Version 2 : trained with base sd 1.5 - for general porpose /anime

Version 1.2 : trained and tested with epicrealism v5 - works best for realistic results

Needs higher resolutions for better faces.