• Stars
    star
    2,508
  • Rank 18,295 (Top 0.4 %)
  • Language
    Python
  • License
    MIT License
  • Created over 5 years ago
  • Updated over 1 year ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

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

U-GAT-IT β€” Official PyTorch Implementation

: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation

Paper | Official Tensorflow code

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

β”œβ”€β”€ dataset
Β Β  └── YOUR_DATASET_NAME
Β Β      β”œβ”€β”€ trainA
 Β  Β  Β  Β  Β  β”œβ”€β”€ xxx.jpg (name, format doesn't matter)
           β”œβ”€β”€ yyy.png
           └── ...
Β Β      β”œβ”€β”€ trainB
           β”œβ”€β”€ zzz.jpg
           β”œβ”€β”€ www.png
           └── ...
Β Β      β”œβ”€β”€ testA
        Β  Β β”œβ”€β”€ aaa.jpg 
           β”œβ”€β”€ bbb.png
           └── ...
Β Β      └── testB
           β”œβ”€β”€ ccc.jpg 
           β”œβ”€β”€ ddd.png
           └── ...

Train

> python main.py --dataset selfie2anime
  • If the memory of gpu is not sufficient, set --light to True

Test

> python main.py --dataset selfie2anime --phase test

Architecture


Results

Ablation study

User study

Comparison

More Repositories

1

pytorch-generative-model-collections

Collection of generative models in Pytorch version.
Python
2,588
star
2

pytorch-MNIST-CelebA-GAN-DCGAN

Pytorch implementation of Generative Adversarial Networks (GAN) and Deep Convolutional Generative Adversarial Networks (DCGAN) for MNIST and CelebA datasets
Python
507
star
3

pytorch-MNIST-CelebA-cGAN-cDCGAN

Pytorch implementation of conditional Generative Adversarial Networks (cGAN) and conditional Deep Convolutional Generative Adversarial Networks (cDCGAN) for MNIST dataset
Python
482
star
4

pytorch-CartoonGAN

Pytorch implementation of CartoonGAN (CVPR 2018)
Python
388
star
5

tensorflow-MNIST-GAN-DCGAN

Tensorflow implementation of Generative Adversarial Networks (GAN) and Deep Convolutional Generative Adversarial Netwokrs for MNIST dataset.
Python
175
star
6

tensorflow-MNIST-cGAN-cDCGAN

Tensorflow implementation of conditional Generative Adversarial Networks (cGAN) and conditional Deep Convolutional Adversarial Networks (cDCGAN) for MANIST dataset.
Python
147
star
7

pytorch-pix2pix

Pytorch implementation of pix2pix for various datasets.
Python
119
star
8

pytorch-Conditional-image-to-image-translation

Pytorch implementation of Conditional image-to-image translation (CVPR 2018)
Python
48
star
9

pytorch-apex-experiment

Simple experiment of Apex (A PyTorch Extension)
Python
47
star
10

pytorch-CycleGAN

Pytorch implementation of CycleGAN.
Python
40
star
11

FUNIT-pytorch

Pytorch implementation of "Few-Shot Unsupervised Image-to-Image Translation" (ICCV 2019)
Python
36
star
12

tensorflow-pix2pix

Tensorflow implementation of pix2pix for various datasets.
Python
6
star