cuda\cudnn\torch安装-526互联

cuda toolkits 是工具包包含cuda和显卡驱动，需要选择显卡合适的版本，到官网查看 cudnn是加速的 torch是需要支持cuda的版本 ------------------------------------------------cuda安装----------------------------------------------------------- 1、https://blog.csdn.net/RenZuoym/article/details/132211730 2、下载：cuda—toolkits https://developer.nvidia.com/cuda-12-2-0-download-archive?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=22.04&target_type=deb_local 3、设置环境变量，依次敲以下命令 export PATH=$PATH:/usr/local/cuda-12.2/bin export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-12.2/lib64 export CUDA_HOME=$CUDA_HOME:/usr/local/cuda-12.2 ---------------------------------------cuda验证-------------------------------------- 验证：命令行输入： nvidia-smi、 nvcc --version ------------------------------------------------cudnn安装--------------------------- 1、先下载文件，wget http://47.105.176.36:9999/down/VWz6gRjGUoqj.deb 2、打开网址：https://docs.nvidia.com/deeplearning/cudnn/install-guide/index.html 用作参考 3、重命名：mv dash.deb cudnn-local-repo-ubuntu2204-8.9.7.29_1.0-1_amd64.deb 4：依次敲以下命令： sudo dpkg -i cudnn-local-repo-ubuntu2204-8.9.7.29_1.0-1_amd64.deb sudo cp /var/cudnn-local-repo-/cudnn-local--keyring.gpg /usr/share/keyrings/ sudo apt-get update sudo apt-cache madison libcudnn8 sudo apt-get install libcudnn8=8.9.7.29-1+cuda12.2 --------------------------------cudnn验证---------------------------- cudnn验证：https://blog.csdn.net/kz_java/article/details/122067916 -------------------------------------------------pytorch安装---------------------- 官网： pip3 install torch torchvision torchaudio ------验证torch和cuda是否可用---- 进入python3命令行：import torch torch.cuda.is_available() 如果返回True 表示ok --------假如torch 验证出现出题------ https://www.cnblogs.com/huadongw/p/16504137.html 依次敲入以下命令 version=535.54.03 main_version=$(echo $version | awk -F '.' '{print $1}') apt-get update apt-get -y install nvidia-fabricmanager-${main_version}=${version}-* systemctl enable nvidia-fabricmanager systemctl restart nvidia-fabricmanager systemctl status nvidia-fabricmanager