{"product_id":"google-coral-m2-accelerator-with-dual-edge-tpu","title":"Google Coral M.2 Accelerator with Dual Edge TPU","description":"\u003cp\u003e\u003cstrong\u003e\u003cspan\u003eThe Coral M.2 Accelerator with Dual Edge TPU integrates two Edge TPUs into existing computer systems with the help of an M.2 E-key interface.\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cspan\u003eThis Edge TPU module is particularly suitable for mobile and embedded systems that can benefit from accelerated machine learning. \u003c\/span\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cspan\u003eEach Edge TPU coprocessor is capable of 4 trillion arithmetic operations per second (4 TOPS) with 2-watt power consumption. For example, modern Mobile Vision models such as MobileNet v2 can run efficiently at close to 400 FPS. \u003c\/span\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cspan\u003eWith two Edge TPUs (and thus 8 TOPS) you can \u003c\/span\u003e\u003cstrong\u003e\u003cspan\u003edouble\u003c\/span\u003e\u003c\/strong\u003e\u003cspan\u003e the performance of the system - for example, by running two models in parallel, or by distributing the processing steps of a model between both Edge TPUs. \u003c\/span\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cspan\u003eThe M.2 Accelerator with Dual Edge TPU is supported under Debian Linux based systems as well as under Windows 10. \u003c\/span\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cspan\u003eWe also carry the Coral USB Accelerator\u003c\/span\u003e\u003cspan\u003e - to be able to retrofit systems that do not offer an M.2 interface with a USB 3.0 interface.\u003c\/span\u003e\u003c\/p\u003e\n\u003ch2\u003eSpecification\u003c\/h2\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cspan\u003eTPU\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003e2 x Google Edge TPU ML accelerator coprocessor\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003e8 TOPS (int8); 2 TOPS per watt\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cspan\u003eInterface \u0026amp; software support\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003eInterface: M.2 E-key (with two PCIe Gen2 x1 lanes)\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eSupports Linux, and Windows 10 on the host system\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cspan\u003eOther data\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cspan\u003eDimensions: 22 mm x 30 mm x 2.8 mm (\u003c\/span\u003eM.2-2230-D3-E)\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eWeight: 2.5 g\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eOperating temperature: -40 to + 85 ° C\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cstrong\u003e\u003c\/strong\u003e\u003cstrong\u003e\u003cspan\u003eRequirements for the host system\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cspan\u003eLinux:\u003c\/span\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003e64-bit Linux Debian 10.0 \/ Ubuntu 16.04 (or newer)\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eCPU architecture: x86-64, or ARMv8 (64-bit)\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cspan\u003eor Windows:\u003c\/span\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003eWindows 10 (64-bit), \u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003ex86-64 CPU architecture\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cstrong\u003eGoogle Part Number\u003c\/strong\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003eG650-06076-01\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003ch2\u003e\u003cstrong\u003e\u003cspan\u003eImportant Notes\u003c\/span\u003e\u003c\/strong\u003e\u003c\/h2\u003e\n\u003cp\u003e\u003cspan\u003eAlthough the M.2 specification requires that E-key sockets provide two PCIe x1 lanes, most manufacturers only provide PCIe once! In order to be able to use both Edge TPUs, please make sure that your socket actually provides \u003c\/span\u003e\u003cstrong\u003e\u003cspan\u003etwo\u003c\/span\u003e\u003c\/strong\u003e\u003cspan\u003e PCIe x1.\u003c\/span\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cspan\u003ePlease also note the information in the datasheet on peak current requirements (up to 3 A per EdgeTPU) and thermal management. Each Edge TPU contains a built-in temperature sensor and allows parameters to be configured when it should be switched off. \u003c\/span\u003e\u003c\/p\u003e\n\u003cp\u003e\u003cstrong\u003eWarning: Overheating of the system can lead to fire or the destruction of hardware!\u003c\/strong\u003e\u003c\/p\u003e\n\u003ch2\u003e\u003cspan\u003ePotential for Industrial Applications\u003c\/span\u003e\u003c\/h2\u003e\n\u003cp\u003e\u003cspan\u003eThe Coral Edge TPU is a revolutionary product for machine learning applications. This enables embedded solutions that can, for example, recognize problems with workpieces, recognize traffic situations and much more.\u003c\/span\u003e\u003c\/p\u003e\n\u003ch2\u003e\u003cstrong\u003e\u003cspan\u003eResources\u003c\/span\u003e\u003c\/strong\u003e\u003c\/h2\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cspan\u003eM.2 accelerator module\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003eCoral M.2 Accelerator with Dual Edge TPU data sheet (PDF)\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eGetting started with the M.2 or PCIe Accelerator\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eTemperature management of the PCIe module\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cspan\u003eEdge TPU \u0026amp; models\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003eIntroduction to Models on the EdgeTPU\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eOverview of inferencing on the Edge TPU\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eOperation of multiple models with multiple Edge TPUs (English)\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eSplit a model across multiple Edge TPUs\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cspan\u003eAPI \u0026amp; Downloads\u003c\/span\u003e\u003c\/strong\u003e\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cspan\u003ePyCoral API (Python)\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eLibcoral API (C ++)\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eLibedgetpu API (C ++)\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eEdge TPU compiler\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eprecompiled models\u003c\/span\u003e\u003c\/li\u003e\n\u003cli\u003e\u003cspan\u003eAll software downloads\u003c\/span\u003e\u003c\/li\u003e\n\u003c\/ul\u003e","brand":"Google","offers":[{"title":"Default Title","offer_id":48624220045530,"sku":"ajg1789686962405","price":47.27,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0824\/8318\/3834\/files\/google-coral-m2-accelerator-with-dual-edge-tpu_1.jpg?v=1789686970","url":"https:\/\/memoryfoamwear.shop\/products\/google-coral-m2-accelerator-with-dual-edge-tpu","provider":"My Store","version":"1.0","type":"link"}