Skip to content

Latest commit

 

History

History
140 lines (115 loc) · 4.6 KB

installation.md

File metadata and controls

140 lines (115 loc) · 4.6 KB

Installation

  1. Install Python and Torch

    conda create -n dino3d python=3.9.20
conda activate dino3d
conda install pytorch==2.4.0 torchvision==0.19.0 torchaudio==2.4.0 pytorch-cuda=12.1 -c pytorch -c nvidia

  2. Install MinkowskiEngine

    conda install openblas-devel -c anaconda
cp ~/anaconda3/envs/dino3d/lib/libopenblas.so* ~/anaconda3/envs/dino3d/lib/python3.9/site-packages/torch/lib/
cd your/path/packages # your path to save packages
git clone https://github.com/NVIDIA/MinkowskiEngine.git
cd MinkowskiEngine

    Before compilation, we need to modify the source code to fix a bug refer to this issue:

    • MinkowskiEngine/src/3rdparty/concurrent_unordered_map.cuh: Add #include <thrust/execution_policy.h>

    • MinkowskiEngine/src/convolution_kernel.cuh: Add #include <thrust/execution_policy.h>

    • MinkowskiEngine/src/coordinate_map_gpu.cu: Add #include <thrust/unique.h> and #include <thrust/remove.h>

    • MinkowskiEngine/src/spmm.cu: Add #include <thrust/execution_policy.h>, #include <thrust/reduce.h>, and #include <thrust/sort.h>

    find ${CONDA_PREFIX}/include -name "cblas.h"
python setup.py install --blas_include_dirs=${CONDA_PREFIX}/include --blas=openblas

  3. Install OpenMMLab projects

    pip install --no-deps mmengine==0.10.6 mmdet==3.3.0
pip install ninja psutil

# install mmcv
pip install setuptools wheel packaging pyyaml # Install the dependencies required for compilation manually
cd your/path/packages # your path to save packages
git clone https://github.com/open-mmlab/mmcv.git
cd mmcv
git checkout v2.1.0
FORCE_CUDA=1 pip install -e . --no-build-isolation -v

pip install --no-deps mmdet3d==1.4

  4. Install torch-scatter

    pip install torch-scatter==2.1.2 -f https://data.pyg.org/whl/torch-2.4.0+cu121.html --no-deps

  5. Install ScanNet superpoint segmentator if you need to prepare superpoint (optional)

    cd your/path/packages # your path to save packages
git clone https://github.com/Karbo123/segmentator.git
cd segmentator

cd csrc && mkdir build && cd build
cmake .. \
-DCMAKE_PREFIX_PATH=`python -c 'import torch;print(torch.utils.cmake_prefix_path)'` \
-DPYTHON_INCLUDE_DIR=$(python -c "from distutils.sysconfig import get_python_inc; print(get_python_inc())")  \
-DPYTHON_LIBRARY=$(python -c "import distutils.sysconfig as sysconfig; print(sysconfig.get_config_var('LIBDIR'))") \
-DCMAKE_INSTALL_PREFIX=`python -c 'from distutils.sysconfig import get_python_lib; print(get_python_lib())'` 

make && make install # after install, please do not delete this folder (as we only create a symbolic link)

  6. Install remaining python packages

    pip install --no-deps \
    spconv-cu121==2.3.8 \
    addict==2.4.0 \
    yapf==0.43.0 \
    termcolor==2.5.0 \
    packaging==24.2 \
    numpy==2.0.1 \
    rich==13.4.2 \
    opencv-python==4.11.0.86 \
    pycocotools==2.0.10 \
    Shapely==1.8.5.post1 \
    scipy==1.13.1 \
    terminaltables==3.1.10 \
    numba==0.60.0 \
    llvmlite==0.43.0 \
    pccm==0.4.16 \
    ccimport==0.4.4 \
    pybind11==2.13.6 \
    ninja==1.11.1.3 \
    lark==1.2.2 \
    cumm-cu121==0.7.11 \
    pyquaternion==0.9.9 \
    lyft-dataset-sdk==0.0.8 \
    pandas==2.2.3 \
    python-dateutil==2.9.0.post0 \
    matplotlib==3.9.4 \
    pyparsing==3.2.1 \
    cycler==0.12.1 \
    kiwisolver==1.4.7 \
    scikit-learn==1.6.1 \
    joblib==1.4.2 \
    threadpoolctl==3.5.0 \
    cachetools==5.5.2 \
    nuscenes-devkit==1.1.11 \
    trimesh==4.6.2 \
    open3d==0.19.0 \
    plotly==6.0.0 \
    dash==2.18.2 \
    plyfile==1.1 \
    flask==3.0.3 \
    werkzeug==3.0.6 \
    click==8.1.8 \
    blinker==1.9.0 \
    itsdangerous==2.2.0 \
    importlib_metadata==8.5.0 \
    zipp==3.21.0 \
    tqdm==4.67.1 \
    transformers==4.48.0 \
    huggingface-hub==0.32.4 \
    regex==2024.11.6 \
    safetensors==0.5.3 \
    tokenizers==0.21.1 \
    portalocker==3.1.1 \
    fire==0.7.0 \
    pytz==2023.4

  7. Modify the source code in mmdet3d

  • Find the file: ~/anaconda3/envs/dino3d/lib/python3.9/site-packages/mmdet3d/models/layers/spconv/overwrite_spconv/write_spconv2.py

  • Change the line version = local metadata.get('version', None) to version = local metadata.get('version', 2)