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
Yes you are right. We are still trying to figure out if the results difference comes from the dataset. Looks like it has the highest possibility. I was thinking imagenet knn and zeroshot might be somewhat transferrable but might be just too arbitrary.
Hello, did you train clip models on coyo and evaluate them on zero shot classification ?
I'm curious to know what zero shot classification on imagenet numbers you got
The text was updated successfully, but these errors were encountered: