Week 2 Assignment Exercise 7 - predict

[DarrenH]

coursera-platform - W2 Assignemt: Logistic_Regression_with_a_Neural_Network_mindset

In exercise 7, the correnct Y_prediction output is [1 0 1], while, no matter what, I always got

A = [0.52241776 0.50960677 0.34597965]

also, in this case, Y_prediction should be [1 1 0], but I got [1, 0, 0] instead, which is very weird…

my code is very simple, A = sigmoid(np.dot(wT, X)+b), as sigmoid function was defined above in Exercise 3 (and All test passed is shownd); the threshold of Y_prediction of 0.5 is set. I got stuck for a while, please help!

[TMosh]

Just for starters:
Does the code you added to sigmoid() or predict() ever use the len() function?

[TMosh]

Note: There are some other tests that are performed in the background when you run the predict_test(). It’s possible your code is failing one of those hidden tests.

[paulinpaloalto]

You must be referencing global variables instead of the actual parameters that are passed to the function. Notice that in the test that fails, your prediction has 3 elements, but the expected value has 6. So you’re not using the correct input X value.

[DarrenH]

Thanks guys, it turns out that I accidentally added a Tab in front of the line “return Y_prediction” of predict function. As I removed the tab space, it shows all tests passed.

Key takeaway: space and Tab space matters a lot in python

the previous version (that made the bug) is shown as below:

[TMosh]

Whitespace in python is critical.

All tabs should be expanded to spaces. Most editors will do that automatically - unless you paste in some code that contained an embedded space.

[DarrenH]

exactly