As a continuation of the previous episode in this one I cover the topic about compressing deep learning models and explain another simple yet fantastic approach that can lead to much smaller models that still perform as good as the original one.