GVM

KO | EN

gitlite — search

#javascript #hacktoberfest #typescript #go #deep-learning #windows #awesome #kubernetes #awesome-list #vue #mcp #ios #openai #open-source

(★ 92)

[SIGGRAPH2025] Generative Video Matting

File Explorer

Download Latest Version (.zip)

- - __init__.py
  - unet_spatio_temporal_condition.py
- - pipeline_gvm.py
- - __init__.py
  - inference_utils.py
- __init__.py
demo.py
README.md
requirements.txt
setup.py

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<div align="center">
<div style="text-align: center;">
    <h1>Generative Video Matting</h1>
</div>

**SIGGRAPH2025**

</div>

##  📖 Table of Contents

- [Generative Video Matting](#-generative-video-matting)
  - [🔥 News](#-news)
  - [🚀 Getting Started](#-getting-started)
    - [Environment Requirement 🌍](#environment-requirement-)
    - [Download Model Weights ⬇️](#download-️model-weights-)
  - [🏃🏼 Run](#-run)
    - [Inference 📜](#inference-)
    - [Evaluation 📏](#evaluation-)
  - [🎫 License](#-license)
  - [📢 Disclaimer](#-disclaimer)
  - [🤝 Cite Us](#-cite-us)

## 🔥 News
- **August 10, 2025:** Release the inference code and model checkpoints.
- **June 11, 2025:** Repo created. The code and dataset for this project are currently being prepared for release and will be available here soon. Please stay tuned!

## 🚀 Getting Started

### Environment Requirement 🌍

First, clone the repo:

```
git clone https://github.com/aim-uofa/GVM.git
cd GVM
```

Then, we recommend you first use `conda` to create virtual environment, and install needed libraries. For example:

```
conda create -n gvm python=3.10 -y
conda activate gvm
pip install -r requirements.txt
python setup.py develop
```

### Download Model Weights ⬇️

You need to download the model weights by:

```
hugginface-cli download geyongtao/gvm --local-dir data/weights
```

The ckpt structure should be like:

## 🏃🏼 Run

### Inference 📜

You can run generative video matting with:

```
python demo.py \
--model_base 'data/weights/' \
--unet_base data/weights/unet \
--lora_base data/weights/unet \
--mode 'matte' \
--num_frames_per_batch 8 \
--num_interp_frames 1 \
--num_overlap_frames 1 \
--denoise_steps 1 \
--decode_chunk_size 8 \
--max_resolution 960 \
--pretrain_type 'svd' \
--data_dir 'data/demo_videos/xxx.mp4' \
--output_dir 'output_path'
```

### Evaluation 📏

```
TODO
```

## 🎫 License

For academic usage, this project is licensed under [the 2-clause BSD License](LICENSE). For commercial inquiries, please contact [Chunhua Shen](mailto:chhshen@gmail.com).

## 📢 Disclaimer

This repository provides a one-step model for faster inference speed. Its performance is slightly different from the results reported in the original SIGRRAPH paper.

## 🤝 Cite Us

If you find this work helpful for your research, please cite:
```
@inproceedings{ge2025gvm,
author = {Ge, Yongtao and Xie, Kangyang and Xu, Guangkai and Ke, Li and Liu, Mingyu and Huang, Longtao and Xue, Hui and Chen, Hao and Shen, Chunhua},
title = {Generative Video Matting},
publisher = {Association for Computing Machinery},
url = {https://doi.org/10.1145/3721238.3730642},
doi = {10.1145/3721238.3730642},
booktitle = {Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers},
series = {SIGGRAPH Conference Papers '25}
}
```

GVM

File Explorer

Today's picks

Was this content helpful?

Recent Feedback

GVM

File Explorer

Today's picks

Was this content helpful?

Recent Feedback

Download README