I am running this in a docker container on Unraid and it does work for a bit of time so I think it is installed and detecting the coral TPU correctly.
In the logs it just says:
EDIT: Switched to YoloV8 model and disabled multitpu and it seems to not be using my processor anymore and has some decent speeds but why is it says CPU (tf-lite) instead of TPU? Sometimes it says it at the beginning before it processes any requests then switched immediately
EDIT 2: eventually even in that CPU (tf-lite) mode it loses contact
In the logs it just says:
Code:
10:12:34:objectdetection_coral_adapter.py: WARNING:root:Queue stalled; refreshing interpreters.
10:12:48:objectdetection_coral_adapter.py: WARNING:root:Pipe thread didn't join!EDIT: Switched to YoloV8 model and disabled multitpu and it seems to not be using my processor anymore and has some decent speeds but why is it says CPU (tf-lite) instead of TPU? Sometimes it says it at the beginning before it processes any requests then switched immediately
EDIT 2: eventually even in that CPU (tf-lite) mode it loses contact
Last edited: