Coral USB Accelerator with Edge TPU

TheCoral USB Accelerator is a USB accelerator designed to add additional computing power to your system to run machine learning programs. This device has a standard interface (USB port) for connection, making it suitable for many systems. The Coral USB Accelerator adds an Edge TPU coprocessor to your existing system that is capable of performing 4 trillion operations (terra operations) per second (TOPS) using 0.5 watts per TOPS (2 TOPS per watt). For example, it can deliver MobileNet v2 mobile vision tasks at nearly 400 frames per second with very low power consumption. The USB accelerator receives 5 V power from the USB interface via an integrated power management chip (PMIC). The device must be connected to the host computer using a USB cable that provides at least 500 mA at 5 V. Please note that when the device is running at maximum speed, it can consume a peak current of 900 mA. Depending on the computing load and operating frequency, the USB accelerator may become hot during normal operation.

Specifications:
Machine Learning Accelerator: The Google Edge TPU coprocessor is an ASIC developed by Google that delivers high performance machine logic inference for the TensorFlow Lite Model:
● 4 trillion operations per second (TOPS)
● 2 TOPS per watt

Interface: USB 3.0 (USB 3.1 Gen 1) port (SuperSpeed, 5 Gbps)

System requirements:
● A computer supporting one of the following operating systems:
○ Linux Debian 10 or its derivative (e.g. Ubuntu 18.04) and system architecture x86-64, Armv7 (32-bit) or Armv8 (64-bit) (Raspberry Pi is supported, but only Raspberry Pi 3B+ and Raspberry Pi 4 models were tested)
macOS 10.15 with MacPorts or Homebrew installed
○ Windows 10
One available USB Type-C port (use USB 3.0 port for maximum performance)
Python 3.5, 3.6, or 3.7.
Supports TensorFlow Lite.

Supports AutoML Vision Edge.

Overall dimensions: 65 mm x 30 mm x 8 mm.

Package contents:

1. USB accelerator - 1 pc.
2. USB Type-C to Type-A cable 300 mm (12 inches) long - 1 pc.