Error while calling a Python Module

The Setup: I have a deep net model trained in python (using keras and tensorflow packages). I saved the model file as a json file and weights in h5 format. I wrote a simple Python script that takes two arrays as input, runs the learned model and returns a prediction. My main file in matlab is providing the inputs to the python file, calling it to return a prediction, and then continues to do further computation on it.

The Problem: Everytime I call the python module from matlab, I get the following error: "Python Error: AttributeError: 'array.array' object has no attribute 'reshape'" My model in python requires a 4D array. Since I figured earlier, the hard way, that one can't pass a multi-dimensional array from matlab to python, I'm passing a 1D array to python and doing the necessary conversion there itself. While, the error says I can't reshape, when I run the same commands through the python interpreter, they work.

Any help to resolve this is much appreciated. Thanks!

[Robert Snoeberger]

To respond to your comment "when I run the same commands through the python interpreter, they work", the following is a way to hit the same error with the python interpreter.

>>> import array
>>> a = array.array('d', range(20))
>>> a.reshape((2,10))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'array.array' object has no attribute 'reshape'
>>>

Maybe a solution for the above will fix your issue.

[ARUN]

Hello, I am doing exactly the same, but when I call the python script with the keras model loaded as .h5 from Matlab, I am getting a Matlab error. Can you suggest me how I can solve this issue?

Thanks in advance!