GitHub - geng-haoran/GAPartNet_env: GAPartNet conda environment (original) (raw)

GAPartNet conda environment

An worked environment:

build env for RL-Pose

Step 1: CUDA and basic conda environment

python == 3.8

cuda, nvcc >= 11.3

torch >= 1.11: e.g. pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu117

Step 2: cmake and open3d

Notice: here you may need su/sudo cmake >= 3.24: following the instruction here

if your cmake version is lower than 3.24: follow this

Some possible problems here:

Could not find OpenSSL

sudo aptitude install libssl-dev

Could NOT find X11 (missing: X11_X11_INCLUDE_PATH X11_X11_LIB)

sudo aptitude install libx11-dev

The RandR headers were not found

sudo apt-get install libxrandr-dev

The Xinerama headers were not found

sudo apt-get install libxinerama-dev
sudo apt-get install libsdl2-dev

The Xcursor headers were not found

sudo apt-get install libxcursor-dev

Cannot find matching libc++ and libc++abi libraries with version >=7

sudo aptitude install -y clang libc++-dev libc++abi-dev cmake ninja-build

open3d with pytorch extension (need to build from source) install repo

git clone https://github.com/isl-org/Open3D

install dependences

util/install_deps_ubuntu.sh

build:

cmake command (remember to replace YOUR_PATH_TO_INSTALL_FOLDER & YOUR_PATH_TO_ANACONDA_OR_MINICONDA & YOUR_CONDA_ENV_NAME as your real paths or names):

cmake \
-DPython3_ROOT="/YOUR_PATH_TO_ANACONDA_OR_MINICONDA/envs/YOUR_CONDA_ENV_NAME/bin/" \
-DWITH_OPENMP=ON \
-DWITH_SIMD=ON \
-DBUILD_PYTORCH_OPS=ON \
-DBUILD_CUDA_MODULE=ON \
-DBUILD_COMMON_CUDA_ARCHS=ON \
-DGLIBCXX_USE_CXX11_ABI=OFF \
-DBUILD_JUPYTER_EXTENSION=ON \
-DCMAKE_CUDA_COMPILER=$(which nvcc) \
-DCMAKE_INSTALL_PREFIX="YOUR_PATH_TO_INSTALL_FOLDER/open3d_install" \
..

An example of tested cmake command:

cmake \
-DPython3_ROOT="~/miniconda3/envs/rlgpu/bin" \
-DWITH_OPENMP=ON \
-DWITH_SIMD=ON \
-DBUILD_PYTORCH_OPS=ON \
-DBUILD_CUDA_MODULE=ON \
-DBUILD_COMMON_CUDA_ARCHS=ON \
-DGLIBCXX_USE_CXX11_ABI=OFF \
-DBUILD_JUPYTER_EXTENSION=ON \
-DCMAKE_CUDA_COMPILER=$(which nvcc) \
-DCMAKE_INSTALL_PREFIX="/mnt/data/GAPartNet_env/open3d_install" \
..

# Activate the virtualenv first
# Install pip package in the current python environment (recommended)
pip install jupyter_packaging jupyterlab
make install-pip-package

# Create Python package in build/lib
make python-package

# Create pip wheel in build/lib
# This creates a .whl file that you can install manually.
make pip-package

Step 3: install other packages

pointnet_ops

https://github.com/erikwijmans/Pointnet2_PyTorch/tree/master/pointnet2_ops_lib

clone repo

git clone git@github.com:erikwijmans/Pointnet2_PyTorch.git

install this repo

cd pointnet2_ops_lib
python setup.py install

epic_ops

https://github.com/geng-haoran/epic_ops

clone repo

git clone git@github.com:geng-haoran/epic_ops.git

install this repo

cd epic_ops
python setup.py develop

spconv

spconv: pip install spconv-cuxxx (https://github.com/traveller59/spconv)

pip

pip install wandb tensorboard ipdb gym tqdm rich opencv_python pytorch3d pyparsing pytorch_lightning addict yapf h5py sorcery  pynvml torchdata==0.5.1 einops