# Use the Adam optimizer for learning
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

        __
        __