Open
Description
Hi @Yuanshi9815 ,
By default the OminiModel does:
adapter_names = [None, None, "default"]
In every example script the first two slots remain None, so I’m unclear what their intended role is.
- What are the intended roles of each of the three slots?
- Can all three be activated at inference instead of only the "default" slot?
I trained only the "default" adapter for spatial alignment, but varying the text prompt didn’t affect the image style so I wonder if using the other two slots would help.
Any pointers would be much appreciated, thanks for the great work!
Metadata
Metadata
Assignees
Labels
No labels