Nous acceptons la cryptomonnaie — les cryptos que vous détenez peuvent ainsi acheter du matériel réel.Payez en crypto, dépensez-la en matériel réel. Livraison assurée offerte partout dans le monde au-delà de 399 $. Cartons neutres, sans marque, expédiés depuis l’Allemagne.Livraison assurée offerte au-delà de 399 $. Votre taux est bloqué pendant 30 minutes dès l’ouverture de la caisse.Taux bloqué 30 minutes. Chaque article est scellé, son numéro de série est vérifié, et il est couvert par une garantie de 24 mois.Scellé, garanti 24 mois.

Articles suivis Se connecter
Green development board with M.2 B+M key edge connector and silver heatsink

MCDODO

MCDODO Coral M.2 Accelerator B+M Key development board

SKU CH-B32K-JDMMA

Performs 4 trillion operations per second at .5 watts per TOPS, running MobileNet v2 at 400 FPS on Debian Linux with TensorFlow Lite and AutoML Vision Edge support

  • TypeSingle Board Computers
  • Edge TPU delivers 4 trillion ops/sec at .5 watts per TOPS
  • Runs MobileNet v2 at 400 FPS for real-time vision
  • Works with any Debian Linux system with M.2 slot
  • TensorFlow Lite models compile directly for Edge TPU
  • AutoML Vision Edge trains and deploys custom image models

Coral M.2 Accelerator B+M Key

The Coral M.2 Accelerator B+M Key is an IoT edge device designed for developers adding on-device machine-learning inference to Debian-based Linux systems. It targets makers and engineers who need a compact accelerator that fits a compatible M.2 slot without requiring a full single-board computer redesign.

4 TOPS Edge TPU capacity

The on-board Edge TPU coprocessor delivers 4 trillion operations per second and runs MobileNet v2 at 400 FPS, giving ample headroom for real-time vision tasks such as object detection or classification. Workloads that demand heavier models or sustained batch processing will exceed the available compute budget and require a higher-performance accelerator.

Efficient thermal profile

At .5 watts per TOPS the module draws very little power, so passive cooling in a typical enclosure is usually sufficient and audible fan noise is not a concern. Continuous heavy inference will raise the board temperature, but the low power envelope keeps thermal throttling rare in normal ambient conditions.

Points forts

  • Edge TPU delivers 4 trillion ops/sec at .5 watts per TOPS
  • Runs MobileNet v2 at 400 FPS for real-time vision
  • Works with any Debian Linux system with M.2 slot
  • TensorFlow Lite models compile directly for Edge TPU
  • AutoML Vision Edge trains and deploys custom image models

Caractéristiques

TypeSingle Board Computers

Questions sur cet article

Will the Coral M.2 Accelerator B+M Key fit in the M.2 slot on my older Debian Linux board?

It integrates with any Debian-based Linux system that has a compatible card module slot, so an older board with the correct M.2 interface will work.

How can I verify the Edge TPU is performing high-speed ML inferencing after installation?

You can confirm it is working by running a TensorFlow Lite model such as MobileNet v2 and observing inference speeds up to 400 FPS.

Does the box include any cables or a heatsink, or do I need to buy them separately?

The board is supplied as a standalone M.2 accelerator module; any mounting hardware, cooling or adapter cables must be sourced separately.

Également dans ce rayon

Les clients ont aussi comparé

Même rayon, même caisse — huit monnaies et un taux bloqué 30 minutes.