Question about C4_W2A1 coursework

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

Hello!
I worked through the problem using the coursera platform and there is no issue with my code. However, when I downloaded and run the same code on my laptop offline, that doesn’t work any more. What is the problem there? How to solve this?
Here is the error I get for exercise one: AssertionError Traceback (most recent call last)
/var/folders/d2/3mf1ph3s2q9bny72q5vytdcm0000gn/T/ipykernel_72017/3018596845.py in
26 print(np.around(A4.numpy()[:,(0,-1),:,:].mean(axis = 3), 5))
27
—> 28 public_tests.identity_block_test(identity_block)

~/Library/CloudStorage/OneDrive-UniversityCollegeLondon/Documents/Extra curricular/Deep learning/Assignment file/3/W2A1/public_tests.py in identity_block_test(target)
35 assert resume[1, 1, 0] - np.floor(resume[1, 1, 0]) > 0.7, “Looks like the BatchNormalization units are not working”
36
—> 37 assert np.allclose(resume,
38 np.array([[[ 0., 0., 0., 0., ],
39 [ 0., 0., 0., 0., ]],

AssertionError: Wrong values with training=False

2

This could be caused by “versionitis”: all the packages in this space evolve pretty quickly over time and there is no guarantee that they do so in backward compatible ways in all cases. The courses were last updated in April of 2021 and in most of the assignments use the current versions as of that point in time. It’s not a simple matter, but there are ways to use tools like Anaconda (conda) to reproduce the exact versions in a different environment. There are no official instructions for that, but here’s a thread from a fellow student who gives some very useful guidance about how to start down that road.

Of course the first point is that you need all the files, not just the notebook, but I’m guessing you would not have gotten as far as you show without already having done that.

3

Thank you very much! I got it now.