This is a lora to help control emptions and expressions. The data set for this was completely redone and had over 150 images of individual, single and up to four people. No specific trigger words were used. You should be able to describe each person and what emotion they have as well as the specific mouth, eyebrow shapes they are making.
See the example images for prompting.
Strength closer to 1 makes the quality a little overdone and blown out. So maybe a bit less.
Shout out to @antarek for prodding me to actually get this done, I have been lazily working on it for a while.


