Tensorrt model is not getting generated in NVIDIA Jetson AGX Orin

[bm-srikanth]

The tensorrt model is not getting generated in Nvidia Jetson Orin.

2025-02-26 18:47:54.304226100  Process detector:tensorrt:
2025-02-26 18:47:54.304244116  Traceback (most recent call last):
2025-02-26 18:47:54.304246356    File "/usr/lib/python3.9/multiprocessing/process.py", line 315, in _bootstrap
2025-02-26 18:47:54.304264084      self.run()
2025-02-26 18:47:54.304267124    File "/opt/frigate/frigate/util/process.py", line 41, in run_wrapper
2025-02-26 18:47:54.304269300      return run(*args, **kwargs)
2025-02-26 18:47:54.304311828    File "/usr/lib/python3.9/multiprocessing/process.py", line 108, in run
2025-02-26 18:47:54.304313460      self._target(*self._args, **self._kwargs)
2025-02-26 18:47:54.304316308    File "/opt/frigate/frigate/object_detection.py", line 121, in run_detector
2025-02-26 18:47:54.304317972      object_detector = LocalObjectDetector(detector_config=detector_config)
2025-02-26 18:47:54.304319540    File "/opt/frigate/frigate/object_detection.py", line 68, in __init__
2025-02-26 18:47:54.304320916      self.detect_api = create_detector(detector_config)
2025-02-26 18:47:54.304349844    File "/opt/frigate/frigate/detectors/__init__.py", line 18, in create_detector
2025-02-26 18:47:54.304351188      return api(detector_config)
2025-02-26 18:47:54.304365620    File "/opt/frigate/frigate/detectors/plugins/tensorrt.py", line 243, in __init__
2025-02-26 18:47:54.304367156      self.engine = self._load_engine(detector_config.model.path)
2025-02-26 18:47:54.304368916    File "/opt/frigate/frigate/detectors/plugins/tensorrt.py", line 87, in _load_engine
2025-02-26 18:47:54.304370452      with open(model_path, "rb") as f, trt.Runtime(self.trt_logger) as runtime:
2025-02-26 18:47:54.304372244  FileNotFoundError: [Errno 2] No such file or directory: '/config/model_cache/yolov7-320-dla.trt'
2025-02-26 18:47:54.304373780  Exception ignored in: <function TensorRtDetector.__del__ at 0xffff8765eee0>
2025-02-26 18:47:54.304387700  Traceback (most recent call last):
2025-02-26 18:47:54.304390516    File "/opt/frigate/frigate/detectors/plugins/tensorrt.py", line 262, in __del__
2025-02-26 18:47:54.304392212      if self.outputs is not None:
2025-02-26 18:47:54.304394740  AttributeError: 'TensorRtDetector' object has no attribute 'outputs'

The docker-compose file is as follows:

version: "2.4"
services:
  frigate:
    environment:
      - YOLO_MODELS=yolov7-320-dla,yolov3-288-dla
      - USE_FP16=false  
    container_name: frigate
    privileged: true # this may not be necessary for all setups
    restart: unless-stopped
    image: ghcr.io/blakeblackshear/frigate:stable-tensorrt-jp5
    runtime: nvidia
    shm_size: "900mb" # update for your cameras based on calculation above
    group_add:
      - "109" # render
      - "110" # render
      - "44"  # video
      - "46"  # plugdev
    volumes:
      - /etc/localtime:/etc/localtime:ro
      - ./config:/config
      - ./storage:/media/frigate
      - type: tmpfs # Optional: 1GB of memory, reduces SSD/SD Card wear
        target: /tmp/cache
        tmpfs:
          size: 1000000000

    ports:
      - "8971:8971"
      - "5000:5000" # Internal unauthenticated access. Expose carefully.
      - "8554:8554" # RTSP feeds
      - "8555:8555/tcp" # WebRTC over tcp
      - "8555:8555/udp" # WebRTC over udp
    environment:
      FRIGATE_RTSP_PASSWORD: "password"

The config file is as follows:

mqtt:
  enabled: false

cameras:
  camera1:  #<------ Name the camera
    enabled: true
    ffmpeg:
      inputs:
        - path: rtsp://x.x.x.x:x/live.sdp  # <----- The stream you want to use for detection
          roles:
            - detect
    detect:
      enabled: true
    objects:
      track:
        - person
        - car

detectors:
  tensorrt:
    type: tensorrt
    device: 0

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

snapshots:
  enabled: false
  clean_copy: true #false  #true
  timestamp: false
  bounding_box: true
  crop: false
  required_zones: []
  retain:
    default: 5
    objects:
      person: 7
      car: 7
  quality: 100

auth:
  reset_admin_password: true
  
version: 0.16.0

System information:
ARMv8 Processor rev 1 (v8l) × 8, 32GB RAM, 64GB disk capacity,
Ubuntu 20.04.6 LTS, 5.10.104-tegra-0030A1AIM35UIV00002

Any help will be greatly appreciated..

[hawkeye217]

See the docs and ensure your model path is correct in your config: https://docs.frigate.video/configuration/object_detectors/#generate-models

[bm-srikanth]

Yes, the model path is correct in the config. I have checked inside the container and there are no model files generated.. Is there a log file where I can see why the model files are not getting generated inside the container?

[NickM-27]

The generation logs are part of docker logs