Implementation of various GANs with TensorFlow version 2.0

• Final update: 2019. 06. 13.
• All right reserved @ Il Gu Yi 2018

This repository is a collection of various GAN models implemented by TensorFlow version 2.0 style.

This repository moves to ilguyi/generative.models.tensorflow.v2 that is a collection of various generative models including autoregressive models, latent variable models, normalizing flow models as well as GAN. You will see more implementations of generative models.

Getting Started

Prerequisites

• TensorFlow above 1.12
• Python 3.6
• Python libraries:
  • numpy, matplotlib, PIL
• Jupyter notebook
• OS X and Linux (Not validated on Windows but probably it would work)

Contents (TF version 2.x style)

Generative Adversarial Networks

• Original GAN paper arXiv:1406.2661
• gan.ipynb

Contents (TF version 1.x style)

Generative Adversarial Networks

• Original GAN paper arXiv:1406.2661
• gan.ipynb

Deep Convolutional GAN

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks paper arXiv:1511.06434
• dcgan.ipynb

Conditional GAN

• Conditional Generative Adversarial Nets arXiv:1411.1784
• cgan_based_on_dcgan.ipynb

InfoGAN

• InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets arXiv:1606.03657
• infogan.ipynb

Bidirectional GAN

• Adversarial Feature Learning arXiv:1605.09782
• bigan.ipynb

Least Squares GAN

• Least Squares Generative Adversarial Networks arXiv:1611.04076
• lsgan.ipynb

Wasserstein GAN

• Wasserstein GAN arXiv:1701.07875
• wgan.ipynb

Boundary Equilibrium GAN

• BEGAN: Boundary Equilibrium Generative Adversarial Networks arXiv:1703.10717
• began.ipynb

Author

Il Gu Yi