Please look at the version name (above the photos) to make sure you are downloading the correct version

For n00bs the Flux LoRA won't work with SDXL / Pony, and vice versa

This was trained off of images from the book After Man by Dougal Dixon

It works well with the following prompts:

• text, captions, english text

• 2koma , 4koma for layouts

• rabbuck will very cleanly reproduce the Rabbucks from the books

I would recommend starting with the strength set to 0.4 and cranking it up from there.

Update - Version 2.0

This version is much better at generalizing subjects and will cleanly render backgrounds from prompts.

This was trained on Pony Diffusion V6 XL instead of Austism Mix using more photos and much better prompting. V1 relied too heavily on the auto generated WD14 tags, V2 relies primarily on manual tagging.

This will also render Rabbucks quite well (from the book) using the prompt rabbuck

Version 1.0

This was trained on 20 images using AutismMix Confetti as a base. This is a LyCoris/Glora using Prodigy optimizer with cosine LR Scheduler, running for 3200 steps.