This Lora is not very precise, not sure if because of bad prompting or i just removed too many tags on training.

Also, it seems to prefer close-up shots, use upper body or cowboy shot tags.