Error while traininig detectnet_v2 with mobilenet_v2 backbone

Hi,

I’m trying to train a detectnet_v2 object detector with the mobilenet_v2 backbone.
Once the training process has started I get the following error:

Traceback (most recent call last):
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/scripts/train.py", line 797, in <module>
File "<decorator-gen-2>", line 2, in main
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/utilities/timer.py", line 46, in wrapped_fn
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/scripts/train.py", line 790, in main
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/scripts/train.py", line 691, in run_experiment
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/scripts/train.py", line 576, in train_gridbox
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/scripts/train.py", line 387, in build_gridbox_model
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/model/detectnet_model.py", line 106, in construct_model
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/model/detectnet_model.py", line 235, in _construct_feature_extractor
IndexError: tuple index out of range
Traceback (most recent call last):
File "/usr/local/bin/detectnet_v2", line 8, in <module>
sys.exit(main())
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/detectnet_v2/entrypoint/detectnet_v2.py", line 12, in main
File "/home/vpraveen/.cache/dazel/_dazel_vpraveen/216c8b41e526c3295d3b802489ac2034/execroot/ai_infra/bazel-out/k8-fastbuild/bin/magnet/packages/iva/build_wheel.runfiles/ai_infra/iva/common/entrypoint/entrypoint.py", line 296, in launch_job
AssertionError: Process run failed.
2021-04-20 20:57:57,443 [INFO] tlt.components.docker_handler.docker_handler: Stopping container.

I can’t seem to reproduce the issue with any other backbone, even though they’re using roughly the same spec file (aside from the “arch” and “pretrained_model_file” options). The spec file in this case is as follows:

random_seed: 42
dataset_config {
  data_sources {
    tfrecords_path: "/workspace/tlt-experiments/tfrecords/*"
    image_directory_path: "/workspace/tlt-experiments/data/train/"
  }
  image_extension: "png"
  target_class_mapping {
    key: "pedestrian"
    value: "pedestrian"
  }
  target_class_mapping {
    key: "person"
    value: "pedestrian"
  }
  target_class_mapping {
    key: "person_sitting"
    value: "pedestrian"
  }
  validation_fold: 0
}
augmentation_config {
  preprocessing {
    output_image_width: 640
    output_image_height: 480
    min_bbox_width: 1.0
    min_bbox_height: 1.0
    output_image_channel: 3
  }
  spatial_augmentation {
    hflip_probability: 0.5
    zoom_min: 1.0
    zoom_max: 1.0
    translate_max_x: 8.0
    translate_max_y: 8.0
  }
  color_augmentation {
    hue_rotation_max: 25.0
    saturation_shift_max: 0.20000000298
    contrast_scale_max: 0.10000000149
    contrast_center: 0.5
  }
}
postprocessing_config {
  target_class_config {
    key: "car"
    value {
      clustering_config {
        clustering_algorithm: DBSCAN
        dbscan_confidence_threshold: 0.9
        coverage_threshold: 0.00499999988824
        dbscan_eps: 0.20000000298
        dbscan_min_samples: 0.0500000007451
        minimum_bounding_box_height: 20
      }
    }
  }
  target_class_config {
    key: "cyclist"
    value {
      clustering_config {
        clustering_algorithm: DBSCAN
        dbscan_confidence_threshold: 0.9
        coverage_threshold: 0.00499999988824
        dbscan_eps: 0.15000000596
        dbscan_min_samples: 0.0500000007451
        minimum_bounding_box_height: 20
      }
    }
  }
  target_class_config {
    key: "pedestrian"
    value {
      clustering_config {
        clustering_algorithm: DBSCAN
        dbscan_confidence_threshold: 0.9
        coverage_threshold: 0.00749999983236
        dbscan_eps: 0.230000004172
        dbscan_min_samples: 0.0500000007451
        minimum_bounding_box_height: 20
      }
    }
  }
}
model_config {
  pretrained_model_file: "/workspace/tlt-experiments/mobilenet_v2.hdf5"
  use_batch_norm: true
  objective_set {
    bbox {
      scale: 35.0
      offset: 0.5
    }
    cov {
    }
  }
  training_precision {
    backend_floatx: FLOAT32
  }
  arch: "mobilenet_v2"
}
evaluation_config {
  validation_period_during_training: 5
  first_validation_epoch: 30
  minimum_detection_ground_truth_overlap {
    key: "car"
    value: 0.699999988079
  }
  minimum_detection_ground_truth_overlap {
    key: "cyclist"
    value: 0.5
  }
  minimum_detection_ground_truth_overlap {
    key: "pedestrian"
    value: 0.5
  }
  evaluation_box_config {
    key: "car"
    value {
      minimum_height: 10
      maximum_height: 9999
      minimum_width: 10
      maximum_width: 9999
    }
  }
  evaluation_box_config {
    key: "cyclist"
    value {
      minimum_height: 10
      maximum_height: 9999
      minimum_width: 10
      maximum_width: 9999
    }
  }
  evaluation_box_config {
    key: "pedestrian"
    value {
      minimum_height: 10
      maximum_height: 9999
      minimum_width: 10
      maximum_width: 9999
    }
  }
  average_precision_mode: SAMPLE
}
cost_function_config {
  target_classes {
    name: "pedestrian"
    class_weight: 4.0
    coverage_foreground_weight: 0.0500000007451
    objectives {
      name: "cov"
      initial_weight: 1.0
      weight_target: 1.0
    }
    objectives {
      name: "bbox"
      initial_weight: 10.0
      weight_target: 10.0
    }
  }
  enable_autoweighting: true
  max_objective_weight: 0.999899983406
  min_objective_weight: 9.99999974738e-05
}
training_config {
  batch_size_per_gpu: 4
  num_epochs: 40
  learning_rate {
    soft_start_annealing_schedule {
      min_learning_rate: 5e-06
      max_learning_rate: 5e-04
      soft_start: 0.10000000149
      annealing: 0.699999988079
    }
  }
  regularizer {
    type: L1
    weight: 3.00000002618e-09
  }
  optimizer {
    adam {
      epsilon: 9.99999993923e-09
      beta1: 0.899999976158
      beta2: 0.999000012875
    }
  }
  cost_scaling {
    initial_exponent: 20.0
    increment: 0.005
    decrement: 1.0
  }
  checkpoint_interval: 5
}
bbox_rasterizer_config {
  target_class_config {
    key: "car"
    value {
      cov_center_x: 0.5
      cov_center_y: 0.5
      cov_radius_x: 0.40000000596
      cov_radius_y: 0.40000000596
      bbox_min_radius: 1.0
    }
  }
  target_class_config {
    key: "cyclist"
    value {
      cov_center_x: 0.5
      cov_center_y: 0.5
      cov_radius_x: 1.0
      cov_radius_y: 1.0
      bbox_min_radius: 1.0
    }
  }
  target_class_config {
    key: "pedestrian"
    value {
      cov_center_x: 0.5
      cov_center_y: 0.5
      cov_radius_x: 1.0
      cov_radius_y: 1.0
      bbox_min_radius: 1.0
    }
  }
  deadzone_radius: 0.400000154972
}

Is there something wrong with the spec file, or I should provide more information about the setup?

Thanks for helping out.

Edit:
The tlt command I’m using for launching this task is as follows:

tlt detectnet_v2 train -e /workspace/tlt-experiments/detectnet_v2/specs/detectnet_v2_train_mobilenet_v2_kitti.txt -r /workspace/tlt-experiments/unpruned -k tlt_encode --gpus 1

Sorry, this is an issue in TLT 3.0_dp. The fix will be available in next release.

Thanks for the information, is there something that I can do on my end to temporarily work around the issue?

Sorry, there is not a simple workaround yet.

1 Like