Skip to content

Latest commit

 

History

History
54 lines (43 loc) · 1.68 KB

local_setup.md

File metadata and controls

54 lines (43 loc) · 1.68 KB

Installation

Setup a conda environment

git clone https://github.com/mees/vapo.git
cd vapo/
conda create -n vapo_env python=3.8
conda activate vapo_env

Install pytorch

To install the voting layer the cudatoolkit installed with pytorch must match the native CUDA version (in /usr/local/cuda/) which will be used to compile the CUDA code. Otherwise, the compiled CUDA/C++ code may not be compatible with the conda-installed PyTorch.

First check your CUDA version with nvcc --version or in /usr/local/cuda/version.json then install pytorch with the corresponding toolkit version. This code was tested with pytorch 1.10.1 and cuda 11.3.

conda install pytorch==1.10.1 torchvision==0.11.2 torchaudio==0.10.1 cudatoolkit=11.3 -c pytorch -c conda-forge

Install the VRENv

For more details refere to VREnv setup

pip install pybullet
cd /VAPO_ROOT/VREnv
pip install -e .

Install VAPO

For more details refere to VREnv setup

cd /VAPO_ROOT/
pip install -e .

Install the Hough voting layer

To install the voting layer first install Eigen.

git clone https://gitlab.com/libeigen/eigen.git
cd eigen/
mkdir build/
cd build/
cmake ..
sudo make install

Go to the directory of the voting layer and run setup.py. If you do not have sudo privileges, don't run sudo make install instead change the diretory in "include_dirs" to match where the eigen repo was downloaded, then run:

conda activate vapo_env
cd /VAPO_ROOT/vapo/affordance/hough_voting/
python setup.py install