Contents

Unpacks packed tensors into their constituents and removes padding.

This function acts as the inverse of the pack operation. It splits a packed tensor along a specified axis and removes any padding that was added during packing.

Parameters:

  • tensor (Tensor) – The tensor to be unpacked.

  • value (int | float | bool, optional) – The identity of the padding value, by default 0.

  • axis (int, optional) – The axis along which the tensor was packed, by default 0.

returns:

A tuple of the constituent tensors after unpacking and deflating.

rtype:

tuple[Tensor, …]

Examples

Suppose you have a tensor that has been packed along the first axis (axis=0):

>>> packed_tensor = torch.tensor([
>>>     [1, 2, 3, 0, 0],
>>>     [4, 5, 0, 0, 0],
>>>     [6, 7, 8, 9, 0]
>>> ])

Unpacking this tensor would yield:

>>> unpacked_tensors = unpack(packed_tensor, value=0, axis=0)
>>> for tensor in unpacked_tensors:
>>>     print(tensor)
tensor([1, 2, 3])
tensor([4, 5])
tensor([6, 7, 8, 9])
Contents