Tools for training explainable models using attribution priors.
Code submitted as the final revision of the manuscript "Improving performance of deep learning models with axiomatic attribution priors and expected gradients" at Nature Machine Intelligence.