{"id":305,"date":"2026-06-28T00:20:52","date_gmt":"2026-06-27T16:20:52","guid":{"rendered":"https:\/\/www.lixiaodong.com\/?p=305"},"modified":"2026-06-28T00:20:52","modified_gmt":"2026-06-27T16:20:52","slug":"%e4%bd%bf%e7%94%a8nvidia-gpu%e8%bf%90%e8%a1%8cd2ldive-into-deep-learning%e4%b8%ad%e7%9a%84%e6%b7%b1%e5%ba%a6%e5%ad%a6%e4%b9%a0%e4%bb%a3%e7%a0%81","status":"publish","type":"post","link":"https:\/\/www.lixiaodong.com\/?p=305","title":{"rendered":"\u4f7f\u7528Nvidia GPU\u8fd0\u884cd2l(Dive into Deep Learning)\u4e2d\u7684\u6df1\u5ea6\u5b66\u4e60\u4ee3\u7801"},"content":{"rendered":"\n

d2l\u6559\u7a0b\u4e2d\u5b89\u88c5\u7684\u662fcpu\u7248\u672c\u7684pytorch\uff0c\u8fd0\u884c\u67d0\u4e9b\u4f7f\u7528GPU\u7684\u4ee3\u7801\u65f6\u4f1a\u62a5\u9519\u3002\u4e0b\u9762\u5c31\u8bf4\u4e00\u4e0b\u5982\u4f55\u914d\u7f6e\u73af\u5883\u6765\u4f7f\u7528GPU\u6765\u8fd0\u884c\u67d0\u4e9b\u4ee3\u7801\u3002<\/p>\n\n\n\n

    \n
  1. \u4ecenvidia\u5b98\u7f51\u4e0b\u8f7d\u9a71\u52a8\u5e76\u5b89\u88c5\u3002
    \u6211\u4e0b\u8f7d\u7684\u6587\u4ef6\u662f\uff1a582.66-notebook-win10-win11-64bit-international-dch-whql.exe
    \u5b89\u88c5\u5b8c\uff0c\u6253\u5f00\u547d\u4ee4\u884c\u8f93\u5165\uff1anvidia-smi\uff0c\u80fd\u663e\u793a\u4fe1\u606f\u8bf4\u660e\u9a71\u52a8\u5b89\u88c5\u6210\u529f\u3002<\/li>\n\n\n\n
  2. \u5b89\u88c5miniconda3\u3002
    \u5b98\u7f51\u4e0b\u8f7d\u6700\u65b0\u7248\u672cMiniconda3-latest-Windows-x86_64.exe\uff0c\u5b89\u88c5\u3002<\/li>\n\n\n\n
  3. \u521b\u5efaenv\u5e76\u5b89\u88c5\u4f9d\u8d56
    \u6253\u5f00Anaconda Prompt, \u5148\u540c\u610f3\u4e2aterms of service\uff0c\u6267\u884c\u4e0b\u97623\u6761\u547d\u4ee4\uff1a
    conda tos accept –override-channels –channel https:\/\/repo.anaconda.com\/pkgs\/main
    conda tos accept –override-channels –channel https:\/\/repo.anaconda.com\/pkgs\/r
    conda tos accept –override-channels –channel https:\/\/repo.anaconda.com\/pkgs\/msys2<\/li>\n<\/ol>\n\n\n\n

    \u521b\u5efa\u4e00\u4e2a\u65b0\u73af\u5883\uff1aconda create –name d2l python=3.9 -y
    \u6fc0\u6d3b\u4e0a\u9762\u521b\u5efa\u7684\u73af\u5883\uff1aconda activate d2l<\/p>\n\n\n\n

    d2l\u6559\u7a0b\u4e0a\u5b89\u88c5\u7684torch\u7684\u7248\u672c\u662f1.12.0\uff0ctorchvision\u7248\u672c\u662f0.13.0\u3002
    \u5728pytorch\u5b98\u7f51https:\/\/pytorch.org\/get-started\/previous-versions\/ \u6211\u4eec\u80fd\u627e\u5230\u4e0b\u9762\u7684\u8fd9\u884c\uff1a
    conda install pytorch==1.12.0 torchvision==0.13.0 torchaudio==0.12.0 cudatoolkit=11.3 -c pytorch
    \u6211\u4eec\u53ea\u9700\u8981pytorch\u548ctorchvision\uff0c\u4e0d\u9700\u8981torchaudio\uff0c\u6240\u4ee5\u6211\u4eec\u6267\u884c\uff1a<\/p>\n\n\n\n

    conda install pytorch==1.12.0 torchvision==0.13.0 cudatoolkit=11.3 -c pytorch
    \u6ce8\u610f\uff1a\u8fd9\u91cc\u6211\u4eec\u5b89\u88c5\u4e86cudatoolkit\uff0c\u5c31\u4e0d\u9700\u8981\u4ecenvidia\u5b98\u7f51\u4e0b\u8f7d\u5b89\u88c5CUDA\u4e86\u3002<\/p>\n\n\n\n

    pip install d2l==0.17.6<\/p>\n\n\n\n

      \n
    1. \u5b89\u88c5vscode
      \u4e0b\u8f7d\u6700\u65b0\u7248vscode\u5e76\u5b89\u88c5\u3002
      \u6253\u5f00vscode\uff0c\u5b89\u88c5python\u63d2\u4ef6\u3002
      ctrl + shift +p \uff0c\u8f93\u5165python: select interpreter\uff0c\u9009\u62e9\u521a\u624d\u521b\u5efa\u7684d2l\u73af\u5883\u3002
      \u7535\u8111\u4e0a\u65b0\u5efa\u4e00\u4e2a\u76ee\u5f55\uff0c\u4f5c\u4e3apython\u5de5\u7a0b\u7684\u76ee\u5f55\u3002
      vscode\u91cc\u6253\u5f00\u4e0a\u9762\u521b\u5efa\u7684\u76ee\u5f55\u3002
      \u65b0\u5efapy\u6587\u4ef6\uff0c\u7c98\u8d34d2l\u6559\u7a0b\u91cc\u7684\u4ee3\u7801\uff0c\u59826.6\u5c0f\u8282 \u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff08LeNet\uff09\u7684\u4ee3\u7801\u3002<\/li>\n\n\n\n
    2. \u9047\u5230\u7684\u95ee\u9898
      \u8fd0\u884c\u65f6\u9047\u5230\u4ee5\u4e0b\u95ee\u9898\uff1a
      backend = matplotlib.rcParams._get(“backend”)
      AttributeError: ‘RcParams’ object has no attribute ‘_get’<\/li>\n<\/ol>\n\n\n\n

      \u6211\u4eec\u5728\u6587\u4ef6\u6700\u5f00\u59cb\u52a0\u5165\u4e0b\u9762\u51e0\u884c\u5c31\u80fd\u89e3\u51b3\u3002<\/p>\n\n\n\n

      import matplotlib\n# Patch the missing private method to match the standard public .get method\nif not hasattr(matplotlib.rcParams, '_get'):\n    matplotlib.rcParams._get = matplotlib.rcParams.get<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"
      d2l\u6559\u7a0b\u4e2d\u5b89\u88c5\u7684\u662fcpu\u7248\u672c\u7684pytorch\uff0c\u8fd0\u884c\u67d0\u4e9b\u4f7f\u7528GPU\u7684\u4ee3\u7801\u65f6\u4f1a\u62a5\u9519 … \u7ee7\u7eed\u9605\u8bfb →<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[57],"tags":[],"class_list":["post-305","post","type-post","status-publish","format-standard","hentry","category-57"],"_links":{"self":[{"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=\/wp\/v2\/posts\/305","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=305"}],"version-history":[{"count":3,"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=\/wp\/v2\/posts\/305\/revisions"}],"predecessor-version":[{"id":308,"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=\/wp\/v2\/posts\/305\/revisions\/308"}],"wp:attachment":[{"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=305"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=305"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.lixiaodong.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=305"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}