The bad performance when used VelocityAug in bevfusion

[liyih]

hi junjie, In you latest work DAL, you mention a method VelocityAug to balance the velocity distribution. However, when I insert this method in bevfusion the NDS decrease, which seems different with your conclusion in Table 5. Here is my train_pipline config:

train_pipeline = [ dict( type='BEVLoadMultiViewImageFromFiles', to_float32=True, color_type='color', backend_args=backend_args), dict( type='LoadPointsFromFile', coord_type='LIDAR', load_dim=5, use_dim=5, backend_args=backend_args), dict( type='LoadPointsFromMultiSweeps', sweeps_num=9, load_dim=5, use_dim=5, pad_empty_sweeps=True, remove_close=True, backend_args=backend_args), dict( type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True, with_attr_label=False), dict(type='VelocityAug'), dict( type='ImageAug3D', final_dim=[256, 704], resize_lim=[0.38, 0.55], bot_pct_lim=[0.0, 0.0], rot_lim=[-5.4, 5.4], rand_flip=True, is_train=True), dict( type='BEVFusionGlobalRotScaleTrans', scale_ratio_range=[0.9, 1.1], rot_range=[-0.78539816, 0.78539816], translation_std=0.5), dict(type='BEVFusionRandomFlip3D'), dict(type='PointsRangeFilter', point_cloud_range=point_cloud_range), dict(type='ObjectRangeFilter', point_cloud_range=point_cloud_range), dict( type='ObjectNameFilter', classes=[ 'car', 'truck', 'construction_vehicle', 'bus', 'trailer', 'barrier', 'motorcycle', 'bicycle', 'pedestrian', 'traffic_cone' ]),

Actually, 'GridMask' is not used here

dict(
    type='GridMask',
    use_h=True,
    use_w=True,
    max_epoch=max_epoch,
    rotate=1,
    offset=False,
    ratio=0.5,
    mode=1,
    prob=0.0,
    fixed_prob=True),
dict(type='PointShuffle'),
dict(
    type='Pack3DDetInputs',
    keys=[
        'points', 'img', 'gt_bboxes_3d', 'gt_labels_3d', 'gt_bboxes',
        'gt_labels'
    ],
    meta_keys=[
        'cam2img', 'ori_cam2img', 'lidar2cam', 'lidar2img', 'cam2lidar',
        'ori_lidar2img', 'img_aug_matrix', 'box_type_3d', 'sample_idx',
        'lidar_path', 'img_path', 'transformation_3d_flow', 'pcd_rotation',
        'pcd_scale_factor', 'pcd_trans', 'img_aug_matrix',
        'lidar_aug_matrix', 'num_pts_feats'
    ])

]

[HuangJunJie2017]

....emmm, I haven't used the origin repo of bevfusion for any experiments. instead, I use a private implementation on BEVDet repo. If you insist on using the origin repo of bevfusion. Be careful in checking out every line of code to make sure they are reasonable .