Microcontrollers and Machine Learning with MicroPython and emlearn
summary
Shows a $20 ESP32-class board running MicroPython plus emlearn C modules to do on-device inference, not just demos. Pure Python filter for 128 ms audio takes >1 s so emlearn C modules cut that to ~20% CPU, letting the device sleep and still sample audio, run an activity tracker trained on a few hours of data, and install C extensions at runtime via mpy install. Highlights I2S digital mic, npy installable modules.
AI-written summary · about
chapters (12)
- 0:00Microcontrollers and ML intro
- 1:41Presentation goal
- 2:25What is a microcontroller
- 4:09Install MicroPython
- 5:35Simple sensor example
- 7:24Power and package management
- 9:12Advanced sensor processing
- 12:00Optimizing with C modules
- 13:44Activity tracker demo
- 16:32Training ML models
- 17:55Key takeaways and emlearn
- 19:06Q&A
related talks
29mMicrocontrollers + Machine Learning with MicroPython in 1-2-3
44mMachine Learning on Microcontrollers With Zephyr and Emlearn
25mMachine Learning on Microcontrollers; Tiny, Useful and Fun- 44mMachine Learning on Microcontrollers With Zephyr and Emlearn
32mUsing MicroPython in the wild
34mUsing Python to enter the world of Microcontrollers
32mNext level Machine Learning with TinyML and Python
27mMicroPython: Embedded programming for everyone