本模型主要提升FLUX底模的亚洲人像质量，该lora能有效提升人像清晰度，以及真实感，降低脸部皮肤油光，降低AI感，有趣的是，它并不会固定一个人物摸样，每一次出图都是不同的人，不用担心侵权的风险。本模型使用素材均为影楼写真集，素材质量高，出图效果好。V2.0重要说明！！本lora在WEBUI生图，lora权重推荐0.3-0.4！！本lora在COMFYUI生图，lora权重推荐1-1.2！！生成步数25-35步CFG建议3-5建议生成尺寸：1024*1536然后自己发挥提示词吧！

MMPortrait

本模型主要提升FLUX底模的亚洲人像质量，该lora能有效提升人像清晰度，以及真实感，降低脸部皮肤油光，降低AI感，有趣的是，它并不会固定一个人物摸样，每一次出图都是不同的人，不用担心侵权的风险。本模型使用素材均为影楼写真集，素材质量高，出图效果好。

V2.0重要说明！！

本lora在WEBUI生图，lora权重推荐0.3-0.4！！

本lora在COMFYUI生图，lora权重推荐1-1.2！！



生成步数25-35步

CFG建议3-5

建议生成尺寸：1024*1536

然后自己发挥提示词吧！

MMPortrait 女性肖像_v2.0.safetensors

V2.0

MMPortrait Portrait of Women

This model mainly improves the quality of Asian portraits in FLUX base models. The lora can effectively enhance the clarity and realism of portraits, reduce facial skin oiliness, and decrease AI perception. Interestingly, it does not fix a single portrait, and each image is of a different person, so there is no need to worry about the risk of infringement. The materials used in this model are all photo albums from photography studios, with high material quality and good rendering effects.Important note for V2.0!!This Lora is creating images on WEBUI, with a recommended Lora weight of 0.8-1!!This Lora is in COMFYUI production, Lora weight recommendation is 1-1.2!!Generate steps 25-35CFG recommendation 3-5Suggested size for generation: 1024 * 1536Then use your own prompts!

Files with SHA256: 6e65394cfc21c460075d514fd191960d348178f5cfced58364875ba1bcd253d1