Error in generating Generic C/C++ Code for Deep Learning Networks

I'm trying to generate C/C++ code that does not depend on library. Here is the link that I followed.

https://www.mathworks.com/help/coder/ug/generate-generic-cc-code-for-deep-learning-networks.html

While executing following command I'm getting error.

>> dlconfig = coder.DeepLearningConfig('none');
Error using coder.DeepLearningConfig. Target library 'none' is not supported. Supported values for target library are cudnn, tensorrt, mkldnn,
arm-compute, arm-compute-mali.

Answer 1

[Raj Patel]

In matlab 2020b, it's not working, but in 2021a, works perfectly.

Used following code.

cfg = coder.config('lib');
cfg.TargetLang = 'C++';
cfg.DeepLearningConfig = coder.DeepLearningConfig('none');

[Sayan Saha]

Hi Raj,

Generating C/C++ code for deep neural nets without using any third party library by setting the deep learning config to "none" is a new feature introduced in 21a release. That's why you receive an error in 20b and it works fine in 21a.

~Sayan

[伯儒 周]

Hi Sayan,

I also following the link mentioned above, but the coder and the codegen report an error:

expected a character vector

It seems that the error occurs at coder.loadDeepLearningNetwork. I have tried several times but the problem cannot be solved.

I will be grateful if you can provide me suggestion to solve this problem.