EGLOptions<T>
Options & Configuration · AiDotNet.ContinualLearning.Strategies
Configuration options for Expected Gradient Length strategy.
Constructors
| Constructor | Summary |
|---|---|
EGLOptions() |
Properties
| Property | Summary |
|---|---|
T Lambda { get; set; } |
Gets or sets the regularization strength (lambda). |
bool? NormalizeImportance { get; set; } |
Gets or sets whether to normalize importance values. |
bool? UseSquaredLength { get; set; } |
Gets or sets whether to use squared gradient lengths. |
double? DecayFactor { get; set; } |
Gets or sets the decay factor for online accumulation. |
double? MinImportanceValue { get; set; } |
Gets or sets the minimum importance value (to prevent division by zero). |
int? NumSamples { get; set; } |
Gets or sets the number of samples for gradient length estimation. |