ForestHYS commited on
Commit
3842227
·
verified ·
1 Parent(s): 2bf2fe1

Upload grpo-lora/README.md

Browse files
Files changed (1) hide show
  1. grpo-lora/README.md +71 -0
grpo-lora/README.md ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ model_name: grpo-lora
4
+ tags:
5
+ - base_model:adapter:Qwen/Qwen3.5-9B
6
+ - grpo
7
+ - lora
8
+ - transformers
9
+ - trl
10
+ licence: license
11
+ pipeline_tag: text-generation
12
+ base_model: Qwen/Qwen3.5-9B
13
+ ---
14
+
15
+ # Model Card for grpo-lora
16
+
17
+ This model is a fine-tuned version of [None](https://huggingface.co/None).
18
+ It has been trained using [TRL](https://github.com/huggingface/trl).
19
+
20
+ ## Quick start
21
+
22
+ ```python
23
+ from transformers import pipeline
24
+
25
+ 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?"
26
+ generator = pipeline("text-generation", model="None", device="cuda")
27
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
28
+ print(output["generated_text"])
29
+ ```
30
+
31
+ ## Training procedure
32
+
33
+
34
+
35
+
36
+ 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).
37
+
38
+ ### Framework versions
39
+
40
+ - PEFT 0.19.1
41
+ - TRL: 0.28.0
42
+ - Transformers: 5.8.0
43
+ - Pytorch: 2.5.1+cu124
44
+ - Datasets: 4.8.5
45
+ - Tokenizers: 0.22.2
46
+
47
+ ## Citations
48
+
49
+ Cite GRPO as:
50
+
51
+ ```bibtex
52
+ @article{shao2024deepseekmath,
53
+ title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
54
+ 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},
55
+ year = 2024,
56
+ eprint = {arXiv:2402.03300},
57
+ }
58
+
59
+ ```
60
+
61
+ Cite TRL as:
62
+
63
+ ```bibtex
64
+ @software{vonwerra2020trl,
65
+ title = {{TRL: Transformers Reinforcement Learning}},
66
+ 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},
67
+ license = {Apache-2.0},
68
+ url = {https://github.com/huggingface/trl},
69
+ year = {2020}
70
+ }
71
+ ```