Issues

Question

Issues

nil123532 opened this issue a year ago · 7 comments

nil123532 commented a year ago

Hi,
Please update code on setting up.
I have been getting a lot of issues and can't run the code.

Answer 1 · 2024-01-02T05:02:27.000Z

Kind reminder about this project's code.

Answer 2 · 2024-01-20T05:51:35.000Z

why is the code not released yet?

Answer 3 · 2024-01-25T00:16:11.000Z

Hi, thanks for reaching out! Thank you for your paientice!

I was on my winter break after NeurIPS and am now back to work.

Yea I am finalizing the release of the code. Though we have the implementation on our hand, I want to make sure it's most up to date, easy to understand and reproducible. The implementation will be finalized within this week. Please stay tuned!

Answer 4 · 2024-01-25T11:45:57.000Z

Thank you.
Wondering if we will be able to use the code on any gym environments

Answer 5 · 2024-01-25T21:31:23.000Z

we will be able to use the code on any gym environments

It's possible only if the action space of the environment is compatible with the human control device, or at least you should have some approaches to interpret the human input actions into the real action space of the RL environment.

Thanks for reminding me that I should add a section discussing how we can use the same codebase to try new environments. Basically, the user need to implement a gym wrapper that can take human action as input and add some fields into the info dict return by the environment.

Answer 6 · 2024-03-15T06:30:22.000Z

a gym wrapper that resembles the minigridwrapper?

Answer 7 · 2024-03-26T06:49:21.000Z

hi I am having some issue for this.
My environment just freezes and the agent's action always say "None"

NOTE: I am using my own environment