CurriculumLearner<T, TInput, TOutput>

Models & Types · AiDotNet.CurriculumLearning

Main orchestrator for curriculum learning that coordinates difficulty estimation, scheduling, and model training.

Implements: ICurriculumLearner<T, TInput, TOutput>

For Beginners

Curriculum learning is a training strategy that presents samples from easy to hard, mimicking how humans learn. This class coordinates the entire process:

How It Works

Key Components:

Example Usage:

References:

Constructors

Constructor Summary
CurriculumLearner(IFullModel<T, TInput, TOutput> baseModel, ICurriculumLearnerConfig<T> config, IDifficultyEstimator<T, TInput, TOutput> difficultyEstimator, ICurriculumScheduler<T> scheduler) Initializes a new instance of the CurriculumLearner class.

Properties

Property Summary
ICurriculumLearnerConfig<T> Config { get; }
ICurriculumScheduler<T> Scheduler { get; }
IDifficultyEstimator<T, TInput, TOutput> DifficultyEstimator { get; }
IFullModel<T, TInput, TOutput> BaseModel { get; }
T CurrentPhase { get; }
bool IsTraining { get; }
int CurrentEpoch { get; }

Methods

Method Summary
CurriculumLearningResult<T> Train(IDataset<T, TInput, TOutput> trainingData)
CurriculumLearningResult<T> Train(IDataset<T, TInput, TOutput> trainingData, IDataset<T, TInput, TOutput> validationData)
CurriculumLearningResult<T> TrainWithDifficulty(IDataset<T, TInput, TOutput> trainingData, Vector<T> difficultyScores)
IReadOnlyList<CurriculumPhaseResult<T>> GetPhaseHistory()
Vector<T> EstimateDifficulties(IDataset<T, TInput, TOutput> dataset)
bool AdvancePhase()
int[] GetCurrentCurriculumIndices(Vector<T> allDifficulties)
void Load(string path)
void ResetCurriculum()
void Save(string path)