ContinualLearnerBase<T, TInput, TOutput>

Base Classes · AiDotNet.ContinualLearning.Trainers

Base class for continual learning trainers that provides common functionality.

Implements: IContinualLearner<T, TInput, TOutput>

Known derived types: EWCTrainer<T, TInput, TOutput>, GEMTrainer<T, TInput, TOutput>, LwFTrainer<T, TInput, TOutput>, MASTrainer<T, TInput, TOutput>, SITrainer<T, TInput, TOutput>

How It Works

For Algorithm Implementers: To create a new continual learning algorithm:

Common Patterns:

Reference: De Lange et al. “A Continual Learning Survey: Defying Forgetting” (2021)

Properties

Property Summary
IContinualLearnerConfig<T> Config { get; }
IFullModel<T, TInput, TOutput> BaseModel { get; }
bool IsTraining { get; }
int TasksLearned { get; }
long MemoryUsageBytes { get; }

Methods

Method Summary
ContinualEvaluationResult<T> EvaluateAllTasks()
ContinualLearningResult<T> GetTaskHistory(int taskId)
ContinualLearningResult<T> LearnTask(IDataset<T, TInput, TOutput> taskData)
ContinualLearningResult<T> LearnTask(IDataset<T, TInput, TOutput> taskData, IDataset<T, TInput, TOutput> validationData, int? earlyStoppingPatience)
IReadOnlyDictionary<int, T> ComputeForgetting()
IReadOnlyList<ContinualLearningResult<T>> GetAllHistory()
TaskEvaluationResult<T> EvaluateTask(int taskId, IDataset<T, TInput, TOutput> testData)
void Load(string directoryPath)
void Reset()
void Save(string directoryPath)