News

• We re-implement CycleGAN by Tensorflow 2! The old versions are here: v1, v0.

---

---

CycleGAN - Tensorflow 2

Tensorflow 2 implementation of CycleGAN.

Paper: Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks

Author: Jun-Yan Zhu et al.

Exemplar results

summer2winter

row 1: summer -> winter -> reconstructed summer, row 2: winter -> summer -> reconstructed winter

horse2zebra

row 1: horse -> zebra -> reconstructed horse, row 2: zebra -> horse -> reconstructed zebra

apple2orange

row 1: apple -> orange -> reconstructed apple, row 2: orange -> apple -> reconstructed orange

Usage

Setup

conda create -n cyclegan python=3.10

conda activate cyclegan

sh ./setup-macos-conda.sh

• Environment

  • Python 3.10

  • TensorFlow 2.9, TensorFlow Addons 0.17.1

  • OpenCV, scikit-image, tqdm, oyaml

  • we recommend Anaconda or Miniconda, then you can create the TensorFlow 2.9 environment with commands below

    conda create -n cyclegan python=3.10
    
    conda activate cyclegan
    
    conda install -c apple scikit-image tqdm tensorflow=2.9
    
    conda install -c conda-forge oyaml
    
    pip install tensorflow-addons==0.17.1 tensorflow-macos==2.9.0 tensorflow-metal==0.5.0

  • NOTICE: if you create a new conda environment, remember to activate it before any other command

    source activate cyclegan

• Dataset

  • download the summer2winter dataset

    sh ./download_dataset.sh summer2winter_yosemite

  • download the horse2zebra dataset

    sh ./download_dataset.sh horse2zebra

  • see download_dataset.sh for more datasets

• Example of training

  CUDA_VISIBLE_DEVICES=0 python train.py --dataset summer2winter_yosemite

  • tensorboard for loss visualization

    tensorboard --logdir ./output/summer2winter_yosemite/summaries --port 6006

• Example of testing

  CUDA_VISIBLE_DEVICES=0 python test.py --experiment_dir ./output/summer2winter_yosemite

saving a model

The test.py file also provides the ability to save a model for re-running on different machines and through the run.py script

 python test.py --save 1 --model_dir /Users/hcwiley/ml-models/hcwiley/$model_name --experiment_dir ./output/$model_name