C2_W3_Assignment :: Grading: :: Exercise 7 nn_model :: Failed test case: "default_check". Wrong weights matrix W1

I see these grade errors in Exercise 7 nn_model.

Any helps ? it seems a “precision” problem, but I am not able to discover where is my error.

Failed test case: “default_check”. Wrong weights matrix W1…
[[ 2.22641773 -1.95781602]
[ 1.92521557 -1.74008809]],
but got:
[[ 2.53492212 -2.06302136]
[ 2.06885671 -1.96083822]].

Failed test case: “default_check”. Wrong bias vector b1…
[[ 6.31408636]
but got:
[[ 6.29704669]
[-4.8483427 ]].

Failed test case: “default_check”. Wrong weights matrix W2…
[[ 7.12818935 -7.24882361]],
but got:
[[ 7.06228665 -7.20205483]].

Failed test case: “default_check”. Wrong bias vector b2…
but got:

1 Like

Please check your private messages for instructions.


I get the same issue, any updates?

1 Like

This assignment, its utility .py files, and the grader have all been updated in the last two weeks.

Be sure you’re using the current version of this assignment.

Instructions for updating your copy of the assignment are in the “M4ML Resources” area of the discussion forum, in the ‘FAQ’ thread.

1 Like

I’ve done updating the assigment, but the issue still persist



1 Like

I’m getting the same error you are. I will investigate further.

1 Like

I’ve reported the issue to DLAI staff.

1 Like

I have this problem too and error is same. Is there a new solution to this problem?

There is no update, but staff are investigating it.

In compute_cost function i change:
np.dot() to np.multiply()

In sigmoid function i change :
np.power() to np.exp()

It’s work for me, maybe can help you too

@FAIQ_HIDAYAT_DZAKWAN, please do not share your code on the forum. That is not allowed by the Code of Conduct.

I have edited your reply.

Oh i see, i apologize and thank you for telling me.

Thanks for your suggestions.

Note, these statements are both true, but I don’t quite understand why np.power() is any different than np.exp(). Mathematically they should be the same, but the grader certainly seems to prefer np.exp().