Keras-GAN

Why we included this project

For anyone trying to get a practical handle on how GAN architectures work, this repo is a handy starting point. It holds around twenty self-contained Keras scripts, each corresponding to a known GAN paper, from the basic GAN and DCGAN to Pix2Pix, CycleGAN, InfoGAN, SRGAN, and the Wasserstein variants. The author deliberately simplified the models to focus on the central ideas rather than matching every detail of the original papers, which makes them easy to run and read. Just keep in mind the project is no longer maintained, so it targets older Keras and TensorFlow APIs and is best treated as a learning resource, not production code.