Reducing Flipping Errors in Deep Nueral Networks

This repository is a PyTorch implementation for "Reducing Flipping Errors in Deep Nueral Networks".

Requirements

The code was run on

Python 3.6
torch 1.2.0
torchvision 0.4.0

or

Python 3.6
torch 1.4.0
torchvision 0.5.0

Training

To train different networks with FER, run the following command:

sh train.sh

Citation

If you find this code helpful, you may consider citing this paper:

@inproceedings{deng2021fer,
  title={Reducing Flipping Errors in Deep Neural Networks},
  author={Xiang Deng, Yun Xiao, Bo Long, and Zhongfei Zhang},
  booktitle = {Proceedings of the 36th AAAI Conference on Artificial Intelligence},
  year={2022}
}