#
import gym
import numpy as np
from enum import IntEnum
import scipy.ndimage
from rl_coach.graph_managers.graph_manager import ScheduleParameters
from rl_coach.utils import lower_under_to_upper, short_dynamic_import
try:
import roboschool
from OpenGL import GL
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("RoboSchool")
try:
from rl_coach.gym_extensions.continuous import mujoco
except:
from rl_coach.logger import failed_imports
failed_imports.append("GymExtensions")
try:
import pybullet_envs
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("PyBullet")
from typing import Dict, Any, Union
from rl_coach.core_types import RunPhase, EnvironmentSteps
import numpy as np
from rl_coach.filters.observation.observation_move_axis_filter import ObservationMoveAxisFilter
try:
from pysc2 import maps
from pysc2.env import sc2_env
from pysc2.env import available_actions_printer
from pysc2.lib import actions
from pysc2.lib import features
from pysc2.env import environment
from absl import app
from absl import flags
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("PySc2")
from rl_coach.environments.environment import Environment, EnvironmentParameters, LevelSelection
from rl_coach.base_parameters import VisualizationParameters
from rl_coach.spaces import BoxActionSpace, VectorObservationSpace, PlanarMapsObservationSpace, StateSpace, CompoundActionSpace, \
DiscreteActionSpace
from rl_coach.filters.filter import InputFilter, OutputFilter
from rl_coach.filters.observation.observation_rescale_to_size_filter import ObservationRescaleToSizeFilter
from rl_coach.filters.action.linear_box_to_box_map import LinearBoxToBoxMap
from rl_coach.filters.observation.observation_to_uint8_filter import ObservationToUInt8Filter
FLAGS = flags.FLAGS
FLAGS(['coach.py'])
SCREEN_SIZE = 84 # will also impact the action space size
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
try:
import vizdoom
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("ViZDoom")
import os
from enum import Enum
from os import path, environ
from typing import Union, List
import numpy as np
from rl_coach.base_parameters import VisualizationParameters
from rl_coach.environments.environment import Environment, EnvironmentParameters, LevelSelection
from rl_coach.filters.action.full_discrete_action_space_map import FullDiscreteActionSpaceMap
from rl_coach.filters.filter import InputFilter, OutputFilter
from rl_coach.filters.observation.observation_rescale_to_size_filter import ObservationRescaleToSizeFilter
from rl_coach.filters.observation.observation_rgb_to_y_filter import ObservationRGBToYFilter
from rl_coach.filters.observation.observation_stacking_filter import ObservationStackingFilter
# See the License for the specific language governing permissions and
# limitations under the License.
#
import random
from enum import Enum
from typing import Union
import numpy as np
try:
from dm_control import suite
from dm_control.suite.wrappers import pixels
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("DeepMind Control Suite")
from rl_coach.base_parameters import VisualizationParameters
from rl_coach.environments.environment import Environment, EnvironmentParameters, LevelSelection
from rl_coach.filters.filter import NoInputFilter, NoOutputFilter
from rl_coach.spaces import BoxActionSpace, ImageObservationSpace, VectorObservationSpace, StateSpace
class ObservationType(Enum):
Measurements = 1
Image = 2
Image_and_Measurements = 3
# Parameters
class ControlSuiteEnvironmentParameters(EnvironmentParameters):
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
from typing import List, Tuple
from rl_coach.base_parameters import Frameworks, AgentParameters
from rl_coach.logger import failed_imports
from rl_coach.spaces import SpacesDefinition
try:
import tensorflow as tf
from rl_coach.architectures.tensorflow_components.general_network import GeneralTensorFlowNetwork
except ImportError:
failed_imports.append("TensorFlow")
class NetworkWrapper(object):
"""
Contains multiple networks and managers syncing and gradient updates
between them.
"""
def __init__(self,
agent_parameters: AgentParameters,
has_target: bool,
has_global: bool,
name: str,
spaces: SpacesDefinition,
replicated_device=None,
worker_device=None):
if 'CARLA_ROOT' in environ:
sys.path.append(path.join(environ.get('CARLA_ROOT'), 'PythonClient'))
else:
screen.error(
"CARLA_ROOT was not defined. Please set it to point to the CARLA root directory and try again."
)
from carla.client import CarlaClient
from carla.settings import CarlaSettings
from carla.tcp import TCPConnectionError
from carla.sensor import Camera
from carla.client import VehicleControl
from carla.planner.planner import Planner
from carla.driving_benchmark.experiment_suites.experiment_suite import ExperimentSuite
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("CARLA")
import logging
import subprocess
from rl_coach.environments.environment import Environment, EnvironmentParameters, LevelSelection
from rl_coach.spaces import BoxActionSpace, ImageObservationSpace, StateSpace, \
VectorObservationSpace
from rl_coach.utils import get_open_port, force_list
from enum import Enum
import os
import signal
from typing import List, Union
from rl_coach.base_parameters import VisualizationParameters
from rl_coach.filters.filter import InputFilter, NoOutputFilter
from rl_coach.filters.observation.observation_rescale_to_size_filter import ObservationRescaleToSizeFilter
from rl_coach.filters.observation.observation_stacking_filter import ObservationStackingFilter
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
import os
import pickle
import numpy as np
try:
import annoy
from annoy import AnnoyIndex
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("annoy")
class AnnoyDictionary(object):
def __init__(self,
dict_size,
key_width,
new_value_shift_coefficient=0.1,
batch_size=100,
key_error_threshold=0.01,
num_neighbors=50,
override_existing_keys=True,
rebuild_on_every_update=False):
self.rebuild_on_every_update = rebuild_on_every_update
self.max_size = dict_size
self.curr_size = 0
# See the License for the specific language governing permissions and
# limitations under the License.
#
from typing import List, Tuple
from rl_coach.base_parameters import Frameworks, AgentParameters
from rl_coach.logger import failed_imports
from rl_coach.saver import SaverCollection
from rl_coach.spaces import SpacesDefinition
from rl_coach.utils import force_list
try:
import tensorflow as tf
from rl_coach.architectures.tensorflow_components.general_network import GeneralTensorFlowNetwork
except ImportError:
failed_imports.append("tensorflow")
try:
import mxnet as mx
from rl_coach.architectures.mxnet_components.general_network import GeneralMxnetNetwork
except ImportError:
failed_imports.append("mxnet")
class NetworkWrapper(object):
"""
The network wrapper contains multiple copies of the same network, each one with a different set of weights which is
updating in a different time scale. The network wrapper will always contain an online network.
It will contain an additional slow updating target network if it was requested by the user,
and it will contain a global network shared between different workers, if Coach is run in a single-node
multi-process distributed mode. The network wrapper contains functionality for managing these networks and syncing
# limitations under the License.
#
from typing import Union ,Dict, Any
from enum import Enum, Flag, auto
from copy import deepcopy
import numpy as np
import random
from collections import namedtuple
try:
import robosuite
from robosuite.wrappers import Wrapper, DomainRandomizationWrapper
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("Robosuite")
from rl_coach.base_parameters import Parameters, VisualizationParameters
from rl_coach.environments.environment import Environment, EnvironmentParameters, LevelSelection
from rl_coach.spaces import BoxActionSpace, VectorObservationSpace, StateSpace, PlanarMapsObservationSpace
# Importing our custom Robosuite environments here so that they are properly
# registered in Robosuite, and so recognized by 'robosuite.make()' and included
# in 'robosuite.ALL_ENVIRONMENTS'
import rl_coach.environments.robosuite.cube_exp
robosuite_environments = list(robosuite.ALL_ENVIRONMENTS)
robosuite_robots = list(robosuite.ALL_ROBOTS)
robosuite_controllers = list(robosuite.ALL_CONTROLLERS)
try:
import deepmind_lab
except ImportError:
from rl_coach.logger import failed_imports
failed_imports.append("deepmind_lab")
import os
import gym
import random
from enum import Enum
from os import path, environ
from typing import Union, List
import numpy as np
from rl_coach.base_parameters import VisualizationParameters
from rl_coach.environments.environment import Environment, EnvironmentParameters, LevelSelection
from rl_coach.spaces import DiscreteActionSpace
from rl_coach.filters.action.full_discrete_action_space_map import FullDiscreteActionSpaceMap
from rl_coach.filters.filter import InputFilter, OutputFilter
from rl_coach.filters.observation.observation_rescale_to_size_filter import ObservationRescaleToSizeFilter
from rl_coach.filters.observation.observation_rgb_to_y_filter import ObservationRGBToYFilter
from rl_coach.filters.observation.observation_stacking_filter import ObservationStackingFilter
from rl_coach.filters.observation.observation_to_uint8_filter import ObservationToUInt8Filter
from rl_coach.filters.reward.reward_clipping_filter import RewardClippingFilter
from rl_coach.filters.filter import NoOutputFilter, NoInputFilter
from rl_coach.spaces import MultiSelectActionSpace, ImageObservationSpace, VectorObservationSpace, StateSpace, ActionType
from pygame import locals
level_scripts = [
