The Illustrious version of my Chroma lora

Trained on 297 images I manually processed (and stole from the internet respectfully. I mean I did it by hand so that's a little better right?). It's extremely flexible and can generate a broad range of sfw and nsfw material. This lora isn't as good as the Chroma one, but that's probably the 6b parameter difference between Chroma and SDXL.

And it works with CivitAI's on-site generator!!!

Please leave a like and a comment if you enjoy my models!

v3.5

• Use it with WAI-illustrious-SDXL: https://civarchive.com/models/827184/wai-illustrious-sdxl

  • Other Illustrious versions work too. But this is my favourite one.

• N works again!

• Sould be better at everything overall

  • Less character bleed

  • More variety

  • Better at 3D

• But the artstyle is less consistent and jumps wildly.

• In comparission to the Chroma version, these optimal settings aren't required to get a usable output!

v3.5: ComfyUI + 1024p + "DPM++ 2M SDE" + Beta_57 Scheduler + CFG 6.5 + 30 steps

v2

• Use it with WAI-illustrious-SDXL: https://civarchive.com/models/827184/wai-illustrious-sdxl

  • Other Illustrious versions work too. But this is my favourite one.

• The Chroma version had N in it too, but Illustrious merged him with Uzi and V. You can still try to generate him with SD-N though.

• In comparission to the Chroma version, these optimal settings aren't required to get a usable output!

v2: ComfyUI + 1024p + "DPM++ 2M SDE" + Beta_57 Scheduler + CFG 6.5 + 30 steps

Prompting:

• Works best if you start the prompt with the character tag: SD-V, SD-J, Uzidoorman, Cyn

• If a character isn't generated correctly, type more describing traits: SD-V -> SD-V with short white hair, white skin

  • There can be mixups with a too short prompt.

  • Illustrious loves to give the drones human skin. Fix that by typing: white skin

• The Lora likes to forget the tail. Luckily you can remind it by typing: tail

• Natural language prompts work best, but can be combined with tag based prompting

• Sometimes the negative prompt can mess things up. Keep it minimal!

Bias:

• The Lora is heavily biased towards generating V, J and Uzi. N and Cyn need a few more describing words.

• Most APIs I used to get images are on the more nsfw side of things so... write "wearing clothes" to hide that bias.