OMLOptions<T, TInput, TOutput>
Options & Configuration · AiDotNet.MetaLearning.Options
Configuration options for the OML (Online Meta-Learning) algorithm.
Inheritance: ModelOptions → OMLOptions<T, TInput, TOutput>
Implements: IMetaLearnerOptions<T>
How It Works
OML (Javed & White, 2019) partitions the model into a Representation Learning Network (RLN) and a Prediction Learning Network (PLN). Only the PLN parameters are adapted in the inner loop, while the RLN is meta-learned to produce sparse, non-interfering representations that enable continual learning.
Constructors
| Constructor | Summary |
|---|---|
OMLOptions(IFullModel<T, TInput, TOutput> metaModel) |
Creates options for meta-learning metaModel, with the default settings. |
Properties
| Property | Summary |
|---|---|
IEpisodicDataLoader<T, TInput, TOutput> DataLoader { get; set; } |
Episodic data loader that supplies the meta-training tasks; null when tasks are passed to training directly. |
IFullModel<T, TInput, TOutput> MetaModel { get; set; } |
The model whose parameters the meta-learner trains to adapt quickly to new tasks. |
IGradientBasedOptimizer<T, TInput, TOutput> InnerOptimizer { get; set; } |
Optimizer for task adaptation. |
IGradientBasedOptimizer<T, TInput, TOutput> MetaOptimizer { get; set; } |
Optimizer for the meta-update. |
ILossFunction<T> LossFunction { get; set; } |
Loss function for meta-training; null uses the default multi-class classification loss. |
bool EnableCheckpointing { get; set; } |
|
bool OnlineInnerLoop { get; set; } |
Whether the inner loop is online, one update per support example in order (the paper’s Algorithm 1), rather than one full-batch update per step. |
bool UseFirstOrder { get; set; } |
Whether to use the first-order approximation of the meta-gradient. |
double InnerLearningRate { get; set; } |
|
double OuterLearningRate { get; set; } |
|
double PlnFraction { get; set; } |
Fraction of total model parameters that form the PLN (prediction head). |
double SparsityPenalty { get; set; } |
Optional L1 penalty on the RLN parameters, added to the meta-loss. |
double? GradientClipThreshold { get; set; } |
|
int AdaptationSteps { get; set; } |
|
int CheckpointFrequency { get; set; } |
|
int EvaluationFrequency { get; set; } |
|
int EvaluationTasks { get; set; } |
|
int MetaBatchSize { get; set; } |
|
int NumMetaIterations { get; set; } |
|
int? RandomSeed { get; set; } |
Methods
| Method | Summary |
|---|---|
IMetaLearnerOptions<T> Clone() |
|
bool IsValid() |