Implementation of StyleGan3+Clip using microservices (so you can use it directly)
The code is inspired by the official paper and other resources.
A gentle introduction to StyleGan is avaiable here.
Make sure to have docker installed and an accessible GPU.
git clone https://github.com/chumpblocckami/StyleGan3Clip_Service.git
cd StyleGan3Clip_Service
docker-compose up
then the service will be up and running at http://localhost:5000.
The attributes to enter are:
- model: the model used to perform the GAN. For more info: Pretrained models
- text: the text to generate the style from.
- steps: how many cycles to perform.
- seed: same number will reproduce the same result.
MODEL | TEXT | STEP | SEED | START | END | TRANSITION |
---|---|---|---|---|---|---|
FFHQ | Harry Potter trading card game artwork | 200 | - |