Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Missing connector in Cartpole example? #11

Open
gdommergue opened this issue Jan 27, 2022 · 3 comments
Open

Missing connector in Cartpole example? #11

gdommergue opened this issue Jan 27, 2022 · 3 comments

Comments

@gdommergue
Copy link

Hi @krumiaa

2 questions:

Just recently purchased MindMaker: Deep Reinforcement Learning (SingleNode) from UE marketplace and had a look at your Cartpole example, it seems a connector is missing in cart_BP?

UwSEpTtArr

UE4Editor_KNOAcGQFbH

Also, from my understanding, the "Done" boolean is used to notify the RL engine that the current episode has ended, is that correct?

Thanks.

@krumiaa
Copy link
Owner

krumiaa commented Jan 27, 2022

Apologies for the confusion, the DONE node is not necessary for the proper functioning of the example. You don’t need to connect it. The demo will automatically close once the number of training or eval episodes is reached. I experimented with being able to break training prior to reaching the full number of training episodes by implementing a Boolean break. Think I left this in there in the event someone wanted to play around with doing the same. I should have commented that in the blueprints, will do so in an update. Thanks for bringing this to my attention.

@gdommergue
Copy link
Author

Thanks for your reply.
I still don't understand how you handle the end of each episode? In Unity ML Agent, for instance, you can call EndEpisode() to let the AI know that the episode is over.
Should i set Done=True to do that?

@krumiaa
Copy link
Owner

krumiaa commented Feb 1, 2022

After the make observations blueprint function completes, it passes that data back to mindmaker learning engine, completing the episode. After this another episode automatically begins, it basically waits in a holding pattern for the blueprint to pass back reward and observation data. you could break from it at any time by changing the make observations blueprint function, otherwise it continues looping until the specified number of training or eval episodes is completed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants