Description

Shamare - Xiuyunhe character outfit LoRA for WAI v17 / Illustrious-SDXL. Use `shamare_xiuyunhe` at weight 0.8; recommended range is 0.7-0.9. Designed to recall Shamare identity, rose-pink short twintails, hair bun, fox ears, pink fox tail, white-and-blue hanfu outfit, black pleated skirt, grey socks, navy Mary Janes and round fan.

Recommended Usage

Default weight:

0.8

Useful range:

0.7-0.9

Use a lower value around 0.7 when pose or background flexibility is more important. Use around 0.9 only when stronger identity, fox ears/tail or the round fan are needed. Avoid 1.0 and above as a default because it can over-force the fan, red-string ornaments, tail placement and pale studio-background habits.

Basic Prompt

<lora:shamare_xiuyunhe_recommended_e000008:0.8>, shamare_xiuyunhe, shamare (arknights), 1girl, solo, no text, looking at viewer, upper body, simple background

Full Character Prompt

<lora:shamare_xiuyunhe_recommended_e000008:0.8>, masterpiece, best quality, amazing quality, very aesthetic, absurdres, shamare_xiuyunhe, shamare (arknights), 1girl, solo, red eyes, pale skin, rose pink hair, short twintails, hair bun, white flower hair ornament, blue hair ribbon, fox ears, pink fox tail, chinese clothes, hanfu, white dress, blue trim, long sleeves, black pleated skirt, grey socks, navy blue mary janes, holding round fan, 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 fox ears, missing fox tail, wrong outfit, missing hanfu, missing white dress, missing blue trim, missing black skirt, missing grey socks, missing mary janes, missing round fan

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 8
Recommended file: shamare_xiuyunhe_recommended_e000008.safetensors
Recommended weight: 0.8
Final training loss: 0.0629