While training on SDXL (or PDXL) takes more time, it seems to have higher chance (compared to sd1.5) to succeed at making something useful with very limited dataset. So I tried with a light novel waifu who has only a few official illustrations.

DarkMel (Black Summoner)

Made with artificial dataset created with initial version.<code>darkmel</code> for character, <code>darkmel-armor</code> for the default outfit. Also, <code>headgear, hairband, halo, wings</code> can be activated or not. Works well with <code>0.6-0.8</code> strength, I recommend using <code>0.8</code> for txt2img and <code>0.6</code> for upscaling the most.Some details (shape of hairband, feathers on pauldron, halo being black) might require higher strength to get consistently.

Files with SHA256: e760937dae8d64b3fb74801e0107d2d21c8d80c55a1c1800c53037e4be4fb3c6