How to use the NuScenes dataset? Please!!!

[Wangzy-zoey]

~/3Detector/VoxelNeXt/tools: ### python3 test.py --cfg_file /home/wzy/3Detector/VoxelNeXt/tools/cfgs/nuscenes_models/cbgs_voxel0075_voxelnext.yaml --ckpt /home/wzy/3Detector/voxelnext_nuscenes_kernel1.pth

2023-08-09 20:08:21,915 INFO **Start logging** 2023-08-09 20:08:21,915 INFO CUDA_VISIBLE_DEVICES=ALL 2023-08-09 20:08:21,915 INFO cfg_file /home/wzy/3Detector/VoxelNeXt/tools/cfgs/nuscenes_models/cbgs_voxel0075_voxelnext.yaml 2023-08-09 20:08:21,915 INFO batch_size 4 2023-08-09 20:08:21,915 INFO workers 4 2023-08-09 20:08:21,915 INFO extra_tag default 2023-08-09 20:08:21,915 INFO ckpt /home/wzy/3Detector/voxelnext_nuscenes_kernel1.pth 2023-08-09 20:08:21,915 INFO pretrained_model None 2023-08-09 20:08:21,915 INFO launcher none 2023-08-09 20:08:21,915 INFO tcp_port 18888 2023-08-09 20:08:21,915 INFO local_rank 0 2023-08-09 20:08:21,915 INFO set_cfgs None 2023-08-09 20:08:21,915 INFO max_waiting_mins 30 2023-08-09 20:08:21,915 INFO start_epoch 0 2023-08-09 20:08:21,915 INFO eval_tag default 2023-08-09 20:08:21,915 INFO eval_all False 2023-08-09 20:08:21,915 INFO ckpt_dir None 2023-08-09 20:08:21,915 INFO save_to_file False 2023-08-09 20:08:21,915 INFO infer_time False 2023-08-09 20:08:21,915 INFO cfg.ROOT_DIR: /home/wzy/3Detector/VoxelNeXt 2023-08-09 20:08:21,915 INFO cfg.LOCAL_RANK: 0 2023-08-09 20:08:21,915 INFO cfg.CLASS_NAMES: ['car', 'truck', 'construction_vehicle', 'bus', 'trailer', 'barrier', 'motorcycle', 'bicycle', 'pedestrian', 'traffic_cone'] 2023-08-09 20:08:21,915 INFO ----------- DATA_CONFIG ----------- 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.DATASET: NuScenesDataset 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.DATA_PATH: ../data/nuscenes 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.VERSION: v1.0-mini 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.MAX_SWEEPS: 10 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.PRED_VELOCITY: True 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.SET_NAN_VELOCITY_TO_ZEROS: True 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.FILTER_MIN_POINTS_IN_GT: 1 2023-08-09 20:08:21,915 INFO ----------- DATA_SPLIT ----------- 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.DATA_SPLIT.train: train 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.DATA_SPLIT.test: val 2023-08-09 20:08:21,915 INFO ----------- INFO_PATH ----------- 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.INFO_PATH.train: ['nuscenes_infos_10sweeps_train.pkl'] 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.INFO_PATH.test: ['nuscenes_infos_10sweeps_val.pkl'] 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.POINT_CLOUD_RANGE: [-54.0, -54.0, -5.0, 54.0, 54.0, 3.0] 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.BALANCED_RESAMPLING: True 2023-08-09 20:08:21,915 INFO ----------- DATA_AUGMENTOR ----------- 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.DISABLE_AUG_LIST: ['placeholder'] 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.AUG_CONFIG_LIST: [{'NAME': 'gt_sampling', 'DB_INFO_PATH': ['nuscenes_dbinfos_10sweeps_withvelo.pkl'], 'USE_SHARED_MEMORY': False, 'DB_DATA_PATH': ['nuscenes_dbinfos_10sweeps_withvelo_global.pkl.npy'], 'PREPARE': {'filter_by_min_points': ['car:5', 'truck:5', 'construction_vehicle:5', 'bus:5', 'trailer:5', 'barrier:5', 'motorcycle:5', 'bicycle:5', 'pedestrian:5', 'traffic_cone:5']}, 'SAMPLE_GROUPS': ['car:2', 'truck:2', 'construction_vehicle:2', 'bus:2', 'trailer:2', 'barrier:2', 'motorcycle:2', 'bicycle:2', 'pedestrian:2', 'traffic_cone:2'], 'NUM_POINT_FEATURES': 5, 'DATABASE_WITH_FAKELIDAR': False, 'REMOVE_EXTRA_WIDTH': [0.0, 0.0, 0.0], 'LIMIT_WHOLE_SCENE': True}, {'NAME': 'random_world_flip', 'ALONG_AXIS_LIST': ['x', 'y']}, {'NAME': 'random_world_rotation', 'WORLD_ROT_ANGLE': [-0.78539816, 0.78539816]}, {'NAME': 'random_world_scaling', 'WORLD_SCALE_RANGE': [0.9, 1.1]}, {'NAME': 'random_world_translation', 'NOISE_TRANSLATE_STD': [0.5, 0.5, 0.5]}] 2023-08-09 20:08:21,915 INFO ----------- POINT_FEATURE_ENCODING ----------- 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.encoding_type: absolute_coordinates_encoding 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.used_feature_list: ['x', 'y', 'z', 'intensity', 'timestamp'] 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.src_feature_list: ['x', 'y', 'z', 'intensity', 'timestamp'] 2023-08-09 20:08:21,915 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': True}}, {'NAME': 'transform_points_to_voxels', 'VOXEL_SIZE': [0.075, 0.075, 0.2], 'MAX_POINTS_PER_VOXEL': 10, 'MAX_NUMBER_OF_VOXELS': {'train': 120000, 'test': 160000}}] 2023-08-09 20:08:21,915 INFO cfg.DATA_CONFIG._BASECONFIG: cfgs/dataset_configs/nuscenes_dataset.yaml 2023-08-09 20:08:21,915 INFO ----------- MODEL ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.NAME: VoxelNeXt 2023-08-09 20:08:21,916 INFO ----------- VFE ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.VFE.NAME: MeanVFE 2023-08-09 20:08:21,916 INFO ----------- BACKBONE_3D ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.BACKBONE_3D.NAME: VoxelResBackBone8xVoxelNeXt 2023-08-09 20:08:21,916 INFO ----------- DENSE_HEAD ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.NAME: VoxelNeXtHead 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.CLASS_AGNOSTIC: False 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.INPUT_FEATURES: 128 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.CLASS_NAMES_EACH_HEAD: [['car'], ['truck', 'construction_vehicle'], ['bus', 'trailer'], ['barrier'], ['motorcycle', 'bicycle'], ['pedestrian', 'traffic_cone']] 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SHARED_CONV_CHANNEL: 128 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.KERNEL_SIZE_HEAD: 1 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.USE_BIAS_BEFORE_NORM: True 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.NUM_HM_CONV: 2 2023-08-09 20:08:21,916 INFO ----------- SEPARATE_HEAD_CFG ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_ORDER: ['center', 'center_z', 'dim', 'rot', 'vel'] 2023-08-09 20:08:21,916 INFO ----------- HEAD_DICT ----------- 2023-08-09 20:08:21,916 INFO ----------- center ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.center.out_channels: 2 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.center.num_conv: 2 2023-08-09 20:08:21,916 INFO ----------- center_z ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.center_z.out_channels: 1 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.center_z.num_conv: 2 2023-08-09 20:08:21,916 INFO ----------- dim ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.dim.out_channels: 3 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.dim.num_conv: 2 2023-08-09 20:08:21,916 INFO ----------- rot ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.rot.out_channels: 2 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.rot.num_conv: 2 2023-08-09 20:08:21,916 INFO ----------- vel ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.vel.out_channels: 2 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.SEPARATE_HEAD_CFG.HEAD_DICT.vel.num_conv: 2 2023-08-09 20:08:21,916 INFO ----------- TARGET_ASSIGNER_CONFIG ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.FEATURE_MAP_STRIDE: 8 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NUM_MAX_OBJS: 500 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.GAUSSIAN_OVERLAP: 0.1 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.MIN_RADIUS: 2 2023-08-09 20:08:21,916 INFO ----------- LOSS_CONFIG ----------- 2023-08-09 20:08:21,916 INFO ----------- LOSS_WEIGHTS ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.cls_weight: 1.0 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.loc_weight: 0.25 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.code_weights: [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.2, 0.2, 1.0, 1.0] 2023-08-09 20:08:21,916 INFO ----------- POST_PROCESSING ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.POST_PROCESSING.SCORE_THRESH: 0.1 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.POST_PROCESSING.POST_CENTER_LIMIT_RANGE: [-61.2, -61.2, -10.0, 61.2, 61.2, 10.0] 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.POST_PROCESSING.MAX_OBJ_PER_SAMPLE: 500 2023-08-09 20:08:21,916 INFO ----------- NMS_CONFIG ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.POST_PROCESSING.NMS_CONFIG.NMS_TYPE: nms_gpu 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.POST_PROCESSING.NMS_CONFIG.NMS_THRESH: 0.2 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.POST_PROCESSING.NMS_CONFIG.NMS_PRE_MAXSIZE: 1000 2023-08-09 20:08:21,916 INFO cfg.MODEL.DENSE_HEAD.POST_PROCESSING.NMS_CONFIG.NMS_POST_MAXSIZE: 83 2023-08-09 20:08:21,916 INFO ----------- POST_PROCESSING ----------- 2023-08-09 20:08:21,916 INFO cfg.MODEL.POST_PROCESSING.RECALL_THRESH_LIST: [0.3, 0.5, 0.7] 2023-08-09 20:08:21,916 INFO cfg.MODEL.POST_PROCESSING.EVAL_METRIC: kitti 2023-08-09 20:08:21,916 INFO ----------- OPTIMIZATION ----------- 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.BATCH_SIZE_PER_GPU: 4 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.NUM_EPOCHS: 20 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.OPTIMIZER: adam_onecycle 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.LR: 0.001 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.WEIGHT_DECAY: 0.01 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.MOMENTUM: 0.9 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.MOMS: [0.95, 0.85] 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.PCT_START: 0.4 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.DIV_FACTOR: 10 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.DECAY_STEP_LIST: [35, 45] 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.LR_DECAY: 0.1 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.LR_CLIP: 1e-07 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.LR_WARMUP: False 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.WARMUP_EPOCH: 1 2023-08-09 20:08:21,916 INFO cfg.OPTIMIZATION.GRAD_NORM_CLIP: 10 2023-08-09 20:08:21,917 INFO cfg.TAG: cbgs_voxel0075_voxelnext 2023-08-09 20:08:21,917 INFO cfg.EXP_GROUP_PATH: home/wzy/3Detector/VoxelNeXt/tools/cfgs/nuscenes_models 2023-08-09 20:08:21,917 INFO Loading NuScenes dataset 2023-08-09 20:08:21,917 INFO Total samples for NuScenes dataset: 0 2023-08-09 20:08:25,209 INFO ==> Loading parameters from checkpoint /home/wzy/3Detector/voxelnext_nuscenes_kernel1.pth to GPU 2023-08-09 20:08:25,240 INFO ==> Done (loaded 542/542) 2023-08-09 20:08:25,251 INFO * EPOCH 1 EVALUATION *** eval: 0it [00:00, ?it/s] 2023-08-09 20:08:25,420 INFO * Performance of EPOCH 1 *** Traceback (most recent call last): File "test.py", line 207, in main() File "test.py", line 203, in main eval_single_ckpt(model, test_loader, args, eval_output_dir, logger, epoch_id, dist_test=dist_test) File "test.py", line 65, in eval_single_ckpt eval_utils.eval_one_epoch( File "/home/wzy/3Detector/VoxelNeXt/tools/eval_utils/eval_utils.py", line 94, in eval_one_epoch sec_per_example = (time.time() - start_time) / len(dataloader.dataset)

ZeroDivisionError: float division by zero

[Wangzy-zoey]

@yukang2017 Please help me !!! Please!!!

[Zynade]

I attempted to replicate the results on nuScenes test dataset, and I ran into the same issue. Have you found a fix?

[kbar2222]

hello, have you fix this problem plz ?

[pbrandstet]

do you use the v1.0-mini dataset from nuscenes? if so, you need to change the version. In tools/cfgs/dataset_configs/nuscenes_dataset.yaml at line 4 you see VERSION: 'v1.0-trainval' if you use mini you need to change it to VERSION: 'v1.0-mini'

[rpf1019]

do you use the v1.0-mini dataset from nuscenes? if so, you need to change the version. In tools/cfgs/dataset_configs/nuscenes_dataset.yaml at line 4 you see VERSION: 'v1.0-trainval' if you use mini you need to change it to VERSION: 'v1.0-mini'

i tried this but it did not work, is it worked for u?

[pbrandstet]

Yes, after that change it worked for me. You could use the debugger to check if the dataset really got loaded.