code-asc/DISCO-GAN

DISCO-GAN

This project is the implementation of the DISCO-GAN paper. Here we try to understand the cross domain relation between GAN's.

GAN sharing the domain knowledge(Left image is the input and right image is the output)

Installation

Use the package manager pip to install foobar.

pip install cv2
pip install numpy
pip install matplotlib
pip install torch
pip install torchvision

Instructions.

Download the img_align_celeba dataset and paste it in data folder. Also need to download list_attr_celeba.txt and paste it in data folder.

Run the main.py to train both discriminator and generator. It takes a while to complete execution. Finally run samples.py file to generate the fake samples.