metadata
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 video matting model (CVPR 2026) for use with 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 |
⚠️ 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 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
@article{yang2025matanyone2,
title={MatAnyone2: Stable Video Matting with Consistent Memory Propagation},
author={Yang, Peiqing and Shang, Shangchen and others},
year={2025}
}