MetaFDMixupOptions<T, TInput, TOutput>
Options & Configuration · AiDotNet.MetaLearning.Options
Configuration options for Meta-FDMixup: Feature-Distribution Mixup for cross-domain few-shot learning (Xu et al., CVPR 2021).
Inheritance: ModelOptions → MetaFDMixupOptions<T, TInput, TOutput>
Implements: IMetaLearnerOptions<T>
How It Works
Meta-FDMixup improves cross-domain generalization by mixing feature distributions (gradient signals) between tasks in a meta-batch. Instead of mixing raw inputs, it mixes the gradient directions from different tasks, encouraging the meta-learner to find an initialization robust across diverse task distributions.
Constructors
| Constructor | Summary |
|---|---|
MetaFDMixupOptions(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 UseFirstOrder { get; set; } |
Whether to use the first-order approximation of the meta-gradient. |
double AlignmentWeight { get; set; } |
Weight for the feature distribution alignment loss. |
double InnerLearningRate { get; set; } |
|
double MixupAlpha { get; set; } |
Alpha parameter for the Beta distribution used to sample mixup coefficients. |
double MixupProbability { get; set; } |
Probability of applying mixup to each task’s gradient during inner loop. |
double OuterLearningRate { get; set; } |
|
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() |