Skip to content

Latest commit

 

History

History
308 lines (230 loc) · 19.6 KB

GETTING_STARTED_ADA.md

File metadata and controls

308 lines (230 loc) · 19.6 KB

Getting Started & Problem Definition

Different from UDA task, the purpose of an Active Domain Adaptation (ADA) task is to pick up a subset of unlabeled target domain $t$ to perform the manual annotation process, such that we can achieve a good trade-off between high performance and low annotation cost, where labeled training data from the source domain $s$ (such as point cloud or images) are assumed to be available for initializing the training model.

   

Getting Started & Training-Testing for ADA setting

Here, We take Waymo-to-KITTI adaptation as an example.

Pretraining stage: train the source-only model on the labeled source domain:

  • Train FEAT=3 (X,Y,Z) with SN (statistical normalization) using multiple GPUs

    sh scripts/dist_train.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_old_anchor_sn_kitti.yaml
  • Train FEAT=3 (X,Y,Z) with SN (statistical normalization) using multiple machines

    sh scripts/slurm_train.sh ${PARTITION} ${JOB_NAME} ${NUM_NODES} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_old_anchor_sn_kitti.yaml
  • Train FEAT=3 (X,Y,Z) without SN (statistical normalization) using multiple GPUs

    sh scripts/dist_train.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_feat_3_vehi.yaml
  • Train FEAT=3 (X,Y,Z) without SN (statistical normalization) using multiple machines

    sh scripts/slurm_train.sh ${PARTITION} ${JOB_NAME} ${NUM_NODES} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_feat_3_vehi.yaml
  • Train other baseline detectors such as Voxel R-CNN using multiple GPUs

    sh scripts/dist_train.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/voxel_rcnn_feat_3_vehi.yaml
  • Train other baseline detectors such as Voxel R-CNN using multiple machines

    sh scripts/slurm_train.sh ${PARTITION} ${JOB} ${NUM_NODES} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/voxel_rcnn_feat_3_vehi.yaml

Evaluate the source-pretrained model:

  • Note that for the cross-domain setting where the KITTI dataset is regarded as the target domain, please try --set DATA_CONFIG_TAR.FOV_POINTS_ONLY True to enable front view point cloud only. We report the best model for all epochs on the validation set.

  • Test the source-only models using multiple GPUs

    sh scripts/dist_test.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_feat_3_vehi.yaml \ 
     --ckpt ${CKPT} 
  • Test the source-only models using multiple machines

    sh scripts/slurm_test_mgpu.sh ${PARTITION} ${NUM_NODES} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_feat_3_vehi.yaml \ 
     --ckpt ${CKPT}
  • Test the source-only models of all ckpts using multiple GPUs

    sh scripts/dist_test.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_feat_3_vehi.yaml \ 
     --eval_all
  • Test the source-only models of all ckpts using multiple machines

    sh scripts/slurm_test_mgpu.sh ${PARTITION} ${NUM_NODES} \ 
     --cfg_file ./cfgs/DA/waymo_kitti/source_only/pvrcnn_feat_3_vehi.yaml \ 
     --eval_all

Bi3D Adaptation stage 1: active source domain data

  • You need to set the --pretrained_model ${PRETRAINED_MODEL} when finish the pretraining model stage

  • Train with SN (statistical normalization) using multiple GPUs

    sh scripts/ADA/dist_train_active_source.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_source_only.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}
  • Train with SN (statistical normalization) using multiple machines

    sh scripts/ADA/slurm_train_active_source.sh ${PARTITION} ${JOB_NAME} ${NUM_NODES} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_source_only.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}
  • Train without SN (statistical normalization) using multiple GPUs

    sh scripts/ADA/dist_train_active_source.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_source_only_wosn.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}
  • Train without SN (statistical normalization) using multiple machines

    sh scripts/ADA/slurm_train_active_source.sh ${PARTITION} ${JOB_NAME} ${NUM_NODES} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_source_only_wosn.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}

Bi3D Adaptation stage 2: active target domain data

  • You need to set the --pretrained_model ${PRETRAINED_MODEL} when finish the adaptation stage 1

  • Train with 1% annotation budget using multiple GPUs

    sh scripts/ADA/dist_train_active.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_dual_target_01.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}
  • Train with 1% annotation budget using multiple machines

    sh scripts/ADA/slurm_train_active.sh ${PARTITION} ${JOB_NAME} ${NUM_NODES} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_dual_target_01.yaml \  
     --pretrained_model ${PRETRAINED_MODEL}
  • Train with 5% annotation budget using multiple GPUs

    sh scripts/ADA/dist_train_active.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_dual_target_05.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}
  • Train with 5% annotation budget using multiple machines

    sh scripts/ADA/slurm_train_active.sh ${PARTITION} ${JOB_NAME} ${NUM_NODES} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_dual_target_05.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}

Evaluating the model on the labeled target domain

  • Test with a ckpt file:

    python test.py --cfg_file ${CONFIG_FILE} \ 
     --batch_size ${BATCH_SIZE} \ 
     --ckpt ${CKPT}
  • To test all the saved checkpoints of a specific training setting and draw the performance curve on the Tensorboard, add the --eval_all argument:

    python test.py \ 
     --cfg_file ${CONFIG_FILE} \ 
     --batch_size ${BATCH_SIZE} \ 
     --eval_all
  • Notice that if you want to test on the setting with KITTI as target domain, please add --set DATA_CONFIG_TAR.FOV_POINTS_ONLY True to enable front view point cloud only:

    python test.py \ 
     --cfg_file ${CONFIG_FILE} \ 
     --batch_size ${BATCH_SIZE} \ 
     --eval_all \ 
     --set DATA_CONFIG_TAR.FOV_POINTS_ONLY True
  • To test with multiple machines for S-Proj:

    sh scripts/slurm_test_mgpu.sh ${PARTITION} ${NUM_NODES} \ 
      --cfg_file ${CONFIG_FILE} \ 
      --batch_size ${BATCH_SIZE}

Train with other active domain adaptation / active learning methods

  • Train with TQS

    sh scripts/ADA/dist_train_active_TQS.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_TQS.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}
  • Train with CLUE

    sh scripts/ADA/dist_train_active_CLUE.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_CLUE.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}

Combine Bi3D and UDA

  • Train with multiple GPUs

    sh scripts/ADA/dist_train_active_st3d.sh ${NUM_GPUs} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_st3d.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}
  • Train with multiple machines

    sh scripts/ADA/slurm_train_active_st3d.sh ${PARTITION} ${JOB_NAME} ${NUM_NODES} \ 
     --cfg_file ./cfgs/ADA/waymo-kitti/pvrcnn/active_st3d.yaml \ 
     --pretrained_model ${PRETRAINED_MODEL}

   

All ADA Results:

We report the cross-dataset adaptation results including Waymo-to-KITTI, nuScenes-to-KITTI, Waymo-to-nuScenes, and Waymo-to-Lyft.

  • All LiDAR-based models are trained with 2 NVIDIA A100 GPUs and are available for download.

ADA Results for Waymo-to-KITTI:

training time Adaptation Car@R40 download
PV-RCNN ~23h@4 A100 Source Only 67.95 / 27.65 -
PV-RCNN ~1.5h@2 A100 Bi3D (1% annotation budget) 87.12 / 78.03 Model-58M
PV-RCNN ~10h@2 A100 Bi3D (5% annotation budget) 89.53 / 81.32 Model-58M
PV-RCNN ~1.5h@2 A100 TQS 82.00 / 72.04 Model-58M
PV-RCNN ~1.5h@2 A100 CLUE 82.13 / 73.14 Model-50M
PV-RCNN ~10h@2 A100 Bi3D+ST3D 87.83 / 81.23 Model-58M
Voxel R-CNN ~16h@4 A100 Source Only 64.87 / 19.90 -
Voxel R-CNN ~1.5h@2 A100 Bi3D (1% annotation budget) 88.09 / 79.14 Model-72M
Voxel R-CNN ~6h@2 A100 Bi3D (5% annotation budget) 90.18 / 81.34 Model-72M
Voxel R-CNN ~1.5h@2 A100 TQS 78.26 / 67.11 Model-72M
Voxel R-CNN ~1.5h@2 A100 CLUE 81.93 / 70.89 Model-72M

ADA Results for nuScenes-to-KITTI:

training time Adaptation Car@R40 download
PV-RCNN ~23h@4 A100 Source Only 68.15 / 37.17 Model-150M
PV-RCNN ~1.5h@2 A100 Bi3D (1% annotation budget) 87.00 / 77.55 Model-58M
PV-RCNN ~9h@2 A100 Bi3D (5% annotation budget) 89.63 / 81.02 Model-58M
PV-RCNN ~1.5h@2 A100 TQS 84.66 / 75.40 Model-58M
PV-RCNN ~1.5h@2 A100 CLUE 74.77 / 64.43 Model-50M
PV-RCNN ~7h@ 2 A100 Bi3D+ST3D 89.28 / 79.69 Model-58M
Voxel R-CNN ~16h@4 A100 Source Only 68.45 / 33.00 Model-191M
Voxel R-CNN ~1.5h@2 A100 Bi3D (1% annotation budget) 87.33 / 77.24 Model-72M
Voxel R-CNN ~5.5h@2 A100 Bi3D (5% annotation budget) 87.66 / 80.22 Model-72M
Voxel R-CNN ~1.5h@2 A100 TQS 79.12 / 68.02 Model-73M
Voxel R-CNN ~1.5h@2 A100 CLUE 77.98 / 66.02 Model-65M

ADA Results for Waymo-to-nuScenes:

training time Adaptation Car@R40 download
PV-RCNN ~23h@4 A100 Source Only 31.02 / 21.21 -
PV-RCNN ~4h@2 A100 Bi3D (1% annotation budget) 45.00 / 30.81 Model-58M
PV-RCNN ~12h@4 A100 Bi3D (5% annotation budget) 48.03 / 32.02 Model-58M
PV-RCNN ~4h@2 A100 TQS 35.47 / 25.00 Model-58M
PV-RCNN ~3h@2 A100 CLUE 38.18 / 26.96 Model-50M
Voxel R-CNN ~16h@4 A100 Source Only 29.08 / 19.42 -
Voxel R-CNN ~2.5h@2 A100 Bi3D (1% annotation budget) 45.47 / 30.49 Model-72M
Voxel R-CNN ~4h@4 A100 Bi3D (5% annotation budget) 46.78 / 32.14 Model-72M
Voxel R-CNN ~4h@2 A100 TQS 36.38 / 24.18 Model-72M
Voxel R-CNN ~3h@2 A100 CLUE 37.27 / 25.12 Model-65M
SECOND ~3h@2 A100 Bi3D(1%) 46.15 / 26.24 Model-54M

ADA Results for Waymo-to-Lyft:

training time Adaptation Car@R40 download
PV-RCNN ~23h@4 A100 Source Only 70.10 / 53.11 -
PV-RCNN ~7h@2 A100 Bi3D (1% annotation budget) 79.07 / 63.74 Model-58M
PV-RCNN ~22h@2 A100 Bi3D (5% annotation budget) 80.19 / 66.09 Model-58M
PV-RCNN ~7h@2 A100 TQS 70.87 / 55.25 Model-58M
PV-RCNN ~5h@2 A100 CLUE 75.23 / 62.17 Model-50M
Voxel R-CNN ~16h@4 A100 Source Only 70.52 / 53.48 -
Voxel R-CNN ~7h@2 A100 Bi3D (1% annotation budget) 77.00 / 61.23 Model-72M
Voxel R-CNN ~19h@2 A100 Bi3D (5% annotation budget) 79.15 / 65.26 Model-72M
Voxel R-CNN ~8h@2 A100 TQS 71.11 / 56.28 Model-73M
Voxel R-CNN ~5h@2 A100 CLUE 75.61 / 59.34 Model-65M