the func from_tensorflow_frozen_model has no parameter preprocessor

I use this code to convert my tensorflow model to uff

import uff
uff.from_tensorflow_frozen_model("frozen_inference_graph.pb", 
			output_nodes=["NMS"],
			output_filename="frozen.uff",
                        preprocessor='config.py'
)

But I found that the file ‘config.py’ does not run before create uff file which causes failure.

I read the ‘conversion_helpers.py’ located in

/usr/local/lib/python2.7/dist-packages/uff/converters

I found that the function ‘from_tensorflow’ did not process the file ‘config.py’, and what should I do?

Hi,

Is there any plugin layer inside your model?

You don’t need the config.py file if there is no customized layer.
Just output the model with this command:

# generate uff from frozen graph
uff_model = uff.from_tensorflow_frozen_model(
    frozen_file=frozen_graph_filename,
    output_nodes=[output_name],
    output_filename=TMP_UFF_FILENAME,
    text=False,
)

Thanks.