Furries! Version 1 produces images closer to Dagashi's art. Version 2 is more varied
This is trained exclusively using art by Dagashi - expect extremely NSFW results. You may be able to get SFW results with a well worded prompt.
For Best results use the model "ToraFurryMix - v1.0" - The model has 2 different versions, which result in different results you'll have to experiment with. I've tested dozens of furry models against my LoRA and this one by far gives the best "dagashi" results.
(and yes, the cover image is generated by this LORA, one of the first images it ever produced)