This is a log Josh created while following this link: https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/doc/installation.md#operating-systems FAQ: https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/doc/faq.md
- installed GitHub desktop
- cloned most recent open pose master commit
- want the added python functionality
- install Visual Studio 2015 update 3
- install cmake
- install python 3.7 in default location
- add to path
- pip install numpy opencv-python
- download cuda 8.0 and patch 2
- https://developer.nvidia.com/cuda-80-ga2-download-archive
- install cuda toolkit 8.0
- install patch 2
- if CUDA 8 does not work because of the type of graphics card, try CUDA 9
- download cuDNN 5.1
- unzip folder
- copy include and lib contents into CUDA\v8.0 folder
- cmake will downloads the models
- cmake also checks for the zip files of opencv, caffe, and its dependencies in the 3rd party folder
- configure and generate with cmake
- When asked what to use as the generator, pick “Visual Studio 14 2015 win64”
- the config option ‘gpu mode’ is CUDA or CPU_ONLY depending on what you want
- open the VS solution, openpose.sln, in the build directory
- select the Release and x64 configuration, not debug or x86 (32bit) ones
- in the right hand pane, right click the top folder and select build
- also build the python wrapper (pyopenpose)
- right click the pyopenpose folder at the bottom of the listed files and select build
- In order to use OpenPose outside Visual Studio, and assuming you have not unchecked the BUILD_BIN_FOLDER flag in CMake, copy all DLLs from {build_directory}/bin into the folder where the generated openpose.dll and *.exe demos are, e.g., {build_directory}/x64/Release for the 64-bit release version
- do this after every build