IMetaLearnerOptions<T>

Interfaces · AiDotNet.Interfaces

Configuration options interface for meta-learning algorithms.

Known implementations: ACLOptions<T, TInput, TOutput>, ANILOptions<T, TInput, TOutput>, ANPOptions<T, TInput, TOutput>, ATAMLOptions<T, TInput, TOutput>, AWGIMOptions<T, TInput, TOutput>, ActiveTransFSLOptions<T, TInput, TOutput>, AutoLoRAOptions<T, TInput, TOutput>, BMAMLOptions<T, TInput, TOutput>, BOILOptions<T, TInput, TOutput>, BayProNetOptions<T, TInput, TOutput>, BayTransProtoOptions<T, TInput, TOutput>, CAMLOptions<T, TInput, TOutput>, CAVIAOptions<T, TInput, TOutput>, CNAPOptions<T, TInput, TOutput>, CNPOptions<T, TInput, TOutput>, ConstellationNetOptions<T, TInput, TOutput>, ContextMetaRLOptions<T, TInput, TOutput>, ConvCNPOptions<T, TInput, TOutput>, ConvNPOptions<T, TInput, TOutput>, DKTOptions<T, TInput, TOutput>, DPGNOptions<T, TInput, TOutput>, DREAMOptions<T, TInput, TOutput>, DeepEMDOptions<T, TInput, TOutput>, DiscoRLOptions<T, TInput, TOutput>, DynamicTaskSamplingOptions<T, TInput, TOutput>, EPNetOptions<T, TInput, TOutput>, ETPNOptions<T, TInput, TOutput>, EquivCNPOptions<T, TInput, TOutput>, FEATOptions<T, TInput, TOutput>, FRNOptions<T, TInput, TOutput>, FewTUREOptions<T, TInput, TOutput>, FreqPriorOptions<T, TInput, TOutput>, FreqPromptOptions<T, TInput, TOutput>, GNNMetaOptions<T, TInput, TOutput>, HyperCLIPOptions<T, TInput, TOutput>, HyperMAMLOptions<T, TInput, TOutput>, HyperNeRFMetaOptions<T, TInput, TOutput>, HyperNetMetaRLOptions<T, TInput, TOutput>, HyperShotOptions<T, TInput, TOutput>, ICMFusionOptions<T, TInput, TOutput>, InContextRLOptions<T, TInput, TOutput>, JMPOptions<T, TInput, TOutput>, LBANPOptions<T, TInput, TOutput>, LEOOptions<T, TInput, TOutput>, LFTOptions<T, TInput, TOutput>, LaplacianShotOptions<T, TInput, TOutput>, LoRARecycleOptions<T, TInput, TOutput>, MAMLOptions<T, TInput, TOutput>, MAMLPlusPlusOptions<T, TInput, TOutput>, MANNOptions<T, TInput, TOutput>, MCLOptions<T, TInput, TOutput>, MOCAOptions<T, TInput, TOutput>, MPTSOptions<T, TInput, TOutput>, MatchingNetworksOptions<T, TInput, TOutput>, MbPAOptions<T, TInput, TOutput>, MetaBaselineOptions<T, TInput, TOutput>, MetaCollaborativeOptions<T, TInput, TOutput>, MetaDDPMOptions<T, TInput, TOutput>, MetaDMOptions<T, TInput, TOutput>, MetaDiffOptions<T, TInput, TOutput> (+44 more)

For Beginners

Think of meta-learning like learning to study effectively:

  • Inner loop: How you study for a specific exam (practice problems, examples)
  • Outer loop: Learning better study techniques by reflecting across many exams

The configuration controls both loops:

  • InnerLearningRate: How aggressively to adapt to each task
  • OuterLearningRate: How much to update the meta-parameters
  • AdaptationSteps: How many gradient updates per task

How It Works

Meta-learning algorithms use a two-loop optimization structure:

  • Inner loop: Fast adaptation to a specific task using support set
  • Outer loop: Meta-optimization to improve adaptation across all tasks

Properties

Property Summary
bool EnableCheckpointing { get; } Gets whether to save checkpoints during training.
bool UseFirstOrder { get; } Gets whether to use first-order approximation (e.g., FOMAML, Reptile).
double InnerLearningRate { get; } Gets the inner loop learning rate for task-specific adaptation.
double OuterLearningRate { get; } Gets the outer loop learning rate for meta-optimization.
double? GradientClipThreshold { get; } Gets the gradient clipping threshold to prevent exploding gradients.
int AdaptationSteps { get; } Gets the number of gradient descent steps for inner loop adaptation.
int CheckpointFrequency { get; } Gets the checkpoint save frequency in meta-iterations.
int EvaluationFrequency { get; } Gets the evaluation frequency in meta-iterations.
int EvaluationTasks { get; } Gets the number of evaluation tasks for periodic validation.
int MetaBatchSize { get; } Gets the number of tasks to sample per meta-update (meta-batch size).
int NumMetaIterations { get; } Gets the number of meta-training iterations to perform.
int? RandomSeed { get; } Gets the random seed for reproducible task sampling and initialization.

Methods

Method Summary
IMetaLearnerOptions<T> Clone() Creates a copy of this options instance.
bool IsValid() Validates that the configuration is valid and sensible.