Running the package “Quanser QUBE-Servo 2 Pendulum Control Reinforcement Learning” causes an error when extracting the policy from the agent

I downloaded and ran the “Quanser QUBE-Servo 2 Pendulum Control Reinforcement Learning” package.

Learning was successful, but an error occurred in the part where the policy is extracted from the agent.

The error is that the policy is missing.

In the meantime, I would like you to download this package and try it out.

And please let me know how to solve this error.

We look forward to hearing from you!

Quanser QUBE-Servo 2 Pendulum Control Reinforcement Learning

https://jp.mathworks.com/matlabcentral/fileexchange/106935-quanser-qube-servo-2-pendulum-control-reinforcement-learning

バージョン 1.0.3 (1.63 MB) 作成者: Quanser

Shows how to design a reinforcement learning agent to balance the Quanser QUBE-Servo 2 Inverted Pendulum.

% Load pre-defined policy (false) or generate new policy for RT code gen (true)
% doPolicy = false;
doPolicy = true;
% 
if doPolicy
    % Generate policy for deployment
    generatePolicyFunction(agent,'MATFileName','QubeIPBalRLPolicy.mat');
    load("QubeIPBalRLPolicy.mat");
    policy = cut_unnecessary_layers_for_SAC_policy(policy);
    save("QubeIPBalRLPolicy.mat", 'policy');
else
    % Load previously saved policy
    load("QubeIPBalRLPolicy.mat");
end