Tile, emerging through a branch of Coral one of the most recognizable companies in the world - Google. High quality equipment, designed to create any projects based on artificial intelligence AI.
Device USB Coral Accelerator delivers powerful capabilities that the formation of ML in the operating systems Linux, Windows and macOS. The module is equipped with a processor ARM Cortex M0+ and a special IC (ASIC) Edge TPU, designed and created by Google. It can be used to create, for example, modern models of cameras, such as MobileNet v2.
System Som Google Coral system is equipped with NXP.MX 8M SoC, which consists of czerordzeniowy core ARM Cortex A53 with a frequency of 1.5 GHz with support for Cortex M4F, and GPU Vivante GC7000 Lite. The device has 1 GB of LPDDR4 RAM, and 8 GB of internal memory eMMC. Supports multiple interfaces, particularly to Gigabyte Ethernet, USB 3.0, HDMI,...
Tile created by a division of Google Coral, which helps to create intellectual projects, AI. The tile was equipped with a system of SOM, which includes eMMC memory chip based on the NXP.MX 8M, wireless and Google, Edge TPU. Circuit Board has 1 GB of RAM and 8 GB of memory eMMC, USB connector type A and type C, 40 GPIO, Ethernet port, HDMI and more.
Google Coral is equipment that is intended primarily for developers. The device come from one of the most recognizable brands in the world - Google. The logo of the giant in electronics, most often can be associated with phones or computers called Pixel. Describes the hardware are devices that create the environment for the user where possible is machine learning algorithms. In stock Coral Google and Google Dev Board (high tech computer jednopłytkowy) and Coral USB Accelerator (a device with which an ordinary computer takes on a whole new).
Device Coral Dev Board is a small computer jednopłytkowy from Google. It would seem, is little different from the popular mini-computer Raspberry Pi or Arduino, but in reality equipment is far more focused on specific tasks. It is a product designed primarily for developers who want to use advanced machine learning algorithms. The computer is equipped with a working Mendel on Debian Linux, as well as to several libraries Tensor Flow Lite. There you will find a number of ready to compile models. The device is equipped with modern communication systems, wired and wireless, but for normal operation no connection to the Internet. All the processes are performed locally on your device.
The equipment is perfectly protected from overheating by using a heatsink with integrated fan. Although to be aimed at a specific group of clients, this device is perfectly equipped hardware, so that its application is very versatile. Dev Board offers such tools that help you easily implement the processes of the Assembly - a connector for DSI for monitors, connector CSI-2 (so you can connect the camera), a Gigabit Ethernet port, USB 2.0 connector, USB 3.0 connector, and GPIO (40 pins). The “heart” of the device is the CPU NXP iMX 8M (Quad-Core ARM Cortex A53 + ARM Cortex M4F), has 1 GB of RAM (LPDDR4), 8 GB of internal memory eMMC, and Wi-Fi interface (802.11 b/g/n/ac (2.4 GHz/5 GHz), Bluetooth 4.1, a slot for memory cards microSD, HDMI output mini-jack port for microphone and the speaker output (terminal 4-pin 2.54 mm). The most important element, however, is the so-called coprocessor Edge TPU, that is, the system tensorowy, which is used together with the Flow Tensor library for machine learning. The device has small dimensions (the dimensions of the main tiles 88 x 60 x 24 mm) and very low level of energy consumption.
Google Coral USB Accelerator is a small device that can run when connected via the USB port to the computer with the operating system Linux, Windows or MacOS. It should be noted that the hardware works perfectly with devices that are based on Linux, that is, by the way the Raspberry Pi. Computer when you connect USB Coral Accelerator gets access to a unique co-processor Edge (is intended for use in machine learning). We can say that this device, which in combination with a Raspberry Pi can become a cheaper analog of the original Google Dev Board.