Week 1 Gradient Checking programming assignment

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

Hello, I keep getting 60/100 for the gradient checking exercise despite passing all the test cases. I am not sure what’s wrong. Lab id is gwujncfe

[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.

This is the only feedback I get but the notebook doesn’t reflect the error.

Please let me know what can be done. Thank you!

2

Hi @Harika_Pendli,

The notebook tests are different than the autogravder’s, which is why you are passing the test in the assignment, but failing the grader. One reason could be that you are using a hardcoded value in the assignment, which is why you are passing the test in the notebook, but the autograder test uses a different value for that parameter.

I also understand that this feedback, Tests failed on 1 cell(s)! These tests could be hidden. Please check your submission. is not really convenient.

I’m going to direct message you to discuss this further.

Best,
Mubsi

3

Hello @Mubsi thank you for letting me know. But I don’t think I have used any hard coded bits of code. I have replied to you in direct messages.

4

Hi @Harika_Pendli,

The exercise that is failing the autograder for you is Ex 4. In your Ex 4 you have been explicitly mentioning the data type float. But the expected data type for the grader for that particular exercise is numpy.float64.

You can either, change your mentioning of float to np.float64. Or you can leave out the mentioning of float (it automatically then ends up being np.float.64).

Fix that and you should be able to pass.

Best,
Mubsi

5

Thank you @Mubsi. I understand.