Closed jazzydog closed 2 months ago
Do you mean in Windows bat script or windows install steps ? If yes, this is because the default installation of onnxruntime and onnxruntime-gpu will install the version built for CUDA 11.x. But just I mentioned in the ticket https://github.com/AiuniAI/Unique3D/issues/15, my env is CUDA 12.1, and according to onnxruntime official doc here - https://onnxruntime.ai/docs/install/ the correct index for onnxruntime and command should be: pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/ this is the reason why we need to unstall them first, then re-install by this link.
if your env is CUDA 11.8, I believe you don't need this step.
Do you mean in Windows bat script or windows install steps ? If yes, this is because the default installation of onnxruntime and onnxruntime-gpu will install the version built for CUDA 11.x. But just I mentioned in the ticket https://github.com/AiuniAI/Unique3D/issues/15, my env is CUDA 12.1, and according to onnxruntime official doc here - https://onnxruntime.ai/docs/install/
the correct index for onnxruntime and command should be:
pip install onnxruntime-gpu --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/
this is the reason why we need to unstall them first, then re-install by this link.
if your env is CUDA 11.8, I believe you don't need this step.