InvokeAI
InvokeAI copied to clipboard
invoke-ai
[enhancement]: [Controlnet] - IpAdapter Face - Multiple images input
Is there an existing issue for this?
- [X] I have searched the existing issues
Contact Details
No response
What should this feature add?
Multiple images input would help the IpAdapter Face controlnet detect and generate the face more precisely. This is available in A1111. I hope Invoke would compile this feature into controlnet section soon!
Alternatives
No response
Additional Content
No response
@RyanJDick @hipsterusername Isn't this the same as adding multiple IP Adapters with the face model in the linear UI?
@RyanJDick @hipsterusername Isn't this the same as adding multiple IP Adapters with the face model in the linear UI?
Does multi IP-Adapters (one input image for earch Ip-Adapter) work the same as one IP-Adapter (with multi input image).?I'm not a dev person, so I don't know. But in A1111, multi images input with one Ip-Adapter works really well.
Just wanted to add that the IP Adapter node in ComfyUI got updated with new features like setting embeds combine methods that would be only possible in invokeAI with multi-input support in IP Adapter.
InvokeAI already supports:
- Multiple IP-Adapters (in both Linear UI and Workflows).
- Multiple images to an IP-Adapter (Workflows only).
These two features are not equivalent.
@suonnon It sounds like you may already be able to achieve what you want by using "multiple images to an IP-Adapter" in the Workflows UI. Let me know if that is sufficient, or if there is still something missing.
@StellarBeing25 We do not currently support different combine_embeds options - we always use concat when multiple images are passed to an IP-Adapter. Is there a particular workflow that you are trying to implement with this feature? If yes, it's probably worth creating a new Github issue to track that separately.
