Assignment 1 - Gradient Function error : only size-1 arrays can be converted to Python scalars

Course Q&A Natural Language Processing NLP with Classification and Vector Spaces
1

In Assignment 1, for the following Gradient function, the test fails.

(Solution code removed by staff as posting it is against the honour code this community)

When the test is run I get the following error message :

     36     ### END CODE HERE ###
---> 37     J = np.float(J)
     38     return J, theta

TypeError: only size-1 arrays can be converted to Python scalars

Checking the shape of J, i see that its (1,10) in each iteration. Is that the issue?

2

Please check your implementation of sigmoid.
z and h should have the same shape.

1 Like
3

I am getting the same error for J. I checked the shape of z and h, and those are the same. The code I wrote for J is [code removed - moderator]. Is my code wrong?

4

@CaoCao Your code for updating J is correct.

Please check the following:

  1. z.shape == h.shape == y.shape == (10, 1)
  2. theta is updated correctly.

If you are still running into issues, please click my name and message your notebook as an attachment.

5

Oh, I found my mistake, I calculated the sigmoid of x instead of z, and now it works! Thank you!