Table of Contents 1  Load libraries2  Load previously trained model3  Generate predictions4  Find bounding boxes5  Cut the images based on bounding box6  Test set predictions