Aceitamos criptomoeda — para que as moedas que detém possam comprar material informático real.Pague em cripto, gaste-a em material informático real. Envio seguro gratuito para todo o mundo acima de 399 $. Caixas neutras, sem marca, expedidas a partir da Alemanha.Envio seguro gratuito acima de 399 $. A sua taxa fica bloqueada durante 30 minutos assim que abre a finalização da compra.Taxa bloqueada 30 minutos. Cada artigo é selado, o seu número de série é verificado, e está coberto por uma garantia de 24 meses.Selado, garantido 24 meses.

Artigos guardados Iniciar sessão
Green rectangular M.2 development board with gold edge connector and silver heatsink

MCDODO

MCDODO Coral M.2 Accelerator A+E Key development board

SKU CH-QBRH-6NTK5

Performs high-speed ML inferencing at 400 FPS using 0.5 watts per TOPS on Debian Linux with TensorFlow Lite and AutoML Vision Edge support

  • TypeSingle Board Computers
  • High-speed ML inferencing at 4 TOPS using 0.5 watts per TOPS
  • Executes MobileNet v2 at 400 FPS for real-time vision tasks
  • Runs TensorFlow Lite models directly on the Edge TPU coprocessor
  • Builds custom image classifiers with AutoML Vision Edge support
  • Integrates with any Debian-based Linux system via M.2 A+E key slot

High-speed ML inferencing board

The Coral M.2 Accelerator A+E Key adds an Edge TPU coprocessor to any Debian-based Linux system with a compatible M.2 slot. It delivers 4 trillion operations per second while consuming only 0.5 watts per TOPS. Developers can run mobile vision models such as MobileNet v2 at 400 FPS in a power-efficient manner. This board targets makers and engineers building edge AI prototypes.

4 TOPS at 2 TOPS per watt

The accelerator handles TensorFlow Lite models compiled for the Edge TPU and supports AutoML Vision Edge for custom image classification. It runs efficiently on 0.5 watts per TOPS, making it suitable for battery-powered or thermally constrained devices. Workloads requiring higher throughput or different model frameworks will exceed the capabilities of this single coprocessor. Systems without a compatible A+E key M.2 slot cannot use the board.

Debian Linux and TensorFlow Lite integration

The board works with any Debian-based Linux distribution that exposes a compatible M.2 A+E key slot, removing the need for proprietary operating systems. TensorFlow Lite models compile directly for the Edge TPU without rebuilding from scratch. AutoML Vision Edge lets users train and deploy custom image classifiers through a cloud workflow. No additional software licences or bundled development tools are included with the hardware.

Pontos fortes

  • High-speed ML inferencing at 4 TOPS using 0.5 watts per TOPS
  • Executes MobileNet v2 at 400 FPS for real-time vision tasks
  • Runs TensorFlow Lite models directly on the Edge TPU coprocessor
  • Builds custom image classifiers with AutoML Vision Edge support
  • Integrates with any Debian-based Linux system via M.2 A+E key slot

Especificações

TypeSingle Board Computers

Perguntas sobre este artigo

Will the Coral M.2 Accelerator fit in the A+E key slot on my existing single-board computer?

Yes, the board uses an A+E key M.2 interface, so it installs directly into any compatible slot on a Debian-based Linux system.

How can I verify the Edge TPU is running at the rated 4 TOPS after installation?

Run a TensorFlow Lite benchmark such as MobileNet v2; the coprocessor should deliver around 400 FPS at 0.5 watts per TOPS.

Does the package include a heatsink or mounting screws, or must I source them separately?

The board ships as a bare M.2 module; thermal pads, heatsinks and mounting hardware are not included and must be purchased separately.

Também nesta secção

Os clientes também compararam

Mesma prateleira, mesma finalização da compra — oito moedas e uma taxa bloqueada durante 30 minutos.