-
Notifications
You must be signed in to change notification settings - Fork 42
Open
Description
Thank you for sharing this interesting work! I'm having trouble figuring out which modules are being trained even after looking at the VISTA code. Is it correct that only the MLP is trained, and the vision encoder and LLM are frozen, as depicted in Figure 1 of the paper?
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels