This repository provides a collection of example projects demonstrating how to integrate machine learning models into iOS applications using Apple's Core ML and Google's ML Kit. It targets iOS developers looking to implement on-device AI capabilities, offering practical code for various computer vision and natural language processing tasks.

How It Works

The project showcases the typical workflow for integrating ML models on iOS: converting models from frameworks like TensorFlow to Core ML or ML Kit compatible formats, and then implementing pre- and post-processing logic within the iOS app. It highlights the differences and similarities between Core ML and ML Kit for various tasks.

Quick Start & Requirements

• Installation: Clone the repository and open the relevant Xcode project.
• Prerequisites: Xcode, iOS SDK.
• Resources: Projects are self-contained; no external datasets or API keys are required for the provided demos.
• Documentation: Links to Apple Developer documentation for Core ML, ML Kit, and related technologies are provided.

Highlighted Details

• Demonstrates image classification, object detection, pose estimation, depth prediction, and semantic segmentation.
• Includes examples for text recognition and face detection using ML Kit.
• Features an annotation tool for custom dataset creation and projects using Apple's Create ML.
• Provides performance benchmarks (inference time, execution time, FPS) on an iPhone X for various models.

Maintenance & Community

The repository was moved from the @motlabs group, with thanks to @jwkanggist. Further community engagement details are not explicitly provided in the README.

Licensing & Compatibility

The repository's licensing is not explicitly stated in the README. However, the underlying frameworks (Core ML, ML Kit) are generally compatible with commercial iOS applications.

Limitations & Caveats

Some "TODO" items are noted for specific ML tasks (e.g., Object Detection with ML Kit). Performance metrics are specific to an iPhone X and may vary on other devices. The "Tensorflow Mobile" section is marked as DEPRECATED.