C# Class NNX.Core.Training.SimpleGradientTrainer

A SimpleGradientTrainer is an on-line gradient descent trainer that applies backpropagation (with a few adjustments) NumEpochs times. Next input/target pair is selected randomly from the training set.

Inheritance: BaseTrainer

Show file Open project: ikhramts/NNX Class Usage Examples

Public Methods

Method	Description
AdjustWeights ( INeuralNetwork nn, double weightGradients, double prevWeightGradients ) : void
GetValidationSetFraction ( ) : double
Train ( IList trainingSet, IList validationSet, IRandomGenerator rand, INeuralNetwork nn ) : void
Validate ( ) : void

Method Details

AdjustWeights() public method

public AdjustWeights ( INeuralNetwork nn, double weightGradients, double prevWeightGradients ) : void
nn	INeuralNetwork
weightGradients	double
prevWeightGradients	double
return	void

GetValidationSetFraction() public method

public GetValidationSetFraction ( ) : double
return	double

Train() public method

public Train ( IList trainingSet, IList validationSet, IRandomGenerator rand, INeuralNetwork nn ) : void
trainingSet	IList
validationSet	IList
rand	IRandomGenerator
nn	INeuralNetwork
return	void

Validate() public method

public Validate ( ) : void
return	void