Does average_precision
work for multilabel classification?
#1182
Unanswered
KevinMusgrave
asked this question in
Classification
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
By multilabel, I mean that a sample can have a
target
vector like[1, 0, 0, 1]
.I noticed that the score doesn't match
sklearn.metrics.average_precision_score
whenaverage="macro"
.Here's my code:
Here's the output:
Edit: It seems like passing in
num_classes=10
gets rid of the discrepancy. Is that the correct usage?Beta Was this translation helpful? Give feedback.
All reactions