This is my testing model as I work out how to best train sdXL, so it was trained on a much smaller dataset than the 1.5 sd version for testing speed so that being said this model will be pretty limited with characters and pretty much anything that isn't a solo simple pose.