Pausing reinforcement learning by forcing
1 Ansicht (letzte 30 Tage)
Ältere Kommentare anzeigen
SHromaneko
am 2 Jan. 2024
Bearbeitet: SHromaneko
am 10 Jan. 2024
I'm running reinforcement learning, but I think there are times when I think it's not going well and want to stop it.
If you stop today's learning agent executed with the code below with Training stopped, simulink will freeze and stop working.
Repeatedly hitting the escape key doesn't work either.
so that you can stop it properly
Is there something wrong with the code?
numObs = 9;
obsInfo = rlNumericSpec([numObs 1]);
obsInfo.Name = "observations";
mdl = "kineticmodel_wIAVFBC_IncS2Ea_NH3FBC";
open_system(mdl)
numAct = 1;
actInfo = rlNumericSpec([numAct 1],LowerLimit=0,UpperLimit=1);
actInfo.Name = "NH3";
blk = mdl + "/RL agent/RL Agent";
env = rlSimulinkEnv(mdl,blk,obsInfo,actInfo);
Ts = 1
agent = createDDPGAgent(numObs,obsInfo,numAct,actInfo,Ts);
maxEpisodes = 2000;
Tf = 1240*3
maxSteps = floor(Tf/Ts);
trainOpts = rlTrainingOptions(...
MaxEpisodes=maxEpisodes,...
MaxStepsPerEpisode=maxSteps,...
ScoreAveragingWindowLength=250,...
Verbose=false,...
Plots="training-progress",...
StopTrainingCriteria="EpisodeCount",...
StopTrainingValue=maxEpisodes,...
SaveAgentCriteria="EpisodeCount",...
SaveAgentValue=maxEpisodes);
doTraining = true;
if doTraining
% Train the agent.
trainingStats = train(agent,env,trainOpts);
else
% Load a pretrained agent for the selected agent type.
if strcmp(AgentSelection,"DDPG")
load("rlWalkingBipedRobotDDPG.mat","agent")
else
load("rlWalkingBipedRobotTD3.mat","agent")
end
end
0 Kommentare
Akzeptierte Antwort
Emmanouil Tzorakoleftherakis
am 9 Jan. 2024
The proper way to stop it would be through the Episode Manager (top right of the window). Does this not work for you?
1 Kommentar
Weitere Antworten (0)
Siehe auch
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!