This implementation in Keras is based on CycleGAN and pix2pix in PyTorch

Prerequisites

-Windows or Linux or macOS -Python 3 -GPU + CUDA CuDNN

Installation

• install Keras pip install keras

pip install keras

CycleGAN train/test

• Download Datasets

bash ./datasets/download_cyclegan_dataset.sh maps

• Train Model

CUDA_VISIBLE_DEVICES=0 model.py

The test results will be saved to html file: ./results/maps_cyclegan/latest_test/index.html

Acknowledgments

This code and Github ideas are inspired by pytorch-CycleGan, Keras-GAN,Tensorflow-GAN