Learning CoreML with fun
summary
Builds a custom CoreML object detector that recognizes Mario and friends and exports a .mlmodel that runs realtime on iPhone. Uses Image Preparation for annotation, TuriCreate and Python scripts to generate training data and train locally or in cloud. Training takes about 40 minutes on a modest Mac (recommends 4 CPU cores or GPU for faster runs); integration is a drop‑in into Apple’s sample app to produce a working “Mario Finder” that demos live detection.
AI-written summary · about
chapters (17)
- 0:00Introduction to CoreML
- 2:15Why CoreML for iOS
- 3:02Approaches to CoreML models
- 4:54Demo setup overview
- 7:53Installing Python tools
- 10:09Image annotation tools
- 11:38Preparing training images
- 13:25Setting up Python environment
- 15:09Running the training script
- 16:44Training time comparison
- 18:13Cloud training options
- 21:51Integrating custom model
- 23:18Live demo with Mario Bros
- 25:51Hardware benefits for ML
- 26:52CoreML 3 features
- 27:56Privacy with on-device ML
- 28:38Real-time detection workflow
related talks
32mReal-Time Object Detection with CoreML
40mNavigating the Machine Learning Landscape for Apple Devices
55mCore ML: Vision
22mGetting Started with Apple ML
40mDeep Learning on iOS
34mDemystifying ML in iOS
22mDeeper into ARKit with CoreML and Turi Create
31mPower-Up Your App With Apple’s Machine Learning Frameworks