The official PyTorch implementation of the paper On Feature Normalization and Data Augmentation.
*: Equal Contribution
This repo contains the PyTorch implementation of Moment Exchange (MoEx), described in the paper On Feature Normalization and Data Augmentation. For ImageNet and CIFAR experiments, we select Positional Normalization (PONO) as the feature normalization method.
Please follow the instructions in the README.md
in each subfolder to run experiments with MoEx on CIFAR, ImageNet, and ModelNet10/40.
More information and relevant applications will be updated.
If you find this repo useful, please cite:
@article{li2020feature,
title={On Feature Normalization and Data Augmentation},
author={Li, Boyi and Wu, Felix and Lim, Ser-Nam and Belongie, Serge and Weinberger, Kilian Q},
journal={arXiv preprint arXiv:2002.11102},
year={2020}
}
@inproceedings{li2019positional,
title={Positional Normalization},
author={Li, Boyi and Wu, Felix and Weinberger, Kilian Q and Belongie, Serge},
booktitle={Advances in Neural Information Processing Systems},
pages={1620--1632},
year={2019}
}