Data_Augmentations_Pytorch

This repository intergrated various Data Augmentation methods. Our implementation is based on these repositories:

• PyTorch Cifar Models
• Cutout
• RandomErasing
• Mixup
• CutMix
• RICAP
• LA (LocalAugment)

Dataset

• CIFAR10, CIFAR100

Model

• ResNet

Training Start

Requirements

• Python3
• PyTorch (> 1.0)
• torchvision (> 0.2)
• NumPy

Training Command

• Hyperparamters for each augmentation method are fixed to same values on each original paper
• ex) dataset : cifar100, augmentation : cutout, model: resnet110, index of the number of trainings: 1

python3 ./train.py \
--type cifar100 \
--model resnet \
--depth 110 \
--tn 1 \
--augtype cutout \
--length 8