BaseAveragedStochasticGradientDescentTModel, TKernel, TInput, TLossTolerance Property

Gets or sets the maximum relative change in the watched value after an iteration of the algorithm used to detect convergence. Default is 1e-3. If set to 0, the loss will not be computed during learning and execution will be faster.

Namespace: Accord.MachineLearning.VectorMachines.Learning
Assembly: Accord.MachineLearning (in Accord.MachineLearning.dll) Version: 3.8.0

Syntax

Copy

public double Tolerance { get; set; }

Public Property Tolerance As Double
	Get
	Set

Request Example View Source

Property Value

Type: Double

Reference

BaseAveragedStochasticGradientDescentTModel, TKernel, TInput, TLoss Class

Accord.MachineLearning.VectorMachines.Learning Namespace