MLSoundClassifier.ModelParameters.ModelAlgorithmType.transferLearning(featureExtractor:classifier:)
An algorithm that leverages the knowledge of a general-purpose model built into the operating system.
Declaration
case transferLearning(featureExtractor: MLSoundClassifier.ModelParameters.FeatureExtractorType, classifier: MLSoundClassifier.ModelParameters.ClassifierType)Parameters
- featureExtractor:
The extractor type the algorithm uses to detect features from the audio data.
- classifier:
The model type the algorithm uses to classify audio data.
Discussion
You typically use this transfer-learning algorithm to train an object detector in these situations:
Your training dataset has a limited number of examples.
You prefer your object detector’s Core ML model file to be as small as possible.