Characteristics of this model本模型特点：

• While preserving rich details, it reduces skin texture imperfections, resulting in a more natural and clean complexion.

  Enhance realism and increase the texture and clarity of the image.

  Enhance NSFW Prompt compliance capabilities and reduce the probability of random body fluid residue artifacts appearing on the body in NSFW scenes.

  Improve compatibility and performance with external LoRA.While preserving rich details, it reduces skin texture imperfections, resulting in a more natural and clean complexion.

• Enhance realism and increase the texture and clarity of the image.

• Enhance NSFW Prompt compliance capabilities and reduce the probability of random body fluid residue artifacts appearing on the body in NSFW scenes.

• Improve compatibility and performance with external LoRA.

• 在保留丰富细节的同时，减少皮肤纹理瑕疵，让肤质更加自然干净。

• 提升真实感，增加图片的质感和通透感。

• 增强NSFW Prompt 遵循能力，降低NSFW 场景中身体随机出现体液残留伪影的概率。

• 提升对外部LoRA 的兼容性与表现。

  Tip: Use natural language; either pure Chinese or a mix of Chinese and English is fine.Tip: Use natural language; either pure Chinese or a mix of Chinese and English is fine.

  小技巧：使用自然语言，全中文和中英文混合也行。

  Recommended Settings推荐设置：

• 8-15 sampling steps

• CFG 1.0

• er_sde /Euler sampler

• sgm_uniform /Simple scheduler

• 1024 × 1024 default resolution

• Local deployment requires：Qwen3-VL text encoder and Qwen Image VAE loaders