How to deal Found unsupported class for variable using function 'load'. The value at 'mynet1.net' is unsupported for code generation.

I am running a deep learning model to control a system that has been designed using Deep netwoek designer. However, the simulink model gave me an error in this type"

Found unsupported class for variable using function 'load'. The value at 'mynet1.net' is unsupported for code generation."

"Use the command 'whos -file NET.mat' to view the variables in the MAT file."

The code in the block function is given by:

function out = predict(in)
x1=in(1);
x2=in(2);
persistent mynet1;
if isempty(mynet1)
    mynet1 = load("NET.mat","net","-mat");
end
out = predict(mynet1,[x1;x2])

end

Answer 1

[Sayan Saha]

Deep learning networks aren't supported to be loaded via "coder.load" for code generation. You should use "coder.loadDeepLearningNetwork" instead. Here is a reference page that shows the workflow: https://www.mathworks.com/help/coder/ug/load-pretrained-networks-for-code-generation.html