zoukankan      html  css  js  c++  java
  • Caffe source install

    If you use centos: try the command below, which can automatically install the requirements for caffe
    author belong to wangfei@intel.
    yum install epel-release
    Note:
    For make : PREFIX=/your/install/dir make install -n
    For cmake: cmake -DCMAKE_INSTALL_PREFIX=/your/install/dir
    Os: CentOS 7.2
    To install of the software, please refer to the readme of corresponding github repo.

    1. Install cmake 3.6 from github
      git clone https://github.com/Kitware/CMake.git Cmake
      cd Cmake
      ./bootstrap –prefix=<intall_prefix> && make && make install

    2. Install opencv 2.4.13 from github
      git clone https://github.com/opencv/opencv.git Opencv
      cd Opencv
      mkdir release
      cd release
      cmake –D CMAKE_BUILD_TYPE=RELEASE –D CMAKE_INSTALL_PREFIX=/your/install/dir ..
      make && make install

    3. Install protobuf 2.7.0 from github
      git clone https://github.com/google/protobuf.git Protobuf
      cd Protobuf
      ./autogen.sh
      ./configure –prefix=<install_prefix>
      make
      make check
      make install
      cd python
      python setup.py build
      python setup.py install

    4. Install glog 0.3.4 from github
      git clone https://github.com/google/glog.git Glog
      cd Glog
      ./configure –prefix=<install_prefix>
      make && make install

    5. Install gflags2.1.2 from github
      git clone https://github.com/gflags/gflags.git Gflags
      cd Gflags
      mkdir build
      cd build
      cmake –D CMAKE_INSTALL_PREFIX=/your/install/dir ..
      make && make install

    6. Install lmdb 0.9 from github
      git clone https://github.com/LMDB/lmdb.git Lmdb
      cd Lmdb
      make
      PREFIX=/your/install/dir make install –n

    7. Install boost 1.61 from http://www.boost.org/users/history/version_1_61_0.html
      ./bootstrap.sh
      ./b2
      cp boost /usr/local/include
      cp stage/lib/* /usr/local/lib64

    8. Install zlib 1.2.5 from github
      git clone https://github.com/madler/zlib.git Zlib
      ./configure --prefix=<install_dir>
      make && make install

    9. Install snappy 1.1.3 from github
      git clone https://github.com/google/snappy.git Snappy
      cd Snappy
      ./autogen
      ./configure –prefix=<install_dir>
      make && make install

    10. Install leveldb 1.14 from github
      Git clone https://github.com/google/leveldb.git Leveldb
      cd Leveldb
      make
      PREFIX=/your/install/dir make install –n

    11. Install hdf5 (1.8.17for centos) from https://support.hdfgroup.org/downloads/index.html
      cp include/* <.h dir>
      cp lib/* <lib_dir>

    12. Add to ~./bashrc
      export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
      export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib64

    13. Install intel-caffe and pycaffe
      cp Makefile.config.example Makefile.config
      uncomment CPU_ONLY and OPENCV_VERSION and ALLOW_LMDB_NOLOCK
      change the mkl dir and python dir to your own’s
      make
      make runtest
      make pycaffe

    14. export PYTHONPATH=/home/2T/intel-caffe/python:$PYTHONPATH
      Add to ~/.bashrc

    !/bin/sh

    source ~/.bashrc

    source /home/xiaotian/tools/__release_lnx/mkl/bin/mklvars.sh intel64

    KNL

    export LD_LIBRARY_PATH=/usr/local/cuda/lib64:/home/xiaotian/boost_1_62_0/stage/lib:$LD_LIBRARY_PATH

    export LIBS=" -L/home/xiaotian/tools/mkl-2017u2-1/__release_lnx/mkl/lib/intel64 -L/usr/local/cuda/lib64"

    export CPPFLAGS="-I/home/xiaotian/tools/mkl-2017u2-1/__release_lnx/mkl/include -I/usr/local/cuda/include"

    export LIBS="-L/home/xiaotian/boost_1_62_0/stage/lib -L/home/xiaotian/tools/mkl-2017u2-1/__release_lnx/mkl/lib/intel64 -L/usr/local/cuda/lib64"
    export CPPFLAGS="-I/home/xiaotian/boost_1_62_0 -I/home/xiaotian/tools/mkl-2017u2-1/__release_lnx/mkl/include -I/usr/local/cuda/include"

    export CPLUS_INCLUDE_PATH=/home/xiaotian/boost_1_62_0:$CPLUS_INCLUDE_PATH
    make clean

    make -j44
    make py
    ldd build/tools/caffe

  • 相关阅读:
    Loj #6560 小奇取石子
    某谷 P5153 简单的函数
    某谷 P5159 WD与矩阵
    前端ajax访问 django 报错 POST http://127.0.0.1:8001/xxx 403 (Forbidden)
    python
    Java
    Java
    Java
    Java
    java web 向数据库插入中文数据乱码问题
  • 原文地址:https://www.cnblogs.com/Viewsky/p/6714584.html
Copyright © 2011-2022 走看看