Trying to retrain TrafficCamNet - "tao model detectnet_v2 dataset_convert" raises an error

Accelerated Computing Intelligent Video Analytics TAO Toolkit
1

Please provide the following information when requesting support.

• Hardware: T4 (g4dn at AWS)
• Network Type: TrafficCamNet (detectnet_v2)
• TLT Version toolkit_version: 6.25.7

I am trying to retrain TrafficCamNet model with my own dataset using tao notebook (https://api.ngc.nvidia.com/v2/resources/nvidia/tao/tao-getting-started/versions/5.0.0/zip).When I launch the “pip3 install nvidia-tao“ command it installs Tao-toolkit 6.25.7.
Then when I launch the command “tao model detectnet_v2 dataset_convert“ it raises the error:

tao model: error: invalid choice: 'detectnet_v2' (choose from 'list', 'stop', 'info', 'dataset', 'deploy', 'model')

So the command dataset_convert seems to be deprecated in tao v6.
What is the supposed way to retrain TrafficCamNet in TAO v6?

3

Please use the older docker instead of latest tao-launcher.
You can trigger older docker via below ways. For example,
$ docker run --runtime=nvidia -it --rm nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5
$ docker run --runtime=nvidia -it --rm nvcr.io/nvidia/tao/tao-toolkit:4.0.1-tf1.15.5

All dockers are in TAO Toolkit | NVIDIA NGC (https://catalog.ngc.nvidia.com/orgs/nvidia/teams/tao/containers/tao-toolkit/tags).

4

Ok, thanks, I run the older container.
Still I don’t get how to launch tao commands inside the container. When I try to launch “tao model detectnet_v2 dataset_convert“ inside the container, it gives me the error:

bash: tao: command not found

If I explicitly install tao 5.0.0 by “pip3 install nvidia-tao==5.0.0“ inside the container and then launch “tao model detectnet_v2 dataset_convert“, it raises the following error:

root@2a0fab6776ec:/workspace/tao-experiments# tao model detectnet_v2 dataset_convert                   -d specs/detectnet_v2_tfrecords_kitti_trainval.txt                   -o data/tfrecords/kitti_trainval/kitti_trainval                   -r $USER_EXPERIMENT_DIR/
~/.tao_mounts.json wasn't found. Falling back to obtain mount points and docker configs from ~/.tao_mounts.json.
Please note that this will be deprecated going forward.
Traceback (most recent call last):
  File "/usr/local/bin/tao", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.8/dist-packages/nvidia_tao_cli/entrypoint/tao_launcher.py", line 134, in main
    instance.launch_command(
  File "/usr/local/lib/python3.8/dist-packages/nvidia_tao_cli/components/instance_handler/local_instance.py", line 356, in launch_command
    docker_logged_in(required_registry=task_map[task].docker_registry)
  File "/usr/local/lib/python3.8/dist-packages/nvidia_tao_cli/components/instance_handler/utils.py", line 151, in docker_logged_in
    data = load_config_file(docker_config)
  File "/usr/local/lib/python3.8/dist-packages/nvidia_tao_cli/components/instance_handler/utils.py", line 84, in load_config_file
    assert os.path.exists(config_path), (
AssertionError: Config path must be a valid unix path. No file found at: /root/.docker/config.json. Did you run docker login?
5

When you run with docker, it is not needed to add “tao” in the beginning.
You also need not install tao-launcher(“pip3 install nvidia-tao==5.0.0").

Please run
root@2a0fab6776ec:/workspace/tao-experiments# detectnet_v2 dataset_convert xxx

1 Like
6

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.