TF-GAN

TF-GAN is a lightweight library for training and evaluating Generative Adversarial Networks in TensorFlow. It provides core training infrastructure, common GAN operations and normalization techniques, losses and penalties such as the Wasserstein loss and gradient penalty, and standard evaluation metrics including Inception Score, Frechet Distance and Kernel Distance. It can be installed with pip. The repository was archived and is now read-only.

GitHub repo Directory import
0 shouts Added Sep 23, 2026 github.com
github.com/tensorflow/gan
TF-GAN website preview

Features

  • lightweight library for training and evaluating Generative Adversarial Networks Source: github.com

Links

Tags