search

openai / gym

A toolkit for developing and comparing reinforcement learning algorithms.
https://www.gymlibrary.dev
Other
34.77k stars 8.61k forks source link

How do I give human inputs to the environment? #675

Closed 0xDaksh closed 7 years ago

0xDaksh commented 7 years ago

Hey,

I was trying to build the MountainCar-v0 Env but while training I noticed the random steps taken which was in my input data were not efficient and didn't work well, So I decided to give it input on my own but the way to do it is no where to be found? is there any way to give the env inputs from human?

Best. Daksh

olegklimov commented 7 years ago

Please see keyboard_agent.py

eabase commented 5 months ago

@olegklimov Where did this go?

pseudo-rnd-thoughts commented 5 months ago

See utils.play

eabase commented 5 months ago

@pseudo-rnd-thoughts Hi Mark, TY, but that doesn't seem to work. I mean, it runs but nothing at all happens.

I tried:

C:\venvs\lunar\Lib\site-packages\gymnasium\utils
python.exe .\play.py LunarLander-v2

In the old version you could do this: