Step-by-step installation instructions

a. Create a conda virtual environment and activate it.

conda create -n mgmap python=3.8 -y
conda activate mgmap

b. Install PyTorch and torchvision following the official instructions.

pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html

c. Install gcc>=5 in conda env (optional).

conda install -c omgarcia gcc-5 # gcc-6.2

c. Install mmcv-full.

pip install mmcv-full==1.4.0
#  pip install mmcv-full==1.4.0 -f https://download.openmmlab.com/mmcv/dist/cu111/torch1.9.0/index.html

d. Install mmdet and mmseg.

pip install mmdet==2.14.0
pip install mmsegmentation==0.14.1

e. Install timm.

pip install timm

f. Clone MGMap.

git clone https://github.com/xiaolul2/MGMap

g. Install mmdet3d and GKT

cd /path/to/MGMap/mmdetection3d
python setup.py develop

cd /path/to/MGMap/projects/mmdet3d_plugin/gemap/modules/ops/geometric_kernel_attn
python setup.py build install

h. Install other requirements.

cd /path/to/MGMap
pip install -r requirement.txt

i. Prepare pretrained models.

cd /path/to/MGMap
mkdir ckpts

# download ResNet weights
cd ckpts 
wget https://download.pytorch.org/models/resnet50-19c8e357.pth