With the release of its latest devices, Apple is putting the emphasis on machine learning. They designed a custom chip—the neural engine. This can achieve fast deep learning operations while maintaining a low power usage. To fully benefit from this chip, developers must use a set of official APIs called Core ML (refer to the documentation at https://developer.apple.com/documentation/coreml).
To use an existing model with Core ML, developers need to convert it to the .mlmodel format. Thankfully, Apple provides Python tools to convert from Keras or TensorFlow.
In addition to speed and energy efficiency, one of the strengths of Core ML is its integration with other iOS APIs. Powerful native methods exist for augmented reality, face detection, object tracking, and much more.
