This is a Clip Skip 2 anime-model-trained style Lora based on a dataset of essentially every DragonBall Z hentai art image ever released by (long "retired") artist Pandora's Box. Fully capable of accurately reproducing the style of both their colored images and monochrome graphite images. Use pbxdbzstyle to activate the overall Lora, and use monochrome or greyscale if you want to hint the output away from being in color.

This Lora was trained natively at the highest possible resolution relative to the source material (something like up to 704 x 1024 and 1024 x 704, after sizing to match standard SD 1.5 resolutions).

Additionally, this Lora does benefit quite a lot from "after detailing" of faces, however detailing for hands generally does not do anything particularly positive relative to the intended art style, and so isn't necessarily recommended as much.

Feel free to lower the strength down from 1.0 as needed, depending on how any given checkpoint you use this with responds to it.