共用方式為


ISupportSdcaLoss 介面

定義

遺失函式可能知道最佳雙更新 Ref 的封閉式解決方案:秒 (6.2) http://jmlr.org/papers/volume14/shalev-shwartz13a/shalev-shwartz13a.pdf

public interface ISupportSdcaLoss : Microsoft.ML.Trainers.ILossFunction<float,float>, Microsoft.ML.Trainers.IScalarLoss
type ISupportSdcaLoss = interface
    interface IScalarLoss
    interface ILossFunction<single, single>
Public Interface ISupportSdcaLoss
Implements ILossFunction(Of Single, Single), IScalarLoss
衍生
實作

方法

ComputeDualUpdateInvariant(Single)

遺失函式可能知道最佳雙更新 Ref 的封閉式解決方案:秒 (6.2) http://jmlr.org/papers/volume14/shalev-shwartz13a/shalev-shwartz13a.pdf

Derivative(Single, Single)

與輸出相關的遺失函式衍生

(繼承來源 IScalarLoss)
DualLoss(Single, Single)

定型範例的雙重遺失函式。 如果 f (x) 代表個別定型範例上的遺失函式,則此函式會傳回 -f* (-x*) ,其中 f* (x*) 是 f (x) 的 Fenchel 串連。

DualUpdate(Single, Single, Single, Single, Int32)

在 SDCA 中計算雙重更新 (\Delta\Alpha_i)

  • Alpha:指定實例上的雙重變數
  • LambdaN:L2 const x 實例數目
  • 快取的不變異,由上述方法提示
Loss(TOutput, TLabel)

計算輸出和基礎事實的遺失。 請注意,傳回值的類型為 Double,因為遺失通常會累積在許多實例上。

(繼承來源 ILossFunction<TOutput,TLabel>)

適用於