A feature's value and its type bundled as a read-only instance.
- iOS 11.0+
- macOS 10.13+
- Mac Catalyst 13.0+Beta
- tvOS 11.0+
- watchOS 4.0+
- Core ML
A Core ML feature value, represented by
MLFeature, wraps both a data value and its type, represented by
MLFeature. For every
MLModel prediction, every input feature receives exactly one
MLFeature and every output feature produces exactly one
MLFeature for details about input and output features.)
However, the model and your app don't pass
MLFeature instances directly. Instead, your app uses an
MLFeature to pass
MLFeature instances to the model's input features and another
MLFeature to get
MLFeature instances from the model's output features.
If you use the dynamically generated interface for your Core ML model, you will use the model input and output interfaces specific to your model, instead of using
MLFeature, to manipulate values directly.