Internal error in Model Optimizer Converter

[rudakov__mikhail]

I'm trying to convert a Tensorflow model into IR. The model Mask RCNN Inception V2 and I didn't modify it. After running MO I receive the following output: 

[ ERROR ]  ----------------- INTERNAL ERROR ----------------

[ ERROR ]  Unexpected exception happened.

[ ERROR ]  Please contact Model Optimizer developers and forward the following information:

[ ERROR ]  0

[ ERROR ]  Traceback (most recent call last):

  File "C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\mo\main.py", line 321, in main

    return driver(argv)

  File "C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\mo\main.py", line 263, in driver

    mean_scale_values=mean_scale)

  File "C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\mo\pipeline\tf.py", line 171, in

 tf2nx

    class_registration.apply_replacements(graph, class_registration.ClassType.FRONT_REPLACER)

  File "C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\mo\utils\class_registration.py",

 line 102, in apply_replacements

    replacer.find_and_replace_pattern(graph)

  File "C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\mo\front\tf\replacement.py", lin

e 50, in find_and_replace_pattern

    self.transform_graph(graph, desc._replacement_desc['custom_attributes'])

  File "C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\extensions\front\tf\ObjectDetect

ionAPI.py", line 359, in transform_graph

    last_node = Node(graph, op_output).in_node(0)

  File "C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\mo\graph\graph.py", line 267, in

 in_node

    return self.in_nodes()[key]

KeyError: 0

[ ERROR ]  ---------------- END OF BUG REPORT --------------

[ ERROR ]  -------------------------------------------------

The command used is below:

python ./mo_tf.py --input_model=C:\downloads\mask_rcnn_inception_v2_coco_2018_01_28.tar\mask_rcnn_inception_v2_coco_2018_01_28\frozen_inference_graph.pb   --tensorflow_use_custom_operations_config C:\Intel\computer_vision_sdk_2018.3.343\deployment_tools\model_optimizer\extensions\front\tf\mask_rcnn_support.json  --output="num_detections,detection_masks,detection_boxes,detection_classes,detection_scores" --tensorflow_object_detection_api_pipeline_config C:\downloads\mask_rcnn_inception_v2_coco_2018_01_28.tar\mask_rcnn_inception_v2_coco_2018_01_28\pipeline.config

[Severine_H_Intel]

Dear Mikhail, 

it is the argument --output="num_detections,detection_masks,detection_boxes,detection_classes,detection_scores" which breaks the conversion, withdraw it and it will work. 

Best, 

Severine