Module framework 

Meta-learning framework and its subsystems.

MetaLearningFramework wires a concrete meta-learner to the task sampler, the held-out validator, the adaptation engine and the transfer / continual / multi-task / few-shot subsystems. Every subsystem here computes its reported numbers from data it actually saw; see the individual doc comments for what each figure means.

Structs

AdaptationEngine

Adaptation engine for meta-learning

ContinualLearningSystem

Continual learning system backed by crate::continual_learning::ElasticWeightConsolidation.

FewShotLearner

Few-shot learner over the shared linear task model.

MetaLearningFramework

Meta-Learning Framework for Learned Optimizers

MetaOptimizationTracker

Meta-optimization tracker.

MetaValidator

Meta-validation system for meta-learning.

MultiTaskCoordinator

Multi-task coordinator.

TaskDistributionManager

Task distribution manager.

TransferLearningManager

Transfer learning manager backed by crate::cross_domain_transfer::CrossDomainTransfer.