[W1] [One Layer of a Convolutional Network] [10:25] Error on lecture notes

[Changjin_Lee]

For the nH[l] (height of layer l), should the padding be p[l-1], not p[l]?

[reinoudbosch]

Hi Changjin_Lee,

The padding is related to the filter that is being used. So it’s the padding for filter[l], i.e. p[l].