Doubt in the First FF Network Python Code

Hello Ishvinder,

In Deep Learning in PadhAI platform on the chapter Python: Scalar Backpropagation - why different dw and db are divided by m while updating the weights? It was not normally done earlier.

Hi @vkgarg,
m here denotes the batch size, i.e number of training examples in a batch.

Hi @Ishvinder, Thank you for your reply. I understand it that m is the batch size. But why this is considered only in this example to divide the accumulated gradients?