Welcome to the
ML Blocks
Landing Page
ML Blocks are
ML Blocks are
A block-based user interface for creating TinyML models
A block-based user interface for creating TinyML models
With ML Blocks, users can assemble data sets, define neural network classifiers, then train, evaluate, and export their models onto microcontrollers. ML Blocks are for end-to-end development of TinyML models for learners and tinkerers at any level.
Read our paper from the 2022 IEEE Visual Languages / Human-Centered Computing Conference. Our slides from this conference are available as a Microsoft Stream (soon!).
Watch Randi's end-of-summer-internship demo. The video goes through the full process of using ML Blocks to train an accelerometer-based Harry Potter wand.
Try ML Blocks yourself. You will need to setup a micro:bit or other Jacdac compatible microcontroller. Keep this tab open so you can see the quickstart instructions (below).
mlblocks_quickstart_updated.pdf