How do I save Episode Manager training data for *plotting* later

27 views (last 30 days)
Lets say I am using two algorithms to train for a task (DDPG and PPO). How do I save the data to plot a comparion later?
This is data like episode rewards, Q-value etc. plotted at every episode.
(Please note I am not referring to network weights for transfer learning.)

Accepted Answer

Rajesh Siraskar
Rajesh Siraskar on 5 Jan 2021
Edited: Rajesh Siraskar on 5 Jan 2021
Thank you Asvin and Emmanouil.
I didnt realize that I could store trainingStats and I can do anything with the data later. Awesome.
% Train the agent
trainingStats = train(agent, env, trainOpts);

More Answers (1)

Emmanouil Tzorakoleftherakis
Edited: Emmanouil Tzorakoleftherakis on 11 Mar 2021
Hi Rajesh,
As mentioned in the comment above, if you are using R2020b, you can use
>> help rlPlotTrainingResults
to recreate the Episode manager plot.
If you are using R2021a, you may notice that this function is now renamed to "inspectTrainingResult".
Hope that helps

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by