Previous

Next

Train and test the neural network

Completed

Check your knowledge

1.

Why do we need the GradientTape during training?

We need it to record the difference between training and test data.

We need it to calculate derivatives for the gradient descent algorithm.

We need it to help find the right mini-batch size.

You must answer all questions before checking your work.

Continue