Using google coral tpu on Home Assistance for frigate addon

Home assistance with frigate as an addon no longer supports Google Coral TPU due to the Python version above 3.9. Does anybody have a solution to it.
Hardware details:-
Rasspberry pi 5 8GB Ram
Google Coral TPU USB
O.S- Home Assistance version 14.2

Not sure what you mean by that. The addon uses its own python version.
Do you have a reference to that issue?

Ok so I am providing you with my details can you give me the solution to my problem.

lsusb
Bus 003 Device 001: ID 1d6b:0002 Linux 6.6.62-haos-raspi xhci-hcd xHCI Host Controller
Bus 002 Device 002: ID 1a6e:089a
Bus 001 Device 001: ID 1d6b:0002 Linux 6.6.62-haos-raspi xhci-hcd xHCI Host Controller
Bus 004 Device 001: ID 1d6b:0003 Linux 6.6.62-haos-raspi xhci-hcd xHCI Host Controller
Bus 002 Device 001: ID 1d6b:0003 Linux 6.6.62-haos-raspi xhci-hcd xHCI Host Controller

Frigate logs when I am starting the app

F driver/usb/usb_driver.cc:1148] HandleQueuedBulkIn transfer in failed. Not found: USB transfer error 5 [LibUsbDataInCallback]
2025-02-11 06:36:23.552803927 Fatal Python error: Aborted
Attempting to load TPU as usb
2025-02-11 07:01:28.091252110 [2025-02-11 12:31:28] frigate.detectors.plugins.edgetpu_tfl ERROR : No EdgeTPU was detected. If you do not have a Coral device yet, you must configure CPU detectors.

Hardware:-
rasspiberry pi 5 8GB .

I have same hardware and OS. I am not seeing any errors. Coral TPU is recognised and working

image
OS details you are currently using.

frigate.yaml file details

detectors:
coral:
type: edgetpu
device: usb

mqtt:
host: xxx.xxx.xxx.xxx
port: 1883
topic_prefix: frigate
client_id: frigate
user:xxxxxx
password: xxxxxxxxxx
stats_interval: 60

cameras:
#simple camera example
Camera_166:
ffmpeg:
inputs:
- path: rtsp://xxxxxxxxxxxxxxxxxxxxxxxxxx
roles:
- detect
- record

detect:
  width: 352
  height: 288
  fps: 5

record:
enabled: true
retain:
days: 7
mode: motion
alerts:
retain:
days: 14
detections:
retain:
days: 14
ffmpeg:
hwaccel_args: preset-rpi-64-h264

objects:
track:
- person
- dog
- cat
filters:
person:
# Optional: minimum widthheight of the bounding box for the detected object (default: 0)
min_area: 0
# Optional: maximum width
height of the bounding box for the detected object (default: 24000000)
max_area: 100000
# Optional: minimum score for the object to initiate tracking (default: shown below)
min_score: 0.5
# Optional: minimum decimal percentage for tracked object’s computed score to be considered a true positive (default: shown below)
threshold: 0.7
version: 0.15-1

However, I must mention that when I use CPU detectors it’s working.

detectors:
  coral:
    type: edgetpu
    device: usb

Not clear to me what your problem is exactly, but I would make backup copy of your yaml configuration and then start with bare bones configuration. Rebuild as there has been quite a few changes and I am on Frigate 0.15.0 (Full Access) Beta

I had the same issue =>

F driver/usb/usb_driver.cc:1148] HandleQueuedBulkIn transfer in failed. Not found: USB transfer error 5 [LibUsbDataInCallback]
2025-02-17 10:13:38.456967387 Fatal Python error: Aborted


[2025-02-17 11:13:39] frigate.watchdog INFO : Detection appears to have stopped. Exiting Frigate…

This happened 4 times today (I get a notification when Frigate isn’t running). I could however start the service with no problems every time.

I just updated from 14.1 to 15.0.0. No issues so far

Extra info: I’m using a Intel Nuc with Google Edge Tpu and Truenas Server for recordings