---

This repository provides a optimized tensorflow implementation of StarGAN.

Paper

StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018 (Oral)

Dependencies

• Python 2.7
• PyTorch 0.4.0
• TensorFlow 1.3+

Getting started

1. Clone the code

git clone https://github.com/HeroKillerEver/StarGAN-tensorflow.git
cd StarGAN-tensorflow

2. Download the dataset

To download the CelebA dataset:

$ bash download.sh celeba

3. Preprocess the dataset

python util.py

4. Training

python main.py

you may also type

python main.py --help

to see all of the input arguments

Author

Haibin Yu @HeroKillerEver