Skip to content
On Stabilizing Generative Adversarial Training with Noise. In CVPR, 2019.
Python
Branch: master
Clone or download
Latest commit 7be6aa2 Aug 22, 2019

README.md

On Stabilizing Generative Adversarial Training with Noise [Project Page]

This repository contains demo code of our CVPR2019 paper. It contains code for the training and evaluation of a DFGAN with learned noise on the CIFAR-10 dataset.

Requirements

The code is based on Python 2.7 and tensorflow 1.12.

How to use it

1. Setup

  • Set the paths to the data and log directories in constants.py.
  • Run init_datasets.py to download and convert the CIFAR-10 dataset.

2. DFGAN training and evaluation

  • To train and evaluate a DFGAN with learned noise on CIFAR-10 run run_DFGAN_ln.py.
  • To train and evaluate a standard GAN on CIFAR-10 run run_standard_GAN.py.

PyTorch implementation

A PyTorch implementation of DFGAN is provided by: https://github.com/Johnson-yue/pytorch-DFGAN

You can’t perform that action at this time.