SelfSupervisedLearningStepResult<T>

Models & Types · AiDotNet.SelfSupervisedLearning

Result of a single SSL training step.

For Beginners

This contains all the information from one training iteration, including the loss value and any additional metrics specific to the SSL method.

Constructors

Constructor Summary
SelfSupervisedLearningStepResult()

Properties

Property Summary
Dictionary<string, T> Metrics { get; set; } Gets or sets additional metrics specific to the SSL method.
T Loss { get; set; } Gets or sets the primary loss value for this step.
double? CurrentLearningRate { get; set; } Gets or sets the current learning rate (if adaptive).
double? CurrentTemperature { get; set; } Gets or sets the current temperature parameter (if applicable).
int NumNegativePairs { get; set; } Gets or sets the number of negative pairs in this batch.
int NumPositivePairs { get; set; } Gets or sets the number of positive pairs in this batch.