diy projects
Seeed’s reComputer RK3576 Kit Runs YOLO11 at 80 FPS for $219
Seeed's new RK3576 dev kit packs a 6 TOPS NPU, a Pi-compatible 40-pin header and an optional 20 TOPS accelerator. Here's whether your build…
arduino
Arduino VENTUNO Q Runs a Vision Model on a Robot Arm
A $299 Arduino VENTUNO Q runs a vision-language-action model on-board, teaching an SO-101 robot arm to pick and place from just 50 demonstrations.
microcontrollers
Someone Squeezed a Generative AI Image Model Onto a $1 RP2350
Five seconds, 520 KB of RAM and about a dollar of silicon: the RP2350 inside the Raspberry Pi Pico 2 is now generating 128x128…
arduino
This Arduino Watches a Treadmill With a Webcam to Clock Speed
A rented treadmill, a USB webcam, and an Arduino VENTUNO Q: an on-board neural net reads belt speed and broadcasts it over Bluetooth as…
news & releases
Banana Pi’s BPI-AI2N Packs 15 TOPS of Vision AI Into a Module
Banana Pi's BPI-AI2N runs vision models on-device: a Renesas RZ/V2N module with 15 TOPS, 8 GB of RAM, dual GbE and two MIPI CSI…
diy projects
TinyML on a Raspberry Pi 3: LesionIQ Scans Skin Lesions Offline
A PhD student in Edinburgh fit a skin-lesion classifier onto a Raspberry Pi 3 with no internet at all. Here is how that constraint…
raspberry pi
LiteRT on Raspberry Pi 5: Run Gemma and Vision Models Locally
Google's LiteRT CLI puts Gemma and EfficientNet on a Raspberry Pi 5 with three commands. Here is the setup, the file sizes, and the…
news & releases
NVIDIA Jetson Orin Nano 2 Packs 78 TOPS Into a 15-Watt Board
Double the inference of the Orin Nano Super at 78 TOPS, plus 40 percent less power in 15-watt mode. What NVIDIA's new Jetson means…
arduino
Arduino VENTUNO Q: 8-Core Dragonwing SBC With an STM32 Sidekick
Arduino's VENTUNO Q pairs an 8-core Dragonwing with an STM32, 16GB of LPDDR5 and a 40 TOPS DSP for $299. Here's what to have…
news & releases
NVIDIA Retires Jetson Xavier: What ECE Students Should Buy Now
NVIDIA set a July 2027 end of life for every Jetson Xavier module. Here is what that means for your capstone build, and which…
iot & smart home
Run Gemma AI Offline on a Raspberry Pi 5 with Google’s LiteRT
Google's LiteRT runs Gemma models on a Raspberry Pi 5 at 99 tokens/sec prefill in 1432 MB of RAM, with no cloud and no…
microcontrollers
XCORE-VISION: Real-Time Embedded Vision, No Pi Needed
XMOS' XCORE-VISION runs YOLOv8 and MobileNetV2 on-device with an 8MP camera, landing between the ESP32-CAM and a Raspberry Pi for real-time vision.