This is a sample notebook for visual recognition. I used real-world pathology data. Because the data is restricted is has been replaced by a free dataset from Andrew Janowczyk (http://www.andrewjanowczyk.com/deep-learning/). It uses the ResNet50 convolutional network architeture and ImageNet weights.