Modular Mojo claims to be over 36,000 times faster than Python for AI workloads

Modular Mojo vs Python matmul

Modular Mojo is a new programming language designed for AI developers that is said to combine the usability of Python with the performance of C with over 36,000 times the performance of Python on a matrix multiplication workload. Modular Mojo programming language was not in the initial plan of the company but came about when the company’s founders – who focused on building a platform to unify the world’s ML/AI infrastructure – realized that programming across the entire stack was too complicated and also ended up writing a lot of MLIR (Multi-Level Intermediate Representation) by hand. The “over 36,000 times speedup” claim comes with the matmul.py script performing a 128×128 matrix multiplication in Python with a throughput of 0.00215 GFLOP/s and another script doing 512×512 vectorized + parallelized matrix multiplication in Mojo at 79.636 GFLOP/s. The claim looks dubious and that’s odd they used different matrix sizes, but some are […]

TRACEPaw sensorized paw helps legged robots “feel the floor” with Arduino Nicla Vision

TRACEPaw

Our four-legged friends don’t walk on tarmac the same way as they do on ice or sand as they can see and feel the floor with their eyes and nerve endings and adapt accordingly. The TRACEPaw open-source project, which stands for “Terrain Recognition And Contact force Estimation through Sensorized Legged Robot Paw“, aims to bring the same capabilities to legged robots. Autonomous Robots Lab achieves this through the Arduino Nicla Vision board leveraging its camera and microphone to run machine learning models on the STM32H7 Cortex-M7 microcontroller in order to determine the type of terrain and estimate the force exercized on the leg. But the camera is apparently not used to look at the terrain, but instead, at the deformation of the silicone hemisphere – made of “Dragon Skin” – at the end of the leg to estimate 3D force vectors, while the microphone is used to recognize terrain types […]

MediaPipe for Raspberry Pi released – No-code/low-code on-device machine learning solutions

MediaPipe Studio Raspberry Pi 4

Google has just released MediaPipe Solutions for no-code/low-code on-device machine learning for the Raspberry Pi (and an iOS SDK) following the official release in May for Android, web, and Python, but it’s been years in the making as we first wrote about the MediaPipe project back in December 2019. The Raspberry Pi port is an update to the Python SDK and supports audio classification, face landmark detection, object detection, and various natural language processing tasks. MediaPipe Solutions consists of three components: MediaPipe Tasks (low-code) to create and deploy custom end-to-end ML solution pipelines using cross-platform APIs and libraries MediaPipe Model Maker (low-code) to create custom ML models MediaPipe Studio (no-code) webpage to create, evaluate, debug, benchmark, prototype, and deploy production-level solutions. You can try it out directly in your web browser at least on PC and I could quickly test the object detection on Ubuntu 22.04. MediaPipe Tasks can be […]

Mini review of GGtag e-paper display programmable through sound or USB serial

GGTag Price Tag

GGtag is a 3.52-inch e-paper display based on the Raspberry Pi RP2040 microcontroller and programmable through sound or USB serial from your web browser, plus it also supports emulation of 125 kHz RFID tags (ASK and FSK). When Radoslav Gerganov contacted me about the upcoming Crowd Supply campaign for the GGtag e-paper badge, I happened to have just discussed using an e-paper display to sell some samples on Facebook Groups where requirements include using the seller’s name and date in the photos. So I asked for a sample if any were available, and I just received it today… GGTag specifications: MCUs Raspberry Pi RP2040 dual-core Cortex-M0+ microcontroller @ 133 MHz with 264KB SRAM Microchip ATtiny85 8-bit AVR microcontroller (used for RFID) Storage – 2 MB flash Display – 3.52-inc e-paper display with USB – 1x USB Type-C port for power and serial programming Sensor – PDM digital microphone for data-over-sound […]

myCobot Pro 600 Raspberry Pi 4-based robot arm supports 600mm working range, up to 2kg payload

myCobot Pro 600 robotic arm

Elephant Robotics has launched its most advanced 6 DoF robot arm so far with the myCobot Pro 600 equipped with a Raspberry Pi 4 SBC, offering a maximum 600mm working range and support for up to 2kg payloads. We’ve covered Elephant Robotics’ myCobot robotic arms based on Raspberry Pi 4, ESP32, Jetson Nano, or Arduino previously, even reviewed the myCobot 280 Pi using both Python and visual programming, and the new Raspberry Pi 4-based myCobot Pro 600 provides about the same features but its much larger design enables it to be used on larger areas and handles heavier objects. myCobot Pro 600 specifications: SBC – Raspberry Pi 4 single board computer MCU – 240 MHz ESP32 dual-core microcontroller (600 DMIPS) with 520KB SRAM, Wi-Fi & dual-mode Bluetooth Video Output – 2x micro HDMI 2.0 ports Audio – 3.5mm audio jack, digital audio via HDMI Networking – Gigabit Ethernet, dual-band WiFi […]

Yahboom DOFBOT 6 DoF AI Vision robotic arm for Jetson Nano sells for $289 and up

Yahboom AI robot arm for Jetson Nano

Robotic arms can be expensive especially if you want one with AI Vision support, but Yahboom DOFBOT robotic arm designed for NVIDIA Jetson Nano offers a lower cost alternative as the 6 DoF robot arm sells for about $289 with a VGA camera, or $481 with the Jetson Nano SBC included. We previously published a review of the myCobot 280 Pi robotic arm from Elephant Robotics, and while it’s working well, supports computer vision through the Raspberry Pi, and is nicely packaged, it sells for around $800 and up depending on the accessories, and one reader complained the “price tag is still way too high for exploration“. The DOFBOT robotic arm is looking more like a DIY build, but its price may make it more suitable for education and hobbyists. DOFBOT robotic arm main components and specifications: SBC – NVIDIA Jetson Nano B01 developer kit recommended, but Raspberry Pi, Arduino, […]

Review of SunFounder Raspberry Pi UPS Power Supply

PiPower Raspberry Pi UPS charging

SunFounder’s Raspberry Pi UPS Power Supply is a complete UPS kit for the Raspberry Pi 3/4 Model B/B+ with a PiPower board, a 2,000 mAh battery, and all accessories requires for the assembly. It also works with other Raspberry Pi-sized boards that support 5V DC input such as Banana Pi BPI-M5, Libre Computer ROC-RK3328-CC, and other similar SBCs. Many years ago, I bought a Raspberry Pi battery pack for review hoping that it would also work as a UPS, but it was not perfect as the board would sometime reboot during power failure simulations. Since then, there have been many UPS kits launched to the market, but I didn’t try any so far, so when SunFounder contacted CNX Software to review their “Raspberry Pi UPS Power Supply”, I took the opportunity, and I will report my finding in this review. Raspberry Pi UPS Power Supply key features UPS module output […]

UniHiker review – A Linux-based STEM education platform with IoT and AI support, Micro:bit edge connector

UniHiker Review

DFRobot’s UniHiker is a STEM educational platform that was originally launched in China, but now UniHiker is now available worldwide through the DFRobot shop. The company has sent us a UniHiker sample for review, so let’s unpack the kit and learn how to use the UniHiker platform. The main component of the kit is the Linux-powered UniHiker board which features a 2.8-inch resistive touchscreen display and a BBC Micro:bit edge connector, so we can use expansion boards for the Micro:bit board. Let’s start unboxing it together. UniHiker unboxing DFRobot sent us the UniHiker platform by DHL. The package is a familiar-looking DFRobot box in orange color and comes with a plastic box to safely store the UniHiker board and accessories after use. The plastic box contains another plastic box with the board, some 3-pin and 4-pin cables for Gravity ports, and a USB Type-C cable. The UniHiker is like a […]

EmbeddedTS embedded systems design