共用方式為


TensorFeatureDescriptor 類別

定義

張量是值的多維度陣列。

public ref class TensorFeatureDescriptor sealed : ILearningModelFeatureDescriptor
/// [Windows.Foundation.Metadata.ContractVersion(Windows.AI.MachineLearning.MachineLearningContract, 65536)]
/// [Windows.Foundation.Metadata.MarshalingBehavior(Windows.Foundation.Metadata.MarshalingType.Agile)]
class TensorFeatureDescriptor final : ILearningModelFeatureDescriptor
[Windows.Foundation.Metadata.ContractVersion(typeof(Windows.AI.MachineLearning.MachineLearningContract), 65536)]
[Windows.Foundation.Metadata.MarshalingBehavior(Windows.Foundation.Metadata.MarshalingType.Agile)]
public sealed class TensorFeatureDescriptor : ILearningModelFeatureDescriptor
Public NotInheritable Class TensorFeatureDescriptor
Implements ILearningModelFeatureDescriptor
繼承
Object Platform::Object IInspectable TensorFeatureDescriptor
屬性
實作

Windows 需求

裝置系列
Windows 10, version 1809 (已於 10.0.17763.0 引進)
API contract
Windows.AI.MachineLearning.MachineLearningContract (已於 v1.0 引進)

備註

張量是值的多維度陣列。

張量的配置是資料列主要,且緊密封裝的連續資料代表每個維度。 張量的大小總計是每個維度大小的乘積。

Windows Server

若要在 Windows Server 上使用此 API,您必須搭配桌面體驗使用 Windows Server 2019。

執行緒安全

此 API 是安全線程。

屬性

Description

在模型中使用此功能的描述。

IsRequired

如果為 true,您必須先將值系結至此功能,才能呼叫 LearningModelSession.Evaluate

Kind

功能種類;使用這個 來知道要使用的衍生類別。

Name

您用來將值系結至此功能的名稱。

Shape

傳回每個維度的計數和大小。

TensorKind

傳回張量的資料型別。

適用於

另請參閱