W1 final lab: how to build the layers of the lab?

Hello all!

In C2W1, with the final lab (Programming Assignment: binary classification) I had some questions:

What is the logic behind assigning 3 layers (25 units, 15 units, and 1 unit each) to this exercise? When building my own neuronal networks how do i know how many layers and units I’d need to use?

Hi @Mamey ,

This kind of the questions are every NN architect would be asking when they are given a particular problem.
In this course, prof. Ng is showing you an example what the structure looks like. He designed the NN to have that number of layers and units per layer. Prof. Ng has been working in this field for many years, so, for the time being, just hold your thought and try to understand the concepts.

Tyvm for your answer,

Will this course explain the reasoning of this in the next weeks (w2, w3 and w4)?
if not, is there any resources you guys would recommend for understanding this?

Hi @Mamey ,

This post from stackoverflow should give you some idea on how to approach the design of a NN topology.