Style LoRA trained on Rinka Uzuki's AI works.
Base model: Anima Base v1.0
Style Trigger: rinka_uzuki_style
Every caption used for training includes a frame tag — one of portrait, upper body, cowboy shot, feet out of frame, foot out of frame, or head out of frame — plus an additional close-up tag where applicable.
It's also recommended to specify the number of people, as this is present in every training caption (e.g. 1girl).
The LoRA is trained on strictly verified Danbooru-style tags, but also works well with natural language prompts.
Recommended weight: 0.9–1.0