Model save
Browse files- README.md +63 -0
- generation_config.json +4 -0
README.md
ADDED
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: transformers
|
3 |
+
model_name: OLMo-1B-hf-PPO-constitution-1
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
licence: license
|
7 |
+
---
|
8 |
+
|
9 |
+
# Model Card for OLMo-1B-hf-PPO-constitution-1
|
10 |
+
|
11 |
+
This model is a fine-tuned version of [None](https://huggingface.co/None).
|
12 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
13 |
+
|
14 |
+
## Quick start
|
15 |
+
|
16 |
+
```python
|
17 |
+
from transformers import pipeline
|
18 |
+
|
19 |
+
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?"
|
20 |
+
generator = pipeline("text-generation", model="Shahradmz/OLMo-1B-hf-PPO-constitution-1", device="cuda")
|
21 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
22 |
+
print(output["generated_text"])
|
23 |
+
```
|
24 |
+
|
25 |
+
## Training procedure
|
26 |
+
|
27 |
+
[<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/shahrad_m/huggingface/runs/1kr21hea)
|
28 |
+
|
29 |
+
This model was trained with PPO, a method introduced in [Fine-Tuning Language Models from Human Preferences](https://huggingface.co/papers/1909.08593).
|
30 |
+
|
31 |
+
### Framework versions
|
32 |
+
|
33 |
+
- TRL: 0.12.1
|
34 |
+
- Transformers: 4.46.2
|
35 |
+
- Pytorch: 2.5.1
|
36 |
+
- Datasets: 3.1.0
|
37 |
+
- Tokenizers: 0.20.3
|
38 |
+
|
39 |
+
## Citations
|
40 |
+
|
41 |
+
Cite PPO as:
|
42 |
+
|
43 |
+
```bibtex
|
44 |
+
@article{mziegler2019fine-tuning,
|
45 |
+
title = {{Fine-Tuning Language Models from Human Preferences}},
|
46 |
+
author = {Daniel M. Ziegler and Nisan Stiennon and Jeffrey Wu and Tom B. Brown and Alec Radford and Dario Amodei and Paul F. Christiano and Geoffrey Irving},
|
47 |
+
year = 2019,
|
48 |
+
eprint = {arXiv:1909.08593}
|
49 |
+
}
|
50 |
+
```
|
51 |
+
|
52 |
+
Cite TRL as:
|
53 |
+
|
54 |
+
```bibtex
|
55 |
+
@misc{vonwerra2022trl,
|
56 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
57 |
+
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},
|
58 |
+
year = 2020,
|
59 |
+
journal = {GitHub repository},
|
60 |
+
publisher = {GitHub},
|
61 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
62 |
+
}
|
63 |
+
```
|
generation_config.json
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"transformers_version": "4.46.2"
|
4 |
+
}
|