Author This model trained on all the images I've found on internet. He has very interesting style. Lora trained with auto objecet detection, which failed, obviously, so it is more of style concept. For SD 1.5.
https://www.google.com/search?q=fomenko+math