Scale-Bender — Tsunade · AoRA v2
Base model: AnimaYume v4
CHECK CHANGELOG >>>>
What makes this different
This LoRA was trained using my AoRA character workflow.
AoRA is not a new model architecture; it is my dataset preparation and training methodology for character LoRAs.
Most character LoRAs are trained on vibes. This one was trained on structure.
AoRA (Anthropometric-oriented Rank Adaptation) is a proprietary dataset curation method developed for the Scale-Bender project (My project). It operates at the level of physical form — not style, not color. What it produces is a LoRA where the character's identity is load-bearing at a geometric level, which means it survives things that break conventional LoRAs: style transfers, dynamic poses, outfit swaps, lighting extremes.
Tsunade stay Tsunade.
Try it and share your results in the comments — every generation helps improve the next version. What style did you pair it with? What broke? What surprised you? All feedback feeds the next AoRA iteration.
Tag helpfull :
TRIGGER: tsunade_aora, 1girl,
character : mature woman, blonde hair, long hair, forehead diamond, brown eyes, confident expression, sharp eyes, adult woman,
clothe : official outfit, green haori, grey kimono top, dark sash, cropped pants, sandals,
This model not permits users to:
Sell images they generate
Use the model without crediting the creator
This model permits users to:
Run on other service
Run on Civitai
Description
v1 — Aora_tsunade_v1 (current) Trigger compressed to a single token: tsunade_aora. All structural encoding is baked directly into the weights — no tag list at inference. Renamed from BoRA to AoRA to reflect the anthropometric focus of the curation method. Cleaner prompt integration, easier to stack.