lucidrains/stylegan2-pytorch
A PyTorch implementation of StyleGAN2 for training generative adversarial networks that produce high-quality synthetic images.

Velocity · 7d
+1.6
★ / day
Trend
→steady
star history
This repository provides a complete, usable implementation of StyleGAN2 based on the academic paper, implemented in PyTorch. It enables training the GAN model directly from the command line on custom image datasets with no coding required. The model learns to generate photorealistic images by optimizing the adversarial training process between a generator and discriminator network.