https://github.com/znxlwm/ugatit-pytorch

Official PyTorch implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation
https://github.com/znxlwm/ugatit-pytorch

Last synced: 5 months ago
JSON representation

Official PyTorch implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation

Host: GitHub
URL: https://github.com/znxlwm/ugatit-pytorch
Owner: znxlwm
License: mit
Created: 2019-07-29T07:44:56.000Z (about 6 years ago)
Default Branch: master
Last Pushed: 2023-03-16T02:38:05.000Z (over 2 years ago)
Last Synced: 2025-04-14T16:57:59.886Z (6 months ago)
Language: Python
Size: 4.21 MB
Stars: 2,537
Watchers: 62
Forks: 474
Open Issues: 51
Metadata Files:
- Readme: README.md
- License: LICENSE

Awesome Lists containing this project

README

## U-GAT-IT — Official PyTorch Implementation
### : Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation

### [Paper](https://arxiv.org/abs/1907.10830) | [Official Tensorflow code](https://github.com/taki0112/UGATIT)
The results of the paper came from the **Tensorflow code**

> **U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation**

>
> **Abstract** *We propose a novel method for unsupervised image-to-image translation, which incorporates a new attention module and a new learnable normalization function in an end-to-end manner. The attention module guides our model to focus on more important regions distinguishing between source and target domains based on the attention map obtained by the auxiliary classifier. Unlike previous attention-based methods which cannot handle the geometric changes between domains, our model can translate both images requiring holistic changes and images requiring large shape changes. Moreover, our new AdaLIN (Adaptive Layer-Instance Normalization) function helps our attention-guided model to flexibly control the amount of change in shape and texture by learned parameters depending on datasets. Experimental results show the superiority of the proposed method compared to the existing state-of-the-art models with a fixed network architecture and hyper-parameters.*

## Usage
```
├── dataset
   └── YOUR_DATASET_NAME
   ├── trainA
├── xxx.jpg (name, format doesn't matter)
├── yyy.png
└── ...
   ├── trainB
├── zzz.jpg
├── www.png
└── ...
   ├── testA
├── aaa.jpg
├── bbb.png
└── ...
   └── testB
├── ccc.jpg
├── ddd.png
└── ...
```