A class that implements kfold for pytorch for a standard model and an adversarial debiasing model.
adv_main.py contains plotting and is the most complete showing of the kfold and fairness metric analysis.
Adapted from Zhang et al 2018 paper, 'Mitigating unwanted bias with hadversarial networks'
Packages: Python 3.8, aif360 (if an error tensorflow and fairlearn may be needed), pytorch, scipy, numpy, pandas