Dela via

AutoMLExperimentExtension Class

  • Reference

Definition

Namespace:
Microsoft.ML.AutoML
Assembly:
Microsoft.ML.AutoML.dll
Package:
Microsoft.ML.AutoML v0.21.1
public static class AutoMLExperimentExtension
type AutoMLExperimentExtension = class
Public Module AutoMLExperimentExtension
Inheritance
Object
AutoMLExperimentExtension

Methods

SetBinaryClassificationMetric(AutoMLExperiment, BinaryClassificationMetric, String, String)

Set Microsoft.ML.AutoML.BinaryMetricManager as evaluation manager for AutoMLExperiment. This will make AutoMLExperiment uses metric as evaluation metric.
SetCheckpoint(AutoMLExperiment, String)

Set checkpoint folder for AutoMLExperiment. The checkpoint folder will be used to save temporary output, run history and many other stuff which will be used for restoring training process from last checkpoint and continue training.
SetCostFrugalTuner(AutoMLExperiment)

Set Microsoft.ML.AutoML.CostFrugalTuner as tuner for hyper-parameter optimization.
SetDataset(AutoMLExperiment, DataOperationsCatalog+TrainTestData)

Set train and validation dataset for AutoMLExperiment. This will make AutoMLExperiment uses TrainSet from trainValidationSplit to train a model, and use TestSet from trainValidationSplit to evaluate the model.
SetDataset(AutoMLExperiment, IDataView, IDataView, Boolean)

Set train and validation dataset for AutoMLExperiment. This will make AutoMLExperiment uses train to train a model, and use validation to evaluate the model.
SetDataset(AutoMLExperiment, IDataView, Int32, String)

Set cross-validation dataset for AutoMLExperiment. This will make AutoMLExperiment use n=fold cross-validation split on dataset to train and evaluate a model.
SetEciCostFrugalTuner(AutoMLExperiment)

set Microsoft.ML.AutoML.EciCostFrugalTuner as tuner for hyper-parameter optimization. This tuner only works with search space from SweepablePipeline.
SetGridSearchTuner(AutoMLExperiment, Int32)

set Microsoft.ML.AutoML.GridSearchTuner as tuner for hyper parameter optimization.
SetMulticlassClassificationMetric(AutoMLExperiment, MulticlassClassificationMetric, String, String)

Set Microsoft.ML.AutoML.MultiClassMetricManager as evaluation manager for AutoMLExperiment. This will make AutoMLExperiment uses metric as evaluation metric.
SetPerformanceMonitor(AutoMLExperiment, Int32)

Set DefaultPerformanceMonitor as IPerformanceMonitor for AutoMLExperiment.
SetPerformanceMonitor<TPerformanceMonitor>(AutoMLExperiment, Func<IServiceProvider,TPerformanceMonitor>)

Set a custom performance monitor as IPerformanceMonitor for AutoMLExperiment.
SetPerformanceMonitor<TPerformanceMonitor>(AutoMLExperiment)

Set a custom performance monitor as IPerformanceMonitor for AutoMLExperiment.
SetPipeline(AutoMLExperiment, SweepablePipeline)

Set pipeline for training. This also make AutoMLExperiment uses Microsoft.ML.AutoML.SweepablePipelineRunner , Microsoft.ML.AutoML.MLContextMonitor and Microsoft.ML.AutoML.EciCostFrugalTuner for automl traininng as well.
SetRandomSearchTuner(AutoMLExperiment, Nullable<Int32>)

set Microsoft.ML.AutoML.RandomSearchTuner as tuner for hyper parameter optimization. If seed is provided, it will use that seed to initialize Microsoft.ML.AutoML.RandomSearchTuner. Otherwise, Seed will be used.
SetRegressionMetric(AutoMLExperiment, RegressionMetric, String, String)

Set Microsoft.ML.AutoML.RegressionMetricManager as evaluation manager for AutoMLExperiment. This will make AutoMLExperiment uses metric as evaluation metric.
SetSmacTuner(AutoMLExperiment, Int32, Int32, Int32, Int32, Single, Int32, Int32, Double, Int32)

Set Microsoft.ML.AutoML.SmacTuner as tuner for hyper-parameter optimization. The performance of smac is in a large extend determined by numberOfTrees, nMinForSpit and splitRatio, which are used to fit smac's inner regressor.

Applies to