This repository contains the main classes used in the thesis project. The custom augmentors developed in the project can be found in this repository, as well as the code used to partially automate the model selection process. An example of how to use the code can be found on the RGS_CNN.py script. The code is far from optimized, and code related to RamLess could be faulty. Training without loading to RAM was not necessary, but the code was created in case.
Note that this repository does not hold the necessary files to run anything.