C4W2 Resnets - Why batchnormalization axis = 3

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

Hi everyone,

In the first assignment of the second week (Resnets), I don’t quite understand why we set BatchNormalization(axis = 3). Not sure if I missed that in the videos, but I would imagine that the number of axis would be related to the number of channels of the previous volume? That is, the output of the Conv2D, which if I am not mistaken depends on the number of filters (provided by F1, F2, and F3).

Thanks,

Eva

2

This post has the best explanation, so instead of repeating it, I advice you to read all of it yourself:

1 Like
3

Great, thanks a lot!

4

This one is also great:

1 Like