def make_atari_environment(
level: str = 'Pong',
sticky_actions: bool = True,
zero_discount_on_life_loss: bool = False,
oar_wrapper: bool = False,
) -> dm_env.Environment:
"""Loads the Atari environment."""
# Internal logic.
version = 'v0' if sticky_actions else 'v4'
level_name = f'{level}NoFrameskip-{version}'
env = gym.make(level_name, full_action_space=True)
wrapper_list = [
wrappers.GymAtariAdapter,
functools.partial(
wrappers.AtariWrapper,
to_float=True,
max_episode_len=108_000,
zero_discount_on_life_loss=zero_discount_on_life_loss,
),
wrappers.SinglePrecisionWrapper,
]
if oar_wrapper:
# E.g. IMPALA and R2D2 use this particular variant.
wrapper_list.append(wrappers.ObservationActionRewardWrapper)
return wrappers.wrap_all(env, wrapper_list)
def make_environmment():
"""Make environment.
Returns:
env (acme.wrappers.single_precision.SinglePrecisionWrapper).
"""
# Create the environment
environment = gym.make("gym_missile_command:missile-command-v0",
custom_config=ENV_CONFIG)
# Add the necessary ALE function
environment.ale = DummyALE()
# Acme processing
environment = wrappers.wrap_all(environment, [
wrappers.GymAtariAdapter,
functools.partial(
wrappers.AtariWrapper,
to_float=True,
max_episode_len=MAX_EPISODE_LEN,
zero_discount_on_life_loss=True,
),
wrappers.SinglePrecisionWrapper,
])
return environment
def make_dqn_atari_environment(task_and_level: str = 'PongNoFrameskip-v4',
evaluation: bool = False) -> dm_env.Environment:
env = gym.make(task_and_level, full_action_space=True)
max_episode_len = 108_000 if evaluation else 50_000
return wrappers.wrap_all(env, [
wrappers.GymAtariAdapter,
functools.partial(
wrappers.AtariWrapper,
to_float=True,
max_episode_len=max_episode_len,
zero_discount_on_life_loss=True,
),
wrappers.SinglePrecisionWrapper,
])
def make_environment(level: str = 'PongNoFrameskip-v4',
oar_wrapper: bool = False) -> dm_env.Environment:
"""Loads the Atari environment."""
env = gym.make(level, full_action_space=True)
# Always use episodes of 108k steps as this is standard, matching the paper.
max_episode_len = 108_000
wrapper_list = [
wrappers.GymAtariAdapter,
functools.partial(
wrappers.AtariWrapper,
to_float=True,
max_episode_len=max_episode_len,
zero_discount_on_life_loss=True,
),
]
if oar_wrapper:
# E.g. IMPALA and R2D2 use this particular variant.
wrapper_list.append(wrappers.ObservationActionRewardWrapper)
wrapper_list.append(wrappers.SinglePrecisionWrapper)
return wrappers.wrap_all(env, wrapper_list)