Our training set includes 100 images for each tag, with examples of the returned labels for new images passed to Computer Vision shown in Figures 13, 14, and 15 below.