You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
is it possible to compute some kind of uncertainty about the answer given for a comprehension? Like for example if you query with an ambiguous expression or you query for classes not present on the image?
The text was updated successfully, but these errors were encountered:
Probably not with current contrastive ranking loss. You could try replacing it with binary cross-entropy, which could output a [0, 1) score. I would imagine an ambiguous expression would return all boxes with lower than 0.5
Hi,
is it possible to compute some kind of uncertainty about the answer given for a comprehension? Like for example if you query with an ambiguous expression or you query for classes not present on the image?
The text was updated successfully, but these errors were encountered: