These are fusion Lora's based on my other Lora's and are designed to get a prettier face shape, different models represent different face shapes and you can choose them according to preference.
The original lora is based on chilloutmix-based training, usually with weights between 0.4 and 0.8.
这些是基于我的其他Lora的融合Lora，旨在获得更漂亮的脸型，不同型号代表不同的脸型，你可以根据喜好选择她们,
原始lora基于基于chilloutmix训练，通常权重在0.4-0.8之间,
描述:
融了新炼的05年神仙妹妹的lora，相比神仙妹妹，这个lora面部更柔和，但是磨皮有点重
建议权重0.5-0.9，在ChilloutMix、Majic系列里表现不错。
最近比较喜欢把脸型的lora放到adetailer插件里用，可以防止污染到光影、背景和衣物。
训练词语:
名称: GFMix-LY55.safetensors
大小 (KB): 36985
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
