https://github.com/bryandlee/stylegan2-encoder-pytorch

PyTorch Implementation of In-Domain GAN Inversion for StyleGAN2
https://github.com/bryandlee/stylegan2-encoder-pytorch

gan image-generation

Last synced: about 2 months ago
JSON representation

PyTorch Implementation of In-Domain GAN Inversion for StyleGAN2

• Host: GitHub
• URL: https://github.com/bryandlee/stylegan2-encoder-pytorch
• Owner: bryandlee
• License: mit
• Created: 2020-08-24T12:53:46.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2020-12-09T08:33:08.000Z (about 4 years ago)
• Last Synced: 2023-11-07T17:59:55.264Z (about 1 year ago)
• Topics: gan, image-generation
• Language: Jupyter Notebook
• Homepage:
• Size: 15.4 MB
• Stars: 143
• Watchers: 7
• Forks: 24
• Open Issues: 8
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# In-Domain GAN Inversion for Real Image Editing

Based on **Seonghyeon Kim's Pytorch Implementation of StyleGAN2**

[[Paper](https://arxiv.org/pdf/2004.00049.pdf)] [[Official Code](https://github.com/genforce/idinvert)] [[StyleGAN2 Pytorch](https://github.com/rosinality/stylegan2-pytorch)]

## Train Encoder

```

python train_encoder.py

```

**0k iter**\



**1M iter**\

\

[[encoder checkpoint](https://drive.google.com/file/d/1QQuZGtHgD24Dn5E21Z2Ik25EPng58MoU/view?usp=sharing)] [[generator checkpoint](https://drive.google.com/file/d/1TH77dUsqcq50htIZT6DljFYr4T_ziJli/view)]

**Note:** The encoder architecture and loss weights are different from the original implemetation.

## Interpolation

```

interpolate.ipynb

```

**Domain-Guided Encoder (Initial projection)**\



**In-Domain Inversion (500 steps)**\



**Inperpolation Result**\



## Encoder + Model Interpolation

[[Paper](https://arxiv.org/abs/2010.05334)] [[Naver Webtoon Model](https://github.com/bryandlee/naver-webtoon-faces#stylegan2)]