ISupportSdcaRegressionLoss Interface

  • Reference

Definition

Namespace:
Microsoft.ML.Trainers
Assembly:
Microsoft.ML.Data.dll
Package:
Microsoft.ML v3.0.1
Package:
Microsoft.ML v1.0.0
Package:
Microsoft.ML v1.1.0
Package:
Microsoft.ML v1.2.0
Package:
Microsoft.ML v1.3.1
Package:
Microsoft.ML v1.4.0
Package:
Microsoft.ML v1.5.5
Package:
Microsoft.ML v1.6.0
Package:
Microsoft.ML v1.7.0
Package:
Microsoft.ML v2.0.0
public interface ISupportSdcaRegressionLoss : Microsoft.ML.Trainers.ILossFunction<float,float>, Microsoft.ML.Trainers.IRegressionLoss, Microsoft.ML.Trainers.ISupportSdcaLoss
type ISupportSdcaRegressionLoss = interface
    interface ISupportSdcaLoss
    interface IScalarLoss
    interface ILossFunction<single, single>
    interface IRegressionLoss
Public Interface ISupportSdcaRegressionLoss
Implements ILossFunction(Of Single, Single), IRegressionLoss, ISupportSdcaLoss
Derived
Implements
ILossFunction<Single,Single> IRegressionLoss IScalarLoss ISupportSdcaLoss

Methods

ComputeDualUpdateInvariant(Single) (Inherited from ISupportSdcaLoss)
Derivative(Single, Single)

Derivative of the loss function with respect to output

(Inherited from IScalarLoss)
DualLoss(Single, Single)

The dual loss function for a training example. If f(x) denotes the loss function on an individual training example, then this function returns -f*(-x*), where f*(x*) is the Fenchel conjugate of f(x).

(Inherited from ISupportSdcaLoss)
DualUpdate(Single, Single, Single, Single, Int32)

Compute the dual update (\Delta\alpha_i) in SDCA

  • alpha: dual variable at the specified instance
  • lambdaN: L2 const x number of instances
  • cached invariant, hinted by the method above
(Inherited from ISupportSdcaLoss)
Loss(TOutput, TLabel)

Computes the loss given the output and the ground truth. Note that the return value has type Double because the loss is usually accumulated over many instances.

(Inherited from ILossFunction<TOutput,TLabel>)

Applies to