/min_terminated_length": 544.1, + "entropy": 0.42196682561188936, + "epoch": 1.04, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.041748046875, + "kl": 0.08644672441296279, + "learning_rate": 1.9375e-06, + "loss": 0.00017178469570353628, + "num_tokens": 906965.0, + "reward": 1.0, + "reward_std": 0.0, + "rewards/JointRewardFunction/mean": 1.0, + "rewards/JointRewardFunction/std": 0.0, + "step": 130, + "step_time": 31.96281487729957 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0125, + "completions/max_length": 837.6, + "completions/max_terminated_length": 826.9, + "completions/mean_length": 642.5625, + "completions/mean_terminated_length": 637.825, + "completions/min_length": 522.1, + "completions/min_terminated_length": 522.1, + "entropy": 0.39782516546547414, + "epoch": 1.12, + "frac_reward_zero_std": 0.875, + "grad_norm": 0.9375, + "kl": 0.08095719190314413, + "learning_rate": 1.3125000000000001e-06, + "loss": 0.0059084448963403705, + "num_tokens": 979018.0, + "reward": 0.9375, + "reward_std": 0.1767766922712326, + "rewards/JointRewardFunction/mean": 0.9375, + "rewards/JointRewardFunction/std": 0.1767766922712326, + "step": 140, + "step_time": 35.47339765110009 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025, + "completions/max_length": 780.2, + "completions/max_terminated_length": 709.4, + "completions/mean_length": 608.125, + "completions/mean_terminated_length": 597.3017883300781, + "completions/min_length": 514.2, + "completions/min_terminated_length": 514.2, + "entropy": 0.3825716434046626, + "epoch": 1.2, + "frac_reward_zero_std": 0.9, + "grad_norm": 0.0263671875, + "kl": 0.0812916701193899, + "learning_rate": 6.875000000000001e-07, + "loss": 0.014202636480331422, + "num_tokens": 1048244.0, + "reward": 0.95, + "reward_std": 0.11700168251991272, + "rewards/JointRewardFunction/mean": 0.95, + "rewards/JointRewardFunction/std": 0.11700168251991272, + "step": 150, + "step_time": 33.213279404800595 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025, + "completions/max_length": 785.1, + "completions/max_terminated_length": 743.9, + "completions/mean_length": 621.9, + "completions/mean_terminated_length": 612.2803649902344, + "completions/min_length": 508.9, + "completions/min_terminated_length": 508.9, + "entropy": 0.4208029452711344, + "epoch": 1.28, + "frac_reward_zero_std": 0.95, + "grad_norm": 0.92578125, + "kl": 0.08554110652767122, + "learning_rate": 6.250000000000001e-08, + "loss": 0.011326169967651368, + "num_tokens": 1118898.0, + "reward": 0.975, + "reward_std": 0.07071067690849304, + "rewards/JointRewardFunction/mean": 0.975, + "rewards/JointRewardFunction/std": 0.07071067690849304, + "step": 160, + "step_time": 33.45168144740146 + } + ], + "logging_steps": 10, + "max_steps": 160, + "num_input_tokens_seen": 1118898, + "num_train_epochs": 2, + "save_steps": 10, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/last-checkpoint/training_args.bin b/last-checkpoint/training_args.bin new file mode 100644 index 0000000..40c64c0 --- /dev/null +++ b/last-checkpoint/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d40f250c23d01fe898035903f1d852e03e4dfb5197a6039e86d76bf7e9a8eaa8 +size 7249 diff --git a/model.safetensors b/model.safetensors new file mode 100644 index 0000000..c39a6d8 --- /dev/null +++ b/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19ec7b8100fdbb732326b9938aaf5de706ac0cd9a0aa6fbb12e56ff42f02fc49 +size 6171927112 diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_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|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000..40c64c0 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d40f250c23d01fe898035903f1d852e03e4dfb5197a6039e86d76bf7e9a8eaa8 +size 7249