Alright. Tests done. The others either couldn't be done because of the model (not the lora) or I just didn't feel the need for further tests in that area.
Prompts for various things should remain the same as the last one, which can be found in the rentry.
THIS IS A LORA FOR XL MODELS, NOT SD1 MODELS. IF YOU USE AN SD1 MODEL IT WILL NOT WORK. DO NOT TRY IT AND THEN WHINE AT ME, DOWNLOAD AN XL MODEL.
This lora was trained on the Illustrious-xl model specifically,
https://civitai.com/models/795765/illustrious-xl and will work well on other models based off of it or like it, no guarantees it will work well with other XL models. The gens I've been doing have been on the NoobXL model because it looked popular, flexible, and easy enough even an idiot could use it.
https://civitai.com/models/833294?modelVersionId=1022833Under certain conditions the lora efficacy will degrade, this is typically because you're prompting things that I couldn't find proper training images for, so it struggles to insert Lumi into images like that. This is unavoidable.
The gens will be slow while using this lora. That is expected.
https://litter.catbox.moe/jevrnz.safetensorsThat link will only last three days due to size. Download it now or rely on others to repost it for you, or for me to be around to post it again, which is unlikely.
Dress Lumis for that one anon. Only three though.
