Facing issues in generating code for Deep Reinforcement Learning trained network for hardware deployment

Hai,

I am trying to Generate C++ Code for Deep Neural Network Policy using Third-Party Libraries. I used the following link to bulid the MKLDNN library for MATLAB Version 2020b with suggested instructions (MATLAB Coder: How do I build the Intel MKL-DNN library for Deep Learning C++ code generation and deployment? - (mathworks.com)) . When I execute the codeden instruction (codegen('-config','cfg','evaluatePolicy','-args',argstr,'-report')), I am getting the following error.

Error:

??? The library method 'dltargets.mkldnn.mkldnnApi.updateBuildInfo' failed.

Caused by: Insufficient number of outputs from right hand side of equal sign to satisfy assignment.

Code generation failed: View Error Report

Answer 1

[Aman]

Hi Anugula,

I understand that you are having issues in generating code for Deep Reinforcement Learning trained network.

This error occurs when the number of arguments that the function is returning is less than the number of arguments that you are trying to assign. This can be resolved either by making the assignments equal to the number of arguments that the function returns or by modifying the function to return a greater number of arguments.

Please refer to the following thread to know how to resolve this error:

https://www.mathworks.com/support/search.html/answers/698720-insufficient-number-of-outputs-from-right-hand-side-of-equal-sign-to-satisfy-assignment.html?fq%5B%5D=asset_type_name:answer&fq%5B%5D=category:matlab/data-types&page=1

Hope this helps!

Regards,

Aman Mehta