I am getting 80/100 due to the error in the dropout regularization test using keep_prob of 0.8. The error says “Not all tests were passed for backward_propagation_with_dropout.”
The Jupyter notebook passed this one.
Since the exact failed test case is not clear from the autograder response, I am not sure how to debug it. Any hint would be deeply appreciated. Thank you.
It might be helpful to look at the tests file of the assignment for backward_propagation_with_dropout function. It might intuitively help you to see which tests you might be failing.
The tests file should be in the File->Open folder.
When you pass the tests in the notebook, but fail the grader, that usually means that your code is not general in some way. The things to look for would be hard-coding the shape of the inputs or referencing a global variable directly that happens to be passed as a parameter, instead of referencing it through the formal parameter.
@Vinay_Nair1 did you ever manage to figure out the error? I have the same issue. I looked at the test files myself, and couldn’t spot anything there that helped my debugging efforts.
My backward_propagation also fails in the notebook, and I can’t for the life of me see why that is. It should be a real simple implementation, as far as I’m concerned.
EDIT: Nevermind, I found my mistake. A silly typo. 
Hi @Scatterbrain ,
Great to hear you have spotted the error.