I submitted a notebook where all test pass when running them cell by cell but fails in the calcification phase:
[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.
I think that it might be because I fixed the backward_propagation_n() function as suggested at the end of the assignment. Can anyone check this?
Thanks
No-one but the course staff can see your notebooks, so we can’t debug this for you. It is unfortunate that the grader can’t give any useful feedback in this type of case. Are you sure that you checked the “expected values” on all the test cases in the notebook.
Thanks for your response. I could upload It here but I think It is against the “Honor Code”… Anyway this is the test that I think is failing when it shouldn’t:
Is there a way to inform about this kind of errors?
That looks correct. You can tell from the grader that you only missed one of the graded functions, so it’s probably something other than gradient_check_n.
What did you get as the difference value in the univariate case? Here’s what I get:
Your backward propagation works perfectly fine! difference = 2.919335883291695e-10
Note that they don’t give you an expected correct value in that section, but only tell you it should be < 10^{-7}.
