{"id":51613,"date":"2024-10-08T20:06:21","date_gmt":"2024-10-08T12:06:21","guid":{"rendered":"https:\/\/server.hk\/cnblog\/51613\/"},"modified":"2024-11-14T10:42:32","modified_gmt":"2024-11-14T02:42:32","slug":"ubuntu-18-04-n-%e5%8d%a1%e9%a9%85%e5%8b%95%e5%ae%89%e8%a3%9d-cuda-10-0-cudnn-7-5-anaconda-tensorflow-gpu","status":"publish","type":"post","link":"https:\/\/server.hk\/cnblog\/51613\/","title":{"rendered":"Ubuntu 18.04 N \u5361\u9a45\u52d5\u5b89\u88dd + CUDA 10.0 + cuDNN 7.5 + Anaconda + Tensorflow-GPU"},"content":{"rendered":"

Ubuntu 18.04 N \u5361\u9a45\u52d5\u5b89\u88dd + CUDA 10.0 + cuDNN 7.5 + Anaconda + Tensorflow-GPU<\/h1>\n

\u5728\u6df1\u5ea6\u5b78\u7fd2\u548c\u6a5f\u5668\u5b78\u7fd2\u7684\u9818\u57df\u4e2d\uff0cNVIDIA \u7684 GPU \u63d0\u4f9b\u4e86\u5f37\u5927\u7684\u8a08\u7b97\u80fd\u529b\u3002\u672c\u6587\u5c07\u4ecb\u7d39\u5982\u4f55\u5728 Ubuntu 18.04 \u4e0a\u5b89\u88dd NVIDIA \u9a45\u52d5\u3001CUDA 10.0\u3001cuDNN 7.5\u3001Anaconda \u4ee5\u53ca TensorFlow-GPU\uff0c\u5e6b\u52a9\u60a8\u5feb\u901f\u642d\u5efa\u6df1\u5ea6\u5b78\u7fd2\u74b0\u5883\u3002<\/p>\n

\u6b65\u9a5f\u4e00\uff1a\u5b89\u88dd NVIDIA \u9a45\u52d5<\/h2>\n

\u9996\u5148\uff0c\u60a8\u9700\u8981\u5b89\u88dd NVIDIA \u9a45\u52d5\u3002\u53ef\u4ee5\u901a\u904e\u4ee5\u4e0b\u6b65\u9a5f\u9032\u884c\u5b89\u88dd\uff1a<\/p>\n

    \n
  1. \u66f4\u65b0\u7cfb\u7d71\uff1a<\/li>\n
    sudo apt update\nsudo apt upgrade<\/code><\/pre>\n
  2. \u6dfb\u52a0 NVIDIA PPA\uff1a<\/li>\n
    sudo add-apt-repository ppa:graphics-drivers\/ppa\nsudo apt update<\/code><\/pre>\n
  3. \u5b89\u88dd NVIDIA \u9a45\u52d5\uff1a<\/li>\n
    sudo apt install nvidia-driver-460<\/code><\/pre>\n
  4. \u91cd\u555f\u7cfb\u7d71\uff1a<\/li>\n
    sudo reboot<\/code><\/pre>\n
  5. \u78ba\u8a8d\u9a45\u52d5\u5b89\u88dd\uff1a<\/li>\n
    nvidia-smi<\/code><\/pre>\n<\/ol>\n
    \u5982\u679c\u4e00\u5207\u6b63\u5e38\uff0c\u60a8\u61c9\u8a72\u80fd\u770b\u5230 NVIDIA \u9a45\u52d5\u7684\u7248\u672c\u548c GPU \u7684\u4f7f\u7528\u60c5\u6cc1\u3002<\/p>\n
    \u6b65\u9a5f\u4e8c\uff1a\u5b89\u88dd CUDA 10.0<\/h2>\n
    \u63a5\u4e0b\u4f86\uff0c\u60a8\u9700\u8981\u5b89\u88dd CUDA 10.0\u3002\u8acb\u6309\u7167\u4ee5\u4e0b\u6b65\u9a5f\u9032\u884c\uff1a<\/p>\n
      \n
    1. \u4e0b\u8f09 CUDA 10.0 \u5b89\u88dd\u5305\uff1a<\/li>\n
      \u8a2a\u554f NVIDIA CUDA Downloads<\/a>\uff0c\u9078\u64c7 Linux -> x86_64 -> Ubuntu -> 18.04 -> deb (network) \u4e0b\u8f09\u3002<\/p>\n
    2. \u5b89\u88dd CUDA\uff1a<\/li>\n
      sudo dpkg -i cuda-repo-ubuntu1804-_amd64.deb\nsudo apt-key adv --fetch-keys https:\/\/developer.download.nvidia.com\/compute\/cuda\/repos\/ubuntu1804\/x86_64\/7fa2af80.pub\nsudo apt update\nsudo apt install cuda-10-0<\/code><\/pre>\n
    3. \u8a2d\u7f6e\u74b0\u5883\u8b8a\u91cf\uff1a<\/li>\n
      echo 'export PATH=\/usr\/local\/cuda-10.0\/bin:$PATH' >> ~\/.bashrc\necho 'export LD_LIBRARY_PATH=\/usr\/local\/cuda-10.0\/lib64:$LD_LIBRARY_PATH' >> ~\/.bashrc\nsource ~\/.bashrc<\/code><\/pre>\n<\/ol>\n
      \u6b65\u9a5f\u4e09\uff1a\u5b89\u88dd cuDNN 7.5<\/h2>\n
      cuDNN \u662f NVIDIA \u63d0\u4f9b\u7684\u6df1\u5ea6\u5b78\u7fd2\u52a0\u901f\u5eab\u3002\u5b89\u88dd\u6b65\u9a5f\u5982\u4e0b\uff1a<\/p>\n
        \n
      1. \u8a2a\u554f NVIDIA cuDNN<\/a>\uff0c\u8a3b\u518a\u4e26\u4e0b\u8f09 cuDNN 7.5 \u7684 tar \u6a94\u6848\u3002<\/li>\n
      2. \u89e3\u58d3\u4e26\u5b89\u88dd cuDNN\uff1a<\/li>\n
        tar -xzvf cudnn-10.0-linux-x64-v7.5.0.56.tgz\nsudo cp cuda\/include\/cudnn*.h \/usr\/local\/cuda\/include\nsudo cp cuda\/lib64\/libcudnn* \/usr\/local\/cuda\/lib64\nsudo chmod a+r \/usr\/local\/cuda\/include\/cudnn*.h \/usr\/local\/cuda\/lib64\/libcudnn*<\/code><\/pre>\n<\/ol>\n
        \u6b65\u9a5f\u56db\uff1a\u5b89\u88dd Anaconda<\/h2>\n
        Anaconda \u662f\u4e00\u500b\u6d41\u884c\u7684 Python \u767c\u884c\u7248\uff0c\u9069\u5408\u6578\u64da\u79d1\u5b78\u548c\u6a5f\u5668\u5b78\u7fd2\u3002\u5b89\u88dd\u6b65\u9a5f\u5982\u4e0b\uff1a<\/p>\n
          \n
        1. \u4e0b\u8f09 Anaconda \u5b89\u88dd\u8173\u672c\uff1a<\/li>\n
          wget https:\/\/repo.anaconda.com\/archive\/Anaconda3-2020.07-Linux-x86_64.sh<\/code><\/pre>\n
        2. \u904b\u884c\u5b89\u88dd\u8173\u672c\uff1a<\/li>\n
          bash Anaconda3-2020.07-Linux-x86_64.sh<\/code><\/pre>\n
        3. \u6309\u7167\u63d0\u793a\u5b8c\u6210\u5b89\u88dd\uff0c\u4e26\u91cd\u555f\u7d42\u7aef\u3002<\/li>\n<\/ol>\n
          \u6b65\u9a5f\u4e94\uff1a\u5b89\u88dd TensorFlow-GPU<\/h2>\n
          \u6700\u5f8c\uff0c\u60a8\u53ef\u4ee5\u5b89\u88dd TensorFlow-GPU\u3002\u8acb\u4f7f\u7528\u4ee5\u4e0b\u547d\u4ee4\uff1a<\/p>\n
          conda create -n tf_gpu python=3.7\nconda activate tf_gpu\npip install tensorflow-gpu==2.3.0<\/code><\/pre>\n
          \u5b89\u88dd\u5b8c\u6210\u5f8c\uff0c\u60a8\u53ef\u4ee5\u901a\u904e\u4ee5\u4e0b\u547d\u4ee4\u6aa2\u67e5 TensorFlow \u662f\u5426\u80fd\u5920\u4f7f\u7528 GPU\uff1a<\/p>\n
          python -c \"import tensorflow as tf; print(tf.test.is_gpu_available())\"<\/code><\/pre>\n
          \u7e3d\u7d50<\/h2>\n
          \u901a\u904e\u4ee5\u4e0a\u6b65\u9a5f\uff0c\u60a8\u5df2\u7d93\u6210\u529f\u5728 Ubuntu 18.04 \u4e0a\u5b89\u88dd\u4e86 NVIDIA \u9a45\u52d5\u3001CUDA 10.0\u3001cuDNN 7.5\u3001Anaconda \u4ee5\u53ca TensorFlow-GPU\u3002\u9019\u4e9b\u5de5\u5177\u5c07\u5e6b\u52a9\u60a8\u5728\u6df1\u5ea6\u5b78\u7fd2\u7684\u65c5\u7a0b\u4e2d\u66f4\u9ad8\u6548\u5730\u9032\u884c\u8a08\u7b97\u3002\u5982\u679c\u60a8\u9700\u8981\u7a69\u5b9a\u7684 VPS<\/a> \u4f86\u904b\u884c\u60a8\u7684\u6df1\u5ea6\u5b78\u7fd2\u6a21\u578b\uff0c\u8acb\u8003\u616e\u6211\u5011\u7684\u670d\u52d9\uff0c\u63d0\u4f9b\u9ad8\u6548\u80fd\u7684\u8a08\u7b97\u8cc7\u6e90\uff0c\u52a9\u60a8\u4e00\u81c2\u4e4b\u529b\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
          Learn how to install NVIDIA drivers, CUDA 10.0, cuDNN 7.5, and set up Anaconda for TensorFlow-GPU on Ubuntu 18.04 in this comprehensive guide.<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4924],"tags":[],"class_list":["post-51613","post","type-post","status-publish","format-standard","hentry","category-setup-tutorials"],"_links":{"self":[{"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/posts\/51613","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/comments?post=51613"}],"version-history":[{"count":1,"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/posts\/51613\/revisions"}],"predecessor-version":[{"id":51614,"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/posts\/51613\/revisions\/51614"}],"wp:attachment":[{"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/media?parent=51613"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/categories?post=51613"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/server.hk\/cnblog\/wp-json\/wp\/v2\/tags?post=51613"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}