Skip to content

Latest commit

 

History

History
46 lines (38 loc) · 1.72 KB

README.md

File metadata and controls

46 lines (38 loc) · 1.72 KB

sagnet

Style-Agnostic Networks (SagNets)

By Hyeonseob Nam, HyunJae Lee, Jongchan Park, Wonjun Yoon, and Donggeun Yoo.

Lunit, Inc.

Introduction

This repository contains a pytorch implementation of Style-Agnostic Networks (SagNets) for Domain Generalization. It is also an extension of our method which won the first place in Semi-Supervised Domain Adaptation of Visual Domain Adaptation (VisDA)-2019 Challenge. Details are described in Reducing Domain Gap by Reducing Style Bias, CVPR 2021 (Oral).

Citation

If you use this code in your research, please cite:

@inproceedings{nam2021reducing,
  title={Reducing Domain Gap by Reducing Style Bias},
  author={Nam, Hyeonseob and Lee, HyunJae and Park, Jongchan and Yoon, Wonjun and Yoo, Donggeun},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2021}
}

Prerequisites

Setup

Download PACS dataset into ./dataset/pacs

images ->  ./dataset/pacs/images/kfold/art_painting/dog/pic_001.jpg, ...
splits ->  ./dataset/pacs/splits/art_painting_train_kfold.txt, ...

Usage

Multi-Source Domain Generalization

python train.py --sources Rest --targets [domain] --method sagnet --sagnet --batch-size 32 -g [gpus]

Single-Source Domain Generalization

python train.py --sources [domain] --targets Rest --method sagnet --sagnet --batch-size 96 -g [gpus]

Results are saved into ./checkpoint