Coral Mini PCIe Accelerator

TheCoral Mini PCIe Accelerator is a Mini PCIe module that adds the Edge TPU coprocessor to existing systems and products. The Edge TPU is a small ASIC developed by Google that powerfully accelerates TensorFlow Lite models: it is capable of performing 4 trillion operations per second (4 TOPS) while consuming 2 watts of power - that's 2 TOPS per watt. For example, a single Edge TPU can run the most advanced mobile vision models, such as MobileNet v2, at nearly 400 frames per second. On-device ML processing reduces latency, improves data privacy, and eliminates the need for a constant, high-bandwidth Internet connection.

The Mini PCIe Accelerator is a half-size Mini PCIe card designed to fit into any standard Mini PCIe slot. This form factor enables easy integration into ARM andx86 platforms to add local machine learning acceleration to products such as embedded platforms, mini-PCs, and industrial gateways.

Key Features:

• Supports TensorFlow Lite, enabling existing models to be used to run on Edge TPUs.
• Supports AutoML Vision Edge for easy creation and deployment of fast, highly accurate custom image classification models on the device.

The Coral Mini PCIe Accelerator must be connected to a host computer with the following specifications:

• Any Linux computer with a compatible Mini PCIe module slot Debian 6.0 or later or any derivative version (e.g. Ubuntu 10.0+)
• System architecture x86-64 or ARM32/64 with ARMv8 instruction set

Overall dimensions: 30mm x 26.8mm x 2.55mm

Weight: 3.6 g