⚡ Update: the recommended weight is now pre-baked into the file — just drop it in and use it at weight 1.0 (no need to set 0.8). Lower only if stacking many LoRAs.

Zani (Wuthering Waves) LoRA

Trigger word: zni

Trained base: waiIllustriousSDXL_v170 (Illustrious XL family)

Network: dim=32, alpha=16

Optimizer: Prodigy (lr=1.0)

Recommended settings

• Weight: 0.8 (strong identity + costume fidelity). Drop to 0.6 if you want the outfit to stay more reliably closed.

• - Hi-res fix: recommended (~1.7x, denoise ~0.45).

• - IMPORTANT — keep her clothed: the dataset leans NSFW, so at higher weight the shirt tends to open. To keep her canonical closed-shirt look, add these to the NEGATIVE prompt: open shirt, cleavage, underboob, exposed navel, midriff.

Usage

1girl, solo, zni, mature female, grey hair, multicolored hair, red eyes, black curled horns, demon tail,
white shirt, collared shirt, red necktie, black jacket, black pants, key earrings, black choker,
<pose here>, <background here>,
masterpiece, best quality, amazing quality, absurdres, newest
```

Recommended weight **0.8**. Trigger `zni`.

**-- Identity (baked into the trigger) --**
grey / silver hair with red inner streaks (multicolored), red eyes, **black curled horns**, an **arrow-tipped demon tail**, key-shaped earrings, a mole. She reads as a mature woman.

**-- Trademark outfit (add for the canonical look) --**
white collared shirt + **red necktie** + black jacket + **black high-waist pants** (gold buttons) + black gloves + black choker.

**-- Style note (IMPORTANT) --**
For clean anime shading (not 3D / semi-real skin), put anti-realistic terms in the NEGATIVE prompt: `realistic, 3d, photorealistic, shiny skin, oiled body, semi-realistic, render`.

**-- NSFW --**
Undresses cleanly; the horns / tail / red eyes persist through explicit scenes.

## Tested checkpoints

- ✅ `waiIllustriousSDXL_v170` (trained base)
- 🔍 Other Illustrious XL fine-tunes (untested; start weight 0.6-0.8)
- ❌ Pony / SDXL base (architecture mismatch, untested)

## Caveats

- Dataset is NSFW-leaning — use the anti-exposure negatives above for clothed / SFW results.
- On checkpoints other than the training base, the recommended weight shifts (try 0.6-0.9).