1/3
PROMPT RECORD图像记录
模糊、低分辨率、变形、丑陋、解剖结构错误、多余肢体、水印、文字、曝光过度、曝光不足、卡通、绘画、3D 渲染瑕疵 获得更好结果的建议: 模型:使用写实风格的基础模型,例如 Realistic V...
中文说明
模糊、低分辨率、变形、丑陋、解剖结构错误、多余肢体、水印、文字、曝光过度、曝光不足、卡通、绘画、3D 渲染瑕疵 获得更好结果的建议: 模型:使用写实风格的基础模型,例如 Realistic Vision、EpicRealism,或在可获取的情况下使用专门的 "Kalinka Fox" 微调/LoRA(她在角色扮演/AI 社区中很受欢迎)。 设置:CFG 5-7,步数 30-50,分辨率 512x768 或 768x1024(竖版),采样器使用 DPM++ 2M Karras 或 Euler a。 变体:添加 --ar 2:3(Midjourney)或用于全身的宽高比。对服装细节进行加权,例如 (glossy black latex catsuit:1.3)。 如需精确复刻,请加入 "in the style o...
原始 Prompt
blurry, lowres, deformed, ugly, bad anatomy, extra limbs, watermark, text, overexposed, underexposed, cartoon, painting, 3d render artifacts
Tips for better results:
Model: Use a realistic base like Realistic Vision, EpicRealism, or a specific "Kalinka Fox" fine-tune/LoRA if you have access (she's popular in cosplay/AI communities).
Settings: CFG 5-7, steps 30-50, resolution 512x768 or 768x1024 (portrait), sampler like DPM++ 2M Karras or Euler a.
Variations: Add --ar 2:3 (Midjourney) or aspect ratio for full-body. Weight the outfit details like (glossy black latex catsuit:1.3).
For exact replication, include "in the style of KalinkaFox" or reference her cosplay photos. This should get you very close — the shiny black outfit, hair, lighting, and expression are the key elements.