Get a set of frames from a particular sequence args: seq_id - index of sequence frame_ids - a list of frame n

Returns: The size of the per-machine process group, i.e. the number of processes per machine.

Name of the dataset returns: string - Name of the dataset

Returns information about a particular sequences, args: seq_id - index of the sequence returns: Dict

:param boxes1: (N, 4) (x1,y1,x2,y2) :param boxes2: (N, 4) (x1,y1,x2,y2) :return:

Check if there exist a parameter with the given name.

Show the wrong tracks on the input image. Args: backend (str, optional): Backend of visualization. Defaults to 'cv2'.

Initialize a detector from config file. Args: config (str or :obj:`mmcv.Config`): Config file path or the config object.

Initialize the weights in backbone. Args: pretrained (str, optional): Path to pre-trained weights. Defaults to Non

Initialize the weights in backbone. Args: pretrained (str, optional): Path to pre-trained weights. Defaults to No

Dimension that can be used by transforms to set the correct image size, etc. This allows transforms to have a single source of truth

Equivalent to nn.functional.interpolate, but with support for empty batch sizes. This will eventually be supported natively by PyTorch, and t

Returns whether the dataset contains real videos or synthetic returns: bool - True if a video dataset

For each frame in frames, extracts a square crop centered at box_extract, of area search_area_factor^2 times box_extract area. The extracted crop

Image reading using jpeg4py https://github.com/ajkxyz/jpeg4py

Function that collates lists or tuples together into one list (of lists/tuples). Use this as the collate function in a Dataloader, if you wan

Load annotation from annotation file.

Load annotation from annotation file.