BalancedTaskSampler<T, TInput, TOutput>

Models & Types · AiDotNet.MetaLearning.Data

Samples tasks while ensuring that all classes in the meta-dataset appear equally often across the sampled episodes over time. This prevents the meta-learner from overfitting to frequently-sampled classes.

Implements: ITaskSampler<T, TInput, TOutput>

Constructors

Constructor Summary
BalancedTaskSampler(IMetaDataset<T, TInput, TOutput> dataset, int numWays, int numShots, int numQueryPerClass, int? seed) Creates a balanced task sampler that rotates through all classes evenly.

Properties

Property Summary
int NumQueryPerClass { get; }
int NumShots { get; }
int NumWays { get; }

Methods

Method Summary
IEpisode<T, TInput, TOutput> SampleOne()
TaskBatch<T, TInput, TOutput> SampleBatch(int batchSize)
void SetSeed(int seed)
void UpdateWithFeedback(IReadOnlyList<IEpisode<T, TInput, TOutput>> episodes, IReadOnlyList<double> losses)