search

avisingh599 / imitation-dagger

[Reimplementation Ross et al 2011] An implementation of DAGGER using ConvNets for driving from pixels.
MIT License
71 stars 20 forks source link

Refactor for donkeycar #5

Open GoldenWings opened 5 years ago

GoldenWings commented 5 years ago

I am trying to refactor the code to work with 1/10 car http://donkeycar.com. But i wanted to know few things in gym_torcs.py line 72 action_torcs = client.R.d and line 127 obs = client.S.d is n't those actions are the same or i am missing something?

GoldenWings commented 5 years ago

Also i wanted to know what is the use of rewards i cant see where it is been used except for writing it to desk.

avisingh599 commented 5 years ago

Reward are not really used in any way by the algorithm, they are only used for tracking the progress of the algorithm.

avisingh599 commented 5 years ago

I wrote the code a long time ago, so I don't clearly remember the TORCS interface, but I am guessing that that line 72 is used for applying an action to the car, while line 127 is used for obtaining a complete observation (from which images are later extracted) from the current game state.