Model Card for metamath-qwen3.5-0.6b-rl-gamma1

This model is a fine-tuned version of None on the meta-math/MetaMathQA dataset. It has been trained using TRL.

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="None", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Training procedure

This model was trained with MiniLLM, a method introduced in MiniLLM: Knowledge Distillation of Large Language Models.

Framework versions

• TRL: 1.0.0.dev0
• Transformers: 5.5.4+computecanada
• Pytorch: 2.11.0+computecanada
• Datasets: 4.8.4+computecanada
• Tokenizers: 0.22.2+computecanada

Citations

Cite MiniLLM as:

@inproceedings{
    gu2024minillm,
    title={{MiniLLM: Knowledge Distillation of Large Language Models}},
    author={Yuxian Gu and Li Dong and Furu Wei and Minlie Huang},
    booktitle={The Twelfth International Conference on Learning Representations},
    year={2024},
    url={https://openreview.net/forum?id=5h0qf7IBZZ}
}

Cite TRL as:

@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}
}