Contents

MLSoundClassifier.ModelParameters.FeatureExtractorType

The feature-extractor options for a sound-classifier training algorithm.

Declaration

enum FeatureExtractorType

Overview

Use the MLSoundClassifier.ModelParameters.FeatureExtractorType.audioFeaturePrint(type:revision:) feature extractor to create a model with the following advantages over vggish:

  • More accurate predictions

  • Lower latency

  • Smaller model file size

  • Less training time

Use MLSoundClassifier.ModelParameters.FeatureExtractorType.vggish(revision:) to support older OS versions.

Topics

Designating a feature extractor

Describing a feature extractor

See Also

Supporting types