GAN Lab is an interactive visualization tool for observing the training process of a generative adversarial network (GAN) in the browser. On a 2D data distribution, you can watch the generator create samples while the discriminator learns to distinguish real from fake examples through changes in points and decision boundaries.
By changing the distribution shape, model structure, and learning speed, you can try cases where GAN training stabilizes or breaks down. It is a useful learning tool for understanding the roles of the generator and discriminator and the competitive relationship between them.
