# TuiGAN-PyTorch

**Repository Path**: Ruanjiyang/TuiGAN-PyTorch

## Basic Information

- **Project Name**: TuiGAN-PyTorch
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-02-01
- **Last Updated**: 2021-02-01

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# TuiGAN-PyTorch
Official PyTorch Implementation of "[TuiGAN: Learning Versatile Image-to-Image Translation with Two Unpaired Images](https://arxiv.org/abs/2004.04634)" (ECCV 2020 Spotlight)

## TuiGAN's applications
TuiGAN can be use for various computer vision tasks ranging from image style transfer to object transformation and appearance transformation:
 ![](imgs/examples.jpg)

## Usage

### Install dependencies

```
python -m pip install -r requirements.txt
```

Our code was tested with python 3.6  and PyToch 1.0.0 or 1.2.0

###  Train
To train TuiGAN model on two unpaired images, put the first training image under `datas/task_name/trainA` and the second training image under `datas/task_name/trainB`, and run

```
python train.py --input_name <input_name> --root <datas/task_name>
```
For example, 
```
python train.py --input_name apple --root datas/apple
```
##  Comparison Results

###  General Unsupervised Image-to-Image Translation
![](imgs/comparisons.jpg)
###  Image Style Transfer
![](imgs/style.jpg)
###  Animal Face Translation
![](imgs/dog.jpg)
###  Painting-to-Image Translation
![](imgs/trees.jpg)

##  More Results

###  Art Style Transfer 
<img src="imgs/artstyle.jpg" width="50%" />

###  Photorealistic Style Transfer 
![](imgs/photostyle.jpg)

###  Animal Face Translation
<img src="imgs/animalface.jpg" width="50%" />