Awesome

The PyTorch implementation of our work on domain generalization:

S. Zhao, M. Gong, T. Liu, H. Fu and D. Tao. Domain Generalization via Entropy Regularization. NeurIPS 2020.

Environment

• Python 3.6
• PyTorch 1.1.0

Run

• Download the PACS dataset (https://drive.google.com/drive/folders/0B6x7gtvErXgfUU1WcGY5SzdwZVk), and prepare the dataset as follows:

  |- dataset
      |- PACS
          |- art_painting
              |- *.jpg
          |- cartoon
              |- *.jpg
          |- sketch
              |- *.jpg
          |- photo
              |- *.jpg
  

• Execute the following command to run (replace xxx with any one of the four datasets, such as art_painting):

  python train.py --target xxx
  

Citation

@article{zhao2020domain,
  title={Domain Generalization via Entropy Regularization},
  author={Zhao, Shanshan and Gong, Mingming and Liu, Tongliang and Fu, Huan and Tao, Dacheng},
  journal={Advances in Neural Information Processing Systems},
  volume={33},
  year={2020}
}

Contact

Shanshan Zhao: szha4333 AT uni dot sydney dot edu dot au / sshan.zhao00 AT gmail dot com