SolidRun, a leading single-board computer provider, today unveiled the Hummingboard 8P Edge AI SBC. Equipped with an Intel i.MX 8M Plus System on Module (SoM) and Hailo-8 AI processor, this cutting-edge solution aims to empower engineers and innovators and accelerate AI deployment at the edge.
The Hummingboard 8P Edge AI is a game-changer in many industries, serving as an ideal platform for computer vision applications like computer vision and image and object recognition, surveillance, IoT, and more. This versatile SBC allows two cameras (Basler MIPI-CSI or USB 3.0 cameras for computer vision applications) to be connected simultaneously; additionally it performs AI analysis on streaming video as well as providing interactive AI applications using an HDMI/USB touchscreen display for AI analysis purposes.
Sasha Strizhiver, embedded product line manager of SolidRun, articulated the company’s mission: “Our aim has always been to simplify embedded system engineering and enable engineers to innovate freely without constraints. The Hummingboard 8P Edge AI SBC represents this vision by offering cutting-edge AI inference capabilities with broad industry support in one convenient package that’s also user friendly.”
At the core of this SBC is SolidRun’s energy-efficient i.MX 8M Plus SOM, enhanced with Hailo-8 M.2 PCIe module capable of delivering up to 28 TOPS from Hailo-8 and two from i.MX8M Plus). Remarkably, this high-performance package draws only 10W maximum power draw – offering unprecedented performance-to-power ratio in AI industry and enabling engineers to create unique AI solutions with cutting-edge edge inference capabilities that all run via PoE.
Hailo-8 AI accelerator stands out thanks to its proprietary dataflow architecture, excelling in low-power memory access with purpose-built pipeline elements and a distributed memory fabric.
Hummingboard 8P Edge AI supports several popular AI frameworks out of the box, including TensorFlow, TensorFlow Lite, Keras, PyTorch and ONNX. Its versatility helps engineers quickly get underway with their AI projects without worrying about framework compatibility issues.
This kit offers engineers unparalleled flexibility. Lightweight machine learning models can run on the i.MX 8M Plus SoC’s embedded network accelerator while more complex models use Hailo-8 AI accelerator, providing access to an entire world of AI application development and testing.
Yaron Ofer, Regional General Manager at Hailo comments: “SolidRun’s Hummingboard 8P Edge AI SBC marks an exciting advancement in edge AI development solutions. Boasting exceptional performance and energy efficiency, it enables engineers to fully exploit AI applications at the Edge while taking full advantage of Hailo-8 chip capabilities for driving innovation.”
SolidRun goes the extra mile by including everything engineers need for setup with its comprehensive Board Support Package (BSP). Conveniently preloaded onto an SD card, setup is quick and effortless allowing users to quickly test AI models with Hailo-8 processor.
This kit also features the Hummingboard 8P Carrier Board as well as the SolidRun i.MX 8M Plus SOM with either 4GB or 8GB RAM options, the Hailo-8 AI Accelerator M.2 module, WiFi/Bluetooth module with antenna and heatsink, power adapter and optional Basler daA3840-30mc image sensor capable of 4K image and video capture with HDR at 30fps for an 8MP resolution image sensor.
SolidRun’s Hummingboard 8P Edge AI measures just 100mm by 70mm and was engineered to withstand harsh environmental conditions, making it suitable for commercial and industrial applications with temperatures from 0deg C up to 70deg C or even as low as -40deg C.
Basler has rigorously verified and certified this SBC for use with their cameras, making it an excellent choice for surveillance, license plate recognition, gaming machines, smart vehicles, IoT solutions and more.
SolidRun’s Hummingboard 8P Edge AI SBC is now available from its official website and select distributors at an MSRP of $590 USD, complete with BSP, preloaded models and everything necessary for AI projects.
