[Bug]: Stable Diffusion WebUI + OpenVINO on Windows 11 (2026)

[RoyalXXX]

Is there an existing issue for this?

• I have searched the existing issues and checked the recent builds/commits

What happened?

The installation instructions for Stable Diffusion WebUI with OpenVINO on Windows 11 are badly outdated. If you follow them as-is in 2026, you’ll most likely end up with a pile of errors.

First, the Stable Diffusion repository (https://github.com/Stability-AI/stablediffusion.git) has changed. During installation, Git fails to clone it and reports that the repository does not exist.
The solution is to specify a working fork of the Stable Diffusion repository in webui-user.bat, for example:

set STABLE_DIFFUSION_REPO=https://github.com/joypaul162/Stability-AI-stablediffusion.git
set STABLE_DIFFUSION_COMMIT_HASH=f16630a927e00098b524d687640719e4eb469b76

or

set STABLE_DIFFUSION_REPO=https://github.com/w-e-w/stablediffusion.git
set STABLE_DIFFUSION_COMMIT_HASH=cf1d67a6fd5ea1aa600c4df58e5b47da45f6bdbf

Second, there is an issue with huggingface_hub after installation. Some methods used by the OpenVINO acceleration script were removed in huggingface_hub version 0.26.
To fix this, you need to downgrade huggingface_hub. Activate the virtual environment and run:

pip install huggingface-hub==0.25.2

After making these changes, I did not encounter any other installation problems.

On Intel Iris Xe:

• In CPU mode, using openvino_accelerate.py provides a slight speed improvement.

• In GPU mode, generation is noticeably faster—actually several times faster 😺.

I hope this information helps someone.

Sysinfo

Windows 11 25H2 Build 6200.7628

Some tips:

1. In some cases, for the OpenVINO acceleration script to work correctly on devices with integrated graphics, it is necessary to set a very large page file size. Otherwise, SD WebUI may crash. Monitor the increase in committed memory usage in Task Manager.

2. If the caching option is enabled in the OpenVINO script settings (it is enabled by default), a large amount of cache data will accumulate in the cache folder.
   OpenVINO supports caching of the compiled models for subsequent uses of the same model. This reduces the warm up time significantly. It is recommended to enable this setting. The compiled models are saved in "cache" directory in the WebUI root. The user can delete the files that are no longer needed in order to save disk space.