This project is about a general template for pytorch. The goal is to make a scalable deep learning architecture, so that the user can add their models/optimizers/datasets only by adding new files and keey the others intact. To run the code with different settings, the user can call the file with different commandline arguments. It is suggested that the user keep the commandline in a shell script (a file with .sh extension).
python 3, pytorch 1.0, torchvision 0.3