Google's Raspberry Pi-like Coral board lands: Turbo-charged AI on a tiny computer

Google's own developer board helps developers bring machine-learning models to the network edge.

How to set up your Raspberry Pi 3 Model B+ This video walks you through everything you need to know to set up your Raspberry Pi 3 Model B+.

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. 

Google unveiled the tiny Edge TPU ASIC last July as its low-cost chip for bringing machine learning to sensors that can run machine-learning models on the TensorFlow lite framework. 

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 also launched a $25 five-megapixel Coral Camera Module that connects to the Dev Board via a camera flex cable. 

SEE: Sensor'd enterprise: IoT, ML, and big data (ZDNet special report) | Download the report as a PDF (TechRepublic)

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. 

According to to Hackster.io, Google plans to reveal more about the developer board and USB device at the TensorFlow Dev Summit, which kicks off today.

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.   

googlecoralaiboard.jpg

The baseboard has a Raspberry Pi-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.

Image: Google

Previous and related coverage

Monster BeagleBone AI crams computer-vision engine onto Raspberry Pi-style board

BeagleBoard AI offers developers a single-board computer that's customized for artificial-intelligence applications.

Superfast Raspberry Pi rival: Odroid N2 promises blistering speed for only 2x price

Hardkernel's Odroid N2 promises great performance for its price.

Raspberry Pi vs new Pine64 board: Same price but H64 has 2x memory

Pine H64 Model B also available with same memory as the Raspberry Pi 3 Model B+ but for $10 less.     

Raspberry Pi's latest competitor RockPro64 brings more power plus AI processor

The RockPro64 developer board is a Rockchip RK3399-powered alternative to the Raspberry Pi.

Raspberry Pi reveals Compute Module 3+: From $25, cooler, 8x more storage

The latest Raspberry Pi Compute Module runs cooler under heavy loads and is now available with up to 32GB storage.

Raspberry Pi with extra punch: New Orange Pi 3 packs powerful Allwinner H6 from $30

The Orange Pi 3 pushes RAM beyond the 1GB limit on earlier Orange Pi boards running on the Allwinner H6 system on chip.

New Raspberry Pi challenger promises 7x the speed at 2x the price, with Android 9 Pie, USB 3.0, 4K video TechRepublic

The Odroid-N2 single-board computer should be one of the fastest sub-$100 computers available.

A Raspberry Pi-style computer you can build yourself: Blueberry Pi

Provided you can think of something useful to do with a board with only 64MB of on-chip RAM.

Google AI on Raspberry Pi: Now you get official TensorFlow support

Google's TensorFlow team makes it a whole lot easier to get AI up and running on a Raspberry Pi.

Raspberry Pi-style Renegade Elite runs Android Oreo on six-core, 4K board

If you're specifically not looking for a Raspberry Pi, the Renegade Elite could be the board of choice.

Raspberry Pi meets AI: The projects that put machine learning on the $35 board

Explore the projects pushing the limit of what's possible on the budget board.

Raspberry Pi space rover: NASA open-sources its mini Mars robot

Now you can have your own six-wheel, rock-climbing rover robot for Earth exploration.

Raspberry Pi goes Android Auto: Now you can build your own cheap car head unit

Why buy a finished Android Auto head unit when you can hack one together with a Raspberry Pi 3?

Raspberry Pi's 'app store' lands with new Raspbian OS update

Raspberry Pi Foundation is hungry for beginners to try out its device, hence a new setup wizard and app store.

Raspberry Pi-style Renegade Elite runs Android Oreo on six-core, 4K board

If you're specifically not looking for a Raspberry Pi, the Renegade Elite could be the board of choice.

Google includes a Raspberry Pi in a DIY smart speaker kit CNET

The updated kits are rolling out to Target and include everything you need to build your own smart speaker or smart camera.

Raspberry Pi: A cheat sheet TechRepublic

Everything you need to know about the tiny, ultra-cheap computer that has taken the world by storm.