From 8a6d9332f538f54a480e9baf41619bc30033e794 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Tue, 16 Jun 2026 06:23:16 +0800 Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?= =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?= =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Model: harsha070/exp2-qwen-mbpp-s123-lambda-0p25 Source: Original Platform --- .gitattributes | 36 + README.md | 67 ++ chat_template.jinja | 54 ++ completions/completions_00010.parquet | 3 + completions/completions_00020.parquet | 3 + completions/completions_00030.parquet | 3 + completions/completions_00040.parquet | 3 + completions/completions_00050.parquet | 3 + completions/completions_00060.parquet | 3 + completions/completions_00070.parquet | 3 + completions/completions_00080.parquet | 3 + completions/completions_00090.parquet | 3 + completions/completions_00100.parquet | 3 + completions/completions_00110.parquet | 3 + completions/completions_00120.parquet | 3 + completions/completions_00130.parquet | 3 + completions/completions_00140.parquet | 3 + completions/completions_00150.parquet | 3 + completions/completions_00160.parquet | 3 + completions/completions_00170.parquet | 3 + completions/completions_00180.parquet | 3 + completions/completions_00190.parquet | 3 + completions/completions_00200.parquet | 3 + completions/completions_00210.parquet | 3 + completions/completions_00220.parquet | 3 + completions/completions_00230.parquet | 3 + completions/completions_00240.parquet | 3 + completions/completions_00250.parquet | 3 + completions/completions_00260.parquet | 3 + completions/completions_00270.parquet | 3 + completions/completions_00280.parquet | 3 + completions/completions_00290.parquet | 3 + completions/completions_00300.parquet | 3 + config.json | 69 ++ generation_config.json | 13 + last-checkpoint/chat_template.jinja | 54 ++ last-checkpoint/config.json | 69 ++ last-checkpoint/generation_config.json | 13 + last-checkpoint/model.safetensors | 3 + last-checkpoint/tokenizer.json | 3 + last-checkpoint/tokenizer_config.json | 30 + last-checkpoint/trainer_state.json | 874 +++++++++++++++++++++++++ last-checkpoint/training_args.bin | 3 + model.safetensors | 3 + tokenizer.json | 3 + tokenizer_config.json | 30 + training_args.bin | 3 + 47 files changed, 1417 insertions(+) create mode 100644 .gitattributes create mode 100644 README.md create mode 100644 chat_template.jinja create mode 100644 completions/completions_00010.parquet create mode 100644 completions/completions_00020.parquet create mode 100644 completions/completions_00030.parquet create mode 100644 completions/completions_00040.parquet create mode 100644 completions/completions_00050.parquet create mode 100644 completions/completions_00060.parquet create mode 100644 completions/completions_00070.parquet create mode 100644 completions/completions_00080.parquet create mode 100644 completions/completions_00090.parquet create mode 100644 completions/completions_00100.parquet create mode 100644 completions/completions_00110.parquet create mode 100644 completions/completions_00120.parquet create mode 100644 completions/completions_00130.parquet create mode 100644 completions/completions_00140.parquet create mode 100644 completions/completions_00150.parquet create mode 100644 completions/completions_00160.parquet create mode 100644 completions/completions_00170.parquet create mode 100644 completions/completions_00180.parquet create mode 100644 completions/completions_00190.parquet create mode 100644 completions/completions_00200.parquet create mode 100644 completions/completions_00210.parquet create mode 100644 completions/completions_00220.parquet create mode 100644 completions/completions_00230.parquet create mode 100644 completions/completions_00240.parquet create mode 100644 completions/completions_00250.parquet create mode 100644 completions/completions_00260.parquet create mode 100644 completions/completions_00270.parquet create mode 100644 completions/completions_00280.parquet create mode 100644 completions/completions_00290.parquet create mode 100644 completions/completions_00300.parquet create mode 100644 config.json create mode 100644 generation_config.json create mode 100644 last-checkpoint/chat_template.jinja create mode 100644 last-checkpoint/config.json create mode 100644 last-checkpoint/generation_config.json create mode 100644 last-checkpoint/model.safetensors create mode 100644 last-checkpoint/tokenizer.json create mode 100644 last-checkpoint/tokenizer_config.json create mode 100644 last-checkpoint/trainer_state.json create mode 100644 last-checkpoint/training_args.bin create mode 100644 model.safetensors create mode 100644 tokenizer.json create mode 100644 tokenizer_config.json create mode 100644 training_args.bin 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..634ecfd --- /dev/null +++ b/README.md @@ -0,0 +1,67 @@ +--- +base_model: harsha070/sft-warmup-qwen-v2 +library_name: transformers +model_name: exp2-qwen-mbpp-s123-lambda-0p25 +tags: +- generated_from_trainer +- grpo +- trl +licence: license +--- + +# Model Card for exp2-qwen-mbpp-s123-lambda-0p25 + +This model is a fine-tuned version of [harsha070/sft-warmup-qwen-v2](https://huggingface.co/harsha070/sft-warmup-qwen-v2). +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/exp2-qwen-mbpp-s123-lambda-0p25", 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/k77p5nhu) + + + +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..c8aa5f0 --- /dev/null +++ b/completions/completions_00010.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e1c6aaa319ac9d8f7eaf33130b2a7662d93edf6e66f290bff1bbb5435b0ff98d +size 17004 diff --git a/completions/completions_00020.parquet b/completions/completions_00020.parquet new file mode 100644 index 0000000..9a4162d --- /dev/null +++ b/completions/completions_00020.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5142fffbc8831e558aa0f2cf5e8cb4c6382bc50ef3c17b33fb6734986726a6f9 +size 18606 diff --git a/completions/completions_00030.parquet b/completions/completions_00030.parquet new file mode 100644 index 0000000..89f500a --- /dev/null +++ b/completions/completions_00030.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2da295599f9d18b821c652cf44fb2c7be2b7ab3e8a67d352eb2edd380c7386f2 +size 16556 diff --git a/completions/completions_00040.parquet b/completions/completions_00040.parquet new file mode 100644 index 0000000..2d59e3c --- /dev/null +++ b/completions/completions_00040.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b00eb1bea259f36fe97d95f999a61b857120ecf9c370abe5a6b896d0e82822b +size 17096 diff --git a/completions/completions_00050.parquet b/completions/completions_00050.parquet new file mode 100644 index 0000000..c3c52cf --- /dev/null +++ b/completions/completions_00050.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:48e82094f526a06b8757562987b007083889a6bc1543803fe3770664cfee77de +size 12502 diff --git a/completions/completions_00060.parquet b/completions/completions_00060.parquet new file mode 100644 index 0000000..54124f4 --- /dev/null +++ b/completions/completions_00060.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c70644550f56415752059bebd73109960ed5599c69a98a7afc8cc24abb88b008 +size 14342 diff --git a/completions/completions_00070.parquet b/completions/completions_00070.parquet new file mode 100644 index 0000000..03d7f56 --- /dev/null +++ b/completions/completions_00070.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2972c6503169172d6feb9294f75096dc5246b0fff360016e8a5cfa970c668d88 +size 14590 diff --git a/completions/completions_00080.parquet b/completions/completions_00080.parquet new file mode 100644 index 0000000..196c840 --- /dev/null +++ b/completions/completions_00080.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4ceee19eba47963320971167f3fa929b9c33707d3bf7760d94b9d749048daf4 +size 14069 diff --git a/completions/completions_00090.parquet b/completions/completions_00090.parquet new file mode 100644 index 0000000..6029cf7 --- /dev/null +++ b/completions/completions_00090.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b03fe06c8359aa38bde06009308f7b243b762b36610562ccce15501fa356aa6a +size 15463 diff --git a/completions/completions_00100.parquet b/completions/completions_00100.parquet new file mode 100644 index 0000000..bded361 --- /dev/null +++ b/completions/completions_00100.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b0fbb5e71b5107219e0bac496d2fc6ca19b8d5d5a9d10d6915ec71af213fc8cd +size 15223 diff --git a/completions/completions_00110.parquet b/completions/completions_00110.parquet new file mode 100644 index 0000000..61a67a6 --- /dev/null +++ b/completions/completions_00110.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8afcf092d0caf80bd8ea0364dc691c4c0a9adbc9bf81e149bdbbde2ebd6e68ee +size 16226 diff --git a/completions/completions_00120.parquet b/completions/completions_00120.parquet new file mode 100644 index 0000000..a8b42a8 --- /dev/null +++ b/completions/completions_00120.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8de62fc5409a9925a16f04733c77daf5d36085c6ef84aa734722fed98521a1b1 +size 14128 diff --git a/completions/completions_00130.parquet b/completions/completions_00130.parquet new file mode 100644 index 0000000..1b10d92 --- /dev/null +++ b/completions/completions_00130.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:417aa117f16eee02c9499af3f133008faeef9c46da9b3a5c2523519fd8664976 +size 15578 diff --git a/completions/completions_00140.parquet b/completions/completions_00140.parquet new file mode 100644 index 0000000..c418fce --- /dev/null +++ b/completions/completions_00140.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e35c6ffd51c0b4437027a0eee599ecfb050f244f6039340b36e69e41cbdab923 +size 13534 diff --git a/completions/completions_00150.parquet b/completions/completions_00150.parquet new file mode 100644 index 0000000..e6c0a17 --- /dev/null +++ b/completions/completions_00150.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ba99317ac5eabeeae7882eb63a46d9ea8ed1c1f261679dcaf6750bdd1343698c +size 13744 diff --git a/completions/completions_00160.parquet b/completions/completions_00160.parquet new file mode 100644 index 0000000..69681f6 --- /dev/null +++ b/completions/completions_00160.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bb7957c70833bf5502f356dc4c8eafa59389c7361a0964c8535f1f5bdfccd4e3 +size 13852 diff --git a/completions/completions_00170.parquet b/completions/completions_00170.parquet new file mode 100644 index 0000000..271064f --- /dev/null +++ b/completions/completions_00170.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7a0f8454849aac3e920d470d022b82934a8b5505c6ce9c47024bafcafd7d6d8 +size 14643 diff --git a/completions/completions_00180.parquet b/completions/completions_00180.parquet new file mode 100644 index 0000000..397eded --- /dev/null +++ b/completions/completions_00180.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9a00a9762b46418983748252269a725757020ee5f0a1fe8b491327e32d3686fd +size 13721 diff --git a/completions/completions_00190.parquet b/completions/completions_00190.parquet new file mode 100644 index 0000000..03ea1ef --- /dev/null +++ b/completions/completions_00190.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f43b0b18e87737fc54abe81db443dcb62b2cd5823f4e2bbf78fb308777dda19e +size 14581 diff --git a/completions/completions_00200.parquet b/completions/completions_00200.parquet new file mode 100644 index 0000000..ab70928 --- /dev/null +++ b/completions/completions_00200.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:33aee6584fd1dd233dab65f3e1b08222ae1ea1915ae764abecf197626cff9efd +size 14830 diff --git a/completions/completions_00210.parquet b/completions/completions_00210.parquet new file mode 100644 index 0000000..f3ab774 --- /dev/null +++ b/completions/completions_00210.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:22c18f3ff0be3b5ca7ee847582a71f9f9d5ef30332b400c7e89a21414c06e5a4 +size 15067 diff --git a/completions/completions_00220.parquet b/completions/completions_00220.parquet new file mode 100644 index 0000000..dfa2a61 --- /dev/null +++ b/completions/completions_00220.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca67634f5e8c475b667f924d4f723511671b4f031cf8277a0821d1858da58eab +size 14568 diff --git a/completions/completions_00230.parquet b/completions/completions_00230.parquet new file mode 100644 index 0000000..7cc707b --- /dev/null +++ b/completions/completions_00230.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:666114377758d48a982628e62e9c1b8498ab1f5ad0c1b6fcd428dbd6c75cb6f4 +size 14281 diff --git a/completions/completions_00240.parquet b/completions/completions_00240.parquet new file mode 100644 index 0000000..5d3bf75 --- /dev/null +++ b/completions/completions_00240.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e92656dccadc970e79427404b70cb5de28caa8d5369f043bdbfca494c245a6c2 +size 16779 diff --git a/completions/completions_00250.parquet b/completions/completions_00250.parquet new file mode 100644 index 0000000..23896f1 --- /dev/null +++ b/completions/completions_00250.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e443cf82fa45886ea5f58cbf99fb6dcf8fc1cf065a59761bd3849aa477783b7 +size 15259 diff --git a/completions/completions_00260.parquet b/completions/completions_00260.parquet new file mode 100644 index 0000000..7fc7d8e --- /dev/null +++ b/completions/completions_00260.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:465eb27319771a1c25029e37e5da50f1c20f729c023ae74143c5edca40fdb041 +size 14775 diff --git a/completions/completions_00270.parquet b/completions/completions_00270.parquet new file mode 100644 index 0000000..7e368a9 --- /dev/null +++ b/completions/completions_00270.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:90deab49a9c7cfb4a2d2cab6e701b3f181ed3e498dc60d7e84be2d2f152e8a19 +size 15901 diff --git a/completions/completions_00280.parquet b/completions/completions_00280.parquet new file mode 100644 index 0000000..e31163d --- /dev/null +++ b/completions/completions_00280.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d687b1883b804b7b34aaacab2037099dafc4e396586eb82a7e9dea0ffbdb93a +size 15424 diff --git a/completions/completions_00290.parquet b/completions/completions_00290.parquet new file mode 100644 index 0000000..cccc56c --- /dev/null +++ b/completions/completions_00290.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:95873ed55fa379bbdeff777af1fcc504747a02c7aa478afc4d2171ea6fb5d328 +size 15530 diff --git a/completions/completions_00300.parquet b/completions/completions_00300.parquet new file mode 100644 index 0000000..1456a8e --- /dev/null +++ b/completions/completions_00300.parquet @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8cb04ab8de2404fdeb549c668b7f635f64d703cea8f69fb751d769223621ecea +size 17020 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..811e776 --- /dev/null +++ b/last-checkpoint/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f07245ca7d96f54917cf2b2f0f9ff1fe8acd2a0f11ba9aeb9cead5cc798cc928 +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..f19da42 --- /dev/null +++ b/last-checkpoint/trainer_state.json @@ -0,0 +1,874 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.0, + "eval_steps": 500, + "global_step": 300, + "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": 0.525, + "completions/max_length": 512.0, + "completions/max_terminated_length": 378.9, + "completions/mean_length": 457.9125, + "completions/mean_terminated_length": 322.60096435546876, + "completions/min_length": 351.5, + "completions/min_terminated_length": 249.1, + "entropy": 0.3948360087350011, + "epoch": 0.06666666666666667, + "frac_reward_zero_std": 0.05, + "grad_norm": 2.03125, + "kl": 0.04399079610357148, + "learning_rate": 9.7e-06, + "loss": -0.004937342554330826, + "num_tokens": 49697.0, + "reward": 0.65526123046875, + "reward_std": 0.4537597060203552, + "rewards/JointRewardFunction/mean": 0.65526123046875, + "rewards/JointRewardFunction/std": 0.45375972241163254, + "step": 10, + "step_time": 22.810993660400346 + }, + { + "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": 0.5, + "completions/max_length": 508.7, + "completions/max_terminated_length": 417.6, + "completions/mean_length": 443.675, + "completions/mean_terminated_length": 342.31012573242185, + "completions/min_length": 326.9, + "completions/min_terminated_length": 275.7, + "entropy": 0.31964702922850846, + "epoch": 0.13333333333333333, + "frac_reward_zero_std": 0.25, + "grad_norm": 0.1982421875, + "kl": 0.09551988991443068, + "learning_rate": 9.366666666666668e-06, + "loss": 0.00889739990234375, + "num_tokens": 97803.0, + "reward": 0.7364013671875, + "reward_std": 0.4511200308799744, + "rewards/JointRewardFunction/mean": 0.7364013671875, + "rewards/JointRewardFunction/std": 0.4511200487613678, + "step": 20, + "step_time": 23.522369655099464 + }, + { + "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": 0.675, + "completions/max_length": 512.0, + "completions/max_terminated_length": 379.6, + "completions/mean_length": 477.4875, + "completions/mean_terminated_length": 350.68857421875, + "completions/min_length": 371.0, + "completions/min_terminated_length": 319.8, + "entropy": 0.40151252076029775, + "epoch": 0.2, + "frac_reward_zero_std": 0.35, + "grad_norm": 0.1630859375, + "kl": 0.11964645721018315, + "learning_rate": 9.033333333333334e-06, + "loss": -0.0028077010065317154, + "num_tokens": 148394.0, + "reward": 0.9680908203125, + "reward_std": 0.40886188447475436, + "rewards/JointRewardFunction/mean": 0.9680908203125, + "rewards/JointRewardFunction/std": 0.40886189937591555, + "step": 30, + "step_time": 23.04281468010122 + }, + { + "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": 0.4, + "completions/max_length": 512.0, + "completions/max_terminated_length": 469.3, + "completions/mean_length": 438.0875, + "completions/mean_terminated_length": 393.87953491210936, + "completions/min_length": 335.8, + "completions/min_terminated_length": 335.8, + "entropy": 0.35483993515372275, + "epoch": 0.26666666666666666, + "frac_reward_zero_std": 0.25, + "grad_norm": 2.34375, + "kl": 0.0885015036445111, + "learning_rate": 8.700000000000001e-06, + "loss": 0.03651041984558105, + "num_tokens": 197537.0, + "reward": 0.9726318359375, + "reward_std": 0.38605129569768903, + "rewards/JointRewardFunction/mean": 0.9726318359375, + "rewards/JointRewardFunction/std": 0.3860512986779213, + "step": 40, + "step_time": 23.371826623300876 + }, + { + "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": 416.7, + "completions/max_terminated_length": 387.6, + "completions/mean_length": 307.5, + "completions/mean_terminated_length": 296.2386932373047, + "completions/min_length": 221.0, + "completions/min_terminated_length": 221.0, + "entropy": 0.3532023083418608, + "epoch": 0.3333333333333333, + "frac_reward_zero_std": 0.2, + "grad_norm": 2.484375, + "kl": 0.1458138041663915, + "learning_rate": 8.366666666666667e-06, + "loss": 0.058735775947570804, + "num_tokens": 235441.0, + "reward": 0.911224365234375, + "reward_std": 0.44634594917297366, + "rewards/JointRewardFunction/mean": 0.911224365234375, + "rewards/JointRewardFunction/std": 0.4463459521532059, + "step": 50, + "step_time": 19.590083589001733 + }, + { + "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": 439.0, + "completions/max_terminated_length": 411.3, + "completions/mean_length": 314.675, + "completions/mean_terminated_length": 302.16786193847656, + "completions/min_length": 213.5, + "completions/min_terminated_length": 213.5, + "entropy": 0.33970522740855813, + "epoch": 0.4, + "frac_reward_zero_std": 0.25, + "grad_norm": 2.21875, + "kl": 0.13768328949809075, + "learning_rate": 8.033333333333335e-06, + "loss": 0.024235518276691438, + "num_tokens": 273659.0, + "reward": 1.012158203125, + "reward_std": 0.34636138631030916, + "rewards/JointRewardFunction/mean": 1.012158203125, + "rewards/JointRewardFunction/std": 0.3463613982312381, + "step": 60, + "step_time": 20.71416906509876 + }, + { + "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": 405.4, + "completions/max_terminated_length": 397.6, + "completions/mean_length": 306.85, + "completions/mean_terminated_length": 298.9619079589844, + "completions/min_length": 215.4, + "completions/min_terminated_length": 215.4, + "entropy": 0.36216176860034466, + "epoch": 0.4666666666666667, + "frac_reward_zero_std": 0.5, + "grad_norm": 2.375, + "kl": 0.11585550024174154, + "learning_rate": 7.7e-06, + "loss": -0.0011418772861361503, + "num_tokens": 311055.0, + "reward": 1.108551025390625, + "reward_std": 0.31149580857600084, + "rewards/JointRewardFunction/mean": 1.108551025390625, + "rewards/JointRewardFunction/std": 0.311495810066117, + "step": 70, + "step_time": 19.105042513498848 + }, + { + "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": 425.2, + "completions/max_terminated_length": 399.6, + "completions/mean_length": 303.65, + "completions/mean_terminated_length": 296.1714324951172, + "completions/min_length": 221.7, + "completions/min_terminated_length": 221.7, + "entropy": 0.458731491304934, + "epoch": 0.5333333333333333, + "frac_reward_zero_std": 0.25, + "grad_norm": 1.8515625, + "kl": 0.12144780233502388, + "learning_rate": 7.3666666666666676e-06, + "loss": 0.04609571397304535, + "num_tokens": 347759.0, + "reward": 1.0232421875, + "reward_std": 0.39059739662334325, + "rewards/JointRewardFunction/mean": 1.0232421875, + "rewards/JointRewardFunction/std": 0.39059740854427216, + "step": 80, + "step_time": 20.152326823800102 + }, + { + "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": 458.2, + "completions/max_terminated_length": 434.3, + "completions/mean_length": 325.7, + "completions/mean_terminated_length": 320.2678619384766, + "completions/min_length": 231.3, + "completions/min_terminated_length": 231.3, + "entropy": 0.414226458594203, + "epoch": 0.6, + "frac_reward_zero_std": 0.45, + "grad_norm": 2.71875, + "kl": 0.11777536268346012, + "learning_rate": 7.033333333333334e-06, + "loss": 0.017658862471580505, + "num_tokens": 386695.0, + "reward": 1.0622802734375, + "reward_std": 0.37295620441436766, + "rewards/JointRewardFunction/mean": 1.0622802734375, + "rewards/JointRewardFunction/std": 0.3729562103748322, + "step": 90, + "step_time": 21.401791930101172 + }, + { + "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": 419.9, + "completions/max_terminated_length": 405.6, + "completions/mean_length": 305.0625, + "completions/mean_terminated_length": 297.75535888671874, + "completions/min_length": 201.6, + "completions/min_terminated_length": 201.6, + "entropy": 0.4645195137709379, + "epoch": 0.6666666666666666, + "frac_reward_zero_std": 0.45, + "grad_norm": 3.21875, + "kl": 0.11028626803308725, + "learning_rate": 6.700000000000001e-06, + "loss": 0.019249849021434784, + "num_tokens": 423284.0, + "reward": 1.095147705078125, + "reward_std": 0.3025559388101101, + "rewards/JointRewardFunction/mean": 1.095147705078125, + "rewards/JointRewardFunction/std": 0.302555950731039, + "step": 100, + "step_time": 19.82219038610201 + }, + { + "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": 437.4, + "completions/max_terminated_length": 425.9, + "completions/mean_length": 325.925, + "completions/mean_terminated_length": 319.71428833007815, + "completions/min_length": 227.9, + "completions/min_terminated_length": 227.9, + "entropy": 0.43129814630374314, + "epoch": 0.7333333333333333, + "frac_reward_zero_std": 0.5, + "grad_norm": 1.40625, + "kl": 0.1031634843442589, + "learning_rate": 6.366666666666668e-06, + "loss": 0.026633352041244507, + "num_tokens": 462986.0, + "reward": 1.043157958984375, + "reward_std": 0.33559322796645574, + "rewards/JointRewardFunction/mean": 1.043157958984375, + "rewards/JointRewardFunction/std": 0.3355932369071525, + "step": 110, + "step_time": 20.435491250798805 + }, + { + "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": 444.4, + "completions/max_terminated_length": 440.1, + "completions/mean_length": 337.825, + "completions/mean_terminated_length": 330.73964538574216, + "completions/min_length": 233.7, + "completions/min_terminated_length": 233.7, + "entropy": 0.43917944487184285, + "epoch": 0.8, + "frac_reward_zero_std": 0.45, + "grad_norm": 4.03125, + "kl": 0.1373384237056598, + "learning_rate": 6.033333333333335e-06, + "loss": 0.051127517223358156, + "num_tokens": 504032.0, + "reward": 1.03193359375, + "reward_std": 0.326281476020813, + "rewards/JointRewardFunction/mean": 1.03193359375, + "rewards/JointRewardFunction/std": 0.3262814819812775, + "step": 120, + "step_time": 20.735252708198097 + }, + { + "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": 412.0, + "completions/max_terminated_length": 392.5, + "completions/mean_length": 312.775, + "completions/mean_terminated_length": 300.4857147216797, + "completions/min_length": 222.2, + "completions/min_terminated_length": 222.2, + "entropy": 0.4507128844037652, + "epoch": 0.8666666666666667, + "frac_reward_zero_std": 0.5, + "grad_norm": 2.09375, + "kl": 0.11878000139258801, + "learning_rate": 5.7e-06, + "loss": 0.011486038565635681, + "num_tokens": 544434.0, + "reward": 1.053509521484375, + "reward_std": 0.31438804864883424, + "rewards/JointRewardFunction/mean": 1.053509521484375, + "rewards/JointRewardFunction/std": 0.31438804864883424, + "step": 130, + "step_time": 19.64918669259896 + }, + { + "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": 439.1, + "completions/max_terminated_length": 435.6, + "completions/mean_length": 333.05, + "completions/mean_terminated_length": 327.24464416503906, + "completions/min_length": 234.3, + "completions/min_terminated_length": 234.3, + "entropy": 0.40997486785054205, + "epoch": 0.9333333333333333, + "frac_reward_zero_std": 0.4, + "grad_norm": 2.84375, + "kl": 0.10414444564376027, + "learning_rate": 5.366666666666666e-06, + "loss": 0.011959856748580933, + "num_tokens": 584866.0, + "reward": 1.083343505859375, + "reward_std": 0.3227477639913559, + "rewards/JointRewardFunction/mean": 1.083343505859375, + "rewards/JointRewardFunction/std": 0.3227477639913559, + "step": 140, + "step_time": 20.672387204799815 + }, + { + "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": 0.0875, + "completions/max_length": 434.4, + "completions/max_terminated_length": 412.6, + "completions/mean_length": 322.45, + "completions/mean_terminated_length": 308.4941680908203, + "completions/min_length": 227.0, + "completions/min_terminated_length": 227.0, + "entropy": 0.4216183707118034, + "epoch": 1.0, + "frac_reward_zero_std": 0.35, + "grad_norm": 3.234375, + "kl": 0.10422139009460807, + "learning_rate": 5.033333333333333e-06, + "loss": 0.009627737849950791, + "num_tokens": 623390.0, + "reward": 1.072760009765625, + "reward_std": 0.298717375023989, + "rewards/JointRewardFunction/mean": 1.072760009765625, + "rewards/JointRewardFunction/std": 0.29871737800422127, + "step": 150, + "step_time": 20.43400644650028 + }, + { + "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": 477.3, + "completions/max_terminated_length": 441.4, + "completions/mean_length": 332.3375, + "completions/mean_terminated_length": 314.32142944335936, + "completions/min_length": 212.2, + "completions/min_terminated_length": 212.2, + "entropy": 0.4680627126246691, + "epoch": 1.0666666666666667, + "frac_reward_zero_std": 0.45, + "grad_norm": 0.054443359375, + "kl": 0.09179632470477372, + "learning_rate": 4.7e-06, + "loss": 0.01482328325510025, + "num_tokens": 663349.0, + "reward": 1.0511962890625, + "reward_std": 0.33187133751343934, + "rewards/JointRewardFunction/mean": 1.0511962890625, + "rewards/JointRewardFunction/std": 0.33187134347390385, + "step": 160, + "step_time": 22.216746885597967 + }, + { + "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": 431.7, + "completions/max_terminated_length": 406.5, + "completions/mean_length": 307.6625, + "completions/mean_terminated_length": 298.221435546875, + "completions/min_length": 207.6, + "completions/min_terminated_length": 207.6, + "entropy": 0.4726201189681888, + "epoch": 1.1333333333333333, + "frac_reward_zero_std": 0.6, + "grad_norm": 2.078125, + "kl": 0.10375995303038507, + "learning_rate": 4.366666666666667e-06, + "loss": 0.008281752467155457, + "num_tokens": 699830.0, + "reward": 1.084967041015625, + "reward_std": 0.2650434322655201, + "rewards/JointRewardFunction/mean": 1.084967041015625, + "rewards/JointRewardFunction/std": 0.26504343524575236, + "step": 170, + "step_time": 20.110693615402123 + }, + { + "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": 379.8, + "completions/max_terminated_length": 379.8, + "completions/mean_length": 297.7875, + "completions/mean_terminated_length": 297.7875, + "completions/min_length": 209.8, + "completions/min_terminated_length": 209.8, + "entropy": 0.4800571627914906, + "epoch": 1.2, + "frac_reward_zero_std": 0.55, + "grad_norm": 2.859375, + "kl": 0.10157871665433049, + "learning_rate": 4.033333333333333e-06, + "loss": 0.015658122301101685, + "num_tokens": 735313.0, + "reward": 1.125604248046875, + "reward_std": 0.25403361991047857, + "rewards/JointRewardFunction/mean": 1.125604248046875, + "rewards/JointRewardFunction/std": 0.2540336221456528, + "step": 180, + "step_time": 18.15690856939873 + }, + { + "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": 442.3, + "completions/max_terminated_length": 417.9, + "completions/mean_length": 322.4875, + "completions/mean_terminated_length": 312.0000061035156, + "completions/min_length": 221.9, + "completions/min_terminated_length": 221.9, + "entropy": 0.4642798786982894, + "epoch": 1.2666666666666666, + "frac_reward_zero_std": 0.55, + "grad_norm": 2.1875, + "kl": 0.0930135941482149, + "learning_rate": 3.7e-06, + "loss": 0.022286814451217652, + "num_tokens": 773840.0, + "reward": 1.0874267578125, + "reward_std": 0.2804231091577094, + "rewards/JointRewardFunction/mean": 1.0874267578125, + "rewards/JointRewardFunction/std": 0.28042311511817386, + "step": 190, + "step_time": 20.867724073099453 + }, + { + "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": 448.1, + "completions/max_terminated_length": 431.5, + "completions/mean_length": 345.2375, + "completions/mean_terminated_length": 335.31964416503905, + "completions/min_length": 235.1, + "completions/min_terminated_length": 235.1, + "entropy": 0.4752088913694024, + "epoch": 1.3333333333333333, + "frac_reward_zero_std": 0.5, + "grad_norm": 2.140625, + "kl": 0.08606207200791686, + "learning_rate": 3.366666666666667e-06, + "loss": 0.019437333941459654, + "num_tokens": 814411.0, + "reward": 1.0824462890625, + "reward_std": 0.3011944264173508, + "rewards/JointRewardFunction/mean": 1.0824462890625, + "rewards/JointRewardFunction/std": 0.3011944264173508, + "step": 200, + "step_time": 21.037474609902347 + }, + { + "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": 0.1125, + "completions/max_length": 447.2, + "completions/max_terminated_length": 440.3, + "completions/mean_length": 341.525, + "completions/mean_terminated_length": 324.12596130371094, + "completions/min_length": 227.0, + "completions/min_terminated_length": 227.0, + "entropy": 0.45753496792167425, + "epoch": 1.4, + "frac_reward_zero_std": 0.45, + "grad_norm": 1.7265625, + "kl": 0.08462402573786676, + "learning_rate": 3.0333333333333337e-06, + "loss": 0.03857302367687225, + "num_tokens": 854993.0, + "reward": 1.0694580078125, + "reward_std": 0.32212619576603174, + "rewards/JointRewardFunction/mean": 1.0694580078125, + "rewards/JointRewardFunction/std": 0.32212621364742516, + "step": 210, + "step_time": 20.958584549299484 + }, + { + "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": 0.0875, + "completions/max_length": 439.8, + "completions/max_terminated_length": 409.0, + "completions/mean_length": 325.825, + "completions/mean_terminated_length": 312.76429138183596, + "completions/min_length": 230.1, + "completions/min_terminated_length": 230.1, + "entropy": 0.43106496389955284, + "epoch": 1.4666666666666668, + "frac_reward_zero_std": 0.7, + "grad_norm": 1.84375, + "kl": 0.08686103990767151, + "learning_rate": 2.7000000000000004e-06, + "loss": 0.019956868886947633, + "num_tokens": 894887.0, + "reward": 1.037451171875, + "reward_std": 0.27579919978743417, + "rewards/JointRewardFunction/mean": 1.037451171875, + "rewards/JointRewardFunction/std": 0.2757992057478987, + "step": 220, + "step_time": 20.81888391800021 + }, + { + "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": 370.2, + "completions/max_terminated_length": 357.7, + "completions/mean_length": 287.3625, + "completions/mean_terminated_length": 284.64464416503904, + "completions/min_length": 212.9, + "completions/min_terminated_length": 212.9, + "entropy": 0.5203873571008444, + "epoch": 1.5333333333333332, + "frac_reward_zero_std": 0.7, + "grad_norm": 2.4375, + "kl": 0.09830583925358952, + "learning_rate": 2.3666666666666667e-06, + "loss": 0.013263462483882904, + "num_tokens": 929688.0, + "reward": 1.1615234375, + "reward_std": 0.1927501901984215, + "rewards/JointRewardFunction/mean": 1.1615234375, + "rewards/JointRewardFunction/std": 0.19275019615888594, + "step": 230, + "step_time": 17.954591237399292 + }, + { + "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": 0.1375, + "completions/max_length": 455.5, + "completions/max_terminated_length": 442.9, + "completions/mean_length": 355.65, + "completions/mean_terminated_length": 341.68095703125, + "completions/min_length": 251.1, + "completions/min_terminated_length": 251.1, + "entropy": 0.42119636284187434, + "epoch": 1.6, + "frac_reward_zero_std": 0.35, + "grad_norm": 1.78125, + "kl": 0.08470410050358623, + "learning_rate": 2.0333333333333335e-06, + "loss": 0.01756674200296402, + "num_tokens": 972840.0, + "reward": 0.989080810546875, + "reward_std": 0.37197236269712447, + "rewards/JointRewardFunction/mean": 0.989080810546875, + "rewards/JointRewardFunction/std": 0.37197236269712447, + "step": 240, + "step_time": 21.474622961001295 + }, + { + "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": 0.0875, + "completions/max_length": 462.3, + "completions/max_terminated_length": 428.0, + "completions/mean_length": 342.35, + "completions/mean_terminated_length": 324.6696472167969, + "completions/min_length": 234.8, + "completions/min_terminated_length": 234.8, + "entropy": 0.4915314789861441, + "epoch": 1.6666666666666665, + "frac_reward_zero_std": 0.3, + "grad_norm": 1.2109375, + "kl": 0.08292658934369683, + "learning_rate": 1.7000000000000002e-06, + "loss": 0.021336786448955536, + "num_tokens": 1013548.0, + "reward": 1.00546875, + "reward_std": 0.3482314422726631, + "rewards/JointRewardFunction/mean": 1.00546875, + "rewards/JointRewardFunction/std": 0.3482314541935921, + "step": 250, + "step_time": 21.594003204298495 + }, + { + "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": 0.1125, + "completions/max_length": 458.6, + "completions/max_terminated_length": 413.7, + "completions/mean_length": 339.1875, + "completions/mean_terminated_length": 320.0369110107422, + "completions/min_length": 236.0, + "completions/min_terminated_length": 236.0, + "entropy": 0.4891188982874155, + "epoch": 1.7333333333333334, + "frac_reward_zero_std": 0.45, + "grad_norm": 1.515625, + "kl": 0.08876022514887154, + "learning_rate": 1.3666666666666668e-06, + "loss": 0.04018712341785431, + "num_tokens": 1053691.0, + "reward": 1.0763916015625, + "reward_std": 0.33348723202943803, + "rewards/JointRewardFunction/mean": 1.0763916015625, + "rewards/JointRewardFunction/std": 0.33348723202943803, + "step": 260, + "step_time": 21.527129449401038 + }, + { + "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": 424.1, + "completions/max_terminated_length": 398.5, + "completions/mean_length": 326.7125, + "completions/mean_terminated_length": 318.79107971191405, + "completions/min_length": 235.0, + "completions/min_terminated_length": 235.0, + "entropy": 0.5027944987639785, + "epoch": 1.8, + "frac_reward_zero_std": 0.4, + "grad_norm": 1.4140625, + "kl": 0.09775767312385142, + "learning_rate": 1.0333333333333333e-06, + "loss": 0.022160670161247252, + "num_tokens": 1093636.0, + "reward": 1.073553466796875, + "reward_std": 0.2811176300048828, + "rewards/JointRewardFunction/mean": 1.073553466796875, + "rewards/JointRewardFunction/std": 0.2811176344752312, + "step": 270, + "step_time": 19.999046050798643 + }, + { + "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": 418.1, + "completions/max_terminated_length": 399.8, + "completions/mean_length": 323.325, + "completions/mean_terminated_length": 314.7875, + "completions/min_length": 233.0, + "completions/min_terminated_length": 233.0, + "entropy": 0.46924052368849517, + "epoch": 1.8666666666666667, + "frac_reward_zero_std": 0.55, + "grad_norm": 1.890625, + "kl": 0.09717398346401751, + "learning_rate": 7.000000000000001e-07, + "loss": 0.011738638579845428, + "num_tokens": 1132610.0, + "reward": 1.1536865234375, + "reward_std": 0.23539954721927642, + "rewards/JointRewardFunction/mean": 1.1536865234375, + "rewards/JointRewardFunction/std": 0.2353995516896248, + "step": 280, + "step_time": 20.17876625719837 + }, + { + "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": 0.1375, + "completions/max_length": 465.2, + "completions/max_terminated_length": 430.9, + "completions/mean_length": 340.55, + "completions/mean_terminated_length": 316.59702758789064, + "completions/min_length": 220.1, + "completions/min_terminated_length": 220.1, + "entropy": 0.48896827585995195, + "epoch": 1.9333333333333333, + "frac_reward_zero_std": 0.5, + "grad_norm": 0.08544921875, + "kl": 0.09501179384533316, + "learning_rate": 3.666666666666667e-07, + "loss": 0.011296895891427993, + "num_tokens": 1175610.0, + "reward": 1.069940185546875, + "reward_std": 0.29926753640174864, + "rewards/JointRewardFunction/mean": 1.069940185546875, + "rewards/JointRewardFunction/std": 0.2992675393819809, + "step": 290, + "step_time": 21.805314568298854 + }, + { + "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": 427.1, + "completions/max_terminated_length": 417.8, + "completions/mean_length": 321.225, + "completions/mean_terminated_length": 312.3041687011719, + "completions/min_length": 233.6, + "completions/min_terminated_length": 233.6, + "entropy": 0.49822530038654805, + "epoch": 2.0, + "frac_reward_zero_std": 0.5, + "grad_norm": 2.28125, + "kl": 0.09350393395870923, + "learning_rate": 3.333333333333334e-08, + "loss": 0.0006398491561412812, + "num_tokens": 1214508.0, + "reward": 1.014300537109375, + "reward_std": 0.3215805938467383, + "rewards/JointRewardFunction/mean": 1.014300537109375, + "rewards/JointRewardFunction/std": 0.3215805938467383, + "step": 300, + "step_time": 20.291091689400854 + } + ], + "logging_steps": 10, + "max_steps": 300, + "num_input_tokens_seen": 1214508, + "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..4982e87 --- /dev/null +++ b/last-checkpoint/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:554414b01fccbb7ba29d04032d89e3ce7c94de46706bef19ef145185ccfd6623 +size 7249 diff --git a/model.safetensors b/model.safetensors new file mode 100644 index 0000000..811e776 --- /dev/null +++ b/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f07245ca7d96f54917cf2b2f0f9ff1fe8acd2a0f11ba9aeb9cead5cc798cc928 +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..4982e87 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:554414b01fccbb7ba29d04032d89e3ce7c94de46706bef19ef145185ccfd6623 +size 7249