C4,W1,A2 E1 BatchNormalization for axis 3

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

in happyModel exercise 1
BatchNormalization for axis 3
here axis 3 refer to channel?
if yes then are we trying to normalize all the channels of input?

2

Hello @rasgulla.c,

Yes! Axis 3 refers to the channel axis. If you have 3 channels by the time it reaches the BatchNorm layer, it means that the BatchNorm layer remembers a mean and a variance for each of the 3 channels. So you are going to normalize all the pixels along the 3 channels respectively.

Cheers,
Raymond

3

so i’m passing axis = 2
im not passing the test.

4

and why do you use axis = 2?

5

let say input is 32,32,3 then axis = 2 is 3 right???

6

The axis 0 is reserved for meaning the number of samples. So your 32, 32, 3 are in axis 1, 2, and 3 respectively.