Imitation Learning

[xkianteb]

Is there a problem with the gym reseting ?

I see a commend in launch_env that says you dont want the env to reset? (https://github.com/duckietown/gym-duckietown/blob/master/learning/utils/env.py)

[rizavelioglu]

It's just to prevent resetting the environment itself. By default, the environment runs for 500.001 time steps, max_steps parameter in Simulator class (see the link you sent).

This parameter is set such a high value so that the environment runs for a relatively high number of (500k) time steps, in the sense that it never stops. But if you would like to reset an environment, the usual way to do so is by calling reset function within Simulator class:

env = Simulator() env.reset()