build ssd engine failed

[YoungSharp]

[TensorRT] ERROR: Could not register plugin creator: FlattenConcat_TRT in namespace:
WARNING: To create TensorRT plugin nodes, please use the create_plugin_node function instead.
NOTE: UFF has been tested with TensorFlow 1.14.0.
WARNING: The version of TensorFlow installed on this system is not guaranteed to work with UFF.
UFF Version 0.6.5
=== Automatically deduced input nodes ===
[name: "Input"
op: "Placeholder"
attr {
key: "shape"
value {
shape {
dim {
size: 1
}
dim {
size: 3
}
dim {
size: 300
}
dim {
size: 300
}
}
}
}
]

Using output node NMS
Converting to UFF graph
Warning: No conversion function registered for layer: NMS_TRT yet.
Converting NMS as custom op: NMS_TRT
Warning: No conversion function registered for layer: FlattenConcat_TRT yet.
Converting concat_box_conf as custom op: FlattenConcat_TRT
Traceback (most recent call last):
File "build_engine.py", line 204, in
main()
File "build_engine.py", line 187, in main
debug_mode=DEBUG_UFF)
File "/usr/local/lib/python3.6/site-packages/uff/converters/tensorflow/conversion_helpers.py", line 178, in from_tensorflow
debug_mode=debug_mode)
File "/usr/local/lib/python3.6/site-packages/uff/converters/tensorflow/converter.py", line 94, in convert_tf2uff_graph
uff_graph, input_replacements, debug_mode=debug_mode)
File "/usr/local/lib/python3.6/site-packages/uff/converters/tensorflow/converter.py", line 79, in convert_tf2uff_node
op, name, tf_node, inputs, uff_graph, tf_nodes=tf_nodes, debug_mode=debug_mode)
File "/usr/local/lib/python3.6/site-packages/uff/converters/tensorflow/converter.py", line 47, in convert_layer
return cls.registry_[op](name, tf_node, inputs, uff_graph, **kwargs)
File "/usr/local/lib/python3.6/site-packages/uff/converters/tensorflow/converter_functions.py", line 21, in convert_placeholder
dtype = tf2uff.convert_tf2numpy_dtype(tf_node.attr['dtype'].type)
File "/usr/local/lib/python3.6/site-packages/uff/converters/tensorflow/converter.py", line 103, in convert_tf2numpy_dtype
return tf.as_dtype(dtype).as_numpy_dtype
File "/usr/local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py", line 712, in as_dtype
raise TypeError("Cannot convert value %r to a TensorFlow DType." % type_value)
TypeError: Cannot convert value 0 to a TensorFlow DType.

linux 16.04
cuda10.0
cudnn7.6.5
Tensorrt 6.0.1.5
tensorflow 1.12
python 3.6

command line: python3 build_engine.py ssd_mobilenet_v2_egohands

[jkjung-avt]

I haven't tested the code with TensorRT 6.x before...

Are you running the code on a Jetson platform? Or a x86 PC?

[YoungSharp]

on x86 PC

[YoungSharp]

error happend when converting Input node.

[jkjung-avt]

I don't have an answer right now. My code was originally developed and tested on Jetson (arm64) platforms.

I might find time to test that later.

[YoungSharp]

i will keep trying, thanks.

[YoungSharp]

@jkjung-avt
Do you mind to share your training code with?
The code to train ssd_mobilenet_v2_egohands.pb
my code is latest "https://github.com/tensorflow/models/blob/master/research", and my model is at "https://drive.google.com/file/d/1e-9cb_sLmuevA4h1IPP029vUH3IxlIx1/view?usp=sharing"
my email address is 331870428@qq.com
Appreciated.

[jkjung-avt]

I already shared all my training code on GitHub. I was using a slightly older snapshot of TensorFlow Object Detection API. Please refer to README.md in the repository and the associated blog posts for details.

https://github.com/jkjung-avt/hand-detection-tutorial

[YoungSharp]

Thanks really

[jkjung-avt]

By the way, the problem you're facing might be due to some changes in TensorRT 6.0. I think the problem would likely go away if use TensorRT 5.x instead.

[YoungSharp]

Your model works well under Tensorrt 6.x;
anyway, i will give it a try.

[jkjung-avt]

I see, Thanks for the update.

[Ndron]

same problem

[jkjung-avt]

@Ndron Could you check and make sure you have patched "graphsurgeon/node_manipulation.py"? More specifically, you have to add the following line of code into the file:

+    node.attr["dtype"].type = 1

Reference:

• https://github.com/jkjung-avt/tensorrt_demos/blob/master/README_x86.md
• https://github.com/AastaNV/TRT_object_detection ("Update graphsurgeon converter")

[Ndron]

@jkjung-avt Fixed , thanks

[aimuch]

@jkjung-avt Fixed , thanks

Hi， same problem，how did you solved it?
Thanks