This repository contains a python package of neural network interpretability and explainablility methods, focusing on the latent space, that can be easily integrated into a keras training routine using a callback to compute and capture outputs of these methods during training. - View it on GitHub