Who’s afraid of Machine Learning? First steps with TensorFlow for Android.
summary
Builds a simple Android image classifier that actually runs TensorFlow Lite locally using a memory‑mapped .tflite model and a plain labels text file The workflow is concrete not theoretical training happens on servers, models are exported and converted to TFLite because full TensorFlow models can be hundreds of megabytes The app loads the model from assets (do not compress), converts a Bitmap into a ByteBuffer input, invokes the interpreter, reads a probability array and displays the top‑3 labels with probabilities in real time and offline
AI-written summary · about
chapters (37)
- 0:00Intro
- 0:22Who am I
- 1:21Inspiration
- 3:39Brain
- 8:10Strawberry
- 15:47Training
- 17:03Hidden Layers
- 18:34End Goal
- 19:04Data Scientists
- 20:07TensorFlow
- 21:28Why not just send it to the server
- 24:06TensorFlow Mobile
- 25:07TensorFlow Light
- 27:34TensorFlow Lite
- 28:22TensorFlow Lite example
- 29:10Adding assets
- 30:05Adding TensorFlow library
- 30:21Dont compress TensorFlow
- 30:43Create damage classifier class
- 31:01Reference to TensorFlow interpreter
- 31:28Load model
- 31:57Load labels
- 32:09Use labels list
- 32:41Create label reader
- 33:10Human readable labels
- 33:28Calculating probabilities
- 34:39What about the input
- 35:09What about the byte buffer
- 35:53Running the model
- 36:11Creating the classifier
- 36:38Input
- 37:10iterate
- 37:56output
- 38:20labels
- 38:37a whole new profession
- 39:19learn more
- 40:01thank you
related talks
33mWho’s Afraid of Machine Learning? First Steps with TensorFlow for Android
47mMachine Learning on Android Demystified
19mOn-device machine learning: TensorFlow on Android
47mDeep Learning: a practical introduction from an Android dev perspective
23mAI is ready for you. Are you ready for AI?
21mLive Coding: Adding Machine Learning to Android Things in 10 Mins
40mApplied TensorFlow in Android apps
46mTensorFlow Lite in Android