def __init__(self, config):
self.config = config
if config.env == 'mountaincar':
self.env = gym.make('MountainCar-v0')
elif config.env == 'acrobot':
self.env = gym.make('Acrobot-v1')
elif 'maze' in config.env:
self.env = maze.MazeEnv(size=config.maze_size, time=100, holes=0, num_goal=1)
self.state_buffer = []
self.reward_buffer = []
self.counter = 0
self.smax = torch.tensor(self.env.observation_space.high)
self.smin = torch.tensor(self.env.observation_space.low)
def build_env_maze(horizon=200, size=20):
import maze
env = maze.MazeEnv(size=size, time=200, holes=0, num_goal=1)
return env
def build_env(args, extra_args):
if 'Lock-v0' in args.env:
# Build combination lock environment
import Environments
env = gym.make('Lock-v0')
ep_dict = {
'horizon': args.horizon,
'dimension': args.dimension,
'switch': 0.1,
'tabular': False
}
env.init(env_config=ep_dict)
return env
elif 'diabcombolock' in args.env:
return build_env_homer(horizon=args.horizon, seed=args.seed)
elif 'maze' in args.env:
import maze
args.maze_size = int(re.findall(r'\d+', args.env)[0])
env = maze.MazeEnv(size=args.maze_size, time=100, holes=0, num_goal=1)
return env
ncpu = multiprocessing.cpu_count()
if sys.platform == 'darwin': ncpu //= 2
nenv = args.num_env or ncpu
alg = args.alg
seed = args.seed
env_type, env_id = get_env_type(args)
if env_type in {'atari', 'retro'}:
if alg == 'deepq':
env = make_env(env_id,
env_type,
seed=seed,
wrapper_kwargs={'frame_stack': True})
elif alg == 'trpo_mpi':
env = make_env(env_id, env_type, seed=seed)
else:
frame_stack_size = 4
env = make_vec_env(env_id,
env_type,
nenv,
seed,
gamestate=args.gamestate,
reward_scale=args.reward_scale)
env = VecFrameStack(env, frame_stack_size)
else:
config = tf.ConfigProto(allow_soft_placement=True,
intra_op_parallelism_threads=1,
inter_op_parallelism_threads=1)
config.gpu_options.allow_growth = True
get_session(config=config)
flatten_dict_observations = alg not in {'her'}
env = make_vec_env(env_id,
env_type,
args.num_env or 1,
seed,
reward_scale=args.reward_scale,
flatten_dict_observations=flatten_dict_observations)
if env_type == 'mujoco':
env = VecNormalize(env, use_tf=True)
return env
def _thunk():
import maze
env = maze.MazeEnv(config=open('config/' + env_id + '.xml'))
return env