Hints
DL-C05-C-001
Use (batch, input_features) x (input_features, hidden_units).
DL-C05-C-002
The output side of W2 gives the number of scores.
DL-C05-C-003
Count layers with learned parameters. Do not count the input size.
DL-C05-C-004
ReLU maps negative values to zero.
DL-C05-C-005
Regression with one output gives one number per example.
DL-C05-C-006
Classification gives one score per class for each example.