Designing an Encoder for StyleGAN Image Manipulation
Omer Tov, Yuval Alaluf, Yotam Nitzan, Or Patashnik, Daniel Cohen-Or. In SIGGRAPH, 2021.
The original page: [Project Page]
We implement the E4E encoder for facial images, both for StyleGAN2 and StyleGAN3.
The official weight are converted jittor for StyleGAN2. The weights for StyleGAN3 are trained by ourselves.
The Google Drive link: https://drive.google.com/drive/folders/1Vvf9pJ_NGj0oB7dBs8Be-Bfk0T08BEup?usp=sharing
Project real image into StyleGAN2 latent space :
python inference_stylegan2.py --images_path ./imgs/input/69918.png --save_path ./imgs/output/69918_stylegan2.png --ckpt ./weights/e4e_ffhq_encode_stylegan2.pt
Project real image into StyleGAN3 latent space:
python inference_stylegan3.py --images_path ./imgs/input/69918.png --save_path ./imgs/output/69918_stylegan3.png --ckpt ./weights/e4e_ffhq_encode_stylegan3.pt
This repository borrows partially from the original codes and Sketch_GAN repository.