This repository provides an unofficial PyTorch reimplementation of RandAugment, a data augmentation technique originally developed by Google for training deep learning models. It enables inserting augmentation transforms into standard image classification pipelines by specifying two hyperparameters (N and M). The library supports training on CIFAR-10/100, SVHN, and ImageNet datasets, and reproduces the competitive performance of the original RandAugment paper when applied to Wide-ResNet and other CNN architectures.