Week 1 - Programming Assignment: Gradient Checking

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

I have completed C2-W1’s Gradient Checking assignment, all tests passed, fixed the backward_propagation errors, but the grader gives 80/100. I passed, but a would know why don’t give me 100/100. I try to put a regularized term in forward-propagation and backward-propagation, and didn’t work too.

In the gradient_check_n function I used np.linalg.norm. I check all the cells and re-write the code in a local python notebook.

error_C2W1
error_C2W1993×296 40.5 KB 

The grader feedback:
[ValidateApp | INFO] Validating '/home/jovyan/work/submitted/courseraLearner/W1A3/Gradient_Checking.ipynb'
[ValidateApp | INFO] Executing notebook with kernel: python3
Tests failed on 1 cell(s)! These tests could be hidden. Please check your submission.

Is it a system error or am I getting it wrong elsewhere in the code?

2

The grader test cases are more stringent than the ones in the notebook, but unfortunately the grader test cases do not always give you useful feedback when they reject your solution. FWIW I checked my result for the test cell you show after fixing the introduced fake back prop errors and my difference value is exactly the same as yours. So my guess is that the problem is on one of the other 3 graded cells, but I can only suggest you check the results on those other cells again.

1 Like
3

Thanks @paulinpaloalto, I will double-check all my notebook.

Best regards :+1:

4

Fixed by @JulioLima.

Double check you’re computing difference as: image

Good luck with the rest of the course :slight_smile:

1 Like
5

Thanks @nramon !!

Best regards :+1:

1 Like