C4 w1 assignment 1 exercise 3

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

Hello
could someone please explain to me how the weight shape changes from (3,3,4,8) to (3,3,4)
thanks

2

It may be difficult to point out where you are referring with a short sentenceā€¦ But, I will try to guess.

As you wrote, the shape of W, as an input to conv_forward(), is (3,3,4,8). This means,

  • We have 8 filters, i.e, output channel is 8, which is n_C in this assignment.
  • Each filter has the shape of (3,3,4) which is 3x3 filter with a depth (channel) = 4. So, the depth of target image is 4, which is n_C_prev in this assignment.

Hope this is what you are looking for, and helps.

1 Like