Hi @Rvi, and welcome! The layer_sizes function, a function of the data in X and Y, contains three lines of code for you to fill in. The n_h = ... line is the only one which is not inferred from the function arguments (X and Y)–it can’t be. That is where the “hard-coding” must take place. The other two, n_x and n_y, are inferred from the arguments using numpy’s shape attribute. If you are unfamiliar with using that, there is an explanatory hint in Exercise 1. Hope this helps!