SmoothedLossDifficultyEstimator<T, TInput, TOutput>
Models & Types · AiDotNet.CurriculumLearning.DifficultyEstimators
Loss-based difficulty estimator with moving average smoothing.
Inheritance: DifficultyEstimatorBase<T, TInput, TOutput> → LossBasedDifficultyEstimator<T, TInput, TOutput> → SmoothedLossDifficultyEstimator<T, TInput, TOutput>
For Beginners
This variant uses a moving average of losses across training epochs. This helps stabilize difficulty estimates and prevents sudden changes in curriculum ordering due to random fluctuations.
Constructors
| Constructor | Summary |
|---|---|
SmoothedLossDifficultyEstimator(T smoothingFactor, ILossFunction<T> lossFunction, bool normalize) |
Initializes a new instance of the SmoothedLossDifficultyEstimator class. |
Properties
| Property | Summary |
|---|---|
string Name { get; } |
Gets the name of this estimator. |
Methods
| Method | Summary |
|---|---|
Vector<T> EstimateDifficulties(IDataset<T, TInput, TOutput> dataset, IFullModel<T, TInput, TOutput> model) |
Estimates difficulty with exponential moving average smoothing. |
void Reset() |
Resets the estimator including smoothed values. |