So it knows everything about how many
classes i have what type of images i have and all of that that information is effectively used by cnn learner to actually put together a to actually give us a model where we don't even define like how many output neurons do we need to have well why do we need to define we .