Target language problem with coder.load​DeepLearni​ngNetwork in Simulink

Hi everyone,

I am trying to implement a YOLO network in Simulink. For my first project, I used the YOLOV2ObjectDetectionUsingIntelMKLDNNExample. Running & experimenting with the example in Matlab works like charm. But when I try to run the yolov2_detection function in a Matlab Function Block in Simulink, I get a compiler error:

Target selection language property on Code Generation configuration was not set to C++. Simulation with deep learning network requires target language to be set to C++.

Changing the default mex compiler to cpp by using mex -setup cpp (MEX configured to use 'Microsoft Visual C++ 2017' for C++ language compilation.) didn't resolve the problem. Do I have to change the target language somewhere else? Is there another compiler? I would be thankful for any insights!

Minimal code example to reproduce the error (using the files in MATLAB\Examples\R2020a\deeplearning_shared\YOLOV2ObjectDetectionUsingIntelMKLDNNExample)

function out_img = run_yolo(img)

persistent yolov2Obj;

if isempty(yolov2Obj)

yolov2Obj = coder.loadDeepLearningNetwork('yolov2ResNet50VehicleExample.mat');

end

out_img = img

Specs:

• Window10 System
• Using Matlab2020A with the Deeplearning Toolbox
• No Matlab Coder available

[Lucas Ferreira-Correia]

Hi, did you manage to resolve the issue?

[Martin Boehme]

The problem can be fixed. The following post solved the issue: https://www.mathworks.com/matlabcentral/answers/447387-matlab-coder-how-do-i-build-the-intel-mkl-dnn-library-for-deep-learning-c-code-generation-and-dep