Hi Burak,
I realize that you want to get a MATLAB function out of a deep neural network after quantization, just like you can get one out of a shallow network with "genFunction".
As of now, there is no functionality to get a MATLAB function out of a deep neural network post quantization, but there is a workaround if you want to generate a deployable code for a microcontroller or a system; using MATLAB Coder.
The steps are:
- Create your neural network, load your dataset.
- Quantize the net, and validate it to make sure it is working as expected.
- Use the MATLAB Coder functionality to generate a C/C++ code which is deployable on machines.
To know how to quantize the networks, please refer to these documentations:
"quantize" function : https://www.mathworks.com/help/deeplearning/ref/dlquantizer.quantize.html
"dlquantizer" object : https://www.mathworks.com/help/deeplearning/ref/dlquantizer.html
To know how to deploy deep learning models using MATLAB Coder, follow this link:
Hope this helps!
