Views
No views yet
1from transformers import AutoTokenizer, AutoModelForCausalLM
2
3model_id = "Hacoo1234/Qwen2-0.5B-GRPO-test2"
4token = "paste code ere" # your HF access token
5
6trained_model = AutoModelForCausalLM.from_pretrained(
7 model_id,
8 use_auth_token=token,
9 torch_dtype="auto",
10 device_map="auto"
11)
12
13trained_tokenizer = AutoTokenizer.from_pretrained(
14 model_id,
15 use_auth_token=token
16)
17
18
19## Training procedure
20
21
22
23
24This 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).
25
26### Framework versions
27
28- TRL: 0.19.0
29- Transformers: 4.53.0
30- Pytorch: 2.6.0+cu124
31- Datasets: 3.6.0
32- Tokenizers: 0.21.2
33
34## Citations
35
36Cite GRPO as:
37
38```bibtex
39@article{zhihong2024deepseekmath,
40 title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
41 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},
42 year = 2024,
43 eprint = {arXiv:2402.03300},
44}
451@misc{vonwerra2022trl,
2 title = {{TRL: Transformer Reinforcement Learning}},
3 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 Gallou{\'e}dec},
4 year = 2020,
5 journal = {GitHub repository},
6 publisher = {GitHub},
7 howpublished = {\url{https://github.com/huggingface/trl}}
8}