# PyTorch-CycleGAN

**Repository Path**: bighank/PyTorch-CycleGAN

## Basic Information

- **Project Name**: PyTorch-CycleGAN
- **Description**: A clean and readable Pytorch implementation of CycleGAN
- **Primary Language**: Python
- **License**: GPL-3.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 2
- **Created**: 2024-01-16
- **Last Updated**: 2024-01-16

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Pytorch-CycleGAN
A clean and readable Pytorch implementation of CycleGAN (https://arxiv.org/abs/1703.10593)

## Prerequisites
Code is intended to work with ```Python 3.6.x```, it hasn't been tested with previous versions

### [PyTorch & torchvision](http://pytorch.org/)
Follow the instructions in [pytorch.org](http://pytorch.org) for your current setup

### [Visdom](https://github.com/facebookresearch/visdom)
To plot loss graphs and draw images in a nice web browser view
```
pip3 install visdom
```

## Training
### 1. Setup the dataset
First, you will need to download and setup a dataset. The easiest way is to use one of the already existing datasets on UC Berkeley's repository:
```
./download_dataset <dataset_name>
```
Valid <dataset_name> are: apple2orange, summer2winter_yosemite, horse2zebra, monet2photo, cezanne2photo, ukiyoe2photo, vangogh2photo, maps, cityscapes, facades, iphone2dslr_flower, ae_photos

Alternatively you can build your own dataset by setting up the following directory structure:

    .
    ├── datasets                   
    |   ├── <dataset_name>         # i.e. brucewayne2batman
    |   |   ├── train              # Training
    |   |   |   ├── A              # Contains domain A images (i.e. Bruce Wayne)
    |   |   |   └── B              # Contains domain B images (i.e. Batman)
    |   |   └── test               # Testing
    |   |   |   ├── A              # Contains domain A images (i.e. Bruce Wayne)
    |   |   |   └── B              # Contains domain B images (i.e. Batman)
    
### 2. Train!
```
./train --dataroot datasets/<dataset_name>/ --cuda
```
This command will start a training session using the images under the *dataroot/train* directory with the hyperparameters that showed best results according to CycleGAN authors. You are free to change those hyperparameters, see ```./train --help``` for a description of those.

Both generators and discriminators weights will be saved under the output directory.

If you don't own a GPU remove the --cuda option, although I advise you to get one!

You can also view the training progress as well as live output images by running ```python3 -m visdom``` in another terminal and opening [http://localhost:8097/](http://localhost:8097/) in your favourite web browser. This should generate training loss progress as shown below (default params, horse2zebra dataset):

![Generator loss](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/loss_G.png)
![Discriminator loss](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/loss_D.png)
![Generator GAN loss](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/loss_G_GAN.png)
![Generator identity loss](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/loss_G_identity.png)
![Generator cycle loss](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/loss_G_cycle.png)

## Testing
```
./test --dataroot datasets/<dataset_name>/ --cuda
```
This command will take the images under the *dataroot/test* directory, run them through the generators and save the output under the *output/A* and *output/B* directories. As with train, some parameters like the weights to load, can be tweaked, see ```./test --help``` for more information.

Examples of the generated outputs (default params, horse2zebra dataset):

![Real horse](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/real_A.jpg)
![Fake zebra](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/fake_B.png)
![Real zebra](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/real_B.jpg)
![Fake horse](https://github.com/ai-tor/PyTorch-CycleGAN/raw/master/output/fake_A.png)

## License
This project is licensed under the GPL v3 License - see the [LICENSE.md](LICENSE.md) file for details

## Acknowledgments
Code is basically a cleaner and less obscured implementation of [pytorch-CycleGAN-and-pix2pix](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix). All credit goes to the authors of [CycleGAN](https://arxiv.org/abs/1703.10593), Zhu, Jun-Yan and Park, Taesung and Isola, Phillip and Efros, Alexei A.