Fudan-FUXI
commited on
Commit
•
cd4d187
1
Parent(s):
52405fe
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,53 @@
|
|
1 |
-
---
|
2 |
-
license: mit
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
base_model:
|
6 |
+
- Efficient-Large-Model/VILA1.5-40b
|
7 |
+
pipeline_tag: video-text-to-text
|
8 |
+
---
|
9 |
+
|
10 |
+
# LiFT: Leveraging Human Feedback for Text-to-Video Model Alignment
|
11 |
+
|
12 |
+
|
13 |
+
## Summary
|
14 |
+
This is the model checkpoint proposed in our paper "LiFT: Leveraging Human Feedback for Text-to-Video Model Alignment". LiFT-Critic is a novel Video-Text-to-Text Reward Model for synthesized video evaluation.
|
15 |
+
|
16 |
+
Project: https://codegoat24.github.io/LiFT/
|
17 |
+
|
18 |
+
Code: https://github.com/CodeGoat24/LiFT
|
19 |
+
|
20 |
+
## 🔧 Installation
|
21 |
+
|
22 |
+
1. Clone the github repository and navigate to LiFT folder
|
23 |
+
```bash
|
24 |
+
git clone https://github.com/CodeGoat24/LiFT.git
|
25 |
+
cd LiFT
|
26 |
+
```
|
27 |
+
2. Install packages
|
28 |
+
```
|
29 |
+
bash ./environment_setup.sh lift
|
30 |
+
```
|
31 |
+
|
32 |
+
## 🚀 Inference
|
33 |
+
|
34 |
+
### Run
|
35 |
+
Please download this public [LiFT-Critic-40b-lora-v1.5](https://huggingface.co/Fudan-FUXI/LiFT-Critic-13b-lora-v1.5) checkpoints.
|
36 |
+
|
37 |
+
We provide some synthesized videos for quick inference in `./demo` directory.
|
38 |
+
|
39 |
+
```bash
|
40 |
+
python LiFT-Critic/test/run_critic_40b.py --model-path ./LiFT-Critic-40b-lora-v1.5
|
41 |
+
```
|
42 |
+
|
43 |
+
# 🖊️ Citation
|
44 |
+
|
45 |
+
If you find our work helpful, please cite our paper.
|
46 |
+
```bibtex
|
47 |
+
@article{LiFT,
|
48 |
+
title={LiFT: Leveraging Human Feedback for Text-to-Video Model Alignment.},
|
49 |
+
author={Wang, Yibin and Tan, Zhiyu, and Wang, Junyan and Yang, Xiaomeng and Jin, Cheng and Li, Hao},
|
50 |
+
journal={arXiv preprint arXiv:2412.04814},
|
51 |
+
year={2024}
|
52 |
+
}
|
53 |
+
```
|