Developers can now get their hands on Google's souped-up answer to the Raspberry Pi: the $150 Coral Dev Board, which features Google's Edge TPU machine-learning accelerator for low-powered devices that sit on the edge of a network.
The Edge TPU now features in the Coral-branded $75 USB 'thum bdrive' accelerator and as part of a removable 'system on module' that ships with a developer baseboard.
The Edge TPU Module includes an NXP i.MX 8M system on chip that consists of a quad-core Cortex-A53 and Cortex-M4F, a Vivante GC7000 Lite Graphics graphics processor, 8GB of eMMC storage, and 1GB of LDDR4 RAM. It also features Wi-Fi with 2x2 MIMO (802.11b/g/n/ac 2.4/5GHz) and Bluetooth 4.1. The add-on measures 48mm x 40mm x 5mm.
The baseboard has a RPi-like 40-pin GPIO expansion header, microSD slot for flash memory, USB ports, Gigabit Ethernet port, USB 2.0 and 3.0 ports for power and peripherals, a 3.5mm audio jack, and a terminal to wire up stereo speakers. It measures 88mm x 60mm x 24mm. There's of course video and camera interfaces for computer-vision applications.
Google notes that the Dev Board can be used as a single-board computer with accelerated machine-learning processing, but it also serves as an evaluation kit for manufacturers who want to use just the SOM combined with their own custom PCB hardware.
The Edge TPU-powered SOM will be capable of executing "state-of-the-art mobile vision models such as MobileNet v2 at 100+ fps, in a power efficient manner", according to Google.
As Hackster.io notes, the Edge TPU devices would be used in the second stage of developing a machine-learning algorithm after it has already been trained on a large dataset. Developers could then run already trained devices nearer to the data and have devices that use machine learning without the cloud.