Load Pretrained Networks in Matlab function block

Hi,

I want use pretrained LSTM network in simulink matlab function block but ı get this error.

function y= fnc( Va,Vb,Vc,Ia,Ib,Ic)
input= [Va, Vb, Vc, Ia, Ib, Ic];
input=rescale(input); 
XTrain = {input'};  
net = coder.loadDeepLearningNetwork('alphabeta.mat');
output= predict(net, XTrain, 'MiniBatchSize', 1);
y=output{1};

Simulink does not have enough information to determine output sizes for this block. If you think the errors below are inaccurate, try specifying types for the block inputs and/or sizes for the block outputs.

Component:MATLAB Function | Category:Coder error

For deep learning, the simulation target language must be set to C++. Function 'MATLAB Function' (#117.120.166), line 6, column 7: "coder.loadDeepLearningNetwork('alphabeta.mat')" Launch diagnostic report.

Answer 1

[Paul]

According to the doc pages coder.loadDeepLearningNetwork and predict- (not recommended) and the error message this function might work (not tested)

function y= fnc( Va,Vb,Vc,Ia,Ib,Ic)
persistent net
if isempty(net)
    net = coder.loadDeepLearningNetwork('alphabeta.mat');
end
input= [Va, Vb, Vc, Ia, Ib, Ic];
input=rescale(input); 
XTrain = {input'};  
output= predict(net, XTrain, 'MiniBatchSize', 1);
y = ????; % assign a dummy value to y that is of the same size and type of output{1}
y=output{1};
end

Also be sure to go to Model Settings -> Simulation Target -> Language and select C++ from the dropdown menu.