Fine tuning, blending of styleGAN2 models and creation animations using latent directions.
- This repository uses the official NVlabs StyleGAN2-ADA PyTorch implementation.
https://github.com/NVlabs/stylegan2-ada-pytorch - Only 1024 x 1024 styleGAN2 models can be used to create animations.
Fine tuning notebook
Animation notebook
Blending notebook
Path | Description |
---|---|
stylegan2-ada-pytorch | Contains tools to load and train NVlabs stylegan2 models |
datasets | Contains datasets used for fine tuning |
├ hubble_1024x1024.zip | Dataset of 261 Hubble Space Telescope snapshots taken from Nasa Hubble Space Telescope page. |
directions | Contains latent directions |
fine_tuning.ipynb | Notebook used for fine tuning |
blending.ipynb | Notebook used for blending |
animation.ipynb | Notebook used for creating animations |
This project was used for creation Void Face NFT collection.
All models are stored in
huggingface repository.
All news about this project is posted on my Twitter.