stable-diffusion.cpp Windows CUDA 12 runtime
Unmodified mirror of two release assets of
leejet/stable-diffusion.cpp, tag
master-1086-465adf3. The JoyFusion Windows app downloads them on demand as its optional
NVIDIA acceleration component.
| File | SHA-256 | Contents |
|---|---|---|
sd-master-465adf3-bin-win-cuda12-x64.zip |
d775bff22f7677af664c198d0d978098990bf45525aa538bb7f630f8329ea5ad |
sd-server / sd-cli and the ggml CUDA backend, built by upstream |
cudart-sd-bin-win-cu12-x64.zip |
fe20366827d357c00797eebb58244dddab7fd9a348d70090c3871004c320f38d |
NVIDIA CUDA 12 runtime DLLs (cudart, cuBLAS) as packaged by upstream |
The files are byte-identical to the upstream release; verify with the hashes above.
Licenses
- stable-diffusion.cpp and ggml are MIT licensed; the license texts are inside the first archive.
- The CUDA runtime DLLs are NVIDIA software, distributed under the NVIDIA CUDA Toolkit End User License Agreement (https://docs.nvidia.com/cuda/eula/), which lists these libraries as redistributable.
An NVIDIA GPU with a current driver is required; the CUDA Toolkit does not need to be installed.
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support