Module meta_learning_optimizer 

MetaLearningOptimizer — production-quality meta-learning (learning to learn) optimization.

Implements MAML, Reptile, FOMAML, and ProtoNet for few-shot adaptation of a linear regression model (y = w·x + b) over many tasks.

Collision aliases (types already exported at crate root from other modules)

• TaskId → MloTaskId
• TaskExample → MloTaskExample
• MetaTask → MloMetaTask
• MetaError → MloMetaError

Structs

AdaptationStep

Snapshot of the model state after one inner-loop gradient step.

MetaLearningOptimizer

A production-quality meta-learning optimizer.

MetaStats

Accumulated statistics for the meta-learning optimizer.

MetaTask

A meta-learning task holding support and query sets.

ModelParams

Linear model parameters: y = w·x + b.

OptimizerConfig

Configuration for MetaLearningOptimizer.

TaskExample

A single labeled example belonging to a specific task.

TaskId

A unique identifier for a meta-learning task.

Enums

MetaAlgorithm

The meta-learning algorithm variant.

MetaError

Errors produced by MetaLearningOptimizer operations.