Hackaday | Google Launches AI Platform That Looks Remarkably Like A Raspberry Pi

Hackaday | Google Launches AI Platform That Appears Remarkably Like A Raspberry Pi

Google has promised us new {hardware} merchandise for machine studying on the edge, and now it is lastly out. The factor you are going to take away from that is that Google constructed a Raspberry Pi with machine studying. That is Google’s Coral, with an Edge TPU platform, a custom-made ASIC that’s designed to run machine studying algorithms ‘on the edge’. Right here is the hyperlink to the board that appears like a Raspberry Pi. Right here is the hyperlink to purchase it. You may cease studying now.

This new {hardware} was launched forward of the TensorFlow Dev Summit, revolving round machine studying and ‘AI’ in embedded purposes, particularly power- and computationally-limited environments. That is ‘the sting’ in advertising and marketing communicate, and already we have seen a number of merchandise designed from the bottom as much as run ML algorithms and inference in embedded purposes. There are RISC-V microcontrollers with machine studying accelerators out there now, and Nvidia has been engaged on this for years. Now Google is throwing their hat into the ring with a custom-designed ASIC that accelerates TensorFlow. It simply so occurs that the board appears like a Raspberry Pi.

What’s On The Board

On board the Coral dev board is an NXP i.MX 8M SOC with a quad-core Cortex-A53 and a Cortex-M4F. The GPU is listed as ‘Built-in GC7000 Lite Graphics’. RAM is 1 GB of LPDDR4, Flash is supplied with 8GB of eMMC, and WiFi and Bluetooth 4.1 are included. Connectivity is offered via USB, with Kind-C OTG, a Kind-C energy connection, a Kind-A 3.zero host, and a micro-B serial console. Gigabit Ethernet, a 3.5mm audio jack, a microphone, full-size HDMI, 4-lane MIPI-DSI, and 4-lane MIPI-CSI2 digicam help. The GPIO pins are precisely — and I imply precisely — just like the Raspberry Pi GPIO pins. The GPIO pins present the identical alerts in the identical locations, though as a result of totally different SOCs, you will have to vary a line or two of code defining the pin numbers.

You could be asking why Google would construct a Raspberry Pi clone. That reply comes within the type of a machine studying accelerator chip implanted on the board. Machine studying and AI chips had been standard within the 80s and all the pieces outdated is new once more, I suppose. The Google Edge TPU coprocessor has help for TensorFlow Lite, or ‘machine studying on the edge’. The purpose of TensorFlow Lite is not to prepare a system, however to run an present mannequin. It’s going to do facial recognition.

The Coral dev board is on the market for $149.00, and you may order it on Mouser. As of this writing, there are 1320 items on order at Mouser, with a supply date of March sixth (seek for Mouser half quantity 212-193575000077).

Additionally In Dongle Kind

There’s additionally one other system within the {hardware} portfolio, referred to as a USB accelerator, which we are able to solely assume is the Edge TPU related to a USB cable. This USB accelerator will work with the Raspberry Pi — that is from Google’s product copy, by the way in which — and can get you began on machine studying inferencing with the Edge TPU designed by Google. The value for this USB accelerator is $75 USD.

We want to congratulate the Raspberry Pi basis for creating one thing so ubiquitous even Google feels the necessity to experience the coat tails.

through https://ift.tt/2NIZTar


Related Posts

Leave a Reply