Modified from the official mmdet3d getting_started.md.
NeuralMapPrior is developed with the following versions of modules, especially based on mmdetection3d, please refer to mmdetection3d getting_started.md to obtain more instructions.
- Linux or macOS (Windows is not currently officially supported)
- Python 3.8.13
- PyTorch 1.9.0+cu111
- CUDA 11.2
- GCC 7.3.0
- MMCV==1.3.14
- MMDetection==v2.14.0
- MMSegmentation==v0.14.1
- MMDetection3D==v0.17.2
a. Create a conda virtual environment and activate it.
conda create --name npn python=3.8 -y
conda activate npn
b. Install PyTorch and torchvision following the official instructions.
pip install torch==1.9.0+cu111 torchvision==0.10.0+cu111 -f https://download.pytorch.org/whl/torch_stable.html
c. Clone the neural_map_prior repository.
git clone --recursive git@github.com:Tsinghua-MARS-Lab/neural_map_prior.git
cd neural_map_prior
d. Install MMCV, MMDetection, MMSegmentation, and other requirements.
pip install -r requirements.txt
e. Install MMDetection3D.
cd neural_map_prior/mmdetection3d
python setup.py develop