🚀 Elevate Your Tech Game with Google Coral!
The Google Coral USB Accelerator is a powerful machine learning accessory designed to enhance existing systems. With its edge TPU coprocessor, it delivers an impressive 4 trillion operations per second while maintaining energy efficiency. Compatible with Debian Linux and equipped with USB 3.0 Type-C connectivity, it allows for seamless integration and rapid deployment of TensorFlow Lite models, making it an essential tool for developers and tech enthusiasts alike.
RAM | DDR3L |
Brand | |
Series | Coral USB Accelerator |
Item model number | G950-06809-01 |
Operating System | Linux |
Item Weight | 3.52 ounces |
Product Dimensions | 5.51 x 2.36 x 4.72 inches |
Item Dimensions LxWxH | 5.51 x 2.36 x 4.72 inches |
Color | Coral |
Processor Brand | Intel |
Number of Processors | 1 |
Voltage | 5 Volts |
Manufacturer | Google LLC |
ASIN | B07S214S5Y |
Is Discontinued By Manufacturer | No |
Date First Available | July 26, 2019 |
W**T
Works well with Frigate NVR on Debnian
Works immediately with Frigate NVR software and WhosAtMyFeeder, running on an older Lenovo SFF PC with Debian 12. Nerdy fun!
W**M
Exactly what Frigate ordered
Amazingly, your average home user can now build hardware workloads to handle complex outputs. You should add this to your setup if you want to tinker and develop your local loop. I have it attached to a Lenovo ThinClient that runs Home Assistant and Frigate, which runs seven cameras ranging from 1080p to 4K with motion and object detection. I track people, cars, motorcycles, dogs, packages, and bicycles. It runs and detects flawlessly ~11ms with only ~5.0% CPU and 0.8% Memory usage. It takes the load off the CPU to handle the other essential tasks.
J**A
Very Good as TPU for NVR
Used Google Coral as TPU for Frigate NVR with excelent results!
T**K
Frigate Detector
This cheap little device runs circles around my old GTX970 I had been using until then. I just have it plugged in to a USB C port and it does its job.
T**L
Overpriced, but works as intended
I knew the seller was charging more than MSRP, but due to sloppy issues, was willing to pay more to obtain the device. It arrived in original packaging, and works great
G**.
Muy buen producto
¡El Acelerador USB Coral de Google es realmente sorprendente! Lo he utilizado en conjunto con la aplicación FRIGATE en Home Assistant para optimizar la detección de personas, y los resultados han sido impresionantes.Desde el momento en que instalé el Acelerador USB Coral, noté una mejora significativa en el rendimiento de la detección de personas. La velocidad de procesamiento es asombrosa, lo que significa que la detección ocurre en tiempo real y sin retrasos molestos. Esto es esencial para la seguridad del hogar, ya que cada segundo cuenta en situaciones críticas.La facilidad de instalación también es digna de elogio. Google ha proporcionado una documentación clara y fácil de seguir, lo que permitió que incluso alguien sin experiencia técnica como yo pudiera configurarlo sin problemas. Además, la integración con la aplicación FRIGATE en Home Assistant fue perfecta. Las instrucciones proporcionadas eran precisas y detalladas, lo que facilitó la configuración y puesta en marcha del sistema.La precisión de detección es otro punto fuerte del Acelerador USB Coral. No solo reconoce personas de manera confiable, sino que también minimiza las falsas alarmas. Esto es crucial para evitar molestias innecesarias y garantizar que solo se generen alertas cuando realmente importa.Otra ventaja que no puedo pasar por alto es el tamaño compacto del dispositivo. Puedes ubicarlo discretamente en cualquier lugar sin ocupar demasiado espacio ni llamar la atención. Además, su eficiencia energética es impresionante, lo que significa que no notarás un aumento significativo en tu factura eléctrica.
B**.
Works perfect in my Home Assistant + Frigate + RPi4
Must have if you are using Frigate on Home Assistant. Allows you to drop the main CPU usage and process tenfold more video streams with object recognition. With this accelerator, I am able to process streams from my 5 HD cameras with object recognition, on RPI4 and reasonable CPU consumption. Works stable and doesn't cause any problems
B**N
As bad as you've heard
I bought this because it was available, knowing that it was less preferred over the m2 slot and other varieties (that are still backordered everywhere). Can confirm it's not great - disconnects all the time and was a pain to set up and keep going with CodeProject AI. I think the performance is adequate when it's working, but I can't handle the disconnects. It's plugged in to my security camera computer which outputs to my TV. I'm seldom viewing it through the TV which is a good thing - because I'd hear the USB connected/disconnected noise every 20 seconds.I know it's not overheating either, I've had a fan blowing ambient temp room air on it since installation. Tried all of my USB ports and a dozen USB cables. Shame it's so unreliable.
Trustpilot
4 days ago
3 days ago