Release code for gridworld with reinforcement learning?

[streamorning]

Hi, could you release the code for gridworld with reinforcement learning? Thanks a lot!

[kentsommer]

@streamorning,

If you could clarify what you mean that would help. This repository contains all the code used to achieve the gridworld results published in @avivt's paper. If you are wondering which section of the code actually handles training you are probably looking for this: https://github.com/avivt/VIN/blob/master/vin.py#L96-L123

If you are looking for the two comparison approaches (DQN style CNN, FCN network), they can be found in these two files:

https://github.com/avivt/VIN/blob/master/CNN.py
https://github.com/avivt/VIN/blob/master/FC.py

[streamorning]

Thanks a lot for your reply. I could not find the code described in Appendix B in original paper, which used TRPO as basic RL algorithm. I'm not sure whether this repository contains all these code.

[kentsommer]

Ah, I misunderstood what you were asking, sorry about that! That code is indeed not include here, perhaps it will be added at some point. However, I believe @avivt's current aim (related to this work) is getting the mars experiment code up. So, hopefully he has time to respond to this at some point 👍

[streamorning]

Thank you anyway~👍