search

trackmania-rl / tmrl

Reinforcement Learning for real-time applications - host of the TrackMania Roborace League
https://pypi.org/project/tmrl
MIT License
501 stars 68 forks source link

Change MAP for incremental learning #31

Closed Thomas-Raynaud-ENSTA closed 1 year ago

Thomas-Raynaud-ENSTA commented 1 year ago

Hi!

In the scope of a school project, alongside with my team we woud like to use your library to create an AI that drives in TrackMania. We were wondering if it was possible to change the map on which the AI is driving. In fact, we managed to get it run on different map, although it keeps reseting after a couple of second.

Therefore, is there a way to setup new maps for training/testing? If yes, where should we look at?

Thanks!

yannbouteiller commented 1 year ago

Hi Thomas, sure there is a way of doing so ;)

Just follow the instructions in the Getting started guide here.

(Closing as this is not an issue, please use discussions for this type of questions)

yannbouteiller commented 1 year ago

(The reason why it restarts on other maps is because you have not built a reward function for these maps, thus the rtgym environment believes your car is running loose away from the track)

Thomas-Raynaud-ENSTA commented 1 year ago

Thank you for your reply and your instructions. This is much clearer!