Getting Started with Apple ML
summary
Shows how to add real models from Create ML inside Xcode and compares it with Turi Create (Python), training a “healthy snacks” image classifier on 4,800 images with feature extraction taking 35 minutes on an older Mac and under 3 minutes on a six‑core i7 GPU machine, classifier training in seconds, and final export as a tiny 312 KB model because iOS includes the base vision features; reports ~91% training vs ~89% validation and explains augmentation and overfitting controls.
AI-written summary · about
chapters (16)
- 0:00Intro
- 1:44APPLE ML JOURNAL
- 2:12WHAT IS MACHINE LEARNING?
- 4:47TRAINING & INFERENCE
- 5:17DEEP LEARNING?
- 5:58IS IT HARD / A LOT OF WORK?
- 8:20VISION FRAMEWORK
- 9:08READY-TO-USE MODELS
- 10:04CORE ML CONVERTERS
- 12:16DATA AUGMENTATION
- 12:31CREATE ML
- 14:39VALIDATION VS TEST
- 18:33TEXT CLASSIFIER
- 19:04TURI CREATE TASKS
- 20:59TRY IT!
- 21:26MACHINE LEARNING BOOK!
related talks
40mNavigating the Machine Learning Landscape for Apple Devices
34mDemystifying ML in iOS
01h 03mTraining dynamic ML models on iOS 15
25mGetting Started with Machine Learning in your iOS apps
30mLearning CoreML with fun
40mDeep Learning on iOS
22mDeeper into ARKit with CoreML and Turi Create
41mTensorFlow Neural Networks on iOS