Linux install the Python face Recognition face_recognition module

System Environment: CentOS6.7

Anaconda Version: anaconda3-5.1.0-linux-x86_64

Python version: Python-3.6.4

1. Yum installation CMake, boost

Yum Install CMake boost

2. Yum Install gcc gcc-c++

Yum Install gcc gcc-c++

Use Yum to install the existing version of the system, and then upgrade to 4.8.1 after uninstalling the Yum version

3. Install Dlib package with Conda

# Conda Install Dlib

[ROOT@HADOOP7 pkgs]# condainstall dlib solving environment:done ==> warning:a newer versionof Conda. <== version:4.4.10 latest version:4.5.1 Please update Conda by running $ conda update-n BA Se Conda # # Package Plan # Environment location:/root/anaconda3 added/updated Specs:-Dlib T            He following packages would bedownloaded:package |                Build---------------------------|-----------------ca-certificates-2018.1.18 | 0 140 KB conda-forge numpy-1.14.2 |py36_blas_openblas_200 9.0 MB Conda-forge opens                SL-1.0.2O |         0 3.5 MB Conda-forge blas-1.1 |                Openblas 1 KB conda-forge sqlite-3.20.1 |           0 1.3 MB Conda-forge certifi-2018.1.18 |    PY36_0 143 KB Conda-forge scipy-1.0.1            |py36_blas_openblas_200 39.8 MB Conda-forge scikit-learn-0.19.1 |py36_blas_openblas_201                12.8 MB Conda-forge libgfortran-3.0.0 |                1 281 KB Https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free openblas-0.2.20 | 7 17.0 MB Conda-forge dlib-19.10 |py36_blas_openblas_200 2.3 MB conda-forge----

 

--------------------------------------------------------total:86.3 MB                                           

    The following NEW packages Willbe Installed:blas:1.1-openblas Conda-forge                                            Dlib:19.10-py36_blas_openblas_200conda-forge [Blas_openblas] Libgfortran:3.0.0-1 https://mirrors.tuna.tsinghua.edu.cn/ Anaconda/pkgs/free openblas:0.2.20-7 CondA-forge the following packages would beremoved:anaconda:5.1 .0-py36_2 defaults The following packages would beupdated : Ca-certificates:2017.08.26-h1d4fec5_0 Defaults--> 2018                                               .1.18-0 Conda-forge Certifi:2018.1.18-py36_0 Defaults --> 2018.1.18-py36_0 conda-forge numpy:1.14.0-py36h3dfced4_1 de

    Faults-->1.14.2-py36_blas_openblas_200 Conda-forge [Blas_openblas]                     Openssl:1.0.2n-hb7f436b_0 Defaults--> 1.0.2o-0                                               Conda-forge scikit-learn:0.19.1-py36h7aa7ec6_0 Defaults --> 0.19.1-py36_blas_openblas_201 Conda-forge [Blas_openblas] scipy:1.0.0-py36hbf646e7_0 defaults --> 1.0.1-py36_blas_openblas_200 Conda-forge [Blas_openblas] the following PAC                                               Kages would bedowngraded:sqlite:3.22.0-h1bed415_0 defaults --> 3.20.1-0 conda-forge Proceed ([y]/n)? Y

You can see that you need to download a lot of other dependencies, and you will automatically download the installation when you enter Y. If download fails, you can manually download the installation file installation

Such as:

CD/PATH/ANACONDA3/PKGS (package default installation path)

wgethttps://ffmpeg.org/releases/ffmpeg-3.4.2.tar.bz2

Conda Install ffmpeg-3.4.2.tar.bz2

4. PIP installation for Face_recgnition

# pip Install Face_recgnition

Possible error: "/usr/lib64/libstdc++.so.6:version ' glibcxx_3.4.19 ' not Found"

Face_recognition module installation, if the error "unsupportedcompiler--PYBIND11 requires c++11 support!", stating that the GCC version is not high enough to upgrade GCC, method:

Download GCC Latest Version

Wgethttp://ftp.gnu.org/gnu/gcc/gcc-4.8.1/gcc-4.8.1.tar.gz

Then unzip to the folder

Tar-xvzf gcc-4.8.1.tar.gz

The directory after the decompression

CD gcc-4.8.1

Then perform the following run

./contrib/download_prerequisites

Go back to the previous layer and set up the build_gcc_4.8.1 directory, which is parallel to the gcc-4.8.1

Cd..

mkdir build_gcc_4.8.1

Enter the newly created directory and execute the compilation process

CD build_gcc_4.8.1

.. /gcc-4.8.1/configure--enable-checking=release--enable-languages=c,c++--disable-multilib

Make-j23

Make install

OK, GCC is already installed in build_gcc_4.8.1

To determine the path of the newly installed GCC, the path to the last MV when you installed it, I am the default/usr/local/bin

Ls/usr/local/bin | grep gcc

Perform:

/usr/sbin/update-alternatives--INSTALL/USR/BIN/GCC gcc/usr/local/bin/x86_64-unknown-linux-gnu-gcc-4.8.1 40

GCC--version #查看版本

/usr/sbin/update-alternatives--install/usr/bin/g++ g++/usr/local/bin/g++ 40

g++--version #查看版本

After the installation of the high version gcc is complete, use the "Yum Remove gcc-c++" and "yum remove gcc" To uninstall the lower version of the GCC compiler

Re-install face_recgnition will succeed

# pip Install Face_recgnition

5. Test Import Package

[ROOT@HADOOP8 pkgs]# python

Python 3.6.4 | Anaconda, inc.| (Default, Jan16 2018, 18:10:19)

[GCC 7.2.0] on Linux

Type ' help ', ' copyright ', ' credits ' or ' license ' for moreinformation.

>>> Import Dlib

>>> Import Face_recognition

>>>exit ()

6. Other installation problems

Possible unexpected conditions for manual Dlib installation:

Error 1:build, the error "The CXX compiler identification is unknown". This situation requires the installation of GCC (as a cxxcompiler), and the issues to be noted are the context.

Error 2:build, the error "could not find Pythonlibs (missing:python_include_dirs)", in this case, the CMake command to specify:

-dpython_include_dir=/usr/python/include/python3.6m (this path should be the installation path that exists for the Python.h file, depending on the machine path)

-dpython_library=/usr/python/lib/libpython3.so (this path should be the installation path that exists for the libpython.**.so file, depending on the machine path)

The complete command is as follows: Enter cmake-dpython_include_dir=/usr/python/include/python3.6m-dpython_library=/usr/under Build path in Dlib source code Python/lib/libpython3.so.. /.. /tools/python/