===============================================================
This code has the source code for the paper
"멀티 블록 기반 Random Erasing in the Frequency Domain".
Previous literature is "Improving Corruption Robustness with Random Erasing in the Frequency Domain".
Requirements for Pytorch （see Pytorch installation instructions）
You should download 'cifar10-c', 'cifar100-c', 'fmnist-c' dataset if you want to evaluate corruption error
ResNet-20 baseline on CIFAR10：
python cifar.py --dataset cifar10 --arch resnet --depth 20
ResNet-20 + Random Erasing on CIFAR10：
python cifar.py --dataset cifar10 --arch resnet --depth 20 --p 0.5
ResNet-20 + REF on CIFAR10：
python cifar.py --dataset cifar10 --arch resnet --depth 20 --p 0.0 --p2 0.5
ResNet-20 baseline on CIFAR100：
python cifar.py --dataset cifar100 --arch resnet --depth 20
ResNet-20 + Random Erasing on CIFAR100：
python cifar.py --dataset cifar100 --arch resnet --depth 20 --p 0.5
ResNet-20 + REF on CIFAR100：
python cifar.py --dataset cifar100 --arch resnet --depth 20 --p 0.0 --p2 0.5
ResNet-20 baseline on Fashion-MNIST：
python fashionmnist.py --dataset fashionmnist --arch resnet --depth 20
ResNet-20 + Random Erasing on Fashion-MNIST：
python fashionmnist.py --dataset fashionmnist --arch resnet --depth 20 --p 0.5
For ResNet：
--arch resnet --depth (20， 32， 44， 56， 110)
For WRN：
--arch wrn --depth 28 --widen-factor 10
===============================================================
Download CIFAR10-C dataset
https://paperswithcode.com/dataset/cifar-10c
Download CIFAR100-C dataset
https://zenodo.org/records/3555552
Clean Error and Corruption Error(Mean)
CIFAR10
python cifar.py --resume [model_best.pth.tar path] --evaluate
CIFAR100
python cifar.py --resume [model_best.pth.tar path] --evaluate --dataset cifar100
===============================================================