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
I have trained the vision tower module by replacing another vision encoder and gathering new custom data from another domain. But I wonder how I can evaluate pretraining performance since it is the crucial part for image-text alignment, it must be consider? Anyone who can response, thanks!
The text was updated successfully, but these errors were encountered:
enkaranfiles
changed the title
[Question] How ton evaluate pretraining performance?
[Question] How to evaluate pretraining[image-text alignment] performance?
Apr 29, 2024
Question
I have trained the vision tower module by replacing another vision encoder and gathering new custom data from another domain. But I wonder how I can evaluate pretraining performance since it is the crucial part for image-text alignment, it must be consider? Anyone who can response, thanks!
The text was updated successfully, but these errors were encountered: