OpenPCDet
OpenPCDet copied to clipboard
Published
20 hours ago •
open-mmlab
open-mmlab
Low Acc on 3D task, in Custom datasets.
Thank you for your excellent work.
I'm trying to use OpenPCDet's customized dataset feature, but I'm finding that my models are performing particularly poorly for BEV and 3D tasks, does anyone have any idea what this might be? Or how I should troubleshoot the problem?
Another question is that when I check (visualize) custom_infos_train.pkl, I find that some labels that are outside the range of the point cloud are not removed, is this normal?
I am using the TUMTraf dataset, only the point cloud and kitti format label(no image, no cabli). You can check my custom dataset and model config files in my fork if you need.
My log file:
2024-05-20 22:53:41,738 INFO CUDA_VISIBLE_DEVICES=0
2024-05-20 22:53:41,738 INFO Training with a single process
2024-05-20 22:53:41,738 INFO cfg_file /user/<user.name>/<user.name2>/OpenPCDet/tools/cfgs/custom_models/pps.yaml
2024-05-20 22:53:41,738 INFO batch_size 4
2024-05-20 22:53:41,738 INFO epochs 15
2024-05-20 22:53:41,738 INFO workers 4
2024-05-20 22:53:41,738 INFO extra_tag default
2024-05-20 22:53:41,738 INFO ckpt None
2024-05-20 22:53:41,738 INFO pretrained_model None
2024-05-20 22:53:41,738 INFO launcher none
2024-05-20 22:53:41,738 INFO tcp_port 18888
2024-05-20 22:53:41,738 INFO sync_bn False
2024-05-20 22:53:41,738 INFO fix_random_seed False
2024-05-20 22:53:41,738 INFO ckpt_save_interval 1
2024-05-20 22:53:41,738 INFO local_rank None
2024-05-20 22:53:41,738 INFO max_ckpt_save_num 30
2024-05-20 22:53:41,738 INFO merge_all_iters_to_one_epoch False
2024-05-20 22:53:41,738 INFO set_cfgs None
2024-05-20 22:53:41,738 INFO max_waiting_mins 0
2024-05-20 22:53:41,738 INFO start_epoch 0
2024-05-20 22:53:41,738 INFO num_epochs_to_eval 0
2024-05-20 22:53:41,738 INFO save_to_file False
2024-05-20 22:53:41,738 INFO use_tqdm_to_record False
2024-05-20 22:53:41,738 INFO logger_iter_interval 50
2024-05-20 22:53:41,739 INFO ckpt_save_time_interval 300
2024-05-20 22:53:41,739 INFO wo_gpu_stat False
2024-05-20 22:53:41,739 INFO use_amp False
2024-05-20 22:53:41,739 INFO cfg.ROOT_DIR: /home/uni08/hpc/<user.name>/u11423/OpenPCDet
2024-05-20 22:53:41,739 INFO cfg.LOCAL_RANK: 0
2024-05-20 22:53:41,739 INFO cfg.CLASS_NAMES: ['CAR', 'PEDESTRIAN', 'BICYCLE']
2024-05-20 22:53:41,739 INFO ----------- DATA_CONFIG -----------
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.DATASET: CustomDataset
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.DATA_PATH: ../data/custom
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.POINT_CLOUD_RANGE: [0.0, -180.0, -10.0, 153.6, 137.44, 0.0]
2024-05-20 22:53:41,739 INFO ----------- MAP_CLASS_TO_KITTI -----------
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.CAR: Car
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.VAN: Van
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.TRUCK: Truck
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.PEDESTRIAN: Pedestrian
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.BICYCLE: Cyclist
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.MOTORCYCLE: Misc
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.TRAILER: Misc
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.BUS: Misc
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.MAP_CLASS_TO_KITTI.EMERGENCY_VEHICLE: Misc
2024-05-20 22:53:41,739 INFO ----------- DATA_SPLIT -----------
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.DATA_SPLIT.train: train
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.DATA_SPLIT.test: val
2024-05-20 22:53:41,739 INFO ----------- INFO_PATH -----------
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.INFO_PATH.train: ['custom_infos_train.pkl']
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.INFO_PATH.test: ['custom_infos_val.pkl']
2024-05-20 22:53:41,739 INFO ----------- POINT_FEATURE_ENCODING -----------
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.encoding_type: absolute_coordinates_encoding
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.used_feature_list: ['x', 'y', 'z', 'intensity']
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.src_feature_list: ['x', 'y', 'z', 'intensity']
2024-05-20 22:53:41,739 INFO ----------- DATA_AUGMENTOR -----------
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.DISABLE_AUG_LIST: ['placeholder']
2024-05-20 22:53:41,739 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.AUG_CONFIG_LIST: [{'NAME': 'gt_sampling', 'USE_ROAD_PLANE': False, 'DB_INFO_PATH': ['custom_dbinfos_train.pkl'], 'PREPARE': {'filter_by_min_points': ['CAR:5', 'PEDESTRIAN:5', 'BICYCLE:5']}, 'SAMPLE_GROUPS': ['CAR:15', 'PEDESTRIAN:15', 'BICYCLE:15'], 'NUM_POINT_FEATURES': 4, 'DATABASE_WITH_FAKELIDAR': False, 'REMOVE_EXTRA_WIDTH': [0.0, 0.0, 0.0], 'LIMIT_WHOLE_SCENE': False}, {'NAME': 'random_world_flip', 'ALONG_AXIS_LIST': ['x']}, {'NAME': 'random_world_rotation', 'WORLD_ROT_ANGLE': [-0.78539816, 0.78539816]}, {'NAME': 'random_world_scaling', 'WORLD_SCALE_RANGE': [0.95, 1.05]}]
2024-05-20 22:53:41,740 INFO cfg.DATA_CONFIG.DATA_PROCESSOR: [{'NAME': 'mask_points_and_boxes_outside_range', 'REMOVE_OUTSIDE_BOXES': True}, {'NAME': 'shuffle_points', 'SHUFFLE_ENABLED': {'train': True, 'test': False}}, {'NAME': 'transform_points_to_voxels', 'VOXEL_SIZE': [0.64, 0.64, 10], 'MAX_POINTS_PER_VOXEL': 32, 'MAX_NUMBER_OF_VOXELS': {'train': 230000, 'test': 230000}}]
2024-05-20 22:53:41,740 INFO cfg.DATA_CONFIG._BASE_CONFIG_: cfgs/dataset_configs/custom_dataset.yaml
2024-05-20 22:53:41,740 INFO ----------- MODEL -----------
2024-05-20 22:53:41,740 INFO cfg.MODEL.NAME: PointPillar
2024-05-20 22:53:41,740 INFO ----------- VFE -----------
2024-05-20 22:53:41,740 INFO cfg.MODEL.VFE.NAME: PillarVFE
2024-05-20 22:53:41,740 INFO cfg.MODEL.VFE.WITH_DISTANCE: False
2024-05-20 22:53:41,740 INFO cfg.MODEL.VFE.USE_ABSLOTE_XYZ: True
2024-05-20 22:53:41,740 INFO cfg.MODEL.VFE.USE_NORM: True
2024-05-20 22:53:41,740 INFO cfg.MODEL.VFE.NUM_FILTERS: [64]
2024-05-20 22:53:41,740 INFO ----------- MAP_TO_BEV -----------
2024-05-20 22:53:41,740 INFO cfg.MODEL.MAP_TO_BEV.NAME: PointPillarScatter
2024-05-20 22:53:41,740 INFO cfg.MODEL.MAP_TO_BEV.NUM_BEV_FEATURES: 64
2024-05-20 22:53:41,740 INFO ----------- BACKBONE_2D -----------
2024-05-20 22:53:41,740 INFO cfg.MODEL.BACKBONE_2D.NAME: BaseBEVBackbone
2024-05-20 22:53:41,740 INFO cfg.MODEL.BACKBONE_2D.LAYER_NUMS: [3, 5, 5]
2024-05-20 22:53:41,740 INFO cfg.MODEL.BACKBONE_2D.LAYER_STRIDES: [2, 2, 2]
2024-05-20 22:53:41,740 INFO cfg.MODEL.BACKBONE_2D.NUM_FILTERS: [64, 128, 256]
2024-05-20 22:53:41,740 INFO cfg.MODEL.BACKBONE_2D.UPSAMPLE_STRIDES: [1, 2, 4]
2024-05-20 22:53:41,740 INFO cfg.MODEL.BACKBONE_2D.NUM_UPSAMPLE_FILTERS: [128, 128, 128]
2024-05-20 22:53:41,740 INFO ----------- DENSE_HEAD -----------
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.NAME: AnchorHeadSingle
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.CLASS_AGNOSTIC: False
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.USE_DIRECTION_CLASSIFIER: True
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.DIR_OFFSET: 0.78539
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.DIR_LIMIT_OFFSET: 0.0
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.NUM_DIR_BINS: 2
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.ANCHOR_GENERATOR_CONFIG: [{'class_name': 'CAR', 'anchor_sizes': [[3.9, 1.6, 1.56]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-1.78], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.6, 'unmatched_threshold': 0.45}, {'class_name': 'PEDESTRIAN', 'anchor_sizes': [[0.8, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}, {'class_name': 'BICYCLE', 'anchor_sizes': [[1.76, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}]
2024-05-20 22:53:41,740 INFO ----------- TARGET_ASSIGNER_CONFIG -----------
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NAME: AxisAlignedTargetAssigner
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.POS_FRACTION: -1.0
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.SAMPLE_SIZE: 512
2024-05-20 22:53:41,740 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NORM_BY_NUM_EXAMPLES: False
2024-05-20 22:53:41,741 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.MATCH_HEIGHT: False
2024-05-20 22:53:41,741 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER: ResidualCoder
2024-05-20 22:53:41,741 INFO ----------- LOSS_CONFIG -----------
2024-05-20 22:53:41,741 INFO ----------- LOSS_WEIGHTS -----------
2024-05-20 22:53:41,741 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.cls_weight: 1.0
2024-05-20 22:53:41,741 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.loc_weight: 2.0
2024-05-20 22:53:41,741 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.dir_weight: 0.2
2024-05-20 22:53:41,741 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.code_weights: [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
2024-05-20 22:53:41,741 INFO ----------- POST_PROCESSING -----------
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.RECALL_THRESH_LIST: [0.3, 0.5, 0.7]
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.SCORE_THRESH: 0.1
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.OUTPUT_RAW_SCORE: False
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.EVAL_METRIC: kitti
2024-05-20 22:53:41,741 INFO ----------- NMS_CONFIG -----------
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.MULTI_CLASSES_NMS: False
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_TYPE: nms_gpu
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_THRESH: 0.01
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_PRE_MAXSIZE: 4096
2024-05-20 22:53:41,741 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_POST_MAXSIZE: 500
2024-05-20 22:53:41,741 INFO ----------- OPTIMIZATION -----------
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.BATCH_SIZE_PER_GPU: 4
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.NUM_EPOCHS: 15
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.OPTIMIZER: adam_onecycle
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.LR: 0.003
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.WEIGHT_DECAY: 0.01
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.MOMENTUM: 0.9
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.MOMS: [0.95, 0.85]
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.PCT_START: 0.4
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.DIV_FACTOR: 10
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.DECAY_STEP_LIST: [35, 45]
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.LR_DECAY: 0.1
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.LR_CLIP: 1e-07
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.LR_WARMUP: True
2024-05-20 22:53:41,741 INFO cfg.OPTIMIZATION.WARMUP_EPOCH: 1
2024-05-20 22:53:41,742 INFO cfg.OPTIMIZATION.GRAD_NORM_CLIP: 10
2024-05-20 22:53:41,742 INFO cfg.TAG: pps
2024-05-20 22:53:41,742 INFO cfg.EXP_GROUP_PATH: user/<user.name>/u11423/OpenPCDet/tools/cfgs/custom_models
2024-05-20 22:53:41,766 INFO ----------- Create dataloader & network & optimizer -----------
2024-05-20 22:53:42,026 INFO Database filter by min points CAR: 31706 => 16172
2024-05-20 22:53:42,027 INFO Database filter by min points PEDESTRIAN: 3472 => 3012
2024-05-20 22:53:42,027 INFO Database filter by min points BICYCLE: 598 => 542
2024-05-20 22:53:42,040 INFO Loading Custom dataset.
2024-05-20 22:53:42,101 INFO Total samples for CUSTOM dataset: 3840
/user/<user.name>/u11423/.conda/envs/openpcd/lib/python3.8/site-packages/torch/functional.py:512: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3587.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
2024-05-20 22:53:43,713 INFO ----------- Model PointPillar created, param count: 4834888 -----------
2024-05-20 22:53:43,713 INFO PointPillar(
(vfe): PillarVFE(
(pfn_layers): ModuleList(
(0): PFNLayer(
(linear): Linear(in_features=10, out_features=64, bias=False)
(norm): BatchNorm1d(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
)
)
)
(backbone_3d): None
(map_to_bev_module): PointPillarScatter()
(pfe): None
(backbone_2d): BaseBEVBackbone(
(blocks): ModuleList(
(0): Sequential(
(0): ZeroPad2d((1, 1, 1, 1))
(1): Conv2d(64, 64, kernel_size=(3, 3), stride=(2, 2), bias=False)
(2): BatchNorm2d(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(3): ReLU()
(4): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(5): BatchNorm2d(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(6): ReLU()
(7): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(8): BatchNorm2d(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(9): ReLU()
(10): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(11): BatchNorm2d(64, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(12): ReLU()
)
(1): Sequential(
(0): ZeroPad2d((1, 1, 1, 1))
(1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), bias=False)
(2): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(3): ReLU()
(4): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(5): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(6): ReLU()
(7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(8): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(9): ReLU()
(10): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(11): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(12): ReLU()
(13): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(14): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(15): ReLU()
(16): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(17): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(18): ReLU()
)
(2): Sequential(
(0): ZeroPad2d((1, 1, 1, 1))
(1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), bias=False)
(2): BatchNorm2d(256, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(3): ReLU()
(4): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(5): BatchNorm2d(256, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(6): ReLU()
(7): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(8): BatchNorm2d(256, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(9): ReLU()
(10): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(11): BatchNorm2d(256, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(12): ReLU()
(13): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(14): BatchNorm2d(256, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(15): ReLU()
(16): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(17): BatchNorm2d(256, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(18): ReLU()
)
)
(deblocks): ModuleList(
(0): Sequential(
(0): ConvTranspose2d(64, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(2): ReLU()
)
(1): Sequential(
(0): ConvTranspose2d(128, 128, kernel_size=(2, 2), stride=(2, 2), bias=False)
(1): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(2): ReLU()
)
(2): Sequential(
(0): ConvTranspose2d(256, 128, kernel_size=(4, 4), stride=(4, 4), bias=False)
(1): BatchNorm2d(128, eps=0.001, momentum=0.01, affine=True, track_running_stats=True)
(2): ReLU()
)
)
)
(dense_head): AnchorHeadSingle(
(cls_loss_func): SigmoidFocalClassificationLoss()
(reg_loss_func): WeightedSmoothL1Loss()
(dir_loss_func): WeightedCrossEntropyLoss()
(conv_cls): Conv2d(384, 18, kernel_size=(1, 1), stride=(1, 1))
(conv_box): Conv2d(384, 42, kernel_size=(1, 1), stride=(1, 1))
(conv_dir_cls): Conv2d(384, 12, kernel_size=(1, 1), stride=(1, 1))
)
(point_head): None
(roi_head): None
)
2024-05-20 22:53:43,714 INFO **********************Start training user/<user.name>/u11423/OpenPCDet/tools/cfgs/custom_models/pps(default)**********************
2024-05-20 23:26:21,948 INFO Train: 15/15 (100%) [ 709/960 ( 74%)] Loss: 0.3736 (0.404) LR: 6.273e-06 Time cost: 01:38/00:34 [32:38/00:34] Acc_iter 14150 Data time: 0.01(0.01) Forward time: 0.13(0.13) Batch time: 0.14(0.14)
2024-05-20 23:26:28,645 INFO Train: 15/15 (100%) [ 759/960 ( 79%)] Loss: 0.4061 (0.404) LR: 4.034e-06 Time cost: 01:44/00:27 [32:44/00:27] Acc_iter 14200 Data time: 0.00(0.01) Forward time: 0.13(0.13) Batch time: 0.13(0.14)
2024-05-20 23:26:35,310 INFO Train: 15/15 (100%) [ 809/960 ( 84%)] Loss: 0.3798 (0.404) LR: 2.290e-06 Time cost: 01:51/00:20 [32:51/00:20] Acc_iter 14250 Data time: 0.00(0.01) Forward time: 0.14(0.13) Batch time: 0.14(0.14)
2024-05-20 23:26:35,909 INFO agq005 Mon May 20 23:26:35 2024 530.30.02
[0] Quadro RTX 5000 | 65°C, 0 % | 3685 / 16384 MB | u11423(3682M)
2024-05-20 23:26:42,610 INFO Train: 15/15 (100%) [ 859/960 ( 89%)] Loss: 0.3945 (0.404) LR: 1.041e-06 Time cost: 01:58/00:13 [32:58/00:13] Acc_iter 14300 Data time: 0.00(0.01) Forward time: 0.13(0.13) Batch time: 0.13(0.14)
2024-05-20 23:26:49,242 INFO Train: 15/15 (100%) [ 909/960 ( 95%)] Loss: 0.3832 (0.403) LR: 2.879e-07 Time cost: 02:05/00:07 [33:05/00:07] Acc_iter 14350 Data time: 0.00(0.01) Forward time: 0.13(0.13) Batch time: 0.14(0.14)
2024-05-20 23:26:55,881 INFO Train: 15/15 (100%) [ 959/960 (100%)] Loss: 0.4412 (0.403) LR: 3.010e-08 Time cost: 02:12/00:00 [33:12/00:00] Acc_iter 14400 Data time: 0.00(0.01) Forward time: 0.13(0.13) Batch time: 0.13(0.14)
2024-05-20 23:26:56,471 INFO agq005 Mon May 20 23:26:56 2024 530.30.02
[0] Quadro RTX 5000 | 65°C, 0 % | 3685 / 16384 MB | u11423(3682M)
train: 0%| | 0/960 [02:12<?, ?it/s]
epochs: 100%|██████████| 15/15 [33:12<00:00, 132.88s/it]
epochs: 100%|██████████| 15/15 [33:12<00:00, 132.86s/it]
2024-05-20 23:26:56,686 INFO **********************End training user/<user.name>/u11423/OpenPCDet/tools/cfgs/custom_models/pps(default)**********************
2024-05-20 23:26:56,686 INFO **********************Start evaluation user/<user.name>/u11423/OpenPCDet/tools/cfgs/custom_models/pps(default)**********************
2024-05-20 23:26:56,855 INFO Loading Custom dataset.
2024-05-20 23:26:56,893 INFO Total samples for CUSTOM dataset: 480
2024-05-20 23:26:56,921 INFO ==> Loading parameters from checkpoint /home/uni08/hpc/<user.name>/u11423/OpenPCDet/output/user/<user.name>/u11423/OpenPCDet/tools/cfgs/custom_models/pps/default/ckpt/checkpoint_epoch_15.pth to GPU
2024-05-20 23:26:57,136 INFO ==> Checkpoint trained from version: pcdet+0.6.0+37a88e1+py4d88722
2024-05-20 23:26:57,140 INFO ==> Done (loaded 127/127)
2024-05-20 23:26:57,143 INFO *************** EPOCH 15 EVALUATION *****************
eval: 0%| | 0/120 [00:00<?, ?it/s]/home/uni08/hpc/<user.name>/u11423/OpenPCDet/tools/../pcdet/ops/iou3d_nms/iou3d_nms_utils.py:66: UserWarning: The torch.cuda.*DtypeTensor constructors are no longer recommended. It's best to use methods such as torch.tensor(data, dtype=*, device='cuda') to create tensors. (Triggered internally at ../torch/csrc/tensor/python_tensor.cpp:78.)
overlaps_bev = torch.cuda.FloatTensor(torch.Size((boxes_a.shape[0], boxes_b.shape[0]))).zero_() # (N, M)
eval: 100%|██████████| 120/120 [00:05<00:00, 21.90it/s, recall_0.3=(0, 1462) / 4482]
2024-05-20 23:27:02,622 INFO *************** Performance of EPOCH 15 *****************
2024-05-20 23:27:02,622 INFO Generate label finished(sec_per_example: 0.0114 second).
2024-05-20 23:27:02,622 INFO recall_roi_0.3: 0.000000
2024-05-20 23:27:02,622 INFO recall_rcnn_0.3: 0.326194
2024-05-20 23:27:02,622 INFO recall_roi_0.5: 0.000000
2024-05-20 23:27:02,623 INFO recall_rcnn_0.5: 0.221776
2024-05-20 23:27:02,623 INFO recall_roi_0.7: 0.000000
2024-05-20 23:27:02,623 INFO recall_rcnn_0.7: 0.075859
2024-05-20 23:27:02,623 INFO Average predicted number of objects(480 samples): 39.467
File "../pcdet/datasets/kitti/kitti_object_eval_python/eval.py", line 122:
@numba.jit(nopython=True, parallel=True)
def d3_box_overlap_kernel(boxes, qboxes, rinc, criterion=-1):
^
/user/.../.../.conda/envs/openpcd/lib/python3.8/site-packages/numba/cuda/dispatcher.py:536: NumbaPerformanceWarning: Grid size 1 will likely result in GPU under-utilization due to low occupancy.
warn(NumbaPerformanceWarning(msg))
2024-05-20 23:27:12,674 INFO Car [email protected], 0.70, 0.70:
bbox AP:86.2801, 86.2801, 86.2801
bev AP:11.3636, 11.3636, 11.3636
3d AP:9.3492, 9.3492, 9.3492
aos AP:60.26, 60.26, 60.26
Car [email protected], 0.70, 0.70:
bbox AP:87.7226, 87.7226, 87.7226
bev AP:5.3244, 5.3244, 5.3244
3d AP:2.9891, 2.9891, 2.9891
aos AP:58.33, 58.33, 58.33
Car [email protected], 0.50, 0.50:
bbox AP:86.2801, 86.2801, 86.2801
bev AP:19.6765, 19.6765, 19.6765
3d AP:18.0766, 18.0766, 18.0766
aos AP:60.26, 60.26, 60.26
Car [email protected], 0.50, 0.50:
bbox AP:87.7226, 87.7226, 87.7226
bev AP:15.1053, 15.1053, 15.1053
3d AP:13.4069, 13.4069, 13.4069
aos AP:58.33, 58.33, 58.33
Pedestrian [email protected], 0.50, 0.50:
bbox AP:54.9332, 54.9332, 54.9332
bev AP:18.6798, 18.6798, 18.6798
3d AP:5.1013, 5.1013, 5.1013
aos AP:43.72, 43.72, 43.72
Pedestrian [email protected], 0.50, 0.50:
bbox AP:53.1667, 53.1667, 53.1667
bev AP:14.1033, 14.1033, 14.1033
3d AP:2.9532, 2.9532, 2.9532
aos AP:42.44, 42.44, 42.44
Pedestrian [email protected], 0.25, 0.25:
bbox AP:54.9332, 54.9332, 54.9332
bev AP:24.7276, 24.7276, 24.7276
3d AP:23.5982, 23.5982, 23.5982
aos AP:43.72, 43.72, 43.72
Pedestrian [email protected], 0.25, 0.25:
bbox AP:53.1667, 53.1667, 53.1667
bev AP:20.6734, 20.6734, 20.6734
3d AP:19.6189, 19.6189, 19.6189
aos AP:42.44, 42.44, 42.44
Cyclist [email protected], 0.50, 0.50:
bbox AP:22.1108, 22.1108, 22.1108
bev AP:2.0695, 2.0695, 2.0695
3d AP:0.3497, 0.3497, 0.3497
aos AP:21.56, 21.56, 21.56
Cyclist [email protected], 0.50, 0.50:
bbox AP:20.5448, 20.5448, 20.5448
bev AP:1.1436, 1.1436, 1.1436
3d AP:0.1923, 0.1923, 0.1923
aos AP:19.94, 19.94, 19.94
Cyclist [email protected], 0.25, 0.25:
bbox AP:22.1108, 22.1108, 22.1108
bev AP:3.9044, 3.9044, 3.9044
3d AP:2.0695, 2.0695, 2.0695
aos AP:21.56, 21.56, 21.56
Cyclist [email protected], 0.25, 0.25:
bbox AP:20.5448, 20.5448, 20.5448
bev AP:2.2639, 2.2639, 2.2639
3d AP:1.1436, 1.1436, 1.1436
aos AP:19.94, 19.94, 19.94```
This issue is stale because it has been open for 30 days with no activity.
![github-actions[bot] avatar](https://avatars.githubusercontent.com/in/15368?v=4 "Published by a pub.dev verified publisher"){kind=small}
This issue was closed because it has been inactive for 14 days since being marked as stale.
