Create beauty-oriented portraits with a MOSS-Image SFT text-to-image graph. The connected Beauty LoRA lets you explore makeup, skin appearance, and portrait styling from prompts.
Main features:
- MOSS-Image SFT INT8 ConvRot with MOSS-Image-Beauty-LoRA-ComfyUI.safetensors at strength 1, Qwen3.5 9B BF16 text encoder, and FLUX.2 VAE.
- Active Skip Layer Guidance 22 node, 1024 x 1024 latent canvas, and one SaveImage output.
- Euler sampler at 30 steps, CFG 3, simple scheduler, fixed seed 42.
Suggested workflow:
1. Install the named SFT model, Beauty LoRA, text encoder, and VAE; load the JSON in ComfyUI.
2. Edit the positive and negative prompts to describe the portrait, lighting, and makeup you want.
3. Generate and review facial detail, skin texture, and composition before making prompt or strength adjustments.
This is text-to-image generation, not restoration of an uploaded face. The supplied positive prompt does not contain a moss_beauty_style trigger; the file is left as supplied. The paired cover is promotional artwork, not a measured output of this JSON.
RunningHub AI workflow: https://www.runninghub.ai/zh-cn/post/2114399590594588674?inviteCode=rh-v1111
RunningHub signup: https://www.runninghub.ai/?inviteCode=rh-v1111
Bilibili video: https://www.bilibili.com/video/BV1UYnU6qEGK/
Related model files: https://pan.quark.cn/s/9ec095b95838