onnxruntime 不使用 gpu

检查 onnxruntime  和 cuda 和 cudnn 的版本对应关

sudo apt purge -y nvidia* libnvidia*
sudo apt autoremove -y
sudo apt install -y nvidia-driver-520
sudo apt install -y cuda-11-8


nvidia 驱动 cuda cuDNN版本关系
https://docs.nvidia.com/deeplearning/cudnn/support-matrix/index.html

检测cuda版本
cat /usr/local/cuda/version.json

nvida-smi

两者可能不一致 https://stackoverflow.com/questions/53422407/different-cuda-versions-shown-by-nvcc-and-nvidia-smi


# add latest key
curl -sS https://dl.yarnpkg.com/debian/pubkey.gpg | sudo apt-key add -

# try refresh again
sudo apt-get update



https://developer.nvidia.com/cuda-11-8-0-download-archive?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=18.04&target_type=deb_network

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-keyring_1.0-1_all.deb
sudo dpkg -i cuda-keyring_1.0-1_all.deb
sudo apt-get update
sudo apt -y upgrade
sudo apt-get -y install cuda

系

NVIDIA - CUDA | onnxruntime

检查cuda

nvcc --version 

检查安装的 onnx

pip list | grep onnx

检查 cudnn版本

cat /usr/include/x86_64-linux-gnu/cudnn_v*.h | grep CUDNN_MAJOR -A 2