TensorRT: Cannot converted the mask-rcnn model to uff in the sampleUffMaskRCNN sample

Description

I tried to convert the mask-rcnn model to uff at TensorRT-7.0.0.11/samples/sampleUffMaskRCNN/converted.

My command is

python3 mrcnn_to_trt_single.py -w ../../../data/mask_rcnn_coco.h5 -o ../../../data/mask_rcnn_coco.uff -p ./config.py 

First, I met the error about setting a channel dimension.

raise ValueError('The channel dimension of the inputs '
ValueError: The channel dimension of the inputs should be defined. Found `None`.

I already solved this problem.

However, when I tried it again, I met a new error message below.

The output names of tensorflow graph nodes: ['mrcnn_mask/Reshape_1']
WARNING:tensorflow:From mrcnn_to_trt_single.py:142: convert_variables_to_constants (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.convert_variables_to_constants`
WARNING:tensorflow:From /home/chieh/.local/lib/python3.6/site-packages/tensorflow/python/framework/graph_util_impl.py:270: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.extract_sub_graph`
WARNING:tensorflow:From /home/chieh/.local/lib/python3.6/site-packages/uff/converters/tensorflow/conversion_helpers.py:227: The name tf.gfile.GFile is deprecated. Please use tf.io.gfile.GFile instead.

WARNING:tensorflow:From /home/chieh/.local/lib/python3.6/site-packages/graphsurgeon/_utils.py:2: The name tf.NodeDef is deprecated. Please use tf.compat.v1.NodeDef instead.

Traceback (most recent call last):
  File "mrcnn_to_trt_single.py", line 169, in <module>
    main()
  File "mrcnn_to_trt_single.py", line 128, in main
    text=True, list_nodes=list_nodes)
  File "mrcnn_to_trt_single.py", line 162, in convert_model
    debug_mode = False
  File "/home/chieh/.local/lib/python3.6/site-packages/uff/converters/tensorflow/conversion_helpers.py", line 229, in from_tensorflow_frozen_model
    return from_tensorflow(graphdef, output_nodes, preprocessor, **kwargs)
  File "/home/chieh/.local/lib/python3.6/site-packages/uff/converters/tensorflow/conversion_helpers.py", line 106, in from_tensorflow
    pre.preprocess(dynamic_graph)
  File "./config.py", line 145, in preprocess
    dynamic_graph.remove(timedistributed_remove_list)
  File "/home/chieh/.local/lib/python3.6/site-packages/graphsurgeon/DynamicGraph.py", line 470, in remove
    nodes = self._force_to_nodes(nodes)
  File "/home/chieh/.local/lib/python3.6/site-packages/graphsurgeon/DynamicGraph.py", line 324, in _force_to_nodes
    buf = self._force_to_names(buf)
  File "/home/chieh/.local/lib/python3.6/site-packages/graphsurgeon/DynamicGraph.py", line 307, in _force_to_names
    assert False, "The name %s does not exist" % el
AssertionError: The name mrcnn_class_bn1/Reshape_5/shape does not exist

Even I got rid of those not existing name from the config.py.

timedistributed_remove_list = [
        "mrcnn_class_conv1/Reshape/shape", "mrcnn_class_conv1/Reshape", "mrcnn_class_conv1/Reshape_1/shape", "mrcnn_class_conv1/Reshape_1",
        "mrcnn_class_bn1/Reshape/shape", "mrcnn_class_bn1/Reshape",
        "mrcnn_class_conv2/Reshape/shape", "mrcnn_class_conv2/Reshape", "mrcnn_class_conv2/Reshape_1/shape", "mrcnn_class_conv2/Reshape_1",
        "mrcnn_class_bn2/Reshape/shape", "mrcnn_class_bn2/Reshape", 
        "mrcnn_class_logits/Reshape/shape", "mrcnn_class_logits/Reshape","mrcnn_class_logits/Reshape_1/shape", "mrcnn_class_logits/Reshape_1",
        "mrcnn_class/Reshape/shape", "mrcnn_class/Reshape","mrcnn_class/Reshape_1/shape", "mrcnn_class/Reshape_1",
        "mrcnn_bbox_fc/Reshape/shape", "mrcnn_bbox_fc/Reshape","mrcnn_bbox_fc/Reshape_1/shape", "mrcnn_bbox_fc/Reshape_1",

        "mrcnn_mask_conv1/Reshape/shape", "mrcnn_mask_conv1/Reshape", "mrcnn_mask_conv1/Reshape_1/shape", "mrcnn_mask_conv1/Reshape_1",
        "mrcnn_mask_bn1/Reshape/shape", "mrcnn_mask_bn1/Reshape", 
        "mrcnn_mask_conv2/Reshape/shape", "mrcnn_mask_conv2/Reshape", "mrcnn_mask_conv2/Reshape_1/shape", "mrcnn_mask_conv2/Reshape_1",
        "mrcnn_mask_bn2/Reshape/shape", "mrcnn_mask_bn2/Reshape", 
        "mrcnn_mask_conv3/Reshape/shape", "mrcnn_mask_conv3/Reshape", "mrcnn_mask_conv3/Reshape_1/shape", "mrcnn_mask_conv3/Reshape_1",
        "mrcnn_mask_bn3/Reshape/shape", "mrcnn_mask_bn3/Reshape", 
        "mrcnn_mask_conv4/Reshape/shape", "mrcnn_mask_conv4/Reshape", "mrcnn_mask_conv4/Reshape_1/shape", "mrcnn_mask_conv4/Reshape_1",
        "mrcnn_mask_bn4/Reshape/shape", "mrcnn_mask_bn4/Reshape", 
        "mrcnn_mask_deconv/Reshape/shape", "mrcnn_mask_deconv/Reshape", "mrcnn_mask_deconv/Reshape_1/shape", "mrcnn_mask_deconv/Reshape_1",
        "mrcnn_mask/Reshape/shape", "mrcnn_mask/Reshape", "mrcnn_mask/Reshape_1/shape", "mrcnn_mask/Reshape_1"
# "mrcnn_class_bn1/Reshape_5/shape" "mrcnn_class_bn1/Reshape_5" "mrcnn_class_bn2/Reshape_5/shape", "mrcnn_class_bn2/Reshape_5",
# "mrcnn_mask_bn1/Reshape_5/shape", "mrcnn_mask_bn1/Reshape_5" "mrcnn_mask_bn2/Reshape_5/shape", "mrcnn_mask_bn2/Reshape_5",
# "mrcnn_mask_bn3/Reshape_5/shape", "mrcnn_mask_bn3/Reshape_5", "mrcnn_mask_bn4/Reshape_5/shape", "mrcnn_mask_bn4/Reshape_5"
]

I still got the strange error.

Traceback (most recent call last):
  File "mrcnn_to_trt_single.py", line 169, in <module>
    main()
  File "mrcnn_to_trt_single.py", line 128, in main
    text=True, list_nodes=list_nodes)
  File "mrcnn_to_trt_single.py", line 162, in convert_model
    debug_mode = False
  File "/home/chieh/.local/lib/python3.6/site-packages/uff/converters/tensorflow/conversion_helpers.py", line 229, in from_tensorflow_frozen_model
    return from_tensorflow(graphdef, output_nodes, preprocessor, **kwargs)
  File "/home/chieh/.local/lib/python3.6/site-packages/uff/converters/tensorflow/conversion_helpers.py", line 106, in from_tensorflow
    pre.preprocess(dynamic_graph)
  File "./config.py", line 149, in preprocess
    connect(dynamic_graph, timedistributed_connect_pairs)
  File "./config.py", line 139, in connect
    if node_a_name not in dynamic_graph.node_map[node_b_name].input:
KeyError: 'mrcnn_mask_bn4/batchnorm/mul_1'

I don’t think so that I should get rid of those names from config.py actually. Is there any solution or thought about this problem?

Thank you.

Environment

• TensorRT version: 7.0.0.11 (Downloaded the tar file from NVIDIA website.)
• Cuda version: 10.2
• TensorFlow-gpu: 1.14.0
• Cudnn version: 7.6.5
• GPU: GTX1060
• Ubuntu: 18.04
• Keras : 2.3.1

PS　I didn’t use the NVidia docker and NGC.