ML Kit: Turnkey APIs to use on-device ML in mobile apps

on YouTube ↗12m

" summary

Demo shows a pose estimator working with about 12 lines of code and ML Kit now exposes 12 turnkey APIs including new digital ink (300+ languages), entity extraction (11 entity types, 15 languages), selfie segmentation and a 33‑point pose detector with 33‑point skeletal landmarks and z depth. SDK runs on-device, offers custom model swapping via TF Hub / TensorFlow Lite and a standalone SDK (no Firebase), cuts object-detection latency up to 75% on Pixel 3 XL, 62.5% on Pixel 4, and claims use in 40,000 apps / 500M users with model bundle sizes like 6.1 MB for image labeling.

AI-written summary · about

" chapters (4)

1. 0:00Introduction
2. 3:52New APIs
3. 6:11MLKit SDK
4. 10:48Recap

" related talks

• 24mML Kit: Machine learning for mobile developersSachin Kotwani, Isabella Chen · Firebase Summit 2018
• 38mML Kit: Machine Learning for Mobile with FirebaseChristiaan Prins, Ann Zimmer, Shiyu Hu · Google I/O 2019
• 04mTop 5 reasons to add ML Kit to your Android or iOS appSachin Kotwani · Firebase Summit 2018
• 21mGetting the best of on-device ML with ML KitDong Chen · Android Dev Summit 2018
• 30mMachine Learning in Android Apps with ML KitFelipe Pedroso · AndroidFest Campinas 2018
• 39mCustomize Your App with MLKitVictoria Gonda · Øredev 2019
• 29mFirebase MLKit — AI in Android Town!!Sweety Bertilla Francis Xavier · Android Summit 2019
• 10mMachine Learning for mobileKemal El Moujahid · Google's Machine Learning Virtual Community Day 2021