KPConv-PyTorch/INSTALL.md
2020-04-27 18:46:34 -04:00

2 KiB

Installation instructions

Ubuntu 18.04

  • Make sure CUDA and cuDNN are installed. One configuration has been tested:

    • PyTorch 1.4.0, CUDA 10.1 and cuDNN 7.6
  • Ensure all python packages are installed :

        sudo apt update
        sudo apt install python3-dev python3-pip python3-tk
    
  • Follow PyTorch installation procedure.

  • Install the other dependencies with pip:

    • numpy
    • scikit-learn
    • PyYAML
    • matplotlib (for visualization)
    • mayavi (for visualization)
    • PyQt5 (for visualization)
  • Compile the C++ extension modules for python located in cpp_wrappers. Open a terminal in this folder, and run:

        sh compile_wrappers.sh
    

You should now be able to train Kernel-Point Convolution models

Windows 10

  • Make sure CUDA and cuDNN are installed. One configuration has been tested:

    • PyTorch 1.4.0, CUDA 10.1 and cuDNN 7.5
  • Follow PyTorch installation procedure.

  • We used the PyCharm IDE to pip install all python dependencies (including PyTorch) in a venv:

    • torch
    • torchvision
    • numpy
    • scikit-learn
    • PyYAML
    • matplotlib (for visualization)
    • mayavi (for visualization)
    • PyQt5 (for visualization)
  • Compile the C++ extension modules for python located in cpp_wrappers. You just have to execute two .bat files:

      cpp_wrappers/cpp_neighbors/build.bat
    

    and

      cpp_wrappers/cpp_subsampling/build.bat
    

You should now be able to train Kernel-Point Convolution models