RT from Pradeep (@pradeeplogu0)

Cellular based ML Person Detection on @Hacksterio using @EdgeImpulse @buildwithblues and @dfrobotcn FireBeetle ESP32S3 Sense. #TinyMl #EdgeImpulse #ESP32S3 #cellularIoT. https://www.hackster.io/pradeeplogu0/ml-person-detection-with-firebeetle-2-and-blues-cellular-c28f7a

Original tweet : https://twitter.com/pradeeplogu0/status/1671737540897042433

.@particle Photon 2 board and P2 module are based on a Realtek RTL8721DM Arm Cortex-M33 #WiFi and #Bluetooth microcontroller and both support #machinelearning through Edge Impulse. #Iot #lowpower #tinyml
https://www.cnx-software.com/2023/06/22/particle-photon-2-realtek-rtl8721dm-dual-band-wifi-and-ble-iot-board-particle-p2-module/

Original tweet : https://twitter.com/cnxsoft/status/1671731937478864896

RT from Marcelo Jose Rovai (@mjrovai)

Here is my 3rd tutorial on #TinyML Made Easy: w/ @seeedstudio XIAO ESP32S3 Sense & @EdgeImpulse. Now, working with audio: "KeyWord Spotting (KWS)" https://www.hackster.io/mjrobot/tinyml-made-easy-keyword-spotting-kws-5fa6e7

Original tweet : https://twitter.com/mjrovai/status/1663253018903445524

RT from Marcelo Jose Rovai (@mjrovai)

Great #TinyML project from my @unifei_oficial students. They used the @seeedstudio XIAO BLE sense and @EdgeImpulse Studio for DSP, training, and deployment. This project was selected as #1 on the @ictpnews 2023 SciTiny Workshop Show&Tell section. Congratulations!!!!https://twitter.com/EdgeImpulse/status/1661783307812581405

Original tweet : https://twitter.com/mjrovai/status/1661797948630466568

RT from Marcelo Jose Rovai (@mjrovai)

The new @seeedstudio XIAO ESP32S3 Sense, with its integrated microphone, works very smoothly with a #TinyML Keyword Spotting application, trained and deployed using the @EdgeImpulse Studio. A new KWS tutorial soon! 😜

Original tweet : https://twitter.com/mjrovai/status/1661323285064237058

RT from Marcelo Jose Rovai (@mjrovai)

#TinyML Made Easy: Image Classification using the new @seeedstudio XIAO ESP32S3 Sense and @EdgeImpulse. The XIAO is a giant tiny device now exploring Computer Vision applications. https://www.hackster.io/mjrobot/tinyml-made-easy-image-classification-cb42ae

Original tweet : https://twitter.com/mjrovai/status/1654609630712983555

RT from Marcelo Jose Rovai (@mjrovai)

The new @seeedstudio XIAO ESP32S3 Sense is a powerful and versatile device, well-prepared for #tinyML projects. Here is a preview of my Web application to help with image dataset collection. The next step will be to run it with @EdgeImpulse Studio and publish it on @Hacksterio.

Original tweet : https://twitter.com/mjrovai/status/1653858764951281668

Open-source hardware earable design based on the same @ublox BLE module (@NordicTweets nRF52840-based) found in the @arduino Nano 33 BLE Sense board. #arduino #OSHW #Hearables #audio #sensors #TinyML #wearables
https://www.cnx-software.com/2023/05/02/openearable-open-source-hardware-earable-platform-leverages-arduino-nano-33-ble-sense-ecosystem/

Original tweet : https://twitter.com/cnxsoft/status/1653309964876931074

.@dfrobotcn tiny UART/I2C offline voice recognition module with over 120 preset voice commands, customizable wake word and voice commands, and easily programmable with the #Arduino IDE. #tinyml #machineleaning #smartspeaker #diy #smartaudio
https://www.cnx-software.com/2023/04/21/offline-voice-recognition-module-arduino-programming/

Original tweet : https://twitter.com/cnxsoft/status/1649308232740143107

.@LucidVisionLabs SENSAiZ SZP123S-001 smart #camera based on a Sony IMX500 camera sensor with DSP and SRAM works with @SonySemicon_JP's AITRIOS edge AI software platform. #AI #machinevision #tinyml #raspberrypi
https://www.cnx-software.com/2023/04/17/sony-imx500-based-smart-camera-works-with-aitrios-software/

Original tweet : https://twitter.com/cnxsoft/status/1647863163713630209

Le'ts count 🐝 and 🔩!

On the latest episode of Let's Get Personal: Computing I'm joined by @shawnhymel from Edge Impulse who has a screw counter using object detection on a microcontroller with FOMO.

#AI #TinyML

https://youtu.be/bpVejDLZxYY

#TIL that 87% of #ML projects do not succeed!

Sometimes that is just because #ML is simply not the right solution.

Sometimes it is because they did not get it to production in a sustainable fashion: with a process to keep on collecting data, re-training, re-validating and re-deploying automatically #CICD.

This is called #MLOps (analog to #DevOps) and I have just completed this last #training to get my future clients started well with #ML or #TinyML!

https://credentials.edx.org/credentials/97ee94165e59477aad0323eb55ad1018/

RT from Marcelo Jose Rovai (@mjrovai)

"#TinyML under the hood: Spectral Analysis". Learn how does @EdgeImpulse Spectral Block work, extracting meaningful features from time-series signals. https://www.hackster.io/mjrobot/tinyml-under-the-hood-spectral-analysis-94676c

Original tweet : https://twitter.com/mjrovai/status/1638272155388346368

For the first time ever, tinyML Summit is opening up the sponsor exhibit and demo tables to the public! Join the expo on March 27th from 1:30 pm -5:30 pm in Burlingame, California: https://form.jotform.com/230337458061150
#tinyMLSummit #tinyML

Original tweet : https://twitter.com/Hacksterio/status/1638262596741615621

#arduino shield based on @aspinity AML100 analog #machinelearning learning chip that - when combined with an Arduino Uno compatible board - is supposed to consume much less power than traditional digital ML solutions. #tinyml #embeddedworld2023 #lowpower
https://www.cnx-software.com/2023/03/13/aspinity-ab2-aml100-arduino-shield-supports-ultra-low-power-analog-machine-learning/

Original tweet : https://twitter.com/cnxsoft/status/1635107553532342273

…and here I go, I completed the last course, and gain that sweet professional certificate: https://credentials.edx.org/credentials/b95f9f8743b94a13965fc4a10bf9d75f/

But more importantly, I learned to cram these #TinyML models into few dozens of kilobytes to do keyword spotting, people detection, movement recognition on a tiny device. #AI

It was in C++ and self-contained to run anywhere #TensorFlow Micro runs.

I’m done with the second course on #TinyML #edX.
https://courses.edx.org/certificates/11156eea6ed34ec69694d82d3960be26

This one was more focused on how to reduce the models while keeping them accurate enough. I also learned how to compensate for bias, especially when social aspects are involved.

Next course: we’re getting them running on #Arduino.

I have just started taking an #onlinecourse for #TinyML, the one from #Harvard on #edX: https://www.edx.org/professional-certificate/harvardx-tiny-machine-learning.

Tiny Machine Learning on as Little as 2 KB of RAM https://hackaday.com/2023/02/24/tiny-machine-learning-on-as-little-as-2-kb-of-ram/ #ArtificialIntelligence #SoftwareDevelopment #machinelearning #tensorflow #TinyML

New revision of the @arduino Nano 33 BLE Sense board with different #sensors and a few other updates. #tinyml #machinelearning #arduino #opensource #OSHW
https://www.cnx-software.com/2023/02/24/arduino-nano-33-ble-sense-rev2-switches-to-bmi270-bmm150-imus-hs3003-temperature-humidity-sensor/

Original tweet : https://twitter.com/cnxsoft/status/1629048098109681664