Q: More than one image to train avatar?
Hi, I’d like to ask about improving avatar and character consistency when generating from images.
Is it currently possible, or planned for the future, to use multiple reference images when creating a character? For example, instead of relying on a single image, could users upload several views such as front, 3/4, and side, or multiple variations of the same face to improve likeness and consistency?
I’ve tried using a single head-to-knees image as recommended, but the resulting character still looks noticeably different. Having the ability to provide additional reference images seems like it could help produce more accurate results.
Could you let me know if this is supported now (how?), or if it might be added in the future? I've had zero luck getting anyone reply to my emails in the past, so I'm asking here.