美文网首页
使用OpenBLAS集成编译numpy

使用OpenBLAS集成编译numpy

作者: 梁间 | 来源:发表于2018-11-26 17:50 被阅读0次

    遇到这样一个错误

    /usr/local/lib/python2.7/site-packages/chainer/_environment_check.py:38: UserWarning: Accelerate has been detected as a NumPy backend library.
    vecLib, which is a part of Accelerate, is known not to work correctly with Chainer.
    We recommend using other BLAS libraries such as OpenBLAS.
    For details of the issue, please see
    https://docs.chainer.org/en/stable/tips.html#mnist-example-does-not-converge-in-cpu-mode-on-mac-os-x.
    

    需要使用OpenBLAS集成编译numpy
    首先安装openblas

    $ brew install openblas
    
    $ cd ~/github
    $ git clone git://github.com/numpy/numpy.git numpy
    $ cd numpy
    $ cp site.cfg.example site.cfg
    

    编辑site.cfg

    [ALL]
    library_dirs = /usr/local/opt/openblas/lib
    include_dirs = /usr/local/opt/openblas/include
    [atlas]
    library_dirs = openblas
    include_dirs = openblas
    [openblas]
    libraries = openblas
    library_dirs = /usr/local/opt/openblas/lib
    include_dirs = /usr/local/opt/openblas/include
    runtime_library_dirs = /usr/local/opt/openblas/lib
    

    编辑 ~/.bash_profile 添加

    export LDFLAGS=-I/usr/local/opt/openblas/lib
    export CPPFLAGS=-L/usr/local/opt/openblas/include
    export PKG_CONFIG_PATH=/usr/local/opt/openblas/lib/pkgconfig
    

    执行

    python setup.py build && python setup.py install
    

    出现错误提示

    fatal error: 'endian.h' file not found
    

    首先需要安装Command Line Tools (macOS 10.X) for XCode 10.X

    xcode-select --install
    

    MacOS 10.14之前版本,同时会自动生成"/usr/include"文件, 包括 "/usr/include/machine/endian.h".
    MacOS 10.14起,文件不会自动升成,需要手动安装 /Library/Developer/CommandLineTools/Packages/macOS_SDK_headers_for_macOS_10.14.pkg

    如果出现权限不足错误

    Adding numpy 1.16.0.dev0+11a316e to easy-install.pth file
    error: [Errno 13] Permission denied: '/usr/local/lib/python2.7/site-packages/easy-install.pth'
    

    可以改用

    sudo python setup.py build && sudo python setup.py install
    

    相关文章

      网友评论

          本文标题:使用OpenBLAS集成编译numpy

          本文链接:https://www.haomeiwen.com/subject/ftvbqqtx.html