[Support]: After upgrade from 16.0 to 16.1, no longer detects TPU Google Coral

[wingmanx]

Checklist

• I have updated to the latest available Frigate version.
• I have cleared the cache of my browser.
• I have tried a different browser to see if it is related to my browser.
• I have tried reproducing the issue in incognito mode to rule out problems with any third party extensions or plugins I have installed.
• I have asked the AI at https://docs.frigate.video about my issue.

Describe the problem you are having

After upgrade from 16.0 to 16.1, no longer detects TPU Google Coral

Steps to reproduce

1. Working flawless with Home Assistant Add-On 16.0
2. Upgraded to 16.1
3. Logs show error loading TPU
4. Restored 16.0
5. All working again

Version

0.16.1 (Full Access)

In which browser(s) are you experiencing the issue with?

No response

Frigate config file

mqtt:
  host: 192.168.0.100
  user: user
  password: ha

detectors:
  coral:
    type: edgetpu
    device: usb

cameras:
  cam1:
    ffmpeg:
      inputs:
        - path: rtsp://user:[email protected]:554/stream2
          roles:
            - detect
    detect:
      width: 640
      height: 360
      fps: 5
    objects:
      track:
        - person


  cam2:
    ffmpeg:
      inputs:
        - path: rtsp://user:[email protected]:554/Preview_01_sub
          roles:
            - detect
    detect:
      width: 640  # Adjust to match your sub-stream resolution
      height: 480
      fps: 10


  cam3:
    ffmpeg:
      inputs:
        - path: rtsp://user:[email protected]:554/stream2
          roles:
            - detect
    detect:
      width: 640
      height: 360
      fps: 5

docker-compose file or Docker CLI command

I have Full HA OS

Relevant Frigate log output

frigate.detectors.plugins.edgetpu_tfl ERROR   : No EdgeTPU was detected. If you do not have a Coral device yet, you must configure CPU detectors.
2025-10-06 12:46:54.975395049  Process detector:coral:
2025-10-06 12:46:54.975398409  Traceback (most recent call last):
2025-10-06 12:46:54.975399997    File "/usr/local/lib/python3.11/dist-packages/tflite_runtime/interpreter.py", line 166, in load_delegate
2025-10-06 12:46:54.975404346      delegate = Delegate(library, options)
2025-10-06 12:46:54.975406111                 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-10-06 12:46:54.975407789    File "/usr/local/lib/python3.11/dist-packages/tflite_runtime/interpreter.py", line 104, in __init__
2025-10-06 12:46:54.975424238      raise ValueError(capture.message)
2025-10-06 12:46:54.975425176  ValueError
2025-10-06 12:46:54.975426496  
2025-10-06 12:46:54.975427598  During handling of the above exception, another exception occurred:
2025-10-06 12:46:54.975428319  
2025-10-06 12:46:54.975429174  Traceback (most recent call last):
2025-10-06 12:46:54.975430336    File "/usr/lib/python3.11/multiprocessing/process.py", line 314, in _bootstrap
2025-10-06 12:46:54.975451179      self.run()
2025-10-06 12:46:54.975452500    File "/opt/frigate/frigate/util/process.py", line 41, in run_wrapper
2025-10-06 12:46:54.975453468      return run(*args, **kwargs)
2025-10-06 12:46:54.975454309             ^^^^^^^^^^^^^^^^^^^^
2025-10-06 12:46:54.975455416    File "/usr/lib/python3.11/multiprocessing/process.py", line 108, in run
2025-10-06 12:46:54.975474285      self._target(*self._args, **self._kwargs)
2025-10-06 12:46:54.975475684    File "/opt/frigate/frigate/object_detection/base.py", line 112, in run_detector
2025-10-06 12:46:54.975477028      object_detector = LocalObjectDetector(detector_config=detector_config)
2025-10-06 12:46:54.975478130                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-10-06 12:46:54.975479512    File "/opt/frigate/frigate/object_detection/base.py", line 57, in __init__
2025-10-06 12:46:54.975480535      self.detect_api = create_detector(detector_config)
2025-10-06 12:46:54.975481475                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-10-06 12:46:54.975492474    File "/opt/frigate/frigate/detectors/__init__.py", line 18, in create_detector
2025-10-06 12:46:54.975493567      return api(detector_config)
2025-10-06 12:46:54.975494444             ^^^^^^^^^^^^^^^^^^^^
2025-10-06 12:46:54.975495639    File "/opt/frigate/frigate/detectors/plugins/edgetpu_tfl.py", line 42, in __init__
2025-10-06 12:46:54.975497311      edge_tpu_delegate = load_delegate("libedgetpu.so.1.0", device_config)
2025-10-06 12:46:54.975498457                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-10-06 12:46:54.975499789    File "/usr/local/lib/python3.11/dist-packages/tflite_runtime/interpreter.py", line 168, in load_delegate
2025-10-06 12:46:54.975514974      raise ValueError('Failed to load delegate from {}\n{}'.format(
2025-10-06 12:46:54.975516117  ValueError: Failed to load delegate from libedgetpu.so.1.0
2025-10-06 12:46:54.975516790  
2025-10-06 12:47:00.530855321  [2025-10-06 13:47:00] frigate.watchdog               INFO    : Detection appears to have stopped. Exiting Frigate...

Relevant go2rtc log output

No relevant output

Operating system

Home Assistant OS

Install method

Home Assistant Add-on

Network connection

Wired

Camera make and model

Tapo C51A

Screenshots of the Frigate UI's System metrics pages

Any other information that may be helpful

No response

[inkeep-ai-assistant[bot]]

Based on the error message and your configuration, this is a known issue with the Google Coral EdgeTPU not being detected after the upgrade. Here are several troubleshooting steps to resolve this:

Common Solutions

1. Check Protection Mode (Home Assistant Add-on)

If you're using the Home Assistant add-on, try disabling protection mode (1). This has resolved the issue for many users experiencing Coral detection problems after upgrades.

2. Verify Device Configuration

Your current configuration looks correct for a USB Coral (2):

detectors:
  coral:
    type: edgetpu
    device: usb

(2)

3. Check for Conflicting Services

Ensure no other services are using the Coral device (3). On QNAP NAS devices, QuMagie with AI Core extension can conflict with Frigate's Coral access (3). You may need to disable AI recognition features in QuMagie or remove the QNAP AI Core extension (3).

4. Try Alternative Device Specifications

If the basic device: usb doesn't work, try specifying the device index (2):

detectors:
  coral:
    type: edgetpu
    device: usb:0

(2)

5. Cable and Hardware Issues

The USB Coral can become stuck and need to be restarted (3). Some users have found that switching to a different USB cable fixed the problem entirely (3). Running Frigate in a VM may also cause communication with the device to be lost and need to be reset (3).

6. Restart Your System

For some NAS systems, a complete system restart while leaving the TPU device plugged in can resolve detection issues (3).

Since you mentioned that downgrading to version 16.0 resolves the issue, this suggests a compatibility problem introduced in version 16.1. The troubleshooting steps above should help resolve the detection problem without needing to downgrade.

📚 Sources:

1. [Detector Support]: Google Coral USB accelerator not detected on Home Assistant OS 2023.8.2
2. Supported Hardware
3. Troubleshooting EdgeTPU

---

Have another question? Just tag @inkeep.

[hawkeye217]

This is not a Frigate bug, and the AI's response is correct.

See the 0.16 FAQ, Upgrade issues question 3.

[wingmanx]

Solved, the protection was enabled.