works with most styles - weight could be 0.6- 0.9 - trained for 768 (can have better results)
any humanoid prompts may apply a face
TW - Plank, - location/placement. camera angle prompts, seem to work
smile, blue eyes, blue lips, red circle, still life, wood, no humans, - these may help or hinder results,
if you can figure ways to properly prompt then your ahead of the game - there are some examples below - V1
Description
FAQ
Comments (2)
this may be the greatest lora of all time.
the diverse array of emotions, and subtle body language shifts that this lora can achieve are magnitudes above the filth that infests our museums.
Wow. I didn't think you could improve a masterpiece, but I was wrong! Bravo, V2 is excellent.