Tsukinogi "Dancing" outfit LoRA for WAI v17 / Illustrious-SDXL. Use tsukinogi_wu with tsukinogi (arknights) and 1girl; recommended default weight is 0.80, useful range 0.7-0.9.

This LoRA focuses on Tsukinogi's dancing-style outfit from Arknights: red plaid strapless dress, white fur stole, long black hair, yellow eyes, deer antlers, animal ears, black gloves, asymmetrical legwear, single thighhigh, high heels, necklace, thigh strap, and Arknights-style oripathy details.

Recommended Usage
<lora:tsukinogi_wu_strict_imagegen_recommended_e000008:0.80>, tsukinogi_wu, tsukinogi (arknights), 1girl, solo, no text, looking at viewer

Character / Outfit Prompt
<lora:tsukinogi_wu_strict_imagegen_recommended_e000008:0.80>, tsukinogi_wu, tsukinogi (arknights), 1girl, solo, long hair, black hair, yellow eyes, deer antlers, animal ears, red dress, plaid dress, strapless dress, white fur stole, black gloves, asymmetrical legwear, single thighhigh, high heels, very long hair, horns, dress, sleeveless dress, coat, pencil dress, thighhighs, black thighhighs, brown thighhighs, gloves, necklace, deer ears, deer girl, antlers, black footwear, jewelry, sleeveless, bare shoulders, anklet, oripathy lesion (arknights), thigh strap, thighlet, feather boa, no text, no logo, no watermark

Negative Prompt
low quality, worst quality, lowres, blurry, bad anatomy, bad hands, bad face, bad proportions, extra limbs, extra fingers, missing fingers, deformed hands, disfigured, duplicate, multiple girls, 2girls, cropped head, cropped body, out of frame, messy background, bad composition, text, watermark, logo, signature, username, artist name, speech bubble, caption, wrong outfit, alternate outfit, costume change, missing antlers, missing animal ears

Recommended Generation Settings
Checkpoint: WAI-Nsfw-Illustrious-17
Resolution: 1024 x 1024
Sampler: DPM++ 2M
Scheduler: Karras
Steps: 30
CFG scale: 6.5
Default LoRA weight: 0.80
Motion-priority LoRA weight: 0.70-0.75

Training Summary
Training images: 112 PNG
Caption files: 112 TXT
SFW samples: 112
NSFW samples: 0
Base model: WAI-Nsfw-Illustrious-17
Training type: SDXL LoRA, U-Net only
Network dim / alpha: 32 / 16
Network dropout: 0.05
Optimizer: AdamW8bit
Precision: bf16
Recommended checkpoint: epoch 8

Known Limitations
Higher weights can improve outfit recall but may reduce pose flexibility. For difficult full-body action prompts, use 0.70-0.75 and keep the key outfit tags. Other SDXL model families may change face shape, outfit color, or accessory details.

Rights / Usage Note
Unofficial fan-made LoRA for a character from Arknights. Follow the source IP holder fan-content policy, Civitai platform rules, and the license or policy of any checkpoint/model you combine with it.