{"id":3741,"date":"2024-05-11T19:27:53","date_gmt":"2024-05-11T11:27:53","guid":{"rendered":"https:\/\/www.aqwu.net\/wp\/?p=3741"},"modified":"2024-05-25T08:50:13","modified_gmt":"2024-05-25T00:50:13","slug":"windows-%e7%8e%af%e5%a2%83%e4%b8%8b-ollama-%e7%9a%84-gpu-%e6%94%af%e6%8c%81","status":"publish","type":"post","link":"https:\/\/www.aqwu.net\/wp\/?p=3741","title":{"rendered":"Windows \u73af\u5883\u4e0b Ollama \u7684 GPU \u652f\u6301"},"content":{"rendered":"\n

1. \u8bbe\u7f6e<\/h2>\n\n\n\n

1.1  NVidia \u9a71\u52a8\u7a0b\u5e8f\u548cCUDA \u5de5\u5177\u5305<\/strong><\/h3>\n\n\n\n

1.1.1  NVidia \u9a71\u52a8\u7a0b\u5e8f<\/strong><\/h4>\n\n\n\n

\u5728\u9009\u62e9NVIDIA\u9a71\u52a8\u7a0b\u5e8f\u65f6\uff0c\u6709\u4e24\u79cd\u4e3b\u8981\u7c7b\u578b\uff1aGame Ready Driver (GRD) \u548c Studio Driver (SD)\u3002\u8fd9\u4e24\u8005\u7684\u9009\u62e9\u53d6\u51b3\u4e8e\u60a8\u7684\u4e3b\u8981\u7528\u9014\uff1a<\/p>\n\n\n\n

    \n
  1. Game Ready Driver (GRD)<\/strong>\uff1a\u8fd9\u79cd\u7c7b\u578b\u7684\u9a71\u52a8\u7a0b\u5e8f\u4e3b\u8981\u9488\u5bf9\u6e38\u620f\u73a9\u5bb6\u8bbe\u8ba1\uff0c\u786e\u4fdd\u6700\u65b0\u7684\u6e38\u620f\u80fd\u591f\u5728\u53d1\u5e03\u65f6\u83b7\u5f97\u6700\u4f73\u6027\u80fd\u548c\u4f53\u9a8c\u3002\u5982\u679c\u60a8\u4e3b\u8981\u7528\u663e\u5361\u73a9\u6e38\u620f\uff0c\u90a3\u4e48GRD\u662f\u8f83\u597d\u7684\u9009\u62e9\u3002<\/li>\n\n\n\n
  2. Studio Driver (SD)<\/strong>\uff1a\u8fd9\u79cd\u9a71\u52a8\u7a0b\u5e8f\u662f\u4e3a\u521b\u610f\u4e13\u4e1a\u4eba\u58eb\u8bbe\u8ba1\u7684\uff0c\u4f8b\u5982\u90a3\u4e9b\u4ece\u4e8b\u89c6\u9891\u7f16\u8f91\u30013D\u6e32\u67d3\u3001\u56fe\u5f62\u8bbe\u8ba1\u548c\u5176\u4ed6\u5f62\u5f0f\u7684\u521b\u4f5c\u5185\u5bb9\u7684\u4eba\u3002Studio Driver\u4e3a\u8fd9\u4e9b\u5e94\u7528\u7a0b\u5e8f\u63d0\u4f9b\u4e86\u66f4\u7a33\u5b9a\u548c\u4f18\u5316\u7684\u652f\u6301\u3002<\/li>\n<\/ol>\n\n\n\n

    \u7531\u4e8e\u60a8\u63d0\u5230\u662f\u7528\u4e8eAI\u8bad\u7ec3\uff0cStudio Driver (SD) \u66f4\u9002\u5408\u60a8\u7684\u9700\u6c42\u3002\u5b83\u4e3a\u4e13\u4e1a\u8f6f\u4ef6\u548c\u5de5\u4f5c\u8d1f\u8f7d\u63d0\u4f9b\u4e86\u4f18\u5316\u548c\u7a33\u5b9a\u6027\uff0c\u7279\u522b\u662f\u5728\u5904\u7406\u56fe\u5f62\u548c\u8ba1\u7b97\u5bc6\u96c6\u578b\u4efb\u52a1\u65f6\uff0c\u8fd9\u5bf9AI\u8bad\u7ec3\u662f\u975e\u5e38\u91cd\u8981\u7684\u3002<\/p>\n\n\n\n

    https:\/\/www.nvidia.com\/download\/index.aspx<\/a><\/p>\n\n\n\n

    \"\"<\/figure>\n\n\n\n

    \u5982\u679c\u4f7f\u7528\u4e2d\u6587\u754c\u9762\uff0c\u53ef\u4ee5\u4f7f\u7528\u4e0b\u9762\u7684\u94fe\u63a5<\/p>\n\n\n\n

    https:\/\/www.nvidia.cn\/Download\/index.aspx?lang=zh-cn<\/a><\/p>\n\n\n\n

    \"\"<\/figure>\n\n\n\n

    1.1.2  NVidia CUDA \u5de5\u5177\u5305<\/strong><\/h4>\n\n\n\n

    https:\/\/developer.nvidia.com\/cuda-downloads<\/a><\/p>\n\n\n\n

    \u68c0\u67e5 nvidia-smi.exe \u548c nvcc.exe \u4e2d\u5bf9 cuda \u7f16\u8bd1\u5de5\u5177\u7684 GPU \u652f\u6301\u3002\/11\/12.<\/p>\n\n\n\n

    1.2  Git<\/strong><\/h3>\n\n\n\n

    https:\/\/git-scm.com\/download\/win<\/a><\/p>\n\n\n\n

    1.3  Python<\/strong><\/h3>\n\n\n\n

    https:\/\/www.python.org\/downloads\/windows\/<\/a><\/p>\n\n\n\n

    1.4 Go<\/strong><\/h3>\n\n\n\n

    https:\/\/go.dev\/doc\/install<\/a><\/p>\n\n\n\n

    1.5 Gcc<\/strong><\/h3>\n\n\n\n

    https:\/\/www.msys2.org\/<\/a><\/p>\n\n\n\n

    \u4e0b\u6ed1\u627e\u5230 .exe \u6587\u4ef6\uff0c\u6bd4\u5982\u4e0b\u9762\u7684\u94fe\u63a5\uff1a<\/p>\n\n\n\n

    https:\/\/github.com\/msys2\/msys2-installer\/releases\/download\/2024-05-07\/msys2-x86_64-20240507.exe<\/a><\/p>\n\n\n\n

    \u4e0b\u8f7d\u5b8c\u6210\u540e\uff0c\u70b9\u51fb msys2-x86_64-20240507.exe<\/a> \u8fd0\u884c\uff1a<\/p>\n\n\n\n

    \u9009\u62e9\u9ed8\u8ba4\u5b89\u88c5\u8def\u5f84\u5c31\u53ef\u4ee5\uff0c\u9ed8\u8ba4\u5b89\u88c5\u8def\u5f84\u662f C:\\msys64<\/p>\n\n\n\n

    \u70b9\u51fb\u8fd0\u884c mingw64.exe<\/strong><\/p>\n\n\n\n

    \u5b89\u88c5 gcc \u548c mingw32-make<\/p>\n\n\n\n

    pacman -S mingw-w64-x86_64-toolchain<\/pre><\/div>\n\n\n\n
    \u5982\u679c\u8981\u66f4\u65b0\u7cfb\u7edf\uff1a<\/p>\n\n\n\n
    pacman -Syu\n<\/pre><\/div>\n\n\n\n
    \u9a8c\u8bc1 gcc<\/strong><\/p>\n\n\n\n
    where gcc\nC:\\msys64\\mingw64\\bin\\gcc.exe\n\n$ gcc --version\ngcc.exe (Rev2, Built by MSYS2 project) 14.1.0\nCopyright (C) 2024 Free Software Foundation, Inc.\nThis is free software; see the source for copying conditions.  There is NO\nwarranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\nwhere mingw32-make\nC:\\msys64\\mingw64\\bin\\mingw32-make.exe\n\nmingw32-make.exe --version\nGNU Make 4.4.1\nBuilt for Windows32\nCopyright (C) 1988-2023 Free Software Foundation, Inc.\nLicense GPLv3+: GNU GPL version 3 or later <https:\/\/gnu.org\/licenses\/gpl.html>\nThis is free software: you are free to change and redistribute it.\nThere is NO WARRANTY, to the extent permitted by law.\n\n<\/pre><\/div>\n\n\n\n
    \u70b9\u51fb gcc.exe \u8def\u5f84\u5230windows \u7cfb\u7edf\u8def\u5f84\u4e0b<\/p>\n\n\n\n
    C:\\msys64\\mingw64\\bin<\/pre><\/div>\n\n\n\n
    1.6 Cmake<\/strong><\/h3>\n\n\n\n
    https:\/\/cmake.org\/download\/<\/a><\/p>\n\n\n\n
    1.7 ollama<\/strong><\/h3>\n\n\n\n
    \u9700\u8981\u5b89\u88c5\u4e00\u4e2a ollama \u7684\u8fd0\u884c\u73af\u5883\uff0c\u4e0b\u9762\u7684\u6b65\u9aa4\u4f1a\u9700\u8981\u3002<\/p>\n\n\n\n
    https:\/\/ollama.com\/download\/OllamaSetup.exe<\/a><\/p>\n\n\n\n
    2. \u9a8c\u8bc1\u5b89\u88c5\u5305<\/strong><\/h2>\n\n\n\n
    2.1 \u514b\u9686 ollama<\/strong><\/h3>\n\n\n\n
    gigit clone --recurse-submodules https:\/\/github.com\/jmorganca\/ollama\nCloning into 'ollama'...\nremote: Enumerating objects: 16306, done.\nremote: Counting objects: 100% (1012\/1012), done.\nremote: Compressing objects: 100% (564\/564), done.\nremote: Total 16306 (delta 584), reused 812 (delta 445), pack-reused 15294\nReceiving objects: 100% (16306\/16306), 10.05 MiB | 2.49 MiB\/s, done.\nResolving deltas: 100% (10212\/10212), done.\nSubmodule 'llama.cpp' (https:\/\/github.com\/ggerganov\/llama.cpp.git) registered for path 'llm\/llama.cpp'\nCloning into '\/mnt\/e\/ollama\/ollama-gpu\/ollama\/llm\/llama.cpp'...\nremote: Enumerating objects: 15930, done.\nremote: Counting objects: 100% (15930\/15930), done.\nremote: Compressing objects: 100% (5519\/5519), done.\nremote: Total 15930 (delta 11617), reused 14309 (delta 10217), pack-reused 0\nReceiving objects: 100% (15930\/15930), 35.29 MiB | 1.53 MiB\/s, done.\nResolving deltas: 100% (11617\/11617), done.\nremote: Enumerating objects: 11, done.\nremote: Counting objects: 100% (11\/11), done.\nremote: Compressing objects: 100% (2\/2), done.\nremote: Total 2 (delta 1), reused 1 (delta 0), pack-reused 0\nUnpacking objects: 100% (2\/2), 915 bytes | 35.00 KiB\/s, done.\nFrom https:\/\/github.com\/ggerganov\/llama.cpp\n * branch            74f33adf5f8b20b08fc5a6aa17ce081abe86ef2f -> FETCH_HEAD\nSubmodule path 'llm\/llama.cpp': checked out '74f33adf5f8b20b08fc5a6aa17ce081abe86ef2f'\nSubmodule 'kompute' (https:\/\/github.com\/nomic-ai\/kompute.git) registered for path 'llm\/llama.cpp\/kompute'\nCloning into '\/mnt\/e\/ollama\/ollama-gpu\/ollama\/llm\/llama.cpp\/kompute'...\nremote: Enumerating objects: 9090, done.\nremote: Counting objects: 100% (225\/225), done.\nremote: Compressing objects: 100% (134\/134), done.\nremote: Total 9090 (delta 99), reused 174 (delta 81), pack-reused 8865\nReceiving objects: 100% (9090\/9090), 17.58 MiB | 3.79 MiB\/s, done.\nResolving deltas: 100% (5706\/5706), done.\nSubmodule path 'llm\/llama.cpp\/kompute': checked out '4565194ed7c32d1d2efa32ceab4d3c6cae006306'<\/pre><\/div>\n\n\n\n
    2.2 \u9a8c\u8bc1\u524d\u9762\u7684\u5b89\u88c5\u5305<\/strong><\/h3>\n\n\n\n
    2.2.1 \u9a8c\u8bc1 nvcc<\/strong><\/h4>\n\n\n\n
    where nvcc\nC:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v12.4\\bin\\nvcc.exe\n\nnvcc --version\nnvcc: NVIDIA (R) Cuda compiler driver\nCopyright (c) 2005-2024 NVIDIA Corporation\nBuilt on Thu_Mar_28_02:30:10_Pacific_Daylight_Time_2024\nCuda compilation tools, release 12.4, V12.4.131\nBuild cuda_12.4.r12.4\/compiler.34097967_0\n<\/pre><\/div>\n\n\n\n
    2.2.2 \u9a8c\u8bc1 git<\/strong><\/h4>\n\n\n\n
    where git\nC:\\Program Files\\Git\\cmd\\git.exe\n\ngit -v\ngit version 2.45.0.windows.1<\/pre><\/div>\n\n\n\n
    2.2.3 \u9a8c\u8bc1 python<\/p>\n\n\n\n
    python --version\nPython 3.11.9<\/pre><\/div>\n\n\n\n
    2.2.3 \u9a8c\u8bc1 cmake<\/strong><\/h4>\n\n\n\n
    where cmake\nC:\\Program Files\\CMake\\bin\\cmake.exe\n\ncmake --version\ncmake version 3.29.3\n\nCMake suite maintained and supported by Kitware (kitware.com\/cmake).<\/pre><\/div>\n\n\n\n
    2.2.4 \u9a8c\u8bc1 go<\/strong><\/h4>\n\n\n\n
    where go\nC:\\Program Files\\Go\\bin\\go.exe\n\ngo version\ngo version go1.22.3 windows\/amd64<\/pre><\/div>\n\n\n\n
    2.2.5 \u9a8c\u8bc1\u4f8b\u5b50<\/strong><\/h4>\n\n\n\n
    \u8fd9\u4e00\u6b65\u53ef\u4ee5\u4e0d\u505a\uff0c\u5982\u679c\u987a\u5229\u7684\u8bdd<\/strong><\/p>\n\n\n\n
    \u901a\u8fc7\u9a8c\u8bc1 .\\examples\\langchain-document\\main.py \u6765\u8fbe\u5230\u9a8c\u8bc1\u8fc7\u7a0b<\/p>\n\n\n\n
    \u6211\u4eec\u521b\u5efa\u4e00\u4e2a\u73af\u5883<\/strong>\uff0c\u8fd9\u91cc\u7528conda<\/p>\n\n\n\n
    conda create -yn ollama-gpu\nconda activate ollama-gpu\n\ncd ollama\ncd examples\\langchain-python-rag-document\n\npip install -r requirements.txt\n<\/pre><\/div>\n\n\n\n
    \u53ef\u80fd\u4f1a\u51fa\u73b0\u8b66\u544a\uff1a<\/p>\n\n\n\n
    ERROR: Could not find a version that satisfies the requirement tensorflow-macos==2.13.0 (from versions: none)\nERROR: No matching distribution found for tensorflow-macos==2.13.0<\/pre><\/div>\n\n\n\n
    \u8fd9\u4e2a\u662f\u9488\u5bf9 macos\u7684\uff0c\u4e0d\u7528\u7406\u3002\u6ce8\u91ca\u6389 requirements.txt \u4e2d\u62a5\u9519\u7684\u90e8\u5206,\u4e0b\u9762\u662f\u6ce8\u91ca\u7684\u90e8\u5206<\/p>\n\n\n\n
    #pdfminer==20191125\n#tensorflow-macos==2.13.0\n#uvloop==0.17.0\n<\/pre><\/div>\n\n\n\n
    \u53ef\u80fd\u8fd8\u9700\u8981\u5b89\u88c5\u522b\u7684\u5305\uff0c\u6839\u636e\u8fd0\u884c\u4e0b\u9762\u7684 main.py \u589e\u52a0\u5c31\u53ef\u4ee5\u3002<\/p>\n\n\n\n
    \u8fd0\u884c main.py<\/strong><\/p>\n\n\n\n
    python main.py<\/pre><\/div>\n\n\n\n
    \u6b63\u5e38\u8fd0\u884c\u540e\uff0c\u5e94\u8be5\u6ca1\u6709\u4efb\u4f55\u8b66\u544a\uff0c\u5728\u67e5\u8be2\u540e\u9762\u8f93\u5165\uff1a<\/p>\n\n\n\n
    How many locations does WeWork have?<\/pre>\n\n\n\n
    python main.py\n\nQuery: How many locations does WeWork have?\nAs of June 2023, WeWork has 777 locations around the world.<\/pre><\/div>\n\n\n\n
    3. \u7f16\u8bd1 ollama<\/strong><\/h2>\n\n\n\n
    \u6253\u5f00 Windows PowerShell<\/p>\n\n\n\n
    cd ollama\n\n$env:CGO_ENABLED=\"1\"\n$env:GIN_MODE=\"release\"\n\ngo generate .\\...\ngo build .\n<\/pre><\/div>\n\n\n\n
    \u5982\u679c\u4e0b\u8f7d\u5f88\u6162\uff0c\u53ef\u4ee5\u4f7f\u7528\u56fd\u5185\u7684 golang \u4ee3\u7406<\/p>\n\n\n\n
    $env:GOPROXY = \"https:\/\/goproxy.cn,direct\"<\/pre><\/div>\n\n\n\n
    4. \u6d4b\u8bd5\u7f16\u8bd1\u540e\u7684\u7ed3\u679c<\/strong><\/h2>\n\n\n\n
    <\/p>\n\n\n\n
    ls -l .\\ollama.exe\n\n\n    \u76ee\u5f55: E:\\ollama-gpu\\ollama\n\n\nMode                 LastWriteTime         Length Name\n----                 -------------         ------ ----\n-a----        2024-05-12      5:25       64805673 ollama.exe<\/pre><\/div>\n\n\n\n
    \u901f\u5ea6\u5feb\u5f88\u591a<\/p>\n\n\n\n
    5. \u7f16\u8bd1 ollama CLI, ollama app \u548c ollama installer<\/strong><\/h2>\n\n\n\n
    \u9700\u8981\u4e0b\u8f7d https:\/\/jrsoftware.org\/isinfo.php<\/a><\/p>\n\n\n\n
    \u70b9\u51fb \uff0cDownload Inno Setup<\/a> \u8fdb\u5165\u5230 https:\/\/jrsoftware.org\/isdl.php<\/a><\/p>\n\n\n\n
    \u70b9\u51fb https:\/\/jrsoftware.org\/download.php\/is.exe<\/a><\/p>\n\n\n\n
    \u7136\u540e\u6267\u884c\u7f16\u8bd1\u547d\u4ee4\uff1a<\/p>\n\n\n\n
    powershell -ExecutionPolicy Bypass -File .\\scripts\\build_windows.ps1<\/pre><\/div>\n\n\n\n
    \u5728 dist \u76ee\u5f55\u4e0b\u4f1a\u53c8\u4e24\u4e2a\u6587\u4ef6<\/p>\n\n\n\n
    OllamaSetup.exe<\/strong>\uff0c \u5b89\u88c5\u7a0b\u5e8f<\/p>\n\n\n\n
    ollama-windows-amd64.zip<\/strong> \u538b\u7f29\u5305<\/p>\n\n\n\n
    \u5efa\u8bae\u624b\u52a8\u5148\u5220\u9664\u539f\u6765\u7684\u5b89\u88c5\u5305\uff0c\u4e0d\u7136\uff0c\u53ef\u80fd\u8fd8\u4f1a\u8c03\u7528\u539f\u5148\u7684\u5305\u3002\u5b89\u88c5\u8def\u5f84\u5728\u4f60\u7684\u7528\u6237\u540d\u4e0b\uff0c\u628aUserName \u66ff\u6362\u4e3a\u4f60\u7684\u7528\u6237\u540d\u3002<\/p>\n\n\n\n
    C:\\Users\\UserName\\AppData\\Local\\Programs\\Ollama<\/p>\n\n\n\n
    ollama \u7684\u65e5\u5fd7<\/strong><\/p>\n\n\n\n
    C:\\Users\\UserName\\AppData\\Local\\Ollama\\ \u76ee\u5f55\u4e0b<\/p>\n\n\n\n
    <\/p>\n","protected":false},"excerpt":{"rendered":"
    1. \u8bbe\u7f6e 1.1  NVidia \u9a71\u52a8\u7a0b\u5e8f\u548cCUDA \u5de5\u5177\u5305 1.1.1  NVidia […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[444],"tags":[461,433],"class_list":["post-3741","post","type-post","status-publish","format-standard","hentry","category-ai","tag-gpu","tag-ollama"],"views":9205,"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts\/3741","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=3741"}],"version-history":[{"count":51,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts\/3741\/revisions"}],"predecessor-version":[{"id":3991,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts\/3741\/revisions\/3991"}],"wp:attachment":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=3741"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=3741"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=3741"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}