GFP-GAN aims at developing a Practical Algorithm for Real-world Face Restoration. It leverages rich and diverse priors encapsulated in a pretrained face GAN (e.g., StyleGAN2) for face restoration. This Generative Facial Prior (GFP) is incorporated into the face restoration process via novel channel-split spatial feature transform layers, which allow our method to achieve a good balance of realness and fidelity.