Openrs3 GRPO Ja

🚀 OpenRS3-GRPO-ja

OpenRS3-GRPO-ja is a fine - tuned language model, leveraging the power of GRPO to enhance mathematical reasoning capabilities on specific datasets.

🚀 Quick Start

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="stardust-eques/OpenRS-GRPO-ja", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

✨ Features

• This model is a fine - tuned version of SakanaAI/TinySwallow-1.5B-Instruct on the kunishou/OpenMathInstruct-1-1.8m-ja dataset.
• It has been trained using TRL, with 300 training steps.
• Trained with GRPO, a method from DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models.

🔧 Technical Details

Framework versions

📚 Documentation

Training procedure

This model was trained with GRPO, a method introduced in DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models.

📄 License

The model is released under the license license.

📄 Citations

Cite GRPO as:

@article{zhihong2024deepseekmath,
    title        = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
    author       = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
    year         = 2024,
    eprint       = {arXiv:2402.03300},
}

Cite TRL as:

@misc{vonwerra2022trl,
	title        = {{TRL: Transformer Reinforcement Learning}},
	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
	year         = 2020,
	journal      = {GitHub repository},
	publisher    = {GitHub},
	howpublished = {\url{https://github.com/huggingface/trl}}
}