DiT.1 and XL versions are available. Their respective preview artworks (if any) and trigger words (if differs) can be inspected separately via selecting the versions at the top, below the title (on desktop browser, need to open a full separate page of this LoRA first to see them)
LoRA model was trained primarily on the source material of Robo Fizz from concept art appearance image specifically, prior to his official debut in S1E2 on December 9, 2020 (it differs slightly in color palette and other few details), as I find this image to be more distinctive and generally just more appealing. I may try to implement different sets of trigger words in the future that would enable changing of the palette from "classic yellow" one to "release green (lime)" one, and vice versa (or just create additional version of LoRA), but I doubt that I'll be really able come around to implementing this effectively enough
Generates most consistently on its corresponding base models, "Hoshino" (XL) and "Tsubaki" (DiT.1). On other models, it may have severe artifacts and/or discrepancies from the oiginal appearance image, unfortunately
