MR FLUX KREA

【模型简介】： 此版本为SDXL1.0模型微调版本V0.1：基础模型+LoraV0.2：基础模型+Lora+LoraV0.3：基础模型+Lora+Lora+ RunDiffusionXL/0.25V0.4：基础模型+Lora+Lora+ RunDiffusionXL/0.25+LoraV0.5：修修补补，扩充场景V0.6：调整亚洲人面部细节V0.7：同尺寸扩展画面细节 【生成建议】：与基础模型使用一致，未特别固定面部，出脸随缘，对新手不是很友好【参数建议】：1、步数：一般20步，上不封顶2、采样：E a，D &nbsp;2M S K（个人用的比较多3、提示词强度：1-3（模型自由发挥多一些）；4-8；9++，4、Clip Skip：25、SDXL Refiner 支持（建议基础模型满意后使用6、高勤修复支持（建议基础模型满意后使用7、面部修复个人一般不开8、配合lora：建议使用自行测试，个人一般很少用lora【创意概述】：SDXL基础原型已比较优秀，主要在模型基础上向真实、人像方向优化提示词建议使用SDXL的：照片模式，模拟胶片、电影、霓虹、细节加强模式（可混合使用亚洲面孔：chinese girl，asian girl…… 例图全随机出图，照片模式不固定面部，保持模型的普适性__SDXL/xlstyle__, __SDXL/filmxl__, __clothes_female/Luxauly_outfit 奢华_服装/lux__, __composition_构图__, __品牌/Designer__, girl, 1girl, solo

MR 真实 0.6 面部细节优化 beta.safetensors

v0.7

XL V4

FLUX

FLUX II

FLUX III

KREA

krea III

MR FLUX VV

```html
[Model Introduction]:This version is the fine-tuned version of SDXL1.0 modelV0.1: Basic model+LoraV0.2: Basic model+Lora+LoraV0.3: Basic model+Lora+Lora+ RunDiffusionXL/0.25V0.4: Basic model+Lora+Lora+ RunDiffusionXL/0.25+LoraV0.5: Minor fixes, expand scenesV0.6: Adjust facial details of AsiansV0.7: Expand image details of the same size [Generated Suggestions]:Consistent with the basic model, no fixed facial features, randomness in facial expression, not very beginner-friendly[Parameter Suggestions]:1. Steps: Generally 20 steps, with no upper limit2. Sampling: E a,D &nbsp;2M S K (commonly used by individuals)3. Prompt word intensity: 1-3 (model has more freedom to play); 4-8; 9++,4. Clip Skip: 25. SDXL Refiner support (recommended to use after being satisfied with the basic model)6. High-frequency repair support (recommended to use after being satisfied with the basic model)7. Facial repair is generally not enabled for personal use8. Coordinating with Lora: It is recommended to test by yourself, Lora is rarely used personally[Creative Overview]:The SDXL basic prototype is already very good, mainly focusing on optimizing towards realism and portraiture on the modelSuggested prompt words usage from SDXL: photo mode, simulate film, movies, neon, detail enhancement modes (can be used in combination)Asian faces: Chinese girl, Asian girl…… Example images are randomly generated, photo mode does not fix facial features, maintaining the universality of the model__SDXL/xlstyle__, __SDXL/filmxl__, __clothes_female/Luxauly_outfit Luxury_Clothing/lux__, __composition_Composition__, __brand/Designer__, girl, 1girl, solo 
```

Files with SHA256: a40c3646a1c22a37a0f6efc82e09b95b2aeca34ce5f038aeb059156db2fb063f