This is an over the mouth gag for SDXL. It has been trained on over 130 selected images, some realistic, some from graphic novels and a few from comics (images are AI made mostly). To use, add 'otm gag'. I guess that by adding 'realistic' and\or 'photograph' it will probably tend towards realism. Using 'graphics novel illustration' or 'comic' should help you use the LoRA in the respective settings.I will try it and update the description (it's trained on civitai).Feel free to use and share.

Files with SHA256: a325238bcab126951b730aee078619acfbd569884e35fba5938cb0ceb9f586a7