[Detector Support]: installing detector on truenas

[dcktp37]

Describe the problem you are having

I’ve been trying to figure out how to get detectors up and running with my nvidia 1050ti. I can’t figure out how to download the models using the scripts in the docs. TrueNAS doesn’t recognize the docker commands.
Please bear with me. This is a bit outside my experience

Version

0.16.1

Frigate config file

Na

docker-compose file or Docker CLI command

Na

Relevant Frigate log output

2025-10-15 12:52:22.118971000  [INFO] Preparing Frigate...
2025-10-15 12:52:22.502772931  [INFO] Starting Frigate...
2025-10-15 12:52:24.459080597  [2025-10-15 08:52:24] frigate.util.config            INFO    : Checking if frigate config needs migration...
2025-10-15 12:52:24.478048934  [2025-10-15 08:52:24] frigate.util.config            INFO    : frigate config does not need migration...
2025-10-15 12:52:24.507825979  [2025-10-15 08:52:24] frigate.app                    INFO    : Starting Frigate (0.16.1-e664cb2)
2025-10-15 12:52:24.523774594  [2025-10-15 08:52:24] peewee_migrate.logs            INFO    : Starting migrations
2025-10-15 12:52:24.524734197  [2025-10-15 08:52:24] peewee_migrate.logs            INFO    : There is nothing to migrate
2025-10-15 12:52:24.537870840  [2025-10-15 08:52:24] frigate.app                    INFO    : Recording process started: 394
2025-10-15 12:52:24.538479448  [2025-10-15 08:52:24] frigate.app                    INFO    : Review process started: 401
2025-10-15 12:52:24.540567256  [2025-10-15 08:52:24] frigate.app                    INFO    : go2rtc process pid: 124
2025-10-15 12:52:24.552726303  [2025-10-15 08:52:24] detector.onnx                  INFO    : Starting detection process: 418
2025-10-15 12:52:24.554239169  [2025-10-15 08:52:24] frigate.detectors.plugins.onnx INFO    : ONNX: loaded onnxruntime module
2025-10-15 12:52:24.554601053  [2025-10-15 08:52:24] frigate.detectors.plugins.onnx INFO    : ONNX: loading /mnt/Frigate_Rec/Frigate/model_cache/yolo_nas_s.onnx
2025-10-15 12:52:24.557447954  Process detector:onnx:
2025-10-15 12:52:24.557450944  Traceback (most recent call last):
2025-10-15 12:52:24.557452354    File "/usr/lib/python3.11/multiprocessing/process.py", line 314, in _bootstrap
2025-10-15 12:52:24.557455361      self.run()
2025-10-15 12:52:24.557457177    File "/opt/frigate/frigate/util/process.py", line 41, in run_wrapper
2025-10-15 12:52:24.557458310      return run(*args, **kwargs)
2025-10-15 12:52:24.557471043             ^^^^^^^^^^^^^^^^^^^^
2025-10-15 12:52:24.557472389    File "/usr/lib/python3.11/multiprocessing/process.py", line 108, in run
2025-10-15 12:52:24.557473738      self._target(*self._args, **self._kwargs)
2025-10-15 12:52:24.557474945    File "/opt/frigate/frigate/object_detection/base.py", line 112, in run_detector
2025-10-15 12:52:24.557476177      object_detector = LocalObjectDetector(detector_config=detector_config)
2025-10-15 12:52:24.557477384                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-10-15 12:52:24.557491816    File "/opt/frigate/frigate/object_detection/base.py", line 57, in __init__
2025-10-15 12:52:24.557492976      self.detect_api = create_detector(detector_config)
2025-10-15 12:52:24.557507962                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-10-15 12:52:24.557509289    File "/opt/frigate/frigate/detectors/__init__.py", line 18, in create_detector
2025-10-15 12:52:24.557510338      return api(detector_config)
2025-10-15 12:52:24.557511289             ^^^^^^^^^^^^^^^^^^^^
2025-10-15 12:52:24.557512479    File "/opt/frigate/frigate/detectors/plugins/onnx.py", line 53, in __init__
2025-10-15 12:52:24.557513504      self.model = ort.InferenceSession(
2025-10-15 12:52:24.557514467                   ^^^^^^^^^^^^^^^^^^^^^
2025-10-15 12:52:24.557515905    File "/usr/local/lib/python3.11/dist-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 465, in __init__
2025-10-15 12:52:24.557525452      self._create_inference_session(providers, provider_options, disabled_optimizers)
2025-10-15 12:52:24.557527365    File "/usr/local/lib/python3.11/dist-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 526, in _create_inference_session
2025-10-15 12:52:24.557528767      sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model)
2025-10-15 12:52:24.557530116             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025-10-15 12:52:24.557532610  onnxruntime.capi.onnxruntime_pybind11_state.NoSuchFile: [ONNXRuntimeError] : 3 : NO_SUCHFILE : Load model from /mnt/Frigate_Rec/Frigate/model_cache/yolo_nas_s.onnx failed:Load model /mnt/Frigate_Rec/Frigate/model_cache/yolo_nas_s.onnx failed. File doesn't exist

Install method

Docker CLI

Object Detector

TensorRT

Screenshots of the Frigate UI's System metrics pages

Na

Any other information that may be helpful

No response

[inkeep-ai-assistant[bot]]

For TensorRT detectors on TrueNAS with your Nvidia 1050ti, you need to use the -tensorrt Frigate image and set the YOLO_MODELS environment variable to automatically download and generate the required models (1).

The TensorRT detector requires preprocessed model files that are generated on the same hardware platform where they'll run (1). Frigate will automatically generate these models during startup if they're not found, storing them in /config/model_cache (1).

To configure this in TrueNAS:

1. Use the TensorRT image: ghcr.io/blakeblackshear/frigate:stable-tensorrt (1)
2. Add the environment variable YOLO_MODELS with your desired models (e.g., yolov7-320,yolov4-416) (1)
3. Ensure GPU passthrough is configured with --gpus=all (2)

Your detector configuration should look like this (1):

detectors:
  tensorrt:
    type: tensorrt
    device: 0

model:
  path: /config/model_cache/tensorrt/yolov7-320.trt
  input_tensor: nchw
  input_pixel_format: rgb
  width: 320
  height: 320

(1)

The available models include yolov7-320, yolov7-416, yolov4-416, and others (1). Models will only be generated if the corresponding .trt file doesn't exist in the model cache folder (1).

A community user successfully resolved a similar TrueNAS setup by adding the environment variable YOLO_MODELS=yolov7-320 in their TrueNAS configuration interface (3).

📚 Sources:

1. Supported Hardware
2. Video Decoding
3. [Detector Support]: TensorRT Crashing [Truenas SCALE] Nvidia GTX 1060

---

Have another question? Just tag @inkeep.

[dcktp37]

detectors:
   onnx:
    type: onnx

model:
  model_type: yolonas
  width: 320 
  height: 320 
  input_pixel_format: bgr
  input_tensor: nchw
  path: /config/model_cache/yolo_nas_s.onnx
  labelmap_path: /labelmap/coco-80.txt

File moved from model_config to Frigate.

log:

2025-10-15 14:25:04.826530623 onnxruntime.capi.onnxruntime_pybind11_state.NoSuchFile: [ONNXRuntimeError] : 3 : NO_SUCHFILE : Load model from /config/model_cache/yolo_nas_s.onnx failed:Load model /config/model_cache/yolo_nas_s.onnx failed. File doesn't exist

[hawkeye217]

I'm not familiar with TrueNAS, but if you've generated the model file according to the documentation and placed it in the correct directory, Frigate should find it without issue. The log message indicates you still have path/volume related issues.

[dcktp37]

Okay, after completely redoing the trunas pools and datasets, I believe I have it running and may have found the original issue. When I set up the datasets on trunas I capitalized them. So config was Config. I'm also not sure that I had the host location and mount locations the same. That could have been causing me issues as well.

Edit:
I spoke too soon. It's still not seeing the file. I'm not sure why.

Error:
"onnxruntime.capi.onnxruntime_pybind11_state.NoSuchFile: [ONNXRuntimeError] : 3 : NO_SUCHFILE : Load model from Frigate/Config/yolo_nas_s.onnx failed:Load model Frigate/Config/yolo_nas_s.onnx failed. File doesn't exist"

As far as I can tell, that's the correct path.
I have tried mnt/Frigate/Config/yolo_nas_s.onnx, Frigate/Config/yolo_nas_s.onnx, Config/yolo_nas_s.onnx.
I also tried copying the file into model_cache and trying Frigate/Config/model_cache/yolo_nas_s.onnx

[dcktp37]

So, I had to add “additional storage” in the Frigate app and mount the exact file location, not just the Config folder. Once I did that, it sees and uses the model. Thanks for your help!!

[hawkeye217]

Glad it's working for you now.