Early demo: Metis M.2 running YOLOv8 on a Raspberry Pi 5

Wanted to share a quick demo we’ve been playing around with — we’ve got the Metis M.2 up and running on a Raspberry Pi 5, doing real-time vehicle detection using an unmodified YOLOv8 model.

Pretty smooth experience, getting them working together. Definitely still early days, so no promises yet on timelines or any kind of future support, but figured it was worth showing that it can be done!

Curious — would people here be interested in standalone edge AI running on something like a Pi5? Any particular use cases that come to mind?

Page 1 / 1

This is really exciting, thanks @Victor Labian! And great work!

I’ve already got four Raspberry Pis running in my house - need to add a fifth one running inference on my CCTV or something!

Pretty cool stuff!

I just got my Metis M.2 and I was hoping to connect it to my Raspberry Pi 5 at home as well — what was your set-up like? Did you plug it into a Raspberry Pi M.2 HAT+ adapter board? I was already pleasantly surprised to see in the Metis M.2 docs that it's been confirmed to work with Raspberry Pi 5, I was prepared for a bit of adventure but it sounds like this may be pretty straightforward!

@Victor Labian knows a lot more about the set up than me, so I’ll let him provide a bit of guidance there, but you’re right that it was tested using the official RPi M.2 hat.

I might be wrong here, but Victor might also have been testing out a different model of hat as well?

Reply

Sign up

Log in, or create an Axelera AI account

Login to the community

Log in, or create an Axelera AI account

Scanning file for viruses.

This file cannot be downloaded