Pytorch Deep Learning Introduction

This is kind of introduction to Pytorch. Before one run this program he should download the dataset to PetImages/, split to Cat & Dog directories, and set the REBUILD_DATA flag to True. Alternatively, one can split the data to

project  
└───train
│   └───cats
│   └───dogs
│   
└───test
    └───cats
    └───dogs
	

like the Pytorch's data loader expects, and set the REBUILD_DATA to false. One can follow easily the main function and following the stages:

1. building data if needed
2. create Pytorch Dataloaders
3. train
4. save model to models directory. If the current model has higher accuracy from the one in the models/best_model directory then replace the older one.