GAN Lab
lastmod 2026-11-10created 2021-11-302 edits
GAN Lab is an educational tool for visually experimenting with the training process of generative adversarial networks (GANs) in a web browser. It draws, in real time, the generator and discriminator training in turn on a two-dimensional data distribution, and lets you observe the effects of changing settings such as the learning rate or loss function. It is implemented with TensorFlow.js, so it runs directly in the browser without any installation.
GAN Lab: An Interactive, Visual Experimentation Tool for Generative Adversarial Networks