ExperimentSettings Class

Definition

Base class for experiment settings. All task-specific AutoML experiment settings (like BinaryExperimentSettings) inherit from this class.

public abstract class ExperimentSettings
type ExperimentSettings = class
Public MustInherit Class ExperimentSettings
Inheritance
ExperimentSettings
Derived

Constructors

ExperimentSettings()

Initializes a new instance of ExperimentSettings.

Properties

CacheBeforeTrainer

Whether AutoML should cache before ML.NET trainers. See WantCaching for more information on caching.

CacheDirectory

This is a pointer to a directory where all models trained during the AutoML experiment will be saved. If null, models will be kept in memory instead of written to disk. (Please note: for an experiment with high runtime operating on a large dataset, opting to keep models in memory could cause a system to run out of memory.)

CancellationToken

Cancellation token for the AutoML experiment. It propagates the notification that the experiment should be canceled.

MaxExperimentTimeInSeconds

Maximum time in seconds the experiment is allowed to run.

Applies to