This Lora was trained on IllustriousXL.
It's not intended to be used with Pony, Animagine, SD1.5, etc.

I've been playing around with Illustrious for the past week or so and have had a pretty good time with it. One thing I couldn't solve was that the check was very bad with backgrounds.
The aim of this Tweaker/Enhancer/whatever is to improve these.
It is by no means perfect and some artist tags are very strong and just force a very blurry background.
You can look here for a direct comparison with the lora and without it for my showcase images.

Feedback is very much appreciated.
For basic prompting refer to the model description of the checkpoints in suggested resources.