Save net for later genFunction function call

I'm using the Neural Network framework to generate with genFunction the weights and other parameters that I use to encapsulate into another project. In my case, the parameters I get with genFunction are x1_step1, y1_step1, b1,b2, IW_1, LW_2.

I'd like to run a batch of trainings, then get those parameters from every network I train, and using genFunction all over the batch would be unusefull. However, creating an array of nets and concatenating them into the for loop seems not working: I only get an unreadable network cell. I tried using struct(net) and I successfully got an array row that can be concatenated, but I cannot cast the rows it to a net with cast(netarray, 'net').

The purpose I want to achieve is using the best MAE in the batch to get the corresponding net's weights.

How do I solve this problem?

here it is the code

    trainFcn = 'trainbr'; %Bayesan backprop.
    netMAE_OUT = [];
    netArray = [];
    hiddenLayerSize = 3;
    trainsize = 2;
    step = 12;
    net = fitnet(hiddenLayerSize,trainFcn);
    for  c = 1:trainsize

    % Setup Division of Data for Training, Validation, Testing
    net.divideParam.trainRatio = 65/100;
    net.divideParam.valRatio = 20/100;
    net.divideParam.testRatio = 15/100;

    for  i = 1:5

        %divide training and target set
        xloop = [x(:,1:step*(i-1),:) x(:,step*i+1:end,:)];
        tloop = [t(:,1:step*(i-1)) t(:,step*i+1:end)];
        % Train the Network
        [net,tr] = train(net,xloop,tloop);
        y = net(x);
        e = gsubtract(t,y);

        %%here there is my try to save the network.
        netArray = [netArray struct(net)];

        mae_target = mae(e)
        netMAE_OUT = [netMAE_OUT mae_target];

    end
end