Something wrong with programming assignment week 2

In the update_parameters_with_adam I do not understand what is the adam parameter ‘t’ does.

In the ‘mini-batch grabient descent’ vadio ‘t’ was used as a pointer for each batch. Also in ‘Adam Optimization Algorithm’ vadio ‘t’ was used as a pointer for each update for ‘W’ and ‘b’ (which in my understanding is a pointer for each batch).

So in update_parameters_with_adam function in the assignment when we calculate the bias-correction ‘v_corrected’, shouldn’t we just divide ‘l’ instead of global adam parameter ‘t’ (which is the correct answer)?

Is that means even we update ‘W’ and ‘d’ within each batch, it only counted as one update untill we go over all batches?

The value t is not a “pointer”: it is the iteration number in both cases you mention. And l (ell) is the layer number, which is a completely different thing, right? So the question is (as always): what do the formulas say? They are given in the instructions and your job is to translate those into python code.