Saving Reinforcement learning agent does not work properly
2 views (last 30 days)
Show older comments
Hi I'm working on a RL project with DDPG agent.
I have troble with saving the agent.
I had used the rlTrainingOptions.save for saving the agent in the middle of the traing but it kept saving the wrong weights (I guess)
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/701257/image.png)
These are my training options.
It saves the agent whenever episode reward exceeds "SaveAgentValue', but everytime I load the saved agent, it never exceeds the reward that I set for saving, rather it shows very poor performance..
Weirdly, the saved Final agent works properly. I don't know what is going on with this
Can anybody teaches me the right answer for this problem?? :-(
0 Comments
Answers (0)
See Also
Categories
Find more on Training and Simulation in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!