I recently put together a blog post featuring a tutorial on how to practically get started with TinyML, using the AZ3166 IOT Developer Kit. In the article and associated video, you’ll learn how to effectively build your own AI model using real-sensor data, in less than 10 minutes!
Have a look and let me know if you have any questions!
If you just want to check the video, here you go:
WoW, this looks great, can this be used on the RISC-V architecture? @benjamin537
Yes I believe this would work, as TensorFlow Lite has experimental support for RISC-V.
TensorFlow Lite is designed for all architectures and is heavy on the microcontroller. If we used a chip with less than 128K RAM, it would be very difficult to run smoothly, if optimized for a specific architecture, such as ARM. The overall running speed and footprint would be excellent. @benjamin537