First, thank you @LittleJelly for your guide, https://civitai.com/articles/19224/guidelocal-training-creating-anime-character-lora-illu-or-noobai. It has been 2 years since I last was active here and creating LoRA's, and those were for SD 1.5. And back then, screen capping was indeed the way for me too!

Making LoRA's is a thing I do for fun. So if you like what I do? You can help me out by just using the LoRA and leaving a review!

This LoRA has been almost exclusively trained on anime screencaps. except 1 full body image from a wiki. If you want more of the anime style use anime screencap as keyword, otherwise ommit it and optionally lower LoRA weigth.

This LoRA does not have a character trigger word, this will provide you with more flexibility. Take a look at the trigger words at the panel to the right, or the images I've provived how to prompt this LoRA. You're free to remove keywords from the provided trigger words, that just gives you flexibility in how you want to promt the LoRA. I've provided the given trigger words as a base set. E.g. if you want to have her wear a different outfit, remove her base outfit keywords.

All WAI-NSFW-illustrious-SDXL sample images used the following setting:

• Use anime screencap as base keyword

• Hires. fix,

  • Upscaler: R-ESRGAN 4x+ Anime6B

  • Steps: 20

  • Denoising: 0.4

  • Scale factor: 1.5

All NoobAI-XL (v-prediction model) sample images used the following setting:

• Don't use anime screencap as base keyword, but rather my own styling I like, see the prompts.

• Hires. fix,

  • Upscaler: R-ESRGAN 4x+ Anime6B

  • Steps: 16

  • Denoising: 0.35/0.4

  • Scale factor: 1.5