DLS Course 2 week 1 assignment 1 EX2

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
1

I am not passing all test cases. I have initialized parameters[W + str(l)] to a random number and multiplied it by 10, with dimensions of its current layer and the previous one, (i, i-1), respectively. And b, I am initializing to zero.

Here is the output:

image
image1250×686 40.7 KB

2

Hi @Angel_Ortiz ,

I will try to assist you here. My first question: did you by any chance change the random.seed? or is it still equal to 3?

The seed should look like this (as given by the exercise):

np.random.seed(3)

Thanks,

Juan

3

I just figured my error was that I was using random.rand() instead of random.randn() :sweat_smile:

Thank you for the help anyways

4

Yes, that was going to be my guess based on the clue that all your values are positive, instead of being a mix of positive and negative values. Glad you figured it out under your own power!