Share via


KMeansTrainer.Options Class

Definition

Options for the KMeansTrainer as used in KMeansTrainer(Options).

public sealed class KMeansTrainer.Options : Microsoft.ML.Trainers.UnsupervisedTrainerInputBaseWithWeight
type KMeansTrainer.Options = class
    inherit UnsupervisedTrainerInputBaseWithWeight
Public NotInheritable Class KMeansTrainer.Options
Inherits UnsupervisedTrainerInputBaseWithWeight
Inheritance

Constructors

KMeansTrainer.Options()

Fields

AccelerationMemoryBudgetMb

Memory budget (in MBs) to use for KMeans acceleration.

ExampleWeightColumnName

Column to use for example weight.

(Inherited from UnsupervisedTrainerInputBaseWithWeight)
FeatureColumnName

Column to use for features.

(Inherited from TrainerInputBase)
InitializationAlgorithm

Cluster initialization algorithm.

MaximumNumberOfIterations

Maximum number of iterations.

NumberOfClusters

The number of clusters.

NumberOfThreads

Degree of lock-free parallelism. Defaults to automatic. Determinism not guaranteed.

OptimizationTolerance

Tolerance parameter for trainer convergence. Low = slower, more accurate.

Applies to