There’s an error in the mean in the conv_forward() function. I have checked some previous threads on this exercise and I’ve tried matching my process with theirs but I can’t find anything different or wrong in my logic. Can anyone help?
I added some print statements in my code and here’s what I see from the first test case for conv_forward:
stride 2 pad 1
New dimensions = 3 by 4
Shape Z = (2, 3, 4, 8)
Shape A_prev_pad = (2, 7, 9, 4)
Z[0,0,0,0] = -2.651123629553914
Z[1,2,3,7] = 0.4427056509973153
Z's mean =
0.5511276474566768
Z[0,2,1] =
[-2.17796037 8.07171329 -0.5772704 3.36286738 4.48113645 -2.89198428
10.99288867 3.03171932]
cache_conv[0][1][2][3] =
[-1.1191154 1.9560789 -0.3264995 -1.34267579]
First Test: All tests passed!
Interestingly your Z[0,2,1] values and your cache_conv values agree, but the Z mean is wrong there as well.
The most common mistakes here are in handling the stride. Note that it is 2 in this test case and that the stride applies in the input space (A_prev), not in the output space (Z), right?
1 Like
I did an absolutely foolish mistake. I didn’t take the loop for w(horizontal start and end) inside the loop for h(vertical).
Thanks for replying
It’s great news that you were able to find the problem under your own power. Lots more interesting material ahead. Onward! 
1 Like