Nvidia GPU and CUDA Notes

SD.Next will auto-detect nVidia GPUs and attempt to install latest stable torch with CUDA support on launch

Important

Do not manually install CUDA on your system as torch will install the appropriate CUDA version in its own environment
System-wide installation of CUDA may cause conflicts and is not required for SD.Next to utilize nVidia GPUs.

Tip

Disable system memory fallback to avoid massive performance degradation when exhausting GPU memory: Instructions

Older GPUs

If you have older nVidia GPU that is not supported by latest torch releases, you may need to install specific older version of torch with appropriate CUDA version that supports your GPU.

Example: Pascal architecture (e.g., RTX 10xx series) or older require torch==2.9 with cuda==12.6 Use following before first startup to force installation of specific torch/cuda version:

set TORCH_COMMAND='torch==2.9.1 torchvision==0.24.1 torchaudio==2.9.1 --index-url https://download.pytorch.org/whl/cu126'

Older Drivers

If you're attempting to run on a system with older nVidia drivers that do not support latest CUDA versions, you may need to install specific older version of torch with appropriate CUDA version that is compatible with your drivers.

This is common when running on cloud instances which do not update drivers frequently.

Example: cuda==13.0 requires newer nVidia drivers
Use following before first startup to force installation of torch==2.11.0 with cuda==12.68:

set TORCH_COMMAND='torch torchvision --index-url https://download.pytorch.org/whl/cu128