How to save pretrained DQN agent and extract the weights inside the network?

The following is part of the program. I want to know how to extract the weight values from the trained DQN network.

DQNnet = [
    imageInputLayer([1 520 1],"Name","ImageFeatureInput","Normalization","none")
    fullyConnectedLayer(1024,"Name","fc1")
    reluLayer("Name","relu1")
    % fullyConnectedLayer(512,"Name","fc2")
    % reluLayer("Name","relu2")
    fullyConnectedLayer(14,"Name","fc3")
    softmaxLayer("Name","softmax")
    classificationLayer("Name","ActionOutput")];

ObsInfo = getObservationInfo(env);
ActInfo = getActionInfo(env);

DQNOpts = rlRepresentationOptions('LearnRate',0.0001,'GradientThreshold',1,'UseDevice','gpu');
DQNagent = rlQValueRepresentation(DQNnet,ObsInfo,ActInfo,'Observation',{'ImageFeatureInput'},'ActionInputNames',{'BoundingBox Actions'},DQNOpts);

agentOpts = rlDQNAgentOptions(...
    'UseDoubleDQN',true ...
    ,'MiniBatchSize',256);

agentOpts.EpsilonGreedyExploration.Epsilon = 1;

agent = rlDQNAgent(DQNagent,agentOpts);

%% Agent Training
% Training options

trainOpts = rlTrainingOptions(...
    'MaxEpisodes', 100, ...
    'MaxStepsPerEpisode', 100, ...
    'Verbose', true, ...
    'Plots','training-progress',...
    'ScoreAveragingWindowLength',400,...
    'StopTrainingCriteria','AverageSteps',...
    'StopTrainingValue',1000000000,...
    'SaveAgentDirectory', pwd + "\agents\");

% Agent training
trainingStats = train(agent,env,trainOpts);

Answer 1

[praguna manvi]

Hi @Kuan Yi Li,

For saving and loading pretrained “DQN” agent, you could use “load” and “save” functions as:

doTraining = false; 
if doTraining 
    % Train the agent. 
    trainingStats = train(agent,env,trainOpts); 

    save myagent.mat agent 
else 
    % Load the pretrained agent for the example. 
    load("MATLABCartpoleDQNMulti.mat","agent") 
end 

Refer to the example illustrated here:

https://www.mathworks.com/help/reinforcement-learning/ug/train-dqn-agent-to-balance-cart-pole-system.html#MATLABCartPoleDQNExample-4

And to extract weights from the saved agent you can use “getLearnableParameters“ function refer:

https://www.mathworks.com/matlabcentral/answers/513136-how-can-i-extract-a-trained-rl-agent-s-network-s-weights-and-biases