Skip to content

Latest commit

 

History

History
83 lines (68 loc) · 2.5 KB

INSTALL.md

File metadata and controls

83 lines (68 loc) · 2.5 KB

Installation

Modified from CenterPoint's codebase.

Requirements

  • Linux
  • Python 3.6+
  • PyTorch 1.1 or higher
  • CUDA 10.0 or higher
  • CMake 3.13.2 or higher
  • APEX
  • SpConv

Notes

  • Our ProposalContrast support both SpConv v1.2 and SpConv v2.
  • Orther versions of SpConv may consume more GPU memory.
Our project has beed tested in the following environment:
  • OS: Ubuntu 18.04
  • Python: 3.6.13
  • PyTorch: 1.10.1
  • CUDA: 11.1

Basic Installation

# basic python libraries
conda create --name proposalcontrast python=3.6
conda activate proposalcontrast
pip install torch==1.10.1+cu111 torchvision==0.11.2+cu111 torchaudio==0.10.1 -f https://download.pytorch.org/whl/torch_stable.html
git clone https://github.com/yinjunbo/ProposalContrast.git
cd ProposalContrast
pip install -r requirements.txt

# add ProposalContrast to PYTHONPATH by adding the following line to ~/.bashrc (change the path accordingly)
export PYTHONPATH="${PYTHONPATH}:PATH_TO_CENTERPOINT"

Advanced Installation

Cuda Extensions

# set the cuda path(change the path to your own cuda location) 
export PATH=/usr/local/cuda-11.1/bin:$PATH
export CUDA_PATH=/usr/local/cuda-11.1
export CUDA_HOME=/usr/local/cuda-11.1
export LD_LIBRARY_PATH=/usr/local/cuda-11.1/lib64:$LD_LIBRARY_PATH
bash setup.sh

APEX

git clone https://github.com/NVIDIA/apex
cd apex
git checkout 5633f6  
pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./

If your use torch>1.8, please remember to modify the code in ~/envs/proposalcontrast/lib/python3.6/site-packages/apex/amp/_amp_state.py:

if TORCH_MAJOR == 1 and TORCH_MINOR < 8:
    from torch._six import container_abcs
else:
    import collections.abc as container_abcs

SpConv v1.2

sudo apt-get install libboost-all-dev
git clone https://github.com/traveller59/spconv.git --recursive
cd spconv && git checkout 7342772
python setup.py bdist_wheel
cd ./dist && pip install *

SpConv v2

sudo apt-get install libboost-all-dev
pip install spconv-cu111

Please prepare the dataset following WAYMO and start the self-supervised learning of 3D models accroding to RUN_MODEL.