Concatenates a list of Boxes into a single Boxes Args: boxes_list (list[Boxes]) Returns: Boxes: the

Send all values to device by calling `v.to(device)` (PyTorch only). Args: device (`str` or `torch.device`): The device t

Remove the node of the key, and return its value, its function is similar to the `get` of dict.

r"""Get an item from registry with key 'name' Args: name (string): Key whose value needs to be retreived. default: If

Insert a value into the history series according to the key of `update_dict`, and we will convert the :external:py:class:`torch.Tenso

Returns: A list of (key, value) pair, and the function is similar to `dict`

Copy tensor and all other fields in the way of deepcopy. Returns: an instance of SizedDataStructure.

Similar to ``.to`` function on a `torch.Tensor`. Moves all of the tensors present inside the ``SampleList`` to a particular device. If an

r"""Register an item to registry with key 'name' Args: name: Key with which the item will be registered. Usage::

Add an attribute ``field`` with value ``data`` to the SampleList Args: field (str): Key under which the data will be added.

r"""Returns the state of the optimizer as a :class:`dict`. It contains two entries: * state - a dict holding current optimization st

This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring o

Register a module which must be a class. Example1:: @ModuleRegistry.register class MLP(torch.nn.Module):

Freeze the configuration, and all nodes in any depth can not be editable.

Clip (in place) the boxes by limiting x coordinates to the range [0, width] and y coordinates to the range [0, height]. Args

Print information about the VQA annotation file. :return:

r""" Returns: use_item (bool): whether successfully get image image (Tensor): image tensor ```

Convert mode of the boxes, for example, convert the mode of boxes from xyxy to xywh, vice versa. Args: box_mode (int): o

Decompose NestedTensor to a tuple of tensors Returns: tensors, mask (torch.Tensor):

Create NestedTensor from image tensors. Args: tensor_list (torch.Tensor): a list of images :external:py:class:`Tensor <t

Main load function use by antmmf. This will internally call ``_load`` function. Calls ``init_processors`` and ``try_fast_read`` on the

return the device of the tensor.

transformer pretrained weights and vocabs dir, weights can be downloaded from https://huggingface.co/models **Note**: for online serving, pr

Returns: All keys of this configuration with depth of 1.

Run all_gather on arbitrary picklable data (not necessarily tensors) Args: data: any picklable object Returns: list[data]

Encode a list of BlockArgs to a list of strings. Args: blocks_args (list[namedtuples]): A list of BlockArgs namedtuples of block

r"""The initial call to start propagating messages. Args: adj (Tensor): A :obj:`torch.LongTensor` that defines the underlying

Defrost the configuration, make it can be editable.