/DragGAN

Unofficial Implementation of DragGAN - "Drag Your GAN: Interactive Point-based Manipulation on the Generative Image Manifold" （DragGAN 全功能实现，在线Demo，本地部署试用，代码、模型已全部开源，支持Windows, macOS, Linux）

Primary LanguagePython

DragGAN

PyPI support

💥 Colab Demo | InternGPT Demo | Local Deployment

🤗 Huggingface Demo1 | Huggingface Demo2

Note for Colab, remember to select a GPU via Runtime/Change runtime type (代码执行程序/更改运行时类型).

If you want to upload custom image, please install 1.1.0 via pip install draggan==1.1.0.

Unofficial implementation of Drag Your GAN: Interactive Point-based Manipulation on the Generative Image Manifold

How it Work ?

Here is a simple tutorial video showing how to use our implementation.

demo.mp4

Check out the original paper for the backend algorithm and math.

demo

News

🌟 What's New

  • [2023/6/25] Relase version 1.1.1, it includes a major bug fix and speed improvement.
  • [2023/6/25] Official Code is released, check it out.
  • [2023/5/29] A new version is in beta, install via pip install draggan==1.1.0b2, includes speed improvement and more models.
  • [2023/5/25] DragGAN is on PyPI, simple install via pip install draggan. Also addressed the common CUDA problems #38 #12
  • [2023/5/25] We now support StyleGAN2-ada with much higher quality and more types of images. Try it by selecting models started with "ada".
  • [2023/5/24] An out-of-box online demo is integrated in InternGPT - a super cool pointing-language-driven visual interactive system. Enjoy for free.🍭
  • [2023/5/24] Custom Image with GAN inversion is supported, but it is possible that your custom images are distorted due to the limitation of GAN inversion. Besides, it is also possible the manipulations fail due to the limitation of our implementation.

🌟 Changelog

  • Add a docker image, thanks @egbaydarov.
  • PTI GAN inversion #71 (comment)
  • Tweak performance, See v2.
  • Improving installation experience, DragGAN is now on PyPI.
  • Automatically determining the number of iterations, See v2.
  • Allow to save video without point annotations, custom image size.
  • Support StyleGAN2-ada.
  • Integrate into InternGPT
  • Custom Image with GAN inversion.
  • Download generated image and generation trajectory.
  • Controlling generation process with GUI.
  • Automatically download stylegan2 checkpoint.
  • Support movable region, multiple handle points.
  • Gradio and Colab Demo.

This project is now a sub-project of InternGPT for interactive image editing. Future updates of more cool tools beyond DragGAN would be added in InternGPT.

Running Locally

Please refer to INSTALL.md.

Citation

@inproceedings{pan2023draggan,
    title={Drag Your GAN: Interactive Point-based Manipulation on the Generative Image Manifold}, 
    author={Pan, Xingang and Tewari, Ayush, and Leimk{\"u}hler, Thomas and Liu, Lingjie and Meka, Abhimitra and Theobalt, Christian},
    booktitle = {ACM SIGGRAPH 2023 Conference Proceedings},
    year={2023}
}

Acknowledgement

Official DragGAN   DragGAN-Streamlit   StyleGAN2   StyleGAN2-pytorch   StyleGAN2-Ada   StyleGAN-Human   Self-Distilled-StyleGAN

image