Machine Learning on Your Device: The Options
summary
Shows the exact workflow: train in TensorFlow, save a SavedModel, run the TensorFlow.js converter to get model.json plus its binary, or run the TF Lite Converter to produce a 612‑byte .tflite for mobile. Demonstrates bundling that .tflite into Android/iOS apps or deploying via ML Kit, and a microcontroller demo using a SparkFun Edge running a speech model <20 KB successfully on coin‑cell power. Also highlights Coral’s accelerator achieving Edge TPU 60 fps inference for video.
AI-written summary · about
chapters (15)
- 0:00Intro
- 0:46Agenda
- 1:10Overview
- 1:48Tensorflow
- 2:20How to get started
- 2:56Building a model
- 7:18Saving a model
- 8:13Code
- 8:50In Action
- 10:06The Options
- 11:39MLKit
- 15:00TensorFlow Lite Runtime
- 21:45Kotlin Demo
- 29:30Summary
- 31:55Demo
related talks
30mOn-device machine learning in mobile and web apps
17mDeploy a custom Machine Learning model to mobile
47mMachine Learning on Android Demystified
27mMachine Learning and Mobile
15mSimplified Machine Learning on mobile, browsers, and edge devices with on-device ML
39mMachine Learning on Android Demystified
34mTensorFlow Lite
09mMachine Learning on the Edge