Google Coral USB Accelerator – Edge TPU ML – ARM Cortex M0

From the shop's description

Google Coral USB Accelerator – Edge TPU ML – ARM Cortex M0 The Google Coral USB Accelerator is a powerful USB accessory designed to enhance your existing systems with accelerated machine learning (ML) inferencing capabilities. Featuring an ARM Cortex M0+ processor and Google’s proprietary Edge TPU chip, this compact device delivers high-speed, energy-efficient ML computations. It’s perfect for tasks like real-time object detection and image classification, capable of running advanced models such as MobileNet v2 at nearly 400 frames per second. Compatible with Linux, Windows, and macOS, it integrates seamlessly with devices like the Raspberry Pi via a USB 3.0 Type-C connection. Ideal for hobbyists, developers, and businesses, the Coral USB Accelerator brings on-device AI processing to your projects with reduced latency and enhanced privacy—no constant internet connection required. Coral USB Accelerator – Edge TPU ML accelerator. Features Powerful ML Inferencing : Accelerates TensorFlow Lite models with exceptional speed and efficiency. Edge TPU Chip : Google-designed ASIC for high-performance, low-power ML computations. Cross-Platform Compatibility : Supports Linux (Debian-based), macOS, and Windows 10 systems. Raspberry Pi Ready : Works effortlessly with popular single-board computers. Low Power Consumption : Delivers 2 TOPS per watt for energy-efficient operation. Easy Integration : Plugs into any USB 3.0 port for quick setup and use. Compact Design : Portable and unobtrusive, ideal for a variety of applications. Content of Coral USB accelerator kit. Google Coral Accelerator Requirements Hardware Requirements Host Computer: A computer running one of these operating systems: macOS: Version 10.15 (Catalina) or later. Requires MacPorts or Homebrew for installation. Linux: Debian 10 or derivatives (e.g., Ubuntu 18.04). Compatible with Raspberry Pi 3 Model B+ and Raspberry Pi 4. Supported architectures: x86‑64, Armv7 (32‑bit), or Armv8 (64‑bit). Windows: Version 10 or later. USB Port: At least one available USB port. For optimal performance, use a USB 3.0 port (USB 3.1 Gen 1). Software Requirements Python : Version 3.5, 3.6, 3.7 or later is required to run TensorFlow Lite models and use the Edge TPU API. Edge TPU Runtime and API Library : These must be installed on your host computer to communicate with the USB Accelerator. Follow the installation steps in the official guide . TensorFlow Lite Models : TensorFlow Lite model specifically compiled for the Edge TPU. Specifications Specification Details Processor ARM Cortex M0+ Co-processor Google Edge TPU Peak Performance 4 TOPS (int8) Power Efficiency 2 TOPS per watt Power Consumption Powered via USB, up to 900 mA at 5 V (max speed) Interface USB 3.0 (USB 3.1 Gen 1) Type-C Supported Operating Systems Linux Debian 10 or derivatives (e.g., Ubuntu 18.04; supports x86-64, Armv7, Armv8) macOS 10.15 (with MacPorts or Homebrew) Windows 10 Model Compatibility TensorFlow Lite models compiled for Edge TPU Dimensions 65mm x 30mm Operating Temperature Reduced frequency: Up to 35°C Maximum frequency: Up to 25°C Applications Real-Time Object Detection: Identify objects in video feeds for security or monitoring systems. Image Classification: Sort and categorize images for automation or analysis. Pose Estimation: Track movements for fitness, gaming, or healthcare solutions. Keyphrase Detection: Enable voice-activated features in smart devices. On-Device Transfer Learning: Customize ML models directly on the device for unique projects. TensorFlow Lite Projects: Accelerate any compatible ML inferencing task. Package Content 1 x Google Coral USB Accelerator – Edge TPU ML – ARM Cortex M0 1 x USB Cable type C – USB A Documents User guide Documentation