def thunk():
env = gym.make(gym_id)
env = ImgObsWrapper(env)
env = gym.wrappers.RecordEpisodeStatistics(env)
if args.capture_video:
if idx == 0:
env = Monitor(env, f'videos/{experiment_name}')
env.seed(seed)
env.action_space.seed(seed)
env.observation_space.seed(seed)
return env
#env = gym.wrappers.RecordEpisodeStatistics(env) # records episode reward in `info['episode']['r']`
if args.capture_video:
env = Monitor(env, f'videos/{experiment_name}')
#env = wrap_deepmind(
# env,
# clip_rewards=True,
# frame_stack=True,
# scale=False,
#)
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
torch.backends.cudnn.deterministic = args.torch_deterministic
env.seed(args.seed)
env.action_space.seed(args.seed)
env.observation_space.seed(args.seed)
# respect the default timelimit
assert isinstance(env.action_space,
Discrete), "only discrete action space is supported"
# modified from https://github.com/seungeunrho/minimalRL/blob/master/dqn.py#
class ReplayBuffer():
def __init__(self, buffer_limit):
self.buffer = collections.deque(maxlen=buffer_limit)
def put(self, transition):
self.buffer.append(transition)