File size: 2,295 Bytes
2049877
 
bba06d9
2049877
 
 
2b5940a
2049877
a134aba
 
2049877
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6f9b7f8
2049877
 
 
 
 
 
 
 
 
 
2eb1233
2049877
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
---
tags:
- image-to-video
- video-generation
- personalized-video
- identity-preservation
pipeline_tag: image-to-video
license: apache-2.0
base_model:
- Wan-AI/Wan2.1-T2V-14B
---

<div align="center">

# Lynx: High-Fidelity Personalized Video Generation

<a href="https://arxiv.org/abs/2509.15496"><img src="https://img.shields.io/badge/arXiv-2509.15496-b31b1b.svg"/></a>
<a href="https://byteaigc.github.io/Lynx/"><img src="https://img.shields.io/badge/Project-Page-green.svg"/></a>
<a href="https://github.com/bytedance/lynx"><img src="https://img.shields.io/badge/GitHub-Repo-blue.svg"/></a>

<h5 style="font-size:1.1em; letter-spacing:0.5px;">
<a href="https://ssangx.github.io/">Shen Sang*</a>&nbsp;&nbsp;&nbsp;
<a href="https://tiancheng-zhi.github.io/">Tiancheng Zhi*</a>&nbsp;&nbsp;&nbsp;
<a href="https://gutianpei.github.io/">Tianpei Gu</a>&nbsp;&nbsp;&nbsp;
<a href="https://www.jingliu.net/">Jing Liu</a>&nbsp;&nbsp;&nbsp;
<a href="https://linjieluo.github.io/">Linjie Luo</a>
</h5>

<p style="font-size: 1.05em; margin: 8px 0;">
Intelligent Creation, ByteDance
</p>

<p style="font-size: 0.95em; font-style: italic;">
* Equal Contribution
</p>

<br>
<img src="assets/teaser.jpg" width="400"/>

</div>

Lynx is a state-of-the-art high-fidelity personalized video generation model that creates videos from a single input image while preserving the subject's identity. Built on a Diffusion Transformer (DiT) foundation model with lightweight ID-adapters and Ref-adapters for identity preservation and spatial detail enhancement.

## Model Variants

This repository contains two model variants:

- **Lynx Full Model** (`lynx_full`): Complete version with all advanced features and best performance
- **Lynx Lite Model** (`lynx_lite`): Lightweight model with fewer parameters (no Ref-adapter), tailored for efficient 24fps (121-frame) video generation.

## Citation

If you use this model in your research, please cite:

```bibtex
@article{sang2025lynx,
  title={Lynx: Towards High-Fidelity Personalized Video Generation},
  author={Sang, Shen and Zhi, Tiancheng and Gu, Tianpei and Liu, Jing and Luo, Linjie},
  journal={arXiv preprint arXiv:2509.15496},
  year={2025}
}
```

## License

This model is licensed under the Apache License 2.0. See the [LICENSE](LICENSE) file for details.