In this repository, the project folder is PETR, the other files are implementation of PETR for Kitti Dataset, which lead to bad result. The update for double camera case will be done in the future I hope.
Before running the project, remember to download NuScenes Dataset following the official git repository of PETR first. The report of this project is at DN_PETR.pdf
Tung Yen Chiang
Tung Hsiao
cd PETR
pip install ninja wheel lyft_dataset_sdk networkx==2.2 numba==0.48.0 nuscenes-devkit plyfile scikit-image tensorboard trimesh==2.35.39 numpy==1.19.5 open3d einops yapf==0.40.0
pip install --upgrade pip
pip install -U openmim
mim install mmcv-full==1.4.0
pip install mmdet==2.24.1
pip install mmsegmentation==0.20.2
apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/3bf863cc.pub
apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/7fa2af80.pub
ln -fs /usr/share/zoneinfo/America/Los_Angeles /etc/localtime
DEBIAN_FRONTEND=noninteractive apt-get install gconf-service libasound2 libatk1.0-0 libc6 libcairo2 libcups2 libdbus-1-3 libexpat1 libfontconfig1 libgcc1 libgconf-2-4 libgdk-pixbuf2.0-0 libglib2.0-0 libgtk-3-0 libnspr4 libpango-1.0-0 libpangocairo-1.0-0 libstdc++6 libx11-6 libx11-xcb1 libxcb1 libxcomposite1 libxcursor1 libxdamage1 libxext6 libxfixes3 libxi6 libxrandr2 libxrender1 libxss1 libxtst6 ca-certificates fonts-liberation libappindicator1 libnss3 lsb-release xdg-utils wget -y
apt-get update && apt-get install -y git
cd mmdetection3d
git checkout v0.17.1
pip install -r requirements/build.txt
python3 setup.py develop
tar -xvzf /path/to/yourfile.tgz
pip install setuptools==59.5.0
PYTHONPATH=${PWD}:$PYTHONPATH python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./data/nuscenes --extra-tag nuscenes
tools/dist_train.sh projects/configs/petr/petr_vovnet_gridmask_p4_800x320.py 1 --work-dir output/epoch24/petr_vovnet_gridmask_p4_800x320/
tools/dist_train.sh projects/configs/denoise/petr_vovnet_gridmask_p4_800x320_dn.py 1 --work-dir output/epoch24/petr_vovnet_gridmask_p4_800x320_dn/