Hello,
Lets say I have picture 5x5 pixels, RGB, and my data m=10
Then the dimension in Python for this data set should be: (10, 3, 5, 5)
Where:
m = 10, number of examples
c = 3, number of channels
h, w = 5, the height and width of a picture
However, for Week 01 - Exercise 03, we define it in a “strange” way:
Where A_prev.shape = (m, h, w, c).
A = np.ones((10, 3, 5, 5))
B = np.ones((10, 5, 5, 3))
Compare the code A and B, where A follows the dimension as I propose, and B follows the dimension in the exercise. The output A match the expectation. while B is not
I believe that dimension should follow A, not B as in the exercise. However, I am not sure if I misunderstand anything here
Thank you
Phan