Skip to content

Latest commit

 

History

History
30 lines (19 loc) · 732 Bytes

README.md

File metadata and controls

30 lines (19 loc) · 732 Bytes

Learning to Generate Novel Domains for Domain Generalization

This is an unofficial PyTorch implementation of Learning to Generate Novel Domains for Domain Generalization (ECCV2020).

[arxiv]

Some of the code is borrowed from https://github.com/HAHA-DL/Episodic-DG and https://github.com/yunjey/stargan.

data

Please download the data from https://drive.google.com/drive/folders/1i23DCs4TJ8LQsmBiMxsxo6qZsbhiX0gw?usp=sharing and use the official train/val split.

Domain-Generalization-via-Image-Stylization
├── data
│   ├── Train val splits and h5py files pre-read

train

python main.py --train True

test

python main.py --train False