xu-chris / vnect-lib-windows-installation-guide

This summarizes my learnings when installing the VNect Pose Estimation Lib on Windows

VNect Lib installation guide for Windows

VNect is a promising pose estimation library written and developed by Dushyant Mehta and many other colleagues on the Max Planck Institute, Saarland University and Universidad Rey Juan Carlos. It generates a 3D pose estimation of a person from a given 2D image in real time either live or offline.

Through their website, you can request access to their library. Note, that they're working on a revised system, called XNect which seems to be super robust. This guide is concerning only VNect (for now).

---

NOTE Note, that this library is only working with NVIDIA CUDA, not with OpenCL or Metal due to the dependency on Caffe v1 (which supports only CUDA for now).

Building dependencies

VNect uses Caffe for Windows to build it up. This contains by itself again multiple dependencies. We will go through each one of these, including:

1. Install Visual Studio 14 (2015)

1. Go to the Visul Studio version archive and download Visual Studio 14 (2015)
2. Install it

2. Installing CUDA 8.0

1. Go to NVIDIA page and download CUDA 8.0
2. Install it

3. Install cuDNN

For this you will need an NVIDIA developer account which you can easily create during the process.

1. Go to NVIDIA developer archive for cuDNN and download cuDNN v7.1.4 (May 16, 2018), for CUDA 8.0 and Windows 10
2. Unzip the ZIP archive
3. Copy the following files into the CUDA Toolkit directory (according to this guide).
   1. Copy <installpath>\cuda\bin\cudnn*.dll to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin.
   2. Copy <installpath>\cuda\include\cudnn*.h to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\include.
   3. Copy <installpath>\cuda\lib\x64\cudnn*.lib to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\lib\x64

4. Install Anaconda (for Python 2.7)

Caffe is quite antique and VNect is not built completely for the future, so we continue here digging in archives of projects.

1. Go to Anaconda page and scroll to the very bottom. Select Python 2.7 64-Bit Graphical Installer for Windows. Alternatively, you can use this shortcut which might have an outdated version linked (but it's a shortcut, isn't it?)

• Alternatively, you can also install Anaconda 3 for later use if you think you will use this.

2. Install Anaconda

5. Install git (if not given)

1. Go to the official Git installer download page and download Git
2. Install git

6. Set your environment variables correct

We will now check if the libraries we use can find each other.

1. Open Windows Menu (the Windows flag) and enter Environment Variables. You should be able to select Edit the system environment variables. Select it
2. Click on Environment variables
3. Check the following system variables, following in the scheme Variable name : Variable value:
   • CUDA_PATH : C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0
   • CUDA_PATH_V8_0 : C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0
   • CUDNN_PATH : C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\8.0
   • CuDnnPath : C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\8.0
   • Path:
     • C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin
     • C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\libnvvp
     • C:\Python27\
     • C:\Python27\Scripts
   • VS140COMNTOOLS : C:\Program Files (x86)\Microsoft Visual Studio 14.0\Common7\Tools\
4. Confirm all changes and if you have the Powershell open, restart it.

7. Install Chocolatey library manager

We will follow the guide from Chocolatey

1. Open Powershell as admin
2. Install Chocolatey by entering the following command:

Set-ExecutionPolicy Bypass -Scope Process -Force; [System.Net.ServicePointManager]::SecurityProtocol = [System.Net.ServicePointManager]::SecurityProtocol -bor 3072; iex ((New-Object System.Net.WebClient).DownloadString('https://chocolatey.org/install.ps1'))

3. Check your installation by typing the following command into the Powershell:

choco

8. Install cmake

Entering the following command in your Powershell window:

choco install -y cmake

9. Install Caffe

We will install Caffe without the Ninja builder because Caffe is antique and won't be updated to support the changed API of newer Ninja versions. This removes luckily a couple of other dependencies for now (which is good) but leads to a very long building time of Caffe (which is bad).

1. Open Powershell with Admin privilidges (right click -> Run as Administrator)
2. Fetch the Caffe Git repository

git clone https://github.com/BVLC/caffe.git

3. Go into the repository and checkout the Windows branch

cd caffe
git checkout windows

Don't close the Powershell. We will need it later! 4. Open the script scripts/build_win.cmd with an editor of your choice. 5. Replace everything with the following code:

@echo off
@setlocal EnableDelayedExpansion

:: Default values
if DEFINED APPVEYOR (
    echo Setting Appveyor defaults
    if NOT DEFINED MSVC_VERSION set MSVC_VERSION=14
    if NOT DEFINED WITH_NINJA set WITH_NINJA=0
    if NOT DEFINED CPU_ONLY set CPU_ONLY=0
    if NOT DEFINED CUDA_ARCH_NAME set CUDA_ARCH_NAME=Auto
    if NOT DEFINED CMAKE_CONFIG set CMAKE_CONFIG=Release
    if NOT DEFINED USE_NCCL set USE_NCCL=0
    if NOT DEFINED CMAKE_BUILD_SHARED_LIBS set CMAKE_BUILD_SHARED_LIBS=0
    if NOT DEFINED PYTHON_VERSION set PYTHON_VERSION=2
    if NOT DEFINED BUILD_PYTHON set BUILD_PYTHON=1
    if NOT DEFINED BUILD_PYTHON_LAYER set BUILD_PYTHON_LAYER=1
    if NOT DEFINED BUILD_MATLAB set BUILD_MATLAB=0
    if NOT DEFINED PYTHON_EXE set PYTHON_EXE=python
    if NOT DEFINED RUN_TESTS set RUN_TESTS=1
    if NOT DEFINED RUN_LINT set RUN_LINT=1
    if NOT DEFINED RUN_INSTALL set RUN_INSTALL=1

    :: Set python 2.7 with conda as the default python
    if !PYTHON_VERSION! EQU 2 (
        set CONDA_ROOT=C:\Miniconda-x64
    )
    :: Set python 3.5 with conda as the default python
    if !PYTHON_VERSION! EQU 3 (
        set CONDA_ROOT=C:\Miniconda35-x64
    )
    set PATH=!CONDA_ROOT!;!CONDA_ROOT!\Scripts;!CONDA_ROOT!\Library\bin;!PATH!

    :: Check that we have the right python version
    !PYTHON_EXE! --version
    :: Add the required channels
    conda config --add channels conda-forge
    conda config --add channels willyd
    :: Update conda
    conda update conda -y
    :: Download other required packages
    conda install --yes cmake ninja numpy scipy protobuf==3.1.0 six scikit-image pyyaml pydotplus graphviz

    if ERRORLEVEL 1  (
      echo ERROR: Conda update or install failed
      exit /b 1
    )

    :: Install cuda and disable tests if needed
    if !WITH_CUDA! == 1 (
        call %~dp0\appveyor\appveyor_install_cuda.cmd
        set CPU_ONLY=0
        set RUN_TESTS=0
        set USE_NCCL=1
    ) else (
        set CPU_ONLY=1
    )

    :: Disable the tests in debug config
    if "%CMAKE_CONFIG%" == "Debug" (
        echo Disabling tests on appveyor with config == %CMAKE_CONFIG%
        set RUN_TESTS=0
    )

    :: Disable linting with python 3 until we find why the script fails
    if !PYTHON_VERSION! EQU 3 (
        set RUN_LINT=0
    )

) else (
    :: Change the settings here to match your setup
    :: Change MSVC_VERSION to 12 to use VS 2013
    if NOT DEFINED MSVC_VERSION set MSVC_VERSION=14
    :: Change to 1 to use Ninja generator (builds much faster)
    if NOT DEFINED WITH_NINJA set WITH_NINJA=0
    :: Change to 1 to build caffe without CUDA support
    if NOT DEFINED CPU_ONLY set CPU_ONLY=0
    :: Change to generate CUDA code for one of the following GPU architectures
    :: [Fermi  Kepler  Maxwell  Pascal  All]
    if NOT DEFINED CUDA_ARCH_NAME set CUDA_ARCH_NAME=Auto
    :: Change to Debug to build Debug. This is only relevant for the Ninja generator the Visual Studio generator will generate both Debug and Release configs
    if NOT DEFINED CMAKE_CONFIG set CMAKE_CONFIG=Release
    :: Set to 1 to use NCCL
    if NOT DEFINED USE_NCCL set USE_NCCL=1
    :: Change to 1 to build a caffe.dll
    if NOT DEFINED CMAKE_BUILD_SHARED_LIBS set CMAKE_BUILD_SHARED_LIBS=0
    :: Change to 3 if using python 3.5 (only 2.7 and 3.5 are supported)
    if NOT DEFINED PYTHON_VERSION set PYTHON_VERSION=2
    :: Change these options for your needs.
    if NOT DEFINED BUILD_PYTHON set BUILD_PYTHON=1
    if NOT DEFINED BUILD_PYTHON_LAYER set BUILD_PYTHON_LAYER=1
    if NOT DEFINED BUILD_MATLAB set BUILD_MATLAB=0
    :: If python is on your path leave this alone
    if NOT DEFINED PYTHON_EXE set PYTHON_EXE=python
    :: Run the tests
    if NOT DEFINED RUN_TESTS set RUN_TESTS=0
    :: Run lint
    if NOT DEFINED RUN_LINT set RUN_LINT=0
    :: Build the install target
    if NOT DEFINED RUN_INSTALL set RUN_INSTALL=1
)

:: Set the appropriate CMake generator
:: Use the exclamation mark ! below to delay the
:: expansion of CMAKE_GENERATOR
if %WITH_NINJA% EQU 0 (
    if "%MSVC_VERSION%"=="14" (
        set CMAKE_GENERATOR=Visual Studio 14 2015 Win64
    )
    if "%MSVC_VERSION%"=="12" (
        set CMAKE_GENERATOR=Visual Studio 12 2013 Win64
    )
    if "!CMAKE_GENERATOR!"=="" (
        echo ERROR: Unsupported MSVC version
        exit /B 1
    )
) else (
    set CMAKE_GENERATOR=Ninja
)

echo INFO: ============================================================
echo INFO: Summary:
echo INFO: ============================================================
echo INFO: MSVC_VERSION               = !MSVC_VERSION!
echo INFO: WITH_NINJA                 = !WITH_NINJA!
echo INFO: CMAKE_GENERATOR            = "!CMAKE_GENERATOR!"
echo INFO: CPU_ONLY                   = !CPU_ONLY!
echo INFO: CUDA_ARCH_NAME             = !CUDA_ARCH_NAME!
echo INFO: CMAKE_CONFIG               = !CMAKE_CONFIG!
echo INFO: USE_NCCL                   = !USE_NCCL!
echo INFO: CMAKE_BUILD_SHARED_LIBS    = !CMAKE_BUILD_SHARED_LIBS!
echo INFO: PYTHON_VERSION             = !PYTHON_VERSION!
echo INFO: BUILD_PYTHON               = !BUILD_PYTHON!
echo INFO: BUILD_PYTHON_LAYER         = !BUILD_PYTHON_LAYER!
echo INFO: BUILD_MATLAB               = !BUILD_MATLAB!
echo INFO: PYTHON_EXE                 = "!PYTHON_EXE!"
echo INFO: RUN_TESTS                  = !RUN_TESTS!
echo INFO: RUN_LINT                   = !RUN_LINT!
echo INFO: RUN_INSTALL                = !RUN_INSTALL!
echo INFO: ============================================================

:: Build and exectute the tests
:: Do not run the tests with shared library
if !RUN_TESTS! EQU 1 (
    if %CMAKE_BUILD_SHARED_LIBS% EQU 1 (
        echo WARNING: Disabling tests with shared library build
        set RUN_TESTS=0
    )
)

if NOT EXIST build mkdir build
pushd build

:: Setup the environement for VS x64
set batch_file=!VS%MSVC_VERSION%0COMNTOOLS!..\..\VC\vcvarsall.bat
call "%batch_file%" amd64

:: Configure using cmake and using the caffe-builder dependencies
:: Add -DCUDNN_ROOT=C:/Projects/caffe/cudnn-8.0-windows10-x64-v5.1/cuda ^
:: below to use cuDNN
cmake -G"!CMAKE_GENERATOR!" ^
      -DBLAS=Open ^
      -DCMAKE_BUILD_TYPE:STRING=%CMAKE_CONFIG% ^
      -DBUILD_SHARED_LIBS:BOOL=%CMAKE_BUILD_SHARED_LIBS% ^
      -DBUILD_python:BOOL=%BUILD_PYTHON% ^
      -DBUILD_python_layer:BOOL=%BUILD_PYTHON_LAYER% ^
      -DBUILD_matlab:BOOL=%BUILD_MATLAB% ^
      -DCPU_ONLY:BOOL=%CPU_ONLY% ^
      -DCOPY_PREREQUISITES:BOOL=1 ^
      -DINSTALL_PREREQUISITES:BOOL=1 ^
      -DUSE_NCCL:BOOL=!USE_NCCL! ^
      -DCUDA_ARCH_NAME:STRING=%CUDA_ARCH_NAME% ^
      "%~dp0\.."

if ERRORLEVEL 1 (
  echo ERROR: Configure failed
  exit /b 1
)

:: Lint
if %RUN_LINT% EQU 1 (
    cmake --build . --target lint  --config %CMAKE_CONFIG%
)

if ERRORLEVEL 1 (
  echo ERROR: Lint failed
  exit /b 1
)

:: Build the library and tools
cmake --build . --config %CMAKE_CONFIG%

if ERRORLEVEL 1 (
  echo ERROR: Build failed
  exit /b 1
)

:: Build and exectute the tests
if !RUN_TESTS! EQU 1 (
    cmake --build . --target runtest --config %CMAKE_CONFIG%

    if ERRORLEVEL 1 (
        echo ERROR: Tests failed
        exit /b 1
    )

    if %BUILD_PYTHON% EQU 1 (
        if %BUILD_PYTHON_LAYER% EQU 1 (
            :: Run python tests only in Release build since
            :: the _caffe module is _caffe-d is debug
            if "%CMAKE_CONFIG%"=="Release" (
                :: Run the python tests
                cmake --build . --target pytest

                if ERRORLEVEL 1 (
                    echo ERROR: Python tests failed
                    exit /b 1
                )
            )
        )
    )
)

if %RUN_INSTALL% EQU 1 (
    cmake --build . --target install --config %CMAKE_CONFIG%
)

popd
@endlocal

Here's an explanation of the changes:

• if NOT DEFINED MSVC_VERSION set MSVC_VERSION=14: We set the Visual Studio Code version to Visual Studio 14.
• if NOT DEFINED WITH_NINJA set WITH_NINJA=0: We don't use the Ninja builder.
• if NOT DEFINED CPU_ONLY set CPU_ONLY=0: We want to build for the GPU. So disable the CPU only usage.
• if NOT DEFINED USE_NCCL set USE_NCCL=1: Enable building with cuDNN support.
• if NOT DEFINED PYTHON_VERSION set PYTHON_VERSION=2: We use old Python version 2.
• if NOT DEFINED BUILD_PYTHON set BUILD_PYTHON=1: We build the Python library as well.
• if NOT DEFINED BUILD_PYTHON_LAYER set BUILD_PYTHON_LAYER=1: Yeah... let's build the Python Caffe layer as well.
• if NOT DEFINED BUILD_MATLAB set BUILD_MATLAB=0: We don't have Matlab installed.
• if NOT DEFINED RUN_TESTS set RUN_TESTS=0: Disable running tests due to cuDNN support (somehow the makers have disabled the tests then).
• if NOT DEFINED RUN_INSTALL set RUN_INSTALL=1: Install it as global library. All these changes are followed right in the block after ) else (.

6. Go back to your Powershell window. Install additional dependencies for running it with Anaconda. Paste each line separately into the Powershell.

conda config --add channels conda-forge
conda config --add channels willyd
conda install --yes cmake ninja numpy scipy protobuf==3.1.0 six scikit-image pyyaml pydotplus graphviz

7. Build Caffe by entering the following command in the Powershell. (I assume you're still in the caffe/ folder of the Git repository. Double check for it!)

./scripts\build_win.cmd

10. Install VNect library

This is quite straightforward, so you just follow the README.md

1. Download VNect library and place it where you can find it. I recommend for now to place it into <your-home-folder>/Projects/vnect-lib
2. After downloading, open your Powershell as Admin again, navigate to the project folder. When you follow my recommendation, it would be:

cd ~/Projects/vnect-lib

3. Create cmake folder and go into that

mkdir build
cd build

4. Run the cmake command

cmake -G "Visual Studio 14 Win64" -C C:\Users\<your-username>\.caffe\dependencies\libraries_v140_x64_py27_1.1.0\libraries\caffe-builder-config.cmake ..