DLS 2 week 3 exercise 6 : Wrong total cost

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
1

I have used the line:

{Moderator Edit: Solution Code Removed}

output : tf.Tensor(0.17102128, shape=(), dtype=float32)
expected output: tf.Tensor(0.810287, shape=(), dtype=float32)

the shapes of labels and logits are the same (6,2) and (6,2). What is wrong with the code?

2

It is given that:
logits – … of shape (6, num_examples) and labels – … same shape as Z3

Also given that
“It’s important to note that the “y_pred” and “y_true” inputs of tf.keras.losses.categorical_crossentropy are expected to be of shape (number of examples, num_classes).”

The bold parts are the hint for you to solve the error. One more hint: Transpose.

PS: I am deleting your code as posting your code is not allowed…

1 Like