diff --git a/nohup.out b/nohup.out new file mode 100644 index 0000000..e19b59d --- /dev/null +++ b/nohup.out @@ -0,0 +1,2988 @@ +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/launch.py:163: DeprecationWarning: The 'warn' method is deprecated, use 'warning' instead + logger.warn( +The module torch.distributed.launch is deprecated and going to be removed in future.Migrate to torch.distributed.run +WARNING:torch.distributed.run:--use_env is deprecated and will be removed in future releases. + Please read local_rank from `os.environ('LOCAL_RANK')` instead. +INFO:torch.distributed.launcher.api:Starting elastic_operator with launch configs: + entrypoint : tools/train.py + min_nodes : 1 + max_nodes : 1 + nproc_per_node : 2 + run_id : none + rdzv_backend : static + rdzv_endpoint : 127.0.0.1:29500 + rdzv_configs : {'rank': 0, 'timeout': 900} + max_restarts : 3 + monitor_interval : 5 + log_dir : None + metrics_cfg : {} + +INFO:torch.distributed.elastic.agent.server.local_elastic_agent:log directory set to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf +INFO:torch.distributed.elastic.agent.server.api:[default] starting workers for entrypoint: python +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous'ing worker group +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/elastic/utils/store.py:52: FutureWarning: This is an experimental API and will be changed in future. + warnings.warn( +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous complete for workers. Result: + restart_count=0 + master_addr=127.0.0.1 + master_port=29500 + group_rank=0 + group_world_size=1 + local_ranks=[0, 1] + role_ranks=[0, 1] + global_ranks=[0, 1] + role_world_sizes=[2, 2] + global_world_sizes=[2, 2] + +INFO:torch.distributed.elastic.agent.server.api:[default] Starting worker group +INFO:torch.distributed.elastic.multiprocessing:Setting worker0 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_0/0/error.json +INFO:torch.distributed.elastic.multiprocessing:Setting worker1 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_0/1/error.json +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +10/10 17:31:07 - mmengine - INFO - +------------------------------------------------------------ +System environment: + sys.platform: linux + Python: 3.8.18 (default, Sep 11 2023, 13:40:15) [GCC 11.2.0] + CUDA available: True + numpy_random_seed: 496659845 + GPU 0,1: NVIDIA GeForce RTX 4090 + CUDA_HOME: /usr/local/cuda-11.7 + NVCC: Cuda compilation tools, release 11.7, V11.7.64 + GCC: gcc (Ubuntu 9.5.0-3ubuntu1) 9.5.0 + PyTorch: 1.9.0+cu111 + PyTorch compiling details: PyTorch built with: + - GCC 7.3 + - C++ Version: 201402 + - Intel(R) Math Kernel Library Version 2020.0.0 Product Build 20191122 for Intel(R) 64 architecture applications + - Intel(R) MKL-DNN v2.1.2 (Git Hash 98be7e8afa711dc9b66c8ff3504129cb82013cdb) + - OpenMP 201511 (a.k.a. OpenMP 4.5) + - NNPACK is enabled + - CPU capability usage: AVX2 + - CUDA Runtime 11.1 + - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86 + - CuDNN 8.0.5 + - Magma 2.5.2 + - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.1, CUDNN_VERSION=8.0.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.9.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, + + TorchVision: 0.10.0+cu111 + OpenCV: 4.8.0 + MMEngine: 0.8.4 + +Runtime environment: + cudnn_benchmark: False + mp_cfg: {'mp_start_method': 'fork', 'opencv_num_threads': 0} + dist_cfg: {'backend': 'nccl'} + seed: 496659845 + Distributed launcher: pytorch + Distributed training: True + GPU number: 2 +------------------------------------------------------------ + +10/10 17:31:11 - mmengine - INFO - Config: +backend_args = None +class_names = [ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', +] +custom_imports = dict(imports=[ + 'projects.DETR3D.detr3d', +]) +data_prefix = dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts='') +data_root = 'data/nuscenes/' +dataset_type = 'NuScenesDataset' +default_hooks = dict( + checkpoint=dict( + _scope_='mmdet3d', + interval=1, + max_keep_ckpts=1, + save_last=True, + type='CheckpointHook'), + logger=dict(_scope_='mmdet3d', interval=50, type='LoggerHook'), + param_scheduler=dict(_scope_='mmdet3d', type='ParamSchedulerHook'), + sampler_seed=dict(_scope_='mmdet3d', type='DistSamplerSeedHook'), + timer=dict(_scope_='mmdet3d', type='IterTimerHook'), + visualization=dict(_scope_='mmdet3d', type='Det3DVisualizationHook')) +default_scope = 'mmdet3d' +env_cfg = dict( + cudnn_benchmark=False, + dist_cfg=dict(backend='nccl'), + mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0)) +img_norm_cfg = dict( + bgr_to_rgb=False, mean=[ + 103.53, + 116.28, + 123.675, + ], std=[ + 1.0, + 1.0, + 1.0, + ]) +input_modality = dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False) +lang_model_name = 'bert-base-uncased' +launcher = 'pytorch' +load_from = 'pretrained/fcos3d.pth' +log_level = 'INFO' +log_processor = dict( + _scope_='mmdet3d', by_epoch=True, type='LogProcessor', window_size=50) +metainfo = dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', +]) +model = dict( + data_preprocessor=dict( + bgr_to_rgb=False, + mean=[ + 103.53, + 116.28, + 123.675, + ], + pad_size_divisor=32, + std=[ + 1.0, + 1.0, + 1.0, + ], + type='Det3DDataPreprocessor'), + encoder=dict( + fusion_layer_cfg=dict( + embed_dim=1024, + init_values=0.0001, + l_dim=256, + num_heads=4, + v_dim=256), + layer_cfg=dict( + ffn_cfg=dict( + embed_dims=256, feedforward_channels=2048, ffn_drop=0.0), + self_attn_cfg=dict(dropout=0.0, embed_dims=256, num_levels=4)), + num_cp=6, + num_layers=6, + text_layer_cfg=dict( + ffn_cfg=dict( + embed_dims=256, feedforward_channels=1024, ffn_drop=0.0), + self_attn_cfg=dict(dropout=0.0, embed_dims=256, num_heads=4))), + img_backbone=dict( + dcn=dict(deform_groups=1, fallback_on_stride=False, type='DCNv2'), + depth=50, + frozen_stages=1, + norm_cfg=dict(requires_grad=False, type='BN2d'), + norm_eval=True, + num_stages=4, + out_indices=( + 0, + 1, + 2, + 3, + ), + stage_with_dcn=( + False, + False, + True, + True, + ), + style='caffe', + type='mmdet.ResNet'), + img_neck=dict( + add_extra_convs='on_output', + in_channels=[ + 256, + 512, + 1024, + 2048, + ], + num_outs=4, + out_channels=256, + relu_before_extra_convs=True, + start_level=1, + type='mmdet.FPN'), + language_model=dict( + add_pooling_layer=False, + name='bert-base-uncased', + pad_to_max=False, + special_tokens_list=[ + '[CLS]', + '[SEP]', + '.', + '?', + ], + type='BertModel', + use_sub_sentence_represent=True), + positional_encoding_single=dict( + normalize=True, num_feats=128, offset=0.0, temperature=20), + pts_bbox_head=dict( + as_two_stage=False, + bbox_coder=dict( + max_num=300, + num_classes=10, + pc_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + post_center_range=[ + -61.2, + -61.2, + -10.0, + 61.2, + 61.2, + 10.0, + ], + type='NMSFreeCoder', + voxel_size=[ + 0.2, + 0.2, + 8, + ]), + in_channels=256, + loss_bbox=dict(loss_weight=0.25, type='mmdet.L1Loss'), + loss_cls=dict( + alpha=0.25, + gamma=2.0, + loss_weight=2.0, + type='mmdet.FocalLoss', + use_sigmoid=True), + loss_iou=dict(loss_weight=0.0, type='mmdet.GIoULoss'), + num_classes=10, + num_query=900, + positional_encoding=dict( + normalize=True, + num_feats=128, + offset=-0.5, + type='mmdet.SinePositionalEncoding'), + sync_cls_avg_factor=True, + transformer=dict( + decoder=dict( + num_layers=6, + return_intermediate=True, + transformerlayers=dict( + attn_cfgs=[ + dict( + dropout=0.1, + embed_dims=256, + num_heads=8, + type='MultiheadAttention'), + dict( + dropout=0.0, + embed_dims=256, + num_heads=8, + type='MultiheadAttention'), + dict( + embed_dims=256, + num_points=1, + pc_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + type='Detr3DCrossAtten'), + ], + feedforward_channels=512, + ffn_dropout=0.1, + operation_order=( + 'self_attn', + 'norm', + 'self_attn_text', + 'norm', + 'cross_attn', + 'norm', + 'ffn', + 'norm', + ), + type='mmdet.DetrTransformerDecoderLayer'), + type='Detr3DTransformerDecoder'), + type='Detr3DTransformer'), + type='DETR3DHead', + with_box_refine=True), + train_cfg=dict( + pts=dict( + assigner=dict( + cls_cost=dict(type='mmdet.FocalLossCost', weight=2.0), + iou_cost=dict(type='mmdet.IoUCost', weight=0.0), + pc_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + reg_cost=dict(type='BBox3DL1Cost', weight=0.25), + type='HungarianAssigner3D'), + grid_size=[ + 512, + 512, + 1, + ], + out_size_factor=4, + point_cloud_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + voxel_size=[ + 0.2, + 0.2, + 8, + ])), + type='DETR3D', + use_grid_mask=True) +optim_wrapper = dict( + clip_grad=dict(max_norm=35, norm_type=2), + optimizer=dict(lr=0.0002, type='AdamW', weight_decay=0.01), + paramwise_cfg=dict(custom_keys=dict(img_backbone=dict(lr_mult=0.1))), + type='OptimWrapper') +param_scheduler = [ + dict( + begin=0, + by_epoch=False, + end=500, + start_factor=0.3333333333333333, + type='LinearLR'), + dict( + T_max=24, + begin=0, + by_epoch=True, + end=24, + eta_min_ratio=0.001, + type='CosineAnnealingLR'), +] +point_cloud_range = [ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, +] +resume = False +test_cfg = dict(type='TestLoop') +test_dataloader = dict( + batch_size=1, + dataset=dict( + ann_file='nuscenes_infos_val.pkl', + backend_args=None, + box_type_3d='LiDAR', + data_prefix=dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts=''), + data_root='data/nuscenes/', + load_type='frame_based', + metainfo=dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ]), + modality=dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False), + pipeline=[ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + transforms=[ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict(keys=[ + 'img', + ], type='Pack3DDetInputs'), + ], + test_mode=True, + type='NuScenesDataset'), + drop_last=False, + num_workers=4, + persistent_workers=True, + sampler=dict(shuffle=False, type='DefaultSampler')) +test_evaluator = dict( + ann_file='data/nuscenes/nuscenes_infos_val.pkl', + backend_args=None, + data_root='data/nuscenes/', + metric='bbox', + type='NuScenesMetric') +test_pipeline = [ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + transforms=[ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict(keys=[ + 'img', + ], type='Pack3DDetInputs'), +] +test_transforms = [ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), +] +total_epochs = 24 +train_cfg = dict(max_epochs=24, type='EpochBasedTrainLoop', val_interval=2) +train_dataloader = dict( + batch_size=1, + dataset=dict( + ann_file='nuscenes_infos_train.pkl', + backend_args=None, + box_type_3d='LiDAR', + data_prefix=dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts=''), + data_root='data/nuscenes/', + load_interval=2, + load_type='frame_based', + metainfo=dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ]), + modality=dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False), + pipeline=[ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + type='LoadAnnotations3D', + with_attr_label=False, + with_bbox_3d=True, + with_label_3d=True), + dict( + transforms=[ + dict(type='PhotoMetricDistortion3D'), + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict( + point_cloud_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + type='ObjectRangeFilter'), + dict( + classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ], + type='ObjectNameFilter'), + dict( + keys=[ + 'img', + 'gt_bboxes_3d', + 'gt_labels_3d', + ], + type='Pack3DDetInputs'), + ], + test_mode=False, + type='NuScenesDataset'), + drop_last=False, + num_workers=4, + persistent_workers=True, + sampler=dict(shuffle=True, type='DefaultSampler')) +train_pipeline = [ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + type='LoadAnnotations3D', + with_attr_label=False, + with_bbox_3d=True, + with_label_3d=True), + dict( + transforms=[ + dict(type='PhotoMetricDistortion3D'), + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict( + point_cloud_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + type='ObjectRangeFilter'), + dict( + classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ], + type='ObjectNameFilter'), + dict( + keys=[ + 'img', + 'gt_bboxes_3d', + 'gt_labels_3d', + ], type='Pack3DDetInputs'), +] +train_transforms = [ + dict(type='PhotoMetricDistortion3D'), + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), +] +val_cfg = dict(type='ValLoop') +val_dataloader = dict( + batch_size=1, + dataset=dict( + ann_file='nuscenes_infos_val.pkl', + backend_args=None, + box_type_3d='LiDAR', + data_prefix=dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts=''), + data_root='data/nuscenes/', + load_type='frame_based', + metainfo=dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ]), + modality=dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False), + pipeline=[ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + transforms=[ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict(keys=[ + 'img', + ], type='Pack3DDetInputs'), + ], + test_mode=True, + type='NuScenesDataset'), + drop_last=False, + num_workers=4, + persistent_workers=True, + sampler=dict(shuffle=False, type='DefaultSampler')) +val_evaluator = dict( + ann_file='data/nuscenes/nuscenes_infos_val.pkl', + backend_args=None, + data_root='data/nuscenes/', + metric='bbox', + type='NuScenesMetric') +vis_backends = [ + dict(type='TensorboardVisBackend'), +] +visualizer = dict( + name='visualizer', + type='Det3DLocalVisualizer', + vis_backends=[ + dict(type='TensorboardVisBackend'), + ]) +voxel_size = [ + 0.2, + 0.2, + 8, +] +work_dir = './work_dirs/detr3d_r50_bert_gridmask_halfdata_decoder' + +10/10 17:31:25 - mmengine - INFO - Autoplay mode, press [SPACE] to pause. +10/10 17:31:25 - mmengine - INFO - Hooks will be executed in the following order: +before_run: +(VERY_HIGH ) RuntimeInfoHook +(BELOW_NORMAL) LoggerHook + -------------------- +before_train: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(VERY_LOW ) CheckpointHook + -------------------- +before_train_epoch: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(NORMAL ) DistSamplerSeedHook + -------------------- +before_train_iter: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook + -------------------- +after_train_iter: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(BELOW_NORMAL) LoggerHook +(LOW ) ParamSchedulerHook +(VERY_LOW ) CheckpointHook + -------------------- +after_train_epoch: +(NORMAL ) IterTimerHook +(LOW ) ParamSchedulerHook +(VERY_LOW ) CheckpointHook + -------------------- +before_val: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +before_val_epoch: +(NORMAL ) IterTimerHook + -------------------- +before_val_iter: +(NORMAL ) IterTimerHook + -------------------- +after_val_iter: +(NORMAL ) IterTimerHook +(NORMAL ) Det3DVisualizationHook +(BELOW_NORMAL) LoggerHook + -------------------- +after_val_epoch: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(BELOW_NORMAL) LoggerHook +(LOW ) ParamSchedulerHook +(VERY_LOW ) CheckpointHook + -------------------- +after_val: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +after_train: +(VERY_HIGH ) RuntimeInfoHook +(VERY_LOW ) CheckpointHook + -------------------- +before_test: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +before_test_epoch: +(NORMAL ) IterTimerHook + -------------------- +before_test_iter: +(NORMAL ) IterTimerHook + -------------------- +after_test_iter: +(NORMAL ) IterTimerHook +(NORMAL ) Det3DVisualizationHook +(BELOW_NORMAL) LoggerHook + -------------------- +after_test_epoch: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(BELOW_NORMAL) LoggerHook + -------------------- +after_test: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +after_run: +(BELOW_NORMAL) LoggerHook + -------------------- +10/10 17:34:34 - mmengine - INFO - ------------------------------ +10/10 17:34:34 - mmengine - INFO - The length of the dataset: 14065 +10/10 17:34:34 - mmengine - INFO - The number of instances per category in the dataset: ++----------------------+--------+ +| category | number | ++----------------------+--------+ +| car | 206689 | +| truck | 36393 | +| construction_vehicle | 5990 | +| bus | 6576 | +| trailer | 10349 | +| barrier | 62576 | +| motorcycle | 5051 | +| bicycle | 4740 | +| pedestrian | 92925 | +| traffic_cone | 41120 | ++----------------------+--------+ +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.downsample.0.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.downsample.0.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.downsample.0.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.downsample.0.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.downsample.0.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.downsample.0.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.downsample.0.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.downsample.0.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.downsample.0.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv3.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv1.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv1.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv1.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.weight:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.bias:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.bias:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.bias:lr_mult=0.1 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv3.weight:lr=2e-05 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv3.weight:weight_decay=0.01 +10/10 17:34:34 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv3.weight:lr_mult=0.1 +/home/xzt/mmdetection3d-1.1.0/mmdet3d/evaluation/functional/kitti_utils/eval.py:10: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details. + def get_thresholds(scores: np.ndarray, num_gt, num_sample_pts=41): +Loads checkpoint by local backend from path: pretrained/fcos3d.pth +10/10 17:35:32 - mmengine - INFO - ------------------------------ +10/10 17:35:32 - mmengine - INFO - The length of the dataset: 6019 +10/10 17:35:32 - mmengine - INFO - The number of instances per category in the dataset: ++----------------------+--------+ +| category | number | ++----------------------+--------+ +| car | 80004 | +| truck | 15704 | +| construction_vehicle | 2678 | +| bus | 3158 | +| trailer | 4159 | +| barrier | 26992 | +| motorcycle | 2508 | +| bicycle | 2381 | +| pedestrian | 34347 | +| traffic_cone | 15597 | ++----------------------+--------+ +/home/xzt/mmdetection3d-1.1.0/mmdet3d/evaluation/functional/kitti_utils/eval.py:10: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details. + def get_thresholds(scores: np.ndarray, num_gt, num_sample_pts=41): +Loads checkpoint by local backend from path: pretrained/fcos3d.pth +The model and loaded state dict do not match exactly + +unexpected key in source state_dict: bbox_head.bld_alpha, bbox_head.cls_convs.0.conv.weight, bbox_head.cls_convs.0.conv.bias, bbox_head.cls_convs.0.gn.weight, bbox_head.cls_convs.0.gn.bias, bbox_head.cls_convs.1.conv.weight, bbox_head.cls_convs.1.conv.bias, bbox_head.cls_convs.1.conv.conv_offset.weight, bbox_head.cls_convs.1.conv.conv_offset.bias, bbox_head.cls_convs.1.gn.weight, bbox_head.cls_convs.1.gn.bias, bbox_head.reg_convs.0.conv.weight, bbox_head.reg_convs.0.conv.bias, bbox_head.reg_convs.0.gn.weight, bbox_head.reg_convs.0.gn.bias, bbox_head.reg_convs.1.conv.weight, bbox_head.reg_convs.1.conv.bias, bbox_head.reg_convs.1.conv.conv_offset.weight, bbox_head.reg_convs.1.conv.conv_offset.bias, bbox_head.reg_convs.1.gn.weight, bbox_head.reg_convs.1.gn.bias, bbox_head.conv_cls_prev.0.conv.weight, bbox_head.conv_cls_prev.0.conv.bias, bbox_head.conv_cls_prev.0.gn.weight, bbox_head.conv_cls_prev.0.gn.bias, bbox_head.conv_cls.weight, bbox_head.conv_cls.bias, bbox_head.conv_reg_prevs.0.0.conv.weight, bbox_head.conv_reg_prevs.0.0.conv.bias, bbox_head.conv_reg_prevs.0.0.gn.weight, bbox_head.conv_reg_prevs.0.0.gn.bias, bbox_head.conv_reg_prevs.1.0.conv.weight, bbox_head.conv_reg_prevs.1.0.conv.bias, bbox_head.conv_reg_prevs.1.0.gn.weight, bbox_head.conv_reg_prevs.1.0.gn.bias, bbox_head.conv_reg_prevs.2.0.conv.weight, bbox_head.conv_reg_prevs.2.0.conv.bias, bbox_head.conv_reg_prevs.2.0.gn.weight, bbox_head.conv_reg_prevs.2.0.gn.bias, bbox_head.conv_reg_prevs.3.0.conv.weight, bbox_head.conv_reg_prevs.3.0.conv.bias, bbox_head.conv_reg_prevs.3.0.gn.weight, bbox_head.conv_reg_prevs.3.0.gn.bias, bbox_head.conv_reg_prevs.5.0.conv.weight, bbox_head.conv_reg_prevs.5.0.conv.bias, bbox_head.conv_reg_prevs.5.0.gn.weight, bbox_head.conv_reg_prevs.5.0.gn.bias, bbox_head.conv_regs.0.weight, bbox_head.conv_regs.0.bias, bbox_head.conv_regs.1.weight, bbox_head.conv_regs.1.bias, bbox_head.conv_regs.2.weight, bbox_head.conv_regs.2.bias, bbox_head.conv_regs.3.weight, bbox_head.conv_regs.3.bias, bbox_head.conv_regs.4.weight, bbox_head.conv_regs.4.bias, bbox_head.conv_regs.5.weight, bbox_head.conv_regs.5.bias, bbox_head.conv_dir_cls_prev.0.conv.weight, bbox_head.conv_dir_cls_prev.0.conv.bias, bbox_head.conv_dir_cls_prev.0.gn.weight, bbox_head.conv_dir_cls_prev.0.gn.bias, bbox_head.conv_dir_cls.weight, bbox_head.conv_dir_cls.bias, bbox_head.conv_attr_prev.0.conv.weight, bbox_head.conv_attr_prev.0.conv.bias, bbox_head.conv_attr_prev.0.gn.weight, bbox_head.conv_attr_prev.0.gn.bias, bbox_head.conv_attr.weight, bbox_head.conv_attr.bias, bbox_head.conv_depth_cls_prev.0.conv.weight, bbox_head.conv_depth_cls_prev.0.conv.bias, bbox_head.conv_depth_cls_prev.0.gn.weight, bbox_head.conv_depth_cls_prev.0.gn.bias, bbox_head.conv_depth_cls.weight, bbox_head.conv_depth_cls.bias, bbox_head.conv_centerness_prev.0.conv.weight, bbox_head.conv_centerness_prev.0.conv.bias, bbox_head.conv_centerness_prev.0.gn.weight, bbox_head.conv_centerness_prev.0.gn.bias, bbox_head.conv_centerness.weight, bbox_head.conv_centerness.bias, bbox_head.scales.0.0.scale, bbox_head.scales.0.1.scale, bbox_head.scales.0.2.scale, bbox_head.scales.0.3.scale, bbox_head.scales.1.0.scale, bbox_head.scales.1.1.scale, bbox_head.scales.1.2.scale, bbox_head.scales.1.3.scale, bbox_head.scales.2.0.scale, bbox_head.scales.2.1.scale, bbox_head.scales.2.2.scale, bbox_head.scales.2.3.scale, bbox_head.scales.3.0.scale, bbox_head.scales.3.1.scale, bbox_head.scales.3.2.scale, bbox_head.scales.3.3.scale, bbox_head.scales.4.0.scale, bbox_head.scales.4.1.scale, bbox_head.scales.4.2.scale, bbox_head.scales.4.3.scale, img_backbone.layer3.6.conv1.weight, img_backbone.layer3.6.bn1.weight, img_backbone.layer3.6.bn1.bias, img_backbone.layer3.6.bn1.running_mean, img_backbone.layer3.6.bn1.running_var, img_backbone.layer3.6.bn1.num_batches_tracked, img_backbone.layer3.6.conv2.weight, img_backbone.layer3.6.conv2.conv_offset.weight, img_backbone.layer3.6.conv2.conv_offset.bias, img_backbone.layer3.6.bn2.weight, img_backbone.layer3.6.bn2.bias, img_backbone.layer3.6.bn2.running_mean, img_backbone.layer3.6.bn2.running_var, img_backbone.layer3.6.bn2.num_batches_tracked, img_backbone.layer3.6.conv3.weight, img_backbone.layer3.6.bn3.weight, img_backbone.layer3.6.bn3.bias, img_backbone.layer3.6.bn3.running_mean, img_backbone.layer3.6.bn3.running_var, img_backbone.layer3.6.bn3.num_batches_tracked, img_backbone.layer3.7.conv1.weight, img_backbone.layer3.7.bn1.weight, img_backbone.layer3.7.bn1.bias, img_backbone.layer3.7.bn1.running_mean, img_backbone.layer3.7.bn1.running_var, img_backbone.layer3.7.bn1.num_batches_tracked, img_backbone.layer3.7.conv2.weight, img_backbone.layer3.7.conv2.conv_offset.weight, img_backbone.layer3.7.conv2.conv_offset.bias, img_backbone.layer3.7.bn2.weight, img_backbone.layer3.7.bn2.bias, img_backbone.layer3.7.bn2.running_mean, img_backbone.layer3.7.bn2.running_var, img_backbone.layer3.7.bn2.num_batches_tracked, img_backbone.layer3.7.conv3.weight, img_backbone.layer3.7.bn3.weight, img_backbone.layer3.7.bn3.bias, img_backbone.layer3.7.bn3.running_mean, img_backbone.layer3.7.bn3.running_var, img_backbone.layer3.7.bn3.num_batches_tracked, img_backbone.layer3.8.conv1.weight, img_backbone.layer3.8.bn1.weight, img_backbone.layer3.8.bn1.bias, img_backbone.layer3.8.bn1.running_mean, img_backbone.layer3.8.bn1.running_var, img_backbone.layer3.8.bn1.num_batches_tracked, img_backbone.layer3.8.conv2.weight, img_backbone.layer3.8.conv2.conv_offset.weight, img_backbone.layer3.8.conv2.conv_offset.bias, img_backbone.layer3.8.bn2.weight, img_backbone.layer3.8.bn2.bias, img_backbone.layer3.8.bn2.running_mean, img_backbone.layer3.8.bn2.running_var, img_backbone.layer3.8.bn2.num_batches_tracked, img_backbone.layer3.8.conv3.weight, img_backbone.layer3.8.bn3.weight, img_backbone.layer3.8.bn3.bias, img_backbone.layer3.8.bn3.running_mean, img_backbone.layer3.8.bn3.running_var, img_backbone.layer3.8.bn3.num_batches_tracked, img_backbone.layer3.9.conv1.weight, img_backbone.layer3.9.bn1.weight, img_backbone.layer3.9.bn1.bias, img_backbone.layer3.9.bn1.running_mean, img_backbone.layer3.9.bn1.running_var, img_backbone.layer3.9.bn1.num_batches_tracked, img_backbone.layer3.9.conv2.weight, img_backbone.layer3.9.conv2.conv_offset.weight, img_backbone.layer3.9.conv2.conv_offset.bias, img_backbone.layer3.9.bn2.weight, img_backbone.layer3.9.bn2.bias, img_backbone.layer3.9.bn2.running_mean, img_backbone.layer3.9.bn2.running_var, img_backbone.layer3.9.bn2.num_batches_tracked, img_backbone.layer3.9.conv3.weight, img_backbone.layer3.9.bn3.weight, img_backbone.layer3.9.bn3.bias, img_backbone.layer3.9.bn3.running_mean, img_backbone.layer3.9.bn3.running_var, img_backbone.layer3.9.bn3.num_batches_tracked, img_backbone.layer3.10.conv1.weight, img_backbone.layer3.10.bn1.weight, img_backbone.layer3.10.bn1.bias, img_backbone.layer3.10.bn1.running_mean, img_backbone.layer3.10.bn1.running_var, img_backbone.layer3.10.bn1.num_batches_tracked, img_backbone.layer3.10.conv2.weight, img_backbone.layer3.10.conv2.conv_offset.weight, img_backbone.layer3.10.conv2.conv_offset.bias, img_backbone.layer3.10.bn2.weight, img_backbone.layer3.10.bn2.bias, img_backbone.layer3.10.bn2.running_mean, img_backbone.layer3.10.bn2.running_var, img_backbone.layer3.10.bn2.num_batches_tracked, img_backbone.layer3.10.conv3.weight, img_backbone.layer3.10.bn3.weight, img_backbone.layer3.10.bn3.bias, img_backbone.layer3.10.bn3.running_mean, img_backbone.layer3.10.bn3.running_var, img_backbone.layer3.10.bn3.num_batches_tracked, img_backbone.layer3.11.conv1.weight, img_backbone.layer3.11.bn1.weight, img_backbone.layer3.11.bn1.bias, img_backbone.layer3.11.bn1.running_mean, img_backbone.layer3.11.bn1.running_var, img_backbone.layer3.11.bn1.num_batches_tracked, img_backbone.layer3.11.conv2.weight, img_backbone.layer3.11.conv2.conv_offset.weight, img_backbone.layer3.11.conv2.conv_offset.bias, img_backbone.layer3.11.bn2.weight, img_backbone.layer3.11.bn2.bias, img_backbone.layer3.11.bn2.running_mean, img_backbone.layer3.11.bn2.running_var, img_backbone.layer3.11.bn2.num_batches_tracked, img_backbone.layer3.11.conv3.weight, img_backbone.layer3.11.bn3.weight, img_backbone.layer3.11.bn3.bias, img_backbone.layer3.11.bn3.running_mean, img_backbone.layer3.11.bn3.running_var, img_backbone.layer3.11.bn3.num_batches_tracked, img_backbone.layer3.12.conv1.weight, img_backbone.layer3.12.bn1.weight, img_backbone.layer3.12.bn1.bias, img_backbone.layer3.12.bn1.running_mean, img_backbone.layer3.12.bn1.running_var, img_backbone.layer3.12.bn1.num_batches_tracked, img_backbone.layer3.12.conv2.weight, img_backbone.layer3.12.conv2.conv_offset.weight, img_backbone.layer3.12.conv2.conv_offset.bias, img_backbone.layer3.12.bn2.weight, img_backbone.layer3.12.bn2.bias, img_backbone.layer3.12.bn2.running_mean, img_backbone.layer3.12.bn2.running_var, img_backbone.layer3.12.bn2.num_batches_tracked, img_backbone.layer3.12.conv3.weight, img_backbone.layer3.12.bn3.weight, img_backbone.layer3.12.bn3.bias, img_backbone.layer3.12.bn3.running_mean, img_backbone.layer3.12.bn3.running_var, img_backbone.layer3.12.bn3.num_batches_tracked, img_backbone.layer3.13.conv1.weight, img_backbone.layer3.13.bn1.weight, img_backbone.layer3.13.bn1.bias, img_backbone.layer3.13.bn1.running_mean, img_backbone.layer3.13.bn1.running_var, img_backbone.layer3.13.bn1.num_batches_tracked, img_backbone.layer3.13.conv2.weight, img_backbone.layer3.13.conv2.conv_offset.weight, img_backbone.layer3.13.conv2.conv_offset.bias, img_backbone.layer3.13.bn2.weight, img_backbone.layer3.13.bn2.bias, img_backbone.layer3.13.bn2.running_mean, img_backbone.layer3.13.bn2.running_var, img_backbone.layer3.13.bn2.num_batches_tracked, img_backbone.layer3.13.conv3.weight, img_backbone.layer3.13.bn3.weight, img_backbone.layer3.13.bn3.bias, img_backbone.layer3.13.bn3.running_mean, img_backbone.layer3.13.bn3.running_var, img_backbone.layer3.13.bn3.num_batches_tracked, img_backbone.layer3.14.conv1.weight, img_backbone.layer3.14.bn1.weight, img_backbone.layer3.14.bn1.bias, img_backbone.layer3.14.bn1.running_mean, img_backbone.layer3.14.bn1.running_var, img_backbone.layer3.14.bn1.num_batches_tracked, img_backbone.layer3.14.conv2.weight, img_backbone.layer3.14.conv2.conv_offset.weight, img_backbone.layer3.14.conv2.conv_offset.bias, img_backbone.layer3.14.bn2.weight, img_backbone.layer3.14.bn2.bias, img_backbone.layer3.14.bn2.running_mean, img_backbone.layer3.14.bn2.running_var, img_backbone.layer3.14.bn2.num_batches_tracked, img_backbone.layer3.14.conv3.weight, img_backbone.layer3.14.bn3.weight, img_backbone.layer3.14.bn3.bias, img_backbone.layer3.14.bn3.running_mean, img_backbone.layer3.14.bn3.running_var, img_backbone.layer3.14.bn3.num_batches_tracked, img_backbone.layer3.15.conv1.weight, img_backbone.layer3.15.bn1.weight, img_backbone.layer3.15.bn1.bias, img_backbone.layer3.15.bn1.running_mean, img_backbone.layer3.15.bn1.running_var, img_backbone.layer3.15.bn1.num_batches_tracked, img_backbone.layer3.15.conv2.weight, img_backbone.layer3.15.conv2.conv_offset.weight, img_backbone.layer3.15.conv2.conv_offset.bias, img_backbone.layer3.15.bn2.weight, img_backbone.layer3.15.bn2.bias, img_backbone.layer3.15.bn2.running_mean, img_backbone.layer3.15.bn2.running_var, img_backbone.layer3.15.bn2.num_batches_tracked, img_backbone.layer3.15.conv3.weight, img_backbone.layer3.15.bn3.weight, img_backbone.layer3.15.bn3.bias, img_backbone.layer3.15.bn3.running_mean, img_backbone.layer3.15.bn3.running_var, img_backbone.layer3.15.bn3.num_batches_tracked, img_backbone.layer3.16.conv1.weight, img_backbone.layer3.16.bn1.weight, img_backbone.layer3.16.bn1.bias, img_backbone.layer3.16.bn1.running_mean, img_backbone.layer3.16.bn1.running_var, img_backbone.layer3.16.bn1.num_batches_tracked, img_backbone.layer3.16.conv2.weight, img_backbone.layer3.16.conv2.conv_offset.weight, img_backbone.layer3.16.conv2.conv_offset.bias, img_backbone.layer3.16.bn2.weight, img_backbone.layer3.16.bn2.bias, img_backbone.layer3.16.bn2.running_mean, img_backbone.layer3.16.bn2.running_var, img_backbone.layer3.16.bn2.num_batches_tracked, img_backbone.layer3.16.conv3.weight, img_backbone.layer3.16.bn3.weight, img_backbone.layer3.16.bn3.bias, img_backbone.layer3.16.bn3.running_mean, img_backbone.layer3.16.bn3.running_var, img_backbone.layer3.16.bn3.num_batches_tracked, img_backbone.layer3.17.conv1.weight, img_backbone.layer3.17.bn1.weight, img_backbone.layer3.17.bn1.bias, img_backbone.layer3.17.bn1.running_mean, img_backbone.layer3.17.bn1.running_var, img_backbone.layer3.17.bn1.num_batches_tracked, img_backbone.layer3.17.conv2.weight, img_backbone.layer3.17.conv2.conv_offset.weight, img_backbone.layer3.17.conv2.conv_offset.bias, img_backbone.layer3.17.bn2.weight, img_backbone.layer3.17.bn2.bias, img_backbone.layer3.17.bn2.running_mean, img_backbone.layer3.17.bn2.running_var, img_backbone.layer3.17.bn2.num_batches_tracked, img_backbone.layer3.17.conv3.weight, img_backbone.layer3.17.bn3.weight, img_backbone.layer3.17.bn3.bias, img_backbone.layer3.17.bn3.running_mean, img_backbone.layer3.17.bn3.running_var, img_backbone.layer3.17.bn3.num_batches_tracked, img_backbone.layer3.18.conv1.weight, img_backbone.layer3.18.bn1.weight, img_backbone.layer3.18.bn1.bias, img_backbone.layer3.18.bn1.running_mean, img_backbone.layer3.18.bn1.running_var, img_backbone.layer3.18.bn1.num_batches_tracked, img_backbone.layer3.18.conv2.weight, img_backbone.layer3.18.conv2.conv_offset.weight, img_backbone.layer3.18.conv2.conv_offset.bias, img_backbone.layer3.18.bn2.weight, img_backbone.layer3.18.bn2.bias, img_backbone.layer3.18.bn2.running_mean, img_backbone.layer3.18.bn2.running_var, img_backbone.layer3.18.bn2.num_batches_tracked, img_backbone.layer3.18.conv3.weight, img_backbone.layer3.18.bn3.weight, img_backbone.layer3.18.bn3.bias, img_backbone.layer3.18.bn3.running_mean, img_backbone.layer3.18.bn3.running_var, img_backbone.layer3.18.bn3.num_batches_tracked, img_backbone.layer3.19.conv1.weight, img_backbone.layer3.19.bn1.weight, img_backbone.layer3.19.bn1.bias, img_backbone.layer3.19.bn1.running_mean, img_backbone.layer3.19.bn1.running_var, img_backbone.layer3.19.bn1.num_batches_tracked, img_backbone.layer3.19.conv2.weight, img_backbone.layer3.19.conv2.conv_offset.weight, img_backbone.layer3.19.conv2.conv_offset.bias, img_backbone.layer3.19.bn2.weight, img_backbone.layer3.19.bn2.bias, img_backbone.layer3.19.bn2.running_mean, img_backbone.layer3.19.bn2.running_var, img_backbone.layer3.19.bn2.num_batches_tracked, img_backbone.layer3.19.conv3.weight, img_backbone.layer3.19.bn3.weight, img_backbone.layer3.19.bn3.bias, img_backbone.layer3.19.bn3.running_mean, img_backbone.layer3.19.bn3.running_var, img_backbone.layer3.19.bn3.num_batches_tracked, img_backbone.layer3.20.conv1.weight, img_backbone.layer3.20.bn1.weight, img_backbone.layer3.20.bn1.bias, img_backbone.layer3.20.bn1.running_mean, img_backbone.layer3.20.bn1.running_var, img_backbone.layer3.20.bn1.num_batches_tracked, img_backbone.layer3.20.conv2.weight, img_backbone.layer3.20.conv2.conv_offset.weight, img_backbone.layer3.20.conv2.conv_offset.bias, img_backbone.layer3.20.bn2.weight, img_backbone.layer3.20.bn2.bias, img_backbone.layer3.20.bn2.running_mean, img_backbone.layer3.20.bn2.running_var, img_backbone.layer3.20.bn2.num_batches_tracked, img_backbone.layer3.20.conv3.weight, img_backbone.layer3.20.bn3.weight, img_backbone.layer3.20.bn3.bias, img_backbone.layer3.20.bn3.running_mean, img_backbone.layer3.20.bn3.running_var, img_backbone.layer3.20.bn3.num_batches_tracked, img_backbone.layer3.21.conv1.weight, img_backbone.layer3.21.bn1.weight, img_backbone.layer3.21.bn1.bias, img_backbone.layer3.21.bn1.running_mean, img_backbone.layer3.21.bn1.running_var, img_backbone.layer3.21.bn1.num_batches_tracked, img_backbone.layer3.21.conv2.weight, img_backbone.layer3.21.conv2.conv_offset.weight, img_backbone.layer3.21.conv2.conv_offset.bias, img_backbone.layer3.21.bn2.weight, img_backbone.layer3.21.bn2.bias, img_backbone.layer3.21.bn2.running_mean, img_backbone.layer3.21.bn2.running_var, img_backbone.layer3.21.bn2.num_batches_tracked, img_backbone.layer3.21.conv3.weight, img_backbone.layer3.21.bn3.weight, img_backbone.layer3.21.bn3.bias, img_backbone.layer3.21.bn3.running_mean, img_backbone.layer3.21.bn3.running_var, img_backbone.layer3.21.bn3.num_batches_tracked, img_backbone.layer3.22.conv1.weight, img_backbone.layer3.22.bn1.weight, img_backbone.layer3.22.bn1.bias, img_backbone.layer3.22.bn1.running_mean, img_backbone.layer3.22.bn1.running_var, img_backbone.layer3.22.bn1.num_batches_tracked, img_backbone.layer3.22.conv2.weight, img_backbone.layer3.22.conv2.conv_offset.weight, img_backbone.layer3.22.conv2.conv_offset.bias, img_backbone.layer3.22.bn2.weight, img_backbone.layer3.22.bn2.bias, img_backbone.layer3.22.bn2.running_mean, img_backbone.layer3.22.bn2.running_var, img_backbone.layer3.22.bn2.num_batches_tracked, img_backbone.layer3.22.conv3.weight, img_backbone.layer3.22.bn3.weight, img_backbone.layer3.22.bn3.bias, img_backbone.layer3.22.bn3.running_mean, img_backbone.layer3.22.bn3.running_var, img_backbone.layer3.22.bn3.num_batches_tracked, img_neck.fpn_convs.4.conv.weight, img_neck.fpn_convs.4.conv.bias + +missing keys in source state_dict: pts_bbox_head.code_weights, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.0.norms.0.weight, pts_bbox_head.transformer.decoder.layers.0.norms.0.bias, pts_bbox_head.transformer.decoder.layers.0.norms.1.weight, pts_bbox_head.transformer.decoder.layers.0.norms.1.bias, pts_bbox_head.transformer.decoder.layers.0.norms.2.weight, pts_bbox_head.transformer.decoder.layers.0.norms.2.bias, pts_bbox_head.transformer.decoder.layers.0.norms.3.weight, pts_bbox_head.transformer.decoder.layers.0.norms.3.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.1.norms.0.weight, pts_bbox_head.transformer.decoder.layers.1.norms.0.bias, pts_bbox_head.transformer.decoder.layers.1.norms.1.weight, pts_bbox_head.transformer.decoder.layers.1.norms.1.bias, pts_bbox_head.transformer.decoder.layers.1.norms.2.weight, pts_bbox_head.transformer.decoder.layers.1.norms.2.bias, pts_bbox_head.transformer.decoder.layers.1.norms.3.weight, pts_bbox_head.transformer.decoder.layers.1.norms.3.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.2.norms.0.weight, pts_bbox_head.transformer.decoder.layers.2.norms.0.bias, pts_bbox_head.transformer.decoder.layers.2.norms.1.weight, pts_bbox_head.transformer.decoder.layers.2.norms.1.bias, pts_bbox_head.transformer.decoder.layers.2.norms.2.weight, pts_bbox_head.transformer.decoder.layers.2.norms.2.bias, pts_bbox_head.transformer.decoder.layers.2.norms.3.weight, pts_bbox_head.transformer.decoder.layers.2.norms.3.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.3.norms.0.weight, pts_bbox_head.transformer.decoder.layers.3.norms.0.bias, pts_bbox_head.transformer.decoder.layers.3.norms.1.weight, pts_bbox_head.transformer.decoder.layers.3.norms.1.bias, pts_bbox_head.transformer.decoder.layers.3.norms.2.weight, pts_bbox_head.transformer.decoder.layers.3.norms.2.bias, pts_bbox_head.transformer.decoder.layers.3.norms.3.weight, pts_bbox_head.transformer.decoder.layers.3.norms.3.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.4.norms.0.weight, pts_bbox_head.transformer.decoder.layers.4.norms.0.bias, pts_bbox_head.transformer.decoder.layers.4.norms.1.weight, pts_bbox_head.transformer.decoder.layers.4.norms.1.bias, pts_bbox_head.transformer.decoder.layers.4.norms.2.weight, pts_bbox_head.transformer.decoder.layers.4.norms.2.bias, pts_bbox_head.transformer.decoder.layers.4.norms.3.weight, pts_bbox_head.transformer.decoder.layers.4.norms.3.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.5.norms.0.weight, pts_bbox_head.transformer.decoder.layers.5.norms.0.bias, pts_bbox_head.transformer.decoder.layers.5.norms.1.weight, pts_bbox_head.transformer.decoder.layers.5.norms.1.bias, pts_bbox_head.transformer.decoder.layers.5.norms.2.weight, pts_bbox_head.transformer.decoder.layers.5.norms.2.bias, pts_bbox_head.transformer.decoder.layers.5.norms.3.weight, pts_bbox_head.transformer.decoder.layers.5.norms.3.bias, pts_bbox_head.transformer.reference_points.weight, pts_bbox_head.transformer.reference_points.bias, pts_bbox_head.cls_branches.0.0.weight, pts_bbox_head.cls_branches.0.0.bias, pts_bbox_head.cls_branches.0.1.weight, pts_bbox_head.cls_branches.0.1.bias, pts_bbox_head.cls_branches.0.3.weight, pts_bbox_head.cls_branches.0.3.bias, pts_bbox_head.cls_branches.0.4.weight, pts_bbox_head.cls_branches.0.4.bias, pts_bbox_head.cls_branches.0.6.weight, pts_bbox_head.cls_branches.0.6.bias, pts_bbox_head.cls_branches.1.0.weight, pts_bbox_head.cls_branches.1.0.bias, pts_bbox_head.cls_branches.1.1.weight, pts_bbox_head.cls_branches.1.1.bias, pts_bbox_head.cls_branches.1.3.weight, pts_bbox_head.cls_branches.1.3.bias, pts_bbox_head.cls_branches.1.4.weight, pts_bbox_head.cls_branches.1.4.bias, pts_bbox_head.cls_branches.1.6.weight, pts_bbox_head.cls_branches.1.6.bias, pts_bbox_head.cls_branches.2.0.weight, pts_bbox_head.cls_branches.2.0.bias, pts_bbox_head.cls_branches.2.1.weight, pts_bbox_head.cls_branches.2.1.bias, pts_bbox_head.cls_branches.2.3.weight, pts_bbox_head.cls_branches.2.3.bias, pts_bbox_head.cls_branches.2.4.weight, pts_bbox_head.cls_branches.2.4.bias, pts_bbox_head.cls_branches.2.6.weight, pts_bbox_head.cls_branches.2.6.bias, pts_bbox_head.cls_branches.3.0.weight, pts_bbox_head.cls_branches.3.0.bias, pts_bbox_head.cls_branches.3.1.weight, pts_bbox_head.cls_branches.3.1.bias, pts_bbox_head.cls_branches.3.3.weight, pts_bbox_head.cls_branches.3.3.bias, pts_bbox_head.cls_branches.3.4.weight, pts_bbox_head.cls_branches.3.4.bias, pts_bbox_head.cls_branches.3.6.weight, pts_bbox_head.cls_branches.3.6.bias, pts_bbox_head.cls_branches.4.0.weight, pts_bbox_head.cls_branches.4.0.bias, pts_bbox_head.cls_branches.4.1.weight, pts_bbox_head.cls_branches.4.1.bias, pts_bbox_head.cls_branches.4.3.weight, pts_bbox_head.cls_branches.4.3.bias, pts_bbox_head.cls_branches.4.4.weight, pts_bbox_head.cls_branches.4.4.bias, pts_bbox_head.cls_branches.4.6.weight, pts_bbox_head.cls_branches.4.6.bias, pts_bbox_head.cls_branches.5.0.weight, pts_bbox_head.cls_branches.5.0.bias, pts_bbox_head.cls_branches.5.1.weight, pts_bbox_head.cls_branches.5.1.bias, pts_bbox_head.cls_branches.5.3.weight, pts_bbox_head.cls_branches.5.3.bias, pts_bbox_head.cls_branches.5.4.weight, pts_bbox_head.cls_branches.5.4.bias, pts_bbox_head.cls_branches.5.6.weight, pts_bbox_head.cls_branches.5.6.bias, pts_bbox_head.reg_branches.0.0.weight, pts_bbox_head.reg_branches.0.0.bias, pts_bbox_head.reg_branches.0.2.weight, pts_bbox_head.reg_branches.0.2.bias, pts_bbox_head.reg_branches.0.4.weight, pts_bbox_head.reg_branches.0.4.bias, pts_bbox_head.reg_branches.1.0.weight, pts_bbox_head.reg_branches.1.0.bias, pts_bbox_head.reg_branches.1.2.weight, pts_bbox_head.reg_branches.1.2.bias, pts_bbox_head.reg_branches.1.4.weight, pts_bbox_head.reg_branches.1.4.bias, pts_bbox_head.reg_branches.2.0.weight, pts_bbox_head.reg_branches.2.0.bias, pts_bbox_head.reg_branches.2.2.weight, pts_bbox_head.reg_branches.2.2.bias, pts_bbox_head.reg_branches.2.4.weight, pts_bbox_head.reg_branches.2.4.bias, pts_bbox_head.reg_branches.3.0.weight, pts_bbox_head.reg_branches.3.0.bias, pts_bbox_head.reg_branches.3.2.weight, pts_bbox_head.reg_branches.3.2.bias, pts_bbox_head.reg_branches.3.4.weight, pts_bbox_head.reg_branches.3.4.bias, pts_bbox_head.reg_branches.4.0.weight, pts_bbox_head.reg_branches.4.0.bias, pts_bbox_head.reg_branches.4.2.weight, pts_bbox_head.reg_branches.4.2.bias, pts_bbox_head.reg_branches.4.4.weight, pts_bbox_head.reg_branches.4.4.bias, pts_bbox_head.reg_branches.5.0.weight, pts_bbox_head.reg_branches.5.0.bias, pts_bbox_head.reg_branches.5.2.weight, pts_bbox_head.reg_branches.5.2.bias, pts_bbox_head.reg_branches.5.4.weight, pts_bbox_head.reg_branches.5.4.bias, pts_bbox_head.query_embedding.weight, language_model.language_backbone.body.model.embeddings.word_embeddings.weight, language_model.language_backbone.body.model.embeddings.position_embeddings.weight, language_model.language_backbone.body.model.embeddings.token_type_embeddings.weight, language_model.language_backbone.body.model.embeddings.LayerNorm.weight, language_model.language_backbone.body.model.embeddings.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.0.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.0.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.0.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.0.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.0.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.0.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.1.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.1.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.1.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.1.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.1.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.1.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.2.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.2.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.2.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.2.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.2.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.2.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.3.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.3.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.3.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.3.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.3.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.3.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.4.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.4.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.4.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.4.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.4.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.4.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.5.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.5.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.5.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.5.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.5.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.5.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.6.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.6.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.6.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.6.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.6.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.6.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.7.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.7.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.7.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.7.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.7.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.7.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.8.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.8.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.8.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.8.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.8.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.8.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.9.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.9.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.9.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.9.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.9.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.9.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.10.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.10.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.10.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.10.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.10.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.10.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.11.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.11.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.11.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.11.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.11.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.11.output.LayerNorm.bias, text_feat_map.weight, text_feat_map.bias + +10/10 17:35:36 - mmengine - INFO - Load checkpoint from pretrained/fcos3d.pth +10/10 17:35:36 - mmengine - WARNING - "FileClient" will be deprecated in future. Please use io functions in https://mmengine.readthedocs.io/en/latest/api/fileio.html#file-io +10/10 17:35:36 - mmengine - WARNING - "HardDiskBackend" is the alias of "LocalBackend" and the former will be deprecated in future. +10/10 17:35:36 - mmengine - INFO - Checkpoints will be saved to /home/xzt/mmdetection3d-1.1.0/work_dirs/detr3d_r50_bert_gridmask_halfdata_decoder. +/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/grid_mask.py:132: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at /pytorch/torch/csrc/utils/tensor_numpy.cpp:180.) + mask = torch.from_numpy(mask).to(x) +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at /pytorch/c10/core/TensorImpl.h:1156.) + return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode) +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmcv/cnn/bricks/transformer.py:524: UserWarning: position encoding of key ismissing in MultiheadAttention. + warnings.warn(f'position encoding of key is' +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/functional.py:3981: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. + warnings.warn( +/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/grid_mask.py:132: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at /pytorch/torch/csrc/utils/tensor_numpy.cpp:180.) + mask = torch.from_numpy(mask).to(x) +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at /pytorch/c10/core/TensorImpl.h:1156.) + return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode) +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmcv/cnn/bricks/transformer.py:524: UserWarning: position encoding of key ismissing in MultiheadAttention. + warnings.warn(f'position encoding of key is' +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/functional.py:3981: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. + warnings.warn( +10/10 17:36:37 - mmengine - INFO - Epoch(train) [1][ 50/7033] base_lr: 7.9760e-05 lr: 7.9760e-05 eta: 2 days, 9:00:57 time: 1.2164 data_time: 0.0955 memory: 11853 grad_norm: 91.0755 loss: 19.3339 loss_cls: 1.5007 loss_bbox: 1.7604 d0.loss_cls: 1.4829 d0.loss_bbox: 1.7122 d1.loss_cls: 1.4941 d1.loss_bbox: 1.7443 d2.loss_cls: 1.4642 d2.loss_bbox: 1.7357 d3.loss_cls: 1.4655 d3.loss_bbox: 1.7357 d4.loss_cls: 1.4887 d4.loss_bbox: 1.7495 +10/10 17:37:43 - mmengine - INFO - Epoch(train) [1][ 100/7033] base_lr: 9.3120e-05 lr: 9.3120e-05 eta: 2 days, 11:25:32 time: 1.3200 data_time: 0.0519 memory: 11853 grad_norm: 70.0541 loss: 16.4143 loss_cls: 1.1980 loss_bbox: 1.5548 d0.loss_cls: 1.2040 d0.loss_bbox: 1.5131 d1.loss_cls: 1.2072 d1.loss_bbox: 1.5106 d2.loss_cls: 1.1949 d2.loss_bbox: 1.5207 d3.loss_cls: 1.1969 d3.loss_bbox: 1.5895 d4.loss_cls: 1.1876 d4.loss_bbox: 1.5371 +10/10 17:38:47 - mmengine - INFO - Epoch(train) [1][ 150/7033] base_lr: 1.0648e-04 lr: 1.0648e-04 eta: 2 days, 11:29:33 time: 1.2736 data_time: 0.0560 memory: 11853 grad_norm: 57.8964 loss: 17.8822 loss_cls: 1.2607 loss_bbox: 1.7408 d0.loss_cls: 1.2643 d0.loss_bbox: 1.7261 d1.loss_cls: 1.2612 d1.loss_bbox: 1.7138 d2.loss_cls: 1.2562 d2.loss_bbox: 1.7114 d3.loss_cls: 1.2587 d3.loss_bbox: 1.7295 d4.loss_cls: 1.2543 d4.loss_bbox: 1.7053 +10/10 17:39:51 - mmengine - INFO - Epoch(train) [1][ 200/7033] base_lr: 1.1984e-04 lr: 1.1984e-04 eta: 2 days, 11:40:16 time: 1.2868 data_time: 0.0483 memory: 11853 grad_norm: 58.0646 loss: 15.9252 loss_cls: 1.1739 loss_bbox: 1.4770 d0.loss_cls: 1.1801 d0.loss_bbox: 1.5285 d1.loss_cls: 1.1665 d1.loss_bbox: 1.5157 d2.loss_cls: 1.1690 d2.loss_bbox: 1.4825 d3.loss_cls: 1.1556 d3.loss_bbox: 1.4689 d4.loss_cls: 1.1542 d4.loss_bbox: 1.4533 +10/10 17:40:55 - mmengine - INFO - Epoch(train) [1][ 250/7033] base_lr: 1.3320e-04 lr: 1.3320e-04 eta: 2 days, 11:41:48 time: 1.2788 data_time: 0.0450 memory: 11853 grad_norm: 50.6429 loss: 16.6777 loss_cls: 1.2240 loss_bbox: 1.6012 d0.loss_cls: 1.2038 d0.loss_bbox: 1.5982 d1.loss_cls: 1.1943 d1.loss_bbox: 1.5653 d2.loss_cls: 1.1956 d2.loss_bbox: 1.5843 d3.loss_cls: 1.1968 d3.loss_bbox: 1.5651 d4.loss_cls: 1.1838 d4.loss_bbox: 1.5652 +10/10 17:41:59 - mmengine - INFO - Epoch(train) [1][ 300/7033] base_lr: 1.4656e-04 lr: 1.4656e-04 eta: 2 days, 11:40:45 time: 1.2751 data_time: 0.0447 memory: 11853 grad_norm: 57.8544 loss: 18.5497 loss_cls: 1.3733 loss_bbox: 1.7851 d0.loss_cls: 1.3356 d0.loss_bbox: 1.7701 d1.loss_cls: 1.3314 d1.loss_bbox: 1.7734 d2.loss_cls: 1.3550 d2.loss_bbox: 1.7168 d3.loss_cls: 1.3384 d3.loss_bbox: 1.7154 d4.loss_cls: 1.3305 d4.loss_bbox: 1.7249 +10/10 17:43:13 - mmengine - INFO - Epoch(train) [1][ 350/7033] base_lr: 1.5992e-04 lr: 1.5992e-04 eta: 2 days, 13:01:24 time: 1.4788 data_time: 0.0480 memory: 11853 grad_norm: 63.5684 loss: 13.9266 loss_cls: 1.0607 loss_bbox: 1.2319 d0.loss_cls: 1.0515 d0.loss_bbox: 1.3622 d1.loss_cls: 1.0379 d1.loss_bbox: 1.3138 d2.loss_cls: 1.0504 d2.loss_bbox: 1.2657 d3.loss_cls: 1.0437 d3.loss_bbox: 1.2367 d4.loss_cls: 1.0426 d4.loss_bbox: 1.2296 +10/10 17:44:29 - mmengine - INFO - Epoch(train) [1][ 400/7033] base_lr: 1.7328e-04 lr: 1.7328e-04 eta: 2 days, 14:18:52 time: 1.5282 data_time: 0.0487 memory: 11853 grad_norm: 61.5307 loss: 15.4620 loss_cls: 1.1607 loss_bbox: 1.4667 d0.loss_cls: 1.1286 d0.loss_bbox: 1.4983 d1.loss_cls: 1.1181 d1.loss_bbox: 1.4434 d2.loss_cls: 1.1366 d2.loss_bbox: 1.4254 d3.loss_cls: 1.1285 d3.loss_bbox: 1.3845 d4.loss_cls: 1.1389 d4.loss_bbox: 1.4321 +10/10 17:45:45 - mmengine - INFO - Epoch(train) [1][ 450/7033] base_lr: 1.8664e-04 lr: 1.8664e-04 eta: 2 days, 15:18:21 time: 1.5266 data_time: 0.0501 memory: 11853 grad_norm: 59.6309 loss: 17.0077 loss_cls: 1.2936 loss_bbox: 1.5728 d0.loss_cls: 1.2429 d0.loss_bbox: 1.6170 d1.loss_cls: 1.2344 d1.loss_bbox: 1.5552 d2.loss_cls: 1.2877 d2.loss_bbox: 1.5731 d3.loss_cls: 1.2676 d3.loss_bbox: 1.5520 d4.loss_cls: 1.2703 d4.loss_bbox: 1.5410 +10/10 17:46:59 - mmengine - INFO - Epoch(train) [1][ 500/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 15:51:04 time: 1.4745 data_time: 0.0524 memory: 11853 grad_norm: 74.5934 loss: 15.8864 loss_cls: 1.1601 loss_bbox: 1.5060 d0.loss_cls: 1.1234 d0.loss_bbox: 1.5404 d1.loss_cls: 1.1353 d1.loss_bbox: 1.4612 d2.loss_cls: 1.1385 d2.loss_bbox: 1.4648 d3.loss_cls: 1.1910 d3.loss_bbox: 1.5299 d4.loss_cls: 1.1481 d4.loss_bbox: 1.4876 +10/10 17:48:12 - mmengine - INFO - Epoch(train) [1][ 550/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 16:15:14 time: 1.4652 data_time: 0.0427 memory: 11853 grad_norm: 59.5421 loss: 14.8269 loss_cls: 1.1350 loss_bbox: 1.3258 d0.loss_cls: 1.0838 d0.loss_bbox: 1.3825 d1.loss_cls: 1.1034 d1.loss_bbox: 1.3758 d2.loss_cls: 1.1242 d2.loss_bbox: 1.3407 d3.loss_cls: 1.1445 d3.loss_bbox: 1.3514 d4.loss_cls: 1.1295 d4.loss_bbox: 1.3303 +10/10 17:49:25 - mmengine - INFO - Epoch(train) [1][ 600/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 16:34:19 time: 1.4615 data_time: 0.0429 memory: 11853 grad_norm: 51.8184 loss: 15.2389 loss_cls: 1.1428 loss_bbox: 1.4067 d0.loss_cls: 1.1007 d0.loss_bbox: 1.4504 d1.loss_cls: 1.1082 d1.loss_bbox: 1.4061 d2.loss_cls: 1.1323 d2.loss_bbox: 1.4096 d3.loss_cls: 1.1376 d3.loss_bbox: 1.3816 d4.loss_cls: 1.1618 d4.loss_bbox: 1.4012 +10/10 17:50:40 - mmengine - INFO - Epoch(train) [1][ 650/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 16:59:06 time: 1.5024 data_time: 0.0423 memory: 11853 grad_norm: 59.4280 loss: 13.8934 loss_cls: 1.0520 loss_bbox: 1.2981 d0.loss_cls: 1.0098 d0.loss_bbox: 1.3207 d1.loss_cls: 1.0388 d1.loss_bbox: 1.2931 d2.loss_cls: 1.0223 d2.loss_bbox: 1.2541 d3.loss_cls: 1.0379 d3.loss_bbox: 1.2403 d4.loss_cls: 1.0533 d4.loss_bbox: 1.2731 +10/10 17:51:54 - mmengine - INFO - Epoch(train) [1][ 700/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 17:12:43 time: 1.4652 data_time: 0.0418 memory: 11853 grad_norm: 61.5926 loss: 17.3444 loss_cls: 1.2830 loss_bbox: 1.5972 d0.loss_cls: 1.2440 d0.loss_bbox: 1.7030 d1.loss_cls: 1.2525 d1.loss_bbox: 1.6474 d2.loss_cls: 1.2601 d2.loss_bbox: 1.6088 d3.loss_cls: 1.2731 d3.loss_bbox: 1.5779 d4.loss_cls: 1.2950 d4.loss_bbox: 1.6025 +10/10 17:52:58 - mmengine - INFO - Epoch(train) [1][ 750/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 16:49:53 time: 1.2806 data_time: 0.0407 memory: 11853 grad_norm: 53.6512 loss: 15.2474 loss_cls: 1.1735 loss_bbox: 1.3402 d0.loss_cls: 1.1325 d0.loss_bbox: 1.5093 d1.loss_cls: 1.1317 d1.loss_bbox: 1.4075 d2.loss_cls: 1.1545 d2.loss_bbox: 1.3679 d3.loss_cls: 1.1618 d3.loss_bbox: 1.3575 d4.loss_cls: 1.1710 d4.loss_bbox: 1.3400 +10/10 17:54:03 - mmengine - INFO - Epoch(train) [1][ 800/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 16:34:01 time: 1.3048 data_time: 0.0410 memory: 11853 grad_norm: 45.4818 loss: 13.9172 loss_cls: 1.0563 loss_bbox: 1.2857 d0.loss_cls: 1.0119 d0.loss_bbox: 1.3427 d1.loss_cls: 1.0054 d1.loss_bbox: 1.2882 d2.loss_cls: 1.0224 d2.loss_bbox: 1.2706 d3.loss_cls: 1.0498 d3.loss_bbox: 1.2671 d4.loss_cls: 1.0393 d4.loss_bbox: 1.2778 +10/10 17:55:07 - mmengine - INFO - Epoch(train) [1][ 850/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 16:14:54 time: 1.2746 data_time: 0.0382 memory: 11853 grad_norm: 46.4011 loss: 13.0258 loss_cls: 0.9786 loss_bbox: 1.1873 d0.loss_cls: 0.9337 d0.loss_bbox: 1.2722 d1.loss_cls: 0.9551 d1.loss_bbox: 1.2254 d2.loss_cls: 0.9566 d2.loss_bbox: 1.1968 d3.loss_cls: 0.9749 d3.loss_bbox: 1.1952 d4.loss_cls: 0.9682 d4.loss_bbox: 1.1819 +10/10 17:56:11 - mmengine - INFO - Epoch(train) [1][ 900/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 15:58:58 time: 1.2822 data_time: 0.0388 memory: 11853 grad_norm: 45.1400 loss: 13.4320 loss_cls: 0.9791 loss_bbox: 1.2559 d0.loss_cls: 0.9474 d0.loss_bbox: 1.3445 d1.loss_cls: 0.9586 d1.loss_bbox: 1.2690 d2.loss_cls: 0.9767 d2.loss_bbox: 1.2446 d3.loss_cls: 0.9773 d3.loss_bbox: 1.2516 d4.loss_cls: 0.9796 d4.loss_bbox: 1.2477 +10/10 17:57:15 - mmengine - INFO - Epoch(train) [1][ 950/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 15:43:55 time: 1.2775 data_time: 0.0383 memory: 11853 grad_norm: 49.3366 loss: 14.8552 loss_cls: 1.1309 loss_bbox: 1.3665 d0.loss_cls: 1.0376 d0.loss_bbox: 1.4869 d1.loss_cls: 1.0668 d1.loss_bbox: 1.4048 d2.loss_cls: 1.1018 d2.loss_bbox: 1.3568 d3.loss_cls: 1.1056 d3.loss_bbox: 1.3356 d4.loss_cls: 1.1136 d4.loss_bbox: 1.3483 +10/10 17:58:19 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 17:58:19 - mmengine - INFO - Epoch(train) [1][1000/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 15:30:23 time: 1.2784 data_time: 0.0388 memory: 11853 grad_norm: 48.8183 loss: 14.2933 loss_cls: 1.0805 loss_bbox: 1.3004 d0.loss_cls: 1.0203 d0.loss_bbox: 1.4202 d1.loss_cls: 1.0309 d1.loss_bbox: 1.3592 d2.loss_cls: 1.0461 d2.loss_bbox: 1.3216 d3.loss_cls: 1.0573 d3.loss_bbox: 1.3012 d4.loss_cls: 1.0540 d4.loss_bbox: 1.3016 +10/10 17:59:22 - mmengine - INFO - Epoch(train) [1][1050/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 15:17:34 time: 1.2747 data_time: 0.0382 memory: 11853 grad_norm: 51.1685 loss: 13.2414 loss_cls: 0.9933 loss_bbox: 1.2421 d0.loss_cls: 0.9240 d0.loss_bbox: 1.3098 d1.loss_cls: 0.9474 d1.loss_bbox: 1.2415 d2.loss_cls: 0.9719 d2.loss_bbox: 1.2039 d3.loss_cls: 0.9810 d3.loss_bbox: 1.2172 d4.loss_cls: 0.9824 d4.loss_bbox: 1.2270 +10/10 17:59:59 - mmengine - INFO - Epoch(train) [1][1100/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 13:56:15 time: 0.7272 data_time: 0.0397 memory: 11853 grad_norm: 48.2237 loss: 13.4930 loss_cls: 1.0201 loss_bbox: 1.2293 d0.loss_cls: 0.9429 d0.loss_bbox: 1.3451 d1.loss_cls: 0.9779 d1.loss_bbox: 1.2645 d2.loss_cls: 1.0019 d2.loss_bbox: 1.2326 d3.loss_cls: 1.0183 d3.loss_bbox: 1.2319 d4.loss_cls: 1.0126 d4.loss_bbox: 1.2159 +10/10 18:00:33 - mmengine - INFO - Epoch(train) [1][1150/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 12:36:17 time: 0.6805 data_time: 0.0398 memory: 11853 grad_norm: 47.5950 loss: 13.8037 loss_cls: 1.0102 loss_bbox: 1.2697 d0.loss_cls: 0.9415 d0.loss_bbox: 1.3889 d1.loss_cls: 0.9649 d1.loss_bbox: 1.3421 d2.loss_cls: 0.9926 d2.loss_bbox: 1.3052 d3.loss_cls: 1.0032 d3.loss_bbox: 1.2953 d4.loss_cls: 1.0060 d4.loss_bbox: 1.2840 +10/10 18:01:07 - mmengine - INFO - Epoch(train) [1][1200/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 11:22:41 time: 0.6784 data_time: 0.0400 memory: 11853 grad_norm: 44.1694 loss: 12.9094 loss_cls: 0.9508 loss_bbox: 1.2198 d0.loss_cls: 0.8830 d0.loss_bbox: 1.2684 d1.loss_cls: 0.9217 d1.loss_bbox: 1.2087 d2.loss_cls: 0.9384 d2.loss_bbox: 1.1965 d3.loss_cls: 0.9497 d3.loss_bbox: 1.2121 d4.loss_cls: 0.9507 d4.loss_bbox: 1.2095 +10/10 18:01:41 - mmengine - INFO - Epoch(train) [1][1250/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 10:14:53 time: 0.6780 data_time: 0.0399 memory: 11853 grad_norm: 42.0036 loss: 12.7880 loss_cls: 0.9763 loss_bbox: 1.1801 d0.loss_cls: 0.8761 d0.loss_bbox: 1.2739 d1.loss_cls: 0.9155 d1.loss_bbox: 1.2243 d2.loss_cls: 0.9438 d2.loss_bbox: 1.1694 d3.loss_cls: 0.9500 d3.loss_bbox: 1.1618 d4.loss_cls: 0.9576 d4.loss_bbox: 1.1591 +10/10 18:02:15 - mmengine - INFO - Epoch(train) [1][1300/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 9:13:50 time: 0.6927 data_time: 0.0389 memory: 11853 grad_norm: 49.9060 loss: 12.4056 loss_cls: 0.9287 loss_bbox: 1.1560 d0.loss_cls: 0.8441 d0.loss_bbox: 1.2243 d1.loss_cls: 0.8838 d1.loss_bbox: 1.1682 d2.loss_cls: 0.9031 d2.loss_bbox: 1.1558 d3.loss_cls: 0.9150 d3.loss_bbox: 1.1335 d4.loss_cls: 0.9309 d4.loss_bbox: 1.1624 +10/10 18:02:51 - mmengine - INFO - Epoch(train) [1][1350/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 8:19:37 time: 0.7154 data_time: 0.0377 memory: 11853 grad_norm: 44.8158 loss: 11.1597 loss_cls: 0.8152 loss_bbox: 1.0295 d0.loss_cls: 0.7730 d0.loss_bbox: 1.1542 d1.loss_cls: 0.7914 d1.loss_bbox: 1.0754 d2.loss_cls: 0.8015 d2.loss_bbox: 1.0437 d3.loss_cls: 0.7982 d3.loss_bbox: 1.0348 d4.loss_cls: 0.8186 d4.loss_bbox: 1.0241 +10/10 18:03:28 - mmengine - INFO - Epoch(train) [1][1400/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 7:32:02 time: 0.7437 data_time: 0.0369 memory: 11853 grad_norm: 42.0973 loss: 13.2777 loss_cls: 0.9609 loss_bbox: 1.2573 d0.loss_cls: 0.9020 d0.loss_bbox: 1.3516 d1.loss_cls: 0.9281 d1.loss_bbox: 1.2659 d2.loss_cls: 0.9287 d2.loss_bbox: 1.2628 d3.loss_cls: 0.9543 d3.loss_bbox: 1.2489 d4.loss_cls: 0.9637 d4.loss_bbox: 1.2536 +10/10 18:04:05 - mmengine - INFO - Epoch(train) [1][1450/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 6:46:30 time: 0.7311 data_time: 0.0366 memory: 11853 grad_norm: 40.1634 loss: 13.6969 loss_cls: 1.0071 loss_bbox: 1.2675 d0.loss_cls: 0.9185 d0.loss_bbox: 1.4156 d1.loss_cls: 0.9591 d1.loss_bbox: 1.2995 d2.loss_cls: 0.9786 d2.loss_bbox: 1.2884 d3.loss_cls: 0.9894 d3.loss_bbox: 1.2820 d4.loss_cls: 1.0009 d4.loss_bbox: 1.2902 +10/10 18:04:42 - mmengine - INFO - Epoch(train) [1][1500/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 6:04:41 time: 0.7390 data_time: 0.0363 memory: 11853 grad_norm: 44.9839 loss: 13.7259 loss_cls: 1.0253 loss_bbox: 1.2844 d0.loss_cls: 0.9582 d0.loss_bbox: 1.3928 d1.loss_cls: 0.9780 d1.loss_bbox: 1.2823 d2.loss_cls: 0.9850 d2.loss_bbox: 1.2514 d3.loss_cls: 1.0040 d3.loss_bbox: 1.2617 d4.loss_cls: 1.0202 d4.loss_bbox: 1.2827 +10/10 18:05:18 - mmengine - INFO - Epoch(train) [1][1550/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 5:24:54 time: 0.7320 data_time: 0.0370 memory: 11853 grad_norm: 44.6417 loss: 12.1393 loss_cls: 0.9012 loss_bbox: 1.1224 d0.loss_cls: 0.8364 d0.loss_bbox: 1.2311 d1.loss_cls: 0.8487 d1.loss_bbox: 1.1537 d2.loss_cls: 0.8804 d2.loss_bbox: 1.1174 d3.loss_cls: 0.9019 d3.loss_bbox: 1.1203 d4.loss_cls: 0.8994 d4.loss_bbox: 1.1264 +10/10 18:05:55 - mmengine - INFO - Epoch(train) [1][1600/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 4:48:20 time: 0.7410 data_time: 0.0368 memory: 11853 grad_norm: 47.3428 loss: 13.9865 loss_cls: 1.0172 loss_bbox: 1.3211 d0.loss_cls: 0.9837 d0.loss_bbox: 1.4286 d1.loss_cls: 0.9705 d1.loss_bbox: 1.3342 d2.loss_cls: 0.9860 d2.loss_bbox: 1.3224 d3.loss_cls: 0.9999 d3.loss_bbox: 1.2998 d4.loss_cls: 1.0121 d4.loss_bbox: 1.3109 +10/10 18:06:31 - mmengine - INFO - Epoch(train) [1][1650/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 4:11:34 time: 0.7126 data_time: 0.0382 memory: 11853 grad_norm: 41.7511 loss: 11.0717 loss_cls: 0.8130 loss_bbox: 1.0495 d0.loss_cls: 0.7596 d0.loss_bbox: 1.1287 d1.loss_cls: 0.7845 d1.loss_bbox: 1.0352 d2.loss_cls: 0.7970 d2.loss_bbox: 1.0176 d3.loss_cls: 0.8114 d3.loss_bbox: 1.0243 d4.loss_cls: 0.8110 d4.loss_bbox: 1.0399 +10/10 18:07:05 - mmengine - INFO - Epoch(train) [1][1700/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 3:33:45 time: 0.6739 data_time: 0.0390 memory: 11853 grad_norm: 43.4436 loss: 11.8852 loss_cls: 0.8652 loss_bbox: 1.1104 d0.loss_cls: 0.7932 d0.loss_bbox: 1.2336 d1.loss_cls: 0.8175 d1.loss_bbox: 1.1582 d2.loss_cls: 0.8290 d2.loss_bbox: 1.1185 d3.loss_cls: 0.8563 d3.loss_bbox: 1.1190 d4.loss_cls: 0.8655 d4.loss_bbox: 1.1189 +10/10 18:07:39 - mmengine - INFO - Epoch(train) [1][1750/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 2:58:30 time: 0.6794 data_time: 0.0396 memory: 11853 grad_norm: 43.0396 loss: 11.6313 loss_cls: 0.8681 loss_bbox: 1.0732 d0.loss_cls: 0.8118 d0.loss_bbox: 1.1944 d1.loss_cls: 0.8159 d1.loss_bbox: 1.1039 d2.loss_cls: 0.8285 d2.loss_bbox: 1.0806 d3.loss_cls: 0.8457 d3.loss_bbox: 1.0803 d4.loss_cls: 0.8631 d4.loss_bbox: 1.0656 +10/10 18:08:13 - mmengine - INFO - Epoch(train) [1][1800/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 2:25:38 time: 0.6853 data_time: 0.0392 memory: 11853 grad_norm: 38.4248 loss: 13.5162 loss_cls: 0.9621 loss_bbox: 1.2907 d0.loss_cls: 0.8973 d0.loss_bbox: 1.4105 d1.loss_cls: 0.9227 d1.loss_bbox: 1.3200 d2.loss_cls: 0.9320 d2.loss_bbox: 1.2940 d3.loss_cls: 0.9457 d3.loss_bbox: 1.2911 d4.loss_cls: 0.9601 d4.loss_bbox: 1.2901 +10/10 18:08:47 - mmengine - INFO - Epoch(train) [1][1850/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 1:53:41 time: 0.6743 data_time: 0.0393 memory: 11853 grad_norm: 38.8879 loss: 11.8295 loss_cls: 0.8646 loss_bbox: 1.1171 d0.loss_cls: 0.7828 d0.loss_bbox: 1.2293 d1.loss_cls: 0.8158 d1.loss_bbox: 1.1402 d2.loss_cls: 0.8434 d2.loss_bbox: 1.1126 d3.loss_cls: 0.8478 d3.loss_bbox: 1.1103 d4.loss_cls: 0.8612 d4.loss_bbox: 1.1043 +10/10 18:09:20 - mmengine - INFO - Epoch(train) [1][1900/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 1:23:24 time: 0.6745 data_time: 0.0389 memory: 11853 grad_norm: 39.3316 loss: 11.1450 loss_cls: 0.7836 loss_bbox: 1.0874 d0.loss_cls: 0.7347 d0.loss_bbox: 1.1698 d1.loss_cls: 0.7463 d1.loss_bbox: 1.0811 d2.loss_cls: 0.7587 d2.loss_bbox: 1.0704 d3.loss_cls: 0.7706 d3.loss_bbox: 1.0698 d4.loss_cls: 0.7833 d4.loss_bbox: 1.0893 +10/10 18:09:54 - mmengine - INFO - Epoch(train) [1][1950/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 0:54:37 time: 0.6743 data_time: 0.0390 memory: 11853 grad_norm: 42.3490 loss: 11.9950 loss_cls: 0.8604 loss_bbox: 1.1425 d0.loss_cls: 0.8034 d0.loss_bbox: 1.2472 d1.loss_cls: 0.8083 d1.loss_bbox: 1.1746 d2.loss_cls: 0.8389 d2.loss_bbox: 1.1470 d3.loss_cls: 0.8560 d3.loss_bbox: 1.1323 d4.loss_cls: 0.8533 d4.loss_bbox: 1.1311 +10/10 18:10:28 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 18:10:28 - mmengine - INFO - Epoch(train) [1][2000/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 0:27:17 time: 0.6746 data_time: 0.0389 memory: 11853 grad_norm: 39.9590 loss: 11.4582 loss_cls: 0.8157 loss_bbox: 1.1024 d0.loss_cls: 0.7527 d0.loss_bbox: 1.1912 d1.loss_cls: 0.7718 d1.loss_bbox: 1.1266 d2.loss_cls: 0.7954 d2.loss_bbox: 1.1043 d3.loss_cls: 0.7995 d3.loss_bbox: 1.0874 d4.loss_cls: 0.8072 d4.loss_bbox: 1.1040 +10/10 18:11:02 - mmengine - INFO - Epoch(train) [1][2050/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 2 days, 0:01:14 time: 0.6745 data_time: 0.0389 memory: 11853 grad_norm: 38.8342 loss: 13.2169 loss_cls: 0.9421 loss_bbox: 1.2492 d0.loss_cls: 0.8979 d0.loss_bbox: 1.3820 d1.loss_cls: 0.9025 d1.loss_bbox: 1.2854 d2.loss_cls: 0.9137 d2.loss_bbox: 1.2660 d3.loss_cls: 0.9366 d3.loss_bbox: 1.2535 d4.loss_cls: 0.9375 d4.loss_bbox: 1.2505 +10/10 18:11:35 - mmengine - INFO - Epoch(train) [1][2100/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 23:36:36 time: 0.6774 data_time: 0.0386 memory: 11853 grad_norm: 42.4662 loss: 12.9065 loss_cls: 0.9144 loss_bbox: 1.2261 d0.loss_cls: 0.8497 d0.loss_bbox: 1.3611 d1.loss_cls: 0.8671 d1.loss_bbox: 1.2634 d2.loss_cls: 0.8995 d2.loss_bbox: 1.2379 d3.loss_cls: 0.9135 d3.loss_bbox: 1.2308 d4.loss_cls: 0.9157 d4.loss_bbox: 1.2273 +10/10 18:12:09 - mmengine - INFO - Epoch(train) [1][2150/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 23:13:13 time: 0.6793 data_time: 0.0392 memory: 11853 grad_norm: 41.1454 loss: 12.5726 loss_cls: 0.8768 loss_bbox: 1.1938 d0.loss_cls: 0.8103 d0.loss_bbox: 1.3680 d1.loss_cls: 0.8385 d1.loss_bbox: 1.2476 d2.loss_cls: 0.8499 d2.loss_bbox: 1.2258 d3.loss_cls: 0.8724 d3.loss_bbox: 1.2074 d4.loss_cls: 0.8790 d4.loss_bbox: 1.2033 +10/10 18:12:43 - mmengine - INFO - Epoch(train) [1][2200/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 22:50:39 time: 0.6762 data_time: 0.0390 memory: 11853 grad_norm: 42.3337 loss: 11.1345 loss_cls: 0.7991 loss_bbox: 1.0561 d0.loss_cls: 0.7395 d0.loss_bbox: 1.1672 d1.loss_cls: 0.7622 d1.loss_bbox: 1.0771 d2.loss_cls: 0.7742 d2.loss_bbox: 1.0618 d3.loss_cls: 0.7841 d3.loss_bbox: 1.0615 d4.loss_cls: 0.8021 d4.loss_bbox: 1.0497 +10/10 18:13:17 - mmengine - INFO - Epoch(train) [1][2250/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 22:29:24 time: 0.6813 data_time: 0.0394 memory: 11853 grad_norm: 39.6678 loss: 11.4051 loss_cls: 0.8300 loss_bbox: 1.0613 d0.loss_cls: 0.7676 d0.loss_bbox: 1.1801 d1.loss_cls: 0.7969 d1.loss_bbox: 1.0917 d2.loss_cls: 0.8192 d2.loss_bbox: 1.0681 d3.loss_cls: 0.8264 d3.loss_bbox: 1.0720 d4.loss_cls: 0.8325 d4.loss_bbox: 1.0594 +10/10 18:13:51 - mmengine - INFO - Epoch(train) [1][2300/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 22:08:41 time: 0.6754 data_time: 0.0389 memory: 11853 grad_norm: 46.0182 loss: 12.2008 loss_cls: 0.8817 loss_bbox: 1.1476 d0.loss_cls: 0.8307 d0.loss_bbox: 1.2661 d1.loss_cls: 0.8541 d1.loss_bbox: 1.1755 d2.loss_cls: 0.8616 d2.loss_bbox: 1.1487 d3.loss_cls: 0.8756 d3.loss_bbox: 1.1301 d4.loss_cls: 0.8906 d4.loss_bbox: 1.1384 +10/10 18:14:25 - mmengine - INFO - Epoch(train) [1][2350/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 21:48:48 time: 0.6749 data_time: 0.0391 memory: 11853 grad_norm: 40.0063 loss: 11.3188 loss_cls: 0.8060 loss_bbox: 1.0878 d0.loss_cls: 0.7419 d0.loss_bbox: 1.1876 d1.loss_cls: 0.7720 d1.loss_bbox: 1.1042 d2.loss_cls: 0.7758 d2.loss_bbox: 1.0952 d3.loss_cls: 0.7862 d3.loss_bbox: 1.0854 d4.loss_cls: 0.7956 d4.loss_bbox: 1.0811 +10/10 18:14:59 - mmengine - INFO - Epoch(train) [1][2400/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 21:29:49 time: 0.6768 data_time: 0.0389 memory: 11853 grad_norm: 39.8267 loss: 10.9236 loss_cls: 0.7735 loss_bbox: 1.0533 d0.loss_cls: 0.7206 d0.loss_bbox: 1.1450 d1.loss_cls: 0.7399 d1.loss_bbox: 1.0621 d2.loss_cls: 0.7503 d2.loss_bbox: 1.0514 d3.loss_cls: 0.7631 d3.loss_bbox: 1.0480 d4.loss_cls: 0.7652 d4.loss_bbox: 1.0512 +10/10 18:15:33 - mmengine - INFO - Epoch(train) [1][2450/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 21:11:40 time: 0.6780 data_time: 0.0397 memory: 11853 grad_norm: 43.3457 loss: 12.4174 loss_cls: 0.8943 loss_bbox: 1.1765 d0.loss_cls: 0.8284 d0.loss_bbox: 1.3138 d1.loss_cls: 0.8480 d1.loss_bbox: 1.2004 d2.loss_cls: 0.8656 d2.loss_bbox: 1.1812 d3.loss_cls: 0.8757 d3.loss_bbox: 1.1669 d4.loss_cls: 0.8796 d4.loss_bbox: 1.1871 +10/10 18:16:06 - mmengine - INFO - Epoch(train) [1][2500/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 20:54:14 time: 0.6784 data_time: 0.0393 memory: 11853 grad_norm: 40.1775 loss: 12.6597 loss_cls: 0.8944 loss_bbox: 1.2210 d0.loss_cls: 0.8378 d0.loss_bbox: 1.3229 d1.loss_cls: 0.8587 d1.loss_bbox: 1.2359 d2.loss_cls: 0.8721 d2.loss_bbox: 1.2196 d3.loss_cls: 0.8893 d3.loss_bbox: 1.2072 d4.loss_cls: 0.8825 d4.loss_bbox: 1.2182 +10/10 18:16:40 - mmengine - INFO - Epoch(train) [1][2550/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 20:37:16 time: 0.6745 data_time: 0.0392 memory: 11853 grad_norm: 43.6859 loss: 12.6525 loss_cls: 0.9145 loss_bbox: 1.1979 d0.loss_cls: 0.8359 d0.loss_bbox: 1.3127 d1.loss_cls: 0.8726 d1.loss_bbox: 1.2122 d2.loss_cls: 0.8885 d2.loss_bbox: 1.2030 d3.loss_cls: 0.8985 d3.loss_bbox: 1.2016 d4.loss_cls: 0.9062 d4.loss_bbox: 1.2089 +10/10 18:17:14 - mmengine - INFO - Epoch(train) [1][2600/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 20:20:55 time: 0.6747 data_time: 0.0390 memory: 11853 grad_norm: 38.5513 loss: 10.4649 loss_cls: 0.7550 loss_bbox: 0.9887 d0.loss_cls: 0.7026 d0.loss_bbox: 1.0895 d1.loss_cls: 0.7191 d1.loss_bbox: 1.0066 d2.loss_cls: 0.7301 d2.loss_bbox: 0.9936 d3.loss_cls: 0.7398 d3.loss_bbox: 1.0036 d4.loss_cls: 0.7418 d4.loss_bbox: 0.9946 +10/10 18:17:48 - mmengine - INFO - Epoch(train) [1][2650/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 20:05:18 time: 0.6771 data_time: 0.0399 memory: 11853 grad_norm: 42.1461 loss: 10.7324 loss_cls: 0.7581 loss_bbox: 1.0382 d0.loss_cls: 0.7136 d0.loss_bbox: 1.1124 d1.loss_cls: 0.7321 d1.loss_bbox: 1.0517 d2.loss_cls: 0.7405 d2.loss_bbox: 1.0353 d3.loss_cls: 0.7381 d3.loss_bbox: 1.0316 d4.loss_cls: 0.7422 d4.loss_bbox: 1.0386 +10/10 18:18:22 - mmengine - INFO - Epoch(train) [1][2700/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 19:50:13 time: 0.6765 data_time: 0.0392 memory: 11853 grad_norm: 40.1970 loss: 12.2828 loss_cls: 0.8553 loss_bbox: 1.1860 d0.loss_cls: 0.7957 d0.loss_bbox: 1.3123 d1.loss_cls: 0.8144 d1.loss_bbox: 1.2316 d2.loss_cls: 0.8357 d2.loss_bbox: 1.1958 d3.loss_cls: 0.8333 d3.loss_bbox: 1.1920 d4.loss_cls: 0.8522 d4.loss_bbox: 1.1784 +10/10 18:18:56 - mmengine - INFO - Epoch(train) [1][2750/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 19:35:46 time: 0.6787 data_time: 0.0402 memory: 11853 grad_norm: 39.4816 loss: 11.3133 loss_cls: 0.8061 loss_bbox: 1.0900 d0.loss_cls: 0.7484 d0.loss_bbox: 1.1781 d1.loss_cls: 0.7586 d1.loss_bbox: 1.0945 d2.loss_cls: 0.7783 d2.loss_bbox: 1.0907 d3.loss_cls: 0.7822 d3.loss_bbox: 1.0960 d4.loss_cls: 0.7896 d4.loss_bbox: 1.1007 +10/10 18:19:29 - mmengine - INFO - Epoch(train) [1][2800/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 19:21:48 time: 0.6786 data_time: 0.0395 memory: 11853 grad_norm: 40.5397 loss: 10.6832 loss_cls: 0.7600 loss_bbox: 1.0160 d0.loss_cls: 0.7295 d0.loss_bbox: 1.0898 d1.loss_cls: 0.7466 d1.loss_bbox: 1.0383 d2.loss_cls: 0.7507 d2.loss_bbox: 1.0098 d3.loss_cls: 0.7536 d3.loss_bbox: 1.0176 d4.loss_cls: 0.7539 d4.loss_bbox: 1.0175 +10/10 18:20:03 - mmengine - INFO - Epoch(train) [1][2850/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 19:08:13 time: 0.6766 data_time: 0.0401 memory: 11853 grad_norm: 36.6158 loss: 10.7850 loss_cls: 0.7609 loss_bbox: 1.0403 d0.loss_cls: 0.7202 d0.loss_bbox: 1.1115 d1.loss_cls: 0.7300 d1.loss_bbox: 1.0468 d2.loss_cls: 0.7354 d2.loss_bbox: 1.0519 d3.loss_cls: 0.7501 d3.loss_bbox: 1.0449 d4.loss_cls: 0.7595 d4.loss_bbox: 1.0334 +10/10 18:20:37 - mmengine - INFO - Epoch(train) [1][2900/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 18:55:07 time: 0.6772 data_time: 0.0396 memory: 11853 grad_norm: 39.3451 loss: 10.9998 loss_cls: 0.7537 loss_bbox: 1.0943 d0.loss_cls: 0.7098 d0.loss_bbox: 1.1643 d1.loss_cls: 0.7226 d1.loss_bbox: 1.0817 d2.loss_cls: 0.7413 d2.loss_bbox: 1.0814 d3.loss_cls: 0.7476 d3.loss_bbox: 1.0723 d4.loss_cls: 0.7434 d4.loss_bbox: 1.0874 +10/10 18:21:11 - mmengine - INFO - Epoch(train) [1][2950/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 18:42:29 time: 0.6783 data_time: 0.0397 memory: 11853 grad_norm: 36.6020 loss: 10.2437 loss_cls: 0.6942 loss_bbox: 1.0104 d0.loss_cls: 0.6521 d0.loss_bbox: 1.1123 d1.loss_cls: 0.6681 d1.loss_bbox: 1.0344 d2.loss_cls: 0.6757 d2.loss_bbox: 1.0124 d3.loss_cls: 0.6870 d3.loss_bbox: 0.9987 d4.loss_cls: 0.6866 d4.loss_bbox: 1.0119 +10/10 18:21:45 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 18:21:45 - mmengine - INFO - Epoch(train) [1][3000/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 18:30:14 time: 0.6778 data_time: 0.0399 memory: 11853 grad_norm: 45.2384 loss: 11.8740 loss_cls: 0.8463 loss_bbox: 1.1437 d0.loss_cls: 0.7850 d0.loss_bbox: 1.2490 d1.loss_cls: 0.7889 d1.loss_bbox: 1.1915 d2.loss_cls: 0.7971 d2.loss_bbox: 1.1418 d3.loss_cls: 0.8122 d3.loss_bbox: 1.1529 d4.loss_cls: 0.8151 d4.loss_bbox: 1.1505 +10/10 18:22:19 - mmengine - INFO - Epoch(train) [1][3050/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 18:18:20 time: 0.6774 data_time: 0.0396 memory: 11853 grad_norm: 40.4436 loss: 10.5720 loss_cls: 0.7682 loss_bbox: 0.9962 d0.loss_cls: 0.7080 d0.loss_bbox: 1.0968 d1.loss_cls: 0.7188 d1.loss_bbox: 1.0276 d2.loss_cls: 0.7288 d2.loss_bbox: 1.0136 d3.loss_cls: 0.7514 d3.loss_bbox: 1.0059 d4.loss_cls: 0.7595 d4.loss_bbox: 0.9974 +10/10 18:22:53 - mmengine - INFO - Epoch(train) [1][3100/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 18:06:40 time: 0.6742 data_time: 0.0393 memory: 11853 grad_norm: 36.9461 loss: 10.7546 loss_cls: 0.7354 loss_bbox: 1.0506 d0.loss_cls: 0.6813 d0.loss_bbox: 1.1753 d1.loss_cls: 0.6889 d1.loss_bbox: 1.0871 d2.loss_cls: 0.7118 d2.loss_bbox: 1.0618 d3.loss_cls: 0.7173 d3.loss_bbox: 1.0609 d4.loss_cls: 0.7270 d4.loss_bbox: 1.0573 +10/10 18:23:27 - mmengine - INFO - Epoch(train) [1][3150/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 17:55:34 time: 0.6788 data_time: 0.0395 memory: 11853 grad_norm: 43.5819 loss: 11.8406 loss_cls: 0.8394 loss_bbox: 1.1487 d0.loss_cls: 0.7790 d0.loss_bbox: 1.2563 d1.loss_cls: 0.7784 d1.loss_bbox: 1.1794 d2.loss_cls: 0.8019 d2.loss_bbox: 1.1396 d3.loss_cls: 0.8117 d3.loss_bbox: 1.1433 d4.loss_cls: 0.8197 d4.loss_bbox: 1.1432 +10/10 18:24:00 - mmengine - INFO - Epoch(train) [1][3200/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 17:44:36 time: 0.6745 data_time: 0.0394 memory: 11853 grad_norm: 36.5063 loss: 10.6257 loss_cls: 0.7513 loss_bbox: 1.0173 d0.loss_cls: 0.6932 d0.loss_bbox: 1.1275 d1.loss_cls: 0.7135 d1.loss_bbox: 1.0598 d2.loss_cls: 0.7288 d2.loss_bbox: 1.0253 d3.loss_cls: 0.7342 d3.loss_bbox: 1.0160 d4.loss_cls: 0.7460 d4.loss_bbox: 1.0129 +10/10 18:24:34 - mmengine - INFO - Epoch(train) [1][3250/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 17:34:04 time: 0.6773 data_time: 0.0395 memory: 11853 grad_norm: 46.5882 loss: 9.9934 loss_cls: 0.7564 loss_bbox: 0.8971 d0.loss_cls: 0.7040 d0.loss_bbox: 1.0345 d1.loss_cls: 0.7008 d1.loss_bbox: 0.9672 d2.loss_cls: 0.7241 d2.loss_bbox: 0.9182 d3.loss_cls: 0.7322 d3.loss_bbox: 0.9105 d4.loss_cls: 0.7371 d4.loss_bbox: 0.9113 +10/10 18:25:08 - mmengine - INFO - Epoch(train) [1][3300/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 17:23:47 time: 0.6756 data_time: 0.0396 memory: 11853 grad_norm: 39.0033 loss: 10.6673 loss_cls: 0.7697 loss_bbox: 1.0088 d0.loss_cls: 0.6947 d0.loss_bbox: 1.1363 d1.loss_cls: 0.7207 d1.loss_bbox: 1.0470 d2.loss_cls: 0.7326 d2.loss_bbox: 1.0231 d3.loss_cls: 0.7450 d3.loss_bbox: 1.0217 d4.loss_cls: 0.7491 d4.loss_bbox: 1.0188 +10/10 18:25:42 - mmengine - INFO - Epoch(train) [1][3350/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 17:13:56 time: 0.6795 data_time: 0.0397 memory: 11853 grad_norm: 40.2642 loss: 11.8828 loss_cls: 0.8026 loss_bbox: 1.1682 d0.loss_cls: 0.7566 d0.loss_bbox: 1.2758 d1.loss_cls: 0.7748 d1.loss_bbox: 1.2017 d2.loss_cls: 0.7772 d2.loss_bbox: 1.1884 d3.loss_cls: 0.7947 d3.loss_bbox: 1.1768 d4.loss_cls: 0.7895 d4.loss_bbox: 1.1766 +10/10 18:26:16 - mmengine - INFO - Epoch(train) [1][3400/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 17:04:15 time: 0.6766 data_time: 0.0396 memory: 11853 grad_norm: 43.6356 loss: 12.1214 loss_cls: 0.8514 loss_bbox: 1.1712 d0.loss_cls: 0.8084 d0.loss_bbox: 1.2706 d1.loss_cls: 0.8079 d1.loss_bbox: 1.1995 d2.loss_cls: 0.8203 d2.loss_bbox: 1.1679 d3.loss_cls: 0.8351 d3.loss_bbox: 1.1696 d4.loss_cls: 0.8384 d4.loss_bbox: 1.1812 +10/10 18:26:50 - mmengine - INFO - Epoch(train) [1][3450/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 16:54:51 time: 0.6776 data_time: 0.0399 memory: 11853 grad_norm: 49.5292 loss: 12.3369 loss_cls: 0.8889 loss_bbox: 1.1549 d0.loss_cls: 0.8246 d0.loss_bbox: 1.2802 d1.loss_cls: 0.8408 d1.loss_bbox: 1.2102 d2.loss_cls: 0.8571 d2.loss_bbox: 1.1875 d3.loss_cls: 0.8594 d3.loss_bbox: 1.1838 d4.loss_cls: 0.8715 d4.loss_bbox: 1.1779 +10/10 18:27:23 - mmengine - INFO - Epoch(train) [1][3500/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 16:45:38 time: 0.6754 data_time: 0.0398 memory: 11853 grad_norm: 40.8849 loss: 10.6158 loss_cls: 0.7546 loss_bbox: 1.0151 d0.loss_cls: 0.7025 d0.loss_bbox: 1.1309 d1.loss_cls: 0.6965 d1.loss_bbox: 1.0543 d2.loss_cls: 0.7158 d2.loss_bbox: 1.0316 d3.loss_cls: 0.7296 d3.loss_bbox: 1.0211 d4.loss_cls: 0.7442 d4.loss_bbox: 1.0197 +10/10 18:27:57 - mmengine - INFO - Epoch(train) [1][3550/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 16:36:44 time: 0.6774 data_time: 0.0397 memory: 11853 grad_norm: 43.1710 loss: 11.0487 loss_cls: 0.7859 loss_bbox: 1.0488 d0.loss_cls: 0.7406 d0.loss_bbox: 1.1595 d1.loss_cls: 0.7550 d1.loss_bbox: 1.0785 d2.loss_cls: 0.7608 d2.loss_bbox: 1.0684 d3.loss_cls: 0.7720 d3.loss_bbox: 1.0572 d4.loss_cls: 0.7705 d4.loss_bbox: 1.0515 +10/10 18:28:31 - mmengine - INFO - Epoch(train) [1][3600/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 16:28:07 time: 0.6785 data_time: 0.0396 memory: 11853 grad_norm: 38.9145 loss: 11.3763 loss_cls: 0.8158 loss_bbox: 1.0684 d0.loss_cls: 0.7466 d0.loss_bbox: 1.1910 d1.loss_cls: 0.7616 d1.loss_bbox: 1.1356 d2.loss_cls: 0.7948 d2.loss_bbox: 1.0884 d3.loss_cls: 0.7947 d3.loss_bbox: 1.0885 d4.loss_cls: 0.8028 d4.loss_bbox: 1.0881 +10/10 18:29:05 - mmengine - INFO - Epoch(train) [1][3650/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 16:19:41 time: 0.6781 data_time: 0.0392 memory: 11853 grad_norm: 42.4214 loss: 10.3950 loss_cls: 0.7340 loss_bbox: 1.0019 d0.loss_cls: 0.6870 d0.loss_bbox: 1.0878 d1.loss_cls: 0.6929 d1.loss_bbox: 1.0403 d2.loss_cls: 0.7086 d2.loss_bbox: 1.0068 d3.loss_cls: 0.7138 d3.loss_bbox: 0.9972 d4.loss_cls: 0.7152 d4.loss_bbox: 1.0097 +10/10 18:29:39 - mmengine - INFO - Epoch(train) [1][3700/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 16:11:27 time: 0.6770 data_time: 0.0394 memory: 11853 grad_norm: 41.1067 loss: 10.8356 loss_cls: 0.7717 loss_bbox: 1.0189 d0.loss_cls: 0.7208 d0.loss_bbox: 1.1505 d1.loss_cls: 0.7499 d1.loss_bbox: 1.0638 d2.loss_cls: 0.7463 d2.loss_bbox: 1.0411 d3.loss_cls: 0.7506 d3.loss_bbox: 1.0321 d4.loss_cls: 0.7586 d4.loss_bbox: 1.0315 +10/10 18:30:13 - mmengine - INFO - Epoch(train) [1][3750/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 16:03:23 time: 0.6766 data_time: 0.0396 memory: 11853 grad_norm: 41.8763 loss: 10.7093 loss_cls: 0.7628 loss_bbox: 1.0177 d0.loss_cls: 0.7125 d0.loss_bbox: 1.1128 d1.loss_cls: 0.7327 d1.loss_bbox: 1.0554 d2.loss_cls: 0.7387 d2.loss_bbox: 1.0377 d3.loss_cls: 0.7430 d3.loss_bbox: 1.0259 d4.loss_cls: 0.7473 d4.loss_bbox: 1.0228 +10/10 18:30:47 - mmengine - INFO - Epoch(train) [1][3800/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:55:48 time: 0.6844 data_time: 0.0393 memory: 11853 grad_norm: 45.6190 loss: 11.2612 loss_cls: 0.7751 loss_bbox: 1.0827 d0.loss_cls: 0.7500 d0.loss_bbox: 1.1754 d1.loss_cls: 0.7791 d1.loss_bbox: 1.1127 d2.loss_cls: 0.7754 d2.loss_bbox: 1.0869 d3.loss_cls: 0.7733 d3.loss_bbox: 1.0866 d4.loss_cls: 0.7793 d4.loss_bbox: 1.0847 +10/10 18:31:21 - mmengine - INFO - Epoch(train) [1][3850/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:48:08 time: 0.6767 data_time: 0.0394 memory: 11853 grad_norm: 43.6253 loss: 10.7960 loss_cls: 0.7594 loss_bbox: 1.0487 d0.loss_cls: 0.7053 d0.loss_bbox: 1.1276 d1.loss_cls: 0.7143 d1.loss_bbox: 1.0601 d2.loss_cls: 0.7422 d2.loss_bbox: 1.0385 d3.loss_cls: 0.7405 d3.loss_bbox: 1.0551 d4.loss_cls: 0.7546 d4.loss_bbox: 1.0496 +10/10 18:31:55 - mmengine - INFO - Epoch(train) [1][3900/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:40:45 time: 0.6800 data_time: 0.0396 memory: 11853 grad_norm: 41.2186 loss: 11.4945 loss_cls: 0.8324 loss_bbox: 1.0875 d0.loss_cls: 0.7830 d0.loss_bbox: 1.1657 d1.loss_cls: 0.8089 d1.loss_bbox: 1.1001 d2.loss_cls: 0.8117 d2.loss_bbox: 1.0848 d3.loss_cls: 0.8202 d3.loss_bbox: 1.0848 d4.loss_cls: 0.8321 d4.loss_bbox: 1.0832 +10/10 18:32:29 - mmengine - INFO - Epoch(train) [1][3950/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:33:30 time: 0.6782 data_time: 0.0400 memory: 11853 grad_norm: 40.8335 loss: 10.4033 loss_cls: 0.7005 loss_bbox: 1.0249 d0.loss_cls: 0.6706 d0.loss_bbox: 1.1036 d1.loss_cls: 0.6893 d1.loss_bbox: 1.0383 d2.loss_cls: 0.7009 d2.loss_bbox: 1.0228 d3.loss_cls: 0.7112 d3.loss_bbox: 1.0197 d4.loss_cls: 0.6993 d4.loss_bbox: 1.0221 +10/10 18:33:03 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 18:33:03 - mmengine - INFO - Epoch(train) [1][4000/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:26:23 time: 0.6776 data_time: 0.0399 memory: 11853 grad_norm: 38.2631 loss: 11.1154 loss_cls: 0.7692 loss_bbox: 1.0912 d0.loss_cls: 0.7145 d0.loss_bbox: 1.1790 d1.loss_cls: 0.7357 d1.loss_bbox: 1.0946 d2.loss_cls: 0.7475 d2.loss_bbox: 1.0778 d3.loss_cls: 0.7596 d3.loss_bbox: 1.0922 d4.loss_cls: 0.7618 d4.loss_bbox: 1.0924 +10/10 18:33:37 - mmengine - INFO - Epoch(train) [1][4050/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:19:39 time: 0.6841 data_time: 0.0406 memory: 11853 grad_norm: 38.5025 loss: 11.8008 loss_cls: 0.8325 loss_bbox: 1.1275 d0.loss_cls: 0.7701 d0.loss_bbox: 1.2424 d1.loss_cls: 0.7923 d1.loss_bbox: 1.1720 d2.loss_cls: 0.8064 d2.loss_bbox: 1.1373 d3.loss_cls: 0.8191 d3.loss_bbox: 1.1386 d4.loss_cls: 0.8266 d4.loss_bbox: 1.1359 +10/10 18:34:11 - mmengine - INFO - Epoch(train) [1][4100/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:12:55 time: 0.6797 data_time: 0.0398 memory: 11853 grad_norm: 36.9437 loss: 10.1804 loss_cls: 0.7065 loss_bbox: 0.9855 d0.loss_cls: 0.6568 d0.loss_bbox: 1.0936 d1.loss_cls: 0.6632 d1.loss_bbox: 1.0216 d2.loss_cls: 0.6820 d2.loss_bbox: 0.9958 d3.loss_cls: 0.6996 d3.loss_bbox: 0.9890 d4.loss_cls: 0.6994 d4.loss_bbox: 0.9876 +10/10 18:34:45 - mmengine - INFO - Epoch(train) [1][4150/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 15:06:15 time: 0.6770 data_time: 0.0392 memory: 11853 grad_norm: 43.6822 loss: 10.6372 loss_cls: 0.7504 loss_bbox: 1.0252 d0.loss_cls: 0.7398 d0.loss_bbox: 1.0915 d1.loss_cls: 0.7284 d1.loss_bbox: 1.0233 d2.loss_cls: 0.7433 d2.loss_bbox: 1.0047 d3.loss_cls: 0.7465 d3.loss_bbox: 1.0136 d4.loss_cls: 0.7575 d4.loss_bbox: 1.0129 +10/10 18:35:18 - mmengine - INFO - Epoch(train) [1][4200/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:59:43 time: 0.6768 data_time: 0.0398 memory: 11853 grad_norm: 40.6914 loss: 11.0356 loss_cls: 0.7660 loss_bbox: 1.0650 d0.loss_cls: 0.7347 d0.loss_bbox: 1.1611 d1.loss_cls: 0.7427 d1.loss_bbox: 1.0887 d2.loss_cls: 0.7500 d2.loss_bbox: 1.0693 d3.loss_cls: 0.7571 d3.loss_bbox: 1.0674 d4.loss_cls: 0.7654 d4.loss_bbox: 1.0681 +10/10 18:35:52 - mmengine - INFO - Epoch(train) [1][4250/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:53:22 time: 0.6783 data_time: 0.0401 memory: 11853 grad_norm: 44.0803 loss: 10.2409 loss_cls: 0.7304 loss_bbox: 0.9704 d0.loss_cls: 0.6921 d0.loss_bbox: 1.0525 d1.loss_cls: 0.7034 d1.loss_bbox: 0.9988 d2.loss_cls: 0.7117 d2.loss_bbox: 0.9795 d3.loss_cls: 0.7266 d3.loss_bbox: 0.9804 d4.loss_cls: 0.7239 d4.loss_bbox: 0.9714 +10/10 18:36:26 - mmengine - INFO - Epoch(train) [1][4300/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:47:08 time: 0.6774 data_time: 0.0397 memory: 11853 grad_norm: 43.5833 loss: 9.6646 loss_cls: 0.7050 loss_bbox: 0.9069 d0.loss_cls: 0.6709 d0.loss_bbox: 0.9921 d1.loss_cls: 0.6688 d1.loss_bbox: 0.9286 d2.loss_cls: 0.6747 d2.loss_bbox: 0.9073 d3.loss_cls: 0.6946 d3.loss_bbox: 0.9127 d4.loss_cls: 0.7008 d4.loss_bbox: 0.9021 +10/10 18:37:00 - mmengine - INFO - Epoch(train) [1][4350/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:40:59 time: 0.6762 data_time: 0.0399 memory: 11853 grad_norm: 42.0417 loss: 10.1378 loss_cls: 0.7367 loss_bbox: 0.9507 d0.loss_cls: 0.6943 d0.loss_bbox: 1.0351 d1.loss_cls: 0.6990 d1.loss_bbox: 0.9778 d2.loss_cls: 0.7200 d2.loss_bbox: 0.9490 d3.loss_cls: 0.7300 d3.loss_bbox: 0.9568 d4.loss_cls: 0.7266 d4.loss_bbox: 0.9618 +10/10 18:37:34 - mmengine - INFO - Epoch(train) [1][4400/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:34:57 time: 0.6760 data_time: 0.0396 memory: 11853 grad_norm: 40.9308 loss: 10.5481 loss_cls: 0.7286 loss_bbox: 1.0344 d0.loss_cls: 0.6917 d0.loss_bbox: 1.0994 d1.loss_cls: 0.7045 d1.loss_bbox: 1.0322 d2.loss_cls: 0.7157 d2.loss_bbox: 1.0226 d3.loss_cls: 0.7270 d3.loss_bbox: 1.0319 d4.loss_cls: 0.7245 d4.loss_bbox: 1.0357 +10/10 18:38:08 - mmengine - INFO - Epoch(train) [1][4450/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:29:07 time: 0.6781 data_time: 0.0400 memory: 11853 grad_norm: 41.2725 loss: 10.1216 loss_cls: 0.7042 loss_bbox: 0.9674 d0.loss_cls: 0.6579 d0.loss_bbox: 1.0863 d1.loss_cls: 0.6717 d1.loss_bbox: 1.0069 d2.loss_cls: 0.6833 d2.loss_bbox: 0.9866 d3.loss_cls: 0.7005 d3.loss_bbox: 0.9795 d4.loss_cls: 0.7064 d4.loss_bbox: 0.9708 +10/10 18:38:42 - mmengine - INFO - Epoch(train) [1][4500/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:23:23 time: 0.6777 data_time: 0.0396 memory: 11853 grad_norm: 43.1510 loss: 10.3821 loss_cls: 0.7464 loss_bbox: 0.9812 d0.loss_cls: 0.6899 d0.loss_bbox: 1.0947 d1.loss_cls: 0.7006 d1.loss_bbox: 1.0125 d2.loss_cls: 0.7139 d2.loss_bbox: 0.9873 d3.loss_cls: 0.7286 d3.loss_bbox: 0.9912 d4.loss_cls: 0.7401 d4.loss_bbox: 0.9955 +10/10 18:39:16 - mmengine - INFO - Epoch(train) [1][4550/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:17:47 time: 0.6783 data_time: 0.0390 memory: 11853 grad_norm: 42.4591 loss: 10.1885 loss_cls: 0.7286 loss_bbox: 0.9649 d0.loss_cls: 0.6833 d0.loss_bbox: 1.0618 d1.loss_cls: 0.7008 d1.loss_bbox: 0.9864 d2.loss_cls: 0.7055 d2.loss_bbox: 0.9759 d3.loss_cls: 0.7199 d3.loss_bbox: 0.9784 d4.loss_cls: 0.7142 d4.loss_bbox: 0.9688 +10/10 18:39:49 - mmengine - INFO - Epoch(train) [1][4600/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:12:15 time: 0.6766 data_time: 0.0394 memory: 11853 grad_norm: 43.5540 loss: 10.5430 loss_cls: 0.7568 loss_bbox: 0.9985 d0.loss_cls: 0.6850 d0.loss_bbox: 1.1306 d1.loss_cls: 0.7096 d1.loss_bbox: 1.0287 d2.loss_cls: 0.7411 d2.loss_bbox: 1.0076 d3.loss_cls: 0.7398 d3.loss_bbox: 0.9957 d4.loss_cls: 0.7500 d4.loss_bbox: 0.9996 +10/10 18:40:23 - mmengine - INFO - Epoch(train) [1][4650/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:06:47 time: 0.6760 data_time: 0.0395 memory: 11853 grad_norm: 39.5251 loss: 11.1353 loss_cls: 0.8098 loss_bbox: 1.0716 d0.loss_cls: 0.7377 d0.loss_bbox: 1.1548 d1.loss_cls: 0.7476 d1.loss_bbox: 1.0794 d2.loss_cls: 0.7644 d2.loss_bbox: 1.0596 d3.loss_cls: 0.7894 d3.loss_bbox: 1.0640 d4.loss_cls: 0.7947 d4.loss_bbox: 1.0623 +10/10 18:40:57 - mmengine - INFO - Epoch(train) [1][4700/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 14:01:28 time: 0.6769 data_time: 0.0392 memory: 11853 grad_norm: 39.9127 loss: 10.3399 loss_cls: 0.7191 loss_bbox: 0.9950 d0.loss_cls: 0.6696 d0.loss_bbox: 1.1133 d1.loss_cls: 0.6847 d1.loss_bbox: 1.0388 d2.loss_cls: 0.6950 d2.loss_bbox: 1.0049 d3.loss_cls: 0.7031 d3.loss_bbox: 1.0060 d4.loss_cls: 0.7165 d4.loss_bbox: 0.9940 +10/10 18:41:31 - mmengine - INFO - Epoch(train) [1][4750/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:56:16 time: 0.6780 data_time: 0.0395 memory: 11853 grad_norm: 48.3767 loss: 10.3973 loss_cls: 0.7593 loss_bbox: 0.9695 d0.loss_cls: 0.6847 d0.loss_bbox: 1.0846 d1.loss_cls: 0.7213 d1.loss_bbox: 0.9970 d2.loss_cls: 0.7335 d2.loss_bbox: 0.9836 d3.loss_cls: 0.7443 d3.loss_bbox: 0.9829 d4.loss_cls: 0.7505 d4.loss_bbox: 0.9860 +10/10 18:42:05 - mmengine - INFO - Epoch(train) [1][4800/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:51:11 time: 0.6780 data_time: 0.0404 memory: 11853 grad_norm: 41.5217 loss: 9.2194 loss_cls: 0.6539 loss_bbox: 0.8866 d0.loss_cls: 0.5973 d0.loss_bbox: 0.9795 d1.loss_cls: 0.5959 d1.loss_bbox: 0.9239 d2.loss_cls: 0.6197 d2.loss_bbox: 0.9017 d3.loss_cls: 0.6288 d3.loss_bbox: 0.8954 d4.loss_cls: 0.6456 d4.loss_bbox: 0.8911 +10/10 18:42:39 - mmengine - INFO - Epoch(train) [1][4850/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:46:15 time: 0.6805 data_time: 0.0406 memory: 11853 grad_norm: 44.2651 loss: 10.5149 loss_cls: 0.7173 loss_bbox: 1.0248 d0.loss_cls: 0.6801 d0.loss_bbox: 1.1373 d1.loss_cls: 0.6800 d1.loss_bbox: 1.0720 d2.loss_cls: 0.6937 d2.loss_bbox: 1.0332 d3.loss_cls: 0.7052 d3.loss_bbox: 1.0335 d4.loss_cls: 0.7069 d4.loss_bbox: 1.0309 +10/10 18:43:13 - mmengine - INFO - Epoch(train) [1][4900/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:41:22 time: 0.6791 data_time: 0.0402 memory: 11853 grad_norm: 43.2066 loss: 12.1598 loss_cls: 0.8553 loss_bbox: 1.1609 d0.loss_cls: 0.7857 d0.loss_bbox: 1.3086 d1.loss_cls: 0.8004 d1.loss_bbox: 1.2131 d2.loss_cls: 0.8197 d2.loss_bbox: 1.1843 d3.loss_cls: 0.8441 d3.loss_bbox: 1.1722 d4.loss_cls: 0.8483 d4.loss_bbox: 1.1671 +10/10 18:43:47 - mmengine - INFO - Epoch(train) [1][4950/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:36:29 time: 0.6761 data_time: 0.0397 memory: 11853 grad_norm: 42.1235 loss: 10.5249 loss_cls: 0.7522 loss_bbox: 0.9896 d0.loss_cls: 0.7038 d0.loss_bbox: 1.1131 d1.loss_cls: 0.7138 d1.loss_bbox: 1.0375 d2.loss_cls: 0.7208 d2.loss_bbox: 1.0148 d3.loss_cls: 0.7449 d3.loss_bbox: 1.0019 d4.loss_cls: 0.7441 d4.loss_bbox: 0.9884 +10/10 18:44:21 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 18:44:21 - mmengine - INFO - Epoch(train) [1][5000/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:31:47 time: 0.6790 data_time: 0.0400 memory: 11853 grad_norm: 44.7480 loss: 10.9483 loss_cls: 0.7766 loss_bbox: 1.0429 d0.loss_cls: 0.7236 d0.loss_bbox: 1.1454 d1.loss_cls: 0.7280 d1.loss_bbox: 1.0858 d2.loss_cls: 0.7553 d2.loss_bbox: 1.0639 d3.loss_cls: 0.7600 d3.loss_bbox: 1.0540 d4.loss_cls: 0.7639 d4.loss_bbox: 1.0490 +10/10 18:44:55 - mmengine - INFO - Epoch(train) [1][5050/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:27:09 time: 0.6789 data_time: 0.0402 memory: 11853 grad_norm: 41.4967 loss: 9.9738 loss_cls: 0.6977 loss_bbox: 0.9598 d0.loss_cls: 0.6552 d0.loss_bbox: 1.0602 d1.loss_cls: 0.6673 d1.loss_bbox: 0.9805 d2.loss_cls: 0.6770 d2.loss_bbox: 0.9644 d3.loss_cls: 0.6764 d3.loss_bbox: 0.9785 d4.loss_cls: 0.6825 d4.loss_bbox: 0.9744 +10/10 18:45:29 - mmengine - INFO - Epoch(train) [1][5100/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:22:35 time: 0.6782 data_time: 0.0401 memory: 11853 grad_norm: 39.0752 loss: 11.5744 loss_cls: 0.8090 loss_bbox: 1.1252 d0.loss_cls: 0.7633 d0.loss_bbox: 1.2324 d1.loss_cls: 0.7619 d1.loss_bbox: 1.1530 d2.loss_cls: 0.7688 d2.loss_bbox: 1.1248 d3.loss_cls: 0.7792 d3.loss_bbox: 1.1346 d4.loss_cls: 0.7884 d4.loss_bbox: 1.1340 +10/10 18:46:03 - mmengine - INFO - Epoch(train) [1][5150/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:18:11 time: 0.6821 data_time: 0.0397 memory: 11853 grad_norm: 40.1207 loss: 10.2415 loss_cls: 0.6936 loss_bbox: 0.9988 d0.loss_cls: 0.6596 d0.loss_bbox: 1.1194 d1.loss_cls: 0.6673 d1.loss_bbox: 1.0340 d2.loss_cls: 0.6739 d2.loss_bbox: 1.0081 d3.loss_cls: 0.6787 d3.loss_bbox: 1.0099 d4.loss_cls: 0.6886 d4.loss_bbox: 1.0096 +10/10 18:46:37 - mmengine - INFO - Epoch(train) [1][5200/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:13:49 time: 0.6798 data_time: 0.0399 memory: 11853 grad_norm: 41.2157 loss: 10.8800 loss_cls: 0.7636 loss_bbox: 1.0453 d0.loss_cls: 0.7258 d0.loss_bbox: 1.1389 d1.loss_cls: 0.7372 d1.loss_bbox: 1.0619 d2.loss_cls: 0.7472 d2.loss_bbox: 1.0438 d3.loss_cls: 0.7457 d3.loss_bbox: 1.0565 d4.loss_cls: 0.7522 d4.loss_bbox: 1.0619 +10/10 18:47:11 - mmengine - INFO - Epoch(train) [1][5250/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:09:30 time: 0.6792 data_time: 0.0400 memory: 11853 grad_norm: 41.8715 loss: 11.2577 loss_cls: 0.7691 loss_bbox: 1.1031 d0.loss_cls: 0.7193 d0.loss_bbox: 1.2197 d1.loss_cls: 0.7325 d1.loss_bbox: 1.1271 d2.loss_cls: 0.7520 d2.loss_bbox: 1.0932 d3.loss_cls: 0.7561 d3.loss_bbox: 1.1112 d4.loss_cls: 0.7617 d4.loss_bbox: 1.1127 +10/10 18:47:45 - mmengine - INFO - Epoch(train) [1][5300/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:05:20 time: 0.6829 data_time: 0.0398 memory: 11853 grad_norm: 38.4806 loss: 10.0052 loss_cls: 0.7244 loss_bbox: 0.9278 d0.loss_cls: 0.6751 d0.loss_bbox: 1.0477 d1.loss_cls: 0.6934 d1.loss_bbox: 0.9652 d2.loss_cls: 0.7085 d2.loss_bbox: 0.9489 d3.loss_cls: 0.7087 d3.loss_bbox: 0.9468 d4.loss_cls: 0.7220 d4.loss_bbox: 0.9367 +10/10 18:48:19 - mmengine - INFO - Epoch(train) [1][5350/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 13:01:12 time: 0.6810 data_time: 0.0402 memory: 11853 grad_norm: 41.8446 loss: 11.3227 loss_cls: 0.7885 loss_bbox: 1.0932 d0.loss_cls: 0.7611 d0.loss_bbox: 1.1890 d1.loss_cls: 0.7635 d1.loss_bbox: 1.1174 d2.loss_cls: 0.7704 d2.loss_bbox: 1.0924 d3.loss_cls: 0.7786 d3.loss_bbox: 1.0875 d4.loss_cls: 0.7852 d4.loss_bbox: 1.0958 +10/10 18:48:53 - mmengine - INFO - Epoch(train) [1][5400/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:57:06 time: 0.6793 data_time: 0.0401 memory: 11853 grad_norm: 44.8137 loss: 10.8336 loss_cls: 0.7737 loss_bbox: 1.0362 d0.loss_cls: 0.7087 d0.loss_bbox: 1.1204 d1.loss_cls: 0.7340 d1.loss_bbox: 1.0524 d2.loss_cls: 0.7674 d2.loss_bbox: 1.0250 d3.loss_cls: 0.7677 d3.loss_bbox: 1.0339 d4.loss_cls: 0.7727 d4.loss_bbox: 1.0416 +10/10 18:49:27 - mmengine - INFO - Epoch(train) [1][5450/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:53:02 time: 0.6790 data_time: 0.0402 memory: 11853 grad_norm: 37.9306 loss: 9.6040 loss_cls: 0.6939 loss_bbox: 0.9106 d0.loss_cls: 0.6530 d0.loss_bbox: 0.9808 d1.loss_cls: 0.6625 d1.loss_bbox: 0.9247 d2.loss_cls: 0.6723 d2.loss_bbox: 0.9094 d3.loss_cls: 0.6793 d3.loss_bbox: 0.9162 d4.loss_cls: 0.6837 d4.loss_bbox: 0.9177 +10/10 18:50:01 - mmengine - INFO - Epoch(train) [1][5500/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:49:03 time: 0.6789 data_time: 0.0397 memory: 11853 grad_norm: 44.2818 loss: 11.1741 loss_cls: 0.7920 loss_bbox: 1.0620 d0.loss_cls: 0.7445 d0.loss_bbox: 1.1756 d1.loss_cls: 0.7605 d1.loss_bbox: 1.0774 d2.loss_cls: 0.7896 d2.loss_bbox: 1.0551 d3.loss_cls: 0.7931 d3.loss_bbox: 1.0653 d4.loss_cls: 0.7858 d4.loss_bbox: 1.0732 +10/10 18:50:35 - mmengine - INFO - Epoch(train) [1][5550/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:45:08 time: 0.6793 data_time: 0.0398 memory: 11853 grad_norm: 40.4141 loss: 10.0262 loss_cls: 0.6919 loss_bbox: 0.9743 d0.loss_cls: 0.6528 d0.loss_bbox: 1.0675 d1.loss_cls: 0.6598 d1.loss_bbox: 1.0014 d2.loss_cls: 0.6812 d2.loss_bbox: 0.9727 d3.loss_cls: 0.6927 d3.loss_bbox: 0.9705 d4.loss_cls: 0.6899 d4.loss_bbox: 0.9715 +10/10 18:51:08 - mmengine - INFO - Epoch(train) [1][5600/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:41:13 time: 0.6771 data_time: 0.0398 memory: 11853 grad_norm: 42.3421 loss: 9.9868 loss_cls: 0.7075 loss_bbox: 0.9579 d0.loss_cls: 0.6462 d0.loss_bbox: 1.0969 d1.loss_cls: 0.6726 d1.loss_bbox: 0.9781 d2.loss_cls: 0.6644 d2.loss_bbox: 0.9718 d3.loss_cls: 0.6836 d3.loss_bbox: 0.9529 d4.loss_cls: 0.6941 d4.loss_bbox: 0.9607 +10/10 18:51:42 - mmengine - INFO - Epoch(train) [1][5650/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:37:21 time: 0.6765 data_time: 0.0395 memory: 11853 grad_norm: 44.9418 loss: 10.3759 loss_cls: 0.7370 loss_bbox: 1.0062 d0.loss_cls: 0.6788 d0.loss_bbox: 1.1073 d1.loss_cls: 0.7002 d1.loss_bbox: 1.0170 d2.loss_cls: 0.7150 d2.loss_bbox: 0.9862 d3.loss_cls: 0.7203 d3.loss_bbox: 0.9912 d4.loss_cls: 0.7299 d4.loss_bbox: 0.9867 +10/10 18:52:16 - mmengine - INFO - Epoch(train) [1][5700/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:33:37 time: 0.6804 data_time: 0.0393 memory: 11853 grad_norm: 46.2903 loss: 10.6523 loss_cls: 0.7649 loss_bbox: 1.0339 d0.loss_cls: 0.6887 d0.loss_bbox: 1.1159 d1.loss_cls: 0.7279 d1.loss_bbox: 1.0235 d2.loss_cls: 0.7400 d2.loss_bbox: 1.0073 d3.loss_cls: 0.7551 d3.loss_bbox: 1.0141 d4.loss_cls: 0.7605 d4.loss_bbox: 1.0205 +10/10 18:52:50 - mmengine - INFO - Epoch(train) [1][5750/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:29:53 time: 0.6773 data_time: 0.0390 memory: 11853 grad_norm: 40.6780 loss: 11.1220 loss_cls: 0.7847 loss_bbox: 1.0799 d0.loss_cls: 0.7064 d0.loss_bbox: 1.1624 d1.loss_cls: 0.7401 d1.loss_bbox: 1.0897 d2.loss_cls: 0.7559 d2.loss_bbox: 1.0821 d3.loss_cls: 0.7809 d3.loss_bbox: 1.0835 d4.loss_cls: 0.7802 d4.loss_bbox: 1.0764 +10/10 18:53:24 - mmengine - INFO - Epoch(train) [1][5800/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:26:13 time: 0.6777 data_time: 0.0395 memory: 11853 grad_norm: 42.9323 loss: 9.9641 loss_cls: 0.6972 loss_bbox: 0.9592 d0.loss_cls: 0.6602 d0.loss_bbox: 1.0462 d1.loss_cls: 0.6784 d1.loss_bbox: 0.9742 d2.loss_cls: 0.6750 d2.loss_bbox: 0.9666 d3.loss_cls: 0.6855 d3.loss_bbox: 0.9633 d4.loss_cls: 0.6892 d4.loss_bbox: 0.9691 +10/10 18:53:58 - mmengine - INFO - Epoch(train) [1][5850/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:22:37 time: 0.6787 data_time: 0.0391 memory: 11853 grad_norm: 44.0460 loss: 10.1711 loss_cls: 0.7292 loss_bbox: 0.9696 d0.loss_cls: 0.6913 d0.loss_bbox: 1.0727 d1.loss_cls: 0.6919 d1.loss_bbox: 0.9954 d2.loss_cls: 0.6944 d2.loss_bbox: 0.9759 d3.loss_cls: 0.7100 d3.loss_bbox: 0.9628 d4.loss_cls: 0.7134 d4.loss_bbox: 0.9645 +10/10 18:54:32 - mmengine - INFO - Epoch(train) [1][5900/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:19:03 time: 0.6777 data_time: 0.0389 memory: 11853 grad_norm: 41.7618 loss: 10.6561 loss_cls: 0.7436 loss_bbox: 1.0515 d0.loss_cls: 0.6885 d0.loss_bbox: 1.1338 d1.loss_cls: 0.7097 d1.loss_bbox: 1.0541 d2.loss_cls: 0.7238 d2.loss_bbox: 1.0298 d3.loss_cls: 0.7253 d3.loss_bbox: 1.0297 d4.loss_cls: 0.7288 d4.loss_bbox: 1.0374 +10/10 18:55:06 - mmengine - INFO - Epoch(train) [1][5950/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:15:29 time: 0.6755 data_time: 0.0395 memory: 11853 grad_norm: 44.7422 loss: 10.3734 loss_cls: 0.7405 loss_bbox: 0.9941 d0.loss_cls: 0.6802 d0.loss_bbox: 1.0967 d1.loss_cls: 0.6969 d1.loss_bbox: 1.0200 d2.loss_cls: 0.7184 d2.loss_bbox: 0.9940 d3.loss_cls: 0.7118 d3.loss_bbox: 1.0032 d4.loss_cls: 0.7229 d4.loss_bbox: 0.9946 +10/10 18:55:40 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 18:55:40 - mmengine - INFO - Epoch(train) [1][6000/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:12:02 time: 0.6786 data_time: 0.0395 memory: 11853 grad_norm: 45.3890 loss: 11.8164 loss_cls: 0.8446 loss_bbox: 1.1169 d0.loss_cls: 0.8034 d0.loss_bbox: 1.2141 d1.loss_cls: 0.8184 d1.loss_bbox: 1.1434 d2.loss_cls: 0.8308 d2.loss_bbox: 1.1197 d3.loss_cls: 0.8408 d3.loss_bbox: 1.1158 d4.loss_cls: 0.8473 d4.loss_bbox: 1.1213 +10/10 18:56:14 - mmengine - INFO - Epoch(train) [1][6050/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:08:38 time: 0.6786 data_time: 0.0392 memory: 11853 grad_norm: 52.9529 loss: 10.6904 loss_cls: 0.7525 loss_bbox: 1.0453 d0.loss_cls: 0.7020 d0.loss_bbox: 1.1462 d1.loss_cls: 0.7077 d1.loss_bbox: 1.0539 d2.loss_cls: 0.7176 d2.loss_bbox: 1.0349 d3.loss_cls: 0.7220 d3.loss_bbox: 1.0422 d4.loss_cls: 0.7384 d4.loss_bbox: 1.0278 +10/10 18:56:47 - mmengine - INFO - Epoch(train) [1][6100/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:05:12 time: 0.6756 data_time: 0.0387 memory: 11853 grad_norm: 39.8678 loss: 10.2851 loss_cls: 0.7250 loss_bbox: 0.9779 d0.loss_cls: 0.6787 d0.loss_bbox: 1.0901 d1.loss_cls: 0.6941 d1.loss_bbox: 1.0138 d2.loss_cls: 0.7047 d2.loss_bbox: 0.9974 d3.loss_cls: 0.7199 d3.loss_bbox: 0.9846 d4.loss_cls: 0.7251 d4.loss_bbox: 0.9736 +10/10 18:57:21 - mmengine - INFO - Epoch(train) [1][6150/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 12:01:50 time: 0.6756 data_time: 0.0387 memory: 11853 grad_norm: 45.7940 loss: 9.9052 loss_cls: 0.7058 loss_bbox: 0.9517 d0.loss_cls: 0.6581 d0.loss_bbox: 1.0370 d1.loss_cls: 0.6654 d1.loss_bbox: 0.9611 d2.loss_cls: 0.6818 d2.loss_bbox: 0.9539 d3.loss_cls: 0.6902 d3.loss_bbox: 0.9544 d4.loss_cls: 0.6948 d4.loss_bbox: 0.9510 +10/10 18:57:55 - mmengine - INFO - Epoch(train) [1][6200/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:58:33 time: 0.6774 data_time: 0.0391 memory: 11853 grad_norm: 38.5724 loss: 9.8934 loss_cls: 0.6899 loss_bbox: 0.9489 d0.loss_cls: 0.6453 d0.loss_bbox: 1.0516 d1.loss_cls: 0.6610 d1.loss_bbox: 0.9807 d2.loss_cls: 0.6741 d2.loss_bbox: 0.9617 d3.loss_cls: 0.6827 d3.loss_bbox: 0.9551 d4.loss_cls: 0.6897 d4.loss_bbox: 0.9528 +10/10 18:58:29 - mmengine - INFO - Epoch(train) [1][6250/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:55:21 time: 0.6797 data_time: 0.0397 memory: 11853 grad_norm: 45.6219 loss: 9.9063 loss_cls: 0.7009 loss_bbox: 0.9440 d0.loss_cls: 0.6594 d0.loss_bbox: 1.0554 d1.loss_cls: 0.6668 d1.loss_bbox: 0.9757 d2.loss_cls: 0.6777 d2.loss_bbox: 0.9523 d3.loss_cls: 0.6848 d3.loss_bbox: 0.9497 d4.loss_cls: 0.6913 d4.loss_bbox: 0.9483 +10/10 18:59:03 - mmengine - INFO - Epoch(train) [1][6300/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:52:09 time: 0.6773 data_time: 0.0392 memory: 11853 grad_norm: 43.6010 loss: 11.0672 loss_cls: 0.7904 loss_bbox: 1.0692 d0.loss_cls: 0.7291 d0.loss_bbox: 1.1655 d1.loss_cls: 0.7499 d1.loss_bbox: 1.0668 d2.loss_cls: 0.7596 d2.loss_bbox: 1.0564 d3.loss_cls: 0.7665 d3.loss_bbox: 1.0600 d4.loss_cls: 0.7884 d4.loss_bbox: 1.0653 +10/10 18:59:37 - mmengine - INFO - Epoch(train) [1][6350/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:48:59 time: 0.6772 data_time: 0.0392 memory: 11853 grad_norm: 43.7097 loss: 9.8717 loss_cls: 0.6869 loss_bbox: 0.9543 d0.loss_cls: 0.6479 d0.loss_bbox: 1.0451 d1.loss_cls: 0.6538 d1.loss_bbox: 0.9878 d2.loss_cls: 0.6656 d2.loss_bbox: 0.9620 d3.loss_cls: 0.6784 d3.loss_bbox: 0.9578 d4.loss_cls: 0.6847 d4.loss_bbox: 0.9473 +10/10 19:00:11 - mmengine - INFO - Epoch(train) [1][6400/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:45:54 time: 0.6795 data_time: 0.0393 memory: 11853 grad_norm: 42.4176 loss: 10.8069 loss_cls: 0.7677 loss_bbox: 1.0305 d0.loss_cls: 0.7064 d0.loss_bbox: 1.1532 d1.loss_cls: 0.7127 d1.loss_bbox: 1.0722 d2.loss_cls: 0.7266 d2.loss_bbox: 1.0586 d3.loss_cls: 0.7447 d3.loss_bbox: 1.0450 d4.loss_cls: 0.7511 d4.loss_bbox: 1.0384 +10/10 19:00:45 - mmengine - INFO - Epoch(train) [1][6450/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:42:50 time: 0.6785 data_time: 0.0395 memory: 11853 grad_norm: 45.4923 loss: 10.7052 loss_cls: 0.7641 loss_bbox: 1.0259 d0.loss_cls: 0.7161 d0.loss_bbox: 1.1222 d1.loss_cls: 0.7279 d1.loss_bbox: 1.0419 d2.loss_cls: 0.7353 d2.loss_bbox: 1.0317 d3.loss_cls: 0.7534 d3.loss_bbox: 1.0191 d4.loss_cls: 0.7567 d4.loss_bbox: 1.0108 +10/10 19:01:18 - mmengine - INFO - Epoch(train) [1][6500/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:39:48 time: 0.6777 data_time: 0.0395 memory: 11853 grad_norm: 46.5516 loss: 11.5517 loss_cls: 0.8211 loss_bbox: 1.1022 d0.loss_cls: 0.7751 d0.loss_bbox: 1.2010 d1.loss_cls: 0.7885 d1.loss_bbox: 1.1292 d2.loss_cls: 0.7891 d2.loss_bbox: 1.1079 d3.loss_cls: 0.8233 d3.loss_bbox: 1.0926 d4.loss_cls: 0.8220 d4.loss_bbox: 1.0997 +10/10 19:01:52 - mmengine - INFO - Epoch(train) [1][6550/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:36:48 time: 0.6779 data_time: 0.0404 memory: 11853 grad_norm: 41.1180 loss: 10.8410 loss_cls: 0.7637 loss_bbox: 1.0415 d0.loss_cls: 0.7160 d0.loss_bbox: 1.1466 d1.loss_cls: 0.7316 d1.loss_bbox: 1.0603 d2.loss_cls: 0.7369 d2.loss_bbox: 1.0471 d3.loss_cls: 0.7494 d3.loss_bbox: 1.0459 d4.loss_cls: 0.7545 d4.loss_bbox: 1.0474 +10/10 19:02:26 - mmengine - INFO - Epoch(train) [1][6600/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:33:50 time: 0.6769 data_time: 0.0399 memory: 11853 grad_norm: 44.8439 loss: 8.6414 loss_cls: 0.6183 loss_bbox: 0.8255 d0.loss_cls: 0.5698 d0.loss_bbox: 0.9004 d1.loss_cls: 0.5812 d1.loss_bbox: 0.8443 d2.loss_cls: 0.5930 d2.loss_bbox: 0.8334 d3.loss_cls: 0.6107 d3.loss_bbox: 0.8288 d4.loss_cls: 0.6088 d4.loss_bbox: 0.8272 +10/10 19:03:00 - mmengine - INFO - Epoch(train) [1][6650/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:31:01 time: 0.6834 data_time: 0.0405 memory: 11853 grad_norm: 48.8278 loss: 10.4478 loss_cls: 0.7381 loss_bbox: 0.9873 d0.loss_cls: 0.7124 d0.loss_bbox: 1.0997 d1.loss_cls: 0.7158 d1.loss_bbox: 1.0258 d2.loss_cls: 0.7214 d2.loss_bbox: 1.0082 d3.loss_cls: 0.7181 d3.loss_bbox: 0.9997 d4.loss_cls: 0.7329 d4.loss_bbox: 0.9884 +10/10 19:03:34 - mmengine - INFO - Epoch(train) [1][6700/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:28:06 time: 0.6767 data_time: 0.0397 memory: 11853 grad_norm: 45.8190 loss: 10.8954 loss_cls: 0.7935 loss_bbox: 1.0204 d0.loss_cls: 0.7509 d0.loss_bbox: 1.1211 d1.loss_cls: 0.7511 d1.loss_bbox: 1.0435 d2.loss_cls: 0.7721 d2.loss_bbox: 1.0334 d3.loss_cls: 0.7903 d3.loss_bbox: 1.0125 d4.loss_cls: 0.7864 d4.loss_bbox: 1.0204 +10/10 19:04:08 - mmengine - INFO - Epoch(train) [1][6750/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:25:15 time: 0.6781 data_time: 0.0406 memory: 11853 grad_norm: 41.6473 loss: 9.6763 loss_cls: 0.6619 loss_bbox: 0.9531 d0.loss_cls: 0.6286 d0.loss_bbox: 1.0244 d1.loss_cls: 0.6405 d1.loss_bbox: 0.9630 d2.loss_cls: 0.6394 d2.loss_bbox: 0.9528 d3.loss_cls: 0.6543 d3.loss_bbox: 0.9481 d4.loss_cls: 0.6538 d4.loss_bbox: 0.9565 +10/10 19:04:42 - mmengine - INFO - Epoch(train) [1][6800/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:22:26 time: 0.6779 data_time: 0.0401 memory: 11853 grad_norm: 45.7883 loss: 11.1531 loss_cls: 0.7942 loss_bbox: 1.0558 d0.loss_cls: 0.7607 d0.loss_bbox: 1.1547 d1.loss_cls: 0.7702 d1.loss_bbox: 1.0902 d2.loss_cls: 0.7731 d2.loss_bbox: 1.0638 d3.loss_cls: 0.7963 d3.loss_bbox: 1.0425 d4.loss_cls: 0.8006 d4.loss_bbox: 1.0512 +10/10 19:05:16 - mmengine - INFO - Epoch(train) [1][6850/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:19:44 time: 0.6825 data_time: 0.0397 memory: 11853 grad_norm: 46.5460 loss: 10.6727 loss_cls: 0.7570 loss_bbox: 1.0233 d0.loss_cls: 0.7035 d0.loss_bbox: 1.1297 d1.loss_cls: 0.7204 d1.loss_bbox: 1.0610 d2.loss_cls: 0.7423 d2.loss_bbox: 1.0188 d3.loss_cls: 0.7471 d3.loss_bbox: 1.0098 d4.loss_cls: 0.7475 d4.loss_bbox: 1.0123 +10/10 19:05:50 - mmengine - INFO - Epoch(train) [1][6900/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:16:56 time: 0.6753 data_time: 0.0394 memory: 11853 grad_norm: 38.7100 loss: 9.8178 loss_cls: 0.6945 loss_bbox: 0.9326 d0.loss_cls: 0.6461 d0.loss_bbox: 1.0334 d1.loss_cls: 0.6602 d1.loss_bbox: 0.9723 d2.loss_cls: 0.6802 d2.loss_bbox: 0.9399 d3.loss_cls: 0.6939 d3.loss_bbox: 0.9355 d4.loss_cls: 0.6982 d4.loss_bbox: 0.9309 +10/10 19:06:24 - mmengine - INFO - Epoch(train) [1][6950/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:14:12 time: 0.6778 data_time: 0.0401 memory: 11853 grad_norm: 39.4656 loss: 9.4844 loss_cls: 0.6765 loss_bbox: 0.9014 d0.loss_cls: 0.6394 d0.loss_bbox: 0.9860 d1.loss_cls: 0.6557 d1.loss_bbox: 0.9189 d2.loss_cls: 0.6661 d2.loss_bbox: 0.8995 d3.loss_cls: 0.6685 d3.loss_bbox: 0.9013 d4.loss_cls: 0.6734 d4.loss_bbox: 0.8976 +10/10 19:06:58 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 19:06:58 - mmengine - INFO - Epoch(train) [1][7000/7033] base_lr: 2.0000e-04 lr: 2.0000e-04 eta: 1 day, 11:11:31 time: 0.6784 data_time: 0.0400 memory: 11853 grad_norm: 38.0881 loss: 9.7571 loss_cls: 0.6907 loss_bbox: 0.9388 d0.loss_cls: 0.6473 d0.loss_bbox: 1.0256 d1.loss_cls: 0.6499 d1.loss_bbox: 0.9664 d2.loss_cls: 0.6629 d2.loss_bbox: 0.9399 d3.loss_cls: 0.6719 d3.loss_bbox: 0.9403 d4.loss_cls: 0.6790 d4.loss_bbox: 0.9445 +10/10 19:07:20 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 19:07:20 - mmengine - INFO - Saving checkpoint at 1 epochs +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmcv/cnn/bricks/transformer.py:524: UserWarning: position encoding of key ismissing in MultiheadAttention. + warnings.warn(f'position encoding of key is' +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/functional.py:3981: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details. + warnings.warn( +10/10 19:08:07 - mmengine - INFO - Epoch(train) [2][ 50/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 11:07:31 time: 0.6988 data_time: 0.0597 memory: 11853 grad_norm: 40.5222 loss: 10.7513 loss_cls: 0.7750 loss_bbox: 1.0039 d0.loss_cls: 0.7215 d0.loss_bbox: 1.1261 d1.loss_cls: 0.7383 d1.loss_bbox: 1.0468 d2.loss_cls: 0.7642 d2.loss_bbox: 1.0157 d3.loss_cls: 0.7673 d3.loss_bbox: 1.0141 d4.loss_cls: 0.7687 d4.loss_bbox: 1.0095 +10/10 19:08:41 - mmengine - INFO - Epoch(train) [2][ 100/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 11:04:50 time: 0.6746 data_time: 0.0396 memory: 11853 grad_norm: 43.9808 loss: 10.6537 loss_cls: 0.7478 loss_bbox: 1.0440 d0.loss_cls: 0.6789 d0.loss_bbox: 1.1441 d1.loss_cls: 0.6939 d1.loss_bbox: 1.0680 d2.loss_cls: 0.7209 d2.loss_bbox: 1.0375 d3.loss_cls: 0.7210 d3.loss_bbox: 1.0316 d4.loss_cls: 0.7296 d4.loss_bbox: 1.0362 +10/10 19:09:15 - mmengine - INFO - Epoch(train) [2][ 150/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 11:02:14 time: 0.6773 data_time: 0.0399 memory: 11853 grad_norm: 44.0175 loss: 9.6346 loss_cls: 0.6882 loss_bbox: 0.9020 d0.loss_cls: 0.6366 d0.loss_bbox: 1.0182 d1.loss_cls: 0.6546 d1.loss_bbox: 0.9556 d2.loss_cls: 0.6582 d2.loss_bbox: 0.9356 d3.loss_cls: 0.6736 d3.loss_bbox: 0.9202 d4.loss_cls: 0.6803 d4.loss_bbox: 0.9115 +10/10 19:09:49 - mmengine - INFO - Epoch(train) [2][ 200/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:59:39 time: 0.6765 data_time: 0.0395 memory: 11853 grad_norm: 40.7278 loss: 10.1356 loss_cls: 0.7152 loss_bbox: 0.9616 d0.loss_cls: 0.6628 d0.loss_bbox: 1.0949 d1.loss_cls: 0.6732 d1.loss_bbox: 0.9995 d2.loss_cls: 0.6925 d2.loss_bbox: 0.9804 d3.loss_cls: 0.7078 d3.loss_bbox: 0.9746 d4.loss_cls: 0.7050 d4.loss_bbox: 0.9680 +10/10 19:10:22 - mmengine - INFO - Epoch(train) [2][ 250/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:57:04 time: 0.6756 data_time: 0.0399 memory: 11853 grad_norm: 44.3055 loss: 10.3448 loss_cls: 0.7605 loss_bbox: 0.9799 d0.loss_cls: 0.6906 d0.loss_bbox: 1.0706 d1.loss_cls: 0.7162 d1.loss_bbox: 0.9964 d2.loss_cls: 0.7324 d2.loss_bbox: 0.9645 d3.loss_cls: 0.7418 d3.loss_bbox: 0.9708 d4.loss_cls: 0.7440 d4.loss_bbox: 0.9772 +10/10 19:10:56 - mmengine - INFO - Epoch(train) [2][ 300/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:54:33 time: 0.6771 data_time: 0.0401 memory: 11853 grad_norm: 40.8524 loss: 10.0135 loss_cls: 0.7103 loss_bbox: 0.9541 d0.loss_cls: 0.6570 d0.loss_bbox: 1.0692 d1.loss_cls: 0.6776 d1.loss_bbox: 0.9880 d2.loss_cls: 0.6849 d2.loss_bbox: 0.9640 d3.loss_cls: 0.6956 d3.loss_bbox: 0.9589 d4.loss_cls: 0.7011 d4.loss_bbox: 0.9527 +10/10 19:11:30 - mmengine - INFO - Epoch(train) [2][ 350/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:52:01 time: 0.6748 data_time: 0.0395 memory: 11853 grad_norm: 44.7415 loss: 11.0489 loss_cls: 0.7916 loss_bbox: 1.0346 d0.loss_cls: 0.7364 d0.loss_bbox: 1.1790 d1.loss_cls: 0.7681 d1.loss_bbox: 1.0777 d2.loss_cls: 0.7627 d2.loss_bbox: 1.0608 d3.loss_cls: 0.7779 d3.loss_bbox: 1.0459 d4.loss_cls: 0.7707 d4.loss_bbox: 1.0433 +10/10 19:12:04 - mmengine - INFO - Epoch(train) [2][ 400/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:49:31 time: 0.6752 data_time: 0.0400 memory: 11853 grad_norm: 46.2994 loss: 8.5797 loss_cls: 0.6113 loss_bbox: 0.8129 d0.loss_cls: 0.5611 d0.loss_bbox: 0.9284 d1.loss_cls: 0.5649 d1.loss_bbox: 0.8479 d2.loss_cls: 0.5794 d2.loss_bbox: 0.8318 d3.loss_cls: 0.5970 d3.loss_bbox: 0.8239 d4.loss_cls: 0.6020 d4.loss_bbox: 0.8191 +10/10 19:12:38 - mmengine - INFO - Epoch(train) [2][ 450/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:47:02 time: 0.6753 data_time: 0.0394 memory: 11853 grad_norm: 56.0374 loss: 10.1724 loss_cls: 0.7408 loss_bbox: 0.9332 d0.loss_cls: 0.7073 d0.loss_bbox: 1.0560 d1.loss_cls: 0.7071 d1.loss_bbox: 0.9851 d2.loss_cls: 0.7212 d2.loss_bbox: 0.9522 d3.loss_cls: 0.7379 d3.loss_bbox: 0.9500 d4.loss_cls: 0.7419 d4.loss_bbox: 0.9398 +10/10 19:13:11 - mmengine - INFO - Epoch(train) [2][ 500/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:44:36 time: 0.6753 data_time: 0.0391 memory: 11853 grad_norm: 44.6416 loss: 9.4125 loss_cls: 0.6791 loss_bbox: 0.8754 d0.loss_cls: 0.6591 d0.loss_bbox: 0.9828 d1.loss_cls: 0.6550 d1.loss_bbox: 0.8967 d2.loss_cls: 0.6628 d2.loss_bbox: 0.8911 d3.loss_cls: 0.6682 d3.loss_bbox: 0.8866 d4.loss_cls: 0.6730 d4.loss_bbox: 0.8827 +10/10 19:13:45 - mmengine - INFO - Epoch(train) [2][ 550/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:42:17 time: 0.6822 data_time: 0.0400 memory: 11853 grad_norm: 43.2620 loss: 9.2134 loss_cls: 0.6424 loss_bbox: 0.8783 d0.loss_cls: 0.6010 d0.loss_bbox: 0.9898 d1.loss_cls: 0.6121 d1.loss_bbox: 0.9184 d2.loss_cls: 0.6239 d2.loss_bbox: 0.8935 d3.loss_cls: 0.6389 d3.loss_bbox: 0.8893 d4.loss_cls: 0.6404 d4.loss_bbox: 0.8855 +10/10 19:14:19 - mmengine - INFO - Epoch(train) [2][ 600/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:39:59 time: 0.6807 data_time: 0.0397 memory: 11853 grad_norm: 51.2793 loss: 10.7815 loss_cls: 0.7618 loss_bbox: 1.0435 d0.loss_cls: 0.7151 d0.loss_bbox: 1.1396 d1.loss_cls: 0.7201 d1.loss_bbox: 1.0638 d2.loss_cls: 0.7389 d2.loss_bbox: 1.0365 d3.loss_cls: 0.7477 d3.loss_bbox: 1.0347 d4.loss_cls: 0.7519 d4.loss_bbox: 1.0280 +10/10 19:14:53 - mmengine - INFO - Epoch(train) [2][ 650/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:37:36 time: 0.6751 data_time: 0.0398 memory: 11853 grad_norm: 60.3489 loss: 9.4697 loss_cls: 0.6873 loss_bbox: 0.9049 d0.loss_cls: 0.6414 d0.loss_bbox: 0.9734 d1.loss_cls: 0.6503 d1.loss_bbox: 0.9095 d2.loss_cls: 0.6640 d2.loss_bbox: 0.8910 d3.loss_cls: 0.6707 d3.loss_bbox: 0.8999 d4.loss_cls: 0.6778 d4.loss_bbox: 0.8995 +10/10 19:15:27 - mmengine - INFO - Epoch(train) [2][ 700/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:35:17 time: 0.6770 data_time: 0.0394 memory: 11853 grad_norm: 44.0860 loss: 10.7759 loss_cls: 0.7595 loss_bbox: 1.0329 d0.loss_cls: 0.7237 d0.loss_bbox: 1.1317 d1.loss_cls: 0.7302 d1.loss_bbox: 1.0575 d2.loss_cls: 0.7345 d2.loss_bbox: 1.0453 d3.loss_cls: 0.7439 d3.loss_bbox: 1.0378 d4.loss_cls: 0.7506 d4.loss_bbox: 1.0283 +10/10 19:16:01 - mmengine - INFO - Epoch(train) [2][ 750/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:32:56 time: 0.6743 data_time: 0.0397 memory: 11853 grad_norm: 45.2950 loss: 10.3214 loss_cls: 0.7501 loss_bbox: 0.9735 d0.loss_cls: 0.6968 d0.loss_bbox: 1.0710 d1.loss_cls: 0.7032 d1.loss_bbox: 1.0027 d2.loss_cls: 0.7210 d2.loss_bbox: 0.9770 d3.loss_cls: 0.7295 d3.loss_bbox: 0.9767 d4.loss_cls: 0.7413 d4.loss_bbox: 0.9787 +10/10 19:16:35 - mmengine - INFO - Epoch(train) [2][ 800/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:30:39 time: 0.6764 data_time: 0.0401 memory: 11853 grad_norm: 43.8285 loss: 8.9133 loss_cls: 0.6479 loss_bbox: 0.8371 d0.loss_cls: 0.6045 d0.loss_bbox: 0.9259 d1.loss_cls: 0.6173 d1.loss_bbox: 0.8600 d2.loss_cls: 0.6263 d2.loss_bbox: 0.8445 d3.loss_cls: 0.6325 d3.loss_bbox: 0.8419 d4.loss_cls: 0.6369 d4.loss_bbox: 0.8385 +10/10 19:17:08 - mmengine - INFO - Epoch(train) [2][ 850/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:28:23 time: 0.6766 data_time: 0.0398 memory: 11853 grad_norm: 44.5838 loss: 10.4436 loss_cls: 0.7359 loss_bbox: 1.0007 d0.loss_cls: 0.7133 d0.loss_bbox: 1.0777 d1.loss_cls: 0.7099 d1.loss_bbox: 1.0218 d2.loss_cls: 0.7144 d2.loss_bbox: 1.0059 d3.loss_cls: 0.7298 d3.loss_bbox: 1.0016 d4.loss_cls: 0.7326 d4.loss_bbox: 1.0000 +10/10 19:17:42 - mmengine - INFO - Epoch(train) [2][ 900/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:26:07 time: 0.6756 data_time: 0.0395 memory: 11853 grad_norm: 51.4484 loss: 9.6341 loss_cls: 0.6856 loss_bbox: 0.9179 d0.loss_cls: 0.6227 d0.loss_bbox: 1.0389 d1.loss_cls: 0.6350 d1.loss_bbox: 0.9578 d2.loss_cls: 0.6384 d2.loss_bbox: 0.9460 d3.loss_cls: 0.6620 d3.loss_bbox: 0.9281 d4.loss_cls: 0.6730 d4.loss_bbox: 0.9287 +10/10 19:18:16 - mmengine - INFO - Epoch(train) [2][ 950/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:23:52 time: 0.6751 data_time: 0.0397 memory: 11853 grad_norm: 57.7526 loss: 9.3493 loss_cls: 0.6745 loss_bbox: 0.8748 d0.loss_cls: 0.6314 d0.loss_bbox: 0.9896 d1.loss_cls: 0.6351 d1.loss_bbox: 0.9217 d2.loss_cls: 0.6527 d2.loss_bbox: 0.8877 d3.loss_cls: 0.6579 d3.loss_bbox: 0.8824 d4.loss_cls: 0.6667 d4.loss_bbox: 0.8747 +10/10 19:18:27 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 19:18:50 - mmengine - INFO - Epoch(train) [2][1000/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:21:40 time: 0.6765 data_time: 0.0398 memory: 11853 grad_norm: 43.1115 loss: 9.6126 loss_cls: 0.6698 loss_bbox: 0.9162 d0.loss_cls: 0.6299 d0.loss_bbox: 1.0216 d1.loss_cls: 0.6446 d1.loss_bbox: 0.9568 d2.loss_cls: 0.6533 d2.loss_bbox: 0.9427 d3.loss_cls: 0.6623 d3.loss_bbox: 0.9285 d4.loss_cls: 0.6617 d4.loss_bbox: 0.9252 +10/10 19:19:24 - mmengine - INFO - Epoch(train) [2][1050/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:19:30 time: 0.6766 data_time: 0.0397 memory: 11853 grad_norm: 45.6404 loss: 8.5994 loss_cls: 0.6206 loss_bbox: 0.8013 d0.loss_cls: 0.5710 d0.loss_bbox: 0.9189 d1.loss_cls: 0.5885 d1.loss_bbox: 0.8441 d2.loss_cls: 0.5939 d2.loss_bbox: 0.8275 d3.loss_cls: 0.6010 d3.loss_bbox: 0.8119 d4.loss_cls: 0.6168 d4.loss_bbox: 0.8041 +10/10 19:19:58 - mmengine - INFO - Epoch(train) [2][1100/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:17:23 time: 0.6793 data_time: 0.0397 memory: 11853 grad_norm: 45.0233 loss: 9.3046 loss_cls: 0.6602 loss_bbox: 0.8804 d0.loss_cls: 0.6228 d0.loss_bbox: 0.9901 d1.loss_cls: 0.6176 d1.loss_bbox: 0.9244 d2.loss_cls: 0.6267 d2.loss_bbox: 0.9150 d3.loss_cls: 0.6425 d3.loss_bbox: 0.8914 d4.loss_cls: 0.6545 d4.loss_bbox: 0.8789 +10/10 19:20:32 - mmengine - INFO - Epoch(train) [2][1150/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:15:19 time: 0.6819 data_time: 0.0393 memory: 11853 grad_norm: 47.6138 loss: 8.3709 loss_cls: 0.6117 loss_bbox: 0.7676 d0.loss_cls: 0.5533 d0.loss_bbox: 0.8953 d1.loss_cls: 0.5674 d1.loss_bbox: 0.8246 d2.loss_cls: 0.5770 d2.loss_bbox: 0.8142 d3.loss_cls: 0.5899 d3.loss_bbox: 0.7933 d4.loss_cls: 0.6022 d4.loss_bbox: 0.7741 +10/10 19:21:05 - mmengine - INFO - Epoch(train) [2][1200/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:13:12 time: 0.6764 data_time: 0.0395 memory: 11853 grad_norm: 43.7787 loss: 9.7587 loss_cls: 0.7062 loss_bbox: 0.9044 d0.loss_cls: 0.6579 d0.loss_bbox: 1.0234 d1.loss_cls: 0.6634 d1.loss_bbox: 0.9525 d2.loss_cls: 0.6781 d2.loss_bbox: 0.9326 d3.loss_cls: 0.6911 d3.loss_bbox: 0.9298 d4.loss_cls: 0.7015 d4.loss_bbox: 0.9178 +10/10 19:21:39 - mmengine - INFO - Epoch(train) [2][1250/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:11:05 time: 0.6763 data_time: 0.0397 memory: 11853 grad_norm: 41.7787 loss: 9.6164 loss_cls: 0.6837 loss_bbox: 0.9243 d0.loss_cls: 0.6348 d0.loss_bbox: 1.0167 d1.loss_cls: 0.6325 d1.loss_bbox: 0.9454 d2.loss_cls: 0.6439 d2.loss_bbox: 0.9397 d3.loss_cls: 0.6656 d3.loss_bbox: 0.9287 d4.loss_cls: 0.6713 d4.loss_bbox: 0.9298 +10/10 19:22:13 - mmengine - INFO - Epoch(train) [2][1300/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:09:02 time: 0.6781 data_time: 0.0398 memory: 11853 grad_norm: 49.2389 loss: 11.0542 loss_cls: 0.8037 loss_bbox: 1.0245 d0.loss_cls: 0.7631 d0.loss_bbox: 1.1477 d1.loss_cls: 0.7676 d1.loss_bbox: 1.0750 d2.loss_cls: 0.7775 d2.loss_bbox: 1.0531 d3.loss_cls: 0.7908 d3.loss_bbox: 1.0306 d4.loss_cls: 0.7922 d4.loss_bbox: 1.0284 +10/10 19:22:47 - mmengine - INFO - Epoch(train) [2][1350/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:07:00 time: 0.6793 data_time: 0.0400 memory: 11853 grad_norm: 59.1241 loss: 9.6855 loss_cls: 0.6857 loss_bbox: 0.9215 d0.loss_cls: 0.6500 d0.loss_bbox: 1.0256 d1.loss_cls: 0.6543 d1.loss_bbox: 0.9512 d2.loss_cls: 0.6675 d2.loss_bbox: 0.9242 d3.loss_cls: 0.6712 d3.loss_bbox: 0.9309 d4.loss_cls: 0.6719 d4.loss_bbox: 0.9317 +10/10 19:23:21 - mmengine - INFO - Epoch(train) [2][1400/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:04:55 time: 0.6747 data_time: 0.0398 memory: 11853 grad_norm: 41.2073 loss: 8.7039 loss_cls: 0.6123 loss_bbox: 0.8287 d0.loss_cls: 0.5928 d0.loss_bbox: 0.9210 d1.loss_cls: 0.5980 d1.loss_bbox: 0.8458 d2.loss_cls: 0.5998 d2.loss_bbox: 0.8336 d3.loss_cls: 0.6031 d3.loss_bbox: 0.8337 d4.loss_cls: 0.6084 d4.loss_bbox: 0.8268 +10/10 19:23:55 - mmengine - INFO - Epoch(train) [2][1450/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:02:54 time: 0.6775 data_time: 0.0397 memory: 11853 grad_norm: 40.3473 loss: 9.4632 loss_cls: 0.6768 loss_bbox: 0.8884 d0.loss_cls: 0.6436 d0.loss_bbox: 0.9931 d1.loss_cls: 0.6542 d1.loss_bbox: 0.9175 d2.loss_cls: 0.6610 d2.loss_bbox: 0.9050 d3.loss_cls: 0.6695 d3.loss_bbox: 0.8951 d4.loss_cls: 0.6666 d4.loss_bbox: 0.8924 +10/10 19:24:29 - mmengine - INFO - Epoch(train) [2][1500/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 10:00:54 time: 0.6771 data_time: 0.0399 memory: 11853 grad_norm: 50.4027 loss: 10.1091 loss_cls: 0.7406 loss_bbox: 0.9535 d0.loss_cls: 0.7005 d0.loss_bbox: 1.0397 d1.loss_cls: 0.6879 d1.loss_bbox: 0.9672 d2.loss_cls: 0.7002 d2.loss_bbox: 0.9584 d3.loss_cls: 0.7194 d3.loss_bbox: 0.9637 d4.loss_cls: 0.7269 d4.loss_bbox: 0.9511 +10/10 19:25:03 - mmengine - INFO - Epoch(train) [2][1550/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:58:56 time: 0.6788 data_time: 0.0395 memory: 11853 grad_norm: 47.0508 loss: 10.3652 loss_cls: 0.7526 loss_bbox: 0.9723 d0.loss_cls: 0.7053 d0.loss_bbox: 1.0888 d1.loss_cls: 0.7220 d1.loss_bbox: 1.0042 d2.loss_cls: 0.7221 d2.loss_bbox: 0.9831 d3.loss_cls: 0.7238 d3.loss_bbox: 0.9798 d4.loss_cls: 0.7311 d4.loss_bbox: 0.9800 +10/10 19:25:36 - mmengine - INFO - Epoch(train) [2][1600/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:56:57 time: 0.6761 data_time: 0.0399 memory: 11853 grad_norm: 47.5871 loss: 9.0330 loss_cls: 0.6558 loss_bbox: 0.8479 d0.loss_cls: 0.5860 d0.loss_bbox: 0.9669 d1.loss_cls: 0.6117 d1.loss_bbox: 0.8835 d2.loss_cls: 0.6216 d2.loss_bbox: 0.8621 d3.loss_cls: 0.6318 d3.loss_bbox: 0.8665 d4.loss_cls: 0.6404 d4.loss_bbox: 0.8588 +10/10 19:26:10 - mmengine - INFO - Epoch(train) [2][1650/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:55:00 time: 0.6776 data_time: 0.0399 memory: 11853 grad_norm: 44.6310 loss: 9.6804 loss_cls: 0.6935 loss_bbox: 0.9176 d0.loss_cls: 0.6427 d0.loss_bbox: 1.0324 d1.loss_cls: 0.6455 d1.loss_bbox: 0.9554 d2.loss_cls: 0.6623 d2.loss_bbox: 0.9389 d3.loss_cls: 0.6673 d3.loss_bbox: 0.9287 d4.loss_cls: 0.6812 d4.loss_bbox: 0.9148 +10/10 19:26:44 - mmengine - INFO - Epoch(train) [2][1700/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:53:03 time: 0.6764 data_time: 0.0397 memory: 11853 grad_norm: 47.6444 loss: 9.8999 loss_cls: 0.7125 loss_bbox: 0.9237 d0.loss_cls: 0.6664 d0.loss_bbox: 1.0468 d1.loss_cls: 0.6816 d1.loss_bbox: 0.9660 d2.loss_cls: 0.6819 d2.loss_bbox: 0.9514 d3.loss_cls: 0.6979 d3.loss_bbox: 0.9399 d4.loss_cls: 0.6992 d4.loss_bbox: 0.9327 +10/10 19:27:18 - mmengine - INFO - Epoch(train) [2][1750/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:51:07 time: 0.6770 data_time: 0.0400 memory: 11853 grad_norm: 44.0078 loss: 10.1936 loss_cls: 0.7202 loss_bbox: 0.9606 d0.loss_cls: 0.6681 d0.loss_bbox: 1.0895 d1.loss_cls: 0.6876 d1.loss_bbox: 1.0140 d2.loss_cls: 0.6948 d2.loss_bbox: 0.9928 d3.loss_cls: 0.7028 d3.loss_bbox: 0.9830 d4.loss_cls: 0.7139 d4.loss_bbox: 0.9664 +10/10 19:27:52 - mmengine - INFO - Epoch(train) [2][1800/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:49:17 time: 0.6817 data_time: 0.0400 memory: 11853 grad_norm: 43.9600 loss: 9.7307 loss_cls: 0.7070 loss_bbox: 0.8993 d0.loss_cls: 0.6637 d0.loss_bbox: 1.0245 d1.loss_cls: 0.6682 d1.loss_bbox: 0.9455 d2.loss_cls: 0.6850 d2.loss_bbox: 0.9237 d3.loss_cls: 0.6963 d3.loss_bbox: 0.9133 d4.loss_cls: 0.6979 d4.loss_bbox: 0.9064 +10/10 19:28:27 - mmengine - INFO - Epoch(train) [2][1850/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:47:45 time: 0.7013 data_time: 0.0400 memory: 11853 grad_norm: 47.5530 loss: 9.1849 loss_cls: 0.6791 loss_bbox: 0.8517 d0.loss_cls: 0.6376 d0.loss_bbox: 0.9391 d1.loss_cls: 0.6394 d1.loss_bbox: 0.8864 d2.loss_cls: 0.6509 d2.loss_bbox: 0.8627 d3.loss_cls: 0.6549 d3.loss_bbox: 0.8656 d4.loss_cls: 0.6642 d4.loss_bbox: 0.8533 +10/10 19:29:01 - mmengine - INFO - Epoch(train) [2][1900/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:45:56 time: 0.6818 data_time: 0.0396 memory: 11853 grad_norm: 43.6666 loss: 8.7348 loss_cls: 0.6346 loss_bbox: 0.8071 d0.loss_cls: 0.5881 d0.loss_bbox: 0.9208 d1.loss_cls: 0.6086 d1.loss_bbox: 0.8361 d2.loss_cls: 0.6121 d2.loss_bbox: 0.8284 d3.loss_cls: 0.6272 d3.loss_bbox: 0.8197 d4.loss_cls: 0.6338 d4.loss_bbox: 0.8184 +10/10 19:29:35 - mmengine - INFO - Epoch(train) [2][1950/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:44:06 time: 0.6792 data_time: 0.0401 memory: 11853 grad_norm: 46.1774 loss: 9.5993 loss_cls: 0.6857 loss_bbox: 0.9060 d0.loss_cls: 0.6625 d0.loss_bbox: 0.9920 d1.loss_cls: 0.6715 d1.loss_bbox: 0.9318 d2.loss_cls: 0.6759 d2.loss_bbox: 0.9104 d3.loss_cls: 0.6743 d3.loss_bbox: 0.9037 d4.loss_cls: 0.6810 d4.loss_bbox: 0.9045 +10/10 19:29:47 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 19:30:09 - mmengine - INFO - Epoch(train) [2][2000/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:42:15 time: 0.6772 data_time: 0.0399 memory: 11853 grad_norm: 42.0916 loss: 10.0094 loss_cls: 0.7398 loss_bbox: 0.9143 d0.loss_cls: 0.6940 d0.loss_bbox: 1.0396 d1.loss_cls: 0.6995 d1.loss_bbox: 0.9593 d2.loss_cls: 0.7111 d2.loss_bbox: 0.9366 d3.loss_cls: 0.7264 d3.loss_bbox: 0.9304 d4.loss_cls: 0.7356 d4.loss_bbox: 0.9228 +10/10 19:30:43 - mmengine - INFO - Epoch(train) [2][2050/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:40:26 time: 0.6796 data_time: 0.0400 memory: 11853 grad_norm: 43.4554 loss: 9.2566 loss_cls: 0.6877 loss_bbox: 0.8475 d0.loss_cls: 0.6458 d0.loss_bbox: 0.9682 d1.loss_cls: 0.6399 d1.loss_bbox: 0.8871 d2.loss_cls: 0.6531 d2.loss_bbox: 0.8648 d3.loss_cls: 0.6631 d3.loss_bbox: 0.8666 d4.loss_cls: 0.6717 d4.loss_bbox: 0.8610 +10/10 19:31:17 - mmengine - INFO - Epoch(train) [2][2100/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:38:36 time: 0.6763 data_time: 0.0396 memory: 11853 grad_norm: 43.6641 loss: 10.4465 loss_cls: 0.7445 loss_bbox: 0.9758 d0.loss_cls: 0.6973 d0.loss_bbox: 1.1205 d1.loss_cls: 0.7137 d1.loss_bbox: 1.0297 d2.loss_cls: 0.7256 d2.loss_bbox: 0.9965 d3.loss_cls: 0.7323 d3.loss_bbox: 0.9892 d4.loss_cls: 0.7467 d4.loss_bbox: 0.9745 +10/10 19:31:51 - mmengine - INFO - Epoch(train) [2][2150/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:36:46 time: 0.6754 data_time: 0.0399 memory: 11853 grad_norm: 42.8638 loss: 9.6660 loss_cls: 0.6989 loss_bbox: 0.9055 d0.loss_cls: 0.6372 d0.loss_bbox: 1.0405 d1.loss_cls: 0.6573 d1.loss_bbox: 0.9420 d2.loss_cls: 0.6621 d2.loss_bbox: 0.9255 d3.loss_cls: 0.6700 d3.loss_bbox: 0.9241 d4.loss_cls: 0.6836 d4.loss_bbox: 0.9194 +10/10 19:32:24 - mmengine - INFO - Epoch(train) [2][2200/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:34:58 time: 0.6771 data_time: 0.0399 memory: 11853 grad_norm: 47.2792 loss: 9.0044 loss_cls: 0.6731 loss_bbox: 0.8152 d0.loss_cls: 0.6342 d0.loss_bbox: 0.9529 d1.loss_cls: 0.6313 d1.loss_bbox: 0.8626 d2.loss_cls: 0.6372 d2.loss_bbox: 0.8418 d3.loss_cls: 0.6510 d3.loss_bbox: 0.8275 d4.loss_cls: 0.6532 d4.loss_bbox: 0.8246 +10/10 19:32:58 - mmengine - INFO - Epoch(train) [2][2250/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:33:10 time: 0.6765 data_time: 0.0401 memory: 11853 grad_norm: 83.3386 loss: 9.1070 loss_cls: 0.6603 loss_bbox: 0.8562 d0.loss_cls: 0.6296 d0.loss_bbox: 0.9629 d1.loss_cls: 0.6263 d1.loss_bbox: 0.8786 d2.loss_cls: 0.6454 d2.loss_bbox: 0.8417 d3.loss_cls: 0.6572 d3.loss_bbox: 0.8453 d4.loss_cls: 0.6535 d4.loss_bbox: 0.8500 +10/10 19:33:32 - mmengine - INFO - Epoch(train) [2][2300/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:31:22 time: 0.6751 data_time: 0.0398 memory: 11853 grad_norm: 49.4259 loss: 9.1755 loss_cls: 0.6680 loss_bbox: 0.8561 d0.loss_cls: 0.6224 d0.loss_bbox: 0.9561 d1.loss_cls: 0.6411 d1.loss_bbox: 0.8783 d2.loss_cls: 0.6549 d2.loss_bbox: 0.8600 d3.loss_cls: 0.6665 d3.loss_bbox: 0.8533 d4.loss_cls: 0.6641 d4.loss_bbox: 0.8546 +10/10 19:34:06 - mmengine - INFO - Epoch(train) [2][2350/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:29:38 time: 0.6782 data_time: 0.0396 memory: 11853 grad_norm: 53.7834 loss: 8.9311 loss_cls: 0.6501 loss_bbox: 0.8235 d0.loss_cls: 0.5842 d0.loss_bbox: 0.9501 d1.loss_cls: 0.6003 d1.loss_bbox: 0.8736 d2.loss_cls: 0.6148 d2.loss_bbox: 0.8579 d3.loss_cls: 0.6361 d3.loss_bbox: 0.8526 d4.loss_cls: 0.6486 d4.loss_bbox: 0.8392 +10/10 19:34:40 - mmengine - INFO - Epoch(train) [2][2400/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:27:53 time: 0.6776 data_time: 0.0396 memory: 11853 grad_norm: 46.9390 loss: 9.8067 loss_cls: 0.6916 loss_bbox: 0.9347 d0.loss_cls: 0.6311 d0.loss_bbox: 1.0421 d1.loss_cls: 0.6487 d1.loss_bbox: 0.9703 d2.loss_cls: 0.6783 d2.loss_bbox: 0.9538 d3.loss_cls: 0.6840 d3.loss_bbox: 0.9450 d4.loss_cls: 0.6871 d4.loss_bbox: 0.9400 +10/10 19:35:14 - mmengine - INFO - Epoch(train) [2][2450/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:26:08 time: 0.6760 data_time: 0.0395 memory: 11853 grad_norm: 47.6992 loss: 9.3549 loss_cls: 0.6823 loss_bbox: 0.8713 d0.loss_cls: 0.6317 d0.loss_bbox: 0.9774 d1.loss_cls: 0.6356 d1.loss_bbox: 0.9166 d2.loss_cls: 0.6552 d2.loss_bbox: 0.8897 d3.loss_cls: 0.6632 d3.loss_bbox: 0.8785 d4.loss_cls: 0.6852 d4.loss_bbox: 0.8681 +10/10 19:35:47 - mmengine - INFO - Epoch(train) [2][2500/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:24:24 time: 0.6760 data_time: 0.0395 memory: 11853 grad_norm: 48.6315 loss: 10.2126 loss_cls: 0.7776 loss_bbox: 0.9279 d0.loss_cls: 0.7024 d0.loss_bbox: 1.0615 d1.loss_cls: 0.7219 d1.loss_bbox: 0.9725 d2.loss_cls: 0.7308 d2.loss_bbox: 0.9436 d3.loss_cls: 0.7432 d3.loss_bbox: 0.9404 d4.loss_cls: 0.7631 d4.loss_bbox: 0.9278 +10/10 19:36:21 - mmengine - INFO - Epoch(train) [2][2550/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:22:42 time: 0.6776 data_time: 0.0397 memory: 11853 grad_norm: 55.1276 loss: 8.3730 loss_cls: 0.6269 loss_bbox: 0.7560 d0.loss_cls: 0.5795 d0.loss_bbox: 0.8788 d1.loss_cls: 0.5990 d1.loss_bbox: 0.7940 d2.loss_cls: 0.6057 d2.loss_bbox: 0.7737 d3.loss_cls: 0.6155 d3.loss_bbox: 0.7662 d4.loss_cls: 0.6241 d4.loss_bbox: 0.7537 +10/10 19:36:55 - mmengine - INFO - Epoch(train) [2][2600/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:21:01 time: 0.6788 data_time: 0.0409 memory: 11853 grad_norm: 46.1529 loss: 9.2960 loss_cls: 0.6750 loss_bbox: 0.8539 d0.loss_cls: 0.6369 d0.loss_bbox: 0.9746 d1.loss_cls: 0.6523 d1.loss_bbox: 0.8895 d2.loss_cls: 0.6685 d2.loss_bbox: 0.8714 d3.loss_cls: 0.6722 d3.loss_bbox: 0.8655 d4.loss_cls: 0.6800 d4.loss_bbox: 0.8563 +10/10 19:37:29 - mmengine - INFO - Epoch(train) [2][2650/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:19:23 time: 0.6804 data_time: 0.0394 memory: 11853 grad_norm: 50.1537 loss: 8.8159 loss_cls: 0.6611 loss_bbox: 0.7964 d0.loss_cls: 0.6140 d0.loss_bbox: 0.9310 d1.loss_cls: 0.6196 d1.loss_bbox: 0.8412 d2.loss_cls: 0.6271 d2.loss_bbox: 0.8227 d3.loss_cls: 0.6458 d3.loss_bbox: 0.8085 d4.loss_cls: 0.6541 d4.loss_bbox: 0.7946 +10/10 19:38:03 - mmengine - INFO - Epoch(train) [2][2700/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:17:41 time: 0.6757 data_time: 0.0403 memory: 11853 grad_norm: 42.8227 loss: 9.4944 loss_cls: 0.6938 loss_bbox: 0.8877 d0.loss_cls: 0.6368 d0.loss_bbox: 0.9912 d1.loss_cls: 0.6541 d1.loss_bbox: 0.9158 d2.loss_cls: 0.6678 d2.loss_bbox: 0.9014 d3.loss_cls: 0.6799 d3.loss_bbox: 0.8935 d4.loss_cls: 0.6827 d4.loss_bbox: 0.8897 +10/10 19:38:37 - mmengine - INFO - Epoch(train) [2][2750/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:16:02 time: 0.6778 data_time: 0.0402 memory: 11853 grad_norm: 55.0298 loss: 10.8506 loss_cls: 0.7933 loss_bbox: 1.0015 d0.loss_cls: 0.7308 d0.loss_bbox: 1.1482 d1.loss_cls: 0.7465 d1.loss_bbox: 1.0614 d2.loss_cls: 0.7623 d2.loss_bbox: 1.0264 d3.loss_cls: 0.7792 d3.loss_bbox: 1.0164 d4.loss_cls: 0.7820 d4.loss_bbox: 1.0027 +10/10 19:39:11 - mmengine - INFO - Epoch(train) [2][2800/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:14:22 time: 0.6759 data_time: 0.0397 memory: 11853 grad_norm: 66.8938 loss: 10.4382 loss_cls: 0.7578 loss_bbox: 0.9679 d0.loss_cls: 0.7159 d0.loss_bbox: 1.1079 d1.loss_cls: 0.7230 d1.loss_bbox: 1.0076 d2.loss_cls: 0.7299 d2.loss_bbox: 0.9890 d3.loss_cls: 0.7450 d3.loss_bbox: 0.9770 d4.loss_cls: 0.7503 d4.loss_bbox: 0.9669 +10/10 19:39:45 - mmengine - INFO - Epoch(train) [2][2850/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:12:44 time: 0.6774 data_time: 0.0400 memory: 11853 grad_norm: 44.8791 loss: 11.2420 loss_cls: 0.8052 loss_bbox: 1.0623 d0.loss_cls: 0.7357 d0.loss_bbox: 1.1888 d1.loss_cls: 0.7672 d1.loss_bbox: 1.0982 d2.loss_cls: 0.7740 d2.loss_bbox: 1.0845 d3.loss_cls: 0.7859 d3.loss_bbox: 1.0833 d4.loss_cls: 0.7970 d4.loss_bbox: 1.0598 +10/10 19:40:19 - mmengine - INFO - Epoch(train) [2][2900/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:11:06 time: 0.6774 data_time: 0.0398 memory: 11853 grad_norm: 46.4196 loss: 9.8867 loss_cls: 0.7344 loss_bbox: 0.8838 d0.loss_cls: 0.6815 d0.loss_bbox: 1.0438 d1.loss_cls: 0.7078 d1.loss_bbox: 0.9411 d2.loss_cls: 0.7106 d2.loss_bbox: 0.9276 d3.loss_cls: 0.7215 d3.loss_bbox: 0.9120 d4.loss_cls: 0.7251 d4.loss_bbox: 0.8976 +10/10 19:40:52 - mmengine - INFO - Epoch(train) [2][2950/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:09:27 time: 0.6752 data_time: 0.0399 memory: 11853 grad_norm: 39.2180 loss: 10.1171 loss_cls: 0.7227 loss_bbox: 0.9528 d0.loss_cls: 0.6855 d0.loss_bbox: 1.0776 d1.loss_cls: 0.6921 d1.loss_bbox: 0.9881 d2.loss_cls: 0.6877 d2.loss_bbox: 0.9746 d3.loss_cls: 0.7000 d3.loss_bbox: 0.9674 d4.loss_cls: 0.7127 d4.loss_bbox: 0.9560 +10/10 19:41:04 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 19:41:26 - mmengine - INFO - Epoch(train) [2][3000/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:07:51 time: 0.6774 data_time: 0.0399 memory: 11853 grad_norm: 51.2548 loss: 10.6001 loss_cls: 0.7659 loss_bbox: 0.9954 d0.loss_cls: 0.7105 d0.loss_bbox: 1.1115 d1.loss_cls: 0.7195 d1.loss_bbox: 1.0289 d2.loss_cls: 0.7440 d2.loss_bbox: 1.0176 d3.loss_cls: 0.7512 d3.loss_bbox: 1.0056 d4.loss_cls: 0.7526 d4.loss_bbox: 0.9974 +10/10 19:42:00 - mmengine - INFO - Epoch(train) [2][3050/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:06:14 time: 0.6757 data_time: 0.0399 memory: 11853 grad_norm: 46.5530 loss: 8.2639 loss_cls: 0.5933 loss_bbox: 0.7712 d0.loss_cls: 0.5592 d0.loss_bbox: 0.8738 d1.loss_cls: 0.5796 d1.loss_bbox: 0.7958 d2.loss_cls: 0.5793 d2.loss_bbox: 0.7887 d3.loss_cls: 0.5844 d3.loss_bbox: 0.7794 d4.loss_cls: 0.5860 d4.loss_bbox: 0.7732 +10/10 19:42:34 - mmengine - INFO - Epoch(train) [2][3100/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:04:38 time: 0.6765 data_time: 0.0399 memory: 11853 grad_norm: 43.1126 loss: 9.0094 loss_cls: 0.6501 loss_bbox: 0.8494 d0.loss_cls: 0.5970 d0.loss_bbox: 0.9536 d1.loss_cls: 0.6128 d1.loss_bbox: 0.8704 d2.loss_cls: 0.6250 d2.loss_bbox: 0.8552 d3.loss_cls: 0.6433 d3.loss_bbox: 0.8549 d4.loss_cls: 0.6438 d4.loss_bbox: 0.8540 +10/10 19:43:08 - mmengine - INFO - Epoch(train) [2][3150/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:03:05 time: 0.6786 data_time: 0.0404 memory: 11853 grad_norm: 45.6876 loss: 10.3212 loss_cls: 0.7254 loss_bbox: 0.9826 d0.loss_cls: 0.6742 d0.loss_bbox: 1.0986 d1.loss_cls: 0.6834 d1.loss_bbox: 1.0299 d2.loss_cls: 0.7039 d2.loss_bbox: 1.0062 d3.loss_cls: 0.7146 d3.loss_bbox: 0.9973 d4.loss_cls: 0.7266 d4.loss_bbox: 0.9784 +10/10 19:43:42 - mmengine - INFO - Epoch(train) [2][3200/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 9:01:31 time: 0.6775 data_time: 0.0402 memory: 11853 grad_norm: 46.9115 loss: 10.3922 loss_cls: 0.7756 loss_bbox: 0.9455 d0.loss_cls: 0.7165 d0.loss_bbox: 1.0791 d1.loss_cls: 0.7440 d1.loss_bbox: 0.9825 d2.loss_cls: 0.7568 d2.loss_bbox: 0.9596 d3.loss_cls: 0.7634 d3.loss_bbox: 0.9533 d4.loss_cls: 0.7709 d4.loss_bbox: 0.9451 +10/10 19:44:15 - mmengine - INFO - Epoch(train) [2][3250/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:59:56 time: 0.6759 data_time: 0.0402 memory: 11853 grad_norm: 45.7415 loss: 11.3561 loss_cls: 0.8448 loss_bbox: 1.0459 d0.loss_cls: 0.7796 d0.loss_bbox: 1.1626 d1.loss_cls: 0.7914 d1.loss_bbox: 1.0846 d2.loss_cls: 0.8169 d2.loss_bbox: 1.0627 d3.loss_cls: 0.8279 d3.loss_bbox: 1.0541 d4.loss_cls: 0.8368 d4.loss_bbox: 1.0488 +10/10 19:44:49 - mmengine - INFO - Epoch(train) [2][3300/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:58:24 time: 0.6785 data_time: 0.0405 memory: 11853 grad_norm: 46.5716 loss: 9.2160 loss_cls: 0.6975 loss_bbox: 0.8445 d0.loss_cls: 0.6281 d0.loss_bbox: 0.9487 d1.loss_cls: 0.6365 d1.loss_bbox: 0.8833 d2.loss_cls: 0.6543 d2.loss_bbox: 0.8552 d3.loss_cls: 0.6655 d3.loss_bbox: 0.8662 d4.loss_cls: 0.6746 d4.loss_bbox: 0.8615 +10/10 19:45:23 - mmengine - INFO - Epoch(train) [2][3350/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:56:51 time: 0.6759 data_time: 0.0401 memory: 11853 grad_norm: 48.0274 loss: 8.6218 loss_cls: 0.6528 loss_bbox: 0.7780 d0.loss_cls: 0.5880 d0.loss_bbox: 0.9020 d1.loss_cls: 0.6032 d1.loss_bbox: 0.8264 d2.loss_cls: 0.6254 d2.loss_bbox: 0.7957 d3.loss_cls: 0.6407 d3.loss_bbox: 0.7838 d4.loss_cls: 0.6449 d4.loss_bbox: 0.7808 +10/10 19:45:57 - mmengine - INFO - Epoch(train) [2][3400/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:55:20 time: 0.6777 data_time: 0.0401 memory: 11853 grad_norm: 43.6840 loss: 9.2039 loss_cls: 0.6736 loss_bbox: 0.8550 d0.loss_cls: 0.6239 d0.loss_bbox: 0.9729 d1.loss_cls: 0.6317 d1.loss_bbox: 0.8946 d2.loss_cls: 0.6477 d2.loss_bbox: 0.8678 d3.loss_cls: 0.6575 d3.loss_bbox: 0.8594 d4.loss_cls: 0.6624 d4.loss_bbox: 0.8575 +10/10 19:46:31 - mmengine - INFO - Epoch(train) [2][3450/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:53:48 time: 0.6768 data_time: 0.0403 memory: 11853 grad_norm: 40.8090 loss: 8.9380 loss_cls: 0.6807 loss_bbox: 0.8001 d0.loss_cls: 0.6150 d0.loss_bbox: 0.9432 d1.loss_cls: 0.6185 d1.loss_bbox: 0.8570 d2.loss_cls: 0.6397 d2.loss_bbox: 0.8337 d3.loss_cls: 0.6486 d3.loss_bbox: 0.8242 d4.loss_cls: 0.6656 d4.loss_bbox: 0.8118 +10/10 19:47:05 - mmengine - INFO - Epoch(train) [2][3500/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:52:17 time: 0.6767 data_time: 0.0402 memory: 11853 grad_norm: 40.8828 loss: 9.4192 loss_cls: 0.6793 loss_bbox: 0.8860 d0.loss_cls: 0.6147 d0.loss_bbox: 1.0097 d1.loss_cls: 0.6281 d1.loss_bbox: 0.9298 d2.loss_cls: 0.6517 d2.loss_bbox: 0.9001 d3.loss_cls: 0.6664 d3.loss_bbox: 0.8932 d4.loss_cls: 0.6712 d4.loss_bbox: 0.8891 +10/10 19:47:39 - mmengine - INFO - Epoch(train) [2][3550/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:50:48 time: 0.6781 data_time: 0.0399 memory: 11853 grad_norm: 51.0573 loss: 9.3792 loss_cls: 0.6950 loss_bbox: 0.8547 d0.loss_cls: 0.6531 d0.loss_bbox: 0.9650 d1.loss_cls: 0.6736 d1.loss_bbox: 0.8809 d2.loss_cls: 0.6887 d2.loss_bbox: 0.8616 d3.loss_cls: 0.6852 d3.loss_bbox: 0.8662 d4.loss_cls: 0.6957 d4.loss_bbox: 0.8596 +10/10 19:48:13 - mmengine - INFO - Epoch(train) [2][3600/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:49:22 time: 0.6822 data_time: 0.0396 memory: 11853 grad_norm: 49.5295 loss: 10.1024 loss_cls: 0.7510 loss_bbox: 0.9144 d0.loss_cls: 0.6935 d0.loss_bbox: 1.0705 d1.loss_cls: 0.7145 d1.loss_bbox: 0.9646 d2.loss_cls: 0.7239 d2.loss_bbox: 0.9390 d3.loss_cls: 0.7398 d3.loss_bbox: 0.9238 d4.loss_cls: 0.7438 d4.loss_bbox: 0.9236 +10/10 19:48:47 - mmengine - INFO - Epoch(train) [2][3650/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:47:54 time: 0.6795 data_time: 0.0399 memory: 11853 grad_norm: 41.3008 loss: 9.8281 loss_cls: 0.7289 loss_bbox: 0.8967 d0.loss_cls: 0.6629 d0.loss_bbox: 1.0171 d1.loss_cls: 0.6894 d1.loss_bbox: 0.9424 d2.loss_cls: 0.7066 d2.loss_bbox: 0.9253 d3.loss_cls: 0.7181 d3.loss_bbox: 0.9165 d4.loss_cls: 0.7233 d4.loss_bbox: 0.9010 +10/10 19:49:20 - mmengine - INFO - Epoch(train) [2][3700/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:46:25 time: 0.6758 data_time: 0.0405 memory: 11853 grad_norm: 50.3552 loss: 9.5228 loss_cls: 0.7037 loss_bbox: 0.8677 d0.loss_cls: 0.6366 d0.loss_bbox: 1.0222 d1.loss_cls: 0.6582 d1.loss_bbox: 0.9212 d2.loss_cls: 0.6722 d2.loss_bbox: 0.8894 d3.loss_cls: 0.6844 d3.loss_bbox: 0.8888 d4.loss_cls: 0.7021 d4.loss_bbox: 0.8765 +10/10 19:49:54 - mmengine - INFO - Epoch(train) [2][3750/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:44:58 time: 0.6783 data_time: 0.0399 memory: 11853 grad_norm: 48.5526 loss: 9.1411 loss_cls: 0.6563 loss_bbox: 0.8647 d0.loss_cls: 0.5994 d0.loss_bbox: 0.9805 d1.loss_cls: 0.6019 d1.loss_bbox: 0.9124 d2.loss_cls: 0.6089 d2.loss_bbox: 0.8877 d3.loss_cls: 0.6346 d3.loss_bbox: 0.8774 d4.loss_cls: 0.6447 d4.loss_bbox: 0.8724 +10/10 19:50:28 - mmengine - INFO - Epoch(train) [2][3800/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:43:31 time: 0.6791 data_time: 0.0408 memory: 11853 grad_norm: 46.3550 loss: 9.2766 loss_cls: 0.6921 loss_bbox: 0.8504 d0.loss_cls: 0.6263 d0.loss_bbox: 0.9572 d1.loss_cls: 0.6478 d1.loss_bbox: 0.8879 d2.loss_cls: 0.6658 d2.loss_bbox: 0.8671 d3.loss_cls: 0.6773 d3.loss_bbox: 0.8618 d4.loss_cls: 0.6860 d4.loss_bbox: 0.8569 +10/10 19:51:12 - mmengine - INFO - Epoch(train) [2][3850/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:44:26 time: 0.8720 data_time: 0.0404 memory: 11853 grad_norm: 40.1614 loss: 9.3704 loss_cls: 0.6706 loss_bbox: 0.8692 d0.loss_cls: 0.6208 d0.loss_bbox: 1.0002 d1.loss_cls: 0.6311 d1.loss_bbox: 0.9282 d2.loss_cls: 0.6534 d2.loss_bbox: 0.8967 d3.loss_cls: 0.6651 d3.loss_bbox: 0.8891 d4.loss_cls: 0.6693 d4.loss_bbox: 0.8768 +10/10 19:52:00 - mmengine - INFO - Epoch(train) [2][3900/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:46:28 time: 0.9680 data_time: 0.0411 memory: 11853 grad_norm: 49.7404 loss: 8.6690 loss_cls: 0.6356 loss_bbox: 0.8061 d0.loss_cls: 0.5878 d0.loss_bbox: 0.8928 d1.loss_cls: 0.5969 d1.loss_bbox: 0.8386 d2.loss_cls: 0.6170 d2.loss_bbox: 0.8131 d3.loss_cls: 0.6229 d3.loss_bbox: 0.8186 d4.loss_cls: 0.6274 d4.loss_bbox: 0.8122 +10/10 19:52:48 - mmengine - INFO - Epoch(train) [2][3950/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:48:21 time: 0.9564 data_time: 0.0410 memory: 11853 grad_norm: 53.2098 loss: 9.5569 loss_cls: 0.6961 loss_bbox: 0.8893 d0.loss_cls: 0.6288 d0.loss_bbox: 1.0153 d1.loss_cls: 0.6357 d1.loss_bbox: 0.9422 d2.loss_cls: 0.6638 d2.loss_bbox: 0.9121 d3.loss_cls: 0.6803 d3.loss_bbox: 0.9051 d4.loss_cls: 0.6873 d4.loss_bbox: 0.9008 +10/10 19:53:05 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 19:53:36 - mmengine - INFO - Epoch(train) [2][4000/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:50:16 time: 0.9615 data_time: 0.0407 memory: 11853 grad_norm: 43.6996 loss: 9.0037 loss_cls: 0.6682 loss_bbox: 0.8173 d0.loss_cls: 0.6204 d0.loss_bbox: 0.9431 d1.loss_cls: 0.6462 d1.loss_bbox: 0.8509 d2.loss_cls: 0.6467 d2.loss_bbox: 0.8398 d3.loss_cls: 0.6515 d3.loss_bbox: 0.8331 d4.loss_cls: 0.6655 d4.loss_bbox: 0.8209 +10/10 19:54:19 - mmengine - INFO - Epoch(train) [2][4050/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:50:46 time: 0.8444 data_time: 0.0408 memory: 11853 grad_norm: 53.1764 loss: 9.2872 loss_cls: 0.6949 loss_bbox: 0.8450 d0.loss_cls: 0.6442 d0.loss_bbox: 0.9431 d1.loss_cls: 0.6552 d1.loss_bbox: 0.8868 d2.loss_cls: 0.6642 d2.loss_bbox: 0.8723 d3.loss_cls: 0.6701 d3.loss_bbox: 0.8715 d4.loss_cls: 0.6844 d4.loss_bbox: 0.8552 +10/10 19:54:52 - mmengine - INFO - Epoch(train) [2][4100/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:49:18 time: 0.6780 data_time: 0.0401 memory: 11853 grad_norm: 45.3966 loss: 9.2834 loss_cls: 0.6547 loss_bbox: 0.8765 d0.loss_cls: 0.6135 d0.loss_bbox: 1.0057 d1.loss_cls: 0.6291 d1.loss_bbox: 0.9208 d2.loss_cls: 0.6365 d2.loss_bbox: 0.8907 d3.loss_cls: 0.6397 d3.loss_bbox: 0.8912 d4.loss_cls: 0.6515 d4.loss_bbox: 0.8737 +10/10 19:55:27 - mmengine - INFO - Epoch(train) [2][4150/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:47:53 time: 0.6823 data_time: 0.0403 memory: 11853 grad_norm: 41.9276 loss: 8.5877 loss_cls: 0.6272 loss_bbox: 0.7950 d0.loss_cls: 0.5743 d0.loss_bbox: 0.9075 d1.loss_cls: 0.5883 d1.loss_bbox: 0.8386 d2.loss_cls: 0.6040 d2.loss_bbox: 0.8133 d3.loss_cls: 0.6122 d3.loss_bbox: 0.8046 d4.loss_cls: 0.6197 d4.loss_bbox: 0.8030 +10/10 19:56:00 - mmengine - INFO - Epoch(train) [2][4200/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:46:26 time: 0.6782 data_time: 0.0399 memory: 11853 grad_norm: 47.1861 loss: 9.6528 loss_cls: 0.7176 loss_bbox: 0.8741 d0.loss_cls: 0.6570 d0.loss_bbox: 1.0109 d1.loss_cls: 0.6709 d1.loss_bbox: 0.9292 d2.loss_cls: 0.6853 d2.loss_bbox: 0.9076 d3.loss_cls: 0.6995 d3.loss_bbox: 0.9011 d4.loss_cls: 0.7100 d4.loss_bbox: 0.8893 +10/10 19:56:34 - mmengine - INFO - Epoch(train) [2][4250/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:44:59 time: 0.6779 data_time: 0.0399 memory: 11853 grad_norm: 46.7114 loss: 9.5993 loss_cls: 0.7001 loss_bbox: 0.8842 d0.loss_cls: 0.6409 d0.loss_bbox: 1.0108 d1.loss_cls: 0.6748 d1.loss_bbox: 0.9290 d2.loss_cls: 0.6855 d2.loss_bbox: 0.9052 d3.loss_cls: 0.6866 d3.loss_bbox: 0.9010 d4.loss_cls: 0.6950 d4.loss_bbox: 0.8863 +10/10 19:57:08 - mmengine - INFO - Epoch(train) [2][4300/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:43:33 time: 0.6790 data_time: 0.0404 memory: 11853 grad_norm: 48.1774 loss: 9.3907 loss_cls: 0.6896 loss_bbox: 0.8724 d0.loss_cls: 0.6275 d0.loss_bbox: 0.9818 d1.loss_cls: 0.6457 d1.loss_bbox: 0.9163 d2.loss_cls: 0.6550 d2.loss_bbox: 0.8919 d3.loss_cls: 0.6699 d3.loss_bbox: 0.8839 d4.loss_cls: 0.6837 d4.loss_bbox: 0.8729 +10/10 19:57:42 - mmengine - INFO - Epoch(train) [2][4350/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:42:07 time: 0.6776 data_time: 0.0404 memory: 11853 grad_norm: 47.2596 loss: 9.2376 loss_cls: 0.6847 loss_bbox: 0.8549 d0.loss_cls: 0.6265 d0.loss_bbox: 0.9612 d1.loss_cls: 0.6392 d1.loss_bbox: 0.8857 d2.loss_cls: 0.6509 d2.loss_bbox: 0.8691 d3.loss_cls: 0.6655 d3.loss_bbox: 0.8644 d4.loss_cls: 0.6814 d4.loss_bbox: 0.8540 +10/10 19:58:16 - mmengine - INFO - Epoch(train) [2][4400/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:40:41 time: 0.6782 data_time: 0.0401 memory: 11853 grad_norm: 47.5893 loss: 10.8288 loss_cls: 0.7959 loss_bbox: 0.9961 d0.loss_cls: 0.7484 d0.loss_bbox: 1.1325 d1.loss_cls: 0.7451 d1.loss_bbox: 1.0483 d2.loss_cls: 0.7639 d2.loss_bbox: 1.0231 d3.loss_cls: 0.7689 d3.loss_bbox: 1.0172 d4.loss_cls: 0.7798 d4.loss_bbox: 1.0095 +10/10 19:58:50 - mmengine - INFO - Epoch(train) [2][4450/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:39:16 time: 0.6781 data_time: 0.0399 memory: 11853 grad_norm: 52.3482 loss: 10.1294 loss_cls: 0.7542 loss_bbox: 0.9209 d0.loss_cls: 0.7116 d0.loss_bbox: 1.0501 d1.loss_cls: 0.7239 d1.loss_bbox: 0.9635 d2.loss_cls: 0.7365 d2.loss_bbox: 0.9342 d3.loss_cls: 0.7394 d3.loss_bbox: 0.9284 d4.loss_cls: 0.7475 d4.loss_bbox: 0.9192 +10/10 19:59:24 - mmengine - INFO - Epoch(train) [2][4500/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:37:51 time: 0.6767 data_time: 0.0400 memory: 11853 grad_norm: 46.0567 loss: 9.4342 loss_cls: 0.6943 loss_bbox: 0.8802 d0.loss_cls: 0.6252 d0.loss_bbox: 0.9734 d1.loss_cls: 0.6569 d1.loss_bbox: 0.9058 d2.loss_cls: 0.6738 d2.loss_bbox: 0.8842 d3.loss_cls: 0.6904 d3.loss_bbox: 0.8795 d4.loss_cls: 0.6882 d4.loss_bbox: 0.8822 +10/10 19:59:58 - mmengine - INFO - Epoch(train) [2][4550/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:36:26 time: 0.6775 data_time: 0.0402 memory: 11853 grad_norm: 60.5506 loss: 9.7997 loss_cls: 0.6967 loss_bbox: 0.9240 d0.loss_cls: 0.6441 d0.loss_bbox: 1.0489 d1.loss_cls: 0.6593 d1.loss_bbox: 0.9727 d2.loss_cls: 0.6845 d2.loss_bbox: 0.9303 d3.loss_cls: 0.6849 d3.loss_bbox: 0.9320 d4.loss_cls: 0.6880 d4.loss_bbox: 0.9343 +10/10 20:00:32 - mmengine - INFO - Epoch(train) [2][4600/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:35:02 time: 0.6774 data_time: 0.0399 memory: 11853 grad_norm: 46.0743 loss: 9.2577 loss_cls: 0.6835 loss_bbox: 0.8506 d0.loss_cls: 0.6433 d0.loss_bbox: 0.9491 d1.loss_cls: 0.6462 d1.loss_bbox: 0.8882 d2.loss_cls: 0.6638 d2.loss_bbox: 0.8702 d3.loss_cls: 0.6608 d3.loss_bbox: 0.8703 d4.loss_cls: 0.6789 d4.loss_bbox: 0.8529 +10/10 20:01:05 - mmengine - INFO - Epoch(train) [2][4650/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:33:38 time: 0.6773 data_time: 0.0398 memory: 11853 grad_norm: 51.0259 loss: 9.9459 loss_cls: 0.7233 loss_bbox: 0.9215 d0.loss_cls: 0.6600 d0.loss_bbox: 1.0651 d1.loss_cls: 0.6764 d1.loss_bbox: 0.9731 d2.loss_cls: 0.6977 d2.loss_bbox: 0.9503 d3.loss_cls: 0.7066 d3.loss_bbox: 0.9351 d4.loss_cls: 0.7156 d4.loss_bbox: 0.9214 +10/10 20:01:39 - mmengine - INFO - Epoch(train) [2][4700/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:32:16 time: 0.6782 data_time: 0.0401 memory: 11853 grad_norm: 48.4333 loss: 8.8006 loss_cls: 0.6323 loss_bbox: 0.8166 d0.loss_cls: 0.5938 d0.loss_bbox: 0.9405 d1.loss_cls: 0.5984 d1.loss_bbox: 0.8648 d2.loss_cls: 0.6132 d2.loss_bbox: 0.8325 d3.loss_cls: 0.6228 d3.loss_bbox: 0.8358 d4.loss_cls: 0.6280 d4.loss_bbox: 0.8219 +10/10 20:02:15 - mmengine - INFO - Epoch(train) [2][4750/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:31:15 time: 0.7113 data_time: 0.0398 memory: 11853 grad_norm: 57.6995 loss: 8.0065 loss_cls: 0.5715 loss_bbox: 0.7414 d0.loss_cls: 0.5532 d0.loss_bbox: 0.8440 d1.loss_cls: 0.5501 d1.loss_bbox: 0.7771 d2.loss_cls: 0.5612 d2.loss_bbox: 0.7598 d3.loss_cls: 0.5717 d3.loss_bbox: 0.7567 d4.loss_cls: 0.5665 d4.loss_bbox: 0.7531 +10/10 20:03:03 - mmengine - INFO - Epoch(train) [2][4800/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:32:55 time: 0.9522 data_time: 0.0403 memory: 11853 grad_norm: 55.8774 loss: 9.2722 loss_cls: 0.6746 loss_bbox: 0.8630 d0.loss_cls: 0.6168 d0.loss_bbox: 0.9903 d1.loss_cls: 0.6371 d1.loss_bbox: 0.9015 d2.loss_cls: 0.6461 d2.loss_bbox: 0.8809 d3.loss_cls: 0.6541 d3.loss_bbox: 0.8749 d4.loss_cls: 0.6632 d4.loss_bbox: 0.8698 +10/10 20:03:51 - mmengine - INFO - Epoch(train) [2][4850/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:34:44 time: 0.9679 data_time: 0.0409 memory: 11853 grad_norm: 48.3262 loss: 9.5926 loss_cls: 0.7314 loss_bbox: 0.8604 d0.loss_cls: 0.6876 d0.loss_bbox: 0.9542 d1.loss_cls: 0.6996 d1.loss_bbox: 0.8955 d2.loss_cls: 0.7130 d2.loss_bbox: 0.8757 d3.loss_cls: 0.7236 d3.loss_bbox: 0.8697 d4.loss_cls: 0.7195 d4.loss_bbox: 0.8623 +10/10 20:04:44 - mmengine - INFO - Epoch(train) [2][4900/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:37:32 time: 1.0603 data_time: 0.0508 memory: 11853 grad_norm: 54.0255 loss: 9.2385 loss_cls: 0.6744 loss_bbox: 0.8585 d0.loss_cls: 0.6225 d0.loss_bbox: 0.9504 d1.loss_cls: 0.6360 d1.loss_bbox: 0.8956 d2.loss_cls: 0.6550 d2.loss_bbox: 0.8768 d3.loss_cls: 0.6704 d3.loss_bbox: 0.8671 d4.loss_cls: 0.6756 d4.loss_bbox: 0.8562 +10/10 20:05:35 - mmengine - INFO - Epoch(train) [2][4950/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:39:49 time: 1.0159 data_time: 0.0486 memory: 11853 grad_norm: 48.2365 loss: 8.5762 loss_cls: 0.6337 loss_bbox: 0.7840 d0.loss_cls: 0.5863 d0.loss_bbox: 0.8989 d1.loss_cls: 0.5994 d1.loss_bbox: 0.8232 d2.loss_cls: 0.6141 d2.loss_bbox: 0.7981 d3.loss_cls: 0.6220 d3.loss_bbox: 0.7976 d4.loss_cls: 0.6347 d4.loss_bbox: 0.7843 +10/10 20:05:47 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 20:06:09 - mmengine - INFO - Epoch(train) [2][5000/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:38:32 time: 0.6886 data_time: 0.0406 memory: 11853 grad_norm: 50.5573 loss: 10.4281 loss_cls: 0.7670 loss_bbox: 0.9657 d0.loss_cls: 0.7169 d0.loss_bbox: 1.0847 d1.loss_cls: 0.7334 d1.loss_bbox: 1.0084 d2.loss_cls: 0.7411 d2.loss_bbox: 0.9723 d3.loss_cls: 0.7497 d3.loss_bbox: 0.9659 d4.loss_cls: 0.7587 d4.loss_bbox: 0.9644 +10/10 20:06:43 - mmengine - INFO - Epoch(train) [2][5050/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:37:11 time: 0.6822 data_time: 0.0396 memory: 11853 grad_norm: 56.1532 loss: 9.2565 loss_cls: 0.6718 loss_bbox: 0.8601 d0.loss_cls: 0.6082 d0.loss_bbox: 0.9826 d1.loss_cls: 0.6187 d1.loss_bbox: 0.9128 d2.loss_cls: 0.6357 d2.loss_bbox: 0.8997 d3.loss_cls: 0.6524 d3.loss_bbox: 0.8841 d4.loss_cls: 0.6667 d4.loss_bbox: 0.8636 +10/10 20:07:19 - mmengine - INFO - Epoch(train) [2][5100/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:36:07 time: 0.7084 data_time: 0.0403 memory: 11853 grad_norm: 57.6904 loss: 9.1057 loss_cls: 0.6610 loss_bbox: 0.8458 d0.loss_cls: 0.6121 d0.loss_bbox: 0.9503 d1.loss_cls: 0.6307 d1.loss_bbox: 0.8851 d2.loss_cls: 0.6460 d2.loss_bbox: 0.8638 d3.loss_cls: 0.6504 d3.loss_bbox: 0.8568 d4.loss_cls: 0.6549 d4.loss_bbox: 0.8486 +10/10 20:07:55 - mmengine - INFO - Epoch(train) [2][5150/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:35:09 time: 0.7169 data_time: 0.0411 memory: 11853 grad_norm: 53.0566 loss: 9.2260 loss_cls: 0.6881 loss_bbox: 0.8501 d0.loss_cls: 0.6307 d0.loss_bbox: 0.9643 d1.loss_cls: 0.6433 d1.loss_bbox: 0.8854 d2.loss_cls: 0.6582 d2.loss_bbox: 0.8563 d3.loss_cls: 0.6697 d3.loss_bbox: 0.8509 d4.loss_cls: 0.6782 d4.loss_bbox: 0.8508 +10/10 20:08:30 - mmengine - INFO - Epoch(train) [2][5200/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:34:06 time: 0.7100 data_time: 0.0411 memory: 11853 grad_norm: 60.0672 loss: 9.6895 loss_cls: 0.7084 loss_bbox: 0.8785 d0.loss_cls: 0.6596 d0.loss_bbox: 1.0250 d1.loss_cls: 0.6802 d1.loss_bbox: 0.9422 d2.loss_cls: 0.6948 d2.loss_bbox: 0.9126 d3.loss_cls: 0.6967 d3.loss_bbox: 0.9048 d4.loss_cls: 0.7012 d4.loss_bbox: 0.8855 +10/10 20:09:04 - mmengine - INFO - Epoch(train) [2][5250/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:32:45 time: 0.6806 data_time: 0.0407 memory: 11853 grad_norm: 49.5067 loss: 10.8945 loss_cls: 0.7622 loss_bbox: 1.0438 d0.loss_cls: 0.7019 d0.loss_bbox: 1.1659 d1.loss_cls: 0.7219 d1.loss_bbox: 1.0813 d2.loss_cls: 0.7314 d2.loss_bbox: 1.0782 d3.loss_cls: 0.7430 d3.loss_bbox: 1.0604 d4.loss_cls: 0.7524 d4.loss_bbox: 1.0520 +10/10 20:09:38 - mmengine - INFO - Epoch(train) [2][5300/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:31:23 time: 0.6771 data_time: 0.0401 memory: 11853 grad_norm: 52.2692 loss: 9.8974 loss_cls: 0.7258 loss_bbox: 0.9189 d0.loss_cls: 0.6766 d0.loss_bbox: 1.0427 d1.loss_cls: 0.6763 d1.loss_bbox: 0.9697 d2.loss_cls: 0.6936 d2.loss_bbox: 0.9377 d3.loss_cls: 0.7013 d3.loss_bbox: 0.9296 d4.loss_cls: 0.7069 d4.loss_bbox: 0.9184 +10/10 20:10:12 - mmengine - INFO - Epoch(train) [2][5350/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:30:01 time: 0.6786 data_time: 0.0400 memory: 11853 grad_norm: 53.5942 loss: 10.0649 loss_cls: 0.7208 loss_bbox: 0.9444 d0.loss_cls: 0.6792 d0.loss_bbox: 1.0613 d1.loss_cls: 0.6822 d1.loss_bbox: 0.9899 d2.loss_cls: 0.6990 d2.loss_bbox: 0.9628 d3.loss_cls: 0.7042 d3.loss_bbox: 0.9618 d4.loss_cls: 0.7145 d4.loss_bbox: 0.9447 +10/10 20:10:46 - mmengine - INFO - Epoch(train) [2][5400/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:28:40 time: 0.6783 data_time: 0.0401 memory: 11853 grad_norm: 50.6985 loss: 9.2923 loss_cls: 0.6470 loss_bbox: 0.8817 d0.loss_cls: 0.6121 d0.loss_bbox: 1.0224 d1.loss_cls: 0.6130 d1.loss_bbox: 0.9341 d2.loss_cls: 0.6135 d2.loss_bbox: 0.9142 d3.loss_cls: 0.6190 d3.loss_bbox: 0.9034 d4.loss_cls: 0.6382 d4.loss_bbox: 0.8937 +10/10 20:11:20 - mmengine - INFO - Epoch(train) [2][5450/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:27:20 time: 0.6798 data_time: 0.0405 memory: 11853 grad_norm: 49.8003 loss: 9.0373 loss_cls: 0.6713 loss_bbox: 0.8215 d0.loss_cls: 0.6155 d0.loss_bbox: 0.9243 d1.loss_cls: 0.6349 d1.loss_bbox: 0.8712 d2.loss_cls: 0.6532 d2.loss_bbox: 0.8484 d3.loss_cls: 0.6549 d3.loss_bbox: 0.8470 d4.loss_cls: 0.6688 d4.loss_bbox: 0.8261 +10/10 20:11:54 - mmengine - INFO - Epoch(train) [2][5500/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:25:58 time: 0.6765 data_time: 0.0388 memory: 11853 grad_norm: 54.2921 loss: 10.1808 loss_cls: 0.7575 loss_bbox: 0.9318 d0.loss_cls: 0.6925 d0.loss_bbox: 1.0719 d1.loss_cls: 0.7089 d1.loss_bbox: 0.9846 d2.loss_cls: 0.7203 d2.loss_bbox: 0.9518 d3.loss_cls: 0.7310 d3.loss_bbox: 0.9508 d4.loss_cls: 0.7452 d4.loss_bbox: 0.9342 +10/10 20:12:37 - mmengine - INFO - Epoch(train) [2][5550/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:26:31 time: 0.8603 data_time: 0.0397 memory: 11853 grad_norm: 53.3420 loss: 8.5188 loss_cls: 0.6302 loss_bbox: 0.7886 d0.loss_cls: 0.5716 d0.loss_bbox: 0.8751 d1.loss_cls: 0.5931 d1.loss_bbox: 0.8164 d2.loss_cls: 0.6050 d2.loss_bbox: 0.8021 d3.loss_cls: 0.6194 d3.loss_bbox: 0.8009 d4.loss_cls: 0.6236 d4.loss_bbox: 0.7927 +10/10 20:13:24 - mmengine - INFO - Epoch(train) [2][5600/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:28:02 time: 0.9554 data_time: 0.0408 memory: 11853 grad_norm: 89.9621 loss: 9.7952 loss_cls: 0.7111 loss_bbox: 0.9131 d0.loss_cls: 0.6704 d0.loss_bbox: 1.0216 d1.loss_cls: 0.6843 d1.loss_bbox: 0.9426 d2.loss_cls: 0.6787 d2.loss_bbox: 0.9342 d3.loss_cls: 0.6857 d3.loss_bbox: 0.9319 d4.loss_cls: 0.7007 d4.loss_bbox: 0.9211 +10/10 20:14:13 - mmengine - INFO - Epoch(train) [2][5650/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:29:36 time: 0.9615 data_time: 0.0408 memory: 11853 grad_norm: 47.4914 loss: 10.0848 loss_cls: 0.7098 loss_bbox: 0.9645 d0.loss_cls: 0.6651 d0.loss_bbox: 1.0790 d1.loss_cls: 0.6694 d1.loss_bbox: 1.0106 d2.loss_cls: 0.6891 d2.loss_bbox: 0.9702 d3.loss_cls: 0.6937 d3.loss_bbox: 0.9669 d4.loss_cls: 0.6977 d4.loss_bbox: 0.9687 +10/10 20:15:00 - mmengine - INFO - Epoch(train) [2][5700/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:31:06 time: 0.9577 data_time: 0.0403 memory: 11853 grad_norm: 94.4380 loss: 9.4302 loss_cls: 0.6647 loss_bbox: 0.8959 d0.loss_cls: 0.6428 d0.loss_bbox: 0.9981 d1.loss_cls: 0.6370 d1.loss_bbox: 0.9344 d2.loss_cls: 0.6505 d2.loss_bbox: 0.9038 d3.loss_cls: 0.6516 d3.loss_bbox: 0.8992 d4.loss_cls: 0.6595 d4.loss_bbox: 0.8927 +10/10 20:15:46 - mmengine - INFO - Epoch(train) [2][5750/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:32:11 time: 0.9185 data_time: 0.0404 memory: 11853 grad_norm: 52.7806 loss: 9.0619 loss_cls: 0.6534 loss_bbox: 0.8407 d0.loss_cls: 0.6061 d0.loss_bbox: 0.9550 d1.loss_cls: 0.6195 d1.loss_bbox: 0.8899 d2.loss_cls: 0.6356 d2.loss_bbox: 0.8657 d3.loss_cls: 0.6407 d3.loss_bbox: 0.8587 d4.loss_cls: 0.6499 d4.loss_bbox: 0.8466 +10/10 20:16:22 - mmengine - INFO - Epoch(train) [2][5800/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:31:16 time: 0.7212 data_time: 0.0402 memory: 11853 grad_norm: 70.5031 loss: 10.0664 loss_cls: 0.7242 loss_bbox: 0.9373 d0.loss_cls: 0.6702 d0.loss_bbox: 1.0880 d1.loss_cls: 0.6774 d1.loss_bbox: 0.9935 d2.loss_cls: 0.6832 d2.loss_bbox: 0.9684 d3.loss_cls: 0.7010 d3.loss_bbox: 0.9598 d4.loss_cls: 0.7181 d4.loss_bbox: 0.9453 +10/10 20:17:12 - mmengine - INFO - Epoch(train) [2][5850/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:33:07 time: 0.9959 data_time: 0.0442 memory: 11853 grad_norm: 123.1871 loss: 9.5673 loss_cls: 0.6763 loss_bbox: 0.9026 d0.loss_cls: 0.6333 d0.loss_bbox: 1.0410 d1.loss_cls: 0.6313 d1.loss_bbox: 0.9594 d2.loss_cls: 0.6487 d2.loss_bbox: 0.9247 d3.loss_cls: 0.6516 d3.loss_bbox: 0.9272 d4.loss_cls: 0.6628 d4.loss_bbox: 0.9083 +10/10 20:18:00 - mmengine - INFO - Epoch(train) [2][5900/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:34:36 time: 0.9620 data_time: 0.0407 memory: 11853 grad_norm: 56.1264 loss: 9.2495 loss_cls: 0.6787 loss_bbox: 0.8452 d0.loss_cls: 0.6340 d0.loss_bbox: 0.9760 d1.loss_cls: 0.6365 d1.loss_bbox: 0.8983 d2.loss_cls: 0.6540 d2.loss_bbox: 0.8736 d3.loss_cls: 0.6625 d3.loss_bbox: 0.8642 d4.loss_cls: 0.6728 d4.loss_bbox: 0.8537 +10/10 20:18:48 - mmengine - INFO - Epoch(train) [2][5950/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:36:00 time: 0.9553 data_time: 0.0405 memory: 11853 grad_norm: 44.5113 loss: 9.6387 loss_cls: 0.7072 loss_bbox: 0.8938 d0.loss_cls: 0.6589 d0.loss_bbox: 1.0041 d1.loss_cls: 0.6651 d1.loss_bbox: 0.9339 d2.loss_cls: 0.6785 d2.loss_bbox: 0.9122 d3.loss_cls: 0.6827 d3.loss_bbox: 0.9080 d4.loss_cls: 0.6965 d4.loss_bbox: 0.8978 +10/10 20:19:04 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 20:19:36 - mmengine - INFO - Epoch(train) [2][6000/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:37:23 time: 0.9561 data_time: 0.0405 memory: 11853 grad_norm: 52.2684 loss: 8.7281 loss_cls: 0.6371 loss_bbox: 0.8065 d0.loss_cls: 0.6124 d0.loss_bbox: 0.9126 d1.loss_cls: 0.6108 d1.loss_bbox: 0.8387 d2.loss_cls: 0.6081 d2.loss_bbox: 0.8267 d3.loss_cls: 0.6137 d3.loss_bbox: 0.8210 d4.loss_cls: 0.6256 d4.loss_bbox: 0.8149 +10/10 20:20:13 - mmengine - INFO - Epoch(train) [2][6050/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:36:39 time: 0.7428 data_time: 0.0414 memory: 11853 grad_norm: 52.6743 loss: 8.7662 loss_cls: 0.6612 loss_bbox: 0.7909 d0.loss_cls: 0.6201 d0.loss_bbox: 0.8995 d1.loss_cls: 0.6112 d1.loss_bbox: 0.8339 d2.loss_cls: 0.6287 d2.loss_bbox: 0.8205 d3.loss_cls: 0.6426 d3.loss_bbox: 0.8088 d4.loss_cls: 0.6543 d4.loss_bbox: 0.7944 +10/10 20:20:56 - mmengine - INFO - Epoch(train) [2][6100/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:37:08 time: 0.8659 data_time: 0.0407 memory: 11853 grad_norm: 48.4169 loss: 9.5513 loss_cls: 0.6911 loss_bbox: 0.8840 d0.loss_cls: 0.6515 d0.loss_bbox: 1.0070 d1.loss_cls: 0.6546 d1.loss_bbox: 0.9371 d2.loss_cls: 0.6666 d2.loss_bbox: 0.9066 d3.loss_cls: 0.6794 d3.loss_bbox: 0.8989 d4.loss_cls: 0.6844 d4.loss_bbox: 0.8900 +10/10 20:21:39 - mmengine - INFO - Epoch(train) [2][6150/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:37:33 time: 0.8611 data_time: 0.0402 memory: 11853 grad_norm: 52.3341 loss: 8.5969 loss_cls: 0.6468 loss_bbox: 0.7858 d0.loss_cls: 0.5902 d0.loss_bbox: 0.8825 d1.loss_cls: 0.6001 d1.loss_bbox: 0.8245 d2.loss_cls: 0.6126 d2.loss_bbox: 0.8054 d3.loss_cls: 0.6318 d3.loss_bbox: 0.7939 d4.loss_cls: 0.6420 d4.loss_bbox: 0.7813 +10/10 20:22:25 - mmengine - INFO - Epoch(train) [2][6200/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:38:25 time: 0.9077 data_time: 0.0457 memory: 11853 grad_norm: 50.3472 loss: 10.1711 loss_cls: 0.7304 loss_bbox: 0.9561 d0.loss_cls: 0.6829 d0.loss_bbox: 1.0680 d1.loss_cls: 0.6930 d1.loss_bbox: 1.0011 d2.loss_cls: 0.7047 d2.loss_bbox: 0.9797 d3.loss_cls: 0.7227 d3.loss_bbox: 0.9533 d4.loss_cls: 0.7318 d4.loss_bbox: 0.9474 +10/10 20:23:07 - mmengine - INFO - Epoch(train) [2][6250/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:38:39 time: 0.8432 data_time: 0.0479 memory: 11853 grad_norm: 43.6476 loss: 8.3815 loss_cls: 0.6071 loss_bbox: 0.7691 d0.loss_cls: 0.5821 d0.loss_bbox: 0.8866 d1.loss_cls: 0.5820 d1.loss_bbox: 0.8166 d2.loss_cls: 0.5931 d2.loss_bbox: 0.7897 d3.loss_cls: 0.6020 d3.loss_bbox: 0.7801 d4.loss_cls: 0.6068 d4.loss_bbox: 0.7663 +10/10 20:23:55 - mmengine - INFO - Epoch(train) [2][6300/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:39:59 time: 0.9589 data_time: 0.0430 memory: 11853 grad_norm: 77.0243 loss: 8.8990 loss_cls: 0.6688 loss_bbox: 0.8098 d0.loss_cls: 0.6202 d0.loss_bbox: 0.9102 d1.loss_cls: 0.6434 d1.loss_bbox: 0.8333 d2.loss_cls: 0.6333 d2.loss_bbox: 0.8334 d3.loss_cls: 0.6421 d3.loss_bbox: 0.8288 d4.loss_cls: 0.6558 d4.loss_bbox: 0.8198 +10/10 20:24:43 - mmengine - INFO - Epoch(train) [2][6350/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:41:23 time: 0.9656 data_time: 0.0416 memory: 11853 grad_norm: 52.3991 loss: 8.9883 loss_cls: 0.6658 loss_bbox: 0.8175 d0.loss_cls: 0.6143 d0.loss_bbox: 0.9305 d1.loss_cls: 0.6302 d1.loss_bbox: 0.8592 d2.loss_cls: 0.6461 d2.loss_bbox: 0.8486 d3.loss_cls: 0.6573 d3.loss_bbox: 0.8380 d4.loss_cls: 0.6603 d4.loss_bbox: 0.8203 +10/10 20:25:32 - mmengine - INFO - Epoch(train) [2][6400/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:42:49 time: 0.9710 data_time: 0.0422 memory: 11853 grad_norm: 51.3018 loss: 9.0560 loss_cls: 0.6574 loss_bbox: 0.8311 d0.loss_cls: 0.6236 d0.loss_bbox: 0.9674 d1.loss_cls: 0.6316 d1.loss_bbox: 0.8689 d2.loss_cls: 0.6292 d2.loss_bbox: 0.8618 d3.loss_cls: 0.6440 d3.loss_bbox: 0.8532 d4.loss_cls: 0.6480 d4.loss_bbox: 0.8398 +10/10 20:26:11 - mmengine - INFO - Epoch(train) [2][6450/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:42:31 time: 0.7940 data_time: 0.0419 memory: 11853 grad_norm: 41.3641 loss: 8.6269 loss_cls: 0.6532 loss_bbox: 0.7846 d0.loss_cls: 0.5898 d0.loss_bbox: 0.9047 d1.loss_cls: 0.5956 d1.loss_bbox: 0.8280 d2.loss_cls: 0.5972 d2.loss_bbox: 0.8169 d3.loss_cls: 0.6186 d3.loss_bbox: 0.8104 d4.loss_cls: 0.6363 d4.loss_bbox: 0.7916 +10/10 20:26:53 - mmengine - INFO - Epoch(train) [2][6500/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:42:38 time: 0.8359 data_time: 0.0399 memory: 11853 grad_norm: 47.2568 loss: 9.3695 loss_cls: 0.6689 loss_bbox: 0.8790 d0.loss_cls: 0.6233 d0.loss_bbox: 1.0036 d1.loss_cls: 0.6269 d1.loss_bbox: 0.9339 d2.loss_cls: 0.6389 d2.loss_bbox: 0.9092 d3.loss_cls: 0.6477 d3.loss_bbox: 0.8973 d4.loss_cls: 0.6610 d4.loss_bbox: 0.8798 +10/10 20:27:43 - mmengine - INFO - Epoch(train) [2][6550/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:44:18 time: 0.9990 data_time: 0.0397 memory: 11853 grad_norm: 62.8784 loss: 8.8817 loss_cls: 0.6508 loss_bbox: 0.8209 d0.loss_cls: 0.5925 d0.loss_bbox: 0.9461 d1.loss_cls: 0.6116 d1.loss_bbox: 0.8627 d2.loss_cls: 0.6198 d2.loss_bbox: 0.8482 d3.loss_cls: 0.6252 d3.loss_bbox: 0.8375 d4.loss_cls: 0.6391 d4.loss_bbox: 0.8273 +10/10 20:28:33 - mmengine - INFO - Epoch(train) [2][6600/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:45:57 time: 1.0004 data_time: 0.0398 memory: 11853 grad_norm: 44.3527 loss: 9.5539 loss_cls: 0.6802 loss_bbox: 0.8921 d0.loss_cls: 0.6437 d0.loss_bbox: 1.0136 d1.loss_cls: 0.6460 d1.loss_bbox: 0.9443 d2.loss_cls: 0.6595 d2.loss_bbox: 0.9244 d3.loss_cls: 0.6649 d3.loss_bbox: 0.9150 d4.loss_cls: 0.6710 d4.loss_bbox: 0.8994 +10/10 20:29:24 - mmengine - INFO - Epoch(train) [2][6650/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:47:38 time: 1.0065 data_time: 0.0404 memory: 11853 grad_norm: 104.6963 loss: 9.2786 loss_cls: 0.6542 loss_bbox: 0.8777 d0.loss_cls: 0.6131 d0.loss_bbox: 0.9820 d1.loss_cls: 0.6238 d1.loss_bbox: 0.9145 d2.loss_cls: 0.6309 d2.loss_bbox: 0.9034 d3.loss_cls: 0.6433 d3.loss_bbox: 0.8997 d4.loss_cls: 0.6466 d4.loss_bbox: 0.8894 +10/10 20:30:14 - mmengine - INFO - Epoch(train) [2][6700/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:49:18 time: 1.0055 data_time: 0.0404 memory: 11853 grad_norm: 55.0153 loss: 9.7776 loss_cls: 0.7077 loss_bbox: 0.9058 d0.loss_cls: 0.6549 d0.loss_bbox: 1.0402 d1.loss_cls: 0.6722 d1.loss_bbox: 0.9528 d2.loss_cls: 0.6997 d2.loss_bbox: 0.9137 d3.loss_cls: 0.6991 d3.loss_bbox: 0.9166 d4.loss_cls: 0.7117 d4.loss_bbox: 0.9032 +10/10 20:31:04 - mmengine - INFO - Epoch(train) [2][6750/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:50:55 time: 1.0025 data_time: 0.0395 memory: 11853 grad_norm: 74.6167 loss: 8.0670 loss_cls: 0.5779 loss_bbox: 0.7493 d0.loss_cls: 0.5562 d0.loss_bbox: 0.8434 d1.loss_cls: 0.5557 d1.loss_bbox: 0.7850 d2.loss_cls: 0.5698 d2.loss_bbox: 0.7652 d3.loss_cls: 0.5673 d3.loss_bbox: 0.7636 d4.loss_cls: 0.5745 d4.loss_bbox: 0.7592 +10/10 20:31:54 - mmengine - INFO - Epoch(train) [2][6800/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:52:36 time: 1.0100 data_time: 0.0392 memory: 11853 grad_norm: 44.1857 loss: 9.8931 loss_cls: 0.7360 loss_bbox: 0.9023 d0.loss_cls: 0.6725 d0.loss_bbox: 1.0393 d1.loss_cls: 0.6859 d1.loss_bbox: 0.9645 d2.loss_cls: 0.7027 d2.loss_bbox: 0.9293 d3.loss_cls: 0.7104 d3.loss_bbox: 0.9208 d4.loss_cls: 0.7195 d4.loss_bbox: 0.9101 +10/10 20:32:39 - mmengine - INFO - Epoch(train) [2][6850/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:53:09 time: 0.8913 data_time: 0.0403 memory: 11853 grad_norm: 47.2811 loss: 10.2928 loss_cls: 0.7574 loss_bbox: 0.9416 d0.loss_cls: 0.6942 d0.loss_bbox: 1.0745 d1.loss_cls: 0.7125 d1.loss_bbox: 0.9989 d2.loss_cls: 0.7227 d2.loss_bbox: 0.9803 d3.loss_cls: 0.7367 d3.loss_bbox: 0.9709 d4.loss_cls: 0.7490 d4.loss_bbox: 0.9542 +10/10 20:33:13 - mmengine - INFO - Epoch(train) [2][6900/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:51:45 time: 0.6819 data_time: 0.0422 memory: 11853 grad_norm: 48.4111 loss: 8.3811 loss_cls: 0.6007 loss_bbox: 0.7845 d0.loss_cls: 0.5422 d0.loss_bbox: 0.8999 d1.loss_cls: 0.5580 d1.loss_bbox: 0.8324 d2.loss_cls: 0.5653 d2.loss_bbox: 0.8186 d3.loss_cls: 0.5725 d3.loss_bbox: 0.8173 d4.loss_cls: 0.5875 d4.loss_bbox: 0.8021 +10/10 20:33:47 - mmengine - INFO - Epoch(train) [2][6950/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:50:22 time: 0.6832 data_time: 0.0421 memory: 11853 grad_norm: 56.2914 loss: 9.3668 loss_cls: 0.6948 loss_bbox: 0.8503 d0.loss_cls: 0.6513 d0.loss_bbox: 0.9728 d1.loss_cls: 0.6563 d1.loss_bbox: 0.9035 d2.loss_cls: 0.6686 d2.loss_bbox: 0.8748 d3.loss_cls: 0.6802 d3.loss_bbox: 0.8677 d4.loss_cls: 0.6897 d4.loss_bbox: 0.8568 +10/10 20:33:59 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 20:34:21 - mmengine - INFO - Epoch(train) [2][7000/7033] base_lr: 1.9915e-04 lr: 1.9915e-04 eta: 1 day, 8:48:59 time: 0.6820 data_time: 0.0422 memory: 11853 grad_norm: 51.7973 loss: 9.6038 loss_cls: 0.6797 loss_bbox: 0.9051 d0.loss_cls: 0.6261 d0.loss_bbox: 1.0283 d1.loss_cls: 0.6365 d1.loss_bbox: 0.9516 d2.loss_cls: 0.6534 d2.loss_bbox: 0.9337 d3.loss_cls: 0.6772 d3.loss_bbox: 0.9198 d4.loss_cls: 0.6796 d4.loss_bbox: 0.9129 +10/10 20:34:44 - mmengine - INFO - Exp name: detr3d_r50_bert_gridmask_halfdata_decoder_20231010_173055 +10/10 20:34:44 - mmengine - INFO - Saving checkpoint at 2 epochs +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +Traceback (most recent call last): + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3489, in + main() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3482, in main + globals = debugger.run(setup['file'], None, None, is_module) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2510, in run + return self._exec(is_module, entry_point_fn, module_name, file, globals, locals) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2517, in _exec + globals = pydevd_runpy.run_path(file, globals, '__main__') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path + return _run_module_code(code, init_globals, run_name, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code + _run_code(code, mod_globals, init_globals, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code + exec(code, run_globals) + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 131, in main + runner.train() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 1745, in train + model = self.train_loop.run() # type: ignore + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 102, in run + self.runner.val_loop.run() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 363, in run + self.run_iter(idx, data_batch) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 28, in decorate_context + return func(*args, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 383, in run_iter + outputs = self.runner.model.val_step(data_batch) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/model/wrappers/distributed.py", line 137, in val_step + return self.module.val_step(data) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/model/base_model/base_model.py", line 133, in val_step + return self._run_forward(data, mode='predict') # type: ignore + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/model/base_model/base_model.py", line 340, in _run_forward + results = self(**data, mode=mode) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/mmdet3d/models/detectors/base.py", line 86, in forward + return self.predict(inputs, data_samples, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d.py", line 234, in predict + outs = self.pts_bbox_head(img_feats, batch_input_metas) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) +TypeError: forward() missing 1 required positional argument: 'memory_text' +huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks... +To disable this warning, you can either: + - Avoid using `tokenizers` before the fork if possible + - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false) +ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 1 (pid: 819117) of binary: /home/xzt/miniconda3/envs/detr/bin/python +ERROR:torch.distributed.elastic.agent.server.local_elastic_agent:[default] Worker group failed +INFO:torch.distributed.elastic.agent.server.api:[default] Worker group FAILED. 3/3 attempts left; will restart worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Stopping worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous'ing worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous complete for workers. Result: + restart_count=1 + master_addr=127.0.0.1 + master_port=29500 + group_rank=0 + group_world_size=1 + local_ranks=[0, 1] + role_ranks=[0, 1] + global_ranks=[0, 1] + role_world_sizes=[2, 2] + global_world_sizes=[2, 2] + +INFO:torch.distributed.elastic.agent.server.api:[default] Starting worker group +INFO:torch.distributed.elastic.multiprocessing:Setting worker0 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_1/0/error.json +INFO:torch.distributed.elastic.multiprocessing:Setting worker1 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_1/1/error.json +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +Traceback (most recent call last): + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3489, in + main() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3482, in main + globals = debugger.run(setup['file'], None, None, is_module) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2510, in run + return self._exec(is_module, entry_point_fn, module_name, file, globals, locals) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2517, in _exec + globals = pydevd_runpy.run_path(file, globals, '__main__') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path + return _run_module_code(code, init_globals, run_name, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code + _run_code(code, mod_globals, init_globals, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code + exec(code, run_globals) + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 124, in main + runner = Runner.from_cfg(cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 219, in _store_based_barrier + raise RuntimeError( +RuntimeError: Timed out initializing process group in store based barrier on rank: 0, for key: store_based_barrier_key:1 (world_size=2, worker_count=4, timeout=0:30:00) +Traceback (most recent call last): + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3489, in + main() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3482, in main + globals = debugger.run(setup['file'], None, None, is_module) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2510, in run + return self._exec(is_module, entry_point_fn, module_name, file, globals, locals) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2517, in _exec + globals = pydevd_runpy.run_path(file, globals, '__main__') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path + return _run_module_code(code, init_globals, run_name, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code + _run_code(code, mod_globals, init_globals, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code + exec(code, run_globals) + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 124, in main + runner = Runner.from_cfg(cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 219, in _store_based_barrier + raise RuntimeError( +RuntimeError: Timed out initializing process group in store based barrier on rank: 1, for key: store_based_barrier_key:1 (world_size=2, worker_count=4, timeout=0:30:00) +ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 896145) of binary: /home/xzt/miniconda3/envs/detr/bin/python +ERROR:torch.distributed.elastic.agent.server.local_elastic_agent:[default] Worker group failed +INFO:torch.distributed.elastic.agent.server.api:[default] Worker group FAILED. 2/3 attempts left; will restart worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Stopping worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous'ing worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous complete for workers. Result: + restart_count=2 + master_addr=127.0.0.1 + master_port=29500 + group_rank=0 + group_world_size=1 + local_ranks=[0, 1] + role_ranks=[0, 1] + global_ranks=[0, 1] + role_world_sizes=[2, 2] + global_world_sizes=[2, 2] + +INFO:torch.distributed.elastic.agent.server.api:[default] Starting worker group +INFO:torch.distributed.elastic.multiprocessing:Setting worker0 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_2/0/error.json +INFO:torch.distributed.elastic.multiprocessing:Setting worker1 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_2/1/error.json +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +Traceback (most recent call last): + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3489, in + main() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3482, in main + globals = debugger.run(setup['file'], None, None, is_module) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2510, in run + return self._exec(is_module, entry_point_fn, module_name, file, globals, locals) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2517, in _exec + globals = pydevd_runpy.run_path(file, globals, '__main__') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path + return _run_module_code(code, init_globals, run_name, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code + _run_code(code, mod_globals, init_globals, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code + exec(code, run_globals) + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 124, in main + runner = Runner.from_cfg(cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 219, in _store_based_barrier + raise RuntimeError( +RuntimeError: Timed out initializing process group in store based barrier on rank: 0, for key: store_based_barrier_key:1 (world_size=2, worker_count=6, timeout=0:30:00) +Traceback (most recent call last): + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3489, in + main() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3482, in main + globals = debugger.run(setup['file'], None, None, is_module) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2510, in run + return self._exec(is_module, entry_point_fn, module_name, file, globals, locals) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2517, in _exec + globals = pydevd_runpy.run_path(file, globals, '__main__') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path + return _run_module_code(code, init_globals, run_name, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code + _run_code(code, mod_globals, init_globals, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code + exec(code, run_globals) + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 124, in main + runner = Runner.from_cfg(cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 219, in _store_based_barrier + raise RuntimeError( +RuntimeError: Timed out initializing process group in store based barrier on rank: 1, for key: store_based_barrier_key:1 (world_size=2, worker_count=6, timeout=0:30:00) +ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 911965) of binary: /home/xzt/miniconda3/envs/detr/bin/python +ERROR:torch.distributed.elastic.agent.server.local_elastic_agent:[default] Worker group failed +INFO:torch.distributed.elastic.agent.server.api:[default] Worker group FAILED. 1/3 attempts left; will restart worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Stopping worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous'ing worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous complete for workers. Result: + restart_count=3 + master_addr=127.0.0.1 + master_port=29500 + group_rank=0 + group_world_size=1 + local_ranks=[0, 1] + role_ranks=[0, 1] + global_ranks=[0, 1] + role_world_sizes=[2, 2] + global_world_sizes=[2, 2] + +INFO:torch.distributed.elastic.agent.server.api:[default] Starting worker group +INFO:torch.distributed.elastic.multiprocessing:Setting worker0 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_3/0/error.json +INFO:torch.distributed.elastic.multiprocessing:Setting worker1 reply file to: /tmp/torchelastic_2gr53abd/none_c7j0w9cf/attempt_3/1/error.json +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +Traceback (most recent call last): + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3489, in +Traceback (most recent call last): + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3489, in + main() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3482, in main + main() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 3482, in main + globals = debugger.run(setup['file'], None, None, is_module) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2510, in run + globals = debugger.run(setup['file'], None, None, is_module) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2510, in run + return self._exec(is_module, entry_point_fn, module_name, file, globals, locals) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2517, in _exec + return self._exec(is_module, entry_point_fn, module_name, file, globals, locals) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/pydevd.py", line 2517, in _exec + globals = pydevd_runpy.run_path(file, globals, '__main__') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path + return _run_module_code(code, init_globals, run_name, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code +globals = pydevd_runpy.run_path(file, globals, '__main__') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path + _run_code(code, mod_globals, init_globals, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code + return _run_module_code(code, init_globals, run_name, + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code +exec(code, run_globals) + File "tools/train.py", line 135, in + main()_run_code(code, mod_globals, init_globals, + + File "tools/train.py", line 124, in main + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code + runner = Runner.from_cfg(cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg + exec(code, run_globals) + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 124, in main + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + runner = Runner.from_cfg(cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg + self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist +self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 208, in _store_based_barrier + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + worker_count = store.add(store_key, 0) +RuntimeError: Connection reset by peer + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 208, in _store_based_barrier + worker_count = store.add(store_key, 0) +RuntimeError: Connection reset by peer +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/launch.py:163: DeprecationWarning: The 'warn' method is deprecated, use 'warning' instead + logger.warn( +The module torch.distributed.launch is deprecated and going to be removed in future.Migrate to torch.distributed.run +WARNING:torch.distributed.run:--use_env is deprecated and will be removed in future releases. + Please read local_rank from `os.environ('LOCAL_RANK')` instead. +INFO:torch.distributed.launcher.api:Starting elastic_operator with launch configs: + entrypoint : tools/train.py + min_nodes : 1 + max_nodes : 1 + nproc_per_node : 2 + run_id : none + rdzv_backend : static + rdzv_endpoint : 127.0.0.1:29500 + rdzv_configs : {'rank': 0, 'timeout': 900} + max_restarts : 3 + monitor_interval : 5 + log_dir : None + metrics_cfg : {} + +INFO:torch.distributed.elastic.agent.server.local_elastic_agent:log directory set to: /tmp/torchelastic_9g3to02q/none_u2xm2q1i +INFO:torch.distributed.elastic.agent.server.api:[default] starting workers for entrypoint: python +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous'ing worker group +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/elastic/utils/store.py:52: FutureWarning: This is an experimental API and will be changed in future. + warnings.warn( +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous complete for workers. Result: + restart_count=0 + master_addr=127.0.0.1 + master_port=29500 + group_rank=0 + group_world_size=1 + local_ranks=[0, 1] + role_ranks=[0, 1] + global_ranks=[0, 1] + role_world_sizes=[2, 2] + global_world_sizes=[2, 2] + +INFO:torch.distributed.elastic.agent.server.api:[default] Starting worker group +INFO:torch.distributed.elastic.multiprocessing:Setting worker0 reply file to: /tmp/torchelastic_9g3to02q/none_u2xm2q1i/attempt_0/0/error.json +INFO:torch.distributed.elastic.multiprocessing:Setting worker1 reply file to: /tmp/torchelastic_9g3to02q/none_u2xm2q1i/attempt_0/1/error.json +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +10/10 22:40:16 - mmengine - INFO - +------------------------------------------------------------ +System environment: + sys.platform: linux + Python: 3.8.18 (default, Sep 11 2023, 13:40:15) [GCC 11.2.0] + CUDA available: True + numpy_random_seed: 1132959522 + GPU 0,1: NVIDIA GeForce RTX 4090 + CUDA_HOME: /usr/local/cuda-11.7 + NVCC: Cuda compilation tools, release 11.7, V11.7.64 + GCC: gcc (Ubuntu 9.5.0-3ubuntu1) 9.5.0 + PyTorch: 1.9.0+cu111 + PyTorch compiling details: PyTorch built with: + - GCC 7.3 + - C++ Version: 201402 + - Intel(R) Math Kernel Library Version 2020.0.0 Product Build 20191122 for Intel(R) 64 architecture applications + - Intel(R) MKL-DNN v2.1.2 (Git Hash 98be7e8afa711dc9b66c8ff3504129cb82013cdb) + - OpenMP 201511 (a.k.a. OpenMP 4.5) + - NNPACK is enabled + - CPU capability usage: AVX2 + - CUDA Runtime 11.1 + - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86 + - CuDNN 8.0.5 + - Magma 2.5.2 + - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.1, CUDNN_VERSION=8.0.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.9.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, + + TorchVision: 0.10.0+cu111 + OpenCV: 4.8.0 + MMEngine: 0.8.4 + +Runtime environment: + cudnn_benchmark: False + mp_cfg: {'mp_start_method': 'fork', 'opencv_num_threads': 0} + dist_cfg: {'backend': 'nccl'} + seed: 1132959522 + Distributed launcher: pytorch + Distributed training: True + GPU number: 2 +------------------------------------------------------------ + +10/10 22:40:17 - mmengine - INFO - Config: +backend_args = None +class_names = [ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', +] +custom_imports = dict(imports=[ + 'projects.DETR3D.detr3d', +]) +data_prefix = dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts='') +data_root = 'data/nuscenes/' +dataset_type = 'NuScenesDataset' +default_hooks = dict( + checkpoint=dict( + _scope_='mmdet3d', + interval=1, + max_keep_ckpts=1, + save_last=True, + type='CheckpointHook'), + logger=dict(_scope_='mmdet3d', interval=50, type='LoggerHook'), + param_scheduler=dict(_scope_='mmdet3d', type='ParamSchedulerHook'), + sampler_seed=dict(_scope_='mmdet3d', type='DistSamplerSeedHook'), + timer=dict(_scope_='mmdet3d', type='IterTimerHook'), + visualization=dict(_scope_='mmdet3d', type='Det3DVisualizationHook')) +default_scope = 'mmdet3d' +env_cfg = dict( + cudnn_benchmark=False, + dist_cfg=dict(backend='nccl'), + mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0)) +img_norm_cfg = dict( + bgr_to_rgb=False, mean=[ + 103.53, + 116.28, + 123.675, + ], std=[ + 1.0, + 1.0, + 1.0, + ]) +input_modality = dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False) +lang_model_name = 'bert-base-uncased' +launcher = 'pytorch' +load_from = 'pretrained/fcos3d.pth' +log_level = 'INFO' +log_processor = dict( + _scope_='mmdet3d', by_epoch=True, type='LogProcessor', window_size=50) +metainfo = dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', +]) +model = dict( + data_preprocessor=dict( + bgr_to_rgb=False, + mean=[ + 103.53, + 116.28, + 123.675, + ], + pad_size_divisor=32, + std=[ + 1.0, + 1.0, + 1.0, + ], + type='Det3DDataPreprocessor'), + encoder=dict( + fusion_layer_cfg=dict( + embed_dim=1024, + init_values=0.0001, + l_dim=256, + num_heads=4, + v_dim=256), + layer_cfg=dict( + ffn_cfg=dict( + embed_dims=256, feedforward_channels=2048, ffn_drop=0.0), + self_attn_cfg=dict(dropout=0.0, embed_dims=256, num_levels=4)), + num_cp=6, + num_layers=6, + text_layer_cfg=dict( + ffn_cfg=dict( + embed_dims=256, feedforward_channels=1024, ffn_drop=0.0), + self_attn_cfg=dict(dropout=0.0, embed_dims=256, num_heads=4))), + img_backbone=dict( + dcn=dict(deform_groups=1, fallback_on_stride=False, type='DCNv2'), + depth=50, + frozen_stages=1, + norm_cfg=dict(requires_grad=False, type='BN2d'), + norm_eval=True, + num_stages=4, + out_indices=( + 0, + 1, + 2, + 3, + ), + stage_with_dcn=( + False, + False, + True, + True, + ), + style='caffe', + type='mmdet.ResNet'), + img_neck=dict( + add_extra_convs='on_output', + in_channels=[ + 256, + 512, + 1024, + 2048, + ], + num_outs=4, + out_channels=256, + relu_before_extra_convs=True, + start_level=1, + type='mmdet.FPN'), + language_model=dict( + add_pooling_layer=False, + name='bert-base-uncased', + pad_to_max=False, + special_tokens_list=[ + '[CLS]', + '[SEP]', + '.', + '?', + ], + type='BertModel', + use_sub_sentence_represent=True), + positional_encoding_single=dict( + normalize=True, num_feats=128, offset=0.0, temperature=20), + pts_bbox_head=dict( + as_two_stage=False, + bbox_coder=dict( + max_num=300, + num_classes=10, + pc_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + post_center_range=[ + -61.2, + -61.2, + -10.0, + 61.2, + 61.2, + 10.0, + ], + type='NMSFreeCoder', + voxel_size=[ + 0.2, + 0.2, + 8, + ]), + in_channels=256, + loss_bbox=dict(loss_weight=0.25, type='mmdet.L1Loss'), + loss_cls=dict( + alpha=0.25, + gamma=2.0, + loss_weight=2.0, + type='mmdet.FocalLoss', + use_sigmoid=True), + loss_iou=dict(loss_weight=0.0, type='mmdet.GIoULoss'), + num_classes=10, + num_query=900, + positional_encoding=dict( + normalize=True, + num_feats=128, + offset=-0.5, + type='mmdet.SinePositionalEncoding'), + sync_cls_avg_factor=True, + transformer=dict( + decoder=dict( + num_layers=6, + return_intermediate=True, + transformerlayers=dict( + attn_cfgs=[ + dict( + dropout=0.1, + embed_dims=256, + num_heads=8, + type='MultiheadAttention'), + dict( + dropout=0.0, + embed_dims=256, + num_heads=8, + type='MultiheadAttention'), + dict( + embed_dims=256, + num_points=1, + pc_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + type='Detr3DCrossAtten'), + ], + feedforward_channels=512, + ffn_dropout=0.1, + operation_order=( + 'self_attn', + 'norm', + 'self_attn_text', + 'norm', + 'cross_attn', + 'norm', + 'ffn', + 'norm', + ), + type='mmdet.DetrTransformerDecoderLayer'), + type='Detr3DTransformerDecoder'), + type='Detr3DTransformer'), + type='DETR3DHead', + with_box_refine=True), + train_cfg=dict( + pts=dict( + assigner=dict( + cls_cost=dict(type='mmdet.FocalLossCost', weight=2.0), + iou_cost=dict(type='mmdet.IoUCost', weight=0.0), + pc_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + reg_cost=dict(type='BBox3DL1Cost', weight=0.25), + type='HungarianAssigner3D'), + grid_size=[ + 512, + 512, + 1, + ], + out_size_factor=4, + point_cloud_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + voxel_size=[ + 0.2, + 0.2, + 8, + ])), + type='DETR3D', + use_grid_mask=True) +optim_wrapper = dict( + clip_grad=dict(max_norm=35, norm_type=2), + optimizer=dict(lr=0.0002, type='AdamW', weight_decay=0.01), + paramwise_cfg=dict(custom_keys=dict(img_backbone=dict(lr_mult=0.1))), + type='OptimWrapper') +param_scheduler = [ + dict( + begin=0, + by_epoch=False, + end=500, + start_factor=0.3333333333333333, + type='LinearLR'), + dict( + T_max=24, + begin=0, + by_epoch=True, + end=24, + eta_min_ratio=0.001, + type='CosineAnnealingLR'), +] +point_cloud_range = [ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, +] +resume = False +test_cfg = dict(type='TestLoop') +test_dataloader = dict( + batch_size=1, + dataset=dict( + ann_file='nuscenes_infos_val.pkl', + backend_args=None, + box_type_3d='LiDAR', + data_prefix=dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts=''), + data_root='data/nuscenes/', + load_type='frame_based', + metainfo=dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ]), + modality=dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False), + pipeline=[ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + transforms=[ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict(keys=[ + 'img', + ], type='Pack3DDetInputs'), + ], + test_mode=True, + type='NuScenesDataset'), + drop_last=False, + num_workers=4, + persistent_workers=True, + sampler=dict(shuffle=False, type='DefaultSampler')) +test_evaluator = dict( + ann_file='data/nuscenes/nuscenes_infos_val.pkl', + backend_args=None, + data_root='data/nuscenes/', + metric='bbox', + type='NuScenesMetric') +test_pipeline = [ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + transforms=[ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict(keys=[ + 'img', + ], type='Pack3DDetInputs'), +] +test_transforms = [ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), +] +total_epochs = 24 +train_cfg = dict(max_epochs=24, type='EpochBasedTrainLoop', val_interval=2) +train_dataloader = dict( + batch_size=1, + dataset=dict( + ann_file='nuscenes_infos_train.pkl', + backend_args=None, + box_type_3d='LiDAR', + data_prefix=dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts=''), + data_root='data/nuscenes/', + load_interval=2, + load_type='frame_based', + metainfo=dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ]), + modality=dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False), + pipeline=[ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + type='LoadAnnotations3D', + with_attr_label=False, + with_bbox_3d=True, + with_label_3d=True), + dict( + transforms=[ + dict(type='PhotoMetricDistortion3D'), + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict( + point_cloud_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + type='ObjectRangeFilter'), + dict( + classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ], + type='ObjectNameFilter'), + dict( + keys=[ + 'img', + 'gt_bboxes_3d', + 'gt_labels_3d', + ], + type='Pack3DDetInputs'), + ], + test_mode=False, + type='NuScenesDataset'), + drop_last=False, + num_workers=4, + persistent_workers=True, + sampler=dict(shuffle=True, type='DefaultSampler')) +train_pipeline = [ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + type='LoadAnnotations3D', + with_attr_label=False, + with_bbox_3d=True, + with_label_3d=True), + dict( + transforms=[ + dict(type='PhotoMetricDistortion3D'), + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict( + point_cloud_range=[ + -51.2, + -51.2, + -5.0, + 51.2, + 51.2, + 3.0, + ], + type='ObjectRangeFilter'), + dict( + classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ], + type='ObjectNameFilter'), + dict( + keys=[ + 'img', + 'gt_bboxes_3d', + 'gt_labels_3d', + ], type='Pack3DDetInputs'), +] +train_transforms = [ + dict(type='PhotoMetricDistortion3D'), + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), +] +val_cfg = dict(type='ValLoop') +val_dataloader = dict( + batch_size=1, + dataset=dict( + ann_file='nuscenes_infos_val.pkl', + backend_args=None, + box_type_3d='LiDAR', + data_prefix=dict( + CAM_BACK='samples/CAM_BACK', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + pts=''), + data_root='data/nuscenes/', + load_type='frame_based', + metainfo=dict(classes=[ + 'car', + 'truck', + 'construction_vehicle', + 'bus', + 'trailer', + 'barrier', + 'motorcycle', + 'bicycle', + 'pedestrian', + 'traffic_cone', + ]), + modality=dict( + use_camera=True, + use_external=False, + use_lidar=False, + use_map=False, + use_radar=False), + pipeline=[ + dict( + backend_args=None, + num_views=6, + to_float32=True, + type='LoadMultiViewImageFromFiles'), + dict( + transforms=[ + dict( + keep_ratio=True, + ratio_range=( + 1.0, + 1.0, + ), + scale=( + 1600, + 900, + ), + type='RandomResize3D'), + ], + type='MultiViewWrapper'), + dict(keys=[ + 'img', + ], type='Pack3DDetInputs'), + ], + test_mode=True, + type='NuScenesDataset'), + drop_last=False, + num_workers=4, + persistent_workers=True, + sampler=dict(shuffle=False, type='DefaultSampler')) +val_evaluator = dict( + ann_file='data/nuscenes/nuscenes_infos_val.pkl', + backend_args=None, + data_root='data/nuscenes/', + metric='bbox', + type='NuScenesMetric') +vis_backends = [ + dict(type='TensorboardVisBackend'), +] +visualizer = dict( + name='visualizer', + type='Det3DLocalVisualizer', + vis_backends=[ + dict(type='TensorboardVisBackend'), + ]) +voxel_size = [ + 0.2, + 0.2, + 8, +] +work_dir = './work_dirs/detr3d_r50_bert_gridmask_halfdata_decoder' + +10/10 22:40:25 - mmengine - INFO - Autoplay mode, press [SPACE] to pause. +10/10 22:40:25 - mmengine - INFO - Hooks will be executed in the following order: +before_run: +(VERY_HIGH ) RuntimeInfoHook +(BELOW_NORMAL) LoggerHook + -------------------- +before_train: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(VERY_LOW ) CheckpointHook + -------------------- +before_train_epoch: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(NORMAL ) DistSamplerSeedHook + -------------------- +before_train_iter: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook + -------------------- +after_train_iter: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(BELOW_NORMAL) LoggerHook +(LOW ) ParamSchedulerHook +(VERY_LOW ) CheckpointHook + -------------------- +after_train_epoch: +(NORMAL ) IterTimerHook +(LOW ) ParamSchedulerHook +(VERY_LOW ) CheckpointHook + -------------------- +before_val: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +before_val_epoch: +(NORMAL ) IterTimerHook + -------------------- +before_val_iter: +(NORMAL ) IterTimerHook + -------------------- +after_val_iter: +(NORMAL ) IterTimerHook +(NORMAL ) Det3DVisualizationHook +(BELOW_NORMAL) LoggerHook + -------------------- +after_val_epoch: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(BELOW_NORMAL) LoggerHook +(LOW ) ParamSchedulerHook +(VERY_LOW ) CheckpointHook + -------------------- +after_val: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +after_train: +(VERY_HIGH ) RuntimeInfoHook +(VERY_LOW ) CheckpointHook + -------------------- +before_test: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +before_test_epoch: +(NORMAL ) IterTimerHook + -------------------- +before_test_iter: +(NORMAL ) IterTimerHook + -------------------- +after_test_iter: +(NORMAL ) IterTimerHook +(NORMAL ) Det3DVisualizationHook +(BELOW_NORMAL) LoggerHook + -------------------- +after_test_epoch: +(VERY_HIGH ) RuntimeInfoHook +(NORMAL ) IterTimerHook +(BELOW_NORMAL) LoggerHook + -------------------- +after_test: +(VERY_HIGH ) RuntimeInfoHook + -------------------- +after_run: +(BELOW_NORMAL) LoggerHook + -------------------- +10/10 22:42:01 - mmengine - INFO - ------------------------------ +10/10 22:42:01 - mmengine - INFO - The length of the dataset: 14065 +10/10 22:42:01 - mmengine - INFO - The number of instances per category in the dataset: ++----------------------+--------+ +| category | number | ++----------------------+--------+ +| car | 206689 | +| truck | 36393 | +| construction_vehicle | 5990 | +| bus | 6576 | +| trailer | 10349 | +| barrier | 62576 | +| motorcycle | 5051 | +| bicycle | 4740 | +| pedestrian | 92925 | +| traffic_cone | 41120 | ++----------------------+--------+ +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.downsample.0.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.downsample.0.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.0.downsample.0.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.1.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.2.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer2.3.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.downsample.0.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.downsample.0.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.0.downsample.0.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.1.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.2.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.3.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.4.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer3.5.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.downsample.0.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.downsample.0.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.0.downsample.0.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.1.conv3.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv1.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv1.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv1.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.weight:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.bias:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.bias:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv2.conv_offset.bias:lr_mult=0.1 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv3.weight:lr=2e-05 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv3.weight:weight_decay=0.01 +10/10 22:42:01 - mmengine - INFO - paramwise_options -- img_backbone.layer4.2.conv3.weight:lr_mult=0.1 +10/10 22:42:26 - mmengine - INFO - ------------------------------ +10/10 22:42:26 - mmengine - INFO - The length of the dataset: 6019 +10/10 22:42:26 - mmengine - INFO - The number of instances per category in the dataset: ++----------------------+--------+ +| category | number | ++----------------------+--------+ +| car | 80004 | +| truck | 15704 | +| construction_vehicle | 2678 | +| bus | 3158 | +| trailer | 4159 | +| barrier | 26992 | +| motorcycle | 2508 | +| bicycle | 2381 | +| pedestrian | 34347 | +| traffic_cone | 15597 | ++----------------------+--------+ +/home/xzt/mmdetection3d-1.1.0/mmdet3d/evaluation/functional/kitti_utils/eval.py:10: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details. + def get_thresholds(scores: np.ndarray, num_gt, num_sample_pts=41): +Loads checkpoint by local backend from path: pretrained/fcos3d.pth +/home/xzt/mmdetection3d-1.1.0/mmdet3d/evaluation/functional/kitti_utils/eval.py:10: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/deprecation.html#deprecation-of-object-mode-fall-back-behaviour-when-using-jit for details. + def get_thresholds(scores: np.ndarray, num_gt, num_sample_pts=41): +Loads checkpoint by local backend from path: pretrained/fcos3d.pth +The model and loaded state dict do not match exactly + +unexpected key in source state_dict: bbox_head.bld_alpha, bbox_head.cls_convs.0.conv.weight, bbox_head.cls_convs.0.conv.bias, bbox_head.cls_convs.0.gn.weight, bbox_head.cls_convs.0.gn.bias, bbox_head.cls_convs.1.conv.weight, bbox_head.cls_convs.1.conv.bias, bbox_head.cls_convs.1.conv.conv_offset.weight, bbox_head.cls_convs.1.conv.conv_offset.bias, bbox_head.cls_convs.1.gn.weight, bbox_head.cls_convs.1.gn.bias, bbox_head.reg_convs.0.conv.weight, bbox_head.reg_convs.0.conv.bias, bbox_head.reg_convs.0.gn.weight, bbox_head.reg_convs.0.gn.bias, bbox_head.reg_convs.1.conv.weight, bbox_head.reg_convs.1.conv.bias, bbox_head.reg_convs.1.conv.conv_offset.weight, bbox_head.reg_convs.1.conv.conv_offset.bias, bbox_head.reg_convs.1.gn.weight, bbox_head.reg_convs.1.gn.bias, bbox_head.conv_cls_prev.0.conv.weight, bbox_head.conv_cls_prev.0.conv.bias, bbox_head.conv_cls_prev.0.gn.weight, bbox_head.conv_cls_prev.0.gn.bias, bbox_head.conv_cls.weight, bbox_head.conv_cls.bias, bbox_head.conv_reg_prevs.0.0.conv.weight, bbox_head.conv_reg_prevs.0.0.conv.bias, bbox_head.conv_reg_prevs.0.0.gn.weight, bbox_head.conv_reg_prevs.0.0.gn.bias, bbox_head.conv_reg_prevs.1.0.conv.weight, bbox_head.conv_reg_prevs.1.0.conv.bias, bbox_head.conv_reg_prevs.1.0.gn.weight, bbox_head.conv_reg_prevs.1.0.gn.bias, bbox_head.conv_reg_prevs.2.0.conv.weight, bbox_head.conv_reg_prevs.2.0.conv.bias, bbox_head.conv_reg_prevs.2.0.gn.weight, bbox_head.conv_reg_prevs.2.0.gn.bias, bbox_head.conv_reg_prevs.3.0.conv.weight, bbox_head.conv_reg_prevs.3.0.conv.bias, bbox_head.conv_reg_prevs.3.0.gn.weight, bbox_head.conv_reg_prevs.3.0.gn.bias, bbox_head.conv_reg_prevs.5.0.conv.weight, bbox_head.conv_reg_prevs.5.0.conv.bias, bbox_head.conv_reg_prevs.5.0.gn.weight, bbox_head.conv_reg_prevs.5.0.gn.bias, bbox_head.conv_regs.0.weight, bbox_head.conv_regs.0.bias, bbox_head.conv_regs.1.weight, bbox_head.conv_regs.1.bias, bbox_head.conv_regs.2.weight, bbox_head.conv_regs.2.bias, bbox_head.conv_regs.3.weight, bbox_head.conv_regs.3.bias, bbox_head.conv_regs.4.weight, bbox_head.conv_regs.4.bias, bbox_head.conv_regs.5.weight, bbox_head.conv_regs.5.bias, bbox_head.conv_dir_cls_prev.0.conv.weight, bbox_head.conv_dir_cls_prev.0.conv.bias, bbox_head.conv_dir_cls_prev.0.gn.weight, bbox_head.conv_dir_cls_prev.0.gn.bias, bbox_head.conv_dir_cls.weight, bbox_head.conv_dir_cls.bias, bbox_head.conv_attr_prev.0.conv.weight, bbox_head.conv_attr_prev.0.conv.bias, bbox_head.conv_attr_prev.0.gn.weight, bbox_head.conv_attr_prev.0.gn.bias, bbox_head.conv_attr.weight, bbox_head.conv_attr.bias, bbox_head.conv_depth_cls_prev.0.conv.weight, bbox_head.conv_depth_cls_prev.0.conv.bias, bbox_head.conv_depth_cls_prev.0.gn.weight, bbox_head.conv_depth_cls_prev.0.gn.bias, bbox_head.conv_depth_cls.weight, bbox_head.conv_depth_cls.bias, bbox_head.conv_centerness_prev.0.conv.weight, bbox_head.conv_centerness_prev.0.conv.bias, bbox_head.conv_centerness_prev.0.gn.weight, bbox_head.conv_centerness_prev.0.gn.bias, bbox_head.conv_centerness.weight, bbox_head.conv_centerness.bias, bbox_head.scales.0.0.scale, bbox_head.scales.0.1.scale, bbox_head.scales.0.2.scale, bbox_head.scales.0.3.scale, bbox_head.scales.1.0.scale, bbox_head.scales.1.1.scale, bbox_head.scales.1.2.scale, bbox_head.scales.1.3.scale, bbox_head.scales.2.0.scale, bbox_head.scales.2.1.scale, bbox_head.scales.2.2.scale, bbox_head.scales.2.3.scale, bbox_head.scales.3.0.scale, bbox_head.scales.3.1.scale, bbox_head.scales.3.2.scale, bbox_head.scales.3.3.scale, bbox_head.scales.4.0.scale, bbox_head.scales.4.1.scale, bbox_head.scales.4.2.scale, bbox_head.scales.4.3.scale, img_backbone.layer3.6.conv1.weight, img_backbone.layer3.6.bn1.weight, img_backbone.layer3.6.bn1.bias, img_backbone.layer3.6.bn1.running_mean, img_backbone.layer3.6.bn1.running_var, img_backbone.layer3.6.bn1.num_batches_tracked, img_backbone.layer3.6.conv2.weight, img_backbone.layer3.6.conv2.conv_offset.weight, img_backbone.layer3.6.conv2.conv_offset.bias, img_backbone.layer3.6.bn2.weight, img_backbone.layer3.6.bn2.bias, img_backbone.layer3.6.bn2.running_mean, img_backbone.layer3.6.bn2.running_var, img_backbone.layer3.6.bn2.num_batches_tracked, img_backbone.layer3.6.conv3.weight, img_backbone.layer3.6.bn3.weight, img_backbone.layer3.6.bn3.bias, img_backbone.layer3.6.bn3.running_mean, img_backbone.layer3.6.bn3.running_var, img_backbone.layer3.6.bn3.num_batches_tracked, img_backbone.layer3.7.conv1.weight, img_backbone.layer3.7.bn1.weight, img_backbone.layer3.7.bn1.bias, img_backbone.layer3.7.bn1.running_mean, img_backbone.layer3.7.bn1.running_var, img_backbone.layer3.7.bn1.num_batches_tracked, img_backbone.layer3.7.conv2.weight, img_backbone.layer3.7.conv2.conv_offset.weight, img_backbone.layer3.7.conv2.conv_offset.bias, img_backbone.layer3.7.bn2.weight, img_backbone.layer3.7.bn2.bias, img_backbone.layer3.7.bn2.running_mean, img_backbone.layer3.7.bn2.running_var, img_backbone.layer3.7.bn2.num_batches_tracked, img_backbone.layer3.7.conv3.weight, img_backbone.layer3.7.bn3.weight, img_backbone.layer3.7.bn3.bias, img_backbone.layer3.7.bn3.running_mean, img_backbone.layer3.7.bn3.running_var, img_backbone.layer3.7.bn3.num_batches_tracked, img_backbone.layer3.8.conv1.weight, img_backbone.layer3.8.bn1.weight, img_backbone.layer3.8.bn1.bias, img_backbone.layer3.8.bn1.running_mean, img_backbone.layer3.8.bn1.running_var, img_backbone.layer3.8.bn1.num_batches_tracked, img_backbone.layer3.8.conv2.weight, img_backbone.layer3.8.conv2.conv_offset.weight, img_backbone.layer3.8.conv2.conv_offset.bias, img_backbone.layer3.8.bn2.weight, img_backbone.layer3.8.bn2.bias, img_backbone.layer3.8.bn2.running_mean, img_backbone.layer3.8.bn2.running_var, img_backbone.layer3.8.bn2.num_batches_tracked, img_backbone.layer3.8.conv3.weight, img_backbone.layer3.8.bn3.weight, img_backbone.layer3.8.bn3.bias, img_backbone.layer3.8.bn3.running_mean, img_backbone.layer3.8.bn3.running_var, img_backbone.layer3.8.bn3.num_batches_tracked, img_backbone.layer3.9.conv1.weight, img_backbone.layer3.9.bn1.weight, img_backbone.layer3.9.bn1.bias, img_backbone.layer3.9.bn1.running_mean, img_backbone.layer3.9.bn1.running_var, img_backbone.layer3.9.bn1.num_batches_tracked, img_backbone.layer3.9.conv2.weight, img_backbone.layer3.9.conv2.conv_offset.weight, img_backbone.layer3.9.conv2.conv_offset.bias, img_backbone.layer3.9.bn2.weight, img_backbone.layer3.9.bn2.bias, img_backbone.layer3.9.bn2.running_mean, img_backbone.layer3.9.bn2.running_var, img_backbone.layer3.9.bn2.num_batches_tracked, img_backbone.layer3.9.conv3.weight, img_backbone.layer3.9.bn3.weight, img_backbone.layer3.9.bn3.bias, img_backbone.layer3.9.bn3.running_mean, img_backbone.layer3.9.bn3.running_var, img_backbone.layer3.9.bn3.num_batches_tracked, img_backbone.layer3.10.conv1.weight, img_backbone.layer3.10.bn1.weight, img_backbone.layer3.10.bn1.bias, img_backbone.layer3.10.bn1.running_mean, img_backbone.layer3.10.bn1.running_var, img_backbone.layer3.10.bn1.num_batches_tracked, img_backbone.layer3.10.conv2.weight, img_backbone.layer3.10.conv2.conv_offset.weight, img_backbone.layer3.10.conv2.conv_offset.bias, img_backbone.layer3.10.bn2.weight, img_backbone.layer3.10.bn2.bias, img_backbone.layer3.10.bn2.running_mean, img_backbone.layer3.10.bn2.running_var, img_backbone.layer3.10.bn2.num_batches_tracked, img_backbone.layer3.10.conv3.weight, img_backbone.layer3.10.bn3.weight, img_backbone.layer3.10.bn3.bias, img_backbone.layer3.10.bn3.running_mean, img_backbone.layer3.10.bn3.running_var, img_backbone.layer3.10.bn3.num_batches_tracked, img_backbone.layer3.11.conv1.weight, img_backbone.layer3.11.bn1.weight, img_backbone.layer3.11.bn1.bias, img_backbone.layer3.11.bn1.running_mean, img_backbone.layer3.11.bn1.running_var, img_backbone.layer3.11.bn1.num_batches_tracked, img_backbone.layer3.11.conv2.weight, img_backbone.layer3.11.conv2.conv_offset.weight, img_backbone.layer3.11.conv2.conv_offset.bias, img_backbone.layer3.11.bn2.weight, img_backbone.layer3.11.bn2.bias, img_backbone.layer3.11.bn2.running_mean, img_backbone.layer3.11.bn2.running_var, img_backbone.layer3.11.bn2.num_batches_tracked, img_backbone.layer3.11.conv3.weight, img_backbone.layer3.11.bn3.weight, img_backbone.layer3.11.bn3.bias, img_backbone.layer3.11.bn3.running_mean, img_backbone.layer3.11.bn3.running_var, img_backbone.layer3.11.bn3.num_batches_tracked, img_backbone.layer3.12.conv1.weight, img_backbone.layer3.12.bn1.weight, img_backbone.layer3.12.bn1.bias, img_backbone.layer3.12.bn1.running_mean, img_backbone.layer3.12.bn1.running_var, img_backbone.layer3.12.bn1.num_batches_tracked, img_backbone.layer3.12.conv2.weight, img_backbone.layer3.12.conv2.conv_offset.weight, img_backbone.layer3.12.conv2.conv_offset.bias, img_backbone.layer3.12.bn2.weight, img_backbone.layer3.12.bn2.bias, img_backbone.layer3.12.bn2.running_mean, img_backbone.layer3.12.bn2.running_var, img_backbone.layer3.12.bn2.num_batches_tracked, img_backbone.layer3.12.conv3.weight, img_backbone.layer3.12.bn3.weight, img_backbone.layer3.12.bn3.bias, img_backbone.layer3.12.bn3.running_mean, img_backbone.layer3.12.bn3.running_var, img_backbone.layer3.12.bn3.num_batches_tracked, img_backbone.layer3.13.conv1.weight, img_backbone.layer3.13.bn1.weight, img_backbone.layer3.13.bn1.bias, img_backbone.layer3.13.bn1.running_mean, img_backbone.layer3.13.bn1.running_var, img_backbone.layer3.13.bn1.num_batches_tracked, img_backbone.layer3.13.conv2.weight, img_backbone.layer3.13.conv2.conv_offset.weight, img_backbone.layer3.13.conv2.conv_offset.bias, img_backbone.layer3.13.bn2.weight, img_backbone.layer3.13.bn2.bias, img_backbone.layer3.13.bn2.running_mean, img_backbone.layer3.13.bn2.running_var, img_backbone.layer3.13.bn2.num_batches_tracked, img_backbone.layer3.13.conv3.weight, img_backbone.layer3.13.bn3.weight, img_backbone.layer3.13.bn3.bias, img_backbone.layer3.13.bn3.running_mean, img_backbone.layer3.13.bn3.running_var, img_backbone.layer3.13.bn3.num_batches_tracked, img_backbone.layer3.14.conv1.weight, img_backbone.layer3.14.bn1.weight, img_backbone.layer3.14.bn1.bias, img_backbone.layer3.14.bn1.running_mean, img_backbone.layer3.14.bn1.running_var, img_backbone.layer3.14.bn1.num_batches_tracked, img_backbone.layer3.14.conv2.weight, img_backbone.layer3.14.conv2.conv_offset.weight, img_backbone.layer3.14.conv2.conv_offset.bias, img_backbone.layer3.14.bn2.weight, img_backbone.layer3.14.bn2.bias, img_backbone.layer3.14.bn2.running_mean, img_backbone.layer3.14.bn2.running_var, img_backbone.layer3.14.bn2.num_batches_tracked, img_backbone.layer3.14.conv3.weight, img_backbone.layer3.14.bn3.weight, img_backbone.layer3.14.bn3.bias, img_backbone.layer3.14.bn3.running_mean, img_backbone.layer3.14.bn3.running_var, img_backbone.layer3.14.bn3.num_batches_tracked, img_backbone.layer3.15.conv1.weight, img_backbone.layer3.15.bn1.weight, img_backbone.layer3.15.bn1.bias, img_backbone.layer3.15.bn1.running_mean, img_backbone.layer3.15.bn1.running_var, img_backbone.layer3.15.bn1.num_batches_tracked, img_backbone.layer3.15.conv2.weight, img_backbone.layer3.15.conv2.conv_offset.weight, img_backbone.layer3.15.conv2.conv_offset.bias, img_backbone.layer3.15.bn2.weight, img_backbone.layer3.15.bn2.bias, img_backbone.layer3.15.bn2.running_mean, img_backbone.layer3.15.bn2.running_var, img_backbone.layer3.15.bn2.num_batches_tracked, img_backbone.layer3.15.conv3.weight, img_backbone.layer3.15.bn3.weight, img_backbone.layer3.15.bn3.bias, img_backbone.layer3.15.bn3.running_mean, img_backbone.layer3.15.bn3.running_var, img_backbone.layer3.15.bn3.num_batches_tracked, img_backbone.layer3.16.conv1.weight, img_backbone.layer3.16.bn1.weight, img_backbone.layer3.16.bn1.bias, img_backbone.layer3.16.bn1.running_mean, img_backbone.layer3.16.bn1.running_var, img_backbone.layer3.16.bn1.num_batches_tracked, img_backbone.layer3.16.conv2.weight, img_backbone.layer3.16.conv2.conv_offset.weight, img_backbone.layer3.16.conv2.conv_offset.bias, img_backbone.layer3.16.bn2.weight, img_backbone.layer3.16.bn2.bias, img_backbone.layer3.16.bn2.running_mean, img_backbone.layer3.16.bn2.running_var, img_backbone.layer3.16.bn2.num_batches_tracked, img_backbone.layer3.16.conv3.weight, img_backbone.layer3.16.bn3.weight, img_backbone.layer3.16.bn3.bias, img_backbone.layer3.16.bn3.running_mean, img_backbone.layer3.16.bn3.running_var, img_backbone.layer3.16.bn3.num_batches_tracked, img_backbone.layer3.17.conv1.weight, img_backbone.layer3.17.bn1.weight, img_backbone.layer3.17.bn1.bias, img_backbone.layer3.17.bn1.running_mean, img_backbone.layer3.17.bn1.running_var, img_backbone.layer3.17.bn1.num_batches_tracked, img_backbone.layer3.17.conv2.weight, img_backbone.layer3.17.conv2.conv_offset.weight, img_backbone.layer3.17.conv2.conv_offset.bias, img_backbone.layer3.17.bn2.weight, img_backbone.layer3.17.bn2.bias, img_backbone.layer3.17.bn2.running_mean, img_backbone.layer3.17.bn2.running_var, img_backbone.layer3.17.bn2.num_batches_tracked, img_backbone.layer3.17.conv3.weight, img_backbone.layer3.17.bn3.weight, img_backbone.layer3.17.bn3.bias, img_backbone.layer3.17.bn3.running_mean, img_backbone.layer3.17.bn3.running_var, img_backbone.layer3.17.bn3.num_batches_tracked, img_backbone.layer3.18.conv1.weight, img_backbone.layer3.18.bn1.weight, img_backbone.layer3.18.bn1.bias, img_backbone.layer3.18.bn1.running_mean, img_backbone.layer3.18.bn1.running_var, img_backbone.layer3.18.bn1.num_batches_tracked, img_backbone.layer3.18.conv2.weight, img_backbone.layer3.18.conv2.conv_offset.weight, img_backbone.layer3.18.conv2.conv_offset.bias, img_backbone.layer3.18.bn2.weight, img_backbone.layer3.18.bn2.bias, img_backbone.layer3.18.bn2.running_mean, img_backbone.layer3.18.bn2.running_var, img_backbone.layer3.18.bn2.num_batches_tracked, img_backbone.layer3.18.conv3.weight, img_backbone.layer3.18.bn3.weight, img_backbone.layer3.18.bn3.bias, img_backbone.layer3.18.bn3.running_mean, img_backbone.layer3.18.bn3.running_var, img_backbone.layer3.18.bn3.num_batches_tracked, img_backbone.layer3.19.conv1.weight, img_backbone.layer3.19.bn1.weight, img_backbone.layer3.19.bn1.bias, img_backbone.layer3.19.bn1.running_mean, img_backbone.layer3.19.bn1.running_var, img_backbone.layer3.19.bn1.num_batches_tracked, img_backbone.layer3.19.conv2.weight, img_backbone.layer3.19.conv2.conv_offset.weight, img_backbone.layer3.19.conv2.conv_offset.bias, img_backbone.layer3.19.bn2.weight, img_backbone.layer3.19.bn2.bias, img_backbone.layer3.19.bn2.running_mean, img_backbone.layer3.19.bn2.running_var, img_backbone.layer3.19.bn2.num_batches_tracked, img_backbone.layer3.19.conv3.weight, img_backbone.layer3.19.bn3.weight, img_backbone.layer3.19.bn3.bias, img_backbone.layer3.19.bn3.running_mean, img_backbone.layer3.19.bn3.running_var, img_backbone.layer3.19.bn3.num_batches_tracked, img_backbone.layer3.20.conv1.weight, img_backbone.layer3.20.bn1.weight, img_backbone.layer3.20.bn1.bias, img_backbone.layer3.20.bn1.running_mean, img_backbone.layer3.20.bn1.running_var, img_backbone.layer3.20.bn1.num_batches_tracked, img_backbone.layer3.20.conv2.weight, img_backbone.layer3.20.conv2.conv_offset.weight, img_backbone.layer3.20.conv2.conv_offset.bias, img_backbone.layer3.20.bn2.weight, img_backbone.layer3.20.bn2.bias, img_backbone.layer3.20.bn2.running_mean, img_backbone.layer3.20.bn2.running_var, img_backbone.layer3.20.bn2.num_batches_tracked, img_backbone.layer3.20.conv3.weight, img_backbone.layer3.20.bn3.weight, img_backbone.layer3.20.bn3.bias, img_backbone.layer3.20.bn3.running_mean, img_backbone.layer3.20.bn3.running_var, img_backbone.layer3.20.bn3.num_batches_tracked, img_backbone.layer3.21.conv1.weight, img_backbone.layer3.21.bn1.weight, img_backbone.layer3.21.bn1.bias, img_backbone.layer3.21.bn1.running_mean, img_backbone.layer3.21.bn1.running_var, img_backbone.layer3.21.bn1.num_batches_tracked, img_backbone.layer3.21.conv2.weight, img_backbone.layer3.21.conv2.conv_offset.weight, img_backbone.layer3.21.conv2.conv_offset.bias, img_backbone.layer3.21.bn2.weight, img_backbone.layer3.21.bn2.bias, img_backbone.layer3.21.bn2.running_mean, img_backbone.layer3.21.bn2.running_var, img_backbone.layer3.21.bn2.num_batches_tracked, img_backbone.layer3.21.conv3.weight, img_backbone.layer3.21.bn3.weight, img_backbone.layer3.21.bn3.bias, img_backbone.layer3.21.bn3.running_mean, img_backbone.layer3.21.bn3.running_var, img_backbone.layer3.21.bn3.num_batches_tracked, img_backbone.layer3.22.conv1.weight, img_backbone.layer3.22.bn1.weight, img_backbone.layer3.22.bn1.bias, img_backbone.layer3.22.bn1.running_mean, img_backbone.layer3.22.bn1.running_var, img_backbone.layer3.22.bn1.num_batches_tracked, img_backbone.layer3.22.conv2.weight, img_backbone.layer3.22.conv2.conv_offset.weight, img_backbone.layer3.22.conv2.conv_offset.bias, img_backbone.layer3.22.bn2.weight, img_backbone.layer3.22.bn2.bias, img_backbone.layer3.22.bn2.running_mean, img_backbone.layer3.22.bn2.running_var, img_backbone.layer3.22.bn2.num_batches_tracked, img_backbone.layer3.22.conv3.weight, img_backbone.layer3.22.bn3.weight, img_backbone.layer3.22.bn3.bias, img_backbone.layer3.22.bn3.running_mean, img_backbone.layer3.22.bn3.running_var, img_backbone.layer3.22.bn3.num_batches_tracked, img_neck.fpn_convs.4.conv.weight, img_neck.fpn_convs.4.conv.bias + +missing keys in source state_dict: pts_bbox_head.code_weights, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.0.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.0.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.0.norms.0.weight, pts_bbox_head.transformer.decoder.layers.0.norms.0.bias, pts_bbox_head.transformer.decoder.layers.0.norms.1.weight, pts_bbox_head.transformer.decoder.layers.0.norms.1.bias, pts_bbox_head.transformer.decoder.layers.0.norms.2.weight, pts_bbox_head.transformer.decoder.layers.0.norms.2.bias, pts_bbox_head.transformer.decoder.layers.0.norms.3.weight, pts_bbox_head.transformer.decoder.layers.0.norms.3.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.1.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.1.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.1.norms.0.weight, pts_bbox_head.transformer.decoder.layers.1.norms.0.bias, pts_bbox_head.transformer.decoder.layers.1.norms.1.weight, pts_bbox_head.transformer.decoder.layers.1.norms.1.bias, pts_bbox_head.transformer.decoder.layers.1.norms.2.weight, pts_bbox_head.transformer.decoder.layers.1.norms.2.bias, pts_bbox_head.transformer.decoder.layers.1.norms.3.weight, pts_bbox_head.transformer.decoder.layers.1.norms.3.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.2.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.2.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.2.norms.0.weight, pts_bbox_head.transformer.decoder.layers.2.norms.0.bias, pts_bbox_head.transformer.decoder.layers.2.norms.1.weight, pts_bbox_head.transformer.decoder.layers.2.norms.1.bias, pts_bbox_head.transformer.decoder.layers.2.norms.2.weight, pts_bbox_head.transformer.decoder.layers.2.norms.2.bias, pts_bbox_head.transformer.decoder.layers.2.norms.3.weight, pts_bbox_head.transformer.decoder.layers.2.norms.3.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.3.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.3.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.3.norms.0.weight, pts_bbox_head.transformer.decoder.layers.3.norms.0.bias, pts_bbox_head.transformer.decoder.layers.3.norms.1.weight, pts_bbox_head.transformer.decoder.layers.3.norms.1.bias, pts_bbox_head.transformer.decoder.layers.3.norms.2.weight, pts_bbox_head.transformer.decoder.layers.3.norms.2.bias, pts_bbox_head.transformer.decoder.layers.3.norms.3.weight, pts_bbox_head.transformer.decoder.layers.3.norms.3.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.4.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.4.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.4.norms.0.weight, pts_bbox_head.transformer.decoder.layers.4.norms.0.bias, pts_bbox_head.transformer.decoder.layers.4.norms.1.weight, pts_bbox_head.transformer.decoder.layers.4.norms.1.bias, pts_bbox_head.transformer.decoder.layers.4.norms.2.weight, pts_bbox_head.transformer.decoder.layers.4.norms.2.bias, pts_bbox_head.transformer.decoder.layers.4.norms.3.weight, pts_bbox_head.transformer.decoder.layers.4.norms.3.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.0.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.in_proj_weight, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.in_proj_bias, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.out_proj.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.1.attn.out_proj.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.attention_weights.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.attention_weights.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.output_proj.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.output_proj.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.0.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.0.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.1.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.1.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.3.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.3.bias, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.4.weight, pts_bbox_head.transformer.decoder.layers.5.attentions.2.position_encoder.4.bias, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.0.0.weight, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.0.0.bias, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.1.weight, pts_bbox_head.transformer.decoder.layers.5.ffns.0.layers.1.bias, pts_bbox_head.transformer.decoder.layers.5.norms.0.weight, pts_bbox_head.transformer.decoder.layers.5.norms.0.bias, pts_bbox_head.transformer.decoder.layers.5.norms.1.weight, pts_bbox_head.transformer.decoder.layers.5.norms.1.bias, pts_bbox_head.transformer.decoder.layers.5.norms.2.weight, pts_bbox_head.transformer.decoder.layers.5.norms.2.bias, pts_bbox_head.transformer.decoder.layers.5.norms.3.weight, pts_bbox_head.transformer.decoder.layers.5.norms.3.bias, pts_bbox_head.transformer.reference_points.weight, pts_bbox_head.transformer.reference_points.bias, pts_bbox_head.cls_branches.0.0.weight, pts_bbox_head.cls_branches.0.0.bias, pts_bbox_head.cls_branches.0.1.weight, pts_bbox_head.cls_branches.0.1.bias, pts_bbox_head.cls_branches.0.3.weight, pts_bbox_head.cls_branches.0.3.bias, pts_bbox_head.cls_branches.0.4.weight, pts_bbox_head.cls_branches.0.4.bias, pts_bbox_head.cls_branches.0.6.weight, pts_bbox_head.cls_branches.0.6.bias, pts_bbox_head.cls_branches.1.0.weight, pts_bbox_head.cls_branches.1.0.bias, pts_bbox_head.cls_branches.1.1.weight, pts_bbox_head.cls_branches.1.1.bias, pts_bbox_head.cls_branches.1.3.weight, pts_bbox_head.cls_branches.1.3.bias, pts_bbox_head.cls_branches.1.4.weight, pts_bbox_head.cls_branches.1.4.bias, pts_bbox_head.cls_branches.1.6.weight, pts_bbox_head.cls_branches.1.6.bias, pts_bbox_head.cls_branches.2.0.weight, pts_bbox_head.cls_branches.2.0.bias, pts_bbox_head.cls_branches.2.1.weight, pts_bbox_head.cls_branches.2.1.bias, pts_bbox_head.cls_branches.2.3.weight, pts_bbox_head.cls_branches.2.3.bias, pts_bbox_head.cls_branches.2.4.weight, pts_bbox_head.cls_branches.2.4.bias, pts_bbox_head.cls_branches.2.6.weight, pts_bbox_head.cls_branches.2.6.bias, pts_bbox_head.cls_branches.3.0.weight, pts_bbox_head.cls_branches.3.0.bias, pts_bbox_head.cls_branches.3.1.weight, pts_bbox_head.cls_branches.3.1.bias, pts_bbox_head.cls_branches.3.3.weight, pts_bbox_head.cls_branches.3.3.bias, pts_bbox_head.cls_branches.3.4.weight, pts_bbox_head.cls_branches.3.4.bias, pts_bbox_head.cls_branches.3.6.weight, pts_bbox_head.cls_branches.3.6.bias, pts_bbox_head.cls_branches.4.0.weight, pts_bbox_head.cls_branches.4.0.bias, pts_bbox_head.cls_branches.4.1.weight, pts_bbox_head.cls_branches.4.1.bias, pts_bbox_head.cls_branches.4.3.weight, pts_bbox_head.cls_branches.4.3.bias, pts_bbox_head.cls_branches.4.4.weight, pts_bbox_head.cls_branches.4.4.bias, pts_bbox_head.cls_branches.4.6.weight, pts_bbox_head.cls_branches.4.6.bias, pts_bbox_head.cls_branches.5.0.weight, pts_bbox_head.cls_branches.5.0.bias, pts_bbox_head.cls_branches.5.1.weight, pts_bbox_head.cls_branches.5.1.bias, pts_bbox_head.cls_branches.5.3.weight, pts_bbox_head.cls_branches.5.3.bias, pts_bbox_head.cls_branches.5.4.weight, pts_bbox_head.cls_branches.5.4.bias, pts_bbox_head.cls_branches.5.6.weight, pts_bbox_head.cls_branches.5.6.bias, pts_bbox_head.reg_branches.0.0.weight, pts_bbox_head.reg_branches.0.0.bias, pts_bbox_head.reg_branches.0.2.weight, pts_bbox_head.reg_branches.0.2.bias, pts_bbox_head.reg_branches.0.4.weight, pts_bbox_head.reg_branches.0.4.bias, pts_bbox_head.reg_branches.1.0.weight, pts_bbox_head.reg_branches.1.0.bias, pts_bbox_head.reg_branches.1.2.weight, pts_bbox_head.reg_branches.1.2.bias, pts_bbox_head.reg_branches.1.4.weight, pts_bbox_head.reg_branches.1.4.bias, pts_bbox_head.reg_branches.2.0.weight, pts_bbox_head.reg_branches.2.0.bias, pts_bbox_head.reg_branches.2.2.weight, pts_bbox_head.reg_branches.2.2.bias, pts_bbox_head.reg_branches.2.4.weight, pts_bbox_head.reg_branches.2.4.bias, pts_bbox_head.reg_branches.3.0.weight, pts_bbox_head.reg_branches.3.0.bias, pts_bbox_head.reg_branches.3.2.weight, pts_bbox_head.reg_branches.3.2.bias, pts_bbox_head.reg_branches.3.4.weight, pts_bbox_head.reg_branches.3.4.bias, pts_bbox_head.reg_branches.4.0.weight, pts_bbox_head.reg_branches.4.0.bias, pts_bbox_head.reg_branches.4.2.weight, pts_bbox_head.reg_branches.4.2.bias, pts_bbox_head.reg_branches.4.4.weight, pts_bbox_head.reg_branches.4.4.bias, pts_bbox_head.reg_branches.5.0.weight, pts_bbox_head.reg_branches.5.0.bias, pts_bbox_head.reg_branches.5.2.weight, pts_bbox_head.reg_branches.5.2.bias, pts_bbox_head.reg_branches.5.4.weight, pts_bbox_head.reg_branches.5.4.bias, pts_bbox_head.query_embedding.weight, language_model.language_backbone.body.model.embeddings.word_embeddings.weight, language_model.language_backbone.body.model.embeddings.position_embeddings.weight, language_model.language_backbone.body.model.embeddings.token_type_embeddings.weight, language_model.language_backbone.body.model.embeddings.LayerNorm.weight, language_model.language_backbone.body.model.embeddings.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.0.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.0.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.0.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.0.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.0.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.0.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.0.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.0.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.1.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.1.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.1.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.1.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.1.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.1.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.1.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.1.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.2.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.2.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.2.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.2.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.2.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.2.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.2.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.2.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.3.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.3.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.3.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.3.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.3.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.3.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.3.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.3.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.4.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.4.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.4.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.4.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.4.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.4.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.4.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.4.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.5.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.5.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.5.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.5.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.5.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.5.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.5.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.5.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.6.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.6.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.6.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.6.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.6.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.6.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.6.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.6.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.7.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.7.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.7.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.7.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.7.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.7.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.7.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.7.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.8.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.8.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.8.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.8.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.8.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.8.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.8.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.8.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.9.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.9.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.9.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.9.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.9.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.9.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.9.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.9.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.10.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.10.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.10.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.10.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.10.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.10.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.10.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.10.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.self.query.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.self.query.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.self.key.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.self.key.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.self.value.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.self.value.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.11.attention.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.11.attention.output.LayerNorm.bias, language_model.language_backbone.body.model.encoder.layer.11.intermediate.dense.weight, language_model.language_backbone.body.model.encoder.layer.11.intermediate.dense.bias, language_model.language_backbone.body.model.encoder.layer.11.output.dense.weight, language_model.language_backbone.body.model.encoder.layer.11.output.dense.bias, language_model.language_backbone.body.model.encoder.layer.11.output.LayerNorm.weight, language_model.language_backbone.body.model.encoder.layer.11.output.LayerNorm.bias, text_feat_map.weight, text_feat_map.bias, encoder.text_layers.0.self_attn.attn.in_proj_weight, encoder.text_layers.0.self_attn.attn.in_proj_bias, encoder.text_layers.0.self_attn.attn.out_proj.weight, encoder.text_layers.0.self_attn.attn.out_proj.bias, encoder.text_layers.0.ffn.layers.0.0.weight, encoder.text_layers.0.ffn.layers.0.0.bias, encoder.text_layers.0.ffn.layers.1.weight, encoder.text_layers.0.ffn.layers.1.bias, encoder.text_layers.0.norms.0.weight, encoder.text_layers.0.norms.0.bias, encoder.text_layers.0.norms.1.weight, encoder.text_layers.0.norms.1.bias, encoder.text_layers.1.self_attn.attn.in_proj_weight, encoder.text_layers.1.self_attn.attn.in_proj_bias, encoder.text_layers.1.self_attn.attn.out_proj.weight, encoder.text_layers.1.self_attn.attn.out_proj.bias, encoder.text_layers.1.ffn.layers.0.0.weight, encoder.text_layers.1.ffn.layers.0.0.bias, encoder.text_layers.1.ffn.layers.1.weight, encoder.text_layers.1.ffn.layers.1.bias, encoder.text_layers.1.norms.0.weight, encoder.text_layers.1.norms.0.bias, encoder.text_layers.1.norms.1.weight, encoder.text_layers.1.norms.1.bias, encoder.text_layers.2.self_attn.attn.in_proj_weight, encoder.text_layers.2.self_attn.attn.in_proj_bias, encoder.text_layers.2.self_attn.attn.out_proj.weight, encoder.text_layers.2.self_attn.attn.out_proj.bias, encoder.text_layers.2.ffn.layers.0.0.weight, encoder.text_layers.2.ffn.layers.0.0.bias, encoder.text_layers.2.ffn.layers.1.weight, encoder.text_layers.2.ffn.layers.1.bias, encoder.text_layers.2.norms.0.weight, encoder.text_layers.2.norms.0.bias, encoder.text_layers.2.norms.1.weight, encoder.text_layers.2.norms.1.bias, encoder.text_layers.3.self_attn.attn.in_proj_weight, encoder.text_layers.3.self_attn.attn.in_proj_bias, encoder.text_layers.3.self_attn.attn.out_proj.weight, encoder.text_layers.3.self_attn.attn.out_proj.bias, encoder.text_layers.3.ffn.layers.0.0.weight, encoder.text_layers.3.ffn.layers.0.0.bias, encoder.text_layers.3.ffn.layers.1.weight, encoder.text_layers.3.ffn.layers.1.bias, encoder.text_layers.3.norms.0.weight, encoder.text_layers.3.norms.0.bias, encoder.text_layers.3.norms.1.weight, encoder.text_layers.3.norms.1.bias, encoder.text_layers.4.self_attn.attn.in_proj_weight, encoder.text_layers.4.self_attn.attn.in_proj_bias, encoder.text_layers.4.self_attn.attn.out_proj.weight, encoder.text_layers.4.self_attn.attn.out_proj.bias, encoder.text_layers.4.ffn.layers.0.0.weight, encoder.text_layers.4.ffn.layers.0.0.bias, encoder.text_layers.4.ffn.layers.1.weight, encoder.text_layers.4.ffn.layers.1.bias, encoder.text_layers.4.norms.0.weight, encoder.text_layers.4.norms.0.bias, encoder.text_layers.4.norms.1.weight, encoder.text_layers.4.norms.1.bias, encoder.text_layers.5.self_attn.attn.in_proj_weight, encoder.text_layers.5.self_attn.attn.in_proj_bias, encoder.text_layers.5.self_attn.attn.out_proj.weight, encoder.text_layers.5.self_attn.attn.out_proj.bias, encoder.text_layers.5.ffn.layers.0.0.weight, encoder.text_layers.5.ffn.layers.0.0.bias, encoder.text_layers.5.ffn.layers.1.weight, encoder.text_layers.5.ffn.layers.1.bias, encoder.text_layers.5.norms.0.weight, encoder.text_layers.5.norms.0.bias, encoder.text_layers.5.norms.1.weight, encoder.text_layers.5.norms.1.bias, encoder.fusion_layers.0.gamma_v, encoder.fusion_layers.0.gamma_l, encoder.fusion_layers.0.layer_norm_v.weight, encoder.fusion_layers.0.layer_norm_v.bias, encoder.fusion_layers.0.layer_norm_l.weight, encoder.fusion_layers.0.layer_norm_l.bias, encoder.fusion_layers.0.attn.v_proj.weight, encoder.fusion_layers.0.attn.v_proj.bias, encoder.fusion_layers.0.attn.l_proj.weight, encoder.fusion_layers.0.attn.l_proj.bias, encoder.fusion_layers.0.attn.values_v_proj.weight, encoder.fusion_layers.0.attn.values_v_proj.bias, encoder.fusion_layers.0.attn.values_l_proj.weight, encoder.fusion_layers.0.attn.values_l_proj.bias, encoder.fusion_layers.0.attn.out_v_proj.weight, encoder.fusion_layers.0.attn.out_v_proj.bias, encoder.fusion_layers.0.attn.out_l_proj.weight, encoder.fusion_layers.0.attn.out_l_proj.bias, encoder.fusion_layers.1.gamma_v, encoder.fusion_layers.1.gamma_l, encoder.fusion_layers.1.layer_norm_v.weight, encoder.fusion_layers.1.layer_norm_v.bias, encoder.fusion_layers.1.layer_norm_l.weight, encoder.fusion_layers.1.layer_norm_l.bias, encoder.fusion_layers.1.attn.v_proj.weight, encoder.fusion_layers.1.attn.v_proj.bias, encoder.fusion_layers.1.attn.l_proj.weight, encoder.fusion_layers.1.attn.l_proj.bias, encoder.fusion_layers.1.attn.values_v_proj.weight, encoder.fusion_layers.1.attn.values_v_proj.bias, encoder.fusion_layers.1.attn.values_l_proj.weight, encoder.fusion_layers.1.attn.values_l_proj.bias, encoder.fusion_layers.1.attn.out_v_proj.weight, encoder.fusion_layers.1.attn.out_v_proj.bias, encoder.fusion_layers.1.attn.out_l_proj.weight, encoder.fusion_layers.1.attn.out_l_proj.bias, encoder.fusion_layers.2.gamma_v, encoder.fusion_layers.2.gamma_l, encoder.fusion_layers.2.layer_norm_v.weight, encoder.fusion_layers.2.layer_norm_v.bias, encoder.fusion_layers.2.layer_norm_l.weight, encoder.fusion_layers.2.layer_norm_l.bias, encoder.fusion_layers.2.attn.v_proj.weight, encoder.fusion_layers.2.attn.v_proj.bias, encoder.fusion_layers.2.attn.l_proj.weight, encoder.fusion_layers.2.attn.l_proj.bias, encoder.fusion_layers.2.attn.values_v_proj.weight, encoder.fusion_layers.2.attn.values_v_proj.bias, encoder.fusion_layers.2.attn.values_l_proj.weight, encoder.fusion_layers.2.attn.values_l_proj.bias, encoder.fusion_layers.2.attn.out_v_proj.weight, encoder.fusion_layers.2.attn.out_v_proj.bias, encoder.fusion_layers.2.attn.out_l_proj.weight, encoder.fusion_layers.2.attn.out_l_proj.bias, encoder.fusion_layers.3.gamma_v, encoder.fusion_layers.3.gamma_l, encoder.fusion_layers.3.layer_norm_v.weight, encoder.fusion_layers.3.layer_norm_v.bias, encoder.fusion_layers.3.layer_norm_l.weight, encoder.fusion_layers.3.layer_norm_l.bias, encoder.fusion_layers.3.attn.v_proj.weight, encoder.fusion_layers.3.attn.v_proj.bias, encoder.fusion_layers.3.attn.l_proj.weight, encoder.fusion_layers.3.attn.l_proj.bias, encoder.fusion_layers.3.attn.values_v_proj.weight, encoder.fusion_layers.3.attn.values_v_proj.bias, encoder.fusion_layers.3.attn.values_l_proj.weight, encoder.fusion_layers.3.attn.values_l_proj.bias, encoder.fusion_layers.3.attn.out_v_proj.weight, encoder.fusion_layers.3.attn.out_v_proj.bias, encoder.fusion_layers.3.attn.out_l_proj.weight, encoder.fusion_layers.3.attn.out_l_proj.bias, encoder.fusion_layers.4.gamma_v, encoder.fusion_layers.4.gamma_l, encoder.fusion_layers.4.layer_norm_v.weight, encoder.fusion_layers.4.layer_norm_v.bias, encoder.fusion_layers.4.layer_norm_l.weight, encoder.fusion_layers.4.layer_norm_l.bias, encoder.fusion_layers.4.attn.v_proj.weight, encoder.fusion_layers.4.attn.v_proj.bias, encoder.fusion_layers.4.attn.l_proj.weight, encoder.fusion_layers.4.attn.l_proj.bias, encoder.fusion_layers.4.attn.values_v_proj.weight, encoder.fusion_layers.4.attn.values_v_proj.bias, encoder.fusion_layers.4.attn.values_l_proj.weight, encoder.fusion_layers.4.attn.values_l_proj.bias, encoder.fusion_layers.4.attn.out_v_proj.weight, encoder.fusion_layers.4.attn.out_v_proj.bias, encoder.fusion_layers.4.attn.out_l_proj.weight, encoder.fusion_layers.4.attn.out_l_proj.bias, encoder.fusion_layers.5.gamma_v, encoder.fusion_layers.5.gamma_l, encoder.fusion_layers.5.layer_norm_v.weight, encoder.fusion_layers.5.layer_norm_v.bias, encoder.fusion_layers.5.layer_norm_l.weight, encoder.fusion_layers.5.layer_norm_l.bias, encoder.fusion_layers.5.attn.v_proj.weight, encoder.fusion_layers.5.attn.v_proj.bias, encoder.fusion_layers.5.attn.l_proj.weight, encoder.fusion_layers.5.attn.l_proj.bias, encoder.fusion_layers.5.attn.values_v_proj.weight, encoder.fusion_layers.5.attn.values_v_proj.bias, encoder.fusion_layers.5.attn.values_l_proj.weight, encoder.fusion_layers.5.attn.values_l_proj.bias, encoder.fusion_layers.5.attn.out_v_proj.weight, encoder.fusion_layers.5.attn.out_v_proj.bias, encoder.fusion_layers.5.attn.out_l_proj.weight, encoder.fusion_layers.5.attn.out_l_proj.bias + +10/10 22:42:30 - mmengine - INFO - Load checkpoint from pretrained/fcos3d.pth +10/10 22:42:30 - mmengine - WARNING - "FileClient" will be deprecated in future. Please use io functions in https://mmengine.readthedocs.io/en/latest/api/fileio.html#file-io +10/10 22:42:30 - mmengine - WARNING - "HardDiskBackend" is the alias of "LocalBackend" and the former will be deprecated in future. +10/10 22:42:30 - mmengine - INFO - Checkpoints will be saved to /home/xzt/mmdetection3d-1.1.0/work_dirs/detr3d_r50_bert_gridmask_halfdata_decoder. +/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/grid_mask.py:132: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at /pytorch/torch/csrc/utils/tensor_numpy.cpp:180.) + mask = torch.from_numpy(mask).to(x) +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at /pytorch/c10/core/TensorImpl.h:1156.) + return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode) +/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/grid_mask.py:132: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at /pytorch/torch/csrc/utils/tensor_numpy.cpp:180.) + mask = torch.from_numpy(mask).to(x) +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at /pytorch/c10/core/TensorImpl.h:1156.) + return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode) +Traceback (most recent call last): + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 131, in main + runner.train() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 1745, in train + model = self.train_loop.run() # type: ignore + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 96, in run + self.run_epoch() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 112, in run_epoch + self.run_iter(idx, data_batch) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 128, in run_iter + outputs = self.runner.model.train_step( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/model/wrappers/distributed.py", line 121, in train_step + losses = self._run_forward(data, mode='loss') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/model/wrappers/distributed.py", line 161, in _run_forward + results = self(**data, mode=mode) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 799, in forward + output = self.module(*inputs[0], **kwargs[0]) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/mmdet3d/models/detectors/base.py", line 75, in forward + return self.loss(inputs, data_samples, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d.py", line 196, in loss + outs = self.pts_bbox_head(img_feats, batch_input_metas,memory_text, **kwargs)#text_dict + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d_head.py", line 133, in forward + hs, init_reference, inter_references = self.transformer( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d_transformer.py", line 116, in forward + inter_states, inter_references = self.decoder( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d_transformer.py", line 173, in forward + output = layer( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmcv/cnn/bricks/transformer.py", line 845, in forward + temp_key = temp_value = memory_text.reshape(26,1,256) +RuntimeError: shape '[26, 1, 256]' is invalid for input of size 6144 +Traceback (most recent call last): + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 131, in main + runner.train() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 1745, in train + model = self.train_loop.run() # type: ignore + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 96, in run + self.run_epoch() + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 112, in run_epoch + self.run_iter(idx, data_batch) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/loops.py", line 128, in run_iter + outputs = self.runner.model.train_step( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/model/wrappers/distributed.py", line 121, in train_step + losses = self._run_forward(data, mode='loss') + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/model/wrappers/distributed.py", line 161, in _run_forward + results = self(**data, mode=mode) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 799, in forward + output = self.module(*inputs[0], **kwargs[0]) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/mmdet3d/models/detectors/base.py", line 75, in forward + return self.loss(inputs, data_samples, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d.py", line 196, in loss + outs = self.pts_bbox_head(img_feats, batch_input_metas,memory_text, **kwargs)#text_dict + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d_head.py", line 133, in forward + hs, init_reference, inter_references = self.transformer( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d_transformer.py", line 116, in forward + inter_states, inter_references = self.decoder( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/mmdetection3d-1.1.0/projects/DETR3D/detr3d/detr3d_transformer.py", line 173, in forward + output = layer( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl + return forward_call(*input, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmcv/cnn/bricks/transformer.py", line 845, in forward + temp_key = temp_value = memory_text.reshape(26,1,256) +RuntimeError: shape '[26, 1, 256]' is invalid for input of size 6144 +ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 948950) of binary: /home/xzt/miniconda3/envs/detr/bin/python +ERROR:torch.distributed.elastic.agent.server.local_elastic_agent:[default] Worker group failed +INFO:torch.distributed.elastic.agent.server.api:[default] Worker group FAILED. 3/3 attempts left; will restart worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Stopping worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous'ing worker group +INFO:torch.distributed.elastic.agent.server.api:[default] Rendezvous complete for workers. Result: + restart_count=1 + master_addr=127.0.0.1 + master_port=29500 + group_rank=0 + group_world_size=1 + local_ranks=[0, 1] + role_ranks=[0, 1] + global_ranks=[0, 1] + role_world_sizes=[2, 2] + global_world_sizes=[2, 2] + +INFO:torch.distributed.elastic.agent.server.api:[default] Starting worker group +INFO:torch.distributed.elastic.multiprocessing:Setting worker0 reply file to: /tmp/torchelastic_9g3to02q/none_u2xm2q1i/attempt_1/0/error.json +INFO:torch.distributed.elastic.multiprocessing:Setting worker1 reply file to: /tmp/torchelastic_9g3to02q/none_u2xm2q1i/attempt_1/1/error.json +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/utils/dl_utils/setup_env.py:56: UserWarning: Setting MKL_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. + warnings.warn( +Traceback (most recent call last): + File "tools/train.py", line 135, in + main() + File "tools/train.py", line 124, in main + runner = Runner.from_cfg(cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg +Traceback (most recent call last): + File "tools/train.py", line 135, in + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + main() + File "tools/train.py", line 124, in main + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 208, in _store_based_barrier + runner = Runner.from_cfg(cfg) + worker_count = store.add(store_key, 0) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 445, in from_cfg +RuntimeError: Connection reset by peer + runner = cls( + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 355, in __init__ + self.setup_env(env_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/runner/runner.py", line 658, in setup_env + init_dist(self.launcher, **dist_cfg) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 84, in init_dist + _init_dist_pytorch(backend, init_backend=init_backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/mmengine/dist/utils.py", line 125, in _init_dist_pytorch + torch_dist.init_process_group(backend=backend, **kwargs) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 547, in init_process_group + _store_based_barrier(rank, store, timeout) + File "/home/xzt/miniconda3/envs/detr/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 208, in _store_based_barrier + worker_count = store.add(store_key, 0) +RuntimeError: Connection reset by peer diff --git a/projects/DETR3D/configs/detr3d_r50_bert_gridmask_halfdata_decoder.py b/projects/DETR3D/configs/detr3d_r50_bert_gridmask_halfdata_decoder.py new file mode 100644 index 0000000..58330e7 --- /dev/null +++ b/projects/DETR3D/configs/detr3d_r50_bert_gridmask_halfdata_decoder.py @@ -0,0 +1,298 @@ +_base_ = [ + # 'mmdet3d::_base_/datasets/nus-3d.py', + 'mmdet3d::_base_/default_runtime.py' +] + +custom_imports = dict(imports=['projects.DETR3D.detr3d']) +# If point cloud range is changed, the models should also change their point +# cloud range accordingly +point_cloud_range = [-51.2, -51.2, -5.0, 51.2, 51.2, 3.0] +voxel_size = [0.2, 0.2, 8] + +img_norm_cfg = dict( + mean=[103.530, 116.280, 123.675], std=[1.0, 1.0, 1.0], bgr_to_rgb=False) +# For nuScenes we usually do 10-class detection +class_names = [ + 'car', 'truck', 'construction_vehicle', 'bus', 'trailer', 'barrier', + 'motorcycle', 'bicycle', 'pedestrian', 'traffic_cone' +] +lang_model_name = 'bert-base-uncased' +input_modality = dict( + use_lidar=False, + use_camera=True, + use_radar=False, + use_map=False, + use_external=False) +# this means type='DETR3D' will be processed as 'mmdet3d.DETR3D' +default_scope = 'mmdet3d' +model = dict( + type='DETR3D', + use_grid_mask=True, + data_preprocessor=dict( + type='Det3DDataPreprocessor', **img_norm_cfg, pad_size_divisor=32), + language_model=dict( + type='BertModel', + name=lang_model_name, + pad_to_max=False, + use_sub_sentence_represent=True, + special_tokens_list=['[CLS]', '[SEP]', '.', '?'], + add_pooling_layer=False, + ), + encoder=dict( + num_layers=6, + num_cp=6, + # visual layer config + layer_cfg=dict( + self_attn_cfg=dict(embed_dims=256, num_levels=4, dropout=0.0), + ffn_cfg=dict( + embed_dims=256, feedforward_channels=2048, ffn_drop=0.0)), + # text layer config + text_layer_cfg=dict( + self_attn_cfg=dict(num_heads=4, embed_dims=256, dropout=0.0), + ffn_cfg=dict( + embed_dims=256, feedforward_channels=1024, ffn_drop=0.0)), + # fusion layer config + fusion_layer_cfg=dict( + v_dim=256, + l_dim=256, + embed_dim=1024, + num_heads=4, + init_values=1e-4), + ), + positional_encoding_single=dict( + num_feats=128, + normalize=True, + temperature=20, + offset=0.0), + img_backbone=dict( + type='mmdet.ResNet', + depth=50, + num_stages=4, + out_indices=(0, 1, 2, 3), + frozen_stages=1, + norm_cfg=dict(type='BN2d', requires_grad=False), + norm_eval=True, + style='caffe', + dcn=dict(type='DCNv2', deform_groups=1, fallback_on_stride=False), + stage_with_dcn=(False, False, True, True)), + img_neck=dict( + type='mmdet.FPN', + in_channels=[256, 512, 1024, 2048], + out_channels=256, + start_level=1, + add_extra_convs='on_output', + num_outs=4, + relu_before_extra_convs=True), + pts_bbox_head=dict( + type='DETR3DHead', + num_query=900, + num_classes=10, + in_channels=256, + sync_cls_avg_factor=True, + with_box_refine=True, + as_two_stage=False, + transformer=dict( + type='Detr3DTransformer', + decoder=dict( + type='Detr3DTransformerDecoder', + num_layers=6, + return_intermediate=True, + transformerlayers=dict( + type='mmdet.DetrTransformerDecoderLayer', + attn_cfgs=[ + dict( + type='MultiheadAttention', # mmcv. + embed_dims=256, + num_heads=8, + dropout=0.1), + dict( + type='MultiheadAttention', # mmcv. + embed_dims=256, + num_heads=8, + dropout=0.0), + dict( + type='Detr3DCrossAtten', + pc_range=point_cloud_range, + num_points=1, + embed_dims=256) + ], + feedforward_channels=512, + ffn_dropout=0.1, + operation_order=('self_attn', 'norm','self_attn_text','norm', 'cross_attn', 'norm', + 'ffn', 'norm')))), + bbox_coder=dict( + type='NMSFreeCoder', + post_center_range=[-61.2, -61.2, -10.0, 61.2, 61.2, 10.0], + pc_range=point_cloud_range, + max_num=300, + voxel_size=voxel_size, + num_classes=10), + positional_encoding=dict( + type='mmdet.SinePositionalEncoding', + num_feats=128, + normalize=True, + offset=-0.5), + loss_cls=dict( + type='mmdet.FocalLoss', + use_sigmoid=True, + gamma=2.0, + alpha=0.25, + loss_weight=2.0), + loss_bbox=dict(type='mmdet.L1Loss', loss_weight=0.25), + loss_iou=dict(type='mmdet.GIoULoss', loss_weight=0.0)), + # model training and testing settings + train_cfg=dict( + pts=dict( + grid_size=[512, 512, 1], + voxel_size=voxel_size, + point_cloud_range=point_cloud_range, + out_size_factor=4, + assigner=dict( + type='HungarianAssigner3D', + cls_cost=dict(type='mmdet.FocalLossCost', weight=2.0), + reg_cost=dict(type='BBox3DL1Cost', weight=0.25), + # ↓ Fake cost. This is just to get compatible with DETR head + iou_cost=dict(type='mmdet.IoUCost', weight=0.0), + pc_range=point_cloud_range)))) + +dataset_type = 'NuScenesDataset' +data_root = 'data/nuscenes/' + +test_transforms = [ + dict( + type='RandomResize3D', + scale=(1600, 900), + ratio_range=(1., 1.), + keep_ratio=True) +] +train_transforms = [dict(type='PhotoMetricDistortion3D')] + test_transforms + +backend_args = None +train_pipeline = [ + dict( + type='LoadMultiViewImageFromFiles', + to_float32=True, + num_views=6, + backend_args=backend_args), + dict( + type='LoadAnnotations3D', + with_bbox_3d=True, + with_label_3d=True, + with_attr_label=False), + dict(type='MultiViewWrapper', transforms=train_transforms), + dict(type='ObjectRangeFilter', point_cloud_range=point_cloud_range), + dict(type='ObjectNameFilter', classes=class_names), + dict(type='Pack3DDetInputs', keys=['img', 'gt_bboxes_3d', 'gt_labels_3d']) +] + +test_pipeline = [ + dict( + type='LoadMultiViewImageFromFiles', + to_float32=True, + num_views=6, + backend_args=backend_args), + dict(type='MultiViewWrapper', transforms=test_transforms), + dict(type='Pack3DDetInputs', keys=['img']) +] + +metainfo = dict(classes=class_names) +data_prefix = dict( + pts='', + CAM_FRONT='samples/CAM_FRONT', + CAM_FRONT_LEFT='samples/CAM_FRONT_LEFT', + CAM_FRONT_RIGHT='samples/CAM_FRONT_RIGHT', + CAM_BACK='samples/CAM_BACK', + CAM_BACK_RIGHT='samples/CAM_BACK_RIGHT', + CAM_BACK_LEFT='samples/CAM_BACK_LEFT') + +train_dataloader = dict( + batch_size=1, + num_workers=4, + persistent_workers=True, + drop_last=False, + sampler=dict(type='DefaultSampler', shuffle=True), + dataset=dict( + type=dataset_type, + data_root=data_root, + ann_file='nuscenes_infos_train.pkl', + pipeline=train_pipeline, + load_type='frame_based', + metainfo=metainfo, + modality=input_modality, + test_mode=False, + data_prefix=data_prefix, + # we use box_type_3d='LiDAR' in kitti and nuscenes dataset + # and box_type_3d='Depth' in sunrgbd and scannet dataset. + box_type_3d='LiDAR', + load_interval = 2, + backend_args=backend_args)) + +val_dataloader = dict( + batch_size=1, + num_workers=4, + persistent_workers=True, + drop_last=False, + sampler=dict(type='DefaultSampler', shuffle=False), + dataset=dict( + type=dataset_type, + data_root=data_root, + ann_file='nuscenes_infos_val.pkl', + load_type='frame_based', + pipeline=test_pipeline, + metainfo=metainfo, + modality=input_modality, + test_mode=True, + data_prefix=data_prefix, + box_type_3d='LiDAR', + backend_args=backend_args)) + +test_dataloader = val_dataloader + +val_evaluator = dict( + type='NuScenesMetric', + data_root=data_root, + ann_file=data_root + 'nuscenes_infos_val.pkl', + metric='bbox', + backend_args=backend_args) +test_evaluator = val_evaluator + +optim_wrapper = dict( + type='OptimWrapper', + optimizer=dict(type='AdamW', lr=2e-4, weight_decay=0.01), + paramwise_cfg=dict(custom_keys={'img_backbone': dict(lr_mult=0.1)}), + clip_grad=dict(max_norm=35, norm_type=2), +) + +# learning policy +param_scheduler = [ + dict( + type='LinearLR', + start_factor=1.0 / 3, + by_epoch=False, + begin=0, + end=500), + dict( + type='CosineAnnealingLR', + by_epoch=True, + begin=0, + end=24, + T_max=24, + eta_min_ratio=1e-3) +] + +total_epochs = 24 + +train_cfg = dict( + type='EpochBasedTrainLoop', max_epochs=total_epochs, val_interval=2) +val_cfg = dict(type='ValLoop') +test_cfg = dict(type='TestLoop') +default_hooks = dict( + checkpoint=dict( + type='CheckpointHook', interval=1, max_keep_ckpts=1, save_last=True)) +load_from = 'pretrained/fcos3d.pth' + +# setuptools 65 downgrades to 58. +# In mmlab-node we use setuptools 61 but occurs NO errors +vis_backends = [dict(type='TensorboardVisBackend')] +visualizer = dict( + type='Det3DLocalVisualizer', vis_backends=vis_backends, name='visualizer') diff --git a/projects/DETR3D/detr3d/detr3d.py b/projects/DETR3D/detr3d/detr3d.py index c0a84a0..c57af66 100755 --- a/projects/DETR3D/detr3d/detr3d.py +++ b/projects/DETR3D/detr3d/detr3d.py @@ -72,7 +72,7 @@ def __init__(self, self._special_tokens = '. ' self.positional_encoding = SinePositionalEncoding( **positional_encoding_single) - self.encoder = GroundingDinoTransformerEncoder(**encoder) + # self.encoder = GroundingDinoTransformerEncoder(**encoder) # self.level_embed = nn.Parameter( # torch.Tensor(4, 256)) nn.init.constant_(self.text_feat_map.bias.data, 0) @@ -156,8 +156,8 @@ def loss(self, batch_inputs_dict: Dict[List, Tensor], bsz=len(batch_data_samples) #文本预处理 text_prompts=[ - 'car', 'truck', 'trailer', 'bus', 'construction_vehicle', 'bicycle', - 'motorcycle', 'pedestrian', 'traffic_cone', 'barrier'] + 'car', 'truck', 'trailer', 'bus', 'construction vehicle', 'bicycle', + 'motorcycle', 'pedestrian', 'traffic cone', 'barrier'] batch_gt_instances_3d = [ item.gt_instances_3d for item in batch_data_samples ] @@ -180,19 +180,20 @@ def loss(self, batch_inputs_dict: Dict[List, Tensor], positive_maps.append(positive_map) text_dict = self.language_model(new_text_prompts) - for key, value in text_dict.items(): - text_dict[key] = torch.cat([value] * 6, dim=0) + # for key, value in text_dict.items(): + # text_dict[key] = torch.cat([value] * 6, dim=0) if self.text_feat_map is not None: text_dict['embedded'] = self.text_feat_map(text_dict['embedded']) + memory_text=text_dict['embedded'] ##################################################################### - encoder_inputs_dict = self.pre_transformer( - img_feats, batch_data_samples) - - memory = self.forward_encoder( - **encoder_inputs_dict, text_dict=text_dict) - del img_feats - img_feats = self.restore_img_feats(memory, encoder_inputs_dict['spatial_shapes'], encoder_inputs_dict['level_start_index']) - outs = self.pts_bbox_head(img_feats, batch_input_metas, **kwargs)#text_dict + # encoder_inputs_dict = self.pre_transformer( + # img_feats, batch_data_samples) + + # memory,memory_text = self.forward_encoder( + # **encoder_inputs_dict, text_dict=text_dict)#text和图像特征融合 + # del img_feats + # img_feats = self.restore_img_feats(memory, encoder_inputs_dict['spatial_shapes'], encoder_inputs_dict['level_start_index']) + outs = self.pts_bbox_head(img_feats, batch_input_metas,memory_text, **kwargs)#text_dict loss_inputs = [batch_gt_instances_3d, outs] losses_pts = self.pts_bbox_head.loss_by_feat(*loss_inputs) @@ -490,7 +491,7 @@ def forward_encoder(self, feat: Tensor, feat_mask: Tensor, level_start_index: Tensor, valid_ratios: Tensor, text_dict: Dict) -> Dict: text_token_mask = text_dict['text_token_mask'] - memory, _ = self.encoder( + memory, memory_text = self.encoder( query=feat, # query_pos=feat_pos, key_padding_mask=feat_mask, # for self_attn @@ -509,7 +510,7 @@ def forward_encoder(self, feat: Tensor, feat_mask: Tensor, # memory_text=memory_text, # text_token_mask=text_token_mask) # return encoder_outputs_dict - return memory + return memory,memory_text @staticmethod def get_valid_ratio(mask: Tensor) -> Tensor: """Get the valid radios of feature map in a level. diff --git a/projects/DETR3D/detr3d/detr3d_head.py b/projects/DETR3D/detr3d/detr3d_head.py index d4143ad..ff18334 100755 --- a/projects/DETR3D/detr3d/detr3d_head.py +++ b/projects/DETR3D/detr3d/detr3d_head.py @@ -112,7 +112,7 @@ def init_weights(self): for m in self.cls_branches: nn.init.constant_(m[-1].bias, bias_init) - def forward(self, mlvl_feats: List[Tensor], img_metas: List[Dict], + def forward(self, mlvl_feats: List[Tensor], img_metas: List[Dict],memory_text:Tensor, **kwargs) -> Dict[str, Tensor]: """Forward function. @@ -135,6 +135,7 @@ def forward(self, mlvl_feats: List[Tensor], img_metas: List[Dict], query_embeds, reg_branches=self.reg_branches if self.with_box_refine else None, img_metas=img_metas, + memory_text=memory_text, **kwargs) hs = hs.permute(0, 2, 1, 3) outputs_classes = [] diff --git a/projects/DETR3D/detr3d/detr3d_transformer.py b/projects/DETR3D/detr3d/detr3d_transformer.py index dfe0765..f5ed263 100755 --- a/projects/DETR3D/detr3d/detr3d_transformer.py +++ b/projects/DETR3D/detr3d/detr3d_transformer.py @@ -75,7 +75,7 @@ def init_weights(self): m.init_weight() xavier_init(self.reference_points, distribution='uniform', bias=0.) - def forward(self, mlvl_feats, query_embed, reg_branches=None, **kwargs): + def forward(self, mlvl_feats, query_embed,memory_text, reg_branches=None, **kwargs): """Forward function for `Detr3DTransformer`. Args: mlvl_feats (list(Tensor)): Input queries from @@ -120,6 +120,7 @@ def forward(self, mlvl_feats, query_embed, reg_branches=None, **kwargs): query_pos=query_pos, reference_points=reference_points, reg_branches=reg_branches, + memory_text=memory_text, **kwargs) inter_references_out = inter_references diff --git a/projects/DETR3D/layers/transformer/__init__.py b/projects/DETR3D/layers/transformer/__init__.py index a86091e..c073e1f 100755 --- a/projects/DETR3D/layers/transformer/__init__.py +++ b/projects/DETR3D/layers/transformer/__init__.py @@ -21,6 +21,7 @@ from .utils import (MLP, AdaptivePadding, ConditionalAttention, PatchEmbed, PatchMerging, coordinate_to_encoding, inverse_sigmoid, nchw_to_nlc, nlc_to_nchw) +from .vlfuse_helper import BertEncoderLayer, VLFuse, permute_and_flatten __all__ = [ 'nlc_to_nchw', 'nchw_to_nlc', 'AdaptivePadding', 'PatchEmbed', @@ -37,5 +38,6 @@ 'CdnQueryGenerator', 'Mask2FormerTransformerEncoder', 'Mask2FormerTransformerDecoderLayer', 'Mask2FormerTransformerDecoder', 'GroundingDinoTransformerDecoderLayer', 'GroundingDinoTransformerEncoder', - 'GroundingDinoTransformerDecoder' + 'GroundingDinoTransformerDecoder','VLFuse', 'permute_and_flatten', + 'BertEncoderLayer' ] diff --git a/projects/DETR3D/layers/transformer/grounding_dino_layers.py b/projects/DETR3D/layers/transformer/grounding_dino_layers.py index e559b8e..2182d33 100755 --- a/projects/DETR3D/layers/transformer/grounding_dino_layers.py +++ b/projects/DETR3D/layers/transformer/grounding_dino_layers.py @@ -6,8 +6,8 @@ from mmcv.ops import MultiScaleDeformableAttention from mmengine.model import ModuleList from torch import Tensor - -from mmdet.models.utils.vlfuse_helper import SingleScaleBiAttentionBlock +from .vlfuse_helper import SingleScaleBiAttentionBlock +# from mmdet.models.utils.vlfuse_helper import SingleScaleBiAttentionBlock from mmdet.utils import ConfigType, OptConfigType from .deformable_detr_layers import (DeformableDetrTransformerDecoderLayer, DeformableDetrTransformerEncoder, @@ -146,10 +146,10 @@ def _init_layers(self) -> None: # DeformableDetrTransformerEncoderLayer(**self.layer_cfg) # for _ in range(self.num_layers) # ]) - # self.text_layers = ModuleList([ - # DetrTransformerEncoderLayer(**self.text_layer_cfg) - # for _ in range(self.num_layers) - # ]) + self.text_layers = ModuleList([ + DetrTransformerEncoderLayer(**self.text_layer_cfg) + for _ in range(self.num_layers) + ]) self.fusion_layers = ModuleList([ SingleScaleBiAttentionBlock(**self.fusion_layer_cfg) for _ in range(self.num_layers) @@ -208,21 +208,21 @@ def forward(self, output = query # reference_points = self.get_encoder_reference_points( # spatial_shapes, valid_ratios, device=query.device) - # if self.text_layers: - # # generate pos_text - # bs, n_text, _ = memory_text.shape - # if pos_text is None and position_ids is None: - # pos_text = ( - # torch.arange(n_text, - # device=memory_text.device).float().unsqueeze( - # 0).unsqueeze(-1).repeat(bs, 1, 1)) - # pos_text = get_text_sine_pos_embed( - # pos_text, num_pos_feats=256, exchange_xy=False) - # if position_ids is not None: - # pos_text = get_text_sine_pos_embed( - # position_ids[..., None], - # num_pos_feats=256, - # exchange_xy=False) + if self.text_layers: + # generate pos_text + bs, n_text, _ = memory_text.shape + if pos_text is None and position_ids is None: + pos_text = ( + torch.arange(n_text, + device=memory_text.device).float().unsqueeze( + 0).unsqueeze(-1).repeat(bs, 1, 1)) + pos_text = get_text_sine_pos_embed( + pos_text, num_pos_feats=256, exchange_xy=False) + if position_ids is not None: + pos_text = get_text_sine_pos_embed( + position_ids[..., None], + num_pos_feats=256, + exchange_xy=False) # main process # for layer_id, layer in enumerate(self.layers): @@ -234,16 +234,16 @@ def forward(self, attention_mask_v=key_padding_mask, attention_mask_l=text_attention_mask, ) - # if self.text_layers: - # text_num_heads = self.text_layers[ - # layer_id].self_attn_cfg.num_heads - # memory_text = self.text_layers[layer_id]( - # query=memory_text[0], - # query_pos=(pos_text if pos_text is not None else None), - # attn_mask=~text_self_attention_masks.repeat( - # text_num_heads, 1, 1), # note we use ~ for mask here - # key_padding_mask=None, - # ) + if self.text_layers: + text_num_heads = self.text_layers[ + layer_id].self_attn_cfg.num_heads + memory_text = self.text_layers[layer_id]( + query=memory_text[0], + query_pos=(pos_text if pos_text is not None else None), + attn_mask=~text_self_attention_masks.repeat( + text_num_heads, 1, 1), # note we use ~ for mask here + key_padding_mask=None, + ) # output = layer( # query=output, # query_pos=query_pos, diff --git a/projects/DETR3D/layers/transformer/vlfuse_helper.py b/projects/DETR3D/layers/transformer/vlfuse_helper.py new file mode 100644 index 0000000..cb21483 --- /dev/null +++ b/projects/DETR3D/layers/transformer/vlfuse_helper.py @@ -0,0 +1,780 @@ +# Copyright (c) OpenMMLab. All rights reserved. +# Modified from https://github.com/microsoft/GLIP/blob/main/maskrcnn_benchmark/utils/fuse_helper.py # noqa +# and https://github.com/microsoft/GLIP/blob/main/maskrcnn_benchmark/modeling/rpn/modeling_bert.py # noqa +import math +from typing import Dict, Optional, Tuple + +import torch +import torch.nn as nn +import torch.nn.functional as F +import torch.utils.checkpoint as checkpoint +from mmcv.cnn.bricks import DropPath +from torch import Tensor + +try: + from transformers import BertConfig, BertPreTrainedModel + from transformers.modeling_utils import apply_chunking_to_forward + from transformers.models.bert.modeling_bert import \ + BertAttention as HFBertAttention + from transformers.models.bert.modeling_bert import \ + BertIntermediate as HFBertIntermediate + from transformers.models.bert.modeling_bert import \ + BertOutput as HFBertOutput +except ImportError: + BertConfig = None + BertPreTrainedModel = object + apply_chunking_to_forward = None + HFBertAttention = object + HFBertIntermediate = object + HFBertOutput = object + +MAX_CLAMP_VALUE = 50000 + + +def permute_and_flatten(layer: Tensor, N: int, A: int, C: int, H: int, + W: int) -> Tensor: + """Permute and then flatten a tensor, + + from size (N, A, C, H, W) to (N, H * W * A, C). + + Args: + layer (Tensor): Tensor of shape (N, C, H, W). + N (int): Batch size. + A (int): Number of attention heads. + C (int): Number of channels. + H (int): Height of feature map. + W (int): Width of feature map. + + Returns: + Tensor: A Tensor of shape (N, H * W * A, C). + """ + layer = layer.view(N, A, C, H, W) + layer = layer.permute(0, 3, 4, 1, 2) + layer = layer.reshape(N, -1, C) + return layer + + +def clamp_values(vector: Tensor) -> Tensor: + """Clamp the values of a vector to the range [-MAX_CLAMP_VALUE, + MAX_CLAMP_VALUE]. + + Args: + vector (Tensor): Tensor of shape (N, C, H, W). + + Returns: + Tensor: A Tensor of shape (N, C, H, W) with clamped values. + """ + vector = torch.clamp(vector, min=-MAX_CLAMP_VALUE, max=MAX_CLAMP_VALUE) + return vector + + +class BiMultiHeadAttention(nn.Module): + """Bidirectional fusion Multi-Head Attention layer. + + Args: + v_dim (int): The dimension of the vision input. + l_dim (int): The dimension of the language input. + embed_dim (int): The embedding dimension for the attention operation. + num_heads (int): The number of attention heads. + dropout (float, optional): The dropout probability. Defaults to 0.1. + """ + + def __init__(self, + v_dim: int, + l_dim: int, + embed_dim: int, + num_heads: int, + dropout: float = 0.1): + super(BiMultiHeadAttention, self).__init__() + + self.embed_dim = embed_dim + self.num_heads = num_heads + self.head_dim = embed_dim // num_heads + self.v_dim = v_dim + self.l_dim = l_dim + + assert ( + self.head_dim * self.num_heads == self.embed_dim + ), 'embed_dim must be divisible by num_heads ' \ + f'(got `embed_dim`: {self.embed_dim} ' \ + f'and `num_heads`: {self.num_heads}).' + self.scale = self.head_dim**(-0.5) + self.dropout = dropout + + self.v_proj = nn.Linear(self.v_dim, self.embed_dim) + self.l_proj = nn.Linear(self.l_dim, self.embed_dim) + self.values_v_proj = nn.Linear(self.v_dim, self.embed_dim) + self.values_l_proj = nn.Linear(self.l_dim, self.embed_dim) + + self.out_v_proj = nn.Linear(self.embed_dim, self.v_dim) + self.out_l_proj = nn.Linear(self.embed_dim, self.l_dim) + + self.stable_softmax_2d = False + self.clamp_min_for_underflow = True + self.clamp_max_for_overflow = True + + self._reset_parameters() + + def _shape(self, tensor: Tensor, seq_len: int, bsz: int): + return tensor.view(bsz, seq_len, self.num_heads, + self.head_dim).transpose(1, 2).contiguous() + def _shape1(self, tensor: Tensor, seq_len: int, bsz: int ,num_cam: int=6): + return tensor.view(bsz, num_cam, seq_len, self.num_heads, + self.head_dim).transpose(2, 3).contiguous() + def _reset_parameters(self): + nn.init.xavier_uniform_(self.v_proj.weight) + self.v_proj.bias.data.fill_(0) + nn.init.xavier_uniform_(self.l_proj.weight) + self.l_proj.bias.data.fill_(0) + nn.init.xavier_uniform_(self.values_v_proj.weight) + self.values_v_proj.bias.data.fill_(0) + nn.init.xavier_uniform_(self.values_l_proj.weight) + self.values_l_proj.bias.data.fill_(0) + nn.init.xavier_uniform_(self.out_v_proj.weight) + self.out_v_proj.bias.data.fill_(0) + nn.init.xavier_uniform_(self.out_l_proj.weight) + self.out_l_proj.bias.data.fill_(0) + + def forward( + self, + vision: Tensor, + lang: Tensor, + attention_mask_v: Optional[Tensor] = None, + attention_mask_l: Optional[Tensor] = None, + ) -> Tuple[Tensor, Tensor]: + bsz,num_cam, tgt_len, _ = vision.size() + tmp=[] + for i in range(bsz): + tmp.append(self.v_proj(vision[i]) * self.scale)#6,30825,256 6,30825,1024 + query_states=torch.cat(tmp,dim=0).view(bsz,num_cam,-1,1024)#1,6,30825,1024 + key_states = self._shape(self.l_proj(lang), -1, bsz*num_cam)#6,26,256 6,26,1024 6,26,4,256, 6,4,26,256 + tmp=[] + for i in range(bsz): + tmp.append(self._shape(self.values_v_proj(vision[i]), -1, num_cam)) + value_v_states=torch.cat(tmp,dim=0).view(bsz,num_cam,self.num_heads,-1,256)#1,6,4,30825,256 + value_l_states = self._shape(self.values_l_proj(lang), -1, bsz*num_cam) + proj_shape = (bsz * self.num_heads, -1, self.head_dim) + proj_shape1 = (bsz * self.num_heads * num_cam, -1, self.head_dim) + query_states = self._shape1(query_states, tgt_len, + bsz).view(*proj_shape1)#1,6,30825,1024 24,30825,256 + key_states = key_states.view(*proj_shape1) + value_v_states = value_v_states.view(*proj_shape1) + value_l_states = value_l_states.view(*proj_shape1) + + src_len = key_states.size(1) + attn_weights = torch.bmm(query_states, key_states.transpose(1, 2)) + + if attn_weights.size() != (bsz * self.num_heads * num_cam, tgt_len, src_len): + raise ValueError( + f'Attention weights should be of ' + f'size {(bsz * self.num_heads, tgt_len, src_len)}, ' + f'but is {attn_weights.size()}') + + if self.stable_softmax_2d: + attn_weights = attn_weights - attn_weights.max() + + if self.clamp_min_for_underflow: + # Do not increase -50000, data type half has quite limited range + attn_weights = torch.clamp(attn_weights, min=-MAX_CLAMP_VALUE) + if self.clamp_max_for_overflow: + # Do not increase 50000, data type half has quite limited range + attn_weights = torch.clamp(attn_weights, max=MAX_CLAMP_VALUE) + + attn_weights_T = attn_weights.transpose(1, 2) + attn_weights_l = ( + attn_weights_T - + torch.max(attn_weights_T, dim=-1, keepdim=True)[0]) + if self.clamp_min_for_underflow: + # Do not increase -50000, data type half has quite limited range + attn_weights_l = torch.clamp(attn_weights_l, min=-MAX_CLAMP_VALUE) + if self.clamp_max_for_overflow: + # Do not increase 50000, data type half has quite limited range + attn_weights_l = torch.clamp(attn_weights_l, max=MAX_CLAMP_VALUE) + + if attention_mask_v is not None: + attention_mask_v = ( + attention_mask_v[:, :,None, + None, :].repeat(1, 1, self.num_heads, 1, + 1).flatten(0, 1).flatten(0,1)) + attn_weights_l.masked_fill_(attention_mask_v, float('-inf')) + + attn_weights_l = attn_weights_l.softmax(dim=-1) + + if attention_mask_l is not None: + assert (attention_mask_l.dim() == 2) + attention_mask = attention_mask_l.unsqueeze(1).unsqueeze(1) + attention_mask = attention_mask.expand(bsz*num_cam, 1, tgt_len, src_len) + attention_mask = attention_mask.masked_fill( + attention_mask == 0, -9e15) + + if attention_mask.size() != (bsz * num_cam, 1, tgt_len, src_len): + raise ValueError('Attention mask should be of ' + f'size {(bsz * num_cam, 1, tgt_len, src_len)}') + attn_weights = attn_weights.view(bsz * num_cam, self.num_heads, tgt_len, + src_len) + attention_mask + attn_weights = attn_weights.view(bsz * self.num_heads * num_cam, tgt_len, + src_len) + + attn_weights_v = nn.functional.softmax(attn_weights, dim=-1) + + attn_probs_v = F.dropout( + attn_weights_v, p=self.dropout, training=self.training) + attn_probs_l = F.dropout( + attn_weights_l, p=self.dropout, training=self.training) + + attn_output_v = torch.bmm(attn_probs_v, value_l_states) + attn_output_l = torch.bmm(attn_probs_l, value_v_states) + + if attn_output_v.size() != (bsz * self.num_heads * num_cam, tgt_len, + self.head_dim): + raise ValueError( + '`attn_output_v` should be of ' + f'size {(bsz, self.num_heads, tgt_len, self.head_dim)}, ' + f'but is {attn_output_v.size()}') + + if attn_output_l.size() != (bsz * self.num_heads * num_cam, src_len, + self.head_dim): + raise ValueError( + '`attn_output_l` should be of size ' + f'{(bsz, self.num_heads, src_len, self.head_dim)}, ' + f'but is {attn_output_l.size()}') + + attn_output_v = attn_output_v.view(bsz, num_cam, self.num_heads, tgt_len, + self.head_dim) + attn_output_v = attn_output_v.transpose(2, 3) + attn_output_v = attn_output_v.reshape(bsz, num_cam, tgt_len, self.embed_dim) + + attn_output_l = attn_output_l.view(bsz, num_cam, self.num_heads, src_len, + self.head_dim) + attn_output_l = attn_output_l.transpose(2, 3) + attn_output_l = attn_output_l.reshape(bsz, num_cam, src_len, self.embed_dim) + + attn_output_v = self.out_v_proj(attn_output_v) + attn_output_l = self.out_l_proj(attn_output_l) + + return attn_output_v, attn_output_l + + +class BiAttentionBlock(nn.Module): + """BiAttentionBlock Module: + + First, multi-level visual features are concat; Then the concat visual + feature and lang feature are fused by attention; Finally the newly visual + feature are split into multi levels. + + Args: + v_dim (int): The dimension of the visual features. + l_dim (int): The dimension of the language feature. + embed_dim (int): The embedding dimension for the attention operation. + num_heads (int): The number of attention heads. + dropout (float, optional): The dropout probability. Defaults to 0.1. + drop_path (float, optional): The drop path probability. + Defaults to 0.0. + init_values (float, optional): + The initial value for the scaling parameter. + Defaults to 1e-4. + """ + + def __init__(self, + v_dim: int, + l_dim: int, + embed_dim: int, + num_heads: int, + dropout: float = 0.1, + drop_path: float = .0, + init_values: float = 1e-4): + super().__init__() + + # pre layer norm + self.layer_norm_v = nn.LayerNorm(v_dim) + self.layer_norm_l = nn.LayerNorm(l_dim) + self.attn = BiMultiHeadAttention( + v_dim=v_dim, + l_dim=l_dim, + embed_dim=embed_dim, + num_heads=num_heads, + dropout=dropout) + + # add layer scale for training stability + self.drop_path = DropPath( + drop_path) if drop_path > 0. else nn.Identity() + self.gamma_v = nn.Parameter( + init_values * torch.ones(v_dim), requires_grad=True) + self.gamma_l = nn.Parameter( + init_values * torch.ones(l_dim), requires_grad=True) + + def forward(self, + vf0: Tensor, + vf1: Tensor, + vf2: Tensor, + vf3: Tensor, + vf4: Tensor, + lang_feature: Tensor, + attention_mask_l=None): + visual_features = [vf0, vf1, vf2, vf3, vf4] + size_per_level, visual_features_flatten = [], [] + for i, feat_per_level in enumerate(visual_features): + bs, c, h, w = feat_per_level.shape + size_per_level.append([h, w]) + feat = permute_and_flatten(feat_per_level, bs, -1, c, h, w) + visual_features_flatten.append(feat) + visual_features_flatten = torch.cat(visual_features_flatten, dim=1) + new_v, new_lang_feature = self.single_attention_call( + visual_features_flatten, + lang_feature, + attention_mask_l=attention_mask_l) + # [bs, N, C] -> [bs, C, N] + new_v = new_v.transpose(1, 2).contiguous() + + start = 0 + # fvfs is mean fusion_visual_features + fvfs = [] + for (h, w) in size_per_level: + new_v_per_level = new_v[:, :, + start:start + h * w].view(bs, -1, h, + w).contiguous() + fvfs.append(new_v_per_level) + start += h * w + + return fvfs[0], fvfs[1], fvfs[2], fvfs[3], fvfs[4], new_lang_feature + + def single_attention_call( + self, + visual: Tensor, + lang: Tensor, + attention_mask_v: Optional[Tensor] = None, + attention_mask_l: Optional[Tensor] = None, + ) -> Tuple[Tensor, Tensor]: + """Perform a single attention call between the visual and language + inputs. + + Args: + visual (Tensor): The visual input tensor. + lang (Tensor): The language input tensor. + attention_mask_v (Optional[Tensor]): + An optional attention mask tensor for the visual input. + attention_mask_l (Optional[Tensor]): + An optional attention mask tensor for the language input. + + Returns: + Tuple[Tensor, Tensor]: A tuple containing the updated + visual and language tensors after the attention call. + """ + visual = self.layer_norm_v(visual) + lang = self.layer_norm_l(lang) + delta_v, delta_l = self.attn( + visual, + lang, + attention_mask_v=attention_mask_v, + attention_mask_l=attention_mask_l) + # visual, lang = visual + delta_v, l + delta_l + visual = visual + self.drop_path(self.gamma_v * delta_v) + lang = lang + self.drop_path(self.gamma_l * delta_l) + return visual, lang + + +class SingleScaleBiAttentionBlock(BiAttentionBlock): + """This is a single-scale implementation of `BiAttentionBlock`. + + The only differenece between it and `BiAttentionBlock` is that the + `forward` function of `SingleScaleBiAttentionBlock` only accepts a single + flatten visual feature map, while the `forward` function in + `BiAttentionBlock` accepts multiple visual feature maps. + """ + + def forward(self, + visual_feature: Tensor, + lang_feature: Tensor, + attention_mask_v=None, + attention_mask_l=None): + """Single-scale forward pass. + + Args: + visual_feature (Tensor): The visual input tensor. Tensor of + shape (bs, patch_len, ch). + lang_feature (Tensor): The language input tensor. Tensor of + shape (bs, text_len, ch). + attention_mask_v (_type_, optional): Visual feature attention + mask. Defaults to None. + attention_mask_l (_type_, optional): Language feature attention + mask.Defaults to None. + """ + new_v, new_lang_feature = self.single_attention_call( + visual_feature, + lang_feature, + attention_mask_v=attention_mask_v, + attention_mask_l=attention_mask_l) + return new_v, new_lang_feature + + +class VLFuse(nn.Module): + """Early Fusion Module. + + Args: + v_dim (int): Dimension of visual features. + l_dim (int): Dimension of language features. + embed_dim (int): The embedding dimension for the attention operation. + num_heads (int): Number of attention heads. + dropout (float): Dropout probability. + drop_path (float): Drop path probability. + use_checkpoint (bool): Whether to use PyTorch's checkpoint function. + """ + + def __init__(self, + v_dim: int = 256, + l_dim: int = 768, + embed_dim: int = 2048, + num_heads: int = 8, + dropout: float = 0.1, + drop_path: float = 0.0, + use_checkpoint: bool = False): + super().__init__() + self.use_checkpoint = use_checkpoint + self.b_attn = BiAttentionBlock( + v_dim=v_dim, + l_dim=l_dim, + embed_dim=embed_dim, + num_heads=num_heads, + dropout=dropout, + drop_path=drop_path, + init_values=1.0 / 6.0) + + def forward(self, x: dict) -> dict: + """Forward pass of the VLFuse module.""" + visual_features = x['visual'] + language_dict_features = x['lang'] + + if self.use_checkpoint: + # vf is mean visual_features + # checkpoint does not allow complex data structures as input, + # such as list, so we must split them. + vf0, vf1, vf2, vf3, vf4, language_features = checkpoint.checkpoint( + self.b_attn, *visual_features, + language_dict_features['hidden'], + language_dict_features['masks']) + else: + vf0, vf1, vf2, vf3, vf4, language_features = self.b_attn( + *visual_features, language_dict_features['hidden'], + language_dict_features['masks']) + + language_dict_features['hidden'] = language_features + fused_language_dict_features = language_dict_features + + features_dict = { + 'visual': [vf0, vf1, vf2, vf3, vf4], + 'lang': fused_language_dict_features + } + + return features_dict + + +class BertEncoderLayer(BertPreTrainedModel): + """A modified version of the `BertLayer` class from the + `transformers.models.bert.modeling_bert` module. + + Args: + config (:class:`~transformers.BertConfig`): + The configuration object that + contains various parameters for the model. + clamp_min_for_underflow (bool, optional): + Whether to clamp the minimum value of the hidden states + to prevent underflow. Defaults to `False`. + clamp_max_for_overflow (bool, optional): + Whether to clamp the maximum value of the hidden states + to prevent overflow. Defaults to `False`. + """ + + def __init__(self, + config: BertConfig, + clamp_min_for_underflow: bool = False, + clamp_max_for_overflow: bool = False): + super().__init__(config) + self.config = config + self.chunk_size_feed_forward = config.chunk_size_feed_forward + self.seq_len_dim = 1 + + self.attention = BertAttention(config, clamp_min_for_underflow, + clamp_max_for_overflow) + self.intermediate = BertIntermediate(config) + self.output = BertOutput(config) + + def forward( + self, inputs: Dict[str, Dict[str, torch.Tensor]] + ) -> Dict[str, Dict[str, torch.Tensor]]: + """Applies the BertEncoderLayer to the input features.""" + language_dict_features = inputs['lang'] + hidden_states = language_dict_features['hidden'] + attention_mask = language_dict_features['masks'] + + device = hidden_states.device + input_shape = hidden_states.size()[:-1] + extended_attention_mask = self.get_extended_attention_mask( + attention_mask, input_shape, device) + + self_attention_outputs = self.attention( + hidden_states, + extended_attention_mask, + None, + output_attentions=False, + past_key_value=None) + attention_output = self_attention_outputs[0] + outputs = self_attention_outputs[1:] + layer_output = apply_chunking_to_forward(self.feed_forward_chunk, + self.chunk_size_feed_forward, + self.seq_len_dim, + attention_output) + outputs = (layer_output, ) + outputs + hidden_states = outputs[0] + + language_dict_features['hidden'] = hidden_states + + features_dict = { + 'visual': inputs['visual'], + 'lang': language_dict_features + } + + return features_dict + + def feed_forward_chunk(self, attention_output: Tensor) -> Tensor: + """Applies the intermediate and output layers of the BertEncoderLayer + to a chunk of the input sequence.""" + intermediate_output = self.intermediate(attention_output) + layer_output = self.output(intermediate_output, attention_output) + return layer_output + + +# The following code is the same as the Huggingface code, +# with the only difference being the additional clamp operation. +class BertSelfAttention(nn.Module): + """BERT self-attention layer from Huggingface transformers. + + Compared to the BertSelfAttention of Huggingface, only add the clamp. + + Args: + config (:class:`~transformers.BertConfig`): + The configuration object that + contains various parameters for the model. + clamp_min_for_underflow (bool, optional): + Whether to clamp the minimum value of the hidden states + to prevent underflow. Defaults to `False`. + clamp_max_for_overflow (bool, optional): + Whether to clamp the maximum value of the hidden states + to prevent overflow. Defaults to `False`. + """ + + def __init__(self, + config: BertConfig, + clamp_min_for_underflow: bool = False, + clamp_max_for_overflow: bool = False): + super().__init__() + if config.hidden_size % config.num_attention_heads != 0 and \ + not hasattr(config, 'embedding_size'): + raise ValueError(f'The hidden size ({config.hidden_size}) is ' + 'not a multiple of the number of attention ' + f'heads ({config.num_attention_heads})') + + self.num_attention_heads = config.num_attention_heads + self.attention_head_size = int(config.hidden_size / + config.num_attention_heads) + self.all_head_size = self.num_attention_heads * \ + self.attention_head_size + + self.query = nn.Linear(config.hidden_size, self.all_head_size) + self.key = nn.Linear(config.hidden_size, self.all_head_size) + self.value = nn.Linear(config.hidden_size, self.all_head_size) + + self.dropout = nn.Dropout(config.attention_probs_dropout_prob) + self.position_embedding_type = getattr(config, + 'position_embedding_type', + 'absolute') + if self.position_embedding_type == 'relative_key' or \ + self.position_embedding_type == 'relative_key_query': + self.max_position_embeddings = config.max_position_embeddings + self.distance_embedding = nn.Embedding( + 2 * config.max_position_embeddings - 1, + self.attention_head_size) + self.clamp_min_for_underflow = clamp_min_for_underflow + self.clamp_max_for_overflow = clamp_max_for_overflow + + self.is_decoder = config.is_decoder + + def transpose_for_scores(self, x: Tensor) -> Tensor: + """Transpose the dimensions of `x`.""" + new_x_shape = x.size()[:-1] + (self.num_attention_heads, + self.attention_head_size) + x = x.view(*new_x_shape) + return x.permute(0, 2, 1, 3) + + def forward( + self, + hidden_states: Tensor, + attention_mask: Optional[Tensor] = None, + head_mask: Optional[Tensor] = None, + encoder_hidden_states: Optional[Tensor] = None, + encoder_attention_mask: Optional[Tensor] = None, + past_key_value: Optional[Tuple[Tensor, Tensor]] = None, + output_attentions: bool = False, + ) -> Tuple[Tensor, ...]: + """Perform a forward pass through the BERT self-attention layer.""" + + mixed_query_layer = self.query(hidden_states) + + # If this is instantiated as a cross-attention module, the keys + # and values come from an encoder; the attention mask needs to be + # such that the encoder's padding tokens are not attended to. + is_cross_attention = encoder_hidden_states is not None + + if is_cross_attention and past_key_value is not None: + # reuse k,v, cross_attentions + key_layer = past_key_value[0] + value_layer = past_key_value[1] + attention_mask = encoder_attention_mask + elif is_cross_attention: + key_layer = self.transpose_for_scores( + self.key(encoder_hidden_states)) + value_layer = self.transpose_for_scores( + self.value(encoder_hidden_states)) + attention_mask = encoder_attention_mask + elif past_key_value is not None: + key_layer = self.transpose_for_scores(self.key(hidden_states)) + value_layer = self.transpose_for_scores(self.value(hidden_states)) + key_layer = torch.cat([past_key_value[0], key_layer], dim=2) + value_layer = torch.cat([past_key_value[1], value_layer], dim=2) + else: + key_layer = self.transpose_for_scores(self.key(hidden_states)) + value_layer = self.transpose_for_scores(self.value(hidden_states)) + + query_layer = self.transpose_for_scores(mixed_query_layer) + + if self.is_decoder: + past_key_value = (key_layer, value_layer) + + # Take the dot product between "query" and "key" + # to get the raw attention scores. + attention_scores = torch.matmul(query_layer, + key_layer.transpose(-1, -2)) + + if self.position_embedding_type == 'relative_key' or \ + self.position_embedding_type == 'relative_key_query': + seq_length = hidden_states.size()[1] + position_ids_l = torch.arange( + seq_length, dtype=torch.long, + device=hidden_states.device).view(-1, 1) + position_ids_r = torch.arange( + seq_length, dtype=torch.long, + device=hidden_states.device).view(1, -1) + distance = position_ids_l - position_ids_r + positional_embedding = self.distance_embedding( + distance + self.max_position_embeddings - 1) + positional_embedding = positional_embedding.to( + dtype=query_layer.dtype) # fp16 compatibility + + if self.position_embedding_type == 'relative_key': + relative_position_scores = torch.einsum( + 'bhld,lrd->bhlr', query_layer, positional_embedding) + attention_scores = attention_scores + relative_position_scores + elif self.position_embedding_type == 'relative_key_query': + relative_position_scores_query = torch.einsum( + 'bhld,lrd->bhlr', query_layer, positional_embedding) + relative_position_scores_key = torch.einsum( + 'bhrd,lrd->bhlr', key_layer, positional_embedding) + attention_scores = attention_scores + \ + relative_position_scores_query + \ + relative_position_scores_key + + attention_scores = attention_scores / math.sqrt( + self.attention_head_size) + + if self.clamp_min_for_underflow: + attention_scores = torch.clamp( + attention_scores, min=-MAX_CLAMP_VALUE + ) # Do not increase -50000, data type half has quite limited range + if self.clamp_max_for_overflow: + attention_scores = torch.clamp( + attention_scores, max=MAX_CLAMP_VALUE + ) # Do not increase 50000, data type half has quite limited range + + if attention_mask is not None: + # Apply the attention mask is + # (precomputed for all layers in BertModel forward() function) + attention_scores = attention_scores + attention_mask + + # Normalize the attention scores to probabilities. + attention_probs = nn.Softmax(dim=-1)(attention_scores) + + # This is actually dropping out entire tokens to attend to, which might + # seem a bit unusual, but is taken from the original Transformer paper. + attention_probs = self.dropout(attention_probs) + + # Mask heads if we want to + if head_mask is not None: + attention_probs = attention_probs * head_mask + + context_layer = torch.matmul(attention_probs, value_layer) + + context_layer = context_layer.permute(0, 2, 1, 3).contiguous() + new_context_layer_shape = context_layer.size()[:-2] + ( + self.all_head_size, ) + context_layer = context_layer.view(*new_context_layer_shape) + + outputs = (context_layer, + attention_probs) if output_attentions else (context_layer, ) + + if self.is_decoder: + outputs = outputs + (past_key_value, ) + return outputs + + +class BertAttention(HFBertAttention): + """BertAttention is made up of self-attention and intermediate+output. + + Compared to the BertAttention of Huggingface, only add the clamp. + + Args: + config (:class:`~transformers.BertConfig`): + The configuration object that + contains various parameters for the model. + clamp_min_for_underflow (bool, optional): + Whether to clamp the minimum value of the hidden states + to prevent underflow. Defaults to `False`. + clamp_max_for_overflow (bool, optional): + Whether to clamp the maximum value of the hidden states + to prevent overflow. Defaults to `False`. + """ + + def __init__(self, + config: BertConfig, + clamp_min_for_underflow: bool = False, + clamp_max_for_overflow: bool = False): + super().__init__(config) + self.self = BertSelfAttention(config, clamp_min_for_underflow, + clamp_max_for_overflow) + + +class BertIntermediate(HFBertIntermediate): + """Modified from transformers.models.bert.modeling_bert.BertIntermediate. + + Compared to the BertIntermediate of Huggingface, only add the clamp. + """ + + def forward(self, hidden_states: Tensor) -> Tensor: + hidden_states = self.dense(hidden_states) + hidden_states = clamp_values(hidden_states) + hidden_states = self.intermediate_act_fn(hidden_states) + hidden_states = clamp_values(hidden_states) + return hidden_states + + +class BertOutput(HFBertOutput): + """Modified from transformers.models.bert.modeling_bert.BertOutput. + + Compared to the BertOutput of Huggingface, only add the clamp. + """ + + def forward(self, hidden_states: Tensor, input_tensor: Tensor) -> Tensor: + hidden_states = self.dense(hidden_states) + hidden_states = self.dropout(hidden_states) + hidden_states = clamp_values(hidden_states) + hidden_states = self.LayerNorm(hidden_states + input_tensor) + hidden_states = clamp_values(hidden_states) + return hidden_states