Improve language tag

#1
by lbourdois - opened
Files changed (1) hide show
  1. README.md +80 -66
README.md CHANGED
@@ -1,67 +1,81 @@
1
- ---
2
- base_model: Qwen/Qwen2.5-0.5B
3
- datasets: trl-lib/math_shepherd
4
- library_name: transformers
5
- model_name: Qwen2.5-0.5B-Math-Shepherd-PRM-0.1
6
- tags:
7
- - generated_from_trainer
8
- - trl
9
- - stepwise-reward-trainer
10
- licence: license
11
- ---
12
-
13
- # Model Card for Qwen2.5-0.5B-Math-Shepherd-PRM-0.1
14
-
15
- This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) on the [trl-lib/math_shepherd](https://huggingface.co/datasets/trl-lib/math_shepherd) dataset.
16
- It has been trained using [TRL](https://github.com/huggingface/trl).
17
-
18
- ## Quick start
19
-
20
- ```python
21
- from transformers import pipeline
22
-
23
- 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?"
24
- generator = pipeline("text-generation", model="plaguss/Qwen2.5-0.5B-Math-Shepherd-PRM-0.1", device="cuda")
25
- output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
26
- print(output["generated_text"])
27
- ```
28
-
29
- ## Training procedure
30
-
31
- [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/plaguss/huggingface/runs/k1qvhdfz)
32
-
33
- This model was trained with Stepwise Reward.
34
-
35
- ### Framework versions
36
-
37
- - TRL: 0.13.0.dev0
38
- - Transformers: 4.46.0.dev0
39
- - Pytorch: 2.4.1
40
- - Datasets: 3.0.1
41
- - Tokenizers: 0.20.1
42
-
43
- ## Citations
44
-
45
- Cite Stepwise Reward as:
46
-
47
- ```bibtex
48
- @article{uesato2022solving,
49
- title = {Solving Math Word Problems With Process- and Outcome-Based Feedback},
50
- author = {Uesato, Jonathan and Kushman, Nate and Kumar, Ramana and Song, Francis and Siegel, Noah and Wang, Lisa and Creswell, Antonia and Irving, Geoffrey and Higgins, Irina},
51
- year = 2022,
52
- journal = {arXiv preprint arXiv:2211.14275}
53
- }
54
- ```
55
-
56
- Cite TRL as:
57
-
58
- ```bibtex
59
- @misc{vonwerra2022trl,
60
- title = {{TRL: Transformer Reinforcement Learning}},
61
- 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},
62
- year = 2020,
63
- journal = {GitHub repository},
64
- publisher = {GitHub},
65
- howpublished = {\url{https://github.com/huggingface/trl}}
66
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
67
  ```
 
1
+ ---
2
+ base_model: Qwen/Qwen2.5-0.5B
3
+ datasets: trl-lib/math_shepherd
4
+ library_name: transformers
5
+ model_name: Qwen2.5-0.5B-Math-Shepherd-PRM-0.1
6
+ tags:
7
+ - generated_from_trainer
8
+ - trl
9
+ - stepwise-reward-trainer
10
+ licence: license
11
+ language:
12
+ - zho
13
+ - eng
14
+ - fra
15
+ - spa
16
+ - por
17
+ - deu
18
+ - ita
19
+ - rus
20
+ - jpn
21
+ - kor
22
+ - vie
23
+ - tha
24
+ - ara
25
+ ---
26
+
27
+ # Model Card for Qwen2.5-0.5B-Math-Shepherd-PRM-0.1
28
+
29
+ This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) on the [trl-lib/math_shepherd](https://huggingface.co/datasets/trl-lib/math_shepherd) dataset.
30
+ It has been trained using [TRL](https://github.com/huggingface/trl).
31
+
32
+ ## Quick start
33
+
34
+ ```python
35
+ from transformers import pipeline
36
+
37
+ 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?"
38
+ generator = pipeline("text-generation", model="plaguss/Qwen2.5-0.5B-Math-Shepherd-PRM-0.1", device="cuda")
39
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
40
+ print(output["generated_text"])
41
+ ```
42
+
43
+ ## Training procedure
44
+
45
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/plaguss/huggingface/runs/k1qvhdfz)
46
+
47
+ This model was trained with Stepwise Reward.
48
+
49
+ ### Framework versions
50
+
51
+ - TRL: 0.13.0.dev0
52
+ - Transformers: 4.46.0.dev0
53
+ - Pytorch: 2.4.1
54
+ - Datasets: 3.0.1
55
+ - Tokenizers: 0.20.1
56
+
57
+ ## Citations
58
+
59
+ Cite Stepwise Reward as:
60
+
61
+ ```bibtex
62
+ @article{uesato2022solving,
63
+ title = {Solving Math Word Problems With Process- and Outcome-Based Feedback},
64
+ author = {Uesato, Jonathan and Kushman, Nate and Kumar, Ramana and Song, Francis and Siegel, Noah and Wang, Lisa and Creswell, Antonia and Irving, Geoffrey and Higgins, Irina},
65
+ year = 2022,
66
+ journal = {arXiv preprint arXiv:2211.14275}
67
+ }
68
+ ```
69
+
70
+ Cite TRL as:
71
+
72
+ ```bibtex
73
+ @misc{vonwerra2022trl,
74
+ title = {{TRL: Transformer Reinforcement Learning}},
75
+ 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},
76
+ year = 2020,
77
+ journal = {GitHub repository},
78
+ publisher = {GitHub},
79
+ howpublished = {\url{https://github.com/huggingface/trl}}
80
+ }
81
  ```