Skip to content

Latest commit

 

History

History
114 lines (82 loc) · 5.09 KB

GETTING_STARTED.md

File metadata and controls

114 lines (82 loc) · 5.09 KB
Raw

Getting Started

Data Preparation

  • Download the data from here and organize as follows.
WAYMO_DATASET_ROOT
  ├── tfrecord_training
  ├── tfrecord_validation
  └── tfrecord_testing
  • Convert the tfrecord data to pickle files.
# Train set
CUDA_VISIBLE_DEVICES=-1 python det3d/datasets/waymo/waymo_converter.py --record_path 'WAYMO_DATASET_ROOT/tfrecord_training/*.tfrecord' --root_path 'WAYMO_DATASET_ROOT/train/'
# Validation set
CUDA_VISIBLE_DEVICES=-1 python det3d/datasets/waymo/waymo_converter.py --record_path 'WAYMO_DATASET_ROOT/tfrecord_validation/*.tfrecord' --root_path 'WAYMO_DATASET_ROOT/val/'
# Testing set
CUDA_VISIBLE_DEVICES=-1 python det3d/datasets/waymo/waymo_converter.py --record_path 'WAYMO_DATASET_ROOT/tfrecord_testing/*.tfrecord' --root_path 'WAYMO_DATASET_ROOT/test/'
  • Create a symlink to the dataset root.
# Remember to change the WAYMO_DATASET_ROOT to the actual path in your system.
mkdir data && cd data
ln -s WAYMO_DATASET_ROOT Waymo
  • Create info files.
python tools/create_data.py waymo_data_prep --root_path=data/Waymo --split train --nsweeps=2
python tools/create_data.py waymo_data_prep --root_path=data/Waymo --split val --nsweeps=2
python tools/create_data.py waymo_data_prep --root_path=data/Waymo --split test --nsweeps=2
  • The data and info files should be organized as follows.
3DAL_PyTorch
  └── data
    └── Waymo
      ├── tfrecord_training
      ├── tfrecord_validation
      ├── tfrecord_testing
      ├── train <-- all training frames and annotations
      ├── val   <-- all validation frames and annotations
      ├── test  <-- all testing frames and annotations
      ├── infos_train_02sweeps_filter_zero_gt.pkl
      ├── infos_val_02sweeps_filter_zero_gt.pkl
      └── infos_test_02sweeps_filter_zero_gt.pkl

Training & Testing

Preparation

  1. 3D Object Detection

Download the pre-trained 3D object detector CenterPoint and place it into the directory work_dirs/{config_name}.

python tools/dist_test.py configs/waymo/voxelnet/two_stage/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel.py --work_dir work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/{train or val} --checkpoint work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/CenterPoint.pth --speed_test
  1. 3D Multi-Object Tracking
python tools/waymo_tracking/test.py --work_dir work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/{train or val} --checkpoint work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/{train or val}/prediction.pkl --info_path data/Waymo/infos_{train or val}_02sweeps_filter_zero_gt.pkl
  1. Object Track Data Extraction
python tools/trackData.py --work_dir work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/{train or val}
  1. Motion State Classification
python tools/motionState.py --track_train work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/train --track_val work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/val

Training

  • Static Object Auto-Labeling
python tools/static_train.py --track work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/train --infos data/Waymo/infos_train_02sweeps_filter_zero_gt.pkl --model_type {model_type}
  • Dynamic Object Auto-Labeling
python tools/dynamic_train.py --track work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/train --infos data/Waymo/infos_train_02sweeps_filter_zero_gt.pkl --model_type {model_type}

Testing

  • Static Object Auto-Labeling
python tools/static_eval.py --track work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/val/trackStatic.pkl --infos data/Waymo/infos_val_02sweeps_filter_zero_gt.pkl --model_path work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/train/static/model/{model_type}/{model_path} --model_type {model_type} --det_annos work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/val/det_annos.pkl
  • Dynamic Object Auto-Labeling
python tools/dynamic_eval.py --track work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/val/trackDynamic.pkl --infos data/Waymo/infos_val_02sweeps_filter_zero_gt.pkl --model_path work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/train/dynamic/model/{model_type}/{model_path} --model_type {model_type} --det_annos work_dirs/waymo_centerpoint_voxelnet_two_sweep_two_stage_bev_5point_ft_6epoch_freeze_with_vel/val/det_annos.pkl