nvidia.dali.experimental.dynamic.shapes#
- nvidia.dali.experimental.dynamic.shapes(input, /, *, batch_size=None, device=None, dtype=None, type=None)#
-
Returns the shapes of tensors in the input batch.
This operator allows sequence inputs and supports volumetric data.
- Supported backends
‘cpu’
‘gpu’
- Parameters:
input¶ (Tensor/Batch) – Input to the operator.
- Keyword Arguments:
dtype¶ (
nvidia.dali.types.DALIDataType, optional, default = DALIDataType.INT64) – Data type to which the sizes are converted.type¶ (
nvidia.dali.types.DALIDataType) –