This is an unofficial TensorFlow 2.0 implementation of the paper with full compatibility with the orignal code:
A Style-Based Generator Architecture for Generative Adversarial Networks
Tero Karras (NVIDIA), Samuli Laine (NVIDIA), Timo Aila (NVIDIA)
http://stylegan.xyz/paper
Most of the code is just a translation of the offical repo.
You can see how to load the model and generate samples in this notebook.
Training the model is not yet well supproted (might come later in the future). The main goal of this repo is to quickly load up StyleGAN and experiment with a pretrained model.