Skip to content

Latest commit

 

History

History
executable file
·
95 lines (66 loc) · 2.06 KB

getting_started.md

File metadata and controls

executable file
·
95 lines (66 loc) · 2.06 KB
Raw

Modify from the official mmdet3d getting_started.md

Prerequisites

BEVDet is developed with the following version of modules.

  • Linux or macOS (Windows is not currently officially supported)
  • Python 3.7
  • PyTorch 1.9.0
  • CUDA 11.3.1
  • GCC 7.3.0
  • MMCV==1.3.13
  • MMDetection==2.14.0
  • MMSegmentation==0.14.1

Installation

a. Create a conda virtual environment and activate it.

conda create -n bevdet python=3.7 -y
conda activate bevdet

b. Install PyTorch and torchvision following the official instructions.

conda install -c pytorch pytorch torchvision -y

c. Install MMCV.

pip install mmcv-full==1.3.13

d. Install MMDetection.

pip install mmdet==2.14.0

e. Install MMSegmentation.

pip install mmsegmentation==0.14.1

f. Clone the BEVDet repository.

git clone https://github.com/HuangJunJie2017/BEVDet.git
cd BEVDet

g.Install build requirements and then install BEVDet.

pip install -v -e .  # or "python setup.py develop"

A from-scratch setup script

Here is a full script for setting up MMdetection3D with conda.

conda create -n bevdet python=3.7 -y
conda activate bevdet

# install latest PyTorch prebuilt with the default prebuilt CUDA version (usually the latest)
conda install -c pytorch pytorch torchvision -y

# install mmcv
pip install mmcv-full==1.3.13

# install mmdetection
pip install mmdet==2.14.0

# install mmsegmentation
pip install mmsegmentation==0.14.1

# install BEVDet
git clone https://github.com/HuangJunJie2017/BEVDet.git
cd BEVDet
pip install -v -e .

Data Preparation

a. Please refer to nuScenes for initial preparation.

b. Prepare dataset specific for BEVDet4D.

python tools/data_converter/prepare_nuscenes_for_bevdet4d.py