This is a LoRA model I created quite a while ago...I believe it was around last summer. I realized I had not posted it on this site, so I'm posting it now albeit late. The usage is the same as version 1.5. Using it normally will flatten the image, and setting the strength to a negative value will enhance the image details.