LwFTrainerOptions<T>
Options & Configuration · AiDotNet.ContinualLearning.Trainers
Configuration options for the LwF trainer.
For Beginners
These options control how the LwF trainer operates. LwF uses knowledge distillation to preserve knowledge from previous tasks.
Constructors
| Constructor | Summary |
|---|---|
LwFTrainerOptions() |
Properties
| Property | Summary |
|---|---|
bool? ComputeValidationMetrics { get; set; } |
Whether to compute validation metrics after each epoch. |
bool? IncludeDistillationInGradients { get; set; } |
Whether to include distillation loss in gradient updates (true) or only use it for monitoring (false). |
bool? UseExperienceReplay { get; set; } |
Whether to use experience replay alongside distillation. |
double? DistillationWeight { get; set; } |
Weight for the distillation loss relative to task loss. |
double? GradientClipThreshold { get; set; } |
Gradient clipping threshold. |
double? ReplayLearningRateFactor { get; set; } |
Fraction of learning rate to use for replay samples. |
double? Temperature { get; set; } |
Temperature for softmax in distillation. |
int? DistillationWarmupEpochs { get; set; } |
Warmup epochs before applying full distillation weight. |