In this repo, I have implemented Fmix, Cutmix and Mixup on the Fashion MNIST dataset using PreAct ResNet18. This is an attempt to reprdouce the results from this paper: https://arxiv.org/abs/2002.12047
Link to my medium blog: https://medium.com/@virajbagal12/mixed-sample-data-augmentation-721e65093fcf
python train.py --msda baseline --save_dir path_to_save
