The repository contains GAN-based image editing.
Demo is avaiable on Colab.
The demo video (long version) is available.
We also provide an interface for interactive editing based on StreamLit. This interface can be locally launched with
CUDA_VISIBLE_DEVICES=0 streamlit run interface.py
NOTE: User can upload your own image though the file upload dialog. Please wait a few minutes for finishing the optimization of latent code that suits for the uploaded image.
This demo is based on two algorithms: Image2StyleGAN and SeFa. It borrows some codes from Image2StyleGAN and Sefa repositories.