search

bradyz / 2020_CARLA_challenge

"Learning by Cheating" (CoRL 2019) submission for the 2020 CARLA Challenge
181 stars 49 forks source link

RawController And PidController #25

Closed omerXfaruq closed 3 years ago

omerXfaruq commented 3 years ago

Hello there;

As far as I can see training in carla_project/src/image_model and carla_project/src/map_model is done with RawController which is a small network. However benchmark in leaderboard/team_code/image_agent uses PIDController. What is the reason for difference between training controller and benchmark controller?

Thank you for your attention;

bradyz commented 3 years ago

ah those were some old experiments to see which output would be better - low level actions or mid level waypoints

our submission uses the waypoints, so you can simply ignore the rawcontroller

omerXfaruq commented 3 years ago

Ok, thanks!

omerXfaruq commented 3 years ago

@bradyz a question still lingers in my mind, as far as I can see rawController is used in image_model.py while training, and it also affects the loss. So is training done with a different controller, and benchmark is using pid controller? Is the current state of the training part is correct?

bradyz commented 3 years ago

No problem, for training, just set command_coefficient = 0