Week3 Assignment : AssertionError: Wrong values for A2

Course Q&A Deep Learning Specialization Neural Networks and Deep Learning
1

I am getting error in “# GRADED FUNCTION:forward_propagation”
I wrote lines for getting W1,b1,W2,b2 and for calculating Z1,A1,Z2,A2. I got the following error.

AssertionError: Wrong values for A2

***Expected output***

A2 = [[0.21292656 0.21274673 0.21295976]]

A2 from my code is [[-0.86360952 -0.86388209 -0.86355916]]

Thanks for any clues.

2

Are you sure you applied the sigmoid activation function at the output layer? The A2 values should all be between 0 and 1, right? But your values are all negative. So how could that happen?

3

I used tanh as activation function for generation A1, A2.
Generation of W1,b1,W2, b2 looks fine as all tests are passed for that cell.

4

Just to make sure we’re on the same page here: you have fixed your code to use tanh for the hidden layer and sigmoid for the output layer, right? That’s what the instructions tell you to do.

5

Thanks for reply. It was my mistake. I used tanh for both A1, A2. Now I change A2 formula and it is working.