search

Farama-Foundation / Minigrid

Simple and easily configurable grid world environments for reinforcement learning
https://minigrid.farama.org/
Other
2.09k stars 604 forks source link

Added support for setting a custom max_steps value in doorkey environment #193

Closed zeionara closed 2 years ago

pseudo-rnd-thoughts commented 2 years ago

Hi, could you provide a justification for the change?

zeionara commented 2 years ago

Hi, the justification is that the predefined limit of the number of steps is sometimes too small - my agent doesn't have enough time to train and stops too early not even reaching the goal.

pseudo-rnd-thoughts commented 2 years ago

LGTM