sengBJY/CSE151B_FinalProject

1---2base_model: Qwen/Qwen3-4B-Thinking-25073library_name: transformers4model_name: qwen3_4b_dual_pro6000_g6_len30725tags:6- generated_from_trainer7- grpo8- trl9licence: license10---11 12# Model Card for qwen3_4b_dual_pro6000_g6_len307213 14This model is a fine-tuned version of [Qwen/Qwen3-4B-Thinking-2507](https://huggingface.co/Qwen/Qwen3-4B-Thinking-2507).15It has been trained using [TRL](https://github.com/huggingface/trl).16 17## Quick start18 19```python20from transformers import pipeline21 22question = "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?"23generator = pipeline("text-generation", model="None", device="cuda")24output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]25print(output["generated_text"])26```27 28## Training procedure29 30 31 32 33 34This 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).35 36### Framework versions37 38- TRL: 1.5.139- Transformers: 5.9.040- Pytorch: 2.8.0+cu12841- Datasets: 4.8.542- Tokenizers: 0.22.243 44## Citations45 46Cite GRPO as:47 48```bibtex49@article{shao2024deepseekmath,50    title        = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},51    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},52    year         = 2024,53    eprint       = {arXiv:2402.03300},54}55```56 57Cite TRL as:58    59```bibtex60@software{vonwerra2020trl,61  title   = {{TRL: Transformers Reinforcement Learning}},62  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},63  license = {Apache-2.0},64  url     = {https://github.com/huggingface/trl},65  year    = {2020}66}67```