# MLSoundClassifier.DataSource A representation of a sound-classifier dataset located in the file system or in a data table. ``` enum DataSource ``` ## Overview Use a data source to represent a dataset for training, validating, or testing a sound classifier. ## Topics ### Creating a data source [`case labeledDirectories(at: URL)`](/documentation/CreateML/MLSoundClassifier/DataSource/labeledDirectories(at:)) Creates a data source from a folder with subfolders that each contain audio files. [`case labeledFiles(at: URL)`](/documentation/CreateML/MLSoundClassifier/DataSource/labeledFiles(at:)) Creates a data source from a folder that contains audio files, each named after the sound they represent. [`case filesByLabel([String : [URL]])`](/documentation/CreateML/MLSoundClassifier/DataSource/filesByLabel(_:)) Creates a data source from a dictionary. [`case features(table: MLDataTable, featureColumn: String, labelColumn: String, parameters: MLSoundClassifier.FeatureExtractionParameters)`](/documentation/CreateML/MLSoundClassifier/DataSource/features(table:featureColumn:labelColumn:parameters:)) Creates a data source from a data table of audio features. [`case featuresDataFrame(DataFrame, featureColumn: String, labelColumn: String, parameters: MLSoundClassifier.FeatureExtractionParameters)`](/documentation/CreateML/MLSoundClassifier/DataSource/featuresDataFrame(_:featureColumn:labelColumn:parameters:)) Creates a data source from a data frame of audio features. [`struct FeatureExtractionParameters`](/documentation/CreateML/MLSoundClassifier/FeatureExtractionParameters) Parameters that affect the process of extracting sound features from audio files. ### Retrieving the data [`func labeledSounds() throws -> [String : [URL]]`](/documentation/CreateML/MLSoundClassifier/DataSource/labeledSounds()) Generates a dictionary of the data source’s labeled audio files. ### Partitioning the data [`func stratifiedSplit(proportions: [Double], seed: Int) throws -> [[String : [URL]]]`](/documentation/CreateML/MLSoundClassifier/DataSource/stratifiedSplit(proportions:seed:)) Generates an array of labeled audio dictionaries by splitting the data source into strata. [`func stratifiedSplit(proportions: [Double], generator: inout RNG) throws -> [[String : [URL]]]`](/documentation/CreateML/MLSoundClassifier/DataSource/stratifiedSplit(proportions:generator:)) Generates an array of labeled audio dictionaries by splitting the data source into strata using the random-number generator. --- Copyright © 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)