-
Notifications
You must be signed in to change notification settings - Fork 2.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat(ui): FLUX in linear UI #6846
Conversation
The node user interface is not quite self-explanatory for basic users, it would be great if FLUX models could be used like all the other models via the basic ‘text to image’ user interface. Is this a big challenge to adapt Invoke AI to make this possible or can we expect such a functionality to be available in the future? |
That is what this PR is doing 👍 |
When switching to Flux from SDXL, bounding box can get put into a state that is whack/breaks - It's also strange to have the bounding box change sizes when canvas mode is selected. |
… encoder, CLIP embed
… layer, IP adapters, regional control
…own as invalid if its already selected
14f319e
to
371ee35
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks great!
I changed one thing - drop the default steps from 50 to 30. Ain't nobody got time for that.
Summary
Adds support in linear UI for using FLUX as a main model for text-to-image, image-to-image, inpainting, and outpainting.
Does NOT include: LoRA support, controlnet support, IP adapter support, or upscaling. This means the only available layers on the canvas included in this PR are inpainting and raster layers. Support for the rest will be in subsequent PRs as the back-end is built out.
Related Issues / Discussions
QA Instructions
Options for FLUX generation settings:
Merge Plan
Checklist