ELLA & Compel

[andupotorac]

I'm curious if Compel is compatible with algorithms for better prompt adherence like Ella, and if not, what do you guys suggest we do to make this stuff work? Obviously, we'd like people to set weights for words as they need, but also maintain the ability for their prompts to be as close to the image output as possible.