Creating a CNN model in simulink from matlab code

Goodday,

I successfully ran the code found at: https://www.mathworks.com/matlabcentral/fileexchange/69632-code-for-webinar-deep-learning-for-signals-and-sound

Now I wish to create a simulink model based on the CNN trained in the afforementioned code,I should note that I am new to simulink and have had some difficulties, as far as the research I have done the available deep learning blocks in Simulink would not seem to help in this matter, alterantively I have tried running the CNN as a function but simulink returned numerous errors indicating that it did not recognize majority of the commands utilized in the network testing function. I require some advice in this matter or rather some confirmation whether a CNN trained to denoise audio can even be modeled in Simulink in the first place.

Answer 1

[Shivam Singh]

Hi Arvinda,

It is my understanding that you want to implement a trained CNN in Simulink. To achieve this, you can utilize the MATLAB function block to generate predictions for your inputs. An example of this can be found at: “GPU Code Generation for Deep Learning Networks Using MATLAB Function Block”

Also, you can import a pre-trained network from a MAT-file or MATLAB function (like googlenet function) using the “Predict” block, available in Deep Neural Network Library.

I hope this helps you.