StreamEdit

KO | EN

gitlite — search

#python #ai #typescript #golang #go #deep-learning #docker #macos #windows #cli #kubernetes #vue #security #pytorch

(★ 21)

[ECCV 2026] StreamEdit: Training-Free Video Editing via Few-Step Streaming Video Generation

StreamEdit 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# StreamEdit: Training-Free Video Editing via Few-Step Streaming Video Generation

[Guanlong Jiao](https://scholar.google.com/citations?user=UCPOxvEAAAAJ&hl=en)1,4, [Chenyangguang Zhang](https://zhangcyg.github.io/)2, [Jia Jun Cheng Xian](https://www.linkedin.com/in/jia-jun-cheng-xian-56b00a208/)1,4, [Zewei Zhang](https://scholar.google.com/citations?user=RWGzBeoAAAAJ&hl=en)1,3, [Renjie Liao](https://lrjconan.github.io/)1,4,5

1The University of British Columbia, 2ETH Zürich, 3McMaster University, 4Vector Institute, 5Canada CIFAR AI Chair

[![Paper](https://img.shields.io/badge/ArXiv-Paper-red?style=flat-square&logo=arxiv&logoColor=white)](http://arxiv.org/abs/2605.21466)
[![Code](https://img.shields.io/badge/GitHub-Code-blue?style=flat-square&logo=github&logoColor=white)](https://github.com/DSL-Lab/StreamEdit/)
[![Page](https://img.shields.io/badge/StreamEdit-Project_Page-peru?style=flat-square&logo=googlechrome&logoColor=white)](https://dsl-lab.github.io/StreamEdit/)

Video Results are all shown in our [*Project Page*](https://dsl-lab.github.io/StreamEdit/).

---

## ✨ Highlights

StreamEdit is a training-free video editing framework built on few-step streaming video generation models. Instead of treating editing as data-to-data transformation, StreamEdit formulates editing as source-conditioned noise-to-target generation, thus makeing it possible for few-step fast controllable video editing. StreamEdit supports **few-step text-driven video editing** and **optional first-frame visual prompting** for **videos of any length**. It is implemented on streaming video generation models:

- [`Self-Forcing_StreamEdit`](./Self-Forcing_StreamEdit/README.md): StreamEdit built on [Self Forcing](https://github.com/guandeh17/Self-Forcing).
- [`LongLive_StreamEdit`](./LongLive_StreamEdit/README.md): StreamEdit built on [LongLive v1.0](https://github.com/NVlabs/LongLive/tree/v1.0).

## 🏡 Environment

```bash
conda create -n streamedit python=3.12 -y
conda activate streamedit

# Choose the PyTorch command appropriate for your CUDA version, example for CUDA 12.8:
pip install torch==2.8.0 torchvision==0.23.0 --index-url https://download.pytorch.org/whl/cu128

pip install -r requirements.txt
pip install flash-attn --no-build-isolation
```

For the Self Forcing-based implementation, install the local package in editable mode:

```bash
cd Self-Forcing_StreamEdit
python setup.py develop
cd ..
```

## 🎯 Checkpoints

### Self Forcing

```bash
cd Self-Forcing_StreamEdit
huggingface-cli download Wan-AI/Wan2.1-T2V-1.3B --local-dir-use-symlinks False --local-dir wan_models/Wan2.1-T2V-1.3B
huggingface-cli download gdhe17/Self-Forcing checkpoints/self_forcing_dmd.pt --local-dir .
cd ..
```

### LongLive

```bash
cd LongLive_StreamEdit
huggingface-cli download Wan-AI/Wan2.1-T2V-1.3B --local-dir wan_models/Wan2.1-T2V-1.3B
huggingface-cli download Efficient-Large-Model/LongLive --local-dir longlive_models
cd ..
```

You can share the checkpoints of Wan through soft links, as they are the same across projects.

## 🎨 Running Video Editing

Each implementation provides a ready-to-run example script.

### Self Forcing-based editing

```bash
cd Self-Forcing_StreamEdit
bash inference_edit_streamedit.sh
```

### LongLive-based editing

```bash
cd LongLive_StreamEdit
bash inference_edit_streamedit.sh
```

## 🎉 Acknowledgements

This repository builds on the excellent open-source work of [Self Forcing](https://github.com/guandeh17/Self-Forcing), [LongLive](https://github.com/NVlabs/LongLive), and [Wan2.1](https://github.com/Wan-Video/Wan2.1). We also thank [UniEdit-Flow](https://github.com/DSL-Lab/UniEdit-Flow) and [FiVE-Bench](https://github.com/MinghanLi/FiVE-Bench) for helpful open-soucre code and benchmarks.

## 🔮 Citation

If you find this work useful, please consider citing, thanks!

```bibtex
@inproceedings{jiao2026streamedit,
      title={StreamEdit: Training-Free Video Editing via Few-Step Streaming Video Generation}, 
      author={Guanlong Jiao and Chenyangguang Zhang and Jia Jun Cheng Xian and Zewei Zhang and Renjie Liao},
      booktitle={Proceedings of the European Conference on Computer Vision (ECCV)},
      year={2026}
}
```

StreamEdit

StreamEdit 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

StreamEdit

StreamEdit 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README