一、预处理器和WebUI不同,只有3个。根据具体使用的ip-adapter模型选择对应的img encoder

参考链接:https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/932

二、预处理器的文件下载与安装路径

(1) CLIP-ViT-H-14:https://huggingface.co/h94/IP-Adapter/blob/main/models/image_encoder/model.safetensors

下载完更名为CLIP-ViT-H-14.safetensors,放到./models/ControlNetPreprocessor/路径下

(2) CLIP-ViT-bigG-14:https://huggingface.co/h94/IP-Adapter/blob/main/sdxl_models/image_encoder/model.safetensors

下载完更名为CLIP-ViT-bigG-14.safetensors,放到./models/ControlNetPreprocessor/路径下

(3) InsightFace:https://github.com/deepinsight/insightface/releases/download/v0.7/buffalo_l.zip

在./models/路径下创建insightface/models两级文件夹,将下载的buffalo_l.zip解压在./models/insightface/models/路径下面即可


标签: none

添加新评论