Entelechia - Yaomu Zhilan character outfit LoRA for WAI v17 / Illustrious-SDXL. Use `entelechia_yaomu_zhilan` at weight 0.6; recommended range is 0.55-0.75. Designed to recall the character identity, short dark-brown hair, pointed ears, black beret, cobalt-blue jacket and white lace  outfit.

Recommended Usage

Default weight:

0.6

Useful range:

0.55-0.75

Use a lower value around 0.55 when pose or background flexibility is more important. Use around 0.7 only when stronger identity or accessories are needed. Avoid 0.9 and above as a default because it can increase automatic camera recall, red hat-edge artifacts and waist/leg color blocks.

Basic Prompt

<lora:entelechia_yaomu_zhilan_recommended_e000009:0.6>, entelechia_yaomu_zhilan, entelechia (arknights), 1girl, solo, no text, looking at viewer, upper body, simple background

Full Character Prompt

<lora:entelechia_yaomu_zhilan_recommended_e000009:0.6>, masterpiece, best quality, amazing quality, very aesthetic, absurdres, entelechia_yaomu_zhilan, entelechia (arknights), 1girl, solo, red eyes, pointed ears, pale skin, short dark brown hair, asymmetrical hair, long side bangs, hair over one eye, black beret with blue stripe, black choker, white bowtie, sleeveless white lace , translucent star-pattern fabric, oversized cobalt blue  shoulder jacket, black inner lapel, gold ornaments, chain belt, cross earrings, black thighhighs, lace-trimmed thighhighs, glossy black stockings, black lace-up high heel boots, blue toe caps, no text

Negative Prompt

low quality, worst quality, lowres, bad , bad hands, bad fingers, extra fingers, missing fingers, extra limbs, disfigured, deformed, blurry, jpeg artifacts, text, watermark, logo, signature, arknights logo, chinese text, english text, caption, artist name, bad face, bad eyes, multiple girls, 2girls, cropped head, out of frame, wrong hair color, missing pointed ears, missing beret, wrong outfit, missing blue jacket, missing white , missing black thighhighs

Training Summary

Training images: 100 PNG
Caption files: 100 TXT
SFW samples: 100
 samples: 0
Base model: -17
Training type: SDXL LoRA, U-Net only
Network dim / alpha: 32 / 16
Network dropout: 0.05
Optimizer: AdamW8bit
Precision: bf16
Recommended checkpoint: epoch 9
Recommended file: entelechia_yaomu_zhilan_recommended_e000009.safetensors
Recommended weight: 0.6
Final training loss: 0.0551