YOLO v8 issue with Coral TPU

mailseth

Getting the hang of it
Joined
Dec 22, 2023
Messages
126
Reaction score
88
Location
California

jjj111

n3wb
Joined
Feb 1, 2024
Messages
3
Reaction score
3
Location
Bay Area
@mailseth yeah I'd seen those, but the problem is that it seems like there's a bunch of people trying to figure out how to load custom models in codeproject.ai for using those -- there's even a post on their forum from an hour ago where a team member says "Unfortunately, custom models for Coral are not currently supported."

Would you mind sharing how you were able to accomplish using that pre-release model?
 

mailseth

Getting the hang of it
Joined
Dec 22, 2023
Messages
126
Reaction score
88
Location
California
I'm actually running these tests with the objectdetection_coral_mulitpu.py file on the CLI, independent of most of the rest of CPAI. It's just calling the tpu_runner.py file for my local testing purposes. (I don't even have the CPAI server running on my machine right now.)

So I'm running the pre-release model simply by specifying it as an option to objectdetection_coral_mulitpu.py, but unfortunately it's not going to be a very useful method for anyone else here.
 

Pentagano

Getting comfortable
Joined
Dec 11, 2020
Messages
612
Reaction score
282
Location
Uruguay

mailseth

Getting the hang of it
Joined
Dec 22, 2023
Messages
126
Reaction score
88
Location
California
Possibly. I found it has a CPU-only instruction midway through the model so it doesn’t run as well as you’d think it should. Everything after that instruction ends up also being CPU-only. (Unless you’re running multi-TPUs, because then you can run the remaining instructions on the next TPU segment.)
 
Top