Help with Week 1 Programming assignment: initialize Z

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

I am working on the conv_forward function where we are supposed to initialize the output volume Z with 0s.

I tried initializing Z with np.zeros and using the dimensions of (m, n_H, n_W, n_C). Yielded an Assertion Error of “Wrong shape. Don’t hard code the pad and stride values in the function”

I am just confused about where the dimensions come from for Z.

I’m sorry if this has been answered already. I am finding Discourse not very easy to search, especially since most of the threads consist of “I had this problem, no wait I fixed it”.

Thanks for any help you can give me. I’ll go back to my code now and hopefully have an epiphany.

2

np.zeros() requires that you pass it the dimensions as a tuple.
So in python syntax, you have to enclose the dimensions inside nested pairs of ((…)).

So this works:
Z = np.zeros((m, n_H, n_W, n_C))

1 Like
3

The best way to search the Forum for answers is to include the course, week, and assignment (i.e. “C4 W1 A1”), or to search by the function name.

1 Like
4

Thank you for your help. It turns out that I was doing the initialize Z with np.zeros correctly, but when calculating n_H and n_W I had forgotten to divide by stride.