search

XuyangBai / TransFusion

[PyTorch] Official implementation of CVPR2022 paper "TransFusion: Robust LiDAR-Camera Fusion for 3D Object Detection with Transformers". https://arxiv.org/abs/2203.11496
Apache License 2.0
613 stars 76 forks source link

Weird behavoir on ViewofDelft (KITTI-Format) #66

Open LeonRuddat opened 1 year ago

LeonRuddat commented 1 year ago

Hey @XuyangBai, First of all, I would like to congratulate you on your great work. It is truly an awesome contribution! I've been trying to apply TransFusion on the ViewOfDelft-Dataset (VOD). The data (I Use the radar data) set is in KITTI format. Therefore I have taken the configuration file from #51 and adapted it. My config now looks as follows:

point_cloud_range = [-75.2, -75.2, -4, 75.2, 75.2, 2]
class_names = ['Car', 'Pedestrian', 'Cyclist']
voxel_size = [0.1, 0.1, 0.15]
out_size_factor = 8
evaluation = dict(interval=1)
dataset_type = 'KittiDataset'
data_root = '/home/labor/ProjekteRuddat/TransFusion/data/kitti/'
input_modality = dict(
    use_lidar=True,
    use_camera=False,
    use_radar=False,
    use_map=False,
    use_external=False)
train_pipeline = [
    dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=7, use_dim=4),
    dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True),
    dict(
        type='PointsRangeFilter',
        point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
    dict(
        type='ObjectRangeFilter',
        point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
    dict(type='ObjectNameFilter', classes=['Car', 'Pedestrian', 'Cyclist']),
    dict(type='PointShuffle'),
    dict(
        type='DefaultFormatBundle3D',
        class_names=['Car', 'Pedestrian', 'Cyclist']),
    dict(type='Collect3D', keys=['points', 'gt_bboxes_3d', 'gt_labels_3d'])
]
test_pipeline = [
    dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=7, use_dim=4),
    dict(
        type='MultiScaleFlipAug3D',
        img_scale=(800, 1333),
        pts_scale_ratio=1,
        flip=False,
        transforms=[
            dict(
                type='PointsRangeFilter',
                point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
            dict(
                type='DefaultFormatBundle3D',
                class_names=['Car', 'Pedestrian', 'Cyclist'],
                with_label=False),
            dict(type='Collect3D', keys=['points'])
        ])
]
data = dict(
    samples_per_gpu=1,
    workers_per_gpu=4,
    train=dict(
        type='RepeatDataset',
        times=1,
        dataset=dict(
            type='KittiDataset',
            data_root='/home/labor/ProjekteRuddat/TransFusion/data/kitti/',
            ann_file=
            '/home/labor/ProjekteRuddat/TransFusion/data/kitti/kitti_infos_train.pkl',
            split='training',
            pipeline=[
                dict(
                    type='LoadPointsFromFile',
                    coord_type='LIDAR',
                    load_dim=7,
                    use_dim=4),
                dict(
                    type='LoadAnnotations3D',
                    with_bbox_3d=True,
                    with_label_3d=True),
                dict(
                    type='PointsRangeFilter',
                    point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
                dict(
                    type='ObjectRangeFilter',
                    point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
                dict(
                    type='ObjectNameFilter',
                    classes=['Car', 'Pedestrian', 'Cyclist']),
                dict(type='PointShuffle'),
                dict(
                    type='DefaultFormatBundle3D',
                    class_names=['Car', 'Pedestrian', 'Cyclist']),
                dict(
                    type='Collect3D',
                    keys=['points', 'gt_bboxes_3d', 'gt_labels_3d'])
            ],
            modality=dict(
                use_lidar=True,
                use_camera=False,
                use_radar=False,
                use_map=False,
                use_external=False),
            classes=['Car', 'Pedestrian', 'Cyclist'],
            test_mode=False,
            box_type_3d='LiDAR')),
    val=dict(
        type='KittiDataset',
        data_root='/home/labor/ProjekteRuddat/TransFusion/data/kitti/',
        ann_file=
        '/home/labor/ProjekteRuddat/TransFusion/data/kitti/kitti_infos_val.pkl',
        split='training',
        pipeline=[
            dict(
                type='LoadPointsFromFile',
                coord_type='LIDAR',
                load_dim=7,
                use_dim=4),
            dict(
                type='MultiScaleFlipAug3D',
                img_scale=(800, 1333),
                pts_scale_ratio=1,
                flip=False,
                transforms=[
                    dict(
                        type='PointsRangeFilter',
                        point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
                    dict(
                        type='DefaultFormatBundle3D',
                        class_names=['Car', 'Pedestrian', 'Cyclist'],
                        with_label=False),
                    dict(type='Collect3D', keys=['points'])
                ])
        ],
        modality=dict(
            use_lidar=True,
            use_camera=False,
            use_radar=False,
            use_map=False,
            use_external=False),
        classes=['Car', 'Pedestrian', 'Cyclist'],
        test_mode=True,
        box_type_3d='LiDAR'),
    test=dict(
        type='KittiDataset',
        data_root='/home/labor/ProjekteRuddat/TransFusion/data/kitti/',
        ann_file=
        '/home/labor/ProjekteRuddat/TransFusion/data/kitti/kitti_infos_val.pkl',
        split='training',
        pipeline=[
            dict(
                type='LoadPointsFromFile',
                coord_type='LIDAR',
                load_dim=7,
                use_dim=4),
            dict(
                type='MultiScaleFlipAug3D',
                img_scale=(800, 1333),
                pts_scale_ratio=1,
                flip=False,
                transforms=[
                    dict(
                        type='PointsRangeFilter',
                        point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
                    dict(
                        type='DefaultFormatBundle3D',
                        class_names=['Car', 'Pedestrian', 'Cyclist'],
                        with_label=False),
                    dict(type='Collect3D', keys=['points'])
                ])
        ],
        modality=dict(
            use_lidar=True,
            use_camera=False,
            use_radar=False,
            use_map=False,
            use_external=False),
        classes=['Car', 'Pedestrian', 'Cyclist'],
        test_mode=True,
        box_type_3d='LiDAR'))
model = dict(
    type='TransFusionDetector',
    pts_voxel_layer=dict(
        max_num_points=5,
        voxel_size=[0.1, 0.1, 0.15],
        max_voxels=(16000, 40000),
        point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
    pts_voxel_encoder=dict(
        type='HardVFE',
        in_channels=4,
        feat_channels=[64],
        with_distance=False,
        with_cluster_center=False,
        with_voxel_center=False,
        voxel_size=[0.1, 0.1, 0.15],
        norm_cfg=dict(type='BN1d', eps=0.001, momentum=0.01),
        point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2]),
    pts_middle_encoder=dict(
        type='SparseEncoder',
        in_channels=64,
        output_channels=128,
        sparse_shape=[41, 1504, 1504],
        order=('conv', 'norm', 'act'),
        encoder_channels=((16, 16, 32), (32, 32, 64), (64, 64, 128), (128,
                                                                      128)),
        encoder_paddings=((0, 0, 1), (0, 0, 1), (0, 0, [0, 1, 1]), (0, 0)),
        block_type='basicblock'),
    pts_backbone=dict(
        type='SECOND',
        in_channels=256,
        out_channels=[128, 256],
        layer_nums=[5, 5],
        layer_strides=[1, 2],
        norm_cfg=dict(type='BN', eps=0.001, momentum=0.01),
        conv_cfg=dict(type='Conv2d', bias=False)),
    pts_neck=dict(
        type='SECONDFPN',
        in_channels=[128, 256],
        out_channels=[256, 256],
        upsample_strides=[1, 2],
        norm_cfg=dict(type='BN', eps=0.001, momentum=0.01),
        upsample_cfg=dict(type='deconv', bias=False),
        use_conv_for_no_stride=True),
    pts_bbox_head=dict(
        type='TransFusionHead',
        num_proposals=200,
        auxiliary=True,
        in_channels=512,
        hidden_channel=128,
        num_classes=3,
        num_decoder_layers=1,
        num_heads=8,
        learnable_query_pos=False,
        initialize_by_heatmap=True,
        nms_kernel_size=3,
        ffn_channel=256,
        dropout=0.1,
        bn_momentum=0.1,
        activation='relu',
        common_heads=dict(
            center=(2, 2), height=(1, 2), dim=(3, 2), rot=(2, 2)),
        bbox_coder=dict(
            type='TransFusionBBoxCoder',
            pc_range=[-75.2, -75.2],
            voxel_size=[0.1, 0.1],
            out_size_factor=8,
            post_center_range=[-80, -80, -10.0, 80, 80, 10.0],
            score_threshold=0.0,
            code_size=8),
        loss_cls=dict(
            type='FocalLoss',
            use_sigmoid=True,
            gamma=2,
            alpha=0.25,
            reduction='mean',
            loss_weight=1.0),
        loss_bbox=dict(type='L1Loss', reduction='mean', loss_weight=2.0),
        loss_heatmap=dict(
            type='GaussianFocalLoss', reduction='mean', loss_weight=1.5)),
    train_cfg=dict(
        pts=dict(
            dataset='kitti',
            assigner=dict(
                type='HungarianAssigner3D',
                iou_calculator=dict(type='BboxOverlaps3D', coordinate='lidar'),
                cls_cost=dict(
                    type='FocalLossCost', gamma=2, alpha=0.25, weight=0.6),
                reg_cost=dict(type='BBox3DL1Cost', weight=0.5),
                iou_cost=dict(type='IoU3DCost', weight=2)),
            pos_weight=-1,
            gaussian_overlap=0.1,
            min_radius=2,
            grid_size=[1504, 1504, 40],
            voxel_size=[0.1, 0.1, 0.15],
            out_size_factor=8,
            code_weights=[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0],
            point_cloud_range=[-75.2, -75.2, -4, 75.2, 75.2, 2])),
    test_cfg=dict(
        pts=dict(
            dataset='kitti',
            grid_size=[1504, 1504, 40],
            out_size_factor=8,
            pc_range=[-75.2, -75.2],
            voxel_size=[0.1, 0.1],
            nms_type=None)))
optimizer = dict(type='AdamW', lr=8e-06, weight_decay=0.01)
optimizer_config = dict(grad_clip=dict(max_norm=0.1, norm_type=2))
lr_config = dict(
    policy='cyclic',
    target_ratio=(2, 0.0001),
    cyclic_times=1,
    step_ratio_up=0.4)
momentum_config = dict(
    policy='cyclic',
    target_ratio=(0.8947368421052632, 1),
    cyclic_times=1,
    step_ratio_up=0.4)
total_epochs = 40
checkpoint_config = dict(interval=1)
log_config = dict(
    interval=50,
    hooks=[dict(type='TextLoggerHook'),
           dict(type='TensorboardLoggerHook')])
dist_params = dict(backend='nccl')
log_level = 'INFO'
work_dir = './work_dirs/LC_Radar_Test'
load_from = None
resume_from = None
workflow = [('train', 1)]
gpu_ids = range(0, 1)

Note that the original VOD point_cloud_range is [0, -25.6, -3, 51.2, 25.6, 2]. But if I change it accordingly, the model does not even converge. Might be relatet to #17 and #64 as the bbox_loss does not decrease. So my First Question is if you know how I can configure the loss (or architecture) so that I can run the original point_cloud_range. The Radar Set of VOD provides an Nx7 array as Input. For simplicity I only used RCS in addition to the Points (Nx4). However, the behavior I will now show does not change when I use Nx7. I also disabled augmentaion as some of the methods don't work with the Radar Data -> described in the VOD-Paper

I will try to explain my Second Question based on the log of my last training epoch. Full Log can be found here. Throughout the training (in all epochs) the loss decreases continuously. Also the matched_ious looks not to bad. However in the eval I only get zeros and some weird values. Is there something wrong with my config or do you have any Idea from where this weird behavoir originates frome? 

2022-10-18 05:23:05,156 - mmdet - INFO - Epoch [40][50/5139]    lr: 6.794e-08, eta: 0:22:21, time: 0.311, data_time: 0.046, memory: 1983, loss_heatmap: 0.0103, layer_-1_loss_cls: 0.0381, layer_-1_loss_bbox: 2.6234, matched_ious: 0.7535, loss: 2.6718, grad_norm: 103.8107
2022-10-18 05:23:18,337 - mmdet - INFO - Epoch [40][100/5139]   lr: 6.663e-08, eta: 0:22:08, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0108, layer_-1_loss_cls: 0.0575, layer_-1_loss_bbox: 2.1234, matched_ious: 0.7663, loss: 2.1917, grad_norm: 98.8591
2022-10-18 05:23:31,529 - mmdet - INFO - Epoch [40][150/5139]   lr: 6.533e-08, eta: 0:21:55, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0135, layer_-1_loss_cls: 0.0488, layer_-1_loss_bbox: 3.0146, matched_ious: 0.7286, loss: 3.0769, grad_norm: 113.0453
2022-10-18 05:23:44,721 - mmdet - INFO - Epoch [40][200/5139]   lr: 6.405e-08, eta: 0:21:42, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0136, layer_-1_loss_cls: 0.0429, layer_-1_loss_bbox: 2.5382, matched_ious: 0.7547, loss: 2.5946, grad_norm: 101.9943
2022-10-18 05:23:57,984 - mmdet - INFO - Epoch [40][250/5139]   lr: 6.277e-08, eta: 0:21:29, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0164, layer_-1_loss_cls: 0.0529, layer_-1_loss_bbox: 2.8543, matched_ious: 0.7531, loss: 2.9236, grad_norm: 99.8835
2022-10-18 05:24:11,148 - mmdet - INFO - Epoch [40][300/5139]   lr: 6.152e-08, eta: 0:21:16, time: 0.263, data_time: 0.001, memory: 1983, loss_heatmap: 0.0111, layer_-1_loss_cls: 0.0475, layer_-1_loss_bbox: 3.5269, matched_ious: 0.7434, loss: 3.5855, grad_norm: 102.4033
2022-10-18 05:24:24,422 - mmdet - INFO - Epoch [40][350/5139]   lr: 6.027e-08, eta: 0:21:02, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0140, layer_-1_loss_cls: 0.0724, layer_-1_loss_bbox: 4.0681, matched_ious: 0.7039, loss: 4.1545, grad_norm: 106.6865
2022-10-18 05:24:37,612 - mmdet - INFO - Epoch [40][400/5139]   lr: 5.904e-08, eta: 0:20:49, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0106, layer_-1_loss_cls: 0.0449, layer_-1_loss_bbox: 2.2486, matched_ious: 0.7485, loss: 2.3040, grad_norm: 107.8064
2022-10-18 05:24:50,865 - mmdet - INFO - Epoch [40][450/5139]   lr: 5.781e-08, eta: 0:20:36, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0145, layer_-1_loss_cls: 0.0533, layer_-1_loss_bbox: 1.9624, matched_ious: 0.7838, loss: 2.0302, grad_norm: 93.2923
2022-10-18 05:25:04,106 - mmdet - INFO - Epoch [40][500/5139]   lr: 5.661e-08, eta: 0:20:23, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0089, layer_-1_loss_cls: 0.0447, layer_-1_loss_bbox: 2.4725, matched_ious: 0.7589, loss: 2.5261, grad_norm: 89.1439
2022-10-18 05:25:17,340 - mmdet - INFO - Epoch [40][550/5139]   lr: 5.541e-08, eta: 0:20:10, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0071, layer_-1_loss_cls: 0.0472, layer_-1_loss_bbox: 2.4250, matched_ious: 0.8080, loss: 2.4793, grad_norm: 104.2526
2022-10-18 05:25:30,517 - mmdet - INFO - Epoch [40][600/5139]   lr: 5.423e-08, eta: 0:19:56, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0087, layer_-1_loss_cls: 0.0363, layer_-1_loss_bbox: 1.7756, matched_ious: 0.8029, loss: 1.8206, grad_norm: 76.8236
2022-10-18 05:25:43,780 - mmdet - INFO - Epoch [40][650/5139]   lr: 5.306e-08, eta: 0:19:43, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0195, layer_-1_loss_cls: 0.0599, layer_-1_loss_bbox: 4.9168, matched_ious: 0.6770, loss: 4.9962, grad_norm: 113.0495
2022-10-18 05:25:57,052 - mmdet - INFO - Epoch [40][700/5139]   lr: 5.190e-08, eta: 0:19:30, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0110, layer_-1_loss_cls: 0.0374, layer_-1_loss_bbox: 1.7493, matched_ious: 0.7830, loss: 1.7977, grad_norm: 83.9731
2022-10-18 05:26:10,316 - mmdet - INFO - Epoch [40][750/5139]   lr: 5.076e-08, eta: 0:19:17, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0101, layer_-1_loss_cls: 0.0402, layer_-1_loss_bbox: 2.2662, matched_ious: 0.7525, loss: 2.3165, grad_norm: 106.3180
2022-10-18 05:26:23,525 - mmdet - INFO - Epoch [40][800/5139]   lr: 4.963e-08, eta: 0:19:04, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0101, layer_-1_loss_cls: 0.0496, layer_-1_loss_bbox: 2.4821, matched_ious: 0.7727, loss: 2.5418, grad_norm: 92.4684
2022-10-18 05:26:36,777 - mmdet - INFO - Epoch [40][850/5139]   lr: 4.851e-08, eta: 0:18:51, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0107, layer_-1_loss_cls: 0.0387, layer_-1_loss_bbox: 2.0138, matched_ious: 0.7605, loss: 2.0632, grad_norm: 93.5884
2022-10-18 05:26:49,999 - mmdet - INFO - Epoch [40][900/5139]   lr: 4.741e-08, eta: 0:18:37, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0134, layer_-1_loss_cls: 0.0405, layer_-1_loss_bbox: 2.5588, matched_ious: 0.7525, loss: 2.6128, grad_norm: 104.4979
2022-10-18 05:27:03,277 - mmdet - INFO - Epoch [40][950/5139]   lr: 4.632e-08, eta: 0:18:24, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0123, layer_-1_loss_cls: 0.0547, layer_-1_loss_bbox: 2.7537, matched_ious: 0.7439, loss: 2.8208, grad_norm: 108.3117
2022-10-18 05:27:16,552 - mmdet - INFO - Epoch [40][1000/5139]  lr: 4.524e-08, eta: 0:18:11, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0186, layer_-1_loss_cls: 0.0424, layer_-1_loss_bbox: 3.3234, matched_ious: 0.7347, loss: 3.3844, grad_norm: 97.5142
2022-10-18 05:27:29,863 - mmdet - INFO - Epoch [40][1050/5139]  lr: 4.417e-08, eta: 0:17:58, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0112, layer_-1_loss_cls: 0.0419, layer_-1_loss_bbox: 2.2556, matched_ious: 0.7440, loss: 2.3087, grad_norm: 89.1655
2022-10-18 05:27:43,133 - mmdet - INFO - Epoch [40][1100/5139]  lr: 4.312e-08, eta: 0:17:45, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0104, layer_-1_loss_cls: 0.0492, layer_-1_loss_bbox: 2.8029, matched_ious: 0.7531, loss: 2.8625, grad_norm: 104.3726
2022-10-18 05:27:56,397 - mmdet - INFO - Epoch [40][1150/5139]  lr: 4.208e-08, eta: 0:17:31, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0150, layer_-1_loss_cls: 0.0542, layer_-1_loss_bbox: 4.2748, matched_ious: 0.7577, loss: 4.3440, grad_norm: 101.7675
2022-10-18 05:28:09,721 - mmdet - INFO - Epoch [40][1200/5139]  lr: 4.105e-08, eta: 0:17:18, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0140, layer_-1_loss_cls: 0.0596, layer_-1_loss_bbox: 3.3293, matched_ious: 0.7104, loss: 3.4029, grad_norm: 123.4371
2022-10-18 05:28:22,961 - mmdet - INFO - Epoch [40][1250/5139]  lr: 4.004e-08, eta: 0:17:05, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0168, layer_-1_loss_cls: 0.0773, layer_-1_loss_bbox: 4.5674, matched_ious: 0.6956, loss: 4.6614, grad_norm: 140.9851
2022-10-18 05:28:36,221 - mmdet - INFO - Epoch [40][1300/5139]  lr: 3.904e-08, eta: 0:16:52, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0141, layer_-1_loss_cls: 0.0627, layer_-1_loss_bbox: 3.3123, matched_ious: 0.7350, loss: 3.3890, grad_norm: 124.3009
2022-10-18 05:28:49,443 - mmdet - INFO - Epoch [40][1350/5139]  lr: 3.805e-08, eta: 0:16:39, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0101, layer_-1_loss_cls: 0.0479, layer_-1_loss_bbox: 2.6063, matched_ious: 0.7337, loss: 2.6644, grad_norm: 164.3633
2022-10-18 05:29:02,686 - mmdet - INFO - Epoch [40][1400/5139]  lr: 3.707e-08, eta: 0:16:26, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0159, layer_-1_loss_cls: 0.0537, layer_-1_loss_bbox: 2.5215, matched_ious: 0.7352, loss: 2.5911, grad_norm: 100.4110
2022-10-18 05:29:15,974 - mmdet - INFO - Epoch [40][1450/5139]  lr: 3.611e-08, eta: 0:16:12, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0133, layer_-1_loss_cls: 0.0366, layer_-1_loss_bbox: 2.6909, matched_ious: 0.7415, loss: 2.7408, grad_norm: 119.2465
2022-10-18 05:29:29,235 - mmdet - INFO - Epoch [40][1500/5139]  lr: 3.516e-08, eta: 0:15:59, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0125, layer_-1_loss_cls: 0.0609, layer_-1_loss_bbox: 2.7653, matched_ious: 0.7289, loss: 2.8387, grad_norm: 117.3499
2022-10-18 05:29:42,441 - mmdet - INFO - Epoch [40][1550/5139]  lr: 3.422e-08, eta: 0:15:46, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0112, layer_-1_loss_cls: 0.0394, layer_-1_loss_bbox: 1.9499, matched_ious: 0.7625, loss: 2.0006, grad_norm: 124.2124
2022-10-18 05:29:55,694 - mmdet - INFO - Epoch [40][1600/5139]  lr: 3.330e-08, eta: 0:15:33, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0121, layer_-1_loss_cls: 0.0541, layer_-1_loss_bbox: 3.2342, matched_ious: 0.7495, loss: 3.3003, grad_norm: 113.5560
2022-10-18 05:30:08,939 - mmdet - INFO - Epoch [40][1650/5139]  lr: 3.239e-08, eta: 0:15:20, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0111, layer_-1_loss_cls: 0.0498, layer_-1_loss_bbox: 2.3069, matched_ious: 0.7627, loss: 2.3678, grad_norm: 99.8662
2022-10-18 05:30:22,197 - mmdet - INFO - Epoch [40][1700/5139]  lr: 3.149e-08, eta: 0:15:06, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0104, layer_-1_loss_cls: 0.0504, layer_-1_loss_bbox: 2.3275, matched_ious: 0.7422, loss: 2.3883, grad_norm: 86.6703
2022-10-18 05:30:35,410 - mmdet - INFO - Epoch [40][1750/5139]  lr: 3.060e-08, eta: 0:14:53, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0110, layer_-1_loss_cls: 0.0479, layer_-1_loss_bbox: 2.1438, matched_ious: 0.7589, loss: 2.2027, grad_norm: 103.8670
2022-10-18 05:30:48,652 - mmdet - INFO - Epoch [40][1800/5139]  lr: 2.973e-08, eta: 0:14:40, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0099, layer_-1_loss_cls: 0.0471, layer_-1_loss_bbox: 3.8404, matched_ious: 0.7844, loss: 3.8975, grad_norm: 82.8128
2022-10-18 05:31:01,986 - mmdet - INFO - Epoch [40][1850/5139]  lr: 2.887e-08, eta: 0:14:27, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0165, layer_-1_loss_cls: 0.0385, layer_-1_loss_bbox: 2.1723, matched_ious: 0.7434, loss: 2.2273, grad_norm: 101.6845
2022-10-18 05:31:15,263 - mmdet - INFO - Epoch [40][1900/5139]  lr: 2.803e-08, eta: 0:14:14, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0138, layer_-1_loss_cls: 0.0561, layer_-1_loss_bbox: 3.1756, matched_ious: 0.7197, loss: 3.2456, grad_norm: 124.6407
2022-10-18 05:31:28,491 - mmdet - INFO - Epoch [40][1950/5139]  lr: 2.719e-08, eta: 0:14:00, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0142, layer_-1_loss_cls: 0.0622, layer_-1_loss_bbox: 2.7561, matched_ious: 0.7221, loss: 2.8325, grad_norm: 106.6449
2022-10-18 05:31:41,768 - mmdet - INFO - Epoch [40][2000/5139]  lr: 2.637e-08, eta: 0:13:47, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0089, layer_-1_loss_cls: 0.0469, layer_-1_loss_bbox: 2.4486, matched_ious: 0.7659, loss: 2.5044, grad_norm: 100.2331
2022-10-18 05:31:54,983 - mmdet - INFO - Epoch [40][2050/5139]  lr: 2.557e-08, eta: 0:13:34, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0173, layer_-1_loss_cls: 0.0565, layer_-1_loss_bbox: 3.5373, matched_ious: 0.7011, loss: 3.6111, grad_norm: 101.4196
2022-10-18 05:32:08,146 - mmdet - INFO - Epoch [40][2100/5139]  lr: 2.477e-08, eta: 0:13:21, time: 0.263, data_time: 0.001, memory: 1983, loss_heatmap: 0.0116, layer_-1_loss_cls: 0.0494, layer_-1_loss_bbox: 2.9343, matched_ious: 0.7480, loss: 2.9953, grad_norm: 110.2647
2022-10-18 05:32:21,392 - mmdet - INFO - Epoch [40][2150/5139]  lr: 2.399e-08, eta: 0:13:08, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0136, layer_-1_loss_cls: 0.0526, layer_-1_loss_bbox: 3.9963, matched_ious: 0.7543, loss: 4.0626, grad_norm: 106.7714
2022-10-18 05:32:34,831 - mmdet - INFO - Epoch [40][2200/5139]  lr: 2.322e-08, eta: 0:12:55, time: 0.269, data_time: 0.001, memory: 1983, loss_heatmap: 0.0064, layer_-1_loss_cls: 0.0606, layer_-1_loss_bbox: 3.7189, matched_ious: 0.7866, loss: 3.7860, grad_norm: 105.3137
2022-10-18 05:32:48,048 - mmdet - INFO - Epoch [40][2250/5139]  lr: 2.246e-08, eta: 0:12:41, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0072, layer_-1_loss_cls: 0.0450, layer_-1_loss_bbox: 2.5486, matched_ious: 0.7834, loss: 2.6008, grad_norm: 134.8113
2022-10-18 05:33:01,374 - mmdet - INFO - Epoch [40][2300/5139]  lr: 2.172e-08, eta: 0:12:28, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0095, layer_-1_loss_cls: 0.0416, layer_-1_loss_bbox: 1.7599, matched_ious: 0.7671, loss: 1.8110, grad_norm: 111.5903
2022-10-18 05:33:14,628 - mmdet - INFO - Epoch [40][2350/5139]  lr: 2.099e-08, eta: 0:12:15, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0093, layer_-1_loss_cls: 0.0482, layer_-1_loss_bbox: 2.6937, matched_ious: 0.7173, loss: 2.7512, grad_norm: 127.4551
2022-10-18 05:33:27,882 - mmdet - INFO - Epoch [40][2400/5139]  lr: 2.028e-08, eta: 0:12:02, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0126, layer_-1_loss_cls: 0.0423, layer_-1_loss_bbox: 2.2484, matched_ious: 0.7639, loss: 2.3033, grad_norm: 100.2641
2022-10-18 05:33:41,196 - mmdet - INFO - Epoch [40][2450/5139]  lr: 1.957e-08, eta: 0:11:49, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0204, layer_-1_loss_cls: 0.0452, layer_-1_loss_bbox: 1.8809, matched_ious: 0.7600, loss: 1.9465, grad_norm: 85.8562
2022-10-18 05:33:54,502 - mmdet - INFO - Epoch [40][2500/5139]  lr: 1.888e-08, eta: 0:11:35, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0098, layer_-1_loss_cls: 0.0471, layer_-1_loss_bbox: 3.1378, matched_ious: 0.7893, loss: 3.1947, grad_norm: 90.9200
2022-10-18 05:34:07,774 - mmdet - INFO - Epoch [40][2550/5139]  lr: 1.820e-08, eta: 0:11:22, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0097, layer_-1_loss_cls: 0.0618, layer_-1_loss_bbox: 3.6538, matched_ious: 0.7456, loss: 3.7254, grad_norm: 113.8491
2022-10-18 05:34:21,031 - mmdet - INFO - Epoch [40][2600/5139]  lr: 1.754e-08, eta: 0:11:09, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0079, layer_-1_loss_cls: 0.0383, layer_-1_loss_bbox: 1.7205, matched_ious: 0.7859, loss: 1.7668, grad_norm: 106.0450
2022-10-18 05:34:34,344 - mmdet - INFO - Epoch [40][2650/5139]  lr: 1.688e-08, eta: 0:10:56, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0170, layer_-1_loss_cls: 0.0662, layer_-1_loss_bbox: 4.3113, matched_ious: 0.7074, loss: 4.3945, grad_norm: 121.0071
2022-10-18 05:34:47,600 - mmdet - INFO - Epoch [40][2700/5139]  lr: 1.625e-08, eta: 0:10:43, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0135, layer_-1_loss_cls: 0.0544, layer_-1_loss_bbox: 3.4987, matched_ious: 0.7328, loss: 3.5666, grad_norm: 111.4469
2022-10-18 05:35:00,873 - mmdet - INFO - Epoch [40][2750/5139]  lr: 1.562e-08, eta: 0:10:30, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0120, layer_-1_loss_cls: 0.0488, layer_-1_loss_bbox: 2.3933, matched_ious: 0.7258, loss: 2.4542, grad_norm: 128.3246
2022-10-18 05:35:14,138 - mmdet - INFO - Epoch [40][2800/5139]  lr: 1.501e-08, eta: 0:10:16, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0133, layer_-1_loss_cls: 0.0560, layer_-1_loss_bbox: 2.6332, matched_ious: 0.7410, loss: 2.7025, grad_norm: 138.2998
2022-10-18 05:35:27,399 - mmdet - INFO - Epoch [40][2850/5139]  lr: 1.441e-08, eta: 0:10:03, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0112, layer_-1_loss_cls: 0.0477, layer_-1_loss_bbox: 2.2311, matched_ious: 0.7656, loss: 2.2900, grad_norm: 123.3121
2022-10-18 05:35:40,688 - mmdet - INFO - Epoch [40][2900/5139]  lr: 1.382e-08, eta: 0:09:50, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0103, layer_-1_loss_cls: 0.0569, layer_-1_loss_bbox: 4.2099, matched_ious: 0.7381, loss: 4.2771, grad_norm: 98.5493
2022-10-18 05:35:53,955 - mmdet - INFO - Epoch [40][2950/5139]  lr: 1.324e-08, eta: 0:09:37, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0096, layer_-1_loss_cls: 0.0537, layer_-1_loss_bbox: 2.7439, matched_ious: 0.7576, loss: 2.8072, grad_norm: 114.0438
2022-10-18 05:36:07,195 - mmdet - INFO - Epoch [40][3000/5139]  lr: 1.268e-08, eta: 0:09:24, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0123, layer_-1_loss_cls: 0.0508, layer_-1_loss_bbox: 2.2310, matched_ious: 0.7720, loss: 2.2941, grad_norm: 91.9004
2022-10-18 05:36:20,407 - mmdet - INFO - Epoch [40][3050/5139]  lr: 1.213e-08, eta: 0:09:10, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0146, layer_-1_loss_cls: 0.0416, layer_-1_loss_bbox: 3.0102, matched_ious: 0.7372, loss: 3.0665, grad_norm: 94.5921
2022-10-18 05:36:33,635 - mmdet - INFO - Epoch [40][3100/5139]  lr: 1.160e-08, eta: 0:08:57, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0123, layer_-1_loss_cls: 0.0442, layer_-1_loss_bbox: 2.4673, matched_ious: 0.7365, loss: 2.5238, grad_norm: 110.1415
2022-10-18 05:36:46,941 - mmdet - INFO - Epoch [40][3150/5139]  lr: 1.107e-08, eta: 0:08:44, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0112, layer_-1_loss_cls: 0.0383, layer_-1_loss_bbox: 1.5265, matched_ious: 0.7942, loss: 1.5760, grad_norm: 72.1967
2022-10-18 05:37:00,189 - mmdet - INFO - Epoch [40][3200/5139]  lr: 1.057e-08, eta: 0:08:31, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0112, layer_-1_loss_cls: 0.0484, layer_-1_loss_bbox: 2.7638, matched_ious: 0.7328, loss: 2.8234, grad_norm: 161.9037
2022-10-18 05:37:13,486 - mmdet - INFO - Epoch [40][3250/5139]  lr: 1.007e-08, eta: 0:08:18, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0140, layer_-1_loss_cls: 0.0514, layer_-1_loss_bbox: 3.2418, matched_ious: 0.7438, loss: 3.3072, grad_norm: 105.0062
2022-10-18 05:37:26,729 - mmdet - INFO - Epoch [40][3300/5139]  lr: 9.584e-09, eta: 0:08:04, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0112, layer_-1_loss_cls: 0.0574, layer_-1_loss_bbox: 3.0303, matched_ious: 0.7299, loss: 3.0988, grad_norm: 106.7709
2022-10-18 05:37:40,063 - mmdet - INFO - Epoch [40][3350/5139]  lr: 9.114e-09, eta: 0:07:51, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0124, layer_-1_loss_cls: 0.0490, layer_-1_loss_bbox: 2.9647, matched_ious: 0.7694, loss: 3.0261, grad_norm: 87.9185
2022-10-18 05:37:53,381 - mmdet - INFO - Epoch [40][3400/5139]  lr: 8.656e-09, eta: 0:07:38, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0107, layer_-1_loss_cls: 0.0639, layer_-1_loss_bbox: 3.8878, matched_ious: 0.7152, loss: 3.9624, grad_norm: 115.1538
2022-10-18 05:38:06,703 - mmdet - INFO - Epoch [40][3450/5139]  lr: 8.211e-09, eta: 0:07:25, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0145, layer_-1_loss_cls: 0.0508, layer_-1_loss_bbox: 2.6441, matched_ious: 0.7274, loss: 2.7094, grad_norm: 106.8959
2022-10-18 05:38:19,983 - mmdet - INFO - Epoch [40][3500/5139]  lr: 7.779e-09, eta: 0:07:12, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0137, layer_-1_loss_cls: 0.0594, layer_-1_loss_bbox: 3.8454, matched_ious: 0.7269, loss: 3.9186, grad_norm: 104.6642
2022-10-18 05:38:33,249 - mmdet - INFO - Epoch [40][3550/5139]  lr: 7.360e-09, eta: 0:06:59, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0107, layer_-1_loss_cls: 0.0494, layer_-1_loss_bbox: 2.7452, matched_ious: 0.7404, loss: 2.8052, grad_norm: 115.3463
2022-10-18 05:38:46,580 - mmdet - INFO - Epoch [40][3600/5139]  lr: 6.954e-09, eta: 0:06:45, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0102, layer_-1_loss_cls: 0.0424, layer_-1_loss_bbox: 2.5185, matched_ious: 0.7944, loss: 2.5711, grad_norm: 79.7143
2022-10-18 05:38:59,912 - mmdet - INFO - Epoch [40][3650/5139]  lr: 6.561e-09, eta: 0:06:32, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0160, layer_-1_loss_cls: 0.0597, layer_-1_loss_bbox: 4.0300, matched_ious: 0.7243, loss: 4.1057, grad_norm: 87.6999
2022-10-18 05:39:13,231 - mmdet - INFO - Epoch [40][3700/5139]  lr: 6.181e-09, eta: 0:06:19, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0096, layer_-1_loss_cls: 0.0405, layer_-1_loss_bbox: 2.3228, matched_ious: 0.7616, loss: 2.3730, grad_norm: 87.0931
2022-10-18 05:39:26,573 - mmdet - INFO - Epoch [40][3750/5139]  lr: 5.814e-09, eta: 0:06:06, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0090, layer_-1_loss_cls: 0.0385, layer_-1_loss_bbox: 2.2676, matched_ious: 0.7697, loss: 2.3152, grad_norm: 72.8409
2022-10-18 05:39:39,901 - mmdet - INFO - Epoch [40][3800/5139]  lr: 5.459e-09, eta: 0:05:53, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0114, layer_-1_loss_cls: 0.0449, layer_-1_loss_bbox: 1.3681, matched_ious: 0.7759, loss: 1.4245, grad_norm: 85.6208
2022-10-18 05:39:53,188 - mmdet - INFO - Epoch [40][3850/5139]  lr: 5.118e-09, eta: 0:05:39, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0183, layer_-1_loss_cls: 0.0460, layer_-1_loss_bbox: 2.4514, matched_ious: 0.7688, loss: 2.5156, grad_norm: 80.1081
2022-10-18 05:40:06,397 - mmdet - INFO - Epoch [40][3900/5139]  lr: 4.790e-09, eta: 0:05:26, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0106, layer_-1_loss_cls: 0.0459, layer_-1_loss_bbox: 2.9172, matched_ious: 0.7399, loss: 2.9737, grad_norm: 127.2272
2022-10-18 05:40:19,672 - mmdet - INFO - Epoch [40][3950/5139]  lr: 4.475e-09, eta: 0:05:13, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0128, layer_-1_loss_cls: 0.0562, layer_-1_loss_bbox: 3.9414, matched_ious: 0.7126, loss: 4.0103, grad_norm: 127.4876
2022-10-18 05:40:32,957 - mmdet - INFO - Epoch [40][4000/5139]  lr: 4.172e-09, eta: 0:05:00, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0096, layer_-1_loss_cls: 0.0305, layer_-1_loss_bbox: 1.1200, matched_ious: 0.8049, loss: 1.1601, grad_norm: 82.8884
2022-10-18 05:40:46,235 - mmdet - INFO - Epoch [40][4050/5139]  lr: 3.883e-09, eta: 0:04:47, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0098, layer_-1_loss_cls: 0.0538, layer_-1_loss_bbox: 3.8977, matched_ious: 0.7653, loss: 3.9614, grad_norm: 107.2100
2022-10-18 05:40:59,507 - mmdet - INFO - Epoch [40][4100/5139]  lr: 3.607e-09, eta: 0:04:34, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0157, layer_-1_loss_cls: 0.0518, layer_-1_loss_bbox: 2.7920, matched_ious: 0.7294, loss: 2.8594, grad_norm: 89.3084
2022-10-18 05:41:12,800 - mmdet - INFO - Epoch [40][4150/5139]  lr: 3.343e-09, eta: 0:04:20, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0121, layer_-1_loss_cls: 0.0579, layer_-1_loss_bbox: 3.3466, matched_ious: 0.7544, loss: 3.4166, grad_norm: 110.5538
2022-10-18 05:41:26,107 - mmdet - INFO - Epoch [40][4200/5139]  lr: 3.093e-09, eta: 0:04:07, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0122, layer_-1_loss_cls: 0.0446, layer_-1_loss_bbox: 3.5052, matched_ious: 0.7195, loss: 3.5620, grad_norm: 103.7233
2022-10-18 05:41:39,371 - mmdet - INFO - Epoch [40][4250/5139]  lr: 2.856e-09, eta: 0:03:54, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0107, layer_-1_loss_cls: 0.0448, layer_-1_loss_bbox: 2.0877, matched_ious: 0.7404, loss: 2.1431, grad_norm: 95.2957
2022-10-18 05:41:52,610 - mmdet - INFO - Epoch [40][4300/5139]  lr: 2.631e-09, eta: 0:03:41, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0109, layer_-1_loss_cls: 0.0439, layer_-1_loss_bbox: 2.2331, matched_ious: 0.7637, loss: 2.2879, grad_norm: 102.7286
2022-10-18 05:42:06,009 - mmdet - INFO - Epoch [40][4350/5139]  lr: 2.420e-09, eta: 0:03:28, time: 0.268, data_time: 0.001, memory: 1983, loss_heatmap: 0.0113, layer_-1_loss_cls: 0.0436, layer_-1_loss_bbox: 2.1184, matched_ious: 0.7785, loss: 2.1733, grad_norm: 106.0413
2022-10-18 05:42:19,315 - mmdet - INFO - Epoch [40][4400/5139]  lr: 2.221e-09, eta: 0:03:14, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0135, layer_-1_loss_cls: 0.0425, layer_-1_loss_bbox: 3.1755, matched_ious: 0.7539, loss: 3.2316, grad_norm: 103.6085
2022-10-18 05:42:32,566 - mmdet - INFO - Epoch [40][4450/5139]  lr: 2.036e-09, eta: 0:03:01, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0089, layer_-1_loss_cls: 0.0486, layer_-1_loss_bbox: 2.6517, matched_ious: 0.7710, loss: 2.7092, grad_norm: 101.0593
2022-10-18 05:42:45,786 - mmdet - INFO - Epoch [40][4500/5139]  lr: 1.863e-09, eta: 0:02:48, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0121, layer_-1_loss_cls: 0.0375, layer_-1_loss_bbox: 1.8803, matched_ious: 0.7799, loss: 1.9299, grad_norm: 119.2097
2022-10-18 05:42:59,098 - mmdet - INFO - Epoch [40][4550/5139]  lr: 1.703e-09, eta: 0:02:35, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0123, layer_-1_loss_cls: 0.0438, layer_-1_loss_bbox: 2.4544, matched_ious: 0.7519, loss: 2.5105, grad_norm: 94.3239
2022-10-18 05:43:12,449 - mmdet - INFO - Epoch [40][4600/5139]  lr: 1.557e-09, eta: 0:02:22, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0133, layer_-1_loss_cls: 0.0364, layer_-1_loss_bbox: 1.9549, matched_ious: 0.7663, loss: 2.0046, grad_norm: 87.7325
2022-10-18 05:43:25,714 - mmdet - INFO - Epoch [40][4650/5139]  lr: 1.423e-09, eta: 0:02:08, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0233, layer_-1_loss_cls: 0.0552, layer_-1_loss_bbox: 2.7923, matched_ious: 0.7290, loss: 2.8708, grad_norm: 121.1389
2022-10-18 05:43:39,007 - mmdet - INFO - Epoch [40][4700/5139]  lr: 1.302e-09, eta: 0:01:55, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0113, layer_-1_loss_cls: 0.0583, layer_-1_loss_bbox: 2.3911, matched_ious: 0.7747, loss: 2.4607, grad_norm: 95.2007
2022-10-18 05:43:52,300 - mmdet - INFO - Epoch [40][4750/5139]  lr: 1.195e-09, eta: 0:01:42, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0140, layer_-1_loss_cls: 0.0639, layer_-1_loss_bbox: 3.2993, matched_ious: 0.7234, loss: 3.3772, grad_norm: 105.0360
2022-10-18 05:44:05,629 - mmdet - INFO - Epoch [40][4800/5139]  lr: 1.100e-09, eta: 0:01:29, time: 0.267, data_time: 0.001, memory: 1983, loss_heatmap: 0.0111, layer_-1_loss_cls: 0.0606, layer_-1_loss_bbox: 2.7163, matched_ious: 0.7315, loss: 2.7880, grad_norm: 110.2253
2022-10-18 05:44:18,925 - mmdet - INFO - Epoch [40][4850/5139]  lr: 1.018e-09, eta: 0:01:16, time: 0.266, data_time: 0.001, memory: 1983, loss_heatmap: 0.0137, layer_-1_loss_cls: 0.0691, layer_-1_loss_bbox: 3.8306, matched_ious: 0.7641, loss: 3.9135, grad_norm: 121.2616
2022-10-18 05:44:32,196 - mmdet - INFO - Epoch [40][4900/5139]  lr: 9.495e-10, eta: 0:01:03, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0094, layer_-1_loss_cls: 0.0607, layer_-1_loss_bbox: 2.6819, matched_ious: 0.7375, loss: 2.7520, grad_norm: 122.6431
2022-10-18 05:44:45,421 - mmdet - INFO - Epoch [40][4950/5139]  lr: 8.937e-10, eta: 0:00:49, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0124, layer_-1_loss_cls: 0.0449, layer_-1_loss_bbox: 3.4452, matched_ious: 0.7550, loss: 3.5025, grad_norm: 83.5678
2022-10-18 05:44:58,617 - mmdet - INFO - Epoch [40][5000/5139]  lr: 8.509e-10, eta: 0:00:36, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0112, layer_-1_loss_cls: 0.0396, layer_-1_loss_bbox: 1.7579, matched_ious: 0.7795, loss: 1.8088, grad_norm: 101.3882
2022-10-18 05:45:11,817 - mmdet - INFO - Epoch [40][5050/5139]  lr: 8.210e-10, eta: 0:00:23, time: 0.264, data_time: 0.001, memory: 1983, loss_heatmap: 0.0085, layer_-1_loss_cls: 0.0367, layer_-1_loss_bbox: 1.5538, matched_ious: 0.7956, loss: 1.5989, grad_norm: 118.1544
2022-10-18 05:45:25,052 - mmdet - INFO - Epoch [40][5100/5139]  lr: 8.042e-10, eta: 0:00:10, time: 0.265, data_time: 0.001, memory: 1983, loss_heatmap: 0.0106, layer_-1_loss_cls: 0.0481, layer_-1_loss_bbox: 2.8804, matched_ious: 0.7441, loss: 2.9391, grad_norm: 104.4471
2022-10-18 05:45:35,445 - mmdet - INFO - Saving checkpoint at 40 epochs
2022-10-18 05:47:31,135 - mmdet - INFO - Results of pts_bbox:
Car AP@0.70, 0.70, 0.70:
bbox AP:0.0000, 0.0000, 0.0000
bev  AP:0.0000, 0.0000, 9.0909
3d   AP:0.0000, 0.0000, 9.0909
aos  AP:0.00, 0.00, 0.00
Car AP@0.70, 0.50, 0.50:
bbox AP:0.0000, 0.0000, 0.0000
bev  AP:0.0000, 0.0000, 9.0909
3d   AP:0.0000, 0.0000, 9.0909
aos  AP:0.00, 0.00, 0.00
Pedestrian AP@0.50, 0.50, 0.50:
bbox AP:0.0000, 0.0000, 0.0000
bev  AP:0.0000, 0.0000, 0.0000
3d   AP:0.0000, 0.0000, 0.0000
aos  AP:0.00, 0.00, 0.00
Pedestrian AP@0.50, 0.25, 0.25:
bbox AP:0.0000, 0.0000, 0.0000
bev  AP:0.0000, 0.0000, 0.0000
3d   AP:0.0000, 0.0000, 0.0000
aos  AP:0.00, 0.00, 0.00
Cyclist AP@0.50, 0.50, 0.50:
bbox AP:0.0000, 0.0000, 0.0000
bev  AP:0.0000, 0.0000, 0.0000
3d   AP:0.0000, 0.0000, 0.0000
aos  AP:0.00, 0.00, 0.00
Cyclist AP@0.50, 0.25, 0.25:
bbox AP:0.0000, 0.0000, 0.0000
bev  AP:0.0000, 0.0000, 0.0000
3d   AP:0.0000, 0.0000, 0.0000
aos  AP:0.00, 0.00, 0.00

Overall AP@easy, moderate, hard:
bbox AP:0.0000, 0.0000, 0.0000
bev  AP:0.0000, 0.0000, 3.0303
3d   AP:0.0000, 0.0000, 3.0303
aos  AP:0.00, 0.00, 0.00

Thank you so much in advance! Best Leon