TANestedTensors¶

class torchaug.ta_tensors.nested.TANestedTensors(tensors)[source]¶

TANestedTensors is a generic class to handle nested tensors.

A nested tensor is a list of tensors with the same device, dtype, and number of dimensions.

Parameters:: tensors (Sequence[TypeVar(T, Image, Video, Labels, BoundingBoxes, Mask, Tensor)]) – The sequence of tensors to nest.
Returns:: _description_

add(other)[source]¶

Add a tensor or a scalar to the nested tensors.

Return type:: Self

add_(other)[source]¶

Add a tensor or a scalar to the nested tensors in-place.

Return type:: Self

property batch_size: int¶: Return the batch size of the nested tensors.

clone(memory_format=None)[source]¶

Return a deep copy of the nested tensors.

Return type:: Self

count(tensor)[source]¶

Return the number of occurrences of a tensor in the nested tensors.

Return type:: int

property device: device¶: Return the device of the nested tensors.

div(other)[source]¶

Divide the nested tensors by a tensor or a scalar.

Return type:: Self

div_(other)[source]¶

Divide the nested tensors by a tensor or a scalar in-place.

Return type:: Self

property dtype: dtype¶: Return the dtype of the nested tensors.

floor_divide(other)[source]¶

Floor divide the nested tensors by a tensor or a scalar.

Return type:: Self

floor_divide_(other)[source]¶

Floor divide the nested tensors by a tensor or a scalar in-place.

Return type:: Self

index(tensor)[source]¶

Return the index of a tensor in the nested tensors.

Return type:: int

mul(other)[source]¶

Multiply a tensor or a scalar to the nested tensors.

Return type:: Self

mul_(other)[source]¶

Multiply a tensor or a scalar to the nested tensors in-place.

Return type:: Self

property ndim: int¶: Return the number of dimensions of the nested tensors.

pin_memory(device=None)[source]¶

Pin the memory of the nested tensors.

Return type:: Self

remainder(other)[source]¶

Return the remainder of the nested tensors by a tensor or a scalar.

Return type:: Self

remainder_(other)[source]¶

Return the remainder of the nested tensors by a tensor or a scalar in-place.

Return type:: Self

property shapes: List[Size]¶: Return the shapes of the tensors.

size(dim=None)[source]¶

Return the size of the nested tensors.

Return type:: List[Size]

sub(other)[source]¶

Subtract a tensor or a scalar to the nested tensors.

Return type:: Self

sub_(other)[source]¶

Subtract a tensor or a scalar to the nested tensors in-place.

Return type:: Self

to(device=None, dtype=None, non_blocking=False, copy=False, *, memory_format=None)[source]¶

Move the nested tensors to a new device and/or dtype.

Return type:: Self

to_batch()[source]¶

Return the batched tensor and the shapes of the nested tensors.

Return type:: TypeVar(BATCH_T, BatchImages, BatchBoundingBoxes, BatchLabels, BatchMasks, BatchVideos, Tensor)

to_list()[source]¶

Return the nested tensors as a list.

Return type:: List[TypeVar(T, Image, Video, Labels, BoundingBoxes, Mask, Tensor)]