Next, we will compile our Keras model. Compilation basically refers to the manner in which your neural network will learn. It lets you have hands-on control of implementing the learning process, which is done by using the compile method that's called on our model object. The method takes at least three arguments:
model.compile(optimizer='resprop', #'sgd'
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
Here, we describe the following functions:
- A loss function: This simply measures our performance on the training data, compared to the true output labels. Due to this, the loss function can be used as an indication of our model's errors. As we saw earlier, this metric is actually a function that determines how far our model's predictions are from the actual labels of...
