TFLite model conversion

Describe the bug After converting the model to .pb file using firenet-to-protobuf I am trying to convert it to TFlite model for inferencing on Google Coral USB accelerator, however while using tflite_convert command line to do so I am encountering the error as mentioned below in the output section and there is no resolution as of yet, is it the dropout layer which is the issue?

To Reproduce: (please complete the following information):

1. Convert model to pb file using firenet-to-protobuf.py file.
2. Run tensorflow lite conversion command as ‘tflit_convert --graph_def_file=‘path_to_filr/firenet.pb --output_format=TFLITE --output_file=firenet_v1__new.tflite --input_arrays=input --output_arrays=final_result’’

Error output received

pi@raspberrypi:~ $ tflite_convert --graph_def_file=/home/pi/Downloads/FireDetection/code/firenet.pb --output_format=TFLITE --output_file=firenet_v1__new.tflite  --input_arrays=input --output_arrays=final_result
WARNING: Logging before flag parsing goes to stderr.
W0729 04:04:35.524526 3069684432 deprecation_wrapper.py:118] From /usr/local/lib/python3.7/dist-packages/tensorflow/__init__.py:98: The name tf.AUTO_REUSE is deprecated. Please use tf.compat.v1.AUTO_REUSE instead.

W0729 04:04:35.525118 3069684432 deprecation_wrapper.py:118] From /usr/local/lib/python3.7/dist-packages/tensorflow/__init__.py:98: The name tf.AttrValue is deprecated. Please use tf.compat.v1.AttrValue instead.

W0729 04:04:35.525362 3069684432 deprecation_wrapper.py:118] From /usr/local/lib/python3.7/dist-packages/tensorflow/__init__.py:98: The name tf.COMPILER_VERSION is deprecated. Please use tf.version.COMPILER_VERSION instead.

W0729 04:04:35.525576 3069684432 deprecation_wrapper.py:118] From /usr/local/lib/python3.7/dist-packages/tensorflow/__init__.py:98: The name tf.CXX11_ABI_FLAG is deprecated. Please use tf.sysconfig.CXX11_ABI_FLAG instead.

W0729 04:04:35.525793 3069684432 deprecation_wrapper.py:118] From /usr/local/lib/python3.7/dist-packages/tensorflow/__init__.py:98: The name tf.ConditionalAccumulator is deprecated. Please use tf.compat.v1.ConditionalAccumulator instead.

Traceback (most recent call last):
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/framework/importer.py", line 428, in import_graph_def
    graph._c_graph, serialized, options)  # pylint: disable=protected-access
tensorflow.python.framework.errors_impl.InvalidArgumentError: NodeDef expected inputs '' do not match 1 inputs specified; Op<name=Const; signature= -> output:dtype; attr=value:tensor; attr=dtype:type>; NodeDef: {{node Dropout_1/cond/dropout/random_uniform/max}}

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/bin/tflite_convert", line 10, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/lite/python/tflite_convert.py", line 503, in main
    app.run(main=run_main, argv=sys.argv[:1])
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/platform/app.py", line 40, in run
    _run(main=main, argv=argv, flags_parser=_parse_flags_tolerate_undef)
  File "/usr/local/lib/python3.7/dist-packages/absl/app.py", line 300, in run
    _run_main(main, args)
  File "/usr/local/lib/python3.7/dist-packages/absl/app.py", line 251, in _run_main
    sys.exit(main(argv))
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/lite/python/tflite_convert.py", line 499, in run_main
    _convert_tf1_model(tflite_flags)
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/lite/python/tflite_convert.py", line 124, in _convert_tf1_model
    converter = _get_toco_converter(flags)
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/lite/python/tflite_convert.py", line 111, in _get_toco_converter
    return converter_fn(**converter_kwargs)
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/lite/python/lite.py", line 633, in from_frozen_graph
    _import_graph_def(graph_def, name="")
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/util/deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/framework/importer.py", line 432, in import_graph_def
    raise ValueError(str(e))
ValueError: NodeDef expected inputs '' do not match 1 inputs specified; Op<name=Const; signature= -> output:dtype; attr=value:tensor; attr=dtype:type>; NodeDef: {{node Dropout_1/cond/dropout/random_uniform/max}}

Computing Environment (please complete the following information):

• OS: Raspbian Buster
• Tensorflow Version: 1.13.1
• OpenCV Version: 3.2.0

Additional context

[tobybreckon]

Reading here - https://www.tensorflow.org/lite/convert/cmdline_examples - it appears I may be correct on the command line arguments front.

[ashwin-phadke]

Do you mean “training” with the dropout layers commented out or runtime test/inference using the pre-trained model ?

Training with the dropout layers commented out.

I ask, as this repo doesn’t provide code to do the training part itself.

On my todo list is as follows:

1. change the lines 48/51 in firenet.py to be similarly inside an if statement for the training boolean flag within the construct_firenet() routine (and similarly for other 2 models in repo)
2. test runtime test/inference with and without the training flag set over the first N frames of the sample mp4 video provided with the models to see if the output differs
3. possibly wrap this test into a python unit test framework specific to this repo https://docs.python.org/3/library/unittest.html

That is really great to know.

re: Edit 2 - sorry you are at the end of my tflite_convert knowledge.

It’s okay , I figured it out. The solution is to build tensorflow from source along with bazel with Neon instructions enabled in the flag.