File size: 1,535 Bytes
a2a5d5f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: other
license_name: s-lab-1.0
license_link: https://huggingface.co/AEmotionStudio/matanyone2/blob/main/LICENSE.txt
tags:
  - video-matting
  - alpha-matting
  - comfyui
  - ffmpega
---

# MatAnyone2 — SafeTensors Mirror

This is a **SafeTensors conversion** of the [MatAnyone2](https://github.com/pq-yang/MatAnyone2) video matting model (CVPR 2026) for use with [ComfyUI-FFMPEGA](https://github.com/AEmotionStudio/ComfyUI-FFMPEGA).

## Model Details

| Property | Value |
|----------|-------|
| Architecture | MatAnyone2 (temporal video matting) |
| Parameters | 562 tensors, ~135 MB |
| Format | SafeTensors |
| Original Format | PyTorch (.pth) |
| Source | [pq-yang/MatAnyone2](https://github.com/pq-yang/MatAnyone2) |

## ⚠️ License Notice

This model is released under the **NTU S-Lab License 1.0** which is strictly **non-commercial**.

> Redistribution and use for non-commercial purpose in source and binary forms, with or without modification, are permitted provided that the following conditions are met...

See [LICENSE.txt](LICENSE.txt) for the full license text. **Commercial use requires explicit written permission from the original authors.**

## Usage

This model is automatically downloaded by ComfyUI-FFMPEGA when the `video_matting` no-LLM mode is selected. No manual setup is required.

## Citation

```bibtex
@article{yang2025matanyone2,
  title={MatAnyone2: Stable Video Matting with Consistent Memory Propagation},
  author={Yang, Peiqing and Shang, Shangchen and others},
  year={2025}
}
```