Hello everyone, this is my first attempt at training a LoRA — would love to hear feedback and see anything you generate with it!

Base Model: Anima 1.0 — Base checkpoint (anima-base-v1.0). Not the Aesthetic or Turbo variants.

Recommended Weight: 0.9

Trigger Word & Locked Appearance

kushinabcgo, long red hair, grey eyes, hairclip, blue bracelet

Default Outfit (optional)

green pinafore dress, sleeveless white collared shirt with zipper

Optional add-on -- if the art style drifts too far from her trained look:

(anime screencap)

Known Quirk: Some NSFW images come out with stockings For bare legs, add:

bare legs

For full-body shots where you want her feet/footwear visible:

toeless footwear

Heads-up: there's a small chance of a stray, fake watermark-style text appearing somewhere in the image. Keeping watermark, text, logo in your negative prompt is a safe habit.

Suggested Generation Settings

• Sampler: Euler A, Simple scheduler

• Steps: 12 / CFG: 2.5 (with Turbo LoRA --raise CFG/steps if not using it)

• LoRA weight: 0.9

Trained on a small, hand-captioned 20-image dataset - my first LoRA, so there's plenty of room for a v2. Comments and generated samples are very welcome!