Google CoralUSB Edge TPU ML Accelerator coprocessor for Raspberry Pi and Other Embedded Single Board Computers

Simple plug and play, got recognised immediately and definitely speeds up the AI stuff on QNAP, primarily the TS262 as the TS433 has a Neural Processing Unit built in, but i tested it anyway and it worked.I am now tempted to buy a couple more.

BE AWARE, this is not the easiest item to setup, I am still attempting with the help of Coral tech support, not quite there yet though, Once done you have many uses, I'm focusing on running cpu intensive software inside containers that will use the EDGE processor for the real chewy stuff. There is also the AI side but TBH with ChatGPT online I'd rather use that. So a lovely bit of kit, looks cool, nice packaging BUT IT IS NOT PLUG AND PLAY. Oh yeah the tech support is Chinese which can be challenging!!!

Had heard various positive testimonials about the Coral so pulled the trigger to use with Frigate on Home Assistant. Had one camera running on the host CPU before switching to Coral and noticing the sudden drop in processor usage. I now have eight cameras running on Frigate, four of which are using the TPU to detect objects.Runs perfectly, not a single issue. Was detected and adopted straight away with Frigate. Plug and play. You only have to reference coral in your YAML once, no further steps necessary.

Like 99% of other reviewers, I used the Coral TPU USB with Frigate to offload object inference from the CPU. This it does very well. Amazing that such a small, low cost device can do this but it goes to show how purpose built hardware can be remarkably efficicient at a specific task.I only have a couple of cameras at the moment and the device does not even get warm. Inference speed is slightly disappointing (30ms), but I put this down to the older PC it is running on. EDIT: Switching from USB2 to USB3 port brought inference speed to 8.5ms)Now for the negatives. The device changes USB ID once initialised. This can make virtualisation more difficult or less secure. There seem to be very few applications that make drop-in use of these coral devices. The device is sold as a devloper board, so possibly risks becoming another Google abandon-ware project.Still, it was quite an eye-opener to see what this little device can achieve when compared to the cpu grunt required to do the same. Currently at 65 quid, it's a far more economical proposition than it was a year ago.

For anyone running Frigate on their NAS - this is your missing piece. My Synology DS423+ was struggling like a hamster on a wheel trying to handle person detection, until this compact powerhouse entered the chat. Now it spots humans with scary accuracy while barely breaking a sweat.Initial setup took some tinkering (as all good things do), but the performance difference is night and day. CPU usage plummeted from "small datacenter" to "practically idle" on my NAS.

Setting this up to use for my CCTV setup in Linux was rather a pain, having to mess about with old versions of Python and such. The instructions were far out of date and had to go hunting round for guides and files.That being said it is good at its job and takes a lot of the load off the laptop I use an NVR.

Using on an Unraid setup with Home Assistant and Frigate installed. Perfect for taking the recognition work off the CPU. Initially wasn't recognised when I set it up, but this was because my Unraid drivers were a couple of years out of date. Updated to latest version and immediately started working. Would recommend.

I got this for my Frigate NVR server.Before adding it, I had to disable all motion detection.After adding it, all 5 of my cameras are detection enabled and the CPU usage is down at a nice stable level, beautiful.