Skip to content

Latest commit

 

History

History
339 lines (303 loc) · 8.67 KB

Installation_guide.md

File metadata and controls

339 lines (303 loc) · 8.67 KB
Raw

Install Caffe2, OpenCV3, FFMPEG

CentOS

Install OpenCV 3.4.0

  1. Install and create a virtualenv
sudo yum install python-virtualenv  
cd ~/local  
virtualenv r21d  
source r21d/bin/activate  
pip install pip setuptools -U
  1. cmake
    Run cmake --version to see what version you have, If it doesn't exist or less than 3.7, run
sudo yum remove cmake3
sudo yum install cmake
  1. Get OpenCV
wget https://github.com/opencv/opencv/archive/3.4.0.zip -O opencv-3.4.0.zip
unzip opencv-3.4.0.zip
cd opencv-3.4.0
mkdir build && cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE \
	-D CMAKE_INSTALL_PREFIX=/usr/local \
	-D BUILD_EXAMPLES=ON \
	-D BUILD_SHARED_LIBS=ON ..
make -j8
sudo make install
sudo ldconfig
  • If you’ve reached this step without an error, OpenCV should now be installed in
    /usr/local/lib/python2.7/site-packages
  • To use OpenCV within r21d virtual environment 
    cd ~/local/r21d/lib/python2.7/site-packages/
ln -s /usr/local/lib/python2.7/site-packages/cv2.so cv2.so
  • To confirm your installation (ensure that you are in the r21d virtual environment) 
    $ python
>>> import cv2
>>> cv2.__version__
'3.4.0'

Install FFmpeg

  1. Get the dependencies
sudo yum install autoconf automake bzip2 freetype-devel gcc gcc-c++ git libtool pkgconfig zlib-devel yasm-devel libtheora-devel libvorbis-devel libX11-devel gtk2-devel
  1. H.264 video encoder
cd ~/local
git clone http://git.videolan.org/git/x264.git
cd x264
./configure --enable-shared --enable-pic 
make -j8
sudo make install
  1. Ogg bitstream library
cd ~/local
curl -O -L http://downloads.xiph.org/releases/ogg/libogg-1.3.3.tar.gz
tar xzvf libogg-1.3.3.tar.gz
cd libogg-1.3.3
./configure
make -j8
sudo make install
  1. FFmpeg
git clone http://git.videolan.org/git/ffmpeg.git
cd ffmpeg
 ./configure --enable-gpl --enable-nonfree --enable-libtheora --enable-libvorbis  --enable-libx264  --enable-postproc --enable-version3 --enable-pthreads --enable-shared --enable-pic
make -j8
sudo make install
  • Make sure pkg config and the linker can see ffmpeg 
    export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH 
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH
hash -r 
sudo ldconfig
  • To confirm your installation 
    $ ffmpeg
ffmpeg version N-91042-ga3a6d4d Copyright (c) 2000-2018 the FFmpeg developers
built with gcc 4.8.5 (GCC) 20150623 (Red Hat 4.8.5-16)
configuration: --enable-gpl --enable-nonfree --enable-libtheora --enable-libvorbis --enable-libx264 --enable-postproc --enable-version3 --enable-pthreads --enable-shared --enable-pic

Install Caffe2

  1. Get the dependencies
sudo yum install -y protobuf-devel leveldb-devel snappy-devel opencv-devel lmdb-devel python-devel gflags-devel glog-devel kernel-devel
  1. Install cuDNN
cp cuda/lib64/* /usr/local/cuda/lib64/
cp cuda/include/cudnn.h /usr/local/cuda/include/
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
  • Make sure nvcc is runnable 
    $ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2016 NVIDIA Corporation
Built on Sun_Sep__4_22:14:01_CDT_2016
Cuda compilation tools, release 8.0, V8.0.44
  1. Install Python dependencies
sudo pip install lmdb numpy flask future graphviz hypothesis jupyter matplotlib protobuf pydot python-nvd3 pyyaml requests scikit-image scipy six tornado
  1. Build Caffe2
cd ~/local
git clone --recursive https://github.com/pytorch/pytorch.git
cd pytorch && git submodule update --init
  • Modify CMakeLists.txt to make USE_FFMPEG ON
  • After cmake (see below), check the output log, makesure USE_OPENCV: ON and USE_FFMPEG: ON
mkdir build
cd build
cmake ..
sudo make -j8 install

export PYTHONPATH=$PYTHONPATH:~/local/pytorch/build
  • To confirm your installation
python -c 'from caffe2.python import core' 2>/dev/null && echo "Success" || echo "Failure"

Ubuntu

Install OpenCV 3.4.0

  1. Update apt-get and pre-installed libraries
sudo apt-get update
sudo apt-get upgrade
  1. Install some tools
sudo apt-get install build-essential cmake pkg-config libatlas-base-dev gfortran unzip
sudo apt-get install python2.7-dev
sudo apt install python-pip
  1. Setup virtualenv
sudo pip install virtualenv virtualenvwrapper
echo -e "\n# virtualenv and virtualenvwrapper" >> ~/.bashrc
echo "export WORKON_HOME=$HOME/.virtualenvs" >> ~/.bashrc
echo "source /usr/local/bin/virtualenvwrapper.sh" >> ~/.bashrc
source ~/.bashrc
mkvirtualenv r21d -p python2
pip install numpy
  • Note that we need to install numpy before compile OpenCV
  • Ensure you are in the correct virtual environment (e.g. r21d)
  1. Get OpenCV
wget https://github.com/opencv/opencv/archive/3.4.0.zip -O opencv-3.4.0.zip
unzip opencv-3.4.0.zip
cd opencv-3.4.0
mkdir build && cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE \
        -D CMAKE_INSTALL_PREFIX=/usr/local \
        -D PYTHON_EXECUTABLE=~/.virtualenvs/r21d/bin/python \
        -D BUILD_EXAMPLES=ON \
	-D BUILD_SHARED_LIBS=ON ..
make -j8
sudo make install
sudo ldconfig
  • If you’ve reached this step without an error, OpenCV should now be installed in
    /usr/local/lib/python2.7/site-packages
  • To use OpenCV within r21d virtual environment 
    cd ~/.virtualenvs/r21d/lib/python2.7/site-packages/
ln -s /usr/local/lib/python2.7/site-packages/cv2.so cv2.so
  • To confirm your installation (ensure that you are in the r21d virtual environment) 
    $ python
>>> import cv2
>>> cv2.__version__
'3.4.0'

Install FFmpeg

  1. Get the dependencies
sudo apt-get update -qq && sudo apt-get -y install \
  autoconf \
  automake \
  build-essential \
  cmake \
  git-core \
  libass-dev \
  libfreetype6-dev \
  libsdl2-dev \
  libtool \
  libva-dev \
  libvdpau-dev \
  libvorbis-dev \
  libxcb1-dev \
  libxcb-shm0-dev \
  libxcb-xfixes0-dev \
  pkg-config \
  texinfo \
  wget \
  zlib1g-dev

sudo apt-get install yasm libx264-dev
  1. In your home directory make a new directory to put all of the source code into:
mkdir -p ~/ffmpeg_sources ~/bin
  1. NASM
cd ~/ffmpeg_sources && \
wget https://www.nasm.us/pub/nasm/releasebuilds/2.13.03/nasm-2.13.03.tar.bz2 && \
tar xjvf nasm-2.13.03.tar.bz2 && \
cd nasm-2.13.03 && \
./autogen.sh && \
PATH="$HOME/bin:$PATH" ./configure --prefix="$HOME/ffmpeg_build" --bindir="$HOME/bin" && \
make && \
make install
  1. FFmpeg
cd ~/ffmpeg_sources && \
wget -O ffmpeg-snapshot.tar.bz2 https://ffmpeg.org/releases/ffmpeg-snapshot.tar.bz2 && \
tar xjvf ffmpeg-snapshot.tar.bz2 && \
cd ffmpeg && \
PATH="$HOME/bin:$PATH" PKG_CONFIG_PATH="$HOME/ffmpeg_build/lib/pkgconfig" ./configure \
  --prefix="$HOME/ffmpeg_build" \
  --pkg-config-flags="--static" \
  --extra-cflags="-I$HOME/ffmpeg_build/include" \
  --extra-ldflags="-L$HOME/ffmpeg_build/lib" \
  --extra-libs="-lpthread -lm" \
  --bindir="$HOME/bin" \
  --enable-gpl \
  --enable-libvorbis \
  --enable-libx264 \
  --enable-nonfree
PATH="$HOME/bin:$PATH" 
make -j8
make install
  • Make sure pkg config and the linker can see ffmpeg 
    echo "export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH" >> ~/.bashrc
echo "export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH" >> ~/.bashrc
hash -r 
sudo ldconfig
  • To confirm your installation 
    $ ffmpeg
ffmpeg version N-91283-gdaf38d0 Copyright (c) 2000-2018 the FFmpeg developers
built with gcc 5.4.0 (Ubuntu 5.4.0-6ubuntu1~16.04.9) 20160609

Install Caffe2

  1. Get the dependencies
sudo apt-get install -y \
      libgoogle-glog-dev \
      libgtest-dev \
      libiomp-dev \
      libleveldb-dev \
      liblmdb-dev \
      libopencv-dev \
      libopenmpi-dev \
      libsnappy-dev \
      libprotobuf-dev \
      protobuf-compiler \
      libgflags-dev \
      python-dev

pip install lmdb flask future graphviz hypothesis jupyter matplotlib protobuf pydot python-nvd3 pyyaml requests scikit-image scipy six tornado
  1. Build Caffe2
  • If you have a GPU, consider reference to the steps. (note that we didn't test it in this tutorial)
cd ~
git clone --recursive https://github.com/pytorch/pytorch.git
cd pytorch && git submodule update --init

USE_OPENCV=1 USE_FFMPEG=1 USE_LMDB=1 python setup.py install

export PYTHONPATH=$PYTHONPATH:$HOME/pytorch/build
  • To confirm your installation
python -c 'from caffe2.python import core' 2>/dev/null && echo "Success" || echo "Failure"