Utils.py  provides useful utility functions for building pytorch classifiers and dealing with data.

amicable_pertubations.py Includes tools to create and analyze amicable perturbations.

The .ipynb files contains an example of how to make amicable pertubations on the adult income data set and sufficient information (cost functions etc.) to create amicable perubations for other data sets. 

All datasets can be found in the data_folder

All model may be found in models_folder