commit 804993290c934c9f9a187fe7c56de329831ca798 Author: ModelHub XC Date: Tue Jun 16 05:14:16 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: harsha070/expfinal-qwen-island-s42-lambda-0p50 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..97f3313 --- /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-0p50 +tags: +- generated_from_trainer +- grpo +- trl +licence: license +--- + +# Model Card for expfinal-qwen-island-s42-lambda-0p50 + +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-0p50", 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/rottuhnw) + + + +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..152c5b8 --- /dev/null +++ b/completions/completions_00010.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7bdd0558ee31f69dcef0e047620e18b2b89c28dcfc0ed153e2559d4668f446df +size 24317 diff --git a/completions/completions_00020.parquet b/completions/completions_00020.parquet new file mode 100644 index 0000000..76b8875 --- /dev/null +++ b/completions/completions_00020.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4654c956d4acd8bef59680224087f061d8a93f1c85995e3cd4ee71518836524c +size 21310 diff --git a/completions/completions_00030.parquet b/completions/completions_00030.parquet new file mode 100644 index 0000000..f4fd0df --- /dev/null +++ b/completions/completions_00030.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ec451cb47393fbc2087668544da5e65149d57b9ec7ee25cb365007467e610b3 +size 23320 diff --git a/completions/completions_00040.parquet b/completions/completions_00040.parquet new file mode 100644 index 0000000..03cd97d --- /dev/null +++ b/completions/completions_00040.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd835d1e4bdcdadc21180c21da6634a5333853f82dc2a22f96569cf97335fbf +size 21463 diff --git a/completions/completions_00050.parquet b/completions/completions_00050.parquet new file mode 100644 index 0000000..47b474d --- /dev/null +++ b/completions/completions_00050.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a3b51cf9c72102c43b2d33ab45a47b3f18f0b394fad9dacf5149301789855248 +size 22222 diff --git a/completions/completions_00060.parquet b/completions/completions_00060.parquet new file mode 100644 index 0000000..2ff2cb4 --- /dev/null +++ b/completions/completions_00060.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c3d833af4522b6d246fd99b8048d12144feb2e9970d6380ad2fed8de86195329 +size 23570 diff --git a/completions/completions_00070.parquet b/completions/completions_00070.parquet new file mode 100644 index 0000000..182fc2e --- /dev/null +++ b/completions/completions_00070.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d07dd653bfa6ce0ddc96303c5f50fde6a9dcc1ea388bd43b43cdb7b72774ae6f +size 18972 diff --git a/completions/completions_00080.parquet b/completions/completions_00080.parquet new file mode 100644 index 0000000..b334d66 --- /dev/null +++ b/completions/completions_00080.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:517309bce7ad9fa7859b87ecb9cdf0c3c8cb1193a74f0dbc249410b38d4e9d44 +size 19963 diff --git a/completions/completions_00090.parquet b/completions/completions_00090.parquet new file mode 100644 index 0000000..58c52fc --- /dev/null +++ b/completions/completions_00090.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f978e6b612ac321a31291def0eee55a4aed7d408d86c81a42330f6844ec8c78a +size 19253 diff --git a/completions/completions_00100.parquet b/completions/completions_00100.parquet new file mode 100644 index 0000000..3096bb7 --- /dev/null +++ b/completions/completions_00100.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d6edbb2223e4a42eeb886cc7bf875db5c300ef042c6fe9677fa39f661aca778f +size 21716 diff --git a/completions/completions_00110.parquet b/completions/completions_00110.parquet new file mode 100644 index 0000000..dac97b1 --- /dev/null +++ b/completions/completions_00110.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:db04d88b34288f7d8cc99822a749d160bddabc3401096ddcd9ea3235da977e9b +size 21866 diff --git a/completions/completions_00120.parquet b/completions/completions_00120.parquet new file mode 100644 index 0000000..9667fa0 --- /dev/null +++ b/completions/completions_00120.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c552fa033b8ad870dea41dd8c5a64b96b782ee6060ad1d6751b16e34fecbc7eb +size 20930 diff --git a/completions/completions_00130.parquet b/completions/completions_00130.parquet new file mode 100644 index 0000000..a15b87c --- /dev/null +++ b/completions/completions_00130.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c282ede81636a52bc38248cba7fc6977cbc9f139a79821680c05fd17ad6d6a5 +size 21673 diff --git a/completions/completions_00140.parquet b/completions/completions_00140.parquet new file mode 100644 index 0000000..0d4fd09 --- /dev/null +++ b/completions/completions_00140.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:04552e3b9ba6747e9fbf97f651cfbf041b16ca653b28df9aca318e56287bc5b8 +size 17355 diff --git a/completions/completions_00150.parquet b/completions/completions_00150.parquet new file mode 100644 index 0000000..ded63b5 --- /dev/null +++ b/completions/completions_00150.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7a73a0c4351d15173bdb7b22117abf071f0338d397c0676246b7a623a2b13b7f +size 24100 diff --git a/completions/completions_00160.parquet b/completions/completions_00160.parquet new file mode 100644 index 0000000..d4deb3a --- /dev/null +++ b/completions/completions_00160.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:57f36e5114b168437d03b6ec58afc042f893a85ac2db91d92178c0aacc504ed1 +size 19065 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..4665a23 --- /dev/null +++ b/last-checkpoint/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6032762bcfdf97d0ffef0d438d20af4d7f344890e0668ef6b790396c6cafa1f9 +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..aff6c9a --- /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": 899.7, + "completions/max_terminated_length": 834.6, + "completions/mean_length": 599.55, + "completions/mean_terminated_length": 580.2101257324218, + "completions/min_length": 345.7, + "completions/min_terminated_length": 345.7, + "entropy": 0.9067775651812553, + "epoch": 0.08, + "frac_reward_zero_std": 0.15, + "grad_norm": 2.203125, + "kl": 0.034341241162474036, + "learning_rate": 9.4375e-06, + "loss": -0.016357305645942687, + "num_tokens": 68608.0, + "reward": 0.5690673828125, + "reward_std": 0.5320930480957031, + "rewards/JointRewardFunction/mean": 0.5690673828125, + "rewards/JointRewardFunction/std": 0.5320930480957031, + "step": 10, + "step_time": 36.65178473630149 + }, + { + "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": 915.4, + "completions/max_terminated_length": 890.0, + "completions/mean_length": 654.2625, + "completions/mean_terminated_length": 633.4842956542968, + "completions/min_length": 382.9, + "completions/min_terminated_length": 382.9, + "entropy": 1.2692338980734348, + "epoch": 0.16, + "frac_reward_zero_std": 0.175, + "grad_norm": 2.0625, + "kl": 0.08574351686984301, + "learning_rate": 8.8125e-06, + "loss": 0.011555161327123642, + "num_tokens": 141593.0, + "reward": 0.55255126953125, + "reward_std": 0.5202917702496052, + "rewards/JointRewardFunction/mean": 0.55255126953125, + "rewards/JointRewardFunction/std": 0.5202917769551277, + "step": 20, + "step_time": 38.578124190003294 + }, + { + "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": 831.7, + "completions/max_terminated_length": 792.0, + "completions/mean_length": 567.075, + "completions/mean_terminated_length": 556.1625061035156, + "completions/min_length": 353.2, + "completions/min_terminated_length": 353.2, + "entropy": 1.2077870965003967, + "epoch": 0.24, + "frac_reward_zero_std": 0.15, + "grad_norm": 1.4921875, + "kl": 0.07180349011905492, + "learning_rate": 8.1875e-06, + "loss": 0.021731069684028624, + "num_tokens": 207403.0, + "reward": 0.68575439453125, + "reward_std": 0.4835183143615723, + "rewards/JointRewardFunction/mean": 0.68575439453125, + "rewards/JointRewardFunction/std": 0.48351832032203673, + "step": 30, + "step_time": 35.385694373202206 + }, + { + "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": 899.7, + "completions/max_terminated_length": 841.7, + "completions/mean_length": 571.8875, + "completions/mean_terminated_length": 558.7678588867187, + "completions/min_length": 322.2, + "completions/min_terminated_length": 322.2, + "entropy": 1.0670640360563994, + "epoch": 0.32, + "frac_reward_zero_std": 0.15, + "grad_norm": 1.0859375, + "kl": 0.07311075136531145, + "learning_rate": 7.5625e-06, + "loss": -0.013997538387775421, + "num_tokens": 273728.0, + "reward": 0.74111328125, + "reward_std": 0.5125789776444435, + "rewards/JointRewardFunction/mean": 0.74111328125, + "rewards/JointRewardFunction/std": 0.512578995525837, + "step": 40, + "step_time": 38.20040096230041 + }, + { + "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": 842.8, + "completions/max_terminated_length": 821.3, + "completions/mean_length": 610.425, + "completions/mean_terminated_length": 600.9160766601562, + "completions/min_length": 379.3, + "completions/min_terminated_length": 379.3, + "entropy": 1.1711551163345575, + "epoch": 0.4, + "frac_reward_zero_std": 0.2, + "grad_norm": 1.3203125, + "kl": 0.08198881428688765, + "learning_rate": 6.9375e-06, + "loss": 0.0175897479057312, + "num_tokens": 343392.0, + "reward": 0.69986572265625, + "reward_std": 0.47377224564552306, + "rewards/JointRewardFunction/mean": 0.69986572265625, + "rewards/JointRewardFunction/std": 0.47377225160598757, + "step": 50, + "step_time": 35.86526439490153 + }, + { + "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": 783.2, + "completions/max_terminated_length": 783.2, + "completions/mean_length": 560.7, + "completions/mean_terminated_length": 560.7, + "completions/min_length": 375.6, + "completions/min_terminated_length": 375.6, + "entropy": 1.1601736910641194, + "epoch": 0.48, + "frac_reward_zero_std": 0.15, + "grad_norm": 2.0625, + "kl": 0.07123464278411121, + "learning_rate": 6.3125e-06, + "loss": 0.005981885641813278, + "num_tokens": 408588.0, + "reward": 0.77021484375, + "reward_std": 0.45352104529738424, + "rewards/JointRewardFunction/mean": 0.77021484375, + "rewards/JointRewardFunction/std": 0.4535210572183132, + "step": 60, + "step_time": 33.62324385899883 + }, + { + "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": 859.6, + "completions/max_terminated_length": 791.1, + "completions/mean_length": 576.6875, + "completions/mean_terminated_length": 558.1005981445312, + "completions/min_length": 331.1, + "completions/min_terminated_length": 331.1, + "entropy": 1.0957016348838806, + "epoch": 0.56, + "frac_reward_zero_std": 0.225, + "grad_norm": 1.578125, + "kl": 0.06286744237877429, + "learning_rate": 5.6875e-06, + "loss": 0.02311268150806427, + "num_tokens": 475179.0, + "reward": 0.87257080078125, + "reward_std": 0.4612266093492508, + "rewards/JointRewardFunction/mean": 0.87257080078125, + "rewards/JointRewardFunction/std": 0.4612266182899475, + "step": 70, + "step_time": 36.65166022339836 + }, + { + "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": 801.1, + "completions/max_terminated_length": 801.1, + "completions/mean_length": 529.3, + "completions/mean_terminated_length": 529.3, + "completions/min_length": 325.9, + "completions/min_terminated_length": 325.9, + "entropy": 0.9475126106292009, + "epoch": 0.64, + "frac_reward_zero_std": 0.175, + "grad_norm": 1.6484375, + "kl": 0.0728621585527435, + "learning_rate": 5.0625e-06, + "loss": -0.00028136512264609337, + "num_tokens": 538137.0, + "reward": 0.92275390625, + "reward_std": 0.4701471716165543, + "rewards/JointRewardFunction/mean": 0.92275390625, + "rewards/JointRewardFunction/std": 0.47014717757701874, + "step": 80, + "step_time": 34.54745226030063 + }, + { + "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": 809.0, + "completions/max_terminated_length": 774.7, + "completions/mean_length": 550.9125, + "completions/mean_terminated_length": 544.0785736083984, + "completions/min_length": 338.7, + "completions/min_terminated_length": 338.7, + "entropy": 0.8775507573038339, + "epoch": 0.72, + "frac_reward_zero_std": 0.225, + "grad_norm": 1.6328125, + "kl": 0.0652636706130579, + "learning_rate": 4.4375e-06, + "loss": 0.00958292931318283, + "num_tokens": 602830.0, + "reward": 0.9910400390625, + "reward_std": 0.49945615530014037, + "rewards/JointRewardFunction/mean": 0.9910400390625, + "rewards/JointRewardFunction/std": 0.49945615530014037, + "step": 90, + "step_time": 34.542127643501225 + }, + { + "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": 854.6, + "completions/max_terminated_length": 811.1, + "completions/mean_length": 556.0125, + "completions/mean_terminated_length": 544.571435546875, + "completions/min_length": 342.6, + "completions/min_terminated_length": 342.6, + "entropy": 0.9368259467184543, + "epoch": 0.8, + "frac_reward_zero_std": 0.1, + "grad_norm": 1.5703125, + "kl": 0.06576599276158959, + "learning_rate": 3.8125e-06, + "loss": 0.025794655084609985, + "num_tokens": 667977.0, + "reward": 0.95867919921875, + "reward_std": 0.45399501323699953, + "rewards/JointRewardFunction/mean": 0.95867919921875, + "rewards/JointRewardFunction/std": 0.45399501025676725, + "step": 100, + "step_time": 36.289319235399674 + }, + { + "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": 818.4, + "completions/max_terminated_length": 818.4, + "completions/mean_length": 521.6875, + "completions/mean_terminated_length": 521.6875, + "completions/min_length": 269.7, + "completions/min_terminated_length": 269.7, + "entropy": 0.8984692271798849, + "epoch": 0.88, + "frac_reward_zero_std": 0.15, + "grad_norm": 1.7578125, + "kl": 0.06471977110486478, + "learning_rate": 3.1875e-06, + "loss": 0.011397648602724075, + "num_tokens": 730256.0, + "reward": 0.99234619140625, + "reward_std": 0.42927705943584443, + "rewards/JointRewardFunction/mean": 0.99234619140625, + "rewards/JointRewardFunction/std": 0.4292770564556122, + "step": 110, + "step_time": 35.05004594640013 + }, + { + "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": 848.3, + "completions/max_terminated_length": 848.3, + "completions/mean_length": 537.95, + "completions/mean_terminated_length": 537.95, + "completions/min_length": 328.4, + "completions/min_terminated_length": 328.4, + "entropy": 0.886419289931655, + "epoch": 0.96, + "frac_reward_zero_std": 0.325, + "grad_norm": 1.4609375, + "kl": 0.06785837954375892, + "learning_rate": 2.5625e-06, + "loss": 0.01289355307817459, + "num_tokens": 793664.0, + "reward": 0.94119873046875, + "reward_std": 0.4706839919090271, + "rewards/JointRewardFunction/mean": 0.94119873046875, + "rewards/JointRewardFunction/std": 0.47068399786949155, + "step": 120, + "step_time": 36.370545529201266 + }, + { + "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": 827.7, + "completions/max_terminated_length": 761.1, + "completions/mean_length": 534.6625, + "completions/mean_terminated_length": 516.6375061035156, + "completions/min_length": 321.3, + "completions/min_terminated_length": 321.3, + "entropy": 0.954836780950427, + "epoch": 1.04, + "frac_reward_zero_std": 0.275, + "grad_norm": 1.7265625, + "kl": 0.06585919852368534, + "learning_rate": 1.9375e-06, + "loss": 0.011952327191829681, + "num_tokens": 857055.0, + "reward": 0.96031494140625, + "reward_std": 0.44789557456970214, + "rewards/JointRewardFunction/mean": 0.96031494140625, + "rewards/JointRewardFunction/std": 0.44789558053016665, + "step": 130, + "step_time": 35.311843357800896 + }, + { + "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": 826.3, + "completions/max_terminated_length": 738.5, + "completions/mean_length": 522.15, + "completions/mean_terminated_length": 506.9553619384766, + "completions/min_length": 309.7, + "completions/min_terminated_length": 309.7, + "entropy": 0.8662954691797495, + "epoch": 1.12, + "frac_reward_zero_std": 0.25, + "grad_norm": 2.125, + "kl": 0.07243931000120937, + "learning_rate": 1.3125000000000001e-06, + "loss": 0.03181460499763489, + "num_tokens": 919475.0, + "reward": 0.86949462890625, + "reward_std": 0.5099265575408936, + "rewards/JointRewardFunction/mean": 0.86949462890625, + "rewards/JointRewardFunction/std": 0.5099265664815903, + "step": 140, + "step_time": 35.292809389702235 + }, + { + "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": 779.8, + "completions/max_terminated_length": 779.8, + "completions/mean_length": 524.8875, + "completions/mean_terminated_length": 524.8875, + "completions/min_length": 323.7, + "completions/min_terminated_length": 323.7, + "entropy": 0.9163745902478695, + "epoch": 1.2, + "frac_reward_zero_std": 0.325, + "grad_norm": 1.3359375, + "kl": 0.07890935295727104, + "learning_rate": 6.875000000000001e-07, + "loss": 0.0025221250951290132, + "num_tokens": 982042.0, + "reward": 1.10718994140625, + "reward_std": 0.4580986052751541, + "rewards/JointRewardFunction/mean": 1.10718994140625, + "rewards/JointRewardFunction/std": 0.45809862315654754, + "step": 150, + "step_time": 33.446236442600274 + }, + { + "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": 720.2, + "completions/max_terminated_length": 691.7, + "completions/mean_length": 492.3125, + "completions/mean_terminated_length": 485.6000030517578, + "completions/min_length": 343.0, + "completions/min_terminated_length": 343.0, + "entropy": 0.8506938068196177, + "epoch": 1.28, + "frac_reward_zero_std": 0.225, + "grad_norm": 1.71875, + "kl": 0.06682636064942926, + "learning_rate": 6.250000000000001e-08, + "loss": 0.03128217160701752, + "num_tokens": 1042329.0, + "reward": 0.963037109375, + "reward_std": 0.4769674390554428, + "rewards/JointRewardFunction/mean": 0.963037109375, + "rewards/JointRewardFunction/std": 0.47696744799613955, + "step": 160, + "step_time": 31.293741717099692 + } + ], + "logging_steps": 10, + "max_steps": 160, + "num_input_tokens_seen": 1042329, + "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..1057582 --- /dev/null +++ b/last-checkpoint/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:41733497170620207c2a02121b3cf2b43ec76ffc1db127fa6366ce5f1d1ccfde +size 7249 diff --git a/model.safetensors b/model.safetensors new file mode 100644 index 0000000..4665a23 --- /dev/null +++ b/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6032762bcfdf97d0ffef0d438d20af4d7f344890e0668ef6b790396c6cafa1f9 +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..1057582 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:41733497170620207c2a02121b3cf2b43ec76ffc1db127fa6366ce5f1d1ccfde +size 7249