# Standardize data to have feature values between 0 and 1. train_x = train_x_flatten/255. test_x = test_x_flatten/255.