对天涯明月刀OL游戏内人物外观进行LORA模型训练后的成果，通过LoRA分层移除了对脸部的训练，使其脸部的生成更贴近大模型的效果，减少了崩脸的概率，但依旧推荐你采用面部修复，预览图所用大模型为Guofeng3_v34（不知道为什么，图片在放大过程中会出现竖条纹而不放大就正常）对模型测试以及炼丹感兴趣的，欢迎来Q群850498709（一个人又训练又测试还要跑图，太疲乏了而且容易忽略问题）如果你喜欢本模型，也欢迎你来赞助我，这将是我更新模型的最大动力，赞助链接：<a target="_blank" rel="ugc" href="https://fpoh.usells.com/p/d9Uu7h">https://fpoh.usells.com/p/d9Uu7h</a>以下是展示预览图的生成信息（C站自动识别不了，一个一个填又太麻烦）提示词1girl,solo,&lt;lora:[衣服]-[九嶷琼芳-天刀少女]TDSVjyqf:0.6&gt;,TDSVjyqf,(early teen:1.2),(petite:1.2),baby face,long_skirt,pink dancing shoes,(standing:1.2),(full body:1.5),(simple background:1.2),indoor,(looking_at_viewer:1.2),(best quality),((masterpiece)), 负面提示词(expose thighs:1.2),nsfw,sexy,(worst quality:2),(low quality:2),(normal quality:2), (unclear eyes:1.331),bad hands,missing fingers,bad body,(more than 2 nipples:1.331),(extra legs:1.331),(unclear eyes:1.331),(high_heels:1.4),(multi-view:1.2),其他参数Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 11, Seed: 734606849, Size: 600x752, Model: [3D]-[国风]3Guofeng3_v34, VAE: 560K-EMA_VAE.safetensors, Denoising strength: 0.5, Clip skip: 2, ADetailer model: mediapipe_face_full, ADetailer confidence: 0.3, ADetailer dilate/erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.9.1, Hires upscale: 2, Hires steps: 15, Hires upscaler: 4x-UltraSharp, Lora hashes: "TDSVjyqf_20230904193208-000007: 7fe990a3cee0", Dynamic thresholding enabled: True, Mimic scale: 7, Separate Feature Channels: True, Scaling Startpoint: MEAN, Variability Measure: AD, Interpolate Phi: 1, Threshold percentile: 100, Eta: 0.667, Version: v1.6.0

Files with SHA256: 31c681bb90fc8a5e2d391c7e3b153089d4b24f3e37aed88c74abc5cf6c1e066b