Does this actually do anything? I notice that no one uses this feature but it’s there in text-to-image-plugin.
I’ve set up a little test, and… it generates images. Here. Unclear if they are affected by the referenceImage I’ve given it though. Perhaps there is a particular kind of test I could try to see it in action? Or if you could let me know what I’m doing wrong?
Been studying it further. Seems it is sent all the way through to the generator api. The reference image is posted. But… it’s not used? Maybe?
Would love to explore the capabilities of this if the feature does/can work. Any advice, @[email protected]?
Yeah it is actually (IIRC) one line of code to enable it on the server, but I paused that because I was weary of the backwards compat required given how fast things were changing in that area (controlnets, ipadapter, plain img2img, etc. with the possibility of the next generation of models not supporting one or more of those) so I didn’t want to commit to anything just yet. But at this point things have definitely settled enough that I could enable it quite soon, probably along with a few other things like character reference, style reference, etc. Probably makes sense to do this along with the next big model update (to Flux), which is hopefully within the next month or so. Flux is really cool, and will make the current outputs of the text-to-image-plugin look pretty primitive in comparison (especially for complex prompts, and prompts with text in the image)
Oh cool, I’ll look forward to that! 👍