Tensorrt install error

[Tigeryang93]

python -m pip install --pre -i https://pypi.nvidia.com tensorrt==9.0.1.post11.dev4 Looking in indexes: https://pypi.nvidia.com Collecting tensorrt==9.0.1.post11.dev4 Using cached tensorrt-9.0.1.post11.dev4-py2.py3-none-any.whl Collecting tensorrt-libs==9.0.1.post11.dev4 (from tensorrt==9.0.1.post11.dev4) Using cached https://pypi.nvidia.com/tensorrt-libs/tensorrt_libs-9.0.1.post11.dev4-py2.py3-none-manylinux_2_17_x86_64.whl (1060.6 MB) Collecting tensorrt-bindings==9.0.1.post11.dev4 (from tensorrt==9.0.1.post11.dev4) Using cached https://pypi.nvidia.com/tensorrt-bindings/tensorrt_bindings-9.0.1.post11.dev4-cp310-none-manylinux_2_17_x86_64.whl (956 kB) INFO: pip is looking at multiple versions of tensorrt-libs to determine which version is compatible with other requirements. This could take a while. ERROR: Could not find a version that satisfies the requirement nvidia-cuda-runtime-cu11 (from tensorrt-libs) (from versions: none) ERROR: No matching distribution found for nvidia-cuda-runtime-cu11

[vivekmaru36]

installl compatible cuda version first than cudo toolkit in your case for 9.0.1.post11.dev4 from here

https://docs.nvidia.com/deeplearning/tensorrt/support-matrix/index.html

i could not find which one will specifically work for you but it might be any of these :

https://developer.nvidia.com/cuda-90-download-archive

https://developer.nvidia.com/cuda-91-download-archive

[albireophyhe]

I got this problem when I was using CUDA 12.*, this error probably comes from the different dependencies of the tensorrt package with different versions. An instant solution about this is just to use CUDA 11.8. It is possible to run it by manually installing a set of CUDA 12.1 capable cudnn, tensorrt, polygraphy, and onnx_graphsurgeon, but you may just start another game of package version matching.