The other day STMicroelectronics announced another update of their https://newsroom.st.com/media-center/press-item.html/n4370.html?ecmp=tt22077_gl_social_jun2021&fbclid=IwAR3KQ-vikZ8P7ZSmpnlBXcDJg5aD2jfLWaklaoSar4lY5YMs18_5Y5FmAeg IDE, which
"In addition to enabling development of neural networks for edge inference on STM32 microcontrollers (MCUs), the latest STM32Cube.AI release (version 7.0) supports new supervised and semi-supervised methods that work with smaller data sets and fewer CPU cycles. "
Which leads me to this question: have you ever used scikit-learn in a tinyML context? It is possible to convert sklearn models into ONNX, and to use ONNX with STM32Cube.AI, but has it been done yet?
I believe for some applications it could be a pretty good idea, and contrary to pure NNs those “classic” ML models like decision trees or SVMs are usually easily explainable. And scikit-learn has a really friendly API. Has anyone seen anything like that?