EDIT: So it turns out the BLIP captions are completely unusable as a base ("a cartoon ostrich with a broom and a hat" being my favourite), so I'm having to write them from scratch (my years as a e6 tagger are finally paying off), but between work and passing out on the sofa I've not made much progress, so V2.0 is going to take a little while.
A LoRA to generate images of the Yinglet species by Valsalia.
Use strength 0.7 for the best results, lower values will produce more rodent-like features and higher will cause artefacts. I also recommend adding other species, namely canine, to the negative prompt or you'll get things like fleshy noses.
Check the "About this version" section for longer descriptions and specifics.
Please post any images you feel like sharing created with the LoRA so I can learn what works. Also any tips on training LoRA's will be greatly appeciated.