Apple Introduces the Core ML Framework

Apple introduced Core ML at WWDC 2017 to make trained machine-learning models easier to run inside iPhone, iPad, and Mac applications. Developers could convert models from common frameworks into Apple’s format and execute inference through a high-level API.

Core ML handled details such as using the CPU, GPU, or later the Neural Engine depending on the model and device. Running inference locally reduced network latency and allowed some features to work offline without sending personal data to a server.

The framework supported tasks such as image classification, object recognition, language processing, and recommendation. Core ML became part of Apple’s broader strategy of treating on-device machine learning as an operating-system capability rather than forcing every application developer to write low-level numerical code.