TopDownPoseTrack18VideoDataset

[Tommy-Hsu]

mmcv 1.7.0
mmcv-full 1.6.2
mmengine 0.10.4
mmpose 0.29.0
mmpretrain 1.2.0

---

This is my data config,

data_cfg = dict(
    image_size=[288, 384],
    heatmap_size=[72, 96],
    num_output_channels=channel_cfg['num_output_channels'],
    num_joints=channel_cfg['dataset_joints'],
    dataset_channel=channel_cfg['dataset_channel'],
    inference_channel=channel_cfg['inference_channel'],
    soft_nms=False,
    nms_thr=1.0,
    oks_thr=0.9,
    vis_thr=0.2,
    use_gt_bbox=True,
    det_bbox_thr=0.5,
    bbox_file='data/PoseTrack2018/annotations/'
    'posetrack18_val_thesis_bbox.json',

    # For TopDownPoseTrack18VideoDataset
    frame_weight_train=1.0,
    frame_weight_test=1.0,
    frame_indices_train=[-2, -1, 1, 2],
    frame_index_rand=False,
    frame_indices_test=[-2, -1, 0, 1, 2],
)

But when I dig into the python debugger to pipelines LoadImageFromFile
get the weird items

when i put breakpoint in datasets TopDownPoseTrack18VideoDataset, it's normal.
But not in pipelines LoadImageFromFile