I gave an introduction to Machine Learning for Microcontrollers using MicroPython at PyData Berlin earlier this year. Talk is now available at:

Lately I have been working on #machinelearning for @micropython for running on #microcontrollers and it is starting to get into pretty good shape! Project:

MicroPython v1.24 is here! This release has RP2350 and ESP32-C6 support, a RISC-V native emitter and common TinyUSB code

MicroPython v1.23.0 is here! Dynamic USB devices, revamped webassembly port, openamp, tls, vfs modules https://t.co/Abgfv95bos

New #MachineLearning paper and dataset is out! The AeroSonicDB (YPAD-0523) Dataset for Acoustic Detection and Classification of Aircraft. Get it fresh off #Arxiv now! https://t.co/qHqxmD5vhk #noisepollution #noisemonitoring #audio

Added some colors to the emlearn logo - also add a basic brand color scheme. Basic but workable - pretty happy with it overall. #machinelearning #tinyml #embeddedsystems #microcontroller https://t.co/yh9yhZD0jC

Long time no blog. I wrote about the status and uses of the emlearn #TinyML project, which is 5 years old now! https://t.co/PZMTmHEGcQ #machinelearning #embedded

Today in Jon-does-things-he-does-not-really-know, I have attempted to make a logo for the https://t.co/Og9dnj46eA software project. Not sure if I am happy with it, but it was the best I could come up with... Also need color palette #sideproject #opensource #machinelearning

Added support in emlearn for K Nearest Neighbors (KNN) classifier. KNN models can be surprisingly powerful - and they support on-device learning out-of-the-box. Runs nicely on microcontrollers, albeit RAM hungry. https://t.co/PdwGEsGyet #TinyML #embedded

Hi @herokustatus @heroku We are having login issues. After entering 2 factor auth, just get sent in eternal loop between dashboard. and /login - 1 second between each. Happens across multiple accounts and browsers.

I originally started to hack on this @europython 2019 in Basel, when we got a PewPew badge that could run #CircuitPython

Now there are #MicroPython bindings for https://t.co/AsRoi03dfr - available at https://t.co/2P15k5CVtO This makes it possible to have the familiarity and convenience of #Python while deploying efficient #MachineLearning onto #microcontroller devices. #tinyml

There is now API documentation for the public C API for emlearn - https://t.co/DisdAGfZtW #microcontroller #machinelearning #tinyml

It has been over 10 years since I used MATLAB. Even managed to do my DSP and control systems classes in uni using only open-source Python/Octave/Scilab. But man, am I still envious of the MathWorks documentation pages: practical, clear, extensive

Now the userguide for emlearn, machine learning library for microcontrollers, have sections on Classification, Regression and Anomaly Detection. https://t.co/OmGIkPjDJv Still quite far left to have scikit-learn/librosa/skimage levels of doumentation, but it is a start!

Detecting events in sound/accelerometer/ECG/etc is a common task with #machinelearning on sensors. So I started some documentation on it for emlearn: https://t.co/BrlbNThesY #tinyml #arduino

There are also nice getting-started guides for https://t.co/AsRoi03dfr (TinyML inference engine for microcontrollers), both for Arduino and PC: https://t.co/80PzHaEZdh #scikitlearn #sklearn

https://t.co/AsRoi03dfr Machine Learning models with are portable from #microcontroller to #browser - I have documented how to get started with #Emscripten on https://t.co/efSJr7T4Mx Here is a trivial decision tree model running in browser #TinyML

Website bitrotted and down since November it seems :o Automatic PHP 8.x update got me. Back up again now at least...

Did a major cleanup of my notes on #TinyML / #embedded #machine learning, that I have been collecting over the last 4 years now. https://t.co/0vv9CrY1wE - most content now in subpages