➼Like what I do? Consider commissioning me at https://www.fiverr.com/metalchromex/make-you-a-stable-diffusion-anime-character-model or throwing me a few dollars at https://ko-fi.com/dandysombraes so I can keep making cool crap for you all to enjoy for free!
Update 7-6-26: I made some tweaks to the V2 set and uploaded a V2.1. V2.1 should work a bit better.
Update 6-30-26: I added a V2 with an improved dataset and tagging. The LoRA should be a bit more flexible while still closely resembling the official source. While it was not trained with these keywords, you can try adding asahina mikuru and suzumiya haruhi no yuuutsu to your prompt to potentially get an even more on-model art style.