12kobe / GymCarla

An OpenAI Gym based wrapper for GymCarla. Imitates the rendering mode of the examples for ease of use, modular design for "easy" customization.

Gym Wrapper for Carla 0.9.6

Dependencies and install

Conda env install

Basic usage

import gym
import gym_carla

env = gym.make( 'CarlaDefault-v0', host='localhost', timestep_limit = 10000, render = True, render_mode = 'human')

for ep in range( 10):
  print( "Episode %d:" % ep)
  done = False
  obs = env.reset()

  # print( "# DEBUG: Done before entering episode", done)

  while not done:
    obs, _, done, _ = env.step( env.action_space.sample())
    print( 'Obs at timestep %d: %s' % (env._current_timestep, 'Not none' if obs is not None else 'None'))

    env.render()

env.close()