Getting Started with Keras models

If the training was done with Keras then the model can be saved by using the python script named KerasModeltoJSON.py. It will create a json file which can be read by NNSharp. NNSharp will build the model in C# and provides function to execute it. The python script takes the created and compiled Keras model and the output file name as arguments. Then the json file can be created in the following way in your Python program:

import KerasModeltoJSON as js
wrt = js.JSONwriter(model, file_path)
wrt.save()

Then in the C# program use the following:

// Read the previously created json.
var reader = new ReaderKerasModel(filePath); 
SequentialModel model = reader.GetSequentialExecutor();

// Then create the data to run the executer on.
// batch: should be set in the Keras model.
Data2D input = new Data2D(height, width, channel, batch);

// Calculate the network's output.
IData output = model.ExecuteNetwork(input);

In order to know what data should be the input and the output see the corresponding documentations for the layer at input and the layer at output, respectively.