Class L3Regularizer

Namespace

NeuralNetworks.Regularization

Assembly

NeuralNetworks.dll

Applies L3 regularization by adding a penalty proportional to the cubed magnitude of model parameters.

public class L3Regularizer : BaseRegularizer, IRegularizer

Inheritance

object

BaseRegularizer

L3Regularizer

Implements

IRegularizer

Inherited Members

BaseRegularizer.ReportPenaltyInLoss

BaseRegularizer.ApplyGradient<T>(T, T)

BaseRegularizer.ReportPenalty<T>(T)

BaseRegularizer.UpdateParams<T>(T)

BaseRegularizer.UpdateParamsCore(Span<float>)

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

Remarks

Uses the L3 penalty form l3 * Σ(|w|^3), contributes 3 * l3 * w * |w| to each parameter gradient, and can optionally include the penalty term in the reported loss via the base regularizer configuration.

Constructors

L3Regularizer(float, bool)

public L3Regularizer(float l3, bool reportPenaltyInLoss = false)

Parameters

l3 float

reportPenaltyInLoss bool

Methods

ApplyGradientCore(ReadOnlySpan<float>, Span<float>)

Applies a gradient update to the specified parameter span using the corresponding gradient span.

protected override void ApplyGradientCore(ReadOnlySpan<float> paramSpan, Span<float> paramGradientSpan)

Parameters

paramSpan ReadOnlySpan<float>

The span of parameter values to update.

paramGradientSpan Span<float>

The span of gradient values corresponding to paramSpan.

CalculatePenaltyCore(ReadOnlySpan<float>)

protected override float CalculatePenaltyCore(ReadOnlySpan<float> paramSpan)

Parameters

paramSpan ReadOnlySpan<float>

Returns

float

ToString()

Returns a string that represents the current object.

public override string ToString()

Returns

string

A string that represents the current object.