Pre-cleanup

Repository was getting too messy with different datasets, model types, etc. I don't want to lose all previous work so I'm tagging it as a safety net given that the repository is purely for research purposes. Please ignore.