TensorMarshal.CreateTensorSpan<T> Method
Definition
Important
Some information relates to prerelease product that may be substantially modified before it’s released. Microsoft makes no warranties, express or implied, with respect to the information provided here.
Creates a new tensor span over a portion of a regular managed object.
public:
generic <typename T>
static System::Numerics::Tensors::TensorSpan<T> CreateTensorSpan(T % data, IntPtr dataLength, ReadOnlySpan<IntPtr> lengths, ReadOnlySpan<IntPtr> strides, bool pinned);public static System.Numerics.Tensors.TensorSpan<T> CreateTensorSpan<T>(scoped ref T data, IntPtr dataLength, scoped ReadOnlySpan<IntPtr> lengths, scoped ReadOnlySpan<IntPtr> strides, bool pinned);static member CreateTensorSpan : 'T * nativeint * ReadOnlySpan<nativeint> * ReadOnlySpan<nativeint> * bool -> System.Numerics.Tensors.TensorSpan<'T>Public Shared Function CreateTensorSpan(Of T) (ByRef data As T, dataLength As IntPtr, lengths As ReadOnlySpan(Of IntPtr), strides As ReadOnlySpan(Of IntPtr), pinned As Boolean) As TensorSpan(Of T)Type Parameters
- T
The type of the data items.
Parameters
- data
- T
A reference to data.
- dataLength
-
IntPtr
nativeint
The number of T elements that data contains.
- lengths
-
ReadOnlySpan<nativeint>
The lengths of the dimensions. If an empty span is provided, the created tensor will have a single dimension that is the same length as dataLength.
- strides
-
ReadOnlySpan<nativeint>
The strides of each dimension. If an empty span is provided, then strides will be automatically calculated from lengths.
- pinned
- Boolean
true if data is permanently pinned; otherwise, false.
Returns
The created tensor span.
Remarks
This method should be used with caution. It is dangerous because the inputs may not be fully checked. Even though data is marked as scoped, it will be stored into the returned tensor span, and the lifetime of the returned tensor span will not be validated for safety, even by span-aware languages.
