This document discusses StyleGAN, a style-based generator architecture for generative adversarial networks (GANs). StyleGAN uses adaptive instance normalization (AdaIN) to separate high-level attributes like content from style by modulating the instance statistics of intermediate features. It achieves state-of-the-art results on face generation. The document outlines the StyleGAN architecture, compares it to a BigGAN-like model, and discusses experiments applying StyleGAN2 to face datasets and BigGAN to other domains.