这个是个普遍问题，例如：Windows平台安装stable-diffusion-webui-1.7.0解决少工具包clip-vit-large-patch14 - 知乎

Stable Diffusion WebUI linux部署问题_oserror: can't load tokenizer for 'openai/clip-vit-CSDN博客 

但是我发现很多朋友是改代码中的绝对路径的，这非常不科学，为啥要用硬编码绝对路径呢？

具体的报错信息如下：

OSError: Can't load tokenizer for 'openai/clip-vit-large-patch14'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'openai/clip-vit-large-patch14' is the correct path to a directory containing all relevant files for a CLIPTokenizer tokenizer.

Stable diffusion model failed to load

解决方法实际很简单，在 stable-diffusion-webui根目录下的stable-diffusion-webui文件夹下创建openai文件夹，然后在一些镜像repo，如：clip-vit-large-patch14: Mirror of https://huggingface.co/openai/clip-vit-large-patch14

把repo clone到openai文件夹就行了。何必要用硬编码呢，完全违反软件工程实践。

例如我的根目录结构如下：

.

├── CHANGELOG.md

├── CITATION.cff

├── CODEOWNERS

├── configs

├── embeddings

├── environment-wsl2.yaml

├── extensions

├── extensions-builtin

├── html

├── javascript

├── launch.py

├── LICENSE.txt

├── localizations

├── models

├── modules

├── package.json

├── pyproject.toml

├── README.md

├── requirements_npu.txt

├── requirements-test.txt

├── requirements.txt

├── requirements_versions.txt

├── screenshot.png

├── script.js

├── scripts

├── stable-diffusion-webui

├── style.css

├── test

├── textual_inversion_templates

├── webui.bat

├── webui-macos-env.sh

├── webui.py

├── webui.sh

├── webui-user.bat

└── webui-user.sh

则需要在标红的目录下新建openai文件夹。不知道这么简单的问题为什么大家讲不清楚。