Tensorflow 2.1.0 for microcontrollers benchmarks on STM32F746 #ML #MLmonday @ST_World
Stupid Projects benchmarks Tensorflow 2.1.0 for microcontrollers on the STM32F746.
Over 8 months ago I’ve started writing the Machine Learning for Embedded post series, which starts here. The 3rd post in this series (here) was about using the tensorflow lite for microcontrollers on the STM32746NGH6U (STM32F746-disco board). At that post I’ve did some benchmarks and in the next post, I’ve compared the performance with the X-CUBE-AI framework from ST.
At that time the latest TF version was 1.14.0 and the tflite-micro was still in experimental state. The performance was quite bad compared to X-CUBE-AI. CMSIS-DSP and CMSIS-NN was not supported and also optimized or compressed models weren’t supported, too. When I tried to use an optimized model then I was getting an error that INT8 is not supported in the framework.
So, after almost a year TF is now in version 2.1.0, 2.2.0 is around the corner and also the tflite-micro is no longer experimental. Also, CMSIS-DSP/NN is now supported for many kernels as also optimized models. Therefore, I felt like it make sense to give another try.
Eink, E-paper, Think Ink – Collin shares six segments pondering the unusual low-power display technology that somehow still seems a bit sci-fi – http://adafruit.com/thinkink
Stop breadboarding and soldering – start making immediately! Adafruit’s Circuit Playground is jam-packed with LEDs, sensors, buttons, alligator clip pads and more. Build projects with Circuit Playground in a few minutes with the drag-and-drop MakeCode programming site, learn computer science using the CS Discoveries class on code.org, jump into CircuitPython to learn Python and hardware together, TinyGO, or even use the Arduino IDE. Circuit Playground Express is the newest and best Circuit Playground board, with support for CircuitPython, MakeCode, and Arduino. It has a powerful processor, 10 NeoPixels, mini speaker, InfraRed receive and transmit, two buttons, a switch, 14 alligator clip pads, and lots of sensors: capacitive touch, IR proximity, temperature, light, motion and sound. A whole wide world of electronics and coding is waiting for you, and it fits in the palm of your hand.