Building a Generative Adversarial Network (GAN) in PyTorch.
Loading the pretrained parameters in PyTorch.
Training and evaluating your GAN in PyTorch.
Python 3, Numpy, matplotlib, torch, torchvision
CelebA (CelebFaces Attributes Dataset) is a large-scale face attributes dataset with more than 200K celebrity images, each with 40 attribute annotations. The images in this dataset cover large pose variations and background clutter. In this assignment, we use the CelebA dataset to train a GAN in advance as a pretraining step.
AnimeFace is a dataset consisting of 21551 anime faces scraped from an anime website. All images are resized to 64 * 64 for the sake of convenience. In this assignment, we perform transfer learning and treat the AnimeFace dataset as the downstream dataset.