LAD

Lipschitz Anomaly Discriminator

Contains all of the code needed to run comparisons of the LAD method against existing baselines present in the paper

https://arxiv.org/abs/1905.10710

uses make_runs.py to compile lists of models with parameters to test and uses snakemake to manage jobs.

Snakefile2 is for running AUROC comparisons, and Snakefile3 is for the black image robustness tests.