GaussianClassificationMetaDataset<T, TInput, TOutput>
Models & Types ยท AiDotNet.MetaLearning.Data
A synthetic meta-dataset for classification where each class is a Gaussian blob in feature space. Each task consists of N classes with randomly sampled means and a shared covariance, making it a standard benchmark for few-shot classification.
Inheritance: MetaDatasetBase<T, TInput, TOutput> โ GaussianClassificationMetaDataset<T, TInput, TOutput>
For Beginners
This dataset creates classification tasks where each class is a cluster of points centered around a random location. The meta-learner must learn to quickly identify which cluster a new point belongs to, even with very few examples.
Constructors
| Constructor | Summary |
|---|---|
GaussianClassificationMetaDataset(int numClasses, int examplesPerClass, int featureDim, double classSeparation, double clusterStdDev, int? seed) |
Creates a Gaussian classification meta-dataset. |
Properties
| Property | Summary |
|---|---|
IReadOnlyDictionary<int, int> ClassExampleCounts { get; } |
|
int TotalClasses { get; } |
|
int TotalExamples { get; } |
|
string Name { get; } |