Training and Deploying ML models to STM32 Microcontrollers
Jacob Beningo - Watch Now - Duration: 01:54:12
Machine learning (ML) has often been considered a technology that operates on high-end servers and doesn’t have a place in traditional embedded systems. That perception is quickly changing. This workshop will explore how embedded software engineers can get started with machine learning for microcontroller based systems.
This session balances theory with practical hands-on experience using an STM32 development board.
Attendees will learn:
- How to collect and classify data
- Methods available to embedded developers to train a model
- Hands-on experience training a model
- How to convert a model to run on an STM32 MCU
- How to run an inference on a microcontroller
Additional details for development board and tools will be provided closer to the conference.
11:04:01 From Michael Kirkhart : Yes 11:04:08 From Siva Aduri : yes 11:04:13 From Dave Comer : 10-4 11:09:50 From Dave Comer : What does AI Compatible mean? Sensors? 11:14:59 From rakhel : 0 11:15:01 From KN : 1 11:15:02 From NateWelch : 0 11:15:08 From Dave Comer : Begginer with some knowledge...Tried examples 11:15:09 From Devendra Chaudhari : 0 11:15:10 From John Phillips : 3 11:15:11 From Michael Kirkhart : Trained a model 11:15:16 From iiitmk : 3 11:15:19 From Al Anway : 1.5 11:15:35 From Remco Stoutjesdijk : i bought a book that sits on my desk so i must be a 5 11:15:41 From Radu Pralea : 1.(6) 11:15:53 From Eduardo Nahmad : Have done some NN and GA, don’t know what number would that correspond to. 11:16:02 From Zoltan : 1 11:23:18 From Dave Comer : Jason, Do you see LORA used with your ML IoT adventures. 11:23:49 From Dave Comer : Connected to CLOUD servers? 11:24:18 From Dave Comer : YES! 11:24:34 From Dave Comer : How to sign up access? 11:24:53 From Michael Kirkhart : The Things Network? 11:24:55 From Stephane Boucher : iotonlineconference.com 11:25:00 From Dave Comer : :) 11:25:21 From Stephane Boucher : Will be launched in a few weeks, will happen early December... 11:47:15 From Jacob Beningo : https://github.com/tensorflow/tflite-micro/blob/main/tensorflow/lite/micro/examples/hello_world/create_sine_model.ipynb 11:48:24 From Dave Comer : Contents have move from the link you sent but it explains where 11:48:49 From Dave Comer : yes 11:54:07 From Michael Kirkhart : Note: you need to be logged into a Google account to run the example. 12:01:34 From Dave Comer : How do you save your file(s) locally again? 12:02:10 From Dave Comer : Thanks 12:13:05 From Dave Comer : I'm done, but don't have a system avail. Will follow along as you go... 12:14:32 From Al Anway : watching only 12:18:33 From John Phillips : Not to dampen the enthusiasm too much, but all this is effectively a low-pass filter that cleaned up a sinewave waveform ? 12:19:50 From Dave Comer : But, when applied to an FFT 12:21:55 From KN : yes 12:49:35 From Stephane Boucher to Jacob Beningo(Direct Message) : We'll need to wrap things up very soon. Jenny's talks tarts in 10 minutes and it is done live on Zoom... 12:52:40 From Michael Kirkhart : Nice to see that ST has a version of STM32CubeIDE that runs native on Mac and linux. 12:53:13 From Michael Kirkhart : It was not too long ago that many dev tools were Windows only. 12:54:09 From Michael Kirkhart : What is the difference between Tensorflow Lite and TinyML?