Problem converting probabilities A[0,i] to actual predictions p[0,i] Logistic Regression with a Neural Network Mindset programming assignment

Course Q&A Deep Learning Specialization Neural Networks and Deep Learning
1

What’s wrong with this code?

{moderator edit: code removed}

It gives following errors when trying to compile next cell

ValueError Traceback (most recent call last)
in
2 b = -0.3
3 X = np.array([[1., -1.1, -3.2],[1.2, 2., 0.1]])
----> 4 print ("predictions = " + str(predict(w, b, X)))
5
6 predict_test(predict)

in predict(w, b, X)
23 # YOUR CODE STARTS HERE
24
—> 25 A = sigmoid(np.dot(w.T,x) + b)
26 # YOUR CODE ENDS HERE
27

<array_function internals> in dot(*args, **kwargs)

ValueError: shapes (1,2) and (3,) not aligned: 2 (dim 1) != 3 (dim 0)

2

Python is case sensitive. x is different from X. Check your sigmoid implementation in predict function.

Best,
Saif.

1 Like
3

The error here is in the calculation of sigmoid, the np.dot(w.T, X) their shapes are not as suppesed to be. For eg. If w.T is MxN then X needs be (any size) x M, the column dimension of the second matrix needs be same as the row of the first matrix in the dot product.

4

Thanks a lot. You saved my day.!!!

5

@Kalyana_Chakravarthy, in the future, please do not post your code on the forum. That’s not allowed by the course code of conduct.

If a mentor needs to see your code, we’ll ask you to send it via a private message.

6

Posting the error messages and assert outputs is almost always sufficient for the mentors to understand the mistake.

7

@TMosh can you let me know what to do if I get stuck up with some error in my code at some point?. Is there a way to contact a mentor directly without posting the error/problem, kindly let me know. By far the mentors have been doing a great job troubleshooting.

8

You can post your error messages, and any asserts.
If a mentor needs to see your code, they’ll ask you to send it via a private message.

Mentors are not supposed to encourage sharing code on the forums. It violates the code of conduct.