C4_W1 UNQ_C10 1 test failed

Course Q&A Natural Language Processing NLP with Attention Models
1

here the test result:

image
image1008Ă—490 22.9 KB

my lab ID is cvwueigs

Thanks!

3

Hi @skyrockets_21,

In your Ex 10, when you are finding the max score key, you have to pass in the key argument as well.

Best,
Mubsi

4

Oh I just know there is a key arguments in max()! Thanks so much @Mubsi !

5

@Mubsi
I have a similar issue; here are my scores and output. It seems that the similarity scores are off by only -4.1e-06 between two different phrases. Could this be an error with my implementation of the rouge1_similarity function (even though it passes the unit tests) or potentially a hardware issue?

{0: 0.8571387719251243, 1: 0.8571387719251243, 2: 0.7619111193931957, 3: 0.8571428571428571}
['Ich iss Suppe.', 'Ich iss Suppe.', 'Ich esse Schweine.', 'Ich esse Suppe.']
Ich esse Suppe.
Expected output does not match
1 Like
6

I have the same issue

7

Have you solved it yet?

9

I had the same issue. To resolve this, instead of using the prepared “weighted_avg_overlap” function, use the built function in UNQ_C9, which is the “average_overlap” function. Hope it helps.

1 Like