Python code reformat via [`black`](https://github.com/ambv/black).

Features:
- Reformat code via black.
- Adding circleci configurations.
- Add contribution guidelines.

Steps to reproduce:
- `pip install black`
- `black <source code directory>`

.circleci/config.yml

-version: 2.1
+version: 2.0
 
 jobs:
   build:
             . venv/bin/activate
             cd ml-agents-envs && pip install -e .
             cd ../ml-agents && pip install -e .
-            pip install pytest-cov==2.6.1 codacy-coverage==1.3.11
+            pip install black pytest-cov==2.6.1 codacy-coverage==1.3.11
             cd ../gym-unity && pip install -e .
 
       - save_cache:
-        
+
       - run:
           name: Run Tests for ml-agents and gym_unity
           command: |
             python-codacy-coverage -r coverage.xml
+            black --check ml-agents
+            black --check ml-agents-envs
+            black --check gym-unity
 
       - store_test_results:
           path: test-reports

CONTRIBUTING.md

 
 ## Style Guide
 
-When performing changes to the codebase, ensure that you follow the style guide
-of the file you're modifying. For Python, we follow
-[PEP 8](https://www.python.org/dev/peps/pep-0008/).
-For C#, we will soon be adding a formal style guide for our repository.
+When performing changes to the codebase, please ensure that all python code is reformatted using the [black](https://github.com/ambv/black) formatter. For C#, we will soon be requirements for style and formatting.

gym-unity/gym_unity/envs/unity_env.py

     """
     Any error related to the gym wrapper of ml-agents.
     """
+
     pass
 
 
     https://github.com/openai/multiagent-particle-envs
     """
 
-    def __init__(self, environment_filename: str, worker_id=0, use_visual=False, uint8_visual=False, multiagent=False, flatten_branched=False):
+    def __init__(
+        self,
+        environment_filename: str,
+        worker_id=0,
+        use_visual=False,
+        uint8_visual=False,
+        multiagent=False,
+        flatten_branched=False,
+    ):
         """
         Environment initialization
         :param environment_filename: The UnityEnvironment path or file to be wrapped in the gym.
         self._n_agents = None
         self._multiagent = multiagent
         self._flattener = None
-        self.game_over = False # Hidden flag used by Atari environments to determine if the game is over
+        self.game_over = (
+            False
+        )  # Hidden flag used by Atari environments to determine if the game is over
-                "if it is wrapped in a gym.")
+                "if it is wrapped in a gym."
+            )
-            raise UnityGymException("`use_visual` was set to True, however there are no"
-                                    " visual observations as part of this environment.")
+            raise UnityGymException(
+                "`use_visual` was set to True, however there are no"
+                " visual observations as part of this environment."
+            )
-            logger.warning("`uint8_visual was set to true, but visual observations are not in use. "
-                           "This setting will not have any effect.")
+            logger.warning(
+                "`uint8_visual was set to true, but visual observations are not in use. "
+                "This setting will not have any effect."
+            )
-            logger.warning("The environment contains more than one visual observation. "
-                           "Please note that only the first will be provided in the observation.")
+            logger.warning(
+                "The environment contains more than one visual observation. "
+                "Please note that only the first will be provided in the observation."
+            )
-                "if it is wrapped in a gym.")
+                "if it is wrapped in a gym."
+            )
 
         # Check for number of agents in scene.
         initial_info = self._env.reset()[self.brain_name]
                     self._flattener = ActionFlattener(brain.vector_action_space_size)
                     self._action_space = self._flattener.action_space
                 else:
-                    self._action_space = spaces.MultiDiscrete(brain.vector_action_space_size)
+                    self._action_space = spaces.MultiDiscrete(
+                        brain.vector_action_space_size
+                    )
-                logger.warning("The environment has a non-discrete action space. It will "
-                                "not be flattened.")
+                logger.warning(
+                    "The environment has a non-discrete action space. It will "
+                    "not be flattened."
+                )
             high = np.array([1] * brain.vector_action_space_size[0])
             self._action_space = spaces.Box(-high, high, dtype=np.float32)
         high = np.array([np.inf] * brain.vector_observation_space_size)
                 depth = 1
             else:
                 depth = 3
-            self._observation_space = spaces.Box(0, 1, dtype=np.float32,
-                                                 shape=(brain.camera_resolutions[0]["height"],
-                                                        brain.camera_resolutions[0]["width"],
-                                                        depth))
+            self._observation_space = spaces.Box(
+                0,
+                1,
+                dtype=np.float32,
+                shape=(
+                    brain.camera_resolutions[0]["height"],
+                    brain.camera_resolutions[0]["width"],
+                    depth,
+                ),
+            )
         else:
             self._observation_space = spaces.Box(-high, high, dtype=np.float32)
 
         # Use random actions for all other agents in environment.
         if self._multiagent:
             if not isinstance(action, list):
-                raise UnityGymException("The environment was expecting `action` to be a list.")
+                raise UnityGymException(
+                    "The environment was expecting `action` to be a list."
+                )
-                    "The environment was expecting a list of {} actions.".format(self._n_agents))
+                    "The environment was expecting a list of {} actions.".format(
+                        self._n_agents
+                    )
+                )
             else:
                 if self._flattener is not None:
                     # Action space is discrete and flattened - we expect a list of scalars
 
     def _single_step(self, info):
         if self.use_visual:
-            self.visual_obs = self._preprocess_single(info.visual_observations[0][0, :, :, :])
+            self.visual_obs = self._preprocess_single(
+                info.visual_observations[0][0, :, :, :]
+            )
-        return default_observation, info.rewards[0], info.local_done[0], {
-            "text_observation": info.text_observations[0],
-            "brain_info": info}
+        return (
+            default_observation,
+            info.rewards[0],
+            info.local_done[0],
+            {"text_observation": info.text_observations[0], "brain_info": info},
+        )
-            return (255.0*single_visual_obs).astype(np.uint8)
+            return (255.0 * single_visual_obs).astype(np.uint8)
         else:
             return single_visual_obs
 
             default_observation = self.visual_obs
         else:
             default_observation = info.vector_observations
-        return list(default_observation), info.rewards, info.local_done, {
-            "text_observation": info.text_observations,
-            "brain_info": info}
-    
+        return (
+            list(default_observation),
+            info.rewards,
+            info.local_done,
+            {"text_observation": info.text_observations, "brain_info": info},
+        )
+
-            return [(255.0*_visual_obs).astype(np.uint8) for _visual_obs in multiple_visual_obs]
+            return [
+                (255.0 * _visual_obs).astype(np.uint8)
+                for _visual_obs in multiple_visual_obs
+            ]
-    def render(self, mode='rgb_array'):
+    def render(self, mode="rgb_array"):
         return self.visual_obs
 
     def close(self):
         if not self._multiagent and n_agents > 1:
             raise UnityGymException(
                 "The environment was launched as a single-agent environment, however"
-                "there is more than one agent in the scene.")
+                "there is more than one agent in the scene."
+            )
-                "there is only one agent in the scene.")
+                "there is only one agent in the scene."
+            )
-            raise UnityGymException("The number of agents in the environment has changed since "
-                                    "initialization. This is not supported.")
+            raise UnityGymException(
+                "The number of agents in the environment has changed since "
+                "initialization. This is not supported."
+            )
-        return {'render.modes': ['rgb_array']}
+        return {"render.modes": ["rgb_array"]}
-        return -float('inf'), float('inf')
+        return -float("inf"), float("inf")
 
     @property
     def spec(self):
     def number_agents(self):
         return self._n_agents
 
-class ActionFlattener():
+
+class ActionFlattener:
-    def __init__(self,branched_action_space):
+
+    def __init__(self, branched_action_space):
         """
         Initialize the flattener.
         :param branched_action_space: A List containing the sizes of each branch of the action
         possible_vals = [range(_num) for _num in branched_action_space]
         all_actions = [list(_action) for _action in itertools.product(*possible_vals)]
         # Dict should be faster than List for large action spaces
-        action_lookup = {_scalar: _action for (_scalar, _action) in enumerate(all_actions)}
+        action_lookup = {
+            _scalar: _action for (_scalar, _action) in enumerate(all_actions)
+        }
         return action_lookup
 
     def lookup_action(self, action):

gym-unity/gym_unity/tests/test_gym.py

 from gym_unity.envs import UnityEnv, UnityGymException
 
 
-@mock.patch('gym_unity.envs.unity_env.UnityEnvironment')
+@mock.patch("gym_unity.envs.unity_env.UnityEnvironment")
-    env = UnityEnv(' ', use_visual=False, multiagent=False)
+    env = UnityEnv(" ", use_visual=False, multiagent=False)
     assert isinstance(env, UnityEnv)
     assert isinstance(env.reset(), np.ndarray)
     actions = env.action_space.sample()
     assert isinstance(info, dict)
 
 
-@mock.patch('gym_unity.envs.unity_env.UnityEnvironment')
+@mock.patch("gym_unity.envs.unity_env.UnityEnvironment")
 def test_multi_agent(mock_env):
     mock_brain = create_mock_brainparams()
     mock_braininfo = create_mock_vector_braininfo(num_agents=2)
-        UnityEnv(' ', multiagent=False)
+        UnityEnv(" ", multiagent=False)
-    env = UnityEnv(' ', use_visual=False, multiagent=True)
+    env = UnityEnv(" ", use_visual=False, multiagent=True)
     assert isinstance(env.reset(), list)
     actions = [env.action_space.sample() for i in range(env.number_agents)]
     obs, rew, done, info = env.step(actions)
     assert isinstance(info, dict)
 
 
-@mock.patch('gym_unity.envs.unity_env.UnityEnvironment')
+@mock.patch("gym_unity.envs.unity_env.UnityEnvironment")
-    mock_brain = create_mock_brainparams(vector_action_space_type='discrete', vector_action_space_size=[2,2,3])
+    mock_brain = create_mock_brainparams(
+        vector_action_space_type="discrete", vector_action_space_size=[2, 2, 3]
+    )
-    env = UnityEnv(' ', use_visual=False, multiagent=False, flatten_branched=True)
+    env = UnityEnv(" ", use_visual=False, multiagent=False, flatten_branched=True)
-    assert env.action_space.n==12
-    assert env._flattener.lookup_action(0)==[0,0,0]
-    assert env._flattener.lookup_action(11)==[1,1,2]
+    assert env.action_space.n == 12
+    assert env._flattener.lookup_action(0) == [0, 0, 0]
+    assert env._flattener.lookup_action(11) == [1, 1, 2]
-    env = UnityEnv(' ', use_visual=False, multiagent=False, flatten_branched=False)
+    env = UnityEnv(" ", use_visual=False, multiagent=False, flatten_branched=False)
+
-def create_mock_brainparams(number_visual_observations=0, num_stacked_vector_observations=1,
-                            vector_action_space_type='continuous', vector_observation_space_size=3,
-                            vector_action_space_size=None):
+def create_mock_brainparams(
+    number_visual_observations=0,
+    num_stacked_vector_observations=1,
+    vector_action_space_type="continuous",
+    vector_observation_space_size=3,
+    vector_action_space_size=None,
+):
     """
     Creates a mock BrainParameters object with parameters.
     """
     mock_brain = mock.Mock()
     mock_brain.return_value.number_visual_observations = number_visual_observations
-    mock_brain.return_value.num_stacked_vector_observations = num_stacked_vector_observations
+    mock_brain.return_value.num_stacked_vector_observations = (
+        num_stacked_vector_observations
+    )
-    mock_brain.return_value.vector_observation_space_size = vector_observation_space_size
+    mock_brain.return_value.vector_observation_space_size = (
+        vector_observation_space_size
+    )
-def create_mock_vector_braininfo(num_agents = 1):
+def create_mock_vector_braininfo(num_agents=1):
     """
     Creates a mock BrainInfo with vector observations. Imitates constant
     vector observations, rewards, dones, and agents.
     mock_braininfo = mock.Mock()
-    mock_braininfo.return_value.vector_observations = np.array([num_agents*[1, 2, 3,]])
-    mock_braininfo.return_value.rewards = num_agents*[1.0]
-    mock_braininfo.return_value.local_done = num_agents*[False]
-    mock_braininfo.return_value.text_observations = num_agents*['']
-    mock_braininfo.return_value.agents = range(0,num_agents)
+    mock_braininfo.return_value.vector_observations = np.array([num_agents * [1, 2, 3]])
+    mock_braininfo.return_value.rewards = num_agents * [1.0]
+    mock_braininfo.return_value.local_done = num_agents * [False]
+    mock_braininfo.return_value.text_observations = num_agents * [""]
+    mock_braininfo.return_value.agents = range(0, num_agents)
     return mock_braininfo()
 
 
     :Mock mock_brain: A mock Brain object that specifies the params of this environment.
     :Mock mock_braininfo: A mock BrainInfo object that will be returned at each step and reset.
     """
-    mock_env.return_value.academy_name = 'MockAcademy'
-    mock_env.return_value.brains = {'MockBrain':mock_brain}
-    mock_env.return_value.external_brain_names = ['MockBrain']
-    mock_env.return_value.reset.return_value = {'MockBrain':mock_braininfo}
-    mock_env.return_value.step.return_value = {'MockBrain':mock_braininfo}
+    mock_env.return_value.academy_name = "MockAcademy"
+    mock_env.return_value.brains = {"MockBrain": mock_brain}
+    mock_env.return_value.external_brain_names = ["MockBrain"]
+    mock_env.return_value.reset.return_value = {"MockBrain": mock_braininfo}
+    mock_env.return_value.step.return_value = {"MockBrain": mock_braininfo}

gym-unity/setup.py

 
 from setuptools import setup, find_packages
 
-setup(name='gym_unity',
-      version='0.4.1',
-      description='Unity Machine Learning Agents Gym Interface',
-      license='Apache License 2.0',
-      author='Unity Technologies',
-      author_email='ML-Agents@unity3d.com',
-      url='https://github.com/Unity-Technologies/ml-agents',
-      packages=find_packages(),
-      install_requires=['gym', 'mlagents_envs==0.8.1']
-      )
+setup(
+    name="gym_unity",
+    version="0.4.1",
+    description="Unity Machine Learning Agents Gym Interface",
+    license="Apache License 2.0",
+    author="Unity Technologies",
+    author_email="ML-Agents@unity3d.com",
+    url="https://github.com/Unity-Technologies/ml-agents",
+    packages=find_packages(),
+    install_requires=["gym", "mlagents_envs==0.8.1"],
+)

ml-agents-envs/mlagents/envs/base_unity_environment.py

 
 class BaseUnityEnvironment(ABC):
     @abstractmethod
-    def step(self, vector_action=None, memory=None, text_action=None, value=None) -> AllBrainInfo:
+    def step(
+        self, vector_action=None, memory=None, text_action=None, value=None
+    ) -> AllBrainInfo:
         pass
 
     @abstractmethod

ml-agents-envs/mlagents/envs/brain.py

 
 
 class BrainInfo:
-    def __init__(self, visual_observation, vector_observation, text_observations, memory=None,
-                 reward=None, agents=None, local_done=None,
-                 vector_action=None, text_action=None, max_reached=None, action_mask=None,
-                 custom_observations=None):
+    def __init__(
+        self,
+        visual_observation,
+        vector_observation,
+        text_observations,
+        memory=None,
+        reward=None,
+        agents=None,
+        local_done=None,
+        vector_action=None,
+        text_action=None,
+        max_reached=None,
+        action_mask=None,
+        custom_observations=None,
+    ):
         """
         Describes experience at current step of all agents linked to a brain.
         """
     def merge(self, other):
         for i in range(len(self.visual_observations)):
             self.visual_observations[i].extend(other.visual_observations[i])
-        self.vector_observations = np.append(self.vector_observations, other.vector_observations, axis=0)
+        self.vector_observations = np.append(
+            self.vector_observations, other.vector_observations, axis=0
+        )
-        self.memories = self.merge_memories(self.memories, other.memories, self.agents, other.agents)
+        self.memories = self.merge_memories(
+            self.memories, other.memories, self.agents, other.agents
+        )
         self.rewards = safe_concat_lists(self.rewards, other.rewards)
         self.local_done = safe_concat_lists(self.local_done, other.local_done)
         self.max_reached = safe_concat_lists(self.max_reached, other.max_reached)
         self.previous_text_actions = safe_concat_lists(
             self.previous_text_actions, other.previous_text_actions
         )
-        self.action_masks = safe_concat_np_ndarray(self.action_masks, other.action_masks)
-        self.custom_observations = safe_concat_lists(self.custom_observations, other.custom_observations)
+        self.action_masks = safe_concat_np_ndarray(
+            self.action_masks, other.action_masks
+        )
+        self.custom_observations = safe_concat_lists(
+            self.custom_observations, other.custom_observations
+        )
 
     @staticmethod
     def merge_memories(m1, m2, agents1, agents2):
             m2 = np.zeros((len(agents2), m1.shape[1]))
         elif m2.shape[1] > m1.shape[1]:
             new_m1 = np.zeros((m1.shape[0], m2.shape[1]))
-            new_m1[0:m1.shape[0], 0:m1.shape[1]] = m1
+            new_m1[0 : m1.shape[0], 0 : m1.shape[1]] = m1
-            new_m2[0:m2.shape[0], 0:m2.shape[1]] = m2
+            new_m2[0 : m2.shape[0], 0 : m2.shape[1]] = m2
             return np.append(m1, new_m2, axis=0)
         return np.append(m1, m2, axis=0)
 
         """
         vis_obs = []
         for i in range(brain_params.number_visual_observations):
-            obs = [BrainInfo.process_pixels(x.visual_observations[i],
-                                            brain_params.camera_resolutions[i]['blackAndWhite'])
-                   for x in agent_info_list]
+            obs = [
+                BrainInfo.process_pixels(
+                    x.visual_observations[i],
+                    brain_params.camera_resolutions[i]["blackAndWhite"],
+                )
+                for x in agent_info_list
+            ]
             vis_obs += [obs]
         if len(agent_info_list) == 0:
             memory_size = 0
             memory = np.zeros((0, 0))
         else:
-            [x.memories.extend([0] * (memory_size - len(x.memories))) for x in agent_info_list]
+            [
+                x.memories.extend([0] * (memory_size - len(x.memories)))
+                for x in agent_info_list
+            ]
             memory = np.array([list(x.memories) for x in agent_info_list])
         total_num_actions = sum(brain_params.vector_action_space_size)
         mask_actions = np.ones((len(agent_info_list), total_num_actions))
                     mask_actions[agent_index, :] = [
-                        0 if agent_info.action_mask[k] else 1 for k in range(total_num_actions)]
+                        0 if agent_info.action_mask[k] else 1
+                        for k in range(total_num_actions)
+                    ]
-            logger.warning("An agent had a NaN reward for brain " + brain_params.brain_name)
+            logger.warning(
+                "An agent had a NaN reward for brain " + brain_params.brain_name
+            )
-            logger.warning("An agent had a NaN observation for brain " + brain_params.brain_name)
+            logger.warning(
+                "An agent had a NaN observation for brain " + brain_params.brain_name
+            )
-                (0, brain_params.vector_observation_space_size * brain_params.num_stacked_vector_observations)
+                (
+                    0,
+                    brain_params.vector_observation_space_size
+                    * brain_params.num_stacked_vector_observations,
+                )
             )
         else:
             vector_obs = np.nan_to_num(
             text_action=[list(x.stored_text_actions) for x in agent_info_list],
             max_reached=[x.max_step_reached for x in agent_info_list],
             custom_observations=[x.custom_observation for x in agent_info_list],
-            action_mask=mask_actions
+            action_mask=mask_actions,
         )
         return brain_info
 
 
 
 class BrainParameters:
-    def __init__(self,
-                 brain_name: str,
-                 vector_observation_space_size: int,
-                 num_stacked_vector_observations: int,
-                 camera_resolutions: List[Dict],
-                 vector_action_space_size: List[int],
-                 vector_action_descriptions: List[str],
-                 vector_action_space_type: int):
+    def __init__(
+        self,
+        brain_name: str,
+        vector_observation_space_size: int,
+        num_stacked_vector_observations: int,
+        camera_resolutions: List[Dict],
+        vector_action_space_size: List[int],
+        vector_action_descriptions: List[str],
+        vector_action_space_type: int,
+    ):
         """
         Contains all brain-specific parameters.
         """
         self.camera_resolutions = camera_resolutions
         self.vector_action_space_size = vector_action_space_size
         self.vector_action_descriptions = vector_action_descriptions
-        self.vector_action_space_type = ["discrete", "continuous"][vector_action_space_type]
+        self.vector_action_space_type = ["discrete", "continuous"][
+            vector_action_space_type
+        ]
-        return '''Unity brain name: {}
+        return """Unity brain name: {}
-        Vector Action descriptions: {}'''.format(self.brain_name,
-                                                 str(self.number_visual_observations),
-                                                 str(self.vector_observation_space_size),
-                                                 str(self.num_stacked_vector_observations),
-                                                 self.vector_action_space_type,
-                                                 str(self.vector_action_space_size),
-                                                 ', '.join(self.vector_action_descriptions))
+        Vector Action descriptions: {}""".format(
+            self.brain_name,
+            str(self.number_visual_observations),
+            str(self.vector_observation_space_size),
+            str(self.num_stacked_vector_observations),
+            self.vector_action_space_type,
+            str(self.vector_action_space_size),
+            ", ".join(self.vector_action_descriptions),
+        )
 
     @staticmethod
     def from_proto(brain_param_proto):
         :return: BrainParameter object.
         """
-        resolution = [{
-            "height": x.height,
-            "width": x.width,
-            "blackAndWhite": x.gray_scale
-        } for x in brain_param_proto.camera_resolutions]
-        brain_params = BrainParameters(brain_param_proto.brain_name,
-                                       brain_param_proto.vector_observation_size,
-                                       brain_param_proto.num_stacked_vector_observations,
-                                       resolution,
-                                       list(brain_param_proto.vector_action_size),
-                                       list(brain_param_proto.vector_action_descriptions),
-                                       brain_param_proto.vector_action_space_type)
-        return brain_params
+        resolution = [
+            {"height": x.height, "width": x.width, "blackAndWhite": x.gray_scale}
+            for x in brain_param_proto.camera_resolutions
+        ]
+        brain_params = BrainParameters(
+            brain_param_proto.brain_name,
+            brain_param_proto.vector_observation_size,
+            brain_param_proto.num_stacked_vector_observations,
+            resolution,
+            list(brain_param_proto.vector_action_size),
+            list(brain_param_proto.vector_action_descriptions),
+            brain_param_proto.vector_action_space_type,
+        )
+        return brain_params

ml-agents-envs/mlagents/envs/communicator.py

         """
         Sends a shutdown signal to the unity environment, and closes the connection.
         """
-

ml-agents-envs/mlagents/envs/communicator_objects/agent_action_proto_pb2.py

 # source: mlagents/envs/communicator_objects/agent_action_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import custom_action_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_custom__action__pb2
+from mlagents.envs.communicator_objects import (
+    custom_action_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_custom__action__pb2,
+)
-  name='mlagents/envs/communicator_objects/agent_action_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n;mlagents/envs/communicator_objects/agent_action_proto.proto\x12\x14\x63ommunicator_objects\x1a\x36mlagents/envs/communicator_objects/custom_action.proto\"\x9c\x01\n\x10\x41gentActionProto\x12\x16\n\x0evector_actions\x18\x01 \x03(\x02\x12\x14\n\x0ctext_actions\x18\x02 \x01(\t\x12\x10\n\x08memories\x18\x03 \x03(\x02\x12\r\n\x05value\x18\x04 \x01(\x02\x12\x39\n\rcustom_action\x18\x05 \x01(\x0b\x32\".communicator_objects.CustomActionB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_custom__action__pb2.DESCRIPTOR,])
-
+    name="mlagents/envs/communicator_objects/agent_action_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n;mlagents/envs/communicator_objects/agent_action_proto.proto\x12\x14\x63ommunicator_objects\x1a\x36mlagents/envs/communicator_objects/custom_action.proto"\x9c\x01\n\x10\x41gentActionProto\x12\x16\n\x0evector_actions\x18\x01 \x03(\x02\x12\x14\n\x0ctext_actions\x18\x02 \x01(\t\x12\x10\n\x08memories\x18\x03 \x03(\x02\x12\r\n\x05value\x18\x04 \x01(\x02\x12\x39\n\rcustom_action\x18\x05 \x01(\x0b\x32".communicator_objects.CustomActionB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_custom__action__pb2.DESCRIPTOR
+    ],
+)
-
-  name='AgentActionProto',
-  full_name='communicator_objects.AgentActionProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='vector_actions', full_name='communicator_objects.AgentActionProto.vector_actions', index=0,
-      number=1, type=2, cpp_type=6, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='text_actions', full_name='communicator_objects.AgentActionProto.text_actions', index=1,
-      number=2, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='memories', full_name='communicator_objects.AgentActionProto.memories', index=2,
-      number=3, type=2, cpp_type=6, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='value', full_name='communicator_objects.AgentActionProto.value', index=3,
-      number=4, type=2, cpp_type=6, label=1,
-      has_default_value=False, default_value=float(0),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='custom_action', full_name='communicator_objects.AgentActionProto.custom_action', index=4,
-      number=5, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=142,
-  serialized_end=298,
+    name="AgentActionProto",
+    full_name="communicator_objects.AgentActionProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="vector_actions",
+            full_name="communicator_objects.AgentActionProto.vector_actions",
+            index=0,
+            number=1,
+            type=2,
+            cpp_type=6,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="text_actions",
+            full_name="communicator_objects.AgentActionProto.text_actions",
+            index=1,
+            number=2,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="memories",
+            full_name="communicator_objects.AgentActionProto.memories",
+            index=2,
+            number=3,
+            type=2,
+            cpp_type=6,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="value",
+            full_name="communicator_objects.AgentActionProto.value",
+            index=3,
+            number=4,
+            type=2,
+            cpp_type=6,
+            label=1,
+            has_default_value=False,
+            default_value=float(0),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="custom_action",
+            full_name="communicator_objects.AgentActionProto.custom_action",
+            index=4,
+            number=5,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=142,
+    serialized_end=298,
-_AGENTACTIONPROTO.fields_by_name['custom_action'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_custom__action__pb2._CUSTOMACTION
-DESCRIPTOR.message_types_by_name['AgentActionProto'] = _AGENTACTIONPROTO
+_AGENTACTIONPROTO.fields_by_name[
+    "custom_action"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_custom__action__pb2._CUSTOMACTION
+)
+DESCRIPTOR.message_types_by_name["AgentActionProto"] = _AGENTACTIONPROTO
-AgentActionProto = _reflection.GeneratedProtocolMessageType('AgentActionProto', (_message.Message,), dict(
-  DESCRIPTOR = _AGENTACTIONPROTO,
-  __module__ = 'mlagents.envs.communicator_objects.agent_action_proto_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.AgentActionProto)
-  ))
+AgentActionProto = _reflection.GeneratedProtocolMessageType(
+    "AgentActionProto",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_AGENTACTIONPROTO,
+        __module__="mlagents.envs.communicator_objects.agent_action_proto_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.AgentActionProto)
+    ),
+)
 _sym_db.RegisterMessage(AgentActionProto)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/agent_info_proto_pb2.py

 # source: mlagents/envs/communicator_objects/agent_info_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import custom_observation_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_custom__observation__pb2
+from mlagents.envs.communicator_objects import (
+    custom_observation_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_custom__observation__pb2,
+)
-  name='mlagents/envs/communicator_objects/agent_info_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n9mlagents/envs/communicator_objects/agent_info_proto.proto\x12\x14\x63ommunicator_objects\x1a;mlagents/envs/communicator_objects/custom_observation.proto\"\xd7\x02\n\x0e\x41gentInfoProto\x12\"\n\x1astacked_vector_observation\x18\x01 \x03(\x02\x12\x1b\n\x13visual_observations\x18\x02 \x03(\x0c\x12\x18\n\x10text_observation\x18\x03 \x01(\t\x12\x1d\n\x15stored_vector_actions\x18\x04 \x03(\x02\x12\x1b\n\x13stored_text_actions\x18\x05 \x01(\t\x12\x10\n\x08memories\x18\x06 \x03(\x02\x12\x0e\n\x06reward\x18\x07 \x01(\x02\x12\x0c\n\x04\x64one\x18\x08 \x01(\x08\x12\x18\n\x10max_step_reached\x18\t \x01(\x08\x12\n\n\x02id\x18\n \x01(\x05\x12\x13\n\x0b\x61\x63tion_mask\x18\x0b \x03(\x08\x12\x43\n\x12\x63ustom_observation\x18\x0c \x01(\x0b\x32\'.communicator_objects.CustomObservationB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_custom__observation__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/agent_info_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n9mlagents/envs/communicator_objects/agent_info_proto.proto\x12\x14\x63ommunicator_objects\x1a;mlagents/envs/communicator_objects/custom_observation.proto"\xd7\x02\n\x0e\x41gentInfoProto\x12"\n\x1astacked_vector_observation\x18\x01 \x03(\x02\x12\x1b\n\x13visual_observations\x18\x02 \x03(\x0c\x12\x18\n\x10text_observation\x18\x03 \x01(\t\x12\x1d\n\x15stored_vector_actions\x18\x04 \x03(\x02\x12\x1b\n\x13stored_text_actions\x18\x05 \x01(\t\x12\x10\n\x08memories\x18\x06 \x03(\x02\x12\x0e\n\x06reward\x18\x07 \x01(\x02\x12\x0c\n\x04\x64one\x18\x08 \x01(\x08\x12\x18\n\x10max_step_reached\x18\t \x01(\x08\x12\n\n\x02id\x18\n \x01(\x05\x12\x13\n\x0b\x61\x63tion_mask\x18\x0b \x03(\x08\x12\x43\n\x12\x63ustom_observation\x18\x0c \x01(\x0b\x32\'.communicator_objects.CustomObservationB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_custom__observation__pb2.DESCRIPTOR
+    ],
+)
-  name='AgentInfoProto',
-  full_name='communicator_objects.AgentInfoProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='stacked_vector_observation', full_name='communicator_objects.AgentInfoProto.stacked_vector_observation', index=0,
-      number=1, type=2, cpp_type=6, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='visual_observations', full_name='communicator_objects.AgentInfoProto.visual_observations', index=1,
-      number=2, type=12, cpp_type=9, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='text_observation', full_name='communicator_objects.AgentInfoProto.text_observation', index=2,
-      number=3, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='stored_vector_actions', full_name='communicator_objects.AgentInfoProto.stored_vector_actions', index=3,
-      number=4, type=2, cpp_type=6, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='stored_text_actions', full_name='communicator_objects.AgentInfoProto.stored_text_actions', index=4,
-      number=5, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='memories', full_name='communicator_objects.AgentInfoProto.memories', index=5,
-      number=6, type=2, cpp_type=6, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='reward', full_name='communicator_objects.AgentInfoProto.reward', index=6,
-      number=7, type=2, cpp_type=6, label=1,
-      has_default_value=False, default_value=float(0),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='done', full_name='communicator_objects.AgentInfoProto.done', index=7,
-      number=8, type=8, cpp_type=7, label=1,
-      has_default_value=False, default_value=False,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='max_step_reached', full_name='communicator_objects.AgentInfoProto.max_step_reached', index=8,
-      number=9, type=8, cpp_type=7, label=1,
-      has_default_value=False, default_value=False,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='id', full_name='communicator_objects.AgentInfoProto.id', index=9,
-      number=10, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='action_mask', full_name='communicator_objects.AgentInfoProto.action_mask', index=10,
-      number=11, type=8, cpp_type=7, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='custom_observation', full_name='communicator_objects.AgentInfoProto.custom_observation', index=11,
-      number=12, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=145,
-  serialized_end=488,
+    name="AgentInfoProto",
+    full_name="communicator_objects.AgentInfoProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="stacked_vector_observation",
+            full_name="communicator_objects.AgentInfoProto.stacked_vector_observation",
+            index=0,
+            number=1,
+            type=2,
+            cpp_type=6,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="visual_observations",
+            full_name="communicator_objects.AgentInfoProto.visual_observations",
+            index=1,
+            number=2,
+            type=12,
+            cpp_type=9,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="text_observation",
+            full_name="communicator_objects.AgentInfoProto.text_observation",
+            index=2,
+            number=3,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="stored_vector_actions",
+            full_name="communicator_objects.AgentInfoProto.stored_vector_actions",
+            index=3,
+            number=4,
+            type=2,
+            cpp_type=6,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="stored_text_actions",
+            full_name="communicator_objects.AgentInfoProto.stored_text_actions",
+            index=4,
+            number=5,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="memories",
+            full_name="communicator_objects.AgentInfoProto.memories",
+            index=5,
+            number=6,
+            type=2,
+            cpp_type=6,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="reward",
+            full_name="communicator_objects.AgentInfoProto.reward",
+            index=6,
+            number=7,
+            type=2,
+            cpp_type=6,
+            label=1,
+            has_default_value=False,
+            default_value=float(0),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="done",
+            full_name="communicator_objects.AgentInfoProto.done",
+            index=7,
+            number=8,
+            type=8,
+            cpp_type=7,
+            label=1,
+            has_default_value=False,
+            default_value=False,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="max_step_reached",
+            full_name="communicator_objects.AgentInfoProto.max_step_reached",
+            index=8,
+            number=9,
+            type=8,
+            cpp_type=7,
+            label=1,
+            has_default_value=False,
+            default_value=False,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="id",
+            full_name="communicator_objects.AgentInfoProto.id",
+            index=9,
+            number=10,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="action_mask",
+            full_name="communicator_objects.AgentInfoProto.action_mask",
+            index=10,
+            number=11,
+            type=8,
+            cpp_type=7,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="custom_observation",
+            full_name="communicator_objects.AgentInfoProto.custom_observation",
+            index=11,
+            number=12,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=145,
+    serialized_end=488,
-_AGENTINFOPROTO.fields_by_name['custom_observation'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_custom__observation__pb2._CUSTOMOBSERVATION
-DESCRIPTOR.message_types_by_name['AgentInfoProto'] = _AGENTINFOPROTO
+_AGENTINFOPROTO.fields_by_name[
+    "custom_observation"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_custom__observation__pb2._CUSTOMOBSERVATION
+)
+DESCRIPTOR.message_types_by_name["AgentInfoProto"] = _AGENTINFOPROTO
-AgentInfoProto = _reflection.GeneratedProtocolMessageType('AgentInfoProto', (_message.Message,), dict(
-  DESCRIPTOR = _AGENTINFOPROTO,
-  __module__ = 'mlagents.envs.communicator_objects.agent_info_proto_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.AgentInfoProto)
-  ))
+AgentInfoProto = _reflection.GeneratedProtocolMessageType(
+    "AgentInfoProto",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_AGENTINFOPROTO,
+        __module__="mlagents.envs.communicator_objects.agent_info_proto_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.AgentInfoProto)
+    ),
+)
 _sym_db.RegisterMessage(AgentInfoProto)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/brain_parameters_proto_pb2.py

 # source: mlagents/envs/communicator_objects/brain_parameters_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import resolution_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2
-from mlagents.envs.communicator_objects import space_type_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_space__type__proto__pb2
+from mlagents.envs.communicator_objects import (
+    resolution_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    space_type_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_space__type__proto__pb2,
+)
-  name='mlagents/envs/communicator_objects/brain_parameters_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n?mlagents/envs/communicator_objects/brain_parameters_proto.proto\x12\x14\x63ommunicator_objects\x1a\x39mlagents/envs/communicator_objects/resolution_proto.proto\x1a\x39mlagents/envs/communicator_objects/space_type_proto.proto\"\xd4\x02\n\x14\x42rainParametersProto\x12\x1f\n\x17vector_observation_size\x18\x01 \x01(\x05\x12\'\n\x1fnum_stacked_vector_observations\x18\x02 \x01(\x05\x12\x1a\n\x12vector_action_size\x18\x03 \x03(\x05\x12\x41\n\x12\x63\x61mera_resolutions\x18\x04 \x03(\x0b\x32%.communicator_objects.ResolutionProto\x12\"\n\x1avector_action_descriptions\x18\x05 \x03(\t\x12\x46\n\x18vector_action_space_type\x18\x06 \x01(\x0e\x32$.communicator_objects.SpaceTypeProto\x12\x12\n\nbrain_name\x18\x07 \x01(\t\x12\x13\n\x0bis_training\x18\x08 \x01(\x08\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_space__type__proto__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/brain_parameters_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n?mlagents/envs/communicator_objects/brain_parameters_proto.proto\x12\x14\x63ommunicator_objects\x1a\x39mlagents/envs/communicator_objects/resolution_proto.proto\x1a\x39mlagents/envs/communicator_objects/space_type_proto.proto"\xd4\x02\n\x14\x42rainParametersProto\x12\x1f\n\x17vector_observation_size\x18\x01 \x01(\x05\x12\'\n\x1fnum_stacked_vector_observations\x18\x02 \x01(\x05\x12\x1a\n\x12vector_action_size\x18\x03 \x03(\x05\x12\x41\n\x12\x63\x61mera_resolutions\x18\x04 \x03(\x0b\x32%.communicator_objects.ResolutionProto\x12"\n\x1avector_action_descriptions\x18\x05 \x03(\t\x12\x46\n\x18vector_action_space_type\x18\x06 \x01(\x0e\x32$.communicator_objects.SpaceTypeProto\x12\x12\n\nbrain_name\x18\x07 \x01(\t\x12\x13\n\x0bis_training\x18\x08 \x01(\x08\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_space__type__proto__pb2.DESCRIPTOR,
+    ],
+)
-  name='BrainParametersProto',
-  full_name='communicator_objects.BrainParametersProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='vector_observation_size', full_name='communicator_objects.BrainParametersProto.vector_observation_size', index=0,
-      number=1, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='num_stacked_vector_observations', full_name='communicator_objects.BrainParametersProto.num_stacked_vector_observations', index=1,
-      number=2, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='vector_action_size', full_name='communicator_objects.BrainParametersProto.vector_action_size', index=2,
-      number=3, type=5, cpp_type=1, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='camera_resolutions', full_name='communicator_objects.BrainParametersProto.camera_resolutions', index=3,
-      number=4, type=11, cpp_type=10, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='vector_action_descriptions', full_name='communicator_objects.BrainParametersProto.vector_action_descriptions', index=4,
-      number=5, type=9, cpp_type=9, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='vector_action_space_type', full_name='communicator_objects.BrainParametersProto.vector_action_space_type', index=5,
-      number=6, type=14, cpp_type=8, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='brain_name', full_name='communicator_objects.BrainParametersProto.brain_name', index=6,
-      number=7, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='is_training', full_name='communicator_objects.BrainParametersProto.is_training', index=7,
-      number=8, type=8, cpp_type=7, label=1,
-      has_default_value=False, default_value=False,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=208,
-  serialized_end=548,
+    name="BrainParametersProto",
+    full_name="communicator_objects.BrainParametersProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="vector_observation_size",
+            full_name="communicator_objects.BrainParametersProto.vector_observation_size",
+            index=0,
+            number=1,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="num_stacked_vector_observations",
+            full_name="communicator_objects.BrainParametersProto.num_stacked_vector_observations",
+            index=1,
+            number=2,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="vector_action_size",
+            full_name="communicator_objects.BrainParametersProto.vector_action_size",
+            index=2,
+            number=3,
+            type=5,
+            cpp_type=1,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="camera_resolutions",
+            full_name="communicator_objects.BrainParametersProto.camera_resolutions",
+            index=3,
+            number=4,
+            type=11,
+            cpp_type=10,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="vector_action_descriptions",
+            full_name="communicator_objects.BrainParametersProto.vector_action_descriptions",
+            index=4,
+            number=5,
+            type=9,
+            cpp_type=9,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="vector_action_space_type",
+            full_name="communicator_objects.BrainParametersProto.vector_action_space_type",
+            index=5,
+            number=6,
+            type=14,
+            cpp_type=8,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="brain_name",
+            full_name="communicator_objects.BrainParametersProto.brain_name",
+            index=6,
+            number=7,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="is_training",
+            full_name="communicator_objects.BrainParametersProto.is_training",
+            index=7,
+            number=8,
+            type=8,
+            cpp_type=7,
+            label=1,
+            has_default_value=False,
+            default_value=False,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=208,
+    serialized_end=548,
-_BRAINPARAMETERSPROTO.fields_by_name['camera_resolutions'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2._RESOLUTIONPROTO
-_BRAINPARAMETERSPROTO.fields_by_name['vector_action_space_type'].enum_type = mlagents_dot_envs_dot_communicator__objects_dot_space__type__proto__pb2._SPACETYPEPROTO
-DESCRIPTOR.message_types_by_name['BrainParametersProto'] = _BRAINPARAMETERSPROTO
+_BRAINPARAMETERSPROTO.fields_by_name[
+    "camera_resolutions"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2._RESOLUTIONPROTO
+)
+_BRAINPARAMETERSPROTO.fields_by_name[
+    "vector_action_space_type"
+].enum_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_space__type__proto__pb2._SPACETYPEPROTO
+)
+DESCRIPTOR.message_types_by_name["BrainParametersProto"] = _BRAINPARAMETERSPROTO
-BrainParametersProto = _reflection.GeneratedProtocolMessageType('BrainParametersProto', (_message.Message,), dict(
-  DESCRIPTOR = _BRAINPARAMETERSPROTO,
-  __module__ = 'mlagents.envs.communicator_objects.brain_parameters_proto_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.BrainParametersProto)
-  ))
+BrainParametersProto = _reflection.GeneratedProtocolMessageType(
+    "BrainParametersProto",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_BRAINPARAMETERSPROTO,
+        __module__="mlagents.envs.communicator_objects.brain_parameters_proto_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.BrainParametersProto)
+    ),
+)
 _sym_db.RegisterMessage(BrainParametersProto)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/command_proto_pb2.py

 # source: mlagents/envs/communicator_objects/command_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/command_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n6mlagents/envs/communicator_objects/command_proto.proto\x12\x14\x63ommunicator_objects*-\n\x0c\x43ommandProto\x12\x08\n\x04STEP\x10\x00\x12\t\n\x05RESET\x10\x01\x12\x08\n\x04QUIT\x10\x02\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/command_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        "\n6mlagents/envs/communicator_objects/command_proto.proto\x12\x14\x63ommunicator_objects*-\n\x0c\x43ommandProto\x12\x08\n\x04STEP\x10\x00\x12\t\n\x05RESET\x10\x01\x12\x08\n\x04QUIT\x10\x02\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3"
+    ),
-  name='CommandProto',
-  full_name='communicator_objects.CommandProto',
-  filename=None,
-  file=DESCRIPTOR,
-  values=[
-    _descriptor.EnumValueDescriptor(
-      name='STEP', index=0, number=0,
-      serialized_options=None,
-      type=None),
-    _descriptor.EnumValueDescriptor(
-      name='RESET', index=1, number=1,
-      serialized_options=None,
-      type=None),
-    _descriptor.EnumValueDescriptor(
-      name='QUIT', index=2, number=2,
-      serialized_options=None,
-      type=None),
-  ],
-  containing_type=None,
-  serialized_options=None,
-  serialized_start=80,
-  serialized_end=125,
+    name="CommandProto",
+    full_name="communicator_objects.CommandProto",
+    filename=None,
+    file=DESCRIPTOR,
+    values=[
+        _descriptor.EnumValueDescriptor(
+            name="STEP", index=0, number=0, serialized_options=None, type=None
+        ),
+        _descriptor.EnumValueDescriptor(
+            name="RESET", index=1, number=1, serialized_options=None, type=None
+        ),
+        _descriptor.EnumValueDescriptor(
+            name="QUIT", index=2, number=2, serialized_options=None, type=None
+        ),
+    ],
+    containing_type=None,
+    serialized_options=None,
+    serialized_start=80,
+    serialized_end=125,
 )
 _sym_db.RegisterEnumDescriptor(_COMMANDPROTO)
 
 QUIT = 2
 
 
-DESCRIPTOR.enum_types_by_name['CommandProto'] = _COMMANDPROTO
+DESCRIPTOR.enum_types_by_name["CommandProto"] = _COMMANDPROTO
 _sym_db.RegisterFileDescriptor(DESCRIPTOR)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/custom_action_pb2.py

 # source: mlagents/envs/communicator_objects/custom_action.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/custom_action.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n6mlagents/envs/communicator_objects/custom_action.proto\x12\x14\x63ommunicator_objects\"\x0e\n\x0c\x43ustomActionB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/custom_action.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n6mlagents/envs/communicator_objects/custom_action.proto\x12\x14\x63ommunicator_objects"\x0e\n\x0c\x43ustomActionB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='CustomAction',
-  full_name='communicator_objects.CustomAction',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=80,
-  serialized_end=94,
+    name="CustomAction",
+    full_name="communicator_objects.CustomAction",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=80,
+    serialized_end=94,
-DESCRIPTOR.message_types_by_name['CustomAction'] = _CUSTOMACTION
+DESCRIPTOR.message_types_by_name["CustomAction"] = _CUSTOMACTION
-CustomAction = _reflection.GeneratedProtocolMessageType('CustomAction', (_message.Message,), dict(
-  DESCRIPTOR = _CUSTOMACTION,
-  __module__ = 'mlagents.envs.communicator_objects.custom_action_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.CustomAction)
-  ))
+CustomAction = _reflection.GeneratedProtocolMessageType(
+    "CustomAction",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_CUSTOMACTION,
+        __module__="mlagents.envs.communicator_objects.custom_action_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.CustomAction)
+    ),
+)
 _sym_db.RegisterMessage(CustomAction)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/custom_observation_pb2.py

 # source: mlagents/envs/communicator_objects/custom_observation.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/custom_observation.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n;mlagents/envs/communicator_objects/custom_observation.proto\x12\x14\x63ommunicator_objects\"\x13\n\x11\x43ustomObservationB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/custom_observation.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n;mlagents/envs/communicator_objects/custom_observation.proto\x12\x14\x63ommunicator_objects"\x13\n\x11\x43ustomObservationB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='CustomObservation',
-  full_name='communicator_objects.CustomObservation',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=85,
-  serialized_end=104,
+    name="CustomObservation",
+    full_name="communicator_objects.CustomObservation",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=85,
+    serialized_end=104,
-DESCRIPTOR.message_types_by_name['CustomObservation'] = _CUSTOMOBSERVATION
+DESCRIPTOR.message_types_by_name["CustomObservation"] = _CUSTOMOBSERVATION
-CustomObservation = _reflection.GeneratedProtocolMessageType('CustomObservation', (_message.Message,), dict(
-  DESCRIPTOR = _CUSTOMOBSERVATION,
-  __module__ = 'mlagents.envs.communicator_objects.custom_observation_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.CustomObservation)
-  ))
+CustomObservation = _reflection.GeneratedProtocolMessageType(
+    "CustomObservation",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_CUSTOMOBSERVATION,
+        __module__="mlagents.envs.communicator_objects.custom_observation_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.CustomObservation)
+    ),
+)
 _sym_db.RegisterMessage(CustomObservation)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/custom_reset_parameters_pb2.py

 # source: mlagents/envs/communicator_objects/custom_reset_parameters.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/custom_reset_parameters.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n@mlagents/envs/communicator_objects/custom_reset_parameters.proto\x12\x14\x63ommunicator_objects\"\x17\n\x15\x43ustomResetParametersB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/custom_reset_parameters.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n@mlagents/envs/communicator_objects/custom_reset_parameters.proto\x12\x14\x63ommunicator_objects"\x17\n\x15\x43ustomResetParametersB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='CustomResetParameters',
-  full_name='communicator_objects.CustomResetParameters',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=90,
-  serialized_end=113,
+    name="CustomResetParameters",
+    full_name="communicator_objects.CustomResetParameters",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=90,
+    serialized_end=113,
-DESCRIPTOR.message_types_by_name['CustomResetParameters'] = _CUSTOMRESETPARAMETERS
+DESCRIPTOR.message_types_by_name["CustomResetParameters"] = _CUSTOMRESETPARAMETERS
-CustomResetParameters = _reflection.GeneratedProtocolMessageType('CustomResetParameters', (_message.Message,), dict(
-  DESCRIPTOR = _CUSTOMRESETPARAMETERS,
-  __module__ = 'mlagents.envs.communicator_objects.custom_reset_parameters_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.CustomResetParameters)
-  ))
+CustomResetParameters = _reflection.GeneratedProtocolMessageType(
+    "CustomResetParameters",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_CUSTOMRESETPARAMETERS,
+        __module__="mlagents.envs.communicator_objects.custom_reset_parameters_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.CustomResetParameters)
+    ),
+)
 _sym_db.RegisterMessage(CustomResetParameters)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/demonstration_meta_proto_pb2.py

 # source: mlagents/envs/communicator_objects/demonstration_meta_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/demonstration_meta_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\nAmlagents/envs/communicator_objects/demonstration_meta_proto.proto\x12\x14\x63ommunicator_objects\"\x8d\x01\n\x16\x44\x65monstrationMetaProto\x12\x13\n\x0b\x61pi_version\x18\x01 \x01(\x05\x12\x1a\n\x12\x64\x65monstration_name\x18\x02 \x01(\t\x12\x14\n\x0cnumber_steps\x18\x03 \x01(\x05\x12\x17\n\x0fnumber_episodes\x18\x04 \x01(\x05\x12\x13\n\x0bmean_reward\x18\x05 \x01(\x02\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/demonstration_meta_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\nAmlagents/envs/communicator_objects/demonstration_meta_proto.proto\x12\x14\x63ommunicator_objects"\x8d\x01\n\x16\x44\x65monstrationMetaProto\x12\x13\n\x0b\x61pi_version\x18\x01 \x01(\x05\x12\x1a\n\x12\x64\x65monstration_name\x18\x02 \x01(\t\x12\x14\n\x0cnumber_steps\x18\x03 \x01(\x05\x12\x17\n\x0fnumber_episodes\x18\x04 \x01(\x05\x12\x13\n\x0bmean_reward\x18\x05 \x01(\x02\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='DemonstrationMetaProto',
-  full_name='communicator_objects.DemonstrationMetaProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='api_version', full_name='communicator_objects.DemonstrationMetaProto.api_version', index=0,
-      number=1, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='demonstration_name', full_name='communicator_objects.DemonstrationMetaProto.demonstration_name', index=1,
-      number=2, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='number_steps', full_name='communicator_objects.DemonstrationMetaProto.number_steps', index=2,
-      number=3, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='number_episodes', full_name='communicator_objects.DemonstrationMetaProto.number_episodes', index=3,
-      number=4, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='mean_reward', full_name='communicator_objects.DemonstrationMetaProto.mean_reward', index=4,
-      number=5, type=2, cpp_type=6, label=1,
-      has_default_value=False, default_value=float(0),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=92,
-  serialized_end=233,
+    name="DemonstrationMetaProto",
+    full_name="communicator_objects.DemonstrationMetaProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="api_version",
+            full_name="communicator_objects.DemonstrationMetaProto.api_version",
+            index=0,
+            number=1,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="demonstration_name",
+            full_name="communicator_objects.DemonstrationMetaProto.demonstration_name",
+            index=1,
+            number=2,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="number_steps",
+            full_name="communicator_objects.DemonstrationMetaProto.number_steps",
+            index=2,
+            number=3,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="number_episodes",
+            full_name="communicator_objects.DemonstrationMetaProto.number_episodes",
+            index=3,
+            number=4,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="mean_reward",
+            full_name="communicator_objects.DemonstrationMetaProto.mean_reward",
+            index=4,
+            number=5,
+            type=2,
+            cpp_type=6,
+            label=1,
+            has_default_value=False,
+            default_value=float(0),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=92,
+    serialized_end=233,
-DESCRIPTOR.message_types_by_name['DemonstrationMetaProto'] = _DEMONSTRATIONMETAPROTO
+DESCRIPTOR.message_types_by_name["DemonstrationMetaProto"] = _DEMONSTRATIONMETAPROTO
-DemonstrationMetaProto = _reflection.GeneratedProtocolMessageType('DemonstrationMetaProto', (_message.Message,), dict(
-  DESCRIPTOR = _DEMONSTRATIONMETAPROTO,
-  __module__ = 'mlagents.envs.communicator_objects.demonstration_meta_proto_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.DemonstrationMetaProto)
-  ))
+DemonstrationMetaProto = _reflection.GeneratedProtocolMessageType(
+    "DemonstrationMetaProto",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_DEMONSTRATIONMETAPROTO,
+        __module__="mlagents.envs.communicator_objects.demonstration_meta_proto_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.DemonstrationMetaProto)
+    ),
+)
 _sym_db.RegisterMessage(DemonstrationMetaProto)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/engine_configuration_proto_pb2.py

 # source: mlagents/envs/communicator_objects/engine_configuration_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/engine_configuration_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\nCmlagents/envs/communicator_objects/engine_configuration_proto.proto\x12\x14\x63ommunicator_objects\"\x95\x01\n\x18\x45ngineConfigurationProto\x12\r\n\x05width\x18\x01 \x01(\x05\x12\x0e\n\x06height\x18\x02 \x01(\x05\x12\x15\n\rquality_level\x18\x03 \x01(\x05\x12\x12\n\ntime_scale\x18\x04 \x01(\x02\x12\x19\n\x11target_frame_rate\x18\x05 \x01(\x05\x12\x14\n\x0cshow_monitor\x18\x06 \x01(\x08\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/engine_configuration_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\nCmlagents/envs/communicator_objects/engine_configuration_proto.proto\x12\x14\x63ommunicator_objects"\x95\x01\n\x18\x45ngineConfigurationProto\x12\r\n\x05width\x18\x01 \x01(\x05\x12\x0e\n\x06height\x18\x02 \x01(\x05\x12\x15\n\rquality_level\x18\x03 \x01(\x05\x12\x12\n\ntime_scale\x18\x04 \x01(\x02\x12\x19\n\x11target_frame_rate\x18\x05 \x01(\x05\x12\x14\n\x0cshow_monitor\x18\x06 \x01(\x08\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='EngineConfigurationProto',
-  full_name='communicator_objects.EngineConfigurationProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='width', full_name='communicator_objects.EngineConfigurationProto.width', index=0,
-      number=1, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='height', full_name='communicator_objects.EngineConfigurationProto.height', index=1,
-      number=2, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='quality_level', full_name='communicator_objects.EngineConfigurationProto.quality_level', index=2,
-      number=3, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='time_scale', full_name='communicator_objects.EngineConfigurationProto.time_scale', index=3,
-      number=4, type=2, cpp_type=6, label=1,
-      has_default_value=False, default_value=float(0),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='target_frame_rate', full_name='communicator_objects.EngineConfigurationProto.target_frame_rate', index=4,
-      number=5, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='show_monitor', full_name='communicator_objects.EngineConfigurationProto.show_monitor', index=5,
-      number=6, type=8, cpp_type=7, label=1,
-      has_default_value=False, default_value=False,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=94,
-  serialized_end=243,
+    name="EngineConfigurationProto",
+    full_name="communicator_objects.EngineConfigurationProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="width",
+            full_name="communicator_objects.EngineConfigurationProto.width",
+            index=0,
+            number=1,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="height",
+            full_name="communicator_objects.EngineConfigurationProto.height",
+            index=1,
+            number=2,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="quality_level",
+            full_name="communicator_objects.EngineConfigurationProto.quality_level",
+            index=2,
+            number=3,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="time_scale",
+            full_name="communicator_objects.EngineConfigurationProto.time_scale",
+            index=3,
+            number=4,
+            type=2,
+            cpp_type=6,
+            label=1,
+            has_default_value=False,
+            default_value=float(0),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="target_frame_rate",
+            full_name="communicator_objects.EngineConfigurationProto.target_frame_rate",
+            index=4,
+            number=5,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="show_monitor",
+            full_name="communicator_objects.EngineConfigurationProto.show_monitor",
+            index=5,
+            number=6,
+            type=8,
+            cpp_type=7,
+            label=1,
+            has_default_value=False,
+            default_value=False,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=94,
+    serialized_end=243,
-DESCRIPTOR.message_types_by_name['EngineConfigurationProto'] = _ENGINECONFIGURATIONPROTO
+DESCRIPTOR.message_types_by_name["EngineConfigurationProto"] = _ENGINECONFIGURATIONPROTO
-EngineConfigurationProto = _reflection.GeneratedProtocolMessageType('EngineConfigurationProto', (_message.Message,), dict(
-  DESCRIPTOR = _ENGINECONFIGURATIONPROTO,
-  __module__ = 'mlagents.envs.communicator_objects.engine_configuration_proto_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.EngineConfigurationProto)
-  ))
+EngineConfigurationProto = _reflection.GeneratedProtocolMessageType(
+    "EngineConfigurationProto",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_ENGINECONFIGURATIONPROTO,
+        __module__="mlagents.envs.communicator_objects.engine_configuration_proto_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.EngineConfigurationProto)
+    ),
+)
 _sym_db.RegisterMessage(EngineConfigurationProto)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/environment_parameters_proto_pb2.py

 # source: mlagents/envs/communicator_objects/environment_parameters_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import custom_reset_parameters_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_custom__reset__parameters__pb2
+from mlagents.envs.communicator_objects import (
+    custom_reset_parameters_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_custom__reset__parameters__pb2,
+)
-  name='mlagents/envs/communicator_objects/environment_parameters_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\nEmlagents/envs/communicator_objects/environment_parameters_proto.proto\x12\x14\x63ommunicator_objects\x1a@mlagents/envs/communicator_objects/custom_reset_parameters.proto\"\x83\x02\n\x1a\x45nvironmentParametersProto\x12_\n\x10\x66loat_parameters\x18\x01 \x03(\x0b\x32\x45.communicator_objects.EnvironmentParametersProto.FloatParametersEntry\x12L\n\x17\x63ustom_reset_parameters\x18\x02 \x01(\x0b\x32+.communicator_objects.CustomResetParameters\x1a\x36\n\x14\x46loatParametersEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x02:\x02\x38\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_custom__reset__parameters__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/environment_parameters_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\nEmlagents/envs/communicator_objects/environment_parameters_proto.proto\x12\x14\x63ommunicator_objects\x1a@mlagents/envs/communicator_objects/custom_reset_parameters.proto"\x83\x02\n\x1a\x45nvironmentParametersProto\x12_\n\x10\x66loat_parameters\x18\x01 \x03(\x0b\x32\x45.communicator_objects.EnvironmentParametersProto.FloatParametersEntry\x12L\n\x17\x63ustom_reset_parameters\x18\x02 \x01(\x0b\x32+.communicator_objects.CustomResetParameters\x1a\x36\n\x14\x46loatParametersEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x02:\x02\x38\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_custom__reset__parameters__pb2.DESCRIPTOR
+    ],
+)
-  name='FloatParametersEntry',
-  full_name='communicator_objects.EnvironmentParametersProto.FloatParametersEntry',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='key', full_name='communicator_objects.EnvironmentParametersProto.FloatParametersEntry.key', index=0,
-      number=1, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='value', full_name='communicator_objects.EnvironmentParametersProto.FloatParametersEntry.value', index=1,
-      number=2, type=2, cpp_type=6, label=1,
-      has_default_value=False, default_value=float(0),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=_b('8\001'),
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=367,
-  serialized_end=421,
+    name="FloatParametersEntry",
+    full_name="communicator_objects.EnvironmentParametersProto.FloatParametersEntry",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="key",
+            full_name="communicator_objects.EnvironmentParametersProto.FloatParametersEntry.key",
+            index=0,
+            number=1,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="value",
+            full_name="communicator_objects.EnvironmentParametersProto.FloatParametersEntry.value",
+            index=1,
+            number=2,
+            type=2,
+            cpp_type=6,
+            label=1,
+            has_default_value=False,
+            default_value=float(0),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=_b("8\001"),
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=367,
+    serialized_end=421,
-  name='EnvironmentParametersProto',
-  full_name='communicator_objects.EnvironmentParametersProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='float_parameters', full_name='communicator_objects.EnvironmentParametersProto.float_parameters', index=0,
-      number=1, type=11, cpp_type=10, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='custom_reset_parameters', full_name='communicator_objects.EnvironmentParametersProto.custom_reset_parameters', index=1,
-      number=2, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[_ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY, ],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=162,
-  serialized_end=421,
+    name="EnvironmentParametersProto",
+    full_name="communicator_objects.EnvironmentParametersProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="float_parameters",
+            full_name="communicator_objects.EnvironmentParametersProto.float_parameters",
+            index=0,
+            number=1,
+            type=11,
+            cpp_type=10,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="custom_reset_parameters",
+            full_name="communicator_objects.EnvironmentParametersProto.custom_reset_parameters",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[_ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=162,
+    serialized_end=421,
-_ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY.containing_type = _ENVIRONMENTPARAMETERSPROTO
-_ENVIRONMENTPARAMETERSPROTO.fields_by_name['float_parameters'].message_type = _ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY
-_ENVIRONMENTPARAMETERSPROTO.fields_by_name['custom_reset_parameters'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_custom__reset__parameters__pb2._CUSTOMRESETPARAMETERS
-DESCRIPTOR.message_types_by_name['EnvironmentParametersProto'] = _ENVIRONMENTPARAMETERSPROTO
+_ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY.containing_type = (
+    _ENVIRONMENTPARAMETERSPROTO
+)
+_ENVIRONMENTPARAMETERSPROTO.fields_by_name[
+    "float_parameters"
+].message_type = _ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY
+_ENVIRONMENTPARAMETERSPROTO.fields_by_name[
+    "custom_reset_parameters"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_custom__reset__parameters__pb2._CUSTOMRESETPARAMETERS
+)
+DESCRIPTOR.message_types_by_name[
+    "EnvironmentParametersProto"
+] = _ENVIRONMENTPARAMETERSPROTO
-EnvironmentParametersProto = _reflection.GeneratedProtocolMessageType('EnvironmentParametersProto', (_message.Message,), dict(
-
-  FloatParametersEntry = _reflection.GeneratedProtocolMessageType('FloatParametersEntry', (_message.Message,), dict(
-    DESCRIPTOR = _ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY,
-    __module__ = 'mlagents.envs.communicator_objects.environment_parameters_proto_pb2'
-    # @@protoc_insertion_point(class_scope:communicator_objects.EnvironmentParametersProto.FloatParametersEntry)
-    ))
-  ,
-  DESCRIPTOR = _ENVIRONMENTPARAMETERSPROTO,
-  __module__ = 'mlagents.envs.communicator_objects.environment_parameters_proto_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.EnvironmentParametersProto)
-  ))
+EnvironmentParametersProto = _reflection.GeneratedProtocolMessageType(
+    "EnvironmentParametersProto",
+    (_message.Message,),
+    dict(
+        FloatParametersEntry=_reflection.GeneratedProtocolMessageType(
+            "FloatParametersEntry",
+            (_message.Message,),
+            dict(
+                DESCRIPTOR=_ENVIRONMENTPARAMETERSPROTO_FLOATPARAMETERSENTRY,
+                __module__="mlagents.envs.communicator_objects.environment_parameters_proto_pb2"
+                # @@protoc_insertion_point(class_scope:communicator_objects.EnvironmentParametersProto.FloatParametersEntry)
+            ),
+        ),
+        DESCRIPTOR=_ENVIRONMENTPARAMETERSPROTO,
+        __module__="mlagents.envs.communicator_objects.environment_parameters_proto_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.EnvironmentParametersProto)
+    ),
+)
 _sym_db.RegisterMessage(EnvironmentParametersProto)
 _sym_db.RegisterMessage(EnvironmentParametersProto.FloatParametersEntry)
 

ml-agents-envs/mlagents/envs/communicator_objects/header_pb2.py

 # source: mlagents/envs/communicator_objects/header.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/header.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n/mlagents/envs/communicator_objects/header.proto\x12\x14\x63ommunicator_objects\")\n\x06Header\x12\x0e\n\x06status\x18\x01 \x01(\x05\x12\x0f\n\x07message\x18\x02 \x01(\tB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/header.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n/mlagents/envs/communicator_objects/header.proto\x12\x14\x63ommunicator_objects")\n\x06Header\x12\x0e\n\x06status\x18\x01 \x01(\x05\x12\x0f\n\x07message\x18\x02 \x01(\tB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='Header',
-  full_name='communicator_objects.Header',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='status', full_name='communicator_objects.Header.status', index=0,
-      number=1, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='message', full_name='communicator_objects.Header.message', index=1,
-      number=2, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=73,
-  serialized_end=114,
+    name="Header",
+    full_name="communicator_objects.Header",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="status",
+            full_name="communicator_objects.Header.status",
+            index=0,
+            number=1,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="message",
+            full_name="communicator_objects.Header.message",
+            index=1,
+            number=2,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=73,
+    serialized_end=114,
-DESCRIPTOR.message_types_by_name['Header'] = _HEADER
+DESCRIPTOR.message_types_by_name["Header"] = _HEADER
-Header = _reflection.GeneratedProtocolMessageType('Header', (_message.Message,), dict(
-  DESCRIPTOR = _HEADER,
-  __module__ = 'mlagents.envs.communicator_objects.header_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.Header)
-  ))
+Header = _reflection.GeneratedProtocolMessageType(
+    "Header",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_HEADER,
+        __module__="mlagents.envs.communicator_objects.header_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.Header)
+    ),
+)
 _sym_db.RegisterMessage(Header)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/resolution_proto_pb2.py

 # source: mlagents/envs/communicator_objects/resolution_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/resolution_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n9mlagents/envs/communicator_objects/resolution_proto.proto\x12\x14\x63ommunicator_objects\"D\n\x0fResolutionProto\x12\r\n\x05width\x18\x01 \x01(\x05\x12\x0e\n\x06height\x18\x02 \x01(\x05\x12\x12\n\ngray_scale\x18\x03 \x01(\x08\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/resolution_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n9mlagents/envs/communicator_objects/resolution_proto.proto\x12\x14\x63ommunicator_objects"D\n\x0fResolutionProto\x12\r\n\x05width\x18\x01 \x01(\x05\x12\x0e\n\x06height\x18\x02 \x01(\x05\x12\x12\n\ngray_scale\x18\x03 \x01(\x08\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='ResolutionProto',
-  full_name='communicator_objects.ResolutionProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='width', full_name='communicator_objects.ResolutionProto.width', index=0,
-      number=1, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='height', full_name='communicator_objects.ResolutionProto.height', index=1,
-      number=2, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='gray_scale', full_name='communicator_objects.ResolutionProto.gray_scale', index=2,
-      number=3, type=8, cpp_type=7, label=1,
-      has_default_value=False, default_value=False,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=83,
-  serialized_end=151,
+    name="ResolutionProto",
+    full_name="communicator_objects.ResolutionProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="width",
+            full_name="communicator_objects.ResolutionProto.width",
+            index=0,
+            number=1,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="height",
+            full_name="communicator_objects.ResolutionProto.height",
+            index=1,
+            number=2,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="gray_scale",
+            full_name="communicator_objects.ResolutionProto.gray_scale",
+            index=2,
+            number=3,
+            type=8,
+            cpp_type=7,
+            label=1,
+            has_default_value=False,
+            default_value=False,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=83,
+    serialized_end=151,
-DESCRIPTOR.message_types_by_name['ResolutionProto'] = _RESOLUTIONPROTO
+DESCRIPTOR.message_types_by_name["ResolutionProto"] = _RESOLUTIONPROTO
-ResolutionProto = _reflection.GeneratedProtocolMessageType('ResolutionProto', (_message.Message,), dict(
-  DESCRIPTOR = _RESOLUTIONPROTO,
-  __module__ = 'mlagents.envs.communicator_objects.resolution_proto_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.ResolutionProto)
-  ))
+ResolutionProto = _reflection.GeneratedProtocolMessageType(
+    "ResolutionProto",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_RESOLUTIONPROTO,
+        __module__="mlagents.envs.communicator_objects.resolution_proto_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.ResolutionProto)
+    ),
+)
 _sym_db.RegisterMessage(ResolutionProto)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/space_type_proto_pb2.py

 # source: mlagents/envs/communicator_objects/space_type_proto.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import resolution_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2
+from mlagents.envs.communicator_objects import (
+    resolution_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2,
+)
-  name='mlagents/envs/communicator_objects/space_type_proto.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n9mlagents/envs/communicator_objects/space_type_proto.proto\x12\x14\x63ommunicator_objects\x1a\x39mlagents/envs/communicator_objects/resolution_proto.proto*.\n\x0eSpaceTypeProto\x12\x0c\n\x08\x64iscrete\x10\x00\x12\x0e\n\ncontinuous\x10\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2.DESCRIPTOR,])
+    name="mlagents/envs/communicator_objects/space_type_proto.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        "\n9mlagents/envs/communicator_objects/space_type_proto.proto\x12\x14\x63ommunicator_objects\x1a\x39mlagents/envs/communicator_objects/resolution_proto.proto*.\n\x0eSpaceTypeProto\x12\x0c\n\x08\x64iscrete\x10\x00\x12\x0e\n\ncontinuous\x10\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3"
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_resolution__proto__pb2.DESCRIPTOR
+    ],
+)
-  name='SpaceTypeProto',
-  full_name='communicator_objects.SpaceTypeProto',
-  filename=None,
-  file=DESCRIPTOR,
-  values=[
-    _descriptor.EnumValueDescriptor(
-      name='discrete', index=0, number=0,
-      serialized_options=None,
-      type=None),
-    _descriptor.EnumValueDescriptor(
-      name='continuous', index=1, number=1,
-      serialized_options=None,
-      type=None),
-  ],
-  containing_type=None,
-  serialized_options=None,
-  serialized_start=142,
-  serialized_end=188,
+    name="SpaceTypeProto",
+    full_name="communicator_objects.SpaceTypeProto",
+    filename=None,
+    file=DESCRIPTOR,
+    values=[
+        _descriptor.EnumValueDescriptor(
+            name="discrete", index=0, number=0, serialized_options=None, type=None
+        ),
+        _descriptor.EnumValueDescriptor(
+            name="continuous", index=1, number=1, serialized_options=None, type=None
+        ),
+    ],
+    containing_type=None,
+    serialized_options=None,
+    serialized_start=142,
+    serialized_end=188,
 )
 _sym_db.RegisterEnumDescriptor(_SPACETYPEPROTO)
 
 
 
-DESCRIPTOR.enum_types_by_name['SpaceTypeProto'] = _SPACETYPEPROTO
+DESCRIPTOR.enum_types_by_name["SpaceTypeProto"] = _SPACETYPEPROTO
 _sym_db.RegisterFileDescriptor(DESCRIPTOR)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/unity_input_pb2.py

 # source: mlagents/envs/communicator_objects/unity_input.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import unity_rl_input_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__input__pb2
-from mlagents.envs.communicator_objects import unity_rl_initialization_input_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__input__pb2
+from mlagents.envs.communicator_objects import (
+    unity_rl_input_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__input__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    unity_rl_initialization_input_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__input__pb2,
+)
-  name='mlagents/envs/communicator_objects/unity_input.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n4mlagents/envs/communicator_objects/unity_input.proto\x12\x14\x63ommunicator_objects\x1a\x37mlagents/envs/communicator_objects/unity_rl_input.proto\x1a\x46mlagents/envs/communicator_objects/unity_rl_initialization_input.proto\"\x95\x01\n\nUnityInput\x12\x34\n\x08rl_input\x18\x01 \x01(\x0b\x32\".communicator_objects.UnityRLInput\x12Q\n\x17rl_initialization_input\x18\x02 \x01(\x0b\x32\x30.communicator_objects.UnityRLInitializationInputB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__input__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__input__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/unity_input.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n4mlagents/envs/communicator_objects/unity_input.proto\x12\x14\x63ommunicator_objects\x1a\x37mlagents/envs/communicator_objects/unity_rl_input.proto\x1a\x46mlagents/envs/communicator_objects/unity_rl_initialization_input.proto"\x95\x01\n\nUnityInput\x12\x34\n\x08rl_input\x18\x01 \x01(\x0b\x32".communicator_objects.UnityRLInput\x12Q\n\x17rl_initialization_input\x18\x02 \x01(\x0b\x32\x30.communicator_objects.UnityRLInitializationInputB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__input__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__input__pb2.DESCRIPTOR,
+    ],
+)
-  name='UnityInput',
-  full_name='communicator_objects.UnityInput',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='rl_input', full_name='communicator_objects.UnityInput.rl_input', index=0,
-      number=1, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='rl_initialization_input', full_name='communicator_objects.UnityInput.rl_initialization_input', index=1,
-      number=2, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=208,
-  serialized_end=357,
+    name="UnityInput",
+    full_name="communicator_objects.UnityInput",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="rl_input",
+            full_name="communicator_objects.UnityInput.rl_input",
+            index=0,
+            number=1,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="rl_initialization_input",
+            full_name="communicator_objects.UnityInput.rl_initialization_input",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=208,
+    serialized_end=357,
-_UNITYINPUT.fields_by_name['rl_input'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__input__pb2._UNITYRLINPUT
-_UNITYINPUT.fields_by_name['rl_initialization_input'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__input__pb2._UNITYRLINITIALIZATIONINPUT
-DESCRIPTOR.message_types_by_name['UnityInput'] = _UNITYINPUT
+_UNITYINPUT.fields_by_name[
+    "rl_input"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__input__pb2._UNITYRLINPUT
+)
+_UNITYINPUT.fields_by_name[
+    "rl_initialization_input"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__input__pb2._UNITYRLINITIALIZATIONINPUT
+)
+DESCRIPTOR.message_types_by_name["UnityInput"] = _UNITYINPUT
-UnityInput = _reflection.GeneratedProtocolMessageType('UnityInput', (_message.Message,), dict(
-  DESCRIPTOR = _UNITYINPUT,
-  __module__ = 'mlagents.envs.communicator_objects.unity_input_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.UnityInput)
-  ))
+UnityInput = _reflection.GeneratedProtocolMessageType(
+    "UnityInput",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_UNITYINPUT,
+        __module__="mlagents.envs.communicator_objects.unity_input_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.UnityInput)
+    ),
+)
 _sym_db.RegisterMessage(UnityInput)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/unity_message_pb2.py

 # source: mlagents/envs/communicator_objects/unity_message.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import unity_output_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__output__pb2
-from mlagents.envs.communicator_objects import unity_input_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__input__pb2
-from mlagents.envs.communicator_objects import header_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_header__pb2
+from mlagents.envs.communicator_objects import (
+    unity_output_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__output__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    unity_input_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__input__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    header_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_header__pb2,
+)
-  name='mlagents/envs/communicator_objects/unity_message.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n6mlagents/envs/communicator_objects/unity_message.proto\x12\x14\x63ommunicator_objects\x1a\x35mlagents/envs/communicator_objects/unity_output.proto\x1a\x34mlagents/envs/communicator_objects/unity_input.proto\x1a/mlagents/envs/communicator_objects/header.proto\"\xac\x01\n\x0cUnityMessage\x12,\n\x06header\x18\x01 \x01(\x0b\x32\x1c.communicator_objects.Header\x12\x37\n\x0cunity_output\x18\x02 \x01(\x0b\x32!.communicator_objects.UnityOutput\x12\x35\n\x0bunity_input\x18\x03 \x01(\x0b\x32 .communicator_objects.UnityInputB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_unity__output__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_unity__input__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_header__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/unity_message.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n6mlagents/envs/communicator_objects/unity_message.proto\x12\x14\x63ommunicator_objects\x1a\x35mlagents/envs/communicator_objects/unity_output.proto\x1a\x34mlagents/envs/communicator_objects/unity_input.proto\x1a/mlagents/envs/communicator_objects/header.proto"\xac\x01\n\x0cUnityMessage\x12,\n\x06header\x18\x01 \x01(\x0b\x32\x1c.communicator_objects.Header\x12\x37\n\x0cunity_output\x18\x02 \x01(\x0b\x32!.communicator_objects.UnityOutput\x12\x35\n\x0bunity_input\x18\x03 \x01(\x0b\x32 .communicator_objects.UnityInputB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_unity__output__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_unity__input__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_header__pb2.DESCRIPTOR,
+    ],
+)
-  name='UnityMessage',
-  full_name='communicator_objects.UnityMessage',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='header', full_name='communicator_objects.UnityMessage.header', index=0,
-      number=1, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='unity_output', full_name='communicator_objects.UnityMessage.unity_output', index=1,
-      number=2, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='unity_input', full_name='communicator_objects.UnityMessage.unity_input', index=2,
-      number=3, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=239,
-  serialized_end=411,
+    name="UnityMessage",
+    full_name="communicator_objects.UnityMessage",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="header",
+            full_name="communicator_objects.UnityMessage.header",
+            index=0,
+            number=1,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="unity_output",
+            full_name="communicator_objects.UnityMessage.unity_output",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="unity_input",
+            full_name="communicator_objects.UnityMessage.unity_input",
+            index=2,
+            number=3,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=239,
+    serialized_end=411,
-_UNITYMESSAGE.fields_by_name['header'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_header__pb2._HEADER
-_UNITYMESSAGE.fields_by_name['unity_output'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_unity__output__pb2._UNITYOUTPUT
-_UNITYMESSAGE.fields_by_name['unity_input'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_unity__input__pb2._UNITYINPUT
-DESCRIPTOR.message_types_by_name['UnityMessage'] = _UNITYMESSAGE
+_UNITYMESSAGE.fields_by_name[
+    "header"
+].message_type = mlagents_dot_envs_dot_communicator__objects_dot_header__pb2._HEADER
+_UNITYMESSAGE.fields_by_name[
+    "unity_output"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_unity__output__pb2._UNITYOUTPUT
+)
+_UNITYMESSAGE.fields_by_name[
+    "unity_input"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_unity__input__pb2._UNITYINPUT
+)
+DESCRIPTOR.message_types_by_name["UnityMessage"] = _UNITYMESSAGE
-UnityMessage = _reflection.GeneratedProtocolMessageType('UnityMessage', (_message.Message,), dict(
-  DESCRIPTOR = _UNITYMESSAGE,
-  __module__ = 'mlagents.envs.communicator_objects.unity_message_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.UnityMessage)
-  ))
+UnityMessage = _reflection.GeneratedProtocolMessageType(
+    "UnityMessage",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_UNITYMESSAGE,
+        __module__="mlagents.envs.communicator_objects.unity_message_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.UnityMessage)
+    ),
+)
 _sym_db.RegisterMessage(UnityMessage)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/unity_output_pb2.py

 # source: mlagents/envs/communicator_objects/unity_output.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import unity_rl_output_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__output__pb2
-from mlagents.envs.communicator_objects import unity_rl_initialization_output_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__output__pb2
+from mlagents.envs.communicator_objects import (
+    unity_rl_output_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__output__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    unity_rl_initialization_output_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__output__pb2,
+)
-  name='mlagents/envs/communicator_objects/unity_output.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n5mlagents/envs/communicator_objects/unity_output.proto\x12\x14\x63ommunicator_objects\x1a\x38mlagents/envs/communicator_objects/unity_rl_output.proto\x1aGmlagents/envs/communicator_objects/unity_rl_initialization_output.proto\"\x9a\x01\n\x0bUnityOutput\x12\x36\n\trl_output\x18\x01 \x01(\x0b\x32#.communicator_objects.UnityRLOutput\x12S\n\x18rl_initialization_output\x18\x02 \x01(\x0b\x32\x31.communicator_objects.UnityRLInitializationOutputB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__output__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__output__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/unity_output.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n5mlagents/envs/communicator_objects/unity_output.proto\x12\x14\x63ommunicator_objects\x1a\x38mlagents/envs/communicator_objects/unity_rl_output.proto\x1aGmlagents/envs/communicator_objects/unity_rl_initialization_output.proto"\x9a\x01\n\x0bUnityOutput\x12\x36\n\trl_output\x18\x01 \x01(\x0b\x32#.communicator_objects.UnityRLOutput\x12S\n\x18rl_initialization_output\x18\x02 \x01(\x0b\x32\x31.communicator_objects.UnityRLInitializationOutputB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__output__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__output__pb2.DESCRIPTOR,
+    ],
+)
-  name='UnityOutput',
-  full_name='communicator_objects.UnityOutput',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='rl_output', full_name='communicator_objects.UnityOutput.rl_output', index=0,
-      number=1, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='rl_initialization_output', full_name='communicator_objects.UnityOutput.rl_initialization_output', index=1,
-      number=2, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=211,
-  serialized_end=365,
+    name="UnityOutput",
+    full_name="communicator_objects.UnityOutput",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="rl_output",
+            full_name="communicator_objects.UnityOutput.rl_output",
+            index=0,
+            number=1,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="rl_initialization_output",
+            full_name="communicator_objects.UnityOutput.rl_initialization_output",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=211,
+    serialized_end=365,
-_UNITYOUTPUT.fields_by_name['rl_output'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__output__pb2._UNITYRLOUTPUT
-_UNITYOUTPUT.fields_by_name['rl_initialization_output'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__output__pb2._UNITYRLINITIALIZATIONOUTPUT
-DESCRIPTOR.message_types_by_name['UnityOutput'] = _UNITYOUTPUT
+_UNITYOUTPUT.fields_by_name[
+    "rl_output"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__output__pb2._UNITYRLOUTPUT
+)
+_UNITYOUTPUT.fields_by_name[
+    "rl_initialization_output"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_unity__rl__initialization__output__pb2._UNITYRLINITIALIZATIONOUTPUT
+)
+DESCRIPTOR.message_types_by_name["UnityOutput"] = _UNITYOUTPUT
-UnityOutput = _reflection.GeneratedProtocolMessageType('UnityOutput', (_message.Message,), dict(
-  DESCRIPTOR = _UNITYOUTPUT,
-  __module__ = 'mlagents.envs.communicator_objects.unity_output_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.UnityOutput)
-  ))
+UnityOutput = _reflection.GeneratedProtocolMessageType(
+    "UnityOutput",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_UNITYOUTPUT,
+        __module__="mlagents.envs.communicator_objects.unity_output_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.UnityOutput)
+    ),
+)
 _sym_db.RegisterMessage(UnityOutput)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/unity_rl_initialization_input_pb2.py

 # source: mlagents/envs/communicator_objects/unity_rl_initialization_input.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-
-
-  name='mlagents/envs/communicator_objects/unity_rl_initialization_input.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\nFmlagents/envs/communicator_objects/unity_rl_initialization_input.proto\x12\x14\x63ommunicator_objects\"*\n\x1aUnityRLInitializationInput\x12\x0c\n\x04seed\x18\x01 \x01(\x05\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
+    name="mlagents/envs/communicator_objects/unity_rl_initialization_input.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\nFmlagents/envs/communicator_objects/unity_rl_initialization_input.proto\x12\x14\x63ommunicator_objects"*\n\x1aUnityRLInitializationInput\x12\x0c\n\x04seed\x18\x01 \x01(\x05\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
-
-
-  name='UnityRLInitializationInput',
-  full_name='communicator_objects.UnityRLInitializationInput',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='seed', full_name='communicator_objects.UnityRLInitializationInput.seed', index=0,
-      number=1, type=5, cpp_type=1, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=96,
-  serialized_end=138,
+    name="UnityRLInitializationInput",
+    full_name="communicator_objects.UnityRLInitializationInput",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="seed",
+            full_name="communicator_objects.UnityRLInitializationInput.seed",
+            index=0,
+            number=1,
+            type=5,
+            cpp_type=1,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        )
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=96,
+    serialized_end=138,
-DESCRIPTOR.message_types_by_name['UnityRLInitializationInput'] = _UNITYRLINITIALIZATIONINPUT
+DESCRIPTOR.message_types_by_name[
+    "UnityRLInitializationInput"
+] = _UNITYRLINITIALIZATIONINPUT
-UnityRLInitializationInput = _reflection.GeneratedProtocolMessageType('UnityRLInitializationInput', (_message.Message,), dict(
-  DESCRIPTOR = _UNITYRLINITIALIZATIONINPUT,
-  __module__ = 'mlagents.envs.communicator_objects.unity_rl_initialization_input_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInitializationInput)
-  ))
+UnityRLInitializationInput = _reflection.GeneratedProtocolMessageType(
+    "UnityRLInitializationInput",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_UNITYRLINITIALIZATIONINPUT,
+        __module__="mlagents.envs.communicator_objects.unity_rl_initialization_input_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInitializationInput)
+    ),
+)
 _sym_db.RegisterMessage(UnityRLInitializationInput)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/unity_rl_initialization_output_pb2.py

 # source: mlagents/envs/communicator_objects/unity_rl_initialization_output.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import brain_parameters_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_brain__parameters__proto__pb2
-from mlagents.envs.communicator_objects import environment_parameters_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2
+from mlagents.envs.communicator_objects import (
+    brain_parameters_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_brain__parameters__proto__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    environment_parameters_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2,
+)
-  name='mlagents/envs/communicator_objects/unity_rl_initialization_output.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\nGmlagents/envs/communicator_objects/unity_rl_initialization_output.proto\x12\x14\x63ommunicator_objects\x1a?mlagents/envs/communicator_objects/brain_parameters_proto.proto\x1a\x45mlagents/envs/communicator_objects/environment_parameters_proto.proto\"\xe6\x01\n\x1bUnityRLInitializationOutput\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0f\n\x07version\x18\x02 \x01(\t\x12\x10\n\x08log_path\x18\x03 \x01(\t\x12\x44\n\x10\x62rain_parameters\x18\x05 \x03(\x0b\x32*.communicator_objects.BrainParametersProto\x12P\n\x16\x65nvironment_parameters\x18\x06 \x01(\x0b\x32\x30.communicator_objects.EnvironmentParametersProtoB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_brain__parameters__proto__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/unity_rl_initialization_output.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\nGmlagents/envs/communicator_objects/unity_rl_initialization_output.proto\x12\x14\x63ommunicator_objects\x1a?mlagents/envs/communicator_objects/brain_parameters_proto.proto\x1a\x45mlagents/envs/communicator_objects/environment_parameters_proto.proto"\xe6\x01\n\x1bUnityRLInitializationOutput\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0f\n\x07version\x18\x02 \x01(\t\x12\x10\n\x08log_path\x18\x03 \x01(\t\x12\x44\n\x10\x62rain_parameters\x18\x05 \x03(\x0b\x32*.communicator_objects.BrainParametersProto\x12P\n\x16\x65nvironment_parameters\x18\x06 \x01(\x0b\x32\x30.communicator_objects.EnvironmentParametersProtoB\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_brain__parameters__proto__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2.DESCRIPTOR,
+    ],
+)
-  name='UnityRLInitializationOutput',
-  full_name='communicator_objects.UnityRLInitializationOutput',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='name', full_name='communicator_objects.UnityRLInitializationOutput.name', index=0,
-      number=1, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='version', full_name='communicator_objects.UnityRLInitializationOutput.version', index=1,
-      number=2, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='log_path', full_name='communicator_objects.UnityRLInitializationOutput.log_path', index=2,
-      number=3, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='brain_parameters', full_name='communicator_objects.UnityRLInitializationOutput.brain_parameters', index=3,
-      number=5, type=11, cpp_type=10, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='environment_parameters', full_name='communicator_objects.UnityRLInitializationOutput.environment_parameters', index=4,
-      number=6, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=234,
-  serialized_end=464,
+    name="UnityRLInitializationOutput",
+    full_name="communicator_objects.UnityRLInitializationOutput",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="name",
+            full_name="communicator_objects.UnityRLInitializationOutput.name",
+            index=0,
+            number=1,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="version",
+            full_name="communicator_objects.UnityRLInitializationOutput.version",
+            index=1,
+            number=2,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="log_path",
+            full_name="communicator_objects.UnityRLInitializationOutput.log_path",
+            index=2,
+            number=3,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="brain_parameters",
+            full_name="communicator_objects.UnityRLInitializationOutput.brain_parameters",
+            index=3,
+            number=5,
+            type=11,
+            cpp_type=10,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="environment_parameters",
+            full_name="communicator_objects.UnityRLInitializationOutput.environment_parameters",
+            index=4,
+            number=6,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=234,
+    serialized_end=464,
-_UNITYRLINITIALIZATIONOUTPUT.fields_by_name['brain_parameters'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_brain__parameters__proto__pb2._BRAINPARAMETERSPROTO
-_UNITYRLINITIALIZATIONOUTPUT.fields_by_name['environment_parameters'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2._ENVIRONMENTPARAMETERSPROTO
-DESCRIPTOR.message_types_by_name['UnityRLInitializationOutput'] = _UNITYRLINITIALIZATIONOUTPUT
+_UNITYRLINITIALIZATIONOUTPUT.fields_by_name[
+    "brain_parameters"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_brain__parameters__proto__pb2._BRAINPARAMETERSPROTO
+)
+_UNITYRLINITIALIZATIONOUTPUT.fields_by_name[
+    "environment_parameters"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2._ENVIRONMENTPARAMETERSPROTO
+)
+DESCRIPTOR.message_types_by_name[
+    "UnityRLInitializationOutput"
+] = _UNITYRLINITIALIZATIONOUTPUT
-UnityRLInitializationOutput = _reflection.GeneratedProtocolMessageType('UnityRLInitializationOutput', (_message.Message,), dict(
-  DESCRIPTOR = _UNITYRLINITIALIZATIONOUTPUT,
-  __module__ = 'mlagents.envs.communicator_objects.unity_rl_initialization_output_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInitializationOutput)
-  ))
+UnityRLInitializationOutput = _reflection.GeneratedProtocolMessageType(
+    "UnityRLInitializationOutput",
+    (_message.Message,),
+    dict(
+        DESCRIPTOR=_UNITYRLINITIALIZATIONOUTPUT,
+        __module__="mlagents.envs.communicator_objects.unity_rl_initialization_output_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInitializationOutput)
+    ),
+)
 _sym_db.RegisterMessage(UnityRLInitializationOutput)
 
 

ml-agents-envs/mlagents/envs/communicator_objects/unity_rl_input_pb2.py

 # source: mlagents/envs/communicator_objects/unity_rl_input.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import agent_action_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_agent__action__proto__pb2
-from mlagents.envs.communicator_objects import environment_parameters_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2
-from mlagents.envs.communicator_objects import command_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_command__proto__pb2
+from mlagents.envs.communicator_objects import (
+    agent_action_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_agent__action__proto__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    environment_parameters_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2,
+)
+from mlagents.envs.communicator_objects import (
+    command_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_command__proto__pb2,
+)
-  name='mlagents/envs/communicator_objects/unity_rl_input.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n7mlagents/envs/communicator_objects/unity_rl_input.proto\x12\x14\x63ommunicator_objects\x1a;mlagents/envs/communicator_objects/agent_action_proto.proto\x1a\x45mlagents/envs/communicator_objects/environment_parameters_proto.proto\x1a\x36mlagents/envs/communicator_objects/command_proto.proto\"\xb4\x03\n\x0cUnityRLInput\x12K\n\ragent_actions\x18\x01 \x03(\x0b\x32\x34.communicator_objects.UnityRLInput.AgentActionsEntry\x12P\n\x16\x65nvironment_parameters\x18\x02 \x01(\x0b\x32\x30.communicator_objects.EnvironmentParametersProto\x12\x13\n\x0bis_training\x18\x03 \x01(\x08\x12\x33\n\x07\x63ommand\x18\x04 \x01(\x0e\x32\".communicator_objects.CommandProto\x1aM\n\x14ListAgentActionProto\x12\x35\n\x05value\x18\x01 \x03(\x0b\x32&.communicator_objects.AgentActionProto\x1al\n\x11\x41gentActionsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\x46\n\x05value\x18\x02 \x01(\x0b\x32\x37.communicator_objects.UnityRLInput.ListAgentActionProto:\x02\x38\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_agent__action__proto__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2.DESCRIPTOR,mlagents_dot_envs_dot_communicator__objects_dot_command__proto__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/unity_rl_input.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n7mlagents/envs/communicator_objects/unity_rl_input.proto\x12\x14\x63ommunicator_objects\x1a;mlagents/envs/communicator_objects/agent_action_proto.proto\x1a\x45mlagents/envs/communicator_objects/environment_parameters_proto.proto\x1a\x36mlagents/envs/communicator_objects/command_proto.proto"\xb4\x03\n\x0cUnityRLInput\x12K\n\ragent_actions\x18\x01 \x03(\x0b\x32\x34.communicator_objects.UnityRLInput.AgentActionsEntry\x12P\n\x16\x65nvironment_parameters\x18\x02 \x01(\x0b\x32\x30.communicator_objects.EnvironmentParametersProto\x12\x13\n\x0bis_training\x18\x03 \x01(\x08\x12\x33\n\x07\x63ommand\x18\x04 \x01(\x0e\x32".communicator_objects.CommandProto\x1aM\n\x14ListAgentActionProto\x12\x35\n\x05value\x18\x01 \x03(\x0b\x32&.communicator_objects.AgentActionProto\x1al\n\x11\x41gentActionsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\x46\n\x05value\x18\x02 \x01(\x0b\x32\x37.communicator_objects.UnityRLInput.ListAgentActionProto:\x02\x38\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_agent__action__proto__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2.DESCRIPTOR,
+        mlagents_dot_envs_dot_communicator__objects_dot_command__proto__pb2.DESCRIPTOR,
+    ],
+)
-  name='ListAgentActionProto',
-  full_name='communicator_objects.UnityRLInput.ListAgentActionProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='value', full_name='communicator_objects.UnityRLInput.ListAgentActionProto.value', index=0,
-      number=1, type=11, cpp_type=10, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=519,
-  serialized_end=596,
+    name="ListAgentActionProto",
+    full_name="communicator_objects.UnityRLInput.ListAgentActionProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="value",
+            full_name="communicator_objects.UnityRLInput.ListAgentActionProto.value",
+            index=0,
+            number=1,
+            type=11,
+            cpp_type=10,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        )
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=519,
+    serialized_end=596,
-  name='AgentActionsEntry',
-  full_name='communicator_objects.UnityRLInput.AgentActionsEntry',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='key', full_name='communicator_objects.UnityRLInput.AgentActionsEntry.key', index=0,
-      number=1, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='value', full_name='communicator_objects.UnityRLInput.AgentActionsEntry.value', index=1,
-      number=2, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=_b('8\001'),
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=598,
-  serialized_end=706,
+    name="AgentActionsEntry",
+    full_name="communicator_objects.UnityRLInput.AgentActionsEntry",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="key",
+            full_name="communicator_objects.UnityRLInput.AgentActionsEntry.key",
+            index=0,
+            number=1,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="value",
+            full_name="communicator_objects.UnityRLInput.AgentActionsEntry.value",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=_b("8\001"),
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=598,
+    serialized_end=706,
-  name='UnityRLInput',
-  full_name='communicator_objects.UnityRLInput',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='agent_actions', full_name='communicator_objects.UnityRLInput.agent_actions', index=0,
-      number=1, type=11, cpp_type=10, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='environment_parameters', full_name='communicator_objects.UnityRLInput.environment_parameters', index=1,
-      number=2, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='is_training', full_name='communicator_objects.UnityRLInput.is_training', index=2,
-      number=3, type=8, cpp_type=7, label=1,
-      has_default_value=False, default_value=False,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='command', full_name='communicator_objects.UnityRLInput.command', index=3,
-      number=4, type=14, cpp_type=8, label=1,
-      has_default_value=False, default_value=0,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[_UNITYRLINPUT_LISTAGENTACTIONPROTO, _UNITYRLINPUT_AGENTACTIONSENTRY, ],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=270,
-  serialized_end=706,
+    name="UnityRLInput",
+    full_name="communicator_objects.UnityRLInput",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="agent_actions",
+            full_name="communicator_objects.UnityRLInput.agent_actions",
+            index=0,
+            number=1,
+            type=11,
+            cpp_type=10,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="environment_parameters",
+            full_name="communicator_objects.UnityRLInput.environment_parameters",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="is_training",
+            full_name="communicator_objects.UnityRLInput.is_training",
+            index=2,
+            number=3,
+            type=8,
+            cpp_type=7,
+            label=1,
+            has_default_value=False,
+            default_value=False,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="command",
+            full_name="communicator_objects.UnityRLInput.command",
+            index=3,
+            number=4,
+            type=14,
+            cpp_type=8,
+            label=1,
+            has_default_value=False,
+            default_value=0,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[_UNITYRLINPUT_LISTAGENTACTIONPROTO, _UNITYRLINPUT_AGENTACTIONSENTRY],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=270,
+    serialized_end=706,
-_UNITYRLINPUT_LISTAGENTACTIONPROTO.fields_by_name['value'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_agent__action__proto__pb2._AGENTACTIONPROTO
+_UNITYRLINPUT_LISTAGENTACTIONPROTO.fields_by_name[
+    "value"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_agent__action__proto__pb2._AGENTACTIONPROTO
+)
-_UNITYRLINPUT_AGENTACTIONSENTRY.fields_by_name['value'].message_type = _UNITYRLINPUT_LISTAGENTACTIONPROTO
+_UNITYRLINPUT_AGENTACTIONSENTRY.fields_by_name[
+    "value"
+].message_type = _UNITYRLINPUT_LISTAGENTACTIONPROTO
-_UNITYRLINPUT.fields_by_name['agent_actions'].message_type = _UNITYRLINPUT_AGENTACTIONSENTRY
-_UNITYRLINPUT.fields_by_name['environment_parameters'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2._ENVIRONMENTPARAMETERSPROTO
-_UNITYRLINPUT.fields_by_name['command'].enum_type = mlagents_dot_envs_dot_communicator__objects_dot_command__proto__pb2._COMMANDPROTO
-DESCRIPTOR.message_types_by_name['UnityRLInput'] = _UNITYRLINPUT
+_UNITYRLINPUT.fields_by_name[
+    "agent_actions"
+].message_type = _UNITYRLINPUT_AGENTACTIONSENTRY
+_UNITYRLINPUT.fields_by_name[
+    "environment_parameters"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_environment__parameters__proto__pb2._ENVIRONMENTPARAMETERSPROTO
+)
+_UNITYRLINPUT.fields_by_name[
+    "command"
+].enum_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_command__proto__pb2._COMMANDPROTO
+)
+DESCRIPTOR.message_types_by_name["UnityRLInput"] = _UNITYRLINPUT
-UnityRLInput = _reflection.GeneratedProtocolMessageType('UnityRLInput', (_message.Message,), dict(
-
-  ListAgentActionProto = _reflection.GeneratedProtocolMessageType('ListAgentActionProto', (_message.Message,), dict(
-    DESCRIPTOR = _UNITYRLINPUT_LISTAGENTACTIONPROTO,
-    __module__ = 'mlagents.envs.communicator_objects.unity_rl_input_pb2'
-    # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInput.ListAgentActionProto)
-    ))
-  ,
-
-  AgentActionsEntry = _reflection.GeneratedProtocolMessageType('AgentActionsEntry', (_message.Message,), dict(
-    DESCRIPTOR = _UNITYRLINPUT_AGENTACTIONSENTRY,
-    __module__ = 'mlagents.envs.communicator_objects.unity_rl_input_pb2'
-    # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInput.AgentActionsEntry)
-    ))
-  ,
-  DESCRIPTOR = _UNITYRLINPUT,
-  __module__ = 'mlagents.envs.communicator_objects.unity_rl_input_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInput)
-  ))
+UnityRLInput = _reflection.GeneratedProtocolMessageType(
+    "UnityRLInput",
+    (_message.Message,),
+    dict(
+        ListAgentActionProto=_reflection.GeneratedProtocolMessageType(
+            "ListAgentActionProto",
+            (_message.Message,),
+            dict(
+                DESCRIPTOR=_UNITYRLINPUT_LISTAGENTACTIONPROTO,
+                __module__="mlagents.envs.communicator_objects.unity_rl_input_pb2"
+                # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInput.ListAgentActionProto)
+            ),
+        ),
+        AgentActionsEntry=_reflection.GeneratedProtocolMessageType(
+            "AgentActionsEntry",
+            (_message.Message,),
+            dict(
+                DESCRIPTOR=_UNITYRLINPUT_AGENTACTIONSENTRY,
+                __module__="mlagents.envs.communicator_objects.unity_rl_input_pb2"
+                # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInput.AgentActionsEntry)
+            ),
+        ),
+        DESCRIPTOR=_UNITYRLINPUT,
+        __module__="mlagents.envs.communicator_objects.unity_rl_input_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLInput)
+    ),
+)
 _sym_db.RegisterMessage(UnityRLInput)
 _sym_db.RegisterMessage(UnityRLInput.ListAgentActionProto)
 _sym_db.RegisterMessage(UnityRLInput.AgentActionsEntry)

ml-agents-envs/mlagents/envs/communicator_objects/unity_rl_output_pb2.py

 # source: mlagents/envs/communicator_objects/unity_rl_output.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import agent_info_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_agent__info__proto__pb2
+from mlagents.envs.communicator_objects import (
+    agent_info_proto_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_agent__info__proto__pb2,
+)
-  name='mlagents/envs/communicator_objects/unity_rl_output.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n8mlagents/envs/communicator_objects/unity_rl_output.proto\x12\x14\x63ommunicator_objects\x1a\x39mlagents/envs/communicator_objects/agent_info_proto.proto\"\xa3\x02\n\rUnityRLOutput\x12\x13\n\x0bglobal_done\x18\x01 \x01(\x08\x12G\n\nagentInfos\x18\x02 \x03(\x0b\x32\x33.communicator_objects.UnityRLOutput.AgentInfosEntry\x1aI\n\x12ListAgentInfoProto\x12\x33\n\x05value\x18\x01 \x03(\x0b\x32$.communicator_objects.AgentInfoProto\x1ai\n\x0f\x41gentInfosEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\x45\n\x05value\x18\x02 \x01(\x0b\x32\x36.communicator_objects.UnityRLOutput.ListAgentInfoProto:\x02\x38\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_agent__info__proto__pb2.DESCRIPTOR,])
-
-
+    name="mlagents/envs/communicator_objects/unity_rl_output.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n8mlagents/envs/communicator_objects/unity_rl_output.proto\x12\x14\x63ommunicator_objects\x1a\x39mlagents/envs/communicator_objects/agent_info_proto.proto"\xa3\x02\n\rUnityRLOutput\x12\x13\n\x0bglobal_done\x18\x01 \x01(\x08\x12G\n\nagentInfos\x18\x02 \x03(\x0b\x32\x33.communicator_objects.UnityRLOutput.AgentInfosEntry\x1aI\n\x12ListAgentInfoProto\x12\x33\n\x05value\x18\x01 \x03(\x0b\x32$.communicator_objects.AgentInfoProto\x1ai\n\x0f\x41gentInfosEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\x45\n\x05value\x18\x02 \x01(\x0b\x32\x36.communicator_objects.UnityRLOutput.ListAgentInfoProto:\x02\x38\x01\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_agent__info__proto__pb2.DESCRIPTOR
+    ],
+)
-  name='ListAgentInfoProto',
-  full_name='communicator_objects.UnityRLOutput.ListAgentInfoProto',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='value', full_name='communicator_objects.UnityRLOutput.ListAgentInfoProto.value', index=0,
-      number=1, type=11, cpp_type=10, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=253,
-  serialized_end=326,
+    name="ListAgentInfoProto",
+    full_name="communicator_objects.UnityRLOutput.ListAgentInfoProto",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="value",
+            full_name="communicator_objects.UnityRLOutput.ListAgentInfoProto.value",
+            index=0,
+            number=1,
+            type=11,
+            cpp_type=10,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        )
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=253,
+    serialized_end=326,
-  name='AgentInfosEntry',
-  full_name='communicator_objects.UnityRLOutput.AgentInfosEntry',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='key', full_name='communicator_objects.UnityRLOutput.AgentInfosEntry.key', index=0,
-      number=1, type=9, cpp_type=9, label=1,
-      has_default_value=False, default_value=_b("").decode('utf-8'),
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='value', full_name='communicator_objects.UnityRLOutput.AgentInfosEntry.value', index=1,
-      number=2, type=11, cpp_type=10, label=1,
-      has_default_value=False, default_value=None,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[],
-  enum_types=[
-  ],
-  serialized_options=_b('8\001'),
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=328,
-  serialized_end=433,
+    name="AgentInfosEntry",
+    full_name="communicator_objects.UnityRLOutput.AgentInfosEntry",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="key",
+            full_name="communicator_objects.UnityRLOutput.AgentInfosEntry.key",
+            index=0,
+            number=1,
+            type=9,
+            cpp_type=9,
+            label=1,
+            has_default_value=False,
+            default_value=_b("").decode("utf-8"),
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="value",
+            full_name="communicator_objects.UnityRLOutput.AgentInfosEntry.value",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=1,
+            has_default_value=False,
+            default_value=None,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[],
+    enum_types=[],
+    serialized_options=_b("8\001"),
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=328,
+    serialized_end=433,
-  name='UnityRLOutput',
-  full_name='communicator_objects.UnityRLOutput',
-  filename=None,
-  file=DESCRIPTOR,
-  containing_type=None,
-  fields=[
-    _descriptor.FieldDescriptor(
-      name='global_done', full_name='communicator_objects.UnityRLOutput.global_done', index=0,
-      number=1, type=8, cpp_type=7, label=1,
-      has_default_value=False, default_value=False,
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-    _descriptor.FieldDescriptor(
-      name='agentInfos', full_name='communicator_objects.UnityRLOutput.agentInfos', index=1,
-      number=2, type=11, cpp_type=10, label=3,
-      has_default_value=False, default_value=[],
-      message_type=None, enum_type=None, containing_type=None,
-      is_extension=False, extension_scope=None,
-      serialized_options=None, file=DESCRIPTOR),
-  ],
-  extensions=[
-  ],
-  nested_types=[_UNITYRLOUTPUT_LISTAGENTINFOPROTO, _UNITYRLOUTPUT_AGENTINFOSENTRY, ],
-  enum_types=[
-  ],
-  serialized_options=None,
-  is_extendable=False,
-  syntax='proto3',
-  extension_ranges=[],
-  oneofs=[
-  ],
-  serialized_start=142,
-  serialized_end=433,
+    name="UnityRLOutput",
+    full_name="communicator_objects.UnityRLOutput",
+    filename=None,
+    file=DESCRIPTOR,
+    containing_type=None,
+    fields=[
+        _descriptor.FieldDescriptor(
+            name="global_done",
+            full_name="communicator_objects.UnityRLOutput.global_done",
+            index=0,
+            number=1,
+            type=8,
+            cpp_type=7,
+            label=1,
+            has_default_value=False,
+            default_value=False,
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+        _descriptor.FieldDescriptor(
+            name="agentInfos",
+            full_name="communicator_objects.UnityRLOutput.agentInfos",
+            index=1,
+            number=2,
+            type=11,
+            cpp_type=10,
+            label=3,
+            has_default_value=False,
+            default_value=[],
+            message_type=None,
+            enum_type=None,
+            containing_type=None,
+            is_extension=False,
+            extension_scope=None,
+            serialized_options=None,
+            file=DESCRIPTOR,
+        ),
+    ],
+    extensions=[],
+    nested_types=[_UNITYRLOUTPUT_LISTAGENTINFOPROTO, _UNITYRLOUTPUT_AGENTINFOSENTRY],
+    enum_types=[],
+    serialized_options=None,
+    is_extendable=False,
+    syntax="proto3",
+    extension_ranges=[],
+    oneofs=[],
+    serialized_start=142,
+    serialized_end=433,
-_UNITYRLOUTPUT_LISTAGENTINFOPROTO.fields_by_name['value'].message_type = mlagents_dot_envs_dot_communicator__objects_dot_agent__info__proto__pb2._AGENTINFOPROTO
+_UNITYRLOUTPUT_LISTAGENTINFOPROTO.fields_by_name[
+    "value"
+].message_type = (
+    mlagents_dot_envs_dot_communicator__objects_dot_agent__info__proto__pb2._AGENTINFOPROTO
+)
-_UNITYRLOUTPUT_AGENTINFOSENTRY.fields_by_name['value'].message_type = _UNITYRLOUTPUT_LISTAGENTINFOPROTO
+_UNITYRLOUTPUT_AGENTINFOSENTRY.fields_by_name[
+    "value"
+].message_type = _UNITYRLOUTPUT_LISTAGENTINFOPROTO
-_UNITYRLOUTPUT.fields_by_name['agentInfos'].message_type = _UNITYRLOUTPUT_AGENTINFOSENTRY
-DESCRIPTOR.message_types_by_name['UnityRLOutput'] = _UNITYRLOUTPUT
+_UNITYRLOUTPUT.fields_by_name[
+    "agentInfos"
+].message_type = _UNITYRLOUTPUT_AGENTINFOSENTRY
+DESCRIPTOR.message_types_by_name["UnityRLOutput"] = _UNITYRLOUTPUT
-UnityRLOutput = _reflection.GeneratedProtocolMessageType('UnityRLOutput', (_message.Message,), dict(
-
-  ListAgentInfoProto = _reflection.GeneratedProtocolMessageType('ListAgentInfoProto', (_message.Message,), dict(
-    DESCRIPTOR = _UNITYRLOUTPUT_LISTAGENTINFOPROTO,
-    __module__ = 'mlagents.envs.communicator_objects.unity_rl_output_pb2'
-    # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLOutput.ListAgentInfoProto)
-    ))
-  ,
-
-  AgentInfosEntry = _reflection.GeneratedProtocolMessageType('AgentInfosEntry', (_message.Message,), dict(
-    DESCRIPTOR = _UNITYRLOUTPUT_AGENTINFOSENTRY,
-    __module__ = 'mlagents.envs.communicator_objects.unity_rl_output_pb2'
-    # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLOutput.AgentInfosEntry)
-    ))
-  ,
-  DESCRIPTOR = _UNITYRLOUTPUT,
-  __module__ = 'mlagents.envs.communicator_objects.unity_rl_output_pb2'
-  # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLOutput)
-  ))
+UnityRLOutput = _reflection.GeneratedProtocolMessageType(
+    "UnityRLOutput",
+    (_message.Message,),
+    dict(
+        ListAgentInfoProto=_reflection.GeneratedProtocolMessageType(
+            "ListAgentInfoProto",
+            (_message.Message,),
+            dict(
+                DESCRIPTOR=_UNITYRLOUTPUT_LISTAGENTINFOPROTO,
+                __module__="mlagents.envs.communicator_objects.unity_rl_output_pb2"
+                # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLOutput.ListAgentInfoProto)
+            ),
+        ),
+        AgentInfosEntry=_reflection.GeneratedProtocolMessageType(
+            "AgentInfosEntry",
+            (_message.Message,),
+            dict(
+                DESCRIPTOR=_UNITYRLOUTPUT_AGENTINFOSENTRY,
+                __module__="mlagents.envs.communicator_objects.unity_rl_output_pb2"
+                # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLOutput.AgentInfosEntry)
+            ),
+        ),
+        DESCRIPTOR=_UNITYRLOUTPUT,
+        __module__="mlagents.envs.communicator_objects.unity_rl_output_pb2"
+        # @@protoc_insertion_point(class_scope:communicator_objects.UnityRLOutput)
+    ),
+)
 _sym_db.RegisterMessage(UnityRLOutput)
 _sym_db.RegisterMessage(UnityRLOutput.ListAgentInfoProto)
 _sym_db.RegisterMessage(UnityRLOutput.AgentInfosEntry)

ml-agents-envs/mlagents/envs/communicator_objects/unity_to_external_pb2.py

 # source: mlagents/envs/communicator_objects/unity_to_external.proto
 
 import sys
-_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
+
+_b = sys.version_info[0] < 3 and (lambda x: x) or (lambda x: x.encode("latin1"))
+
-from mlagents.envs.communicator_objects import unity_message_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2
+from mlagents.envs.communicator_objects import (
+    unity_message_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2,
+)
-  name='mlagents/envs/communicator_objects/unity_to_external.proto',
-  package='communicator_objects',
-  syntax='proto3',
-  serialized_options=_b('\252\002\034MLAgents.CommunicatorObjects'),
-  serialized_pb=_b('\n:mlagents/envs/communicator_objects/unity_to_external.proto\x12\x14\x63ommunicator_objects\x1a\x36mlagents/envs/communicator_objects/unity_message.proto2g\n\x0fUnityToExternal\x12T\n\x08\x45xchange\x12\".communicator_objects.UnityMessage\x1a\".communicator_objects.UnityMessage\"\x00\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3')
-  ,
-  dependencies=[mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.DESCRIPTOR,])
-
+    name="mlagents/envs/communicator_objects/unity_to_external.proto",
+    package="communicator_objects",
+    syntax="proto3",
+    serialized_options=_b("\252\002\034MLAgents.CommunicatorObjects"),
+    serialized_pb=_b(
+        '\n:mlagents/envs/communicator_objects/unity_to_external.proto\x12\x14\x63ommunicator_objects\x1a\x36mlagents/envs/communicator_objects/unity_message.proto2g\n\x0fUnityToExternal\x12T\n\x08\x45xchange\x12".communicator_objects.UnityMessage\x1a".communicator_objects.UnityMessage"\x00\x42\x1f\xaa\x02\x1cMLAgents.CommunicatorObjectsb\x06proto3'
+    ),
+    dependencies=[
+        mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.DESCRIPTOR
+    ],
+)
 
 
 _sym_db.RegisterFileDescriptor(DESCRIPTOR)
 
 _UNITYTOEXTERNAL = _descriptor.ServiceDescriptor(
-  name='UnityToExternal',
-  full_name='communicator_objects.UnityToExternal',
-  file=DESCRIPTOR,
-  index=0,
-  serialized_options=None,
-  serialized_start=140,
-  serialized_end=243,
-  methods=[
-  _descriptor.MethodDescriptor(
-    name='Exchange',
-    full_name='communicator_objects.UnityToExternal.Exchange',
+    name="UnityToExternal",
+    full_name="communicator_objects.UnityToExternal",
+    file=DESCRIPTOR,
-    containing_service=None,
-    input_type=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2._UNITYMESSAGE,
-    output_type=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2._UNITYMESSAGE,
-  ),
-])
+    serialized_start=140,
+    serialized_end=243,
+    methods=[
+        _descriptor.MethodDescriptor(
+            name="Exchange",
+            full_name="communicator_objects.UnityToExternal.Exchange",
+            index=0,
+            containing_service=None,
+            input_type=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2._UNITYMESSAGE,
+            output_type=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2._UNITYMESSAGE,
+            serialized_options=None,
+        )
+    ],
+)
-DESCRIPTOR.services_by_name['UnityToExternal'] = _UNITYTOEXTERNAL
+DESCRIPTOR.services_by_name["UnityToExternal"] = _UNITYTOEXTERNAL
 
 # @@protoc_insertion_point(module_scope)

ml-agents-envs/mlagents/envs/communicator_objects/unity_to_external_pb2_grpc.py

 # Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
 import grpc
 
-from mlagents.envs.communicator_objects import unity_message_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2
+from mlagents.envs.communicator_objects import (
+    unity_message_pb2 as mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2,
+)
-  # missing associated documentation comment in .proto file
-  pass
+    # missing associated documentation comment in .proto file
+    pass
-  def __init__(self, channel):
-    """Constructor.
+    def __init__(self, channel):
+        """Constructor.
-    self.Exchange = channel.unary_unary(
-        '/communicator_objects.UnityToExternal/Exchange',
-        request_serializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.SerializeToString,
-        response_deserializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.FromString,
+        self.Exchange = channel.unary_unary(
+            "/communicator_objects.UnityToExternal/Exchange",
+            request_serializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.SerializeToString,
+            response_deserializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.FromString,
-  # missing associated documentation comment in .proto file
-  pass
+    # missing associated documentation comment in .proto file
+    pass
-  def Exchange(self, request, context):
-    """Sends the academy parameters
+    def Exchange(self, request, context):
+        """Sends the academy parameters
-    context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-    context.set_details('Method not implemented!')
-    raise NotImplementedError('Method not implemented!')
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
-  rpc_method_handlers = {
-      'Exchange': grpc.unary_unary_rpc_method_handler(
-          servicer.Exchange,
-          request_deserializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.FromString,
-          response_serializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.SerializeToString,
-      ),
-  }
-  generic_handler = grpc.method_handlers_generic_handler(
-      'communicator_objects.UnityToExternal', rpc_method_handlers)
-  server.add_generic_rpc_handlers((generic_handler,))
+    rpc_method_handlers = {
+        "Exchange": grpc.unary_unary_rpc_method_handler(
+            servicer.Exchange,
+            request_deserializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.FromString,
+            response_serializer=mlagents_dot_envs_dot_communicator__objects_dot_unity__message__pb2.UnityMessage.SerializeToString,
+        )
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+        "communicator_objects.UnityToExternal", rpc_method_handlers
+    )
+    server.add_generic_rpc_handlers((generic_handler,))

ml-agents-envs/mlagents/envs/environment.py

 
 from mlagents.envs.base_unity_environment import BaseUnityEnvironment
 from .brain import AllBrainInfo, BrainInfo, BrainParameters
-from .exception import UnityEnvironmentException, UnityActionException, UnityTimeOutException
+from .exception import (
+    UnityEnvironmentException,
+    UnityActionException,
+    UnityTimeOutException,
+)
-from .communicator_objects import UnityRLInput, UnityRLOutput, AgentActionProto, \
-    EnvironmentParametersProto, UnityRLInitializationInput, UnityRLInitializationOutput, \
-    UnityInput, UnityOutput, CustomResetParameters, CustomAction
+from .communicator_objects import (
+    UnityRLInput,
+    UnityRLOutput,
+    AgentActionProto,
+    EnvironmentParametersProto,
+    UnityRLInitializationInput,
+    UnityRLInitializationOutput,
+    UnityInput,
+    UnityOutput,
+    CustomResetParameters,
+    CustomAction,
+)
 
 from .rpc_communicator import RpcCommunicator
 from sys import platform
     SINGLE_BRAIN_ACTION_TYPES = SCALAR_ACTION_TYPES + (list, np.ndarray)
     SINGLE_BRAIN_TEXT_TYPES = (str, list, np.ndarray)
 
-    def __init__(self,
-                 file_name: Optional[str] = None,
-                 worker_id: int = 0,
-                 base_port: int = 5005,
-                 seed: int = 0,
-                 docker_training: bool = False,
-                 no_graphics: bool = False,
-                 timeout_wait: int = 30):
+    def __init__(
+        self,
+        file_name: Optional[str] = None,
+        worker_id: int = 0,
+        base_port: int = 5005,
+        seed: int = 0,
+        docker_training: bool = False,
+        no_graphics: bool = False,
+        timeout_wait: int = 30,
+    ):
         """
         Starts a new unity environment and establishes a connection with the environment.
         Notice: Currently communication between Unity and Python takes place over an open socket without authentication.
         self.port = base_port + worker_id
         self._buffer_size = 12000
         self._version_ = "API-8"
-        self._loaded = False  # If true, this means the environment was successfully loaded
-        self.proc1 = None  # The process that is started. If None, no process was started
+        self._loaded = (
+            False
+        )  # If true, this means the environment was successfully loaded
+        self.proc1 = (
+            None
+        )  # The process that is started. If None, no process was started
         self.communicator = self.get_communicator(worker_id, base_port, timeout_wait)
 
         # If the environment name is None, a new environment will not be launched
             raise UnityEnvironmentException(
                 "If the environment name is None, "
-                "the worker-id must be 0 in order to connect with the Editor.")
+                "the worker-id must be 0 in order to connect with the Editor."
+            )
-            logger.info("Start training by pressing the Play button in the Unity Editor.")
+            logger.info(
+                "Start training by pressing the Play button in the Unity Editor."
+            )
-        rl_init_parameters_in = UnityRLInitializationInput(
-            seed=seed
-        )
+        rl_init_parameters_in = UnityRLInitializationInput(seed=seed)
         try:
             aca_params = self.send_academy_parameters(rl_init_parameters_in)
         except UnityTimeOutException:
             raise UnityEnvironmentException(
                 "The API number is not compatible between Unity and python. Python API : {0}, Unity API : "
                 "{1}.\nPlease go to https://github.com/Unity-Technologies/ml-agents to download the latest version "
-                "of ML-Agents.".format(self._version_, self._unity_version))
+                "of ML-Agents.".format(self._version_, self._unity_version)
+            )
         self._n_agents = {}
         self._global_done = None
         self._academy_name = aca_params.name
         self._external_brain_names = []
         for brain_param in aca_params.brain_parameters:
             self._brain_names += [brain_param.brain_name]
-            self._brains[brain_param.brain_name] = BrainParameters.from_proto(brain_param)
+            self._brains[brain_param.brain_name] = BrainParameters.from_proto(
+                brain_param
+            )
-        logger.info("\n'{0}' started successfully!\n{1}".format(self._academy_name, str(self)))
+        logger.info(
+            "\n'{0}' started successfully!\n{1}".format(self._academy_name, str(self))
+        )
-            logger.warning(" No Learning Brains set to train found in the Unity Environment. "
-                           "You will not be able to pass actions to your agent(s).")
+            logger.warning(
+                " No Learning Brains set to train found in the Unity Environment. "
+                "You will not be able to pass actions to your agent(s)."
+            )
 
     @property
     def logfile_path(self):
 
     def executable_launcher(self, file_name, docker_training, no_graphics):
         cwd = os.getcwd()
-        file_name = (file_name.strip()
-                     .replace('.app', '').replace('.exe', '').replace('.x86_64', '').replace('.x86',
-                                                                                             ''))
+        file_name = (
+            file_name.strip()
+            .replace(".app", "")
+            .replace(".exe", "")
+            .replace(".x86_64", "")
+            .replace(".x86", "")
+        )
-        logger.debug('The true file name is {}'.format(true_filename))
+        logger.debug("The true file name is {}".format(true_filename))
-            candidates = glob.glob(os.path.join(cwd, file_name) + '.x86_64')
+            candidates = glob.glob(os.path.join(cwd, file_name) + ".x86_64")
-                candidates = glob.glob(os.path.join(cwd, file_name) + '.x86')
+                candidates = glob.glob(os.path.join(cwd, file_name) + ".x86")
-                candidates = glob.glob(file_name + '.x86_64')
+                candidates = glob.glob(file_name + ".x86_64")
-                candidates = glob.glob(file_name + '.x86')
+                candidates = glob.glob(file_name + ".x86")
-        elif platform == 'darwin':
+        elif platform == "darwin":
-                os.path.join(cwd, file_name + '.app', 'Contents', 'MacOS', true_filename))
+                os.path.join(
+                    cwd, file_name + ".app", "Contents", "MacOS", true_filename
+                )
+            )
-                    os.path.join(file_name + '.app', 'Contents', 'MacOS', true_filename))
+                    os.path.join(file_name + ".app", "Contents", "MacOS", true_filename)
+                )
-                    os.path.join(cwd, file_name + '.app', 'Contents', 'MacOS', '*'))
+                    os.path.join(cwd, file_name + ".app", "Contents", "MacOS", "*")
+                )
-                candidates = glob.glob(os.path.join(file_name + '.app', 'Contents', 'MacOS', '*'))
+                candidates = glob.glob(
+                    os.path.join(file_name + ".app", "Contents", "MacOS", "*")
+                )
-        elif platform == 'win32':
-            candidates = glob.glob(os.path.join(cwd, file_name + '.exe'))
+        elif platform == "win32":
+            candidates = glob.glob(os.path.join(cwd, file_name + ".exe"))
-                candidates = glob.glob(file_name + '.exe')
+                candidates = glob.glob(file_name + ".exe")
-            raise UnityEnvironmentException("Couldn't launch the {0} environment. "
-                                            "Provided filename does not match any environments."
-                                            .format(true_filename))
+            raise UnityEnvironmentException(
+                "Couldn't launch the {0} environment. "
+                "Provided filename does not match any environments.".format(
+                    true_filename
+                )
+            )
         else:
             logger.debug("This is the launch string {}".format(launch_string))
             # Launch Unity environment
-                        [launch_string, '-nographics', '-batchmode',
-                         '--port', str(self.port)])
+                        [
+                            launch_string,
+                            "-nographics",
+                            "-batchmode",
+                            "--port",
+                            str(self.port),
+                        ]
+                    )
-                        [launch_string, '--port', str(self.port)])
+                        [launch_string, "--port", str(self.port)]
+                    )
             else:
                 """
                 Comments for future maintenance:
                     launched, the arguments are passed to `xvfb-run`. `exec` replaces the shell
                     we created with `xvfb`.
                 """
-                docker_ls = ("exec xvfb-run --auto-servernum"
-                             " --server-args='-screen 0 640x480x24'"
-                             " {0} --port {1}").format(launch_string, str(self.port))
-                self.proc1 = subprocess.Popen(docker_ls,
-                                              stdout=subprocess.PIPE,
-                                              stderr=subprocess.PIPE,
-                                              shell=True)
+                docker_ls = (
+                    "exec xvfb-run --auto-servernum"
+                    " --server-args='-screen 0 640x480x24'"
+                    " {0} --port {1}"
+                ).format(launch_string, str(self.port))
+                self.proc1 = subprocess.Popen(
+                    docker_ls,
+                    stdout=subprocess.PIPE,
+                    stderr=subprocess.PIPE,
+                    shell=True,
+                )
-        return '''Unity Academy name: {0}
+        return (
+            """Unity Academy name: {0}
-        Reset Parameters :\n\t\t{3}'''.format(self._academy_name, str(self._num_brains),
-                                              str(self._num_external_brains),
-                                              "\n\t\t".join(
-                                                  [str(k) + " -> " + str(self._resetParameters[k])
-                                                   for k in self._resetParameters])) + '\n' + \
-               '\n'.join([str(self._brains[b]) for b in self._brains])
+        Reset Parameters :\n\t\t{3}""".format(
+                self._academy_name,
+                str(self._num_brains),
+                str(self._num_external_brains),
+                "\n\t\t".join(
+                    [
+                        str(k) + " -> " + str(self._resetParameters[k])
+                        for k in self._resetParameters
+                    ]
+                ),
+            )
+            + "\n"
+            + "\n".join([str(self._brains[b]) for b in self._brains])
+        )
-    def reset(self, config=None, train_mode=True, custom_reset_parameters=None) -> AllBrainInfo:
+    def reset(
+        self, config=None, train_mode=True, custom_reset_parameters=None
+    ) -> AllBrainInfo:
         """
         Sends a signal to reset the unity environment.
         :return: AllBrainInfo  : A data structure corresponding to the initial reset state of the environment.
         elif config:
-            logger.info("Academy reset with parameters: {0}"
-                        .format(', '.join([str(x) + ' -> ' + str(config[x]) for x in config])))
+            logger.info(
+                "Academy reset with parameters: {0}".format(
+                    ", ".join([str(x) + " -> " + str(config[x]) for x in config])
+                )
+            )
-                    "The value for parameter '{0}'' must be an Integer or a Float.".format(k))
+                    "The value for parameter '{0}'' must be an Integer or a Float.".format(
+                        k
+                    )
+                )
-                    "The parameter '{0}' is not a valid parameter.".format(k))
+                    "The parameter '{0}' is not a valid parameter.".format(k)
+                )
 
         if self._loaded:
             outputs = self.communicator.exchange(
         else:
             raise UnityEnvironmentException("No Unity environment is loaded.")
 
-    def step(self, vector_action=None, memory=None, text_action=None, value=None, custom_action=None) -> AllBrainInfo:
+    def step(
+        self,
+        vector_action=None,
+        memory=None,
+        text_action=None,
+        value=None,
+        custom_action=None,
+    ) -> AllBrainInfo:
         """
         Provides the environment with an action, moves the environment dynamics forward accordingly,
         and returns observation, state, and reward information to the agent.
                 elif self._num_external_brains > 1:
                     raise UnityActionException(
                         "You have {0} brains, you need to feed a dictionary of brain names a keys, "
-                        "and vector_actions as values".format(self._num_brains))
+                        "and vector_actions as values".format(self._num_brains)
+                    )
-                        "step cannot take a vector_action input")
+                        "step cannot take a vector_action input"
+                    )
 
             if isinstance(memory, self.SINGLE_BRAIN_ACTION_TYPES):
                 if self._num_external_brains == 1:
                         "You have {0} brains, you need to feed a dictionary of brain names as keys "
-                        "and memories as values".format(self._num_brains))
+                        "and memories as values".format(self._num_brains)
+                    )
-                        "step cannot take a memory input")
+                        "step cannot take a memory input"
+                    )
 
             if isinstance(text_action, self.SINGLE_BRAIN_TEXT_TYPES):
                 if self._num_external_brains == 1:
                         "You have {0} brains, you need to feed a dictionary of brain names as keys "
-                        "and text_actions as values".format(self._num_brains))
+                        "and text_actions as values".format(self._num_brains)
+                    )
-                        "step cannot take a value input")
+                        "step cannot take a value input"
+                    )
 
             if isinstance(value, self.SINGLE_BRAIN_ACTION_TYPES):
                 if self._num_external_brains == 1:
                         "You have {0} brains, you need to feed a dictionary of brain names as keys "
-                        "and state/action value estimates as values".format(self._num_brains))
+                        "and state/action value estimates as values".format(
+                            self._num_brains
+                        )
+                    )
-                        "step cannot take a value input")
+                        "step cannot take a value input"
+                    )
 
             if isinstance(custom_action, CustomAction):
                 if self._num_external_brains == 1:
                         "You have {0} brains, you need to feed a dictionary of brain names as keys "
-                        "and CustomAction instances as values".format(self._num_brains))
+                        "and CustomAction instances as values".format(self._num_brains)
+                    )
-                        "step cannot take a custom_action input")
+                        "step cannot take a custom_action input"
+                    )
-            for brain_name in list(vector_action.keys()) + list(memory.keys()) + list(
-                    text_action.keys()):
+            for brain_name in (
+                list(vector_action.keys())
+                + list(memory.keys())
+                + list(text_action.keys())
+            ):
-                        "in the environment".format(brain_name))
+                        "in the environment".format(brain_name)
+                    )
-                        vector_action[brain_name] = [0.0] * n_agent * len(
-                            self._brains[brain_name].vector_action_space_size)
+                        vector_action[brain_name] = (
+                            [0.0]
+                            * n_agent
+                            * len(self._brains[brain_name].vector_action_space_size)
+                        )
-                        vector_action[brain_name] = [0.0] * n_agent * \
-                                                    self._brains[
-                                                        brain_name].vector_action_space_size[0]
+                        vector_action[brain_name] = (
+                            [0.0]
+                            * n_agent
+                            * self._brains[brain_name].vector_action_space_size[0]
+                        )
                 else:
                     vector_action[brain_name] = self._flatten(vector_action[brain_name])
                 if brain_name not in memory:
                     if custom_action[brain_name] is None:
                         custom_action[brain_name] = [None] * n_agent
                     if isinstance(custom_action[brain_name], CustomAction):
-                        custom_action[brain_name] = [custom_action[brain_name]] * n_agent
+                        custom_action[brain_name] = [
+                            custom_action[brain_name]
+                        ] * n_agent
 
                 number_text_actions = len(text_action[brain_name])
                 if not ((number_text_actions == n_agent) or number_text_actions == 0):
                         "The brain {0} expected {1} text_action but was given {2}".format(
-                            brain_name, n_agent, number_text_actions))
+                            brain_name, n_agent, number_text_actions
+                        )
+                    )
-                discrete_check = self._brains[brain_name].vector_action_space_type == "discrete"
+                discrete_check = (
+                    self._brains[brain_name].vector_action_space_type == "discrete"
+                )
-                    self._brains[brain_name].vector_action_space_size)
+                    self._brains[brain_name].vector_action_space_size
+                )
-                continuous_check = self._brains[brain_name].vector_action_space_type == "continuous"
+                continuous_check = (
+                    self._brains[brain_name].vector_action_space_type == "continuous"
+                )
-                expected_continuous_size = self._brains[brain_name].vector_action_space_size[
-                                               0] * n_agent
+                expected_continuous_size = (
+                    self._brains[brain_name].vector_action_space_size[0] * n_agent
+                )
-                if not ((discrete_check and len(
-                        vector_action[brain_name]) == expected_discrete_size) or
-                        (continuous_check and len(
-                            vector_action[brain_name]) == expected_continuous_size)):
+                if not (
+                    (
+                        discrete_check
+                        and len(vector_action[brain_name]) == expected_discrete_size
+                    )
+                    or (
+                        continuous_check
+                        and len(vector_action[brain_name]) == expected_continuous_size
+                    )
+                ):
-                        "The brain {0} expected {1} {2} action(s), but was provided: {3}"
-                            .format(brain_name, str(expected_discrete_size)
-                        if discrete_check
-                        else str(expected_continuous_size),
-                                    self._brains[brain_name].vector_action_space_type,
-                                    str(vector_action[brain_name])))
+                        "The brain {0} expected {1} {2} action(s), but was provided: {3}".format(
+                            brain_name,
+                            str(expected_discrete_size)
+                            if discrete_check
+                            else str(expected_continuous_size),
+                            self._brains[brain_name].vector_action_space_type,
+                            str(vector_action[brain_name]),
+                        )
+                    )
-                self._generate_step_input(vector_action, memory, text_action, value, custom_action))
+                self._generate_step_input(
+                    vector_action, memory, text_action, value, custom_action
+                )
+            )
             if outputs is None:
                 raise KeyboardInterrupt
             rl_output = outputs.rl_output
             raise UnityEnvironmentException("No Unity environment is loaded.")
         elif self._global_done:
             raise UnityActionException(
-                "The episode is completed. Reset the environment with 'reset()'")
+                "The episode is completed. Reset the environment with 'reset()'"
+            )
-                "Reset the environment with 'reset()'")
+                "Reset the environment with 'reset()'"
+            )
 
     def close(self):
         """
         global_done = output.global_done
         for brain_name in output.agentInfos:
             agent_info_list = output.agentInfos[brain_name].value
-            _data[brain_name] = BrainInfo.from_agent_proto(agent_info_list,
-                                                           self.brains[brain_name])
+            _data[brain_name] = BrainInfo.from_agent_proto(
+                agent_info_list, self.brains[brain_name]
+            )
-    def _generate_step_input(self, vector_action, memory, text_action, value, custom_action) -> UnityRLInput:
+    def _generate_step_input(
+        self, vector_action, memory, text_action, value, custom_action
+    ) -> UnityRLInput:
         rl_in = UnityRLInput()
         for b in vector_action:
             n_agents = self._n_agents[b]
             _m_s = len(memory[b]) // n_agents
             for i in range(n_agents):
                 action = AgentActionProto(
-                    vector_actions=vector_action[b][i * _a_s: (i + 1) * _a_s],
-                    memories=memory[b][i * _m_s: (i + 1) * _m_s],
+                    vector_actions=vector_action[b][i * _a_s : (i + 1) * _a_s],
+                    memories=memory[b][i * _m_s : (i + 1) * _m_s],
-                    custom_action=custom_action[b][i]
+                    custom_action=custom_action[b][i],
                 )
                 if b in value:
                     if value[b] is not None:
         return self.wrap_unity_input(rl_in)
 
-    def _generate_reset_input(self, training, config, custom_reset_parameters) -> UnityRLInput:
+    def _generate_reset_input(
+        self, training, config, custom_reset_parameters
+    ) -> UnityRLInput:
         rl_in = UnityRLInput()
         rl_in.is_training = training
         rl_in.environment_parameters.CopyFrom(EnvironmentParametersProto())
-            rl_in.environment_parameters.custom_reset_parameters.CopyFrom(custom_reset_parameters)
+            rl_in.environment_parameters.custom_reset_parameters.CopyFrom(
+                custom_reset_parameters
+            )
-    def send_academy_parameters(self,
-                                init_parameters: UnityRLInitializationInput) -> UnityRLInitializationOutput:
+    def send_academy_parameters(
+        self, init_parameters: UnityRLInitializationInput
+    ) -> UnityRLInitializationOutput:
         inputs = UnityInput()
         inputs.rl_initialization_input.CopyFrom(init_parameters)
         return self.communicator.initialize(inputs).rl_initialization_output

ml-agents-envs/mlagents/envs/exception.py

 import logging
+
+
+
+
+
     pass
 
 
     """
+
+
-    def __init__(self, message, log_file_path = None):
+
+    def __init__(self, message, log_file_path=None):
-                    unity_error = '\n'
+                    unity_error = "\n"
-                        l=l.strip()
-                        if (l == 'Exception') or (l=='Error'):
+                        l = l.strip()
+                        if (l == "Exception") or (l == "Error"):
-                            unity_error += '----------------------\n'
-                        if (l == ''):
+                            unity_error += "----------------------\n"
+                        if l == "":
-                            unity_error += l + '\n'
+                            unity_error += l + "\n"
-                    logger.error("An error might have occured in the environment. "
-                        "You can check the logfile for more information at {}".format(log_file_path))
+                    logger.error(
+                        "An error might have occured in the environment. "
+                        "You can check the logfile for more information at {}".format(
+                            log_file_path
+                        )
+                    )
-                logger.error("An error might have occured in the environment. "
-                             "No UnitySDK.log file could be found.")
+                logger.error(
+                    "An error might have occured in the environment. "
+                    "No UnitySDK.log file could be found."
+                )
         super(UnityTimeOutException, self).__init__(message)
 
 
     MESSAGE_TEMPLATE = (
         "Couldn't start socket communication because worker number {} is still in use. "
         "You may need to manually close a previously opened environment "
-        "or use a different worker number.")
+        "or use a different worker number."
+    )
 
     def __init__(self, worker_id):
         message = self.MESSAGE_TEMPLATE.format(str(worker_id))

ml-agents-envs/mlagents/envs/mock_communicator.py

 from .communicator import Communicator
-from .communicator_objects import UnityOutput, UnityInput, \
-    ResolutionProto, BrainParametersProto, UnityRLInitializationOutput, \
-    AgentInfoProto, UnityRLOutput
+from .communicator_objects import (
+    UnityOutput,
+    UnityInput,
+    ResolutionProto,
+    BrainParametersProto,
+    UnityRLInitializationOutput,
+    AgentInfoProto,
+    UnityRLOutput,
+)
-    def __init__(self, discrete_action=False, visual_inputs=0, stack=True, num_agents=3,
-                 brain_name="RealFakeBrain", vec_obs_size=3):
+    def __init__(
+        self,
+        discrete_action=False,
+        visual_inputs=0,
+        stack=True,
+        num_agents=3,
+        brain_name="RealFakeBrain",
+        vec_obs_size=3,
+    ):
         """
         Python side of the grpc communication. Python is the client and Unity the server
 
             self.num_stacks = 1
 
     def initialize(self, inputs: UnityInput) -> UnityOutput:
-        resolutions = [ResolutionProto(
-            width=30,
-            height=40,
-            gray_scale=False) for i in range(self.visual_inputs)]
+        resolutions = [
+            ResolutionProto(width=30, height=40, gray_scale=False)
+            for i in range(self.visual_inputs)
+        ]
         bp = BrainParametersProto(
             vector_observation_size=self.vec_obs_size,
             num_stacked_vector_observations=self.num_stacks,
             vector_action_space_type=int(not self.is_discrete),
             brain_name=self.brain_name,
-            is_training=True
+            is_training=True,
-            name="RealFakeAcademy",
-            version="API-8",
-            log_path="",
-            brain_parameters=[bp]
+            name="RealFakeAcademy", version="API-8", log_path="", brain_parameters=[bp]
-        return UnityOutput(
-            rl_initialization_output=rl_init
-        )
+        return UnityOutput(rl_initialization_output=rl_init)
 
     def exchange(self, inputs: UnityInput) -> UnityOutput:
         dict_agent_info = {}
                     memories=[],
                     done=(i == 2),
                     max_step_reached=False,
-                    id=i
-                ))
-        dict_agent_info["RealFakeBrain"] = \
-            UnityRLOutput.ListAgentInfoProto(value=list_agent_info)
+                    id=i,
+                )
+            )
+        dict_agent_info["RealFakeBrain"] = UnityRLOutput.ListAgentInfoProto(
+            value=list_agent_info
+        )
-            global_done = (fake_brain.value[0].vector_actions[0] == -1)
+            global_done = fake_brain.value[0].vector_actions[0] == -1
-        result = UnityRLOutput(
-            global_done=global_done,
-            agentInfos=dict_agent_info
-        )
-        return UnityOutput(
-            rl_output=result
-        )
+        result = UnityRLOutput(global_done=global_done, agentInfos=dict_agent_info)
+        return UnityOutput(rl_output=result)
 
     def close(self):
         """

ml-agents-envs/mlagents/envs/rpc_communicator.py

 from concurrent.futures import ThreadPoolExecutor
 
 from .communicator import Communicator
-from .communicator_objects import UnityToExternalServicer, add_UnityToExternalServicer_to_server
+from .communicator_objects import (
+    UnityToExternalServicer,
+    add_UnityToExternalServicer_to_server,
+)
 from .communicator_objects import UnityMessage, UnityInput, UnityOutput
 from .exception import UnityTimeOutException, UnityWorkerInUseException
 
             add_UnityToExternalServicer_to_server(self.unity_to_external, self.server)
             # Using unspecified address, which means that grpc is communicating on all IPs
             # This is so that the docker container can connect.
-            self.server.add_insecure_port('[::]:' + str(self.port))
+            self.server.add_insecure_port("[::]:" + str(self.port))
             self.server.start()
             self.is_open = True
         except:
                 "\t The environment does not need user interaction to launch\n"
                 "\t The Academy's Broadcast Hub is configured correctly\n"
                 "\t The Agents are linked to the appropriate Brains\n"
-                "\t The environment and the Python interface have compatible versions.")
+                "\t The environment and the Python interface have compatible versions."
+            )
         aca_param = self.unity_to_external.parent_conn.recv().unity_output
         message = UnityMessage()
         message.header.status = 200

ml-agents-envs/mlagents/envs/socket_communicator.py

 
 
 class SocketCommunicator(Communicator):
-    def __init__(self, worker_id=0,
-                 base_port=5005):
+    def __init__(self, worker_id=0, base_port=5005):
         """
         Python side of the socket communication
 
             self._socket.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
             self._socket.bind(("localhost", self.port))
         except:
-            raise UnityTimeOutException("Couldn't start socket communication because worker number {} is still in use. "
-                                        "You may need to manually close a previously opened environment "
-                                        "or use a different worker number.".format(str(self.worker_id)))
+            raise UnityTimeOutException(
+                "Couldn't start socket communication because worker number {} is still in use. "
+                "You may need to manually close a previously opened environment "
+                "or use a different worker number.".format(str(self.worker_id))
+            )
-        except :
+        except:
-                "\t The environment and the Python interface have compatible versions.")
+                "\t The environment and the Python interface have compatible versions."
+            )
         message = UnityMessage()
         message.header.status = 200
         message.unity_input.CopyFrom(inputs)
         if self._socket is not None:
             self._conn.close()
             self._conn = None
-

ml-agents-envs/mlagents/envs/subprocess_environment.py

 
     def close(self):
         try:
-            self.conn.send(EnvironmentCommand('close'))
+            self.conn.send(EnvironmentCommand("close"))
         except (BrokenPipeError, EOFError):
             pass
         self.process.join()
-    env_factory: Callable[[int], UnityEnvironment] = cloudpickle.loads(pickled_env_factory)
+    env_factory: Callable[[int], UnityEnvironment] = cloudpickle.loads(
+        pickled_env_factory
+    )
-        parent_conn.send(
-            EnvironmentResponse(cmd_name, worker_id, payload)
-        )
+        parent_conn.send(EnvironmentResponse(cmd_name, worker_id, payload))
+
-            if cmd.name == 'step':
+            if cmd.name == "step":
-                _send_response('step', all_brain_info)
-            elif cmd.name == 'external_brains':
-                _send_response('external_brains', env.external_brains)
-            elif cmd.name == 'reset_parameters':
-                _send_response('reset_parameters', env.reset_parameters)
-            elif cmd.name == 'reset':
+                _send_response("step", all_brain_info)
+            elif cmd.name == "external_brains":
+                _send_response("external_brains", env.external_brains)
+            elif cmd.name == "reset_parameters":
+                _send_response("reset_parameters", env.reset_parameters)
+            elif cmd.name == "reset":
-                _send_response('reset', all_brain_info)
-            elif cmd.name == 'global_done':
-                _send_response('global_done', env.global_done)
-            elif cmd.name == 'close':
+                _send_response("reset", all_brain_info)
+            elif cmd.name == "global_done":
+                _send_response("global_done", env.global_done)
+            elif cmd.name == "close":
-        print('UnityEnvironment worker: keyboard interrupt')
+        print("UnityEnvironment worker: keyboard interrupt")
-    def __init__(self,
-                 env_factory: Callable[[int], BaseUnityEnvironment],
-                 n_env: int = 1):
+    def __init__(
+        self, env_factory: Callable[[int], BaseUnityEnvironment], n_env: int = 1
+    ):
         self.envs = []
         self.env_agent_counts = {}
         self.waiting = False
     @staticmethod
     def create_worker(
-            worker_id: int,
-            env_factory: Callable[[int], BaseUnityEnvironment]
+        worker_id: int, env_factory: Callable[[int], BaseUnityEnvironment]
     ) -> UnityEnvWorker:
         parent_conn, child_conn = Pipe()
 
-        child_process = Process(target=worker, args=(child_conn, pickled_env_factory, worker_id))
+        child_process = Process(
+            target=worker, args=(child_conn, pickled_env_factory, worker_id)
+        )
-    def step_async(self, vector_action, memory=None, text_action=None, value=None) -> None:
+    def step_async(
+        self, vector_action, memory=None, text_action=None, value=None
+    ) -> None:
-                'Tried to take an environment step bore previous step has completed.'
+                "Tried to take an environment step bore previous step has completed."
             )
 
         agent_counts_cum = {}
                     start_ind = agent_counts_cum[brain_name][worker_id - 1]
                 end_ind = agent_counts_cum[brain_name][worker_id]
                 if vector_action.get(brain_name) is not None:
-                    env_actions[brain_name] = vector_action[brain_name][start_ind:end_ind]
+                    env_actions[brain_name] = vector_action[brain_name][
+                        start_ind:end_ind
+                    ]
-                    env_text_action[brain_name] = text_action[brain_name][start_ind:end_ind]
+                    env_text_action[brain_name] = text_action[brain_name][
+                        start_ind:end_ind
+                    ]
-            env.send('step', (env_actions, env_memory, env_text_action, env_value))
+            env.send("step", (env_actions, env_memory, env_text_action, env_value))
-            raise UnityEnvironmentException('Tried to await an environment step, but no async step was taken.')
+            raise UnityEnvironmentException(
+                "Tried to await an environment step, but no async step was taken."
+            )
 
         steps = [self.envs[i].recv() for i in range(len(self.envs))]
         self._get_agent_counts(map(lambda s: s.payload, steps))
 
-    def step(self, vector_action=None, memory=None, text_action=None, value=None) -> AllBrainInfo:
+    def step(
+        self, vector_action=None, memory=None, text_action=None, value=None
+    ) -> AllBrainInfo:
-        self._broadcast_message('reset', (config, train_mode))
+        self._broadcast_message("reset", (config, train_mode))
         reset_results = [self.envs[i].recv() for i in range(len(self.envs))]
         self._get_agent_counts(map(lambda r: r.payload, reset_results))
 
     def global_done(self):
-        self._broadcast_message('global_done')
+        self._broadcast_message("global_done")
         dones: List[EnvironmentResponse] = [
             self.envs[i].recv().payload for i in range(len(self.envs))
         ]
     def external_brains(self):
-        self.envs[0].send('external_brains')
+        self.envs[0].send("external_brains")
-        self.envs[0].send('reset_parameters')
+        self.envs[0].send("reset_parameters")
         return self.envs[0].recv().payload
 
     def close(self):
             all_brain_info: AllBrainInfo = env_step.payload
             for brain_name, brain_info in all_brain_info.items():
                 for i in range(len(brain_info.agents)):
-                    brain_info.agents[i] = str(env_step.worker_id) + '-' + str(brain_info.agents[i])
+                    brain_info.agents[i] = (
+                        str(env_step.worker_id) + "-" + str(brain_info.agents[i])
+                    )
                 if accumulated_brain_info:
                     accumulated_brain_info[brain_name].merge(brain_info)
             if not accumulated_brain_info:
-    def _broadcast_message(self, name: str, payload = None):
+    def _broadcast_message(self, name: str, payload=None):
-            env.send(name, payload)
+            env.send(name, payload)

ml-agents-envs/mlagents/envs/tests/test_envs.py

 
 import numpy as np
 
-from mlagents.envs import UnityEnvironment, UnityEnvironmentException, UnityActionException, \
-    BrainInfo
+from mlagents.envs import (
+    UnityEnvironment,
+    UnityEnvironmentException,
+    UnityActionException,
+    BrainInfo,
+)
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-        UnityEnvironment(' ')
+        UnityEnvironment(" ")
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-        discrete_action=False, visual_inputs=0)
-    env = UnityEnvironment(' ')
+        discrete_action=False, visual_inputs=0
+    )
+    env = UnityEnvironment(" ")
-    assert env.brain_names[0] == 'RealFakeBrain'
+    assert env.brain_names[0] == "RealFakeBrain"
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-        discrete_action=False, visual_inputs=0)
-    env = UnityEnvironment(' ')
-    brain = env.brains['RealFakeBrain']
+        discrete_action=False, visual_inputs=0
+    )
+    env = UnityEnvironment(" ")
+    brain = env.brains["RealFakeBrain"]
-    assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
-    assert isinstance(brain_info['RealFakeBrain'].visual_observations, list)
-    assert isinstance(brain_info['RealFakeBrain'].vector_observations, np.ndarray)
-    assert len(brain_info['RealFakeBrain'].visual_observations) == brain.number_visual_observations
-    assert len(brain_info['RealFakeBrain'].vector_observations) == \
-           len(brain_info['RealFakeBrain'].agents)
-    assert len(brain_info['RealFakeBrain'].vector_observations[0]) == \
-           brain.vector_observation_space_size * brain.num_stacked_vector_observations
+    assert isinstance(brain_info["RealFakeBrain"], BrainInfo)
+    assert isinstance(brain_info["RealFakeBrain"].visual_observations, list)
+    assert isinstance(brain_info["RealFakeBrain"].vector_observations, np.ndarray)
+    assert (
+        len(brain_info["RealFakeBrain"].visual_observations)
+        == brain.number_visual_observations
+    )
+    assert len(brain_info["RealFakeBrain"].vector_observations) == len(
+        brain_info["RealFakeBrain"].agents
+    )
+    assert (
+        len(brain_info["RealFakeBrain"].vector_observations[0])
+        == brain.vector_observation_space_size * brain.num_stacked_vector_observations
+    )
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-        discrete_action=False, visual_inputs=0)
-    env = UnityEnvironment(' ')
-    brain = env.brains['RealFakeBrain']
+        discrete_action=False, visual_inputs=0
+    )
+    env = UnityEnvironment(" ")
+    brain = env.brains["RealFakeBrain"]
-    brain_info = env.step([0] * brain.vector_action_space_size[0] * len(brain_info['RealFakeBrain'].agents))
+    brain_info = env.step(
+        [0]
+        * brain.vector_action_space_size[0]
+        * len(brain_info["RealFakeBrain"].agents)
+    )
-    brain_info = env.step([-1] * brain.vector_action_space_size[0] * len(brain_info['RealFakeBrain'].agents))
+    brain_info = env.step(
+        [-1]
+        * brain.vector_action_space_size[0]
+        * len(brain_info["RealFakeBrain"].agents)
+    )
-        env.step([0] * brain.vector_action_space_size[0] * len(brain_info['RealFakeBrain'].agents))
+        env.step(
+            [0]
+            * brain.vector_action_space_size[0]
+            * len(brain_info["RealFakeBrain"].agents)
+        )
-    assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
-    assert isinstance(brain_info['RealFakeBrain'].visual_observations, list)
-    assert isinstance(brain_info['RealFakeBrain'].vector_observations, np.ndarray)
-    assert len(brain_info['RealFakeBrain'].visual_observations) == brain.number_visual_observations
-    assert len(brain_info['RealFakeBrain'].vector_observations) == \
-           len(brain_info['RealFakeBrain'].agents)
-    assert len(brain_info['RealFakeBrain'].vector_observations[0]) == \
-           brain.vector_observation_space_size * brain.num_stacked_vector_observations
+    assert isinstance(brain_info["RealFakeBrain"], BrainInfo)
+    assert isinstance(brain_info["RealFakeBrain"].visual_observations, list)
+    assert isinstance(brain_info["RealFakeBrain"].vector_observations, np.ndarray)
+    assert (
+        len(brain_info["RealFakeBrain"].visual_observations)
+        == brain.number_visual_observations
+    )
+    assert len(brain_info["RealFakeBrain"].vector_observations) == len(
+        brain_info["RealFakeBrain"].agents
+    )
+    assert (
+        len(brain_info["RealFakeBrain"].vector_observations[0])
+        == brain.vector_observation_space_size * brain.num_stacked_vector_observations
+    )
-    print("\n\n\n\n\n\n\n" + str(brain_info['RealFakeBrain'].local_done))
-    assert not brain_info['RealFakeBrain'].local_done[0]
-    assert brain_info['RealFakeBrain'].local_done[2]
+    print("\n\n\n\n\n\n\n" + str(brain_info["RealFakeBrain"].local_done))
+    assert not brain_info["RealFakeBrain"].local_done[0]
+    assert brain_info["RealFakeBrain"].local_done[2]
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-    comm = MockCommunicator(
-        discrete_action=False, visual_inputs=0)
+    comm = MockCommunicator(discrete_action=False, visual_inputs=0)
-    env = UnityEnvironment(' ')
+    env = UnityEnvironment(" ")
     assert env._loaded
     env.close()
     assert not env._loaded
-if __name__ == '__main__':
+if __name__ == "__main__":
     pytest.main()

ml-agents-envs/mlagents/envs/tests/test_rpc_communicator.py

     second_comm = RpcCommunicator(worker_id=1)
     first_comm.close()
     second_comm.close()
-

ml-agents-envs/mlagents/envs/tests/test_subprocess_unity_environment.py

         SubprocessUnityEnvironment.create_worker = MagicMock()
         env = SubprocessUnityEnvironment(mock_env_factory, 2)
         # Creates two processes
-        self.assertEqual(env.create_worker.call_args_list, [
-            mock.call(0, mock_env_factory),
-            mock.call(1, mock_env_factory)
-        ])
+        self.assertEqual(
+            env.create_worker.call_args_list,
+            [mock.call(0, mock_env_factory), mock.call(1, mock_env_factory)],
+        )
         self.assertEqual(len(env.envs), 2)
 
     def test_step_async_fails_when_waiting(self):
     @staticmethod
     def test_step_async_splits_input_by_agent_count():
         env = SubprocessUnityEnvironment(mock_env_factory, 0)
-        env.env_agent_counts = {
-            'MockBrain': [1, 3, 5]
-        }
-        env.envs = [
-            MockEnvWorker(0),
-            MockEnvWorker(1),
-            MockEnvWorker(2),
-        ]
+        env.env_agent_counts = {"MockBrain": [1, 3, 5]}
+        env.envs = [MockEnvWorker(0), MockEnvWorker(1), MockEnvWorker(2)]
-        env_1_actions = ([[3.0, 4.0]] * 3)
-        env_2_actions = ([[5.0, 6.0]] * 5)
-        vector_action = {
-            'MockBrain': env_0_actions + env_1_actions + env_2_actions
-        }
+        env_1_actions = [[3.0, 4.0]] * 3
+        env_2_actions = [[5.0, 6.0]] * 5
+        vector_action = {"MockBrain": env_0_actions + env_1_actions + env_2_actions}
-        env.envs[0].send.assert_called_with('step', ({'MockBrain': env_0_actions}, {}, {}, {}))
-        env.envs[1].send.assert_called_with('step', ({'MockBrain': env_1_actions}, {}, {}, {}))
-        env.envs[2].send.assert_called_with('step', ({'MockBrain': env_2_actions}, {}, {}, {}))
+        env.envs[0].send.assert_called_with(
+            "step", ({"MockBrain": env_0_actions}, {}, {}, {})
+        )
+        env.envs[1].send.assert_called_with(
+            "step", ({"MockBrain": env_1_actions}, {}, {}, {})
+        )
+        env.envs[2].send.assert_called_with(
+            "step", ({"MockBrain": env_2_actions}, {}, {}, {})
+        )
 
     def test_step_async_sets_waiting(self):
         env = SubprocessUnityEnvironment(mock_env_factory, 0)
 
     def test_step_await_combines_brain_info(self):
         all_brain_info_env0 = {
-            'MockBrain': BrainInfo([], [[1.0, 2.0], [1.0, 2.0]], [], agents=[1, 2], memory=np.zeros((0,0)))
+            "MockBrain": BrainInfo(
+                [], [[1.0, 2.0], [1.0, 2.0]], [], agents=[1, 2], memory=np.zeros((0, 0))
+            )
-            'MockBrain': BrainInfo([], [[3.0, 4.0]], [], agents=[3], memory=np.zeros((0,0)))
+            "MockBrain": BrainInfo(
+                [], [[3.0, 4.0]], [], agents=[3], memory=np.zeros((0, 0))
+            )
-        env_worker_0.recv.return_value = EnvironmentResponse('step', 0, all_brain_info_env0)
+        env_worker_0.recv.return_value = EnvironmentResponse(
+            "step", 0, all_brain_info_env0
+        )
-        env_worker_1.recv.return_value = EnvironmentResponse('step', 1, all_brain_info_env1)
+        env_worker_1.recv.return_value = EnvironmentResponse(
+            "step", 1, all_brain_info_env1
+        )
-        combined_braininfo = env.step_await()['MockBrain']
+        combined_braininfo = env.step_await()["MockBrain"]
-            [[1.0, 2.0], [1.0, 2.0], [3.0, 4.0]]
+            [[1.0, 2.0], [1.0, 2.0], [3.0, 4.0]],
-        self.assertEqual(combined_braininfo.agents, ['0-1', '0-2', '1-3'])
+        self.assertEqual(combined_braininfo.agents, ["0-1", "0-2", "1-3"])

ml-agents-envs/setup.py

 here = path.abspath(path.dirname(__file__))
 
 setup(
-    name='mlagents_envs',
-    version='0.8.1',
-    description='Unity Machine Learning Agents Interface',
-    url='https://github.com/Unity-Technologies/ml-agents',
-    author='Unity Technologies',
-    author_email='ML-Agents@unity3d.com',
-
+    name="mlagents_envs",
+    version="0.8.1",
+    description="Unity Machine Learning Agents Interface",
+    url="https://github.com/Unity-Technologies/ml-agents",
+    author="Unity Technologies",
+    author_email="ML-Agents@unity3d.com",
-        'Intended Audience :: Developers',
-        'Topic :: Scientific/Engineering :: Artificial Intelligence',
-        'License :: OSI Approved :: Apache Software License',
-        'Programming Language :: Python :: 3.6'
+        "Intended Audience :: Developers",
+        "Topic :: Scientific/Engineering :: Artificial Intelligence",
+        "License :: OSI Approved :: Apache Software License",
+        "Programming Language :: Python :: 3.6",
-
-    packages=['mlagents.envs', 'mlagents.envs.communicator_objects'],  # Required
+    packages=["mlagents.envs", "mlagents.envs.communicator_objects"],  # Required
-
-        'Pillow>=4.2.1,<=5.4.1',
-        'numpy>=1.13.3,<=1.16.1',
-        'pytest>=3.2.2,<4.0.0',
-        'protobuf>=3.6,<3.7',
-        'grpcio>=1.11.0,<1.12.0',
-        'cloudpickle==0.8.1'],
+        "Pillow>=4.2.1,<=5.4.1",
+        "numpy>=1.13.3,<=1.16.1",
+        "pytest>=3.2.2,<4.0.0",
+        "protobuf>=3.6,<3.7",
+        "grpcio>=1.11.0,<1.12.0",
+        "cloudpickle==0.8.1",
+    ],
     python_requires=">=3.5,<3.8",
 )

ml-agents/mlagents/trainers/__init__.py

 from .ppo.trainer import *
 from .ppo.policy import *
 from .exception import *
-from .demo_loader import *
+from .demo_loader import *

ml-agents/mlagents/trainers/barracuda.py

 from collections import defaultdict
 import numpy as np
 import json
-import struct # convert from Python values and C structs
+import struct  # convert from Python values and C structs
 import re
 import argparse
 import os.path
         self.globals = []
         self.memories = []
 
+
-    def __init__(self, **entries): self.__dict__.update(entries)
+
+    def __init__(self, **entries):
+        self.__dict__.update(entries)
+
-    parser.add_argument('source_file', help=help)
-    parser.add_argument('target_file', help='output Barracuda binary file')
-    parser.add_argument('-trim', '--trim-unused-by-output')
-    parser.add_argument('--print-layers', action='store_true')
-    parser.add_argument('--print-source-json', action='store_true')
-    parser.add_argument('-json', '--print-barracuda-json', action='store_true')
-    parser.add_argument('--print-layer-links', action='store_true')
-    parser.add_argument('--print-patterns', action='store_true')
-    parser.add_argument('--print-tensors', action='store_true')
-    parser.add_argument('--verbose', action='store_true')
+    parser.add_argument("source_file", help=help)
+    parser.add_argument("target_file", help="output Barracuda binary file")
+    parser.add_argument("-trim", "--trim-unused-by-output")
+    parser.add_argument("--print-layers", action="store_true")
+    parser.add_argument("--print-source-json", action="store_true")
+    parser.add_argument("-json", "--print-barracuda-json", action="store_true")
+    parser.add_argument("--print-layer-links", action="store_true")
+    parser.add_argument("--print-patterns", action="store_true")
+    parser.add_argument("--print-tensors", action="store_true")
+    parser.add_argument("--verbose", action="store_true")
-    args.compress_f16 = False # TEMP: disabled, until properly implemented parser.add_argument('-f16', '--compress-f16', action='store_true')
-    
-    output_extension = '.bc' if not args.compress_f16 else '.f16.bc'
+    args.compress_f16 = (
+        False
+    )  # TEMP: disabled, until properly implemented parser.add_argument('-f16', '--compress-f16', action='store_true')
+
+    output_extension = ".bc" if not args.compress_f16 else ".f16.bc"
-        print('File', args.source_file, 'does not exist.')
+        print("File", args.source_file, "does not exist.")
-        return os.path.splitext(os.path.basename(filename))[0] + newExtenstion;
+        return os.path.splitext(os.path.basename(filename))[0] + newExtenstion
-        args.target_file = os.path.join(args.target_file, replaceFilenameExtension(args.source_file, output_extension))
+        args.target_file = os.path.join(
+            args.target_file,
+            replaceFilenameExtension(args.source_file, output_extension),
+        )
+
 
 # Fuse training time BatchNorm tensors into Scale & Bias
 def fuse_batchnorm_weights(gamma, beta, mean, var, epsilon):
     bias = beta - gamma * mean / np.sqrt(var + epsilon)
     return [scale, bias]
 
+
-    if hasattr(model, 'layers'):
+    if hasattr(model, "layers"):
         model = model.layers
     inputs_and_memories = set(list(inputs) + list(memories[1::3]))
 
             ready.add(l.name)
         return missing
 
-    # Class to represent a graph 
+    # Class to represent a graph
-    class Graph: 
-        def __init__(self,vertices): 
-            self.graph = defaultdict(list) #dictionary containing adjacency List 
-            self.V = vertices #No. of vertices 
-      
-        # function to add an edge to graph 
-        def addEdge(self,u,v): 
-            self.graph[u].append(v) 
-      
-        # A recursive function used by topologicalSort 
-        def topologicalSortUtil(self,v,visited,stack): 
-      
-            # Mark the current node as visited. 
+    class Graph:
+        def __init__(self, vertices):
+            self.graph = defaultdict(list)  # dictionary containing adjacency List
+            self.V = vertices  # No. of vertices
+
+        # function to add an edge to graph
+        def addEdge(self, u, v):
+            self.graph[u].append(v)
+
+        # A recursive function used by topologicalSort
+        def topologicalSortUtil(self, v, visited, stack):
+
+            # Mark the current node as visited.
-      
-            # Recur for all the vertices adjacent to this vertex 
-            for i in self.graph[v]: 
-                if visited[i] == False: 
-                    self.topologicalSortUtil(i,visited,stack) 
-      
-            # Push current vertex to stack which stores result 
-            stack.insert(0,v) 
+
+            # Recur for all the vertices adjacent to this vertex
+            for i in self.graph[v]:
+                if visited[i] == False:
+                    self.topologicalSortUtil(i, visited, stack)
-        # The function to do Topological Sort. It uses recursive  
-        # topologicalSortUtil() 
-        def topologicalSort(self): 
-            # Mark all the vertices as not visited 
-            visited = [False]*self.V 
-            stack =[] 
-      
-            # Call the recursive helper function to store Topological 
-            # Sort starting from all vertices one by one 
-            for i in range(self.V): 
-                if visited[i] == False: 
-                    self.topologicalSortUtil(i,visited,stack) 
-      
-            #print(stack)
+            # Push current vertex to stack which stores result
+            stack.insert(0, v)
+
+        # The function to do Topological Sort. It uses recursive
+        # topologicalSortUtil()
+        def topologicalSort(self):
+            # Mark all the vertices as not visited
+            visited = [False] * self.V
+            stack = []
+
+            # Call the recursive helper function to store Topological
+            # Sort starting from all vertices one by one
+            for i in range(self.V):
+                if visited[i] == False:
+                    self.topologicalSortUtil(i, visited, stack)
+
+            # print(stack)
-    if (len(find_missing_inputs(model, inputs_and_memories)) == 0):
+    if len(find_missing_inputs(model, inputs_and_memories)) == 0:
         return model
 
     g = Graph(len(model))
     for l in model:
-        layers[l.name] = id;
+        layers[l.name] = id
         id += 1
 
     for layer in model:
     print("SORTED:", sorted_layer_indices)
     new_model = [model[idx] for idx in sorted_layer_indices]
 
-    assert(len(find_missing_inputs(new_model, inputs_and_memories)) == 0)
+    assert len(find_missing_inputs(new_model, inputs_and_memories)) == 0
-
-    if hasattr(model, 'layers'):
+    if hasattr(model, "layers"):
-    def flatten(items,enter=lambda x:isinstance(x, list)):
+    def flatten(items, enter=lambda x: isinstance(x, list)):
         # http://stackoverflow.com/a/40857703
         # https://github.com/ctmakro/canton/blob/master/canton/misc.py
         """Yield items from any nested iterable; see REF."""
                 yield x
 
     def trim_model(model, outputs):
-        layers = {l.name:l for l in model}
+        layers = {l.name: l for l in model}
         connected = {o for o in outputs}
         while len(outputs) > 0:
             outputs = set(flatten([layers[o].inputs for o in outputs if o in layers]))
                 connected.add(o)
 
         trimmed = [l.name for l in model if l.name not in connected]
+
-            return str(arr)[1:-1] # array to string without brackets
+            return str(arr)[1:-1]  # array to string without brackets
+
         print("TRIMMED:", array_without_brackets(trimmed))
 
         return [l for l in model if l.name in connected]
         print("Trimming model given outputs to preserve:", preserve_outputs)
         model = trim_model(model, preserve_outputs)
     else:
-        print("WARNING: Trim couldn't find any layers to match:", criteria_regexp_string)
+        print(
+            "WARNING: Trim couldn't find any layers to match:", criteria_regexp_string
+        )
+
-    compress_classes = {
-        'Dense'
-    }
+    compress_classes = {"Dense"}
-        if (l.class_name in compress_classes):
-            print("Compressing %s layer '%s' weights to float16" % (l.class_name, l.name))
+        if l.class_name in compress_classes:
+            print(
+                "Compressing %s layer '%s' weights to float16" % (l.class_name, l.name)
+            )
+
-            if isinstance(o, np.ndarray): # skip binary data packed inside ndarray
+            if isinstance(o, np.ndarray):  # skip binary data packed inside ndarray
-            if getattr(o, '__dict__', None):
+            if getattr(o, "__dict__", None):
-    s = json.dumps(model.layers, cls=StructEncoder, separators=(', ',':'))
+    s = json.dumps(model.layers, cls=StructEncoder, separators=(", ", ":"))
-    s = s.replace(']}, {', ']},\n{')
-    s = s.replace(':[{', ':[\n\t{')
-    s = s.replace('}, {', '},\n\t{')
+    s = s.replace("]}, {", "]},\n{")
+    s = s.replace(":[{", ":[\n\t{")
+    s = s.replace("}, {", "},\n\t{")
+
-        return str(arr)[1:-1] # array to string without brackets
+        return str(arr)[1:-1]  # array to string without brackets
 
     if print_layer_links:
         for l in model.layers:
 
     if model.globals:
         if isinstance(model.globals, dict):
-            model.globals = {x.name:x.shape for x in model.globals}
+            model.globals = {x.name: x.shape for x in model.globals}
-            ins = {i:model.inputs[i] for i in l.inputs if i in model.inputs}
+            ins = {i: model.inputs[i] for i in l.inputs if i in model.inputs}
         else:
             ins = [i for i in l.inputs if i in model.inputs]
         if ins:
     print("OUT:", array_without_brackets(model.outputs))
 
-    if (print_tensors):
+    if print_tensors:
+
-    def __init__(self, scope=''):
+    def __init__(self, scope=""):
-        if attr == '_':
+        if attr == "_":
             return self.layers[-1].name if len(self.layer) > 0 else self.scope
         raise AttributeError(attr)
 
 
         i = 1
         while name in self.names_taken:
-            name = self.layers[-1].op + '_' + str(i)
+            name = self.layers[-1].op + "_" + str(i)
-        self.layers[-1].name = self.scope + ('/' if self.scope else '') + name
+        self.layers[-1].name = self.scope + ("/" if self.scope else "") + name
-    def concat(self, a, b, out=''):
-        self.layers += [Struct(name=out, op='Concat', input=[a, b])]
+    def concat(self, a, b, out=""):
+        self.layers += [Struct(name=out, op="Concat", input=[a, b])]
-    def mad(self, x, kernel, bias, out=''):
-        self.layers += [Struct(name=out, op='Dense', input=[x, kernel, bias])]
+
+    def mad(self, x, kernel, bias, out=""):
+        self.layers += [Struct(name=out, op="Dense", input=[x, kernel, bias])]
-    def mul(self, a, b, out=''):
-        self.layers += [Struct(name=out, op='Mul', input=[a, b])]
+
+    def mul(self, a, b, out=""):
+        self.layers += [Struct(name=out, op="Mul", input=[a, b])]
-    def add(self, a, b, out=''):
-        self.layers += [Struct(name=out, op='Add', input=[a, b])]
+
+    def add(self, a, b, out=""):
+        self.layers += [Struct(name=out, op="Add", input=[a, b])]
-    def sub(self, a, b, out=''):
-        self.layers += [Struct(name=out, op='Sub', input=[a, b])]
+
+    def sub(self, a, b, out=""):
+        self.layers += [Struct(name=out, op="Sub", input=[a, b])]
-    def sigmoid(self, x, out=''):
-        self.layers += [Struct(name=out, op='Sigmoid', input=[x])]
+
+    def sigmoid(self, x, out=""):
+        self.layers += [Struct(name=out, op="Sigmoid", input=[x])]
-    def tanh(self, x, out=''):
-        self.layers += [Struct(name=out, op='Tanh', input=[x])]
+
+    def tanh(self, x, out=""):
+        self.layers += [Struct(name=out, op="Tanh", input=[x])]
-def rnn(name, input, state, kernel, bias, new_state, number_of_gates = 2):
-    ''' - Ht = f(Xt*Wi + Ht_1*Ri + Wbi + Rbi)
-    '''
+
+def rnn(name, input, state, kernel, bias, new_state, number_of_gates=2):
+    """ - Ht = f(Xt*Wi + Ht_1*Ri + Wbi + Rbi)
+    """
-    nn.tanh(
-        nn.mad(kernel=kernel, bias=bias,
-            x=nn.concat(input, state)),
-        out=new_state);
-    return nn.layers;
+    nn.tanh(nn.mad(kernel=kernel, bias=bias, x=nn.concat(input, state)), out=new_state)
+    return nn.layers
-def gru(name, input, state, kernel_r, kernel_u, kernel_c, bias_r, bias_u, bias_c, new_state, number_of_gates = 2):
-    ''' - zt = f(Xt*Wz + Ht_1*Rz        + Wbz + Rbz)
+
+def gru(
+    name,
+    input,
+    state,
+    kernel_r,
+    kernel_u,
+    kernel_c,
+    bias_r,
+    bias_u,
+    bias_c,
+    new_state,
+    number_of_gates=2,
+):
+    """ - zt = f(Xt*Wz + Ht_1*Rz        + Wbz + Rbz)
-    '''
+    """
     nn = Build(name)
     inputs = nn.concat(input, state)
 
 
-    c = nn.tanh(nn.mad(kernel=kernel_c, bias=bias_c,
-        x=nn.concat(input, r_state)))
+    c = nn.tanh(nn.mad(kernel=kernel_c, bias=bias_c, x=nn.concat(input, r_state)))
 
     # new_h = u' * state + (1 - u') * c'
     #       = u' * state + c' - u' * c'
     # - u' * c'
-    nn.sub(nn._, nn.mul(u, c),
-        out=new_state)
+    nn.sub(nn._, nn.mul(u, c), out=new_state)
+
+    return nn.layers
-    return nn.layers;
-def lstm(name, input, state_c, state_h, kernel_i, kernel_j, kernel_f, kernel_o, bias_i, bias_j, bias_f, bias_o, new_state_c, new_state_h):
-    ''' Full:
+def lstm(
+    name,
+    input,
+    state_c,
+    state_h,
+    kernel_i,
+    kernel_j,
+    kernel_f,
+    kernel_o,
+    bias_i,
+    bias_j,
+    bias_f,
+    bias_o,
+    new_state_c,
+    new_state_h,
+):
+    """ Full:
     - it = f(Xt*Wi + Ht_1*Ri + Pi . Ct_1 + Wbi + Rbi)
     - ft = f(Xt*Wf + Ht_1*Rf + Pf . Ct_1 + Wbf + Rbf)
     - ct = g(Xt*Wc + Ht_1*Rc + Wbc + Rbc)
-    '''
+    """
-    ''' No peephole:
+    """ No peephole:
     - it = f(Xt*Wi + Ht_1*Ri + Wbi + Rbi)
     - ft = f(Xt*Wf + Ht_1*Rf + Wbf + Rbf)
     - ct = g(Xt*Wc + Ht_1*Rc + Wbc + Rbc)
-    '''
+    """
-    j =    nn.tanh(nn.mad(inputs, kernel_j, bias_j))
+    j = nn.tanh(nn.mad(inputs, kernel_j, bias_j))
-    nn.add(
-        nn.mul(state_c, f), nn.mul(i, j),
-        out=new_state_c)
+    nn.add(nn.mul(state_c, f), nn.mul(i, j), out=new_state_c)
-    # new_h = 
-    nn.mul(o, nn.tanh(new_state_c),
-        out=new_state_h)
+    # new_h =
+    nn.mul(o, nn.tanh(new_state_c), out=new_state_h)
+
 
 # Serialize
 class BarracudaWriter:
-        self.f = open(filename, 'wb+')
+        self.f = open(filename, "wb+")
 
     def __enter__(self):
         return self
 
     def write_str(self, s):
         self.write_int32(len(s))
-        self.f.write(s.encode('ascii'))
+        self.f.write(s.encode("ascii"))
-        self.f.write(struct.pack('<f', d))
+        self.f.write(struct.pack("<f", d))
-        self.f.write(struct.pack('<i', d))
+        self.f.write(struct.pack("<i", d))
-        self.f.write(struct.pack('<q', d))
+        self.f.write(struct.pack("<q", d))
 
     def write_shape(self, s):
         self.write_int32(len(s))
     def close(self):
         self.f.close()
 
+
-        #VERSION = 0xBA22AC0DA000 + BARRACUDA_VERSION
+        # VERSION = 0xBA22AC0DA000 + BARRACUDA_VERSION
         w.write_int64(BARRACUDA_VERSION)
 
         # inputs
         w.write_str_array(model.outputs)
 
         # memories
-        w.write_int32(len(model.memories)//3)
-        for mem_shape, mem_in, mem_out in zip(model.memories[0::3], model.memories[1::3], model.memories[2::3]):
+        w.write_int32(len(model.memories) // 3)
+        for mem_shape, mem_in, mem_out in zip(
+            model.memories[0::3], model.memories[1::3], model.memories[2::3]
+        ):
             w.write_shape(mem_shape)
             w.write_str(mem_in)
             w.write_str(mem_out)
         w.write_int32(len(model.layers))
         for l in model.layers:
 
-            assert(not l.name in l.inputs)
+            assert not l.name in l.inputs
-            w.write_int32(0) #dummy
-            w.write_int32(0) #dummy
+            w.write_int32(0)  # dummy
+            w.write_int32(0)  # dummy
             w.write_shape(l.pads)
             w.write_shape(l.strides)
             w.write_shape(l.pool_size)
-            w.write_int32(0) #dummy
+            w.write_int32(0)  # dummy
-                assert(len(x.shape) == 4)
-                assert(x.data.nbytes % 4 == 0)
-                length = x.data.nbytes >> 2 # length is measured in float32s (at least for now)
+                assert len(x.shape) == 4
+                assert x.data.nbytes % 4 == 0
+                length = (
+                    x.data.nbytes >> 2
+                )  # length is measured in float32s (at least for now)
 
                 w.write_str(x.name)
                 w.write_shape(x.shape)
 
         for x in all_tensors:
             w.write_array(x.data)
-
-
-

ml-agents/mlagents/trainers/bc/models.py

 
 
 class BehavioralCloningModel(LearningModel):
-    def __init__(self, brain, h_size=128, lr=1e-4, n_layers=2, m_size=128,
-                 normalize=False, use_recurrent=False, seed=0):
+    def __init__(
+        self,
+        brain,
+        h_size=128,
+        lr=1e-4,
+        n_layers=2,
+        m_size=128,
+        normalize=False,
+        use_recurrent=False,
+        seed=0,
+    ):
-        self.dropout_rate = tf.placeholder(dtype=tf.float32, shape=[], name="dropout_rate")
+        self.dropout_rate = tf.placeholder(
+            dtype=tf.float32, shape=[], name="dropout_rate"
+        )
-            tf.Variable(self.m_size, name="memory_size", trainable=False, dtype=tf.int32)
-            self.memory_in = tf.placeholder(shape=[None, self.m_size], dtype=tf.float32, name='recurrent_in')
-            hidden_reg, self.memory_out = self.create_recurrent_encoder(hidden_reg, self.memory_in,
-                                                                        self.sequence_length)
-            self.memory_out = tf.identity(self.memory_out, name='recurrent_out')
+            tf.Variable(
+                self.m_size, name="memory_size", trainable=False, dtype=tf.int32
+            )
+            self.memory_in = tf.placeholder(
+                shape=[None, self.m_size], dtype=tf.float32, name="recurrent_in"
+            )
+            hidden_reg, self.memory_out = self.create_recurrent_encoder(
+                hidden_reg, self.memory_in, self.sequence_length
+            )
+            self.memory_out = tf.identity(self.memory_out, name="recurrent_out")
 
         if brain.vector_action_space_type == "discrete":
             policy_branches = []
                         size,
                         activation=None,
                         use_bias=False,
-                        kernel_initializer=c_layers.variance_scaling_initializer(factor=0.01)))
+                        kernel_initializer=c_layers.variance_scaling_initializer(
+                            factor=0.01
+                        ),
+                    )
+                )
-                [tf.nn.softmax(branch) for branch in policy_branches], axis=1, name="action_probs")
-            self.action_masks = tf.placeholder(shape=[None, sum(self.act_size)], dtype=tf.float32, name="action_masks")
+                [tf.nn.softmax(branch) for branch in policy_branches],
+                axis=1,
+                name="action_probs",
+            )
+            self.action_masks = tf.placeholder(
+                shape=[None, sum(self.act_size)], dtype=tf.float32, name="action_masks"
+            )
-                tf.concat(policy_branches, axis=1), self.action_masks, self.act_size)
-            tf.identity(normalized_logits, name='action')
+                tf.concat(policy_branches, axis=1), self.action_masks, self.act_size
+            )
+            tf.identity(normalized_logits, name="action")
-            self.true_action = tf.placeholder(shape=[None, len(policy_branches)], dtype=tf.int32, name="teacher_action")
-            self.action_oh = tf.concat([
-                tf.one_hot(self.true_action[:, i], self.act_size[i]) for i in range(len(self.act_size))], axis=1)
-            self.loss = tf.reduce_sum(-tf.log(self.action_probs + 1e-10) * self.action_oh)
-            self.action_percent = tf.reduce_mean(tf.cast(
-                tf.equal(tf.cast(tf.argmax(self.action_probs, axis=1), tf.int32), self.sample_action), tf.float32))
+            self.true_action = tf.placeholder(
+                shape=[None, len(policy_branches)],
+                dtype=tf.int32,
+                name="teacher_action",
+            )
+            self.action_oh = tf.concat(
+                [
+                    tf.one_hot(self.true_action[:, i], self.act_size[i])
+                    for i in range(len(self.act_size))
+                ],
+                axis=1,
+            )
+            self.loss = tf.reduce_sum(
+                -tf.log(self.action_probs + 1e-10) * self.action_oh
+            )
+            self.action_percent = tf.reduce_mean(
+                tf.cast(
+                    tf.equal(
+                        tf.cast(tf.argmax(self.action_probs, axis=1), tf.int32),
+                        self.sample_action,
+                    ),
+                    tf.float32,
+                )
+            )
-            self.policy = tf.layers.dense(hidden_reg, self.act_size[0], activation=None, use_bias=False, name='pre_action',
-                                          kernel_initializer=c_layers.variance_scaling_initializer(factor=0.01))
+            self.policy = tf.layers.dense(
+                hidden_reg,
+                self.act_size[0],
+                activation=None,
+                use_bias=False,
+                name="pre_action",
+                kernel_initializer=c_layers.variance_scaling_initializer(factor=0.01),
+            )
-            self.true_action = tf.placeholder(shape=[None, self.act_size[0]], dtype=tf.float32, name="teacher_action")
+            self.true_action = tf.placeholder(
+                shape=[None, self.act_size[0]], dtype=tf.float32, name="teacher_action"
+            )
-            self.loss = tf.reduce_sum(tf.squared_difference(self.clipped_true_action, self.sample_action))
+            self.loss = tf.reduce_sum(
+                tf.squared_difference(self.clipped_true_action, self.sample_action)
+            )
 
         optimizer = tf.train.AdamOptimizer(learning_rate=lr)
         self.update = optimizer.minimize(self.loss)

ml-agents/mlagents/trainers/bc/offline_trainer.py

 class OfflineBCTrainer(BCTrainer):
     """The OfflineBCTrainer is an implementation of Offline Behavioral Cloning."""
 
-    def __init__(self, brain, trainer_parameters, training, load, seed,
-                 run_id):
+    def __init__(self, brain, trainer_parameters, training, load, seed, run_id):
         """
         Responsible for collecting experiences and training PPO model.
         :param  trainer_parameters: The parameters for the trainer (dictionary).
         :param run_id: The The identifier of the current run
         """
         super(OfflineBCTrainer, self).__init__(
-            brain, trainer_parameters, training, load, seed, run_id)
+            brain, trainer_parameters, training, load, seed, run_id
+        )
-        self.param_keys = ['batch_size', 'summary_freq', 'max_steps',
-                           'batches_per_epoch', 'use_recurrent',
-                           'hidden_units', 'learning_rate', 'num_layers',
-                           'sequence_length', 'memory_size', 'model_path',
-                           'demo_path']
+        self.param_keys = [
+            "batch_size",
+            "summary_freq",
+            "max_steps",
+            "batches_per_epoch",
+            "use_recurrent",
+            "hidden_units",
+            "learning_rate",
+            "num_layers",
+            "sequence_length",
+            "memory_size",
+            "model_path",
+            "demo_path",
+        ]
-        self.batches_per_epoch = trainer_parameters['batches_per_epoch']
-        self.n_sequences = max(int(trainer_parameters['batch_size'] / self.policy.sequence_length),
-                               1)
+        self.batches_per_epoch = trainer_parameters["batches_per_epoch"]
+        self.n_sequences = max(
+            int(trainer_parameters["batch_size"] / self.policy.sequence_length), 1
+        )
-            trainer_parameters['demo_path'],
-            self.policy.sequence_length)
+            trainer_parameters["demo_path"], self.policy.sequence_length
+        )
-        policy_brain.pop('brain_name')
-        expert_brain.pop('brain_name')
+        policy_brain.pop("brain_name")
+        expert_brain.pop("brain_name")
-            raise UnityTrainerException("The provided demonstration is not compatible with the "
-                                        "brain being used for performance evaluation.")
+            raise UnityTrainerException(
+                "The provided demonstration is not compatible with the "
+                "brain being used for performance evaluation."
+            )
-        return '''Hyperparameters for the Imitation Trainer of brain {0}: \n{1}'''.format(
-            self.brain_name, '\n'.join(
-                ['\t{0}:\t{1}'.format(x, self.trainer_parameters[x]) for x in self.param_keys]))
+        return """Hyperparameters for the Imitation Trainer of brain {0}: \n{1}""".format(
+            self.brain_name,
+            "\n".join(
+                [
+                    "\t{0}:\t{1}".format(x, self.trainer_parameters[x])
+                    for x in self.param_keys
+                ]
+            ),
+        )

ml-agents/mlagents/trainers/bc/online_trainer.py

 class OnlineBCTrainer(BCTrainer):
     """The OnlineBCTrainer is an implementation of Online Behavioral Cloning."""
 
-    def __init__(self, brain, trainer_parameters, training, load, seed,
-                 run_id):
+    def __init__(self, brain, trainer_parameters, training, load, seed, run_id):
         """
         Responsible for collecting experiences and training PPO model.
         :param  trainer_parameters: The parameters for the trainer (dictionary).
         :param run_id: The The identifier of the current run
         """
-        super(OnlineBCTrainer, self).__init__(brain, trainer_parameters, training, load, seed,
-                                              run_id)
+        super(OnlineBCTrainer, self).__init__(
+            brain, trainer_parameters, training, load, seed, run_id
+        )
-        self.param_keys = ['brain_to_imitate', 'batch_size', 'time_horizon',
-                           'summary_freq', 'max_steps',
-                           'batches_per_epoch', 'use_recurrent',
-                           'hidden_units', 'learning_rate', 'num_layers',
-                           'sequence_length', 'memory_size', 'model_path']
+        self.param_keys = [
+            "brain_to_imitate",
+            "batch_size",
+            "time_horizon",
+            "summary_freq",
+            "max_steps",
+            "batches_per_epoch",
+            "use_recurrent",
+            "hidden_units",
+            "learning_rate",
+            "num_layers",
+            "sequence_length",
+            "memory_size",
+            "model_path",
+        ]
-        self.brain_to_imitate = trainer_parameters['brain_to_imitate']
-        self.batches_per_epoch = trainer_parameters['batches_per_epoch']
-        self.n_sequences = max(int(trainer_parameters['batch_size'] / self.policy.sequence_length),
-                               1)
+        self.brain_to_imitate = trainer_parameters["brain_to_imitate"]
+        self.batches_per_epoch = trainer_parameters["batches_per_epoch"]
+        self.n_sequences = max(
+            int(trainer_parameters["batch_size"] / self.policy.sequence_length), 1
+        )
-        return '''Hyperparameters for the Imitation Trainer of brain {0}: \n{1}'''.format(
-            self.brain_name, '\n'.join(
-                ['\t{0}:\t{1}'.format(x, self.trainer_parameters[x]) for x in self.param_keys]))
+        return """Hyperparameters for the Imitation Trainer of brain {0}: \n{1}""".format(
+            self.brain_name,
+            "\n".join(
+                [
+                    "\t{0}:\t{1}".format(x, self.trainer_parameters[x])
+                    for x in self.param_keys
+                ]
+            ),
+        )
-    def add_experiences(self, curr_info: AllBrainInfo, next_info: AllBrainInfo,
-                        take_action_outputs):
+    def add_experiences(
+        self, curr_info: AllBrainInfo, next_info: AllBrainInfo, take_action_outputs
+    ):
         """
         Adds experiences to each agent's experience history.
         :param curr_info: Current AllBrainInfo (Dictionary of all current brains and corresponding BrainInfo).
                 idx = stored_info_teacher.agents.index(agent_id)
                 next_idx = next_info_teacher.agents.index(agent_id)
                 if stored_info_teacher.text_observations[idx] != "":
-                    info_teacher_record, info_teacher_reset = \
+                    info_teacher_record, info_teacher_reset = (
-                    next_info_teacher_record, next_info_teacher_reset = \
-                    next_info_teacher.text_observations[idx]. \
-                        lower().split(",")
+                    )
+                    next_info_teacher_record, next_info_teacher_reset = (
+                        next_info_teacher.text_observations[idx].lower().split(",")
+                    )
                     if next_info_teacher_reset == "true":
                         self.demonstration_buffer.reset_update_buffer()
                 else:
                         for i in range(self.policy.vis_obs_size):
-                            self.demonstration_buffer[agent_id]['visual_obs%d' % i] \
-                                .append(stored_info_teacher.visual_observations[i][idx])
+                            self.demonstration_buffer[agent_id][
+                                "visual_obs%d" % i
+                            ].append(stored_info_teacher.visual_observations[i][idx])
-                            self.demonstration_buffer[agent_id]['vector_obs'] \
-                                .append(stored_info_teacher.vector_observations[idx])
+                            self.demonstration_buffer[agent_id]["vector_obs"].append(
+                                stored_info_teacher.vector_observations[idx]
+                            )
-                                    (len(stored_info_teacher.agents),
-                                     self.policy.m_size))
-                            self.demonstration_buffer[agent_id]['memory'].append(
-                                stored_info_teacher.memories[idx])
-                        self.demonstration_buffer[agent_id]['actions'].append(
-                            next_info_teacher.previous_vector_actions[next_idx])
+                                    (
+                                        len(stored_info_teacher.agents),
+                                        self.policy.m_size,
+                                    )
+                                )
+                            self.demonstration_buffer[agent_id]["memory"].append(
+                                stored_info_teacher.memories[idx]
+                            )
+                        self.demonstration_buffer[agent_id]["actions"].append(
+                            next_info_teacher.previous_vector_actions[next_idx]
+                        )
-        super(OnlineBCTrainer, self).add_experiences(curr_info, next_info, take_action_outputs)
+        super(OnlineBCTrainer, self).add_experiences(
+            curr_info, next_info, take_action_outputs
+        )
 
     def process_experiences(self, current_info: AllBrainInfo, next_info: AllBrainInfo):
         """
         """
         info_teacher = next_info[self.brain_to_imitate]
         for l in range(len(info_teacher.agents)):
-            teacher_action_list = len(self.demonstration_buffer[info_teacher.agents[l]]['actions'])
-            horizon_reached = teacher_action_list > self.trainer_parameters['time_horizon']
-            teacher_filled = len(self.demonstration_buffer[info_teacher.agents[l]]['actions']) > 0
+            teacher_action_list = len(
+                self.demonstration_buffer[info_teacher.agents[l]]["actions"]
+            )
+            horizon_reached = (
+                teacher_action_list > self.trainer_parameters["time_horizon"]
+            )
+            teacher_filled = (
+                len(self.demonstration_buffer[info_teacher.agents[l]]["actions"]) > 0
+            )
-                    agent_id, batch_size=None, training_length=self.policy.sequence_length)
+                    agent_id,
+                    batch_size=None,
+                    training_length=self.policy.sequence_length,
+                )
                 self.demonstration_buffer[agent_id].reset_agent()
 
         super(OnlineBCTrainer, self).process_experiences(current_info, next_info)

ml-agents/mlagents/trainers/bc/policy.py

         with self.graph.as_default():
             with self.graph.as_default():
                 self.model = BehavioralCloningModel(
-                    h_size=int(trainer_parameters['hidden_units']),
-                    lr=float(trainer_parameters['learning_rate']),
-                    n_layers=int(trainer_parameters['num_layers']),
+                    h_size=int(trainer_parameters["hidden_units"]),
+                    lr=float(trainer_parameters["learning_rate"]),
+                    n_layers=int(trainer_parameters["num_layers"]),
-                    use_recurrent=trainer_parameters['use_recurrent'],
+                    use_recurrent=trainer_parameters["use_recurrent"],
-                    seed=seed)
+                    seed=seed,
+                )
 
         if load:
             self._load_graph()
-        self.inference_dict = {'action': self.model.sample_action}
-        self.update_dict = {'policy_loss': self.model.loss,
-                            'update_batch': self.model.update}
+        self.inference_dict = {"action": self.model.sample_action}
+        self.update_dict = {
+            "policy_loss": self.model.loss,
+            "update_batch": self.model.update,
+        }
-            self.inference_dict['memory_out'] = self.model.memory_out
+            self.inference_dict["memory_out"] = self.model.memory_out
 
         self.evaluate_rate = 1.0
         self.update_rate = 0.5
         :param brain_info: BrainInfo input to network.
         :return: Results of evaluation.
         """
-        feed_dict = {self.model.dropout_rate: self.evaluate_rate,
-                     self.model.sequence_length: 1}
+        feed_dict = {
+            self.model.dropout_rate: self.evaluate_rate,
+            self.model.sequence_length: 1,
+        }
 
         feed_dict = self._fill_eval_dict(feed_dict, brain_info)
         if self.use_recurrent:
         :return: Results of update.
         """
 
-        feed_dict = {self.model.dropout_rate: self.update_rate,
-                     self.model.batch_size: num_sequences,
-                     self.model.sequence_length: self.sequence_length}
+        feed_dict = {
+            self.model.dropout_rate: self.update_rate,
+            self.model.batch_size: num_sequences,
+            self.model.sequence_length: self.sequence_length,
+        }
-            feed_dict[self.model.true_action] = mini_batch['actions']. \
-                reshape([-1, self.brain.vector_action_space_size[0]])
+            feed_dict[self.model.true_action] = mini_batch["actions"].reshape(
+                [-1, self.brain.vector_action_space_size[0]]
+            )
-            feed_dict[self.model.true_action] = mini_batch['actions'].reshape(
-                [-1, len(self.brain.vector_action_space_size)])
+            feed_dict[self.model.true_action] = mini_batch["actions"].reshape(
+                [-1, len(self.brain.vector_action_space_size)]
+            )
-                (num_sequences, sum(self.brain.vector_action_space_size)))
+                (num_sequences, sum(self.brain.vector_action_space_size))
+            )
-            apparent_obs_size = self.brain.vector_observation_space_size * \
-                                self.brain.num_stacked_vector_observations
-            feed_dict[self.model.vector_in] = mini_batch['vector_obs'] \
-                .reshape([-1,apparent_obs_size])
+            apparent_obs_size = (
+                self.brain.vector_observation_space_size
+                * self.brain.num_stacked_vector_observations
+            )
+            feed_dict[self.model.vector_in] = mini_batch["vector_obs"].reshape(
+                [-1, apparent_obs_size]
+            )
-            visual_obs = mini_batch['visual_obs%d' % i]
+            visual_obs = mini_batch["visual_obs%d" % i]
             feed_dict[self.model.visual_in[i]] = visual_obs
         if self.use_recurrent:
             feed_dict[self.model.memory_in] = np.zeros([num_sequences, self.m_size])

ml-agents/mlagents/trainers/bc/trainer.py

 class BCTrainer(Trainer):
     """The BCTrainer is an implementation of Behavioral Cloning."""
 
-    def __init__(self, brain, trainer_parameters, training, load, seed,
-                 run_id):
+    def __init__(self, brain, trainer_parameters, training, load, seed, run_id):
         """
         Responsible for collecting experiences and training PPO model.
         :param  trainer_parameters: The parameters for the trainer (dictionary).
         :param run_id: The The identifier of the current run
         """
-        super(BCTrainer, self).__init__(brain, trainer_parameters, training,
-                                        run_id)
+        super(BCTrainer, self).__init__(brain, trainer_parameters, training, run_id)
-        self.stats = {'Losses/Cloning Loss': [], 'Environment/Episode Length': [],
-                      'Environment/Cumulative Reward': []}
+        self.stats = {
+            "Losses/Cloning Loss": [],
+            "Environment/Episode Length": [],
+            "Environment/Cumulative Reward": [],
+        }
-        self.batches_per_epoch = trainer_parameters['batches_per_epoch']
-
+        self.batches_per_epoch = trainer_parameters["batches_per_epoch"]
 
         self.demonstration_buffer = Buffer()
         self.evaluation_buffer = Buffer()
         Returns the maximum number of steps. Is used to know when the trainer should be stopped.
         :return: The maximum number of steps of the trainer
         """
-        return float(self.trainer_parameters['max_steps'])
+        return float(self.trainer_parameters["max_steps"])
 
     @property
     def get_step(self):
         Returns the last reward the trainer has had
         :return: the new last reward
         """
-        if len(self.stats['Environment/Cumulative Reward']) > 0:
-            return np.mean(self.stats['Environment/Cumulative Reward'])
+        if len(self.stats["Environment/Cumulative Reward"]) > 0:
+            return np.mean(self.stats["Environment/Cumulative Reward"])
         else:
             return 0
 
         self.policy.increment_step()
         return
 
-    def add_experiences(self, curr_info: AllBrainInfo, next_info: AllBrainInfo,
-                        take_action_outputs):
+    def add_experiences(
+        self, curr_info: AllBrainInfo, next_info: AllBrainInfo, take_action_outputs
+    ):
         """
         Adds experiences to each agent's experience history.
         :param curr_info: Current AllBrainInfo (Dictionary of all current brains and corresponding BrainInfo).
         for l in range(len(info_student.agents)):
             if info_student.local_done[l]:
                 agent_id = info_student.agents[l]
-                self.stats['Environment/Cumulative Reward'].append(
-                    self.cumulative_rewards.get(agent_id, 0))
-                self.stats['Environment/Episode Length'].append(
-                    self.episode_steps.get(agent_id, 0))
+                self.stats["Environment/Cumulative Reward"].append(
+                    self.cumulative_rewards.get(agent_id, 0)
+                )
+                self.stats["Environment/Episode Length"].append(
+                    self.episode_steps.get(agent_id, 0)
+                )
                 self.cumulative_rewards[agent_id] = 0
                 self.episode_steps[agent_id] = 0
 
         Returns whether or not the trainer has enough elements to run update model
         :return: A boolean corresponding to whether or not update_model() can be run
         """
-        return len(self.demonstration_buffer.update_buffer['actions']) > self.n_sequences
+        return (
+            len(self.demonstration_buffer.update_buffer["actions"]) > self.n_sequences
+        )
 
     def update_policy(self):
         """
         batch_losses = []
-        num_batches = min(len(self.demonstration_buffer.update_buffer['actions']) //
-                          self.n_sequences, self.batches_per_epoch)
+        num_batches = min(
+            len(self.demonstration_buffer.update_buffer["actions"]) // self.n_sequences,
+            self.batches_per_epoch,
+        )
         for i in range(num_batches):
             update_buffer = self.demonstration_buffer.update_buffer
             start = i * self.n_sequences
-            loss = run_out['policy_loss']
+            loss = run_out["policy_loss"]
-            self.stats['Losses/Cloning Loss'].append(np.mean(batch_losses))
+            self.stats["Losses/Cloning Loss"].append(np.mean(batch_losses))
-            self.stats['Losses/Cloning Loss'].append(0)
+            self.stats["Losses/Cloning Loss"].append(0)

ml-agents/mlagents/trainers/buffer.py

     """
     Related to errors with the Buffer.
     """
+
     pass
 
 
                         # The number of sequences of length training_length taken from a list of len(self) elements
                         # with overlapping is equal to batch_size
                         if (len(self) - training_length + 1) < batch_size:
-                            raise BufferException("The batch size and training length requested for get_batch where"
-                                                  " too large given the current number of data points.")
+                            raise BufferException(
+                                "The batch size and training length requested for get_batch where"
+                                " too large given the current number of data points."
+                            )
-                            tmp_list += [np.array(self[end - training_length:end])]
+                            tmp_list += [np.array(self[end - training_length : end])]
                         return np.array(tmp_list)
                     if sequential:
                         # The sequences will not have overlapping elements (this involves padding)
                             # retrieve the maximum number of elements
-                            batch_size = len(self) // training_length + 1 * (leftover != 0)
+                            batch_size = len(self) // training_length + 1 * (
+                                leftover != 0
+                            )
-                        if batch_size > (len(self) // training_length + 1 * (leftover != 0)):
-                            raise BufferException("The batch size and training length requested for get_batch where"
-                                                  " too large given the current number of data points.")
+                        if batch_size > (
+                            len(self) // training_length + 1 * (leftover != 0)
+                        ):
+                            raise BufferException(
+                                "The batch size and training length requested for get_batch where"
+                                " too large given the current number of data points."
+                            )
-                        for end in range(len(self), len(self) % training_length, -training_length)[:batch_size]:
-                            tmp_list += [np.array(self[end - training_length:end])]
+                        for end in range(
+                            len(self), len(self) % training_length, -training_length
+                        )[:batch_size]:
+                            tmp_list += [np.array(self[end - training_length : end])]
-                            tmp_list += [np.array([padding] * (training_length - leftover) + self[:leftover])]
+                            tmp_list += [
+                                np.array(
+                                    [padding] * (training_length - leftover)
+                                    + self[:leftover]
+                                )
+                            ]
                         tmp_list.reverse()
                         return np.array(tmp_list)
 
             super(Buffer.AgentBuffer, self).__init__()
 
         def __str__(self):
-            return ", ".join(["'{0}' : {1}".format(k, str(self[k])) for k in self.keys()])
+            return ", ".join(
+                ["'{0}' : {1}".format(k, str(self[k])) for k in self.keys()]
+            )
 
         def reset_agent(self):
             """
             if key_list is None:
                 key_list = list(self.keys())
             if not self.check_length(key_list):
-                raise BufferException("Unable to shuffle if the fields are not of same length")
+                raise BufferException(
+                    "Unable to shuffle if the fields are not of same length"
+                )
             s = np.arange(len(self[key_list[0]]))
             np.random.shuffle(s)
             for key in key_list:
         super(Buffer, self).__init__()
 
     def __str__(self):
-        return "update buffer :\n\t{0}\nlocal_buffers :\n{1}".format(str(self.update_buffer),
-                                                                     '\n'.join(
-                                                                         ['\tagent {0} :{1}'.format(k, str(self[k])) for
-                                                                          k in self.keys()]))
+        return "update buffer :\n\t{0}\nlocal_buffers :\n{1}".format(
+            str(self.update_buffer),
+            "\n".join(
+                ["\tagent {0} :{1}".format(k, str(self[k])) for k in self.keys()]
+            ),
+        )
 
     def __getitem__(self, key):
         if key not in self.keys():
         for k in agent_ids:
             self[k].reset_agent()
 
-    def append_update_buffer(self, agent_id, key_list=None, batch_size=None, training_length=None):
+    def append_update_buffer(
+        self, agent_id, key_list=None, batch_size=None, training_length=None
+    ):
         """
         Appends the buffer of an agent to the update buffer.
         :param agent_id: The id of the agent which data will be appended
         if key_list is None:
             key_list = self[agent_id].keys()
         if not self[agent_id].check_length(key_list):
-            raise BufferException("The length of the fields {0} for agent {1} where not of same length"
-                                  .format(key_list, agent_id))
+            raise BufferException(
+                "The length of the fields {0} for agent {1} where not of same length".format(
+                    key_list, agent_id
+                )
+            )
-                self[agent_id][field_key].get_batch(batch_size=batch_size, training_length=training_length)
+                self[agent_id][field_key].get_batch(
+                    batch_size=batch_size, training_length=training_length
+                )
-    def append_all_agent_batch_to_update_buffer(self, key_list=None, batch_size=None, training_length=None):
+    def append_all_agent_batch_to_update_buffer(
+        self, key_list=None, batch_size=None, training_length=None
+    ):
         """
         Appends the buffer of all agents to the update buffer.
         :param key_list: The fields that must be added. If None: all fields will be appended.

ml-agents/mlagents/trainers/curriculum.py

 
 import logging
 
-logger = logging.getLogger('mlagents.trainers')
+logger = logging.getLogger("mlagents.trainers")
 
 
 class Curriculum(object):
         self._lesson_num = 0
         # The name of the brain should be the basename of the file without the
         # extension.
-        self._brain_name = os.path.basename(location).split('.')[0]
+        self._brain_name = os.path.basename(location).split(".")[0]
-            raise CurriculumError(
-                'The file {0} could not be found.'.format(location))
+            raise CurriculumError("The file {0} could not be found.".format(location))
-            raise CurriculumError('There was an error decoding {}'
-                                  .format(location))
+            raise CurriculumError("There was an error decoding {}".format(location))
-        for key in ['parameters', 'measure', 'thresholds',
-                    'min_lesson_length', 'signal_smoothing']:
+        for key in [
+            "parameters",
+            "measure",
+            "thresholds",
+            "min_lesson_length",
+            "signal_smoothing",
+        ]:
-                raise CurriculumError("{0} does not contain a "
-                                      "{1} field."
-                                      .format(location, key))
+                raise CurriculumError(
+                    "{0} does not contain a " "{1} field.".format(location, key)
+                )
-        self.measure = self.data['measure']
-        self.min_lesson_length = self.data['min_lesson_length']
-        self.max_lesson_num = len(self.data['thresholds'])
+        self.measure = self.data["measure"]
+        self.min_lesson_length = self.data["min_lesson_length"]
+        self.max_lesson_num = len(self.data["thresholds"])
-        parameters = self.data['parameters']
+        parameters = self.data["parameters"]
-                    'The parameter {0} in Curriculum {1} is not present in '
-                    'the Environment'.format(key, location))
+                    "The parameter {0} in Curriculum {1} is not present in "
+                    "the Environment".format(key, location)
+                )
-                    'The parameter {0} in Curriculum {1} must have {2} values '
-                    'but {3} were found'.format(key, location,
-                                                self.max_lesson_num + 1,
-                                                len(parameters[key])))
+                    "The parameter {0} in Curriculum {1} must have {2} values "
+                    "but {3} were found".format(
+                        key, location, self.max_lesson_num + 1, len(parameters[key])
+                    )
+                )
 
     @property
     def lesson_num(self):
         """
         if not self.data or not measure_val or math.isnan(measure_val):
             return False
-        if self.data['signal_smoothing']:
+        if self.data["signal_smoothing"]:
-            if measure_val > self.data['thresholds'][self.lesson_num]:
+            if measure_val > self.data["thresholds"][self.lesson_num]:
-                parameters = self.data['parameters']
+                parameters = self.data["parameters"]
-                logger.info('{0} lesson changed. Now in lesson {1}: {2}'
-                            .format(self._brain_name,
-                                    self.lesson_num,
-                                    ', '.join([str(x) + ' -> ' + str(config[x])
-                                        for x in config])))
+                logger.info(
+                    "{0} lesson changed. Now in lesson {1}: {2}".format(
+                        self._brain_name,
+                        self.lesson_num,
+                        ", ".join([str(x) + " -> " + str(config[x]) for x in config]),
+                    )
+                )
                 return True
         return False
 
             lesson = self.lesson_num
         lesson = max(0, min(lesson, self.max_lesson_num))
         config = {}
-        parameters = self.data['parameters']
+        parameters = self.data["parameters"]
         for key in parameters:
             config[key] = parameters[key][lesson]
         return config

ml-agents/mlagents/trainers/demo_loader.py

         current_brain_info = brain_infos[idx]
         next_brain_info = brain_infos[idx + 1]
         demo_buffer[0].last_brain_info = current_brain_info
-        demo_buffer[0]['done'].append(next_brain_info.local_done[0])
-        demo_buffer[0]['rewards'].append(next_brain_info.rewards[0])
+        demo_buffer[0]["done"].append(next_brain_info.local_done[0])
+        demo_buffer[0]["rewards"].append(next_brain_info.rewards[0])
-            demo_buffer[0]['visual_obs%d' % i] \
-                .append(current_brain_info.visual_observations[i][0])
+            demo_buffer[0]["visual_obs%d" % i].append(
+                current_brain_info.visual_observations[i][0]
+            )
-            demo_buffer[0]['vector_obs'] \
-                .append(current_brain_info.vector_observations[0])
-        demo_buffer[0]['actions'].append(next_brain_info.previous_vector_actions[0])
+            demo_buffer[0]["vector_obs"].append(
+                current_brain_info.vector_observations[0]
+            )
+        demo_buffer[0]["actions"].append(next_brain_info.previous_vector_actions[0])
-            demo_buffer.append_update_buffer(0, batch_size=None,
-                                             training_length=sequence_length)
+            demo_buffer.append_update_buffer(
+                0, batch_size=None, training_length=sequence_length
+            )
-    demo_buffer.append_update_buffer(0, batch_size=None,
-                                     training_length=sequence_length)
+    demo_buffer.append_update_buffer(
+        0, batch_size=None, training_length=sequence_length
+    )
     return demo_buffer
 
 
     INITIAL_POS = 33
 
     if not os.path.isfile(file_path):
-        raise FileNotFoundError("The demonstration file {} does not exist.".format(file_path))
+        raise FileNotFoundError(
+            "The demonstration file {} does not exist.".format(file_path)
+        )
-    if file_extension != '.demo':
-        raise ValueError("The file is not a '.demo' file. Please provide a file with the "
-                         "correct extension.")
+    if file_extension != ".demo":
+        raise ValueError(
+            "The file is not a '.demo' file. Please provide a file with the "
+            "correct extension."
+        )
 
     brain_params = None
     brain_infos = []
         next_pos, pos = _DecodeVarint32(data, pos)
         if obs_decoded == 0:
             meta_data_proto = DemonstrationMetaProto()
-            meta_data_proto.ParseFromString(data[pos:pos + next_pos])
+            meta_data_proto.ParseFromString(data[pos : pos + next_pos])
-            brain_param_proto.ParseFromString(data[pos:pos + next_pos])
+            brain_param_proto.ParseFromString(data[pos : pos + next_pos])
-            agent_info.ParseFromString(data[pos:pos + next_pos])
+            agent_info.ParseFromString(data[pos : pos + next_pos])
             brain_info = BrainInfo.from_agent_proto([agent_info], brain_params)
             brain_infos.append(brain_info)
             if len(brain_infos) == total_expected:

ml-agents/mlagents/trainers/exception.py

 Contains exceptions for the trainers package.
 """
 
+
+
+
+
+
 
 class MetaCurriculumError(TrainerError):
     """

ml-agents/mlagents/trainers/learn.py

     :param run_options: Command line arguments for training.
     """
     # Docker Parameters
-    docker_target_name = (run_options['--docker-target-name']
-                          if run_options['--docker-target-name'] != 'None' else None)
+    docker_target_name = (
+        run_options["--docker-target-name"]
+        if run_options["--docker-target-name"] != "None"
+        else None
+    )
-    env_path = (run_options['--env']
-                if run_options['--env'] != 'None' else None)
-    run_id = run_options['--run-id']
-    load_model = run_options['--load']
-    train_model = run_options['--train']
-    save_freq = int(run_options['--save-freq'])
-    keep_checkpoints = int(run_options['--keep-checkpoints'])
-    base_port = int(run_options['--base-port'])
-    num_envs = int(run_options['--num-envs'])
-    curriculum_folder = (run_options['--curriculum']
-                         if run_options['--curriculum'] != 'None' else None)
-    lesson = int(run_options['--lesson'])
-    fast_simulation = not bool(run_options['--slow'])
-    no_graphics = run_options['--no-graphics']
-    trainer_config_path = run_options['<trainer-config-path>']
+    env_path = run_options["--env"] if run_options["--env"] != "None" else None
+    run_id = run_options["--run-id"]
+    load_model = run_options["--load"]
+    train_model = run_options["--train"]
+    save_freq = int(run_options["--save-freq"])
+    keep_checkpoints = int(run_options["--keep-checkpoints"])
+    base_port = int(run_options["--base-port"])
+    num_envs = int(run_options["--num-envs"])
+    curriculum_folder = (
+        run_options["--curriculum"] if run_options["--curriculum"] != "None" else None
+    )
+    lesson = int(run_options["--lesson"])
+    fast_simulation = not bool(run_options["--slow"])
+    no_graphics = run_options["--no-graphics"]
+    trainer_config_path = run_options["<trainer-config-path>"]
-        model_path = './models/{run_id}-{sub_id}'.format(run_id=run_id, sub_id=sub_id)
-        summaries_dir = './summaries'
+        model_path = "./models/{run_id}-{sub_id}".format(run_id=run_id, sub_id=sub_id)
+        summaries_dir = "./summaries"
-        trainer_config_path = \
-            '/{docker_target_name}/{trainer_config_path}'.format(
+        trainer_config_path = "/{docker_target_name}/{trainer_config_path}".format(
+            docker_target_name=docker_target_name,
+            trainer_config_path=trainer_config_path,
+        )
+        if curriculum_folder is not None:
+            curriculum_folder = "/{docker_target_name}/{curriculum_folder}".format(
-                trainer_config_path=trainer_config_path)
-        if curriculum_folder is not None:
-            curriculum_folder = \
-                '/{docker_target_name}/{curriculum_folder}'.format(
-                    docker_target_name=docker_target_name,
-                    curriculum_folder=curriculum_folder)
-        model_path = '/{docker_target_name}/models/{run_id}-{sub_id}'.format(
-            docker_target_name=docker_target_name,
-            run_id=run_id,
-            sub_id=sub_id)
-        summaries_dir = '/{docker_target_name}/summaries'.format(
-            docker_target_name=docker_target_name)
+                curriculum_folder=curriculum_folder,
+            )
+        model_path = "/{docker_target_name}/models/{run_id}-{sub_id}".format(
+            docker_target_name=docker_target_name, run_id=run_id, sub_id=sub_id
+        )
+        summaries_dir = "/{docker_target_name}/summaries".format(
+            docker_target_name=docker_target_name
+        )
 
     trainer_config = load_config(trainer_config_path)
     env_factory = create_environment_factory(
         run_seed,
-        base_port + (sub_id * num_envs)
+        base_port + (sub_id * num_envs),
-    tc = TrainerController(model_path, summaries_dir, run_id + '-' + str(sub_id),
-                           save_freq, maybe_meta_curriculum,
-                           load_model, train_model,
-                           keep_checkpoints, lesson, env.external_brains,
-                           run_seed, fast_simulation)
+    tc = TrainerController(
+        model_path,
+        summaries_dir,
+        run_id + "-" + str(sub_id),
+        save_freq,
+        maybe_meta_curriculum,
+        load_model,
+        train_model,
+        keep_checkpoints,
+        lesson,
+        env.external_brains,
+        run_seed,
+        fast_simulation,
+    )
 
     # Signal that environment has been launched.
     process_queue.put(True)
 
 
-def try_create_meta_curriculum(curriculum_folder: Optional[str], env: BaseUnityEnvironment) -> Optional[MetaCurriculum]:
+def try_create_meta_curriculum(
+    curriculum_folder: Optional[str], env: BaseUnityEnvironment
+) -> Optional[MetaCurriculum]:
     if curriculum_folder is None:
         return None
     else:
                 if brain_name not in env.external_brains.keys():
-                    raise MetaCurriculumError('One of the curricula '
-                                              'defined in ' +
-                                              curriculum_folder + ' '
-                                              'does not have a corresponding '
-                                              'Brain. Check that the '
-                                              'curriculum file has the same '
-                                              'name as the Brain '
-                                              'whose curriculum it defines.')
+                    raise MetaCurriculumError(
+                        "One of the curricula "
+                        "defined in " + curriculum_folder + " "
+                        "does not have a corresponding "
+                        "Brain. Check that the "
+                        "curriculum file has the same "
+                        "name as the Brain "
+                        "whose curriculum it defines."
+                    )
-    for f in glob.glob('/{docker_target_name}/*'.format(
-            docker_target_name=docker_target_name)):
+    for f in glob.glob(
+        "/{docker_target_name}/*".format(docker_target_name=docker_target_name)
+    ):
-                    shutil.copytree(f,
-                                    '/ml-agents/{b}'.format(b=b))
+                    shutil.copytree(f, "/ml-agents/{b}".format(b=b))
-                    src_f = '/{docker_target_name}/{b}'.format(
-                        docker_target_name=docker_target_name, b=b)
-                    dst_f = '/ml-agents/{b}'.format(b=b)
+                    src_f = "/{docker_target_name}/{b}".format(
+                        docker_target_name=docker_target_name, b=b
+                    )
+                    dst_f = "/ml-agents/{b}".format(b=b)
-                logging.getLogger('mlagents.trainers').info(e)
-    env_path = '/ml-agents/{env_path}'.format(env_path=env_path)
+                logging.getLogger("mlagents.trainers").info(e)
+    env_path = "/ml-agents/{env_path}".format(env_path=env_path)
     return env_path
 
 
             trainer_config = yaml.load(data_file)
             return trainer_config
     except IOError:
-        raise UnityEnvironmentException('Parameter file could not be found '
-                                        'at {}.'
-                                        .format(trainer_config_path))
+        raise UnityEnvironmentException(
+            "Parameter file could not be found " "at {}.".format(trainer_config_path)
+        )
-        raise UnityEnvironmentException('There was an error decoding '
-                                        'Trainer Config from this path : {}'
-                                        .format(trainer_config_path))
+        raise UnityEnvironmentException(
+            "There was an error decoding "
+            "Trainer Config from this path : {}".format(trainer_config_path)
+        )
-        env_path: str,
-        docker_target_name: str,
-        no_graphics: bool,
-        seed: Optional[int],
-        start_port: int
+    env_path: str,
+    docker_target_name: str,
+    no_graphics: bool,
+    seed: Optional[int],
+    start_port: int,
-        env_path = (env_path.strip()
-                    .replace('.app', '')
-                    .replace('.exe', '')
-                    .replace('.x86_64', '')
-                    .replace('.x86', ''))
+        env_path = (
+            env_path.strip()
+            .replace(".app", "")
+            .replace(".exe", "")
+            .replace(".x86_64", "")
+            .replace(".x86", "")
+        )
-            """
+        """
             Comments for future maintenance:
                 Some OS/VM instances (e.g. COS GCP Image) mount filesystems
                 with COS flag which prevents execution of the Unity scene,
-            # Navigate in docker path and find env_path and copy it.
-            env_path = prepare_for_docker_run(docker_target_name,
-                                              env_path)
+        # Navigate in docker path and find env_path and copy it.
+        env_path = prepare_for_docker_run(docker_target_name, env_path)
     seed_count = 10000
     seed_pool = [np.random.randint(0, seed_count) for _ in range(seed_count)]
 
             seed=env_seed,
             docker_training=docker_training,
             no_graphics=no_graphics,
-            base_port=start_port
+            base_port=start_port,
+
-        print('''
+        print(
+            """
 
                         ▄▄▄▓▓▓▓
                    ╓▓▓▓▓▓▓█▓▓▓▓▓
                    `▀█▓▓▓▓▓▓▓▓▓▌
                         ¬`▀▀▀█▓
 
-        ''')
+        """
+        )
-        print('\n\n\tUnity Technologies\n')
+        print("\n\n\tUnity Technologies\n")
-    _USAGE = '''
+    _USAGE = """
     Usage:
       mlagents-learn <trainer-config-path> [options]
       mlagents-learn --help
       --docker-target-name=<dt>  Docker volume to store training-specific files [default: None].
       --no-graphics              Whether to run the environment in no-graphics mode [default: False].
       --debug                    Whether to run ML-Agents in debug mode with detailed logging [default: False].
-    '''
+    """
-    trainer_logger = logging.getLogger('mlagents.trainers')
-    env_logger = logging.getLogger('mlagents.envs')
+    trainer_logger = logging.getLogger("mlagents.trainers")
+    env_logger = logging.getLogger("mlagents.envs")
-    if options['--debug']:
-        trainer_logger.setLevel('DEBUG')
-        env_logger.setLevel('DEBUG')
-    num_runs = int(options['--num-runs'])
-    seed = int(options['--seed'])
+    if options["--debug"]:
+        trainer_logger.setLevel("DEBUG")
+        env_logger.setLevel("DEBUG")
+    num_runs = int(options["--num-runs"])
+    seed = int(options["--seed"])
-    if options['--env'] == 'None' and num_runs > 1:
-        raise TrainerError('It is not possible to launch more than one concurrent training session '
-                           'when training from the editor.')
+    if options["--env"] == "None" and num_runs > 1:
+        raise TrainerError(
+            "It is not possible to launch more than one concurrent training session "
+            "when training from the editor."
+        )
 
     jobs = []
     run_seed = seed
             job.join()
     except KeyboardInterrupt:
         pass
+
 
 # For python debugger to directly run this script
 if __name__ == "__main__":

ml-agents/mlagents/trainers/meta_curriculum.py

 
 import logging
 
-logger = logging.getLogger('mlagents.trainers')
+logger = logging.getLogger("mlagents.trainers")
 
 
 class MetaCurriculum(object):
 
         try:
             for curriculum_filename in os.listdir(curriculum_folder):
-                brain_name = curriculum_filename.split('.')[0]
-                curriculum_filepath = \
-                    os.path.join(curriculum_folder, curriculum_filename)
-                curriculum = Curriculum(curriculum_filepath,
-                                        default_reset_parameters)
+                brain_name = curriculum_filename.split(".")[0]
+                curriculum_filepath = os.path.join(
+                    curriculum_folder, curriculum_filename
+                )
+                curriculum = Curriculum(curriculum_filepath, default_reset_parameters)
-                if any([(parameter in curriculum.get_config().keys())
-                    for parameter in used_reset_parameters]):
-                    logger.warning('Two or more curriculums will '
-                                'attempt to change the same reset '
-                                'parameter. The result will be '
-                                'non-deterministic.')
+                if any(
+                    [
+                        (parameter in curriculum.get_config().keys())
+                        for parameter in used_reset_parameters
+                    ]
+                ):
+                    logger.warning(
+                        "Two or more curriculums will "
+                        "attempt to change the same reset "
+                        "parameter. The result will be "
+                        "non-deterministic."
+                    )
-            raise MetaCurriculumError(curriculum_folder + ' is not a '
-                                      'directory. Refer to the ML-Agents '
-                                      'curriculum learning docs.')
-
+            raise MetaCurriculumError(
+                curriculum_folder + " is not a "
+                "directory. Refer to the ML-Agents "
+                "curriculum learning docs."
+            )
 
     @property
     def brains_to_curriculums(self):
             Whether the curriculum of the specified brain should attempt to
             increment its lesson.
         """
-        return reward_buff_size >= (self.brains_to_curriculums[brain_name]
-                                        .min_lesson_length)
+        return reward_buff_size >= (
+            self.brains_to_curriculums[brain_name].min_lesson_length
+        )
 
     def increment_lessons(self, measure_vals, reward_buff_sizes=None):
         """Attempts to increments all the lessons of all the curriculums in this
             for brain_name, buff_size in reward_buff_sizes.items():
                 if self._lesson_ready_to_increment(brain_name, buff_size):
                     measure_val = measure_vals[brain_name]
-                    ret[brain_name] = (self.brains_to_curriculums[brain_name]
-                                           .increment_lesson(measure_val))
+                    ret[brain_name] = self.brains_to_curriculums[
+                        brain_name
+                    ].increment_lesson(measure_val)
-                ret[brain_name] = (self.brains_to_curriculums[brain_name]
-                                       .increment_lesson(measure_val))
+                ret[brain_name] = self.brains_to_curriculums[
+                    brain_name
+                ].increment_lesson(measure_val)
-
 
     def set_all_curriculums_to_lesson_num(self, lesson_num):
         """Sets all the curriculums in this meta curriculum to a specified
         """
         for _, curriculum in self.brains_to_curriculums.items():
             curriculum.lesson_num = lesson_num
-
 
     def get_config(self):
         """Get the combined configuration of all curriculums in this

ml-agents/mlagents/trainers/models.py

         self.vector_in = None
         self.global_step, self.increment_step = self.create_global_steps()
         self.visual_in = []
-        self.batch_size = tf.placeholder(shape=None, dtype=tf.int32, name='batch_size')
-        self.sequence_length = tf.placeholder(shape=None, dtype=tf.int32, name='sequence_length')
-        self.mask_input = tf.placeholder(shape=[None], dtype=tf.float32, name='masks')
+        self.batch_size = tf.placeholder(shape=None, dtype=tf.int32, name="batch_size")
+        self.sequence_length = tf.placeholder(
+            shape=None, dtype=tf.int32, name="sequence_length"
+        )
+        self.mask_input = tf.placeholder(shape=[None], dtype=tf.float32, name="masks")
         self.mask = tf.cast(self.mask_input, tf.int32)
         self.use_recurrent = use_recurrent
         if self.use_recurrent:
         self.normalize = normalize
         self.act_size = brain.vector_action_space_size
-        self.vec_obs_size = brain.vector_observation_space_size * \
-                            brain.num_stacked_vector_observations
+        self.vec_obs_size = (
+            brain.vector_observation_space_size * brain.num_stacked_vector_observations
+        )
-        tf.Variable(int(brain.vector_action_space_type == 'continuous'),
-                    name='is_continuous_control', trainable=False, dtype=tf.int32)
-        tf.Variable(self._version_number_, name='version_number', trainable=False, dtype=tf.int32)
+        tf.Variable(
+            int(brain.vector_action_space_type == "continuous"),
+            name="is_continuous_control",
+            trainable=False,
+            dtype=tf.int32,
+        )
+        tf.Variable(
+            self._version_number_,
+            name="version_number",
+            trainable=False,
+            dtype=tf.int32,
+        )
-        if brain.vector_action_space_type == 'continuous':
-            tf.Variable(self.act_size[0], name="action_output_shape", trainable=False, dtype=tf.int32)
+        if brain.vector_action_space_type == "continuous":
+            tf.Variable(
+                self.act_size[0],
+                name="action_output_shape",
+                trainable=False,
+                dtype=tf.int32,
+            )
-            tf.Variable(sum(self.act_size), name="action_output_shape", trainable=False, dtype=tf.int32)
+            tf.Variable(
+                sum(self.act_size),
+                name="action_output_shape",
+                trainable=False,
+                dtype=tf.int32,
+            )
-        global_step = tf.Variable(0, name="global_step", trainable=False, dtype=tf.int32)
+        global_step = tf.Variable(
+            0, name="global_step", trainable=False, dtype=tf.int32
+        )
         increment_step = tf.assign(global_step, tf.add(global_step, 1))
         return global_step, increment_step
 
         :param name: Desired name of input op.
         :return: input op.
         """
-        o_size_h = camera_parameters['height']
-        o_size_w = camera_parameters['width']
-        bw = camera_parameters['blackAndWhite']
+        o_size_h = camera_parameters["height"]
+        o_size_w = camera_parameters["width"]
+        bw = camera_parameters["blackAndWhite"]
 
         if bw:
             c_channels = 1
-        visual_in = tf.placeholder(shape=[None, o_size_h, o_size_w, c_channels], dtype=tf.float32,
-                                   name=name)
+        visual_in = tf.placeholder(
+            shape=[None, o_size_h, o_size_w, c_channels], dtype=tf.float32, name=name
+        )
-    def create_vector_input(self, name='vector_observation'):
+    def create_vector_input(self, name="vector_observation"):
         """
         Creates ops for vector observation input.
         :param name: Name of the placeholder op.
-        self.vector_in = tf.placeholder(shape=[None, self.vec_obs_size], dtype=tf.float32,
-                                        name=name)
+        self.vector_in = tf.placeholder(
+            shape=[None, self.vec_obs_size], dtype=tf.float32, name=name
+        )
-            self.running_mean = tf.get_variable("running_mean", [self.vec_obs_size],
-                                                trainable=False, dtype=tf.float32,
-                                                initializer=tf.zeros_initializer())
-            self.running_variance = tf.get_variable("running_variance", [self.vec_obs_size],
-                                                    trainable=False,
-                                                    dtype=tf.float32,
-                                                    initializer=tf.ones_initializer())
-            self.update_mean, self.update_variance = self.create_normalizer_update(self.vector_in)
+            self.running_mean = tf.get_variable(
+                "running_mean",
+                [self.vec_obs_size],
+                trainable=False,
+                dtype=tf.float32,
+                initializer=tf.zeros_initializer(),
+            )
+            self.running_variance = tf.get_variable(
+                "running_variance",
+                [self.vec_obs_size],
+                trainable=False,
+                dtype=tf.float32,
+                initializer=tf.ones_initializer(),
+            )
+            self.update_mean, self.update_variance = self.create_normalizer_update(
+                self.vector_in
+            )
-            self.normalized_state = tf.clip_by_value((self.vector_in - self.running_mean) / tf.sqrt(
-                self.running_variance / (tf.cast(self.global_step, tf.float32) + 1)), -5, 5,
-                                                     name="normalized_state")
+            self.normalized_state = tf.clip_by_value(
+                (self.vector_in - self.running_mean)
+                / tf.sqrt(
+                    self.running_variance / (tf.cast(self.global_step, tf.float32) + 1)
+                ),
+                -5,
+                5,
+                name="normalized_state",
+            )
             return self.normalized_state
         else:
             return self.vector_in
-        new_mean = self.running_mean + (mean_current_observation - self.running_mean) / \
-                   tf.cast(tf.add(self.global_step, 1), tf.float32)
-        new_variance = self.running_variance + (mean_current_observation - new_mean) * \
-                       (mean_current_observation - self.running_mean)
+        new_mean = self.running_mean + (
+            mean_current_observation - self.running_mean
+        ) / tf.cast(tf.add(self.global_step, 1), tf.float32)
+        new_variance = self.running_variance + (mean_current_observation - new_mean) * (
+            mean_current_observation - self.running_mean
+        )
-    def create_vector_observation_encoder(observation_input, h_size, activation, num_layers, scope,
-                                          reuse):
+    def create_vector_observation_encoder(
+        observation_input, h_size, activation, num_layers, scope, reuse
+    ):
         """
         Builds a set of hidden state encoders.
         :param reuse: Whether to re-use the weights within the same scope.
         with tf.variable_scope(scope):
             hidden = observation_input
             for i in range(num_layers):
-                hidden = tf.layers.dense(hidden, h_size, activation=activation, reuse=reuse,
-                                         name="hidden_{}".format(i),
-                                         kernel_initializer=c_layers.variance_scaling_initializer(
-                                             1.0))
+                hidden = tf.layers.dense(
+                    hidden,
+                    h_size,
+                    activation=activation,
+                    reuse=reuse,
+                    name="hidden_{}".format(i),
+                    kernel_initializer=c_layers.variance_scaling_initializer(1.0),
+                )
-    def create_visual_observation_encoder(self, image_input, h_size, activation, num_layers, scope,
-                                          reuse):
+    def create_visual_observation_encoder(
+        self, image_input, h_size, activation, num_layers, scope, reuse
+    ):
         """
         Builds a set of visual (CNN) encoders.
         :param reuse: Whether to re-use the weights within the same scope.
         :return: List of hidden layer tensors.
         """
         with tf.variable_scope(scope):
-            conv1 = tf.layers.conv2d(image_input, 16, kernel_size=[8, 8], strides=[4, 4],
-                                     activation=tf.nn.elu, reuse=reuse, name="conv_1")
-            conv2 = tf.layers.conv2d(conv1, 32, kernel_size=[4, 4], strides=[2, 2],
-                                     activation=tf.nn.elu, reuse=reuse, name="conv_2")
+            conv1 = tf.layers.conv2d(
+                image_input,
+                16,
+                kernel_size=[8, 8],
+                strides=[4, 4],
+                activation=tf.nn.elu,
+                reuse=reuse,
+                name="conv_1",
+            )
+            conv2 = tf.layers.conv2d(
+                conv1,
+                32,
+                kernel_size=[4, 4],
+                strides=[2, 2],
+                activation=tf.nn.elu,
+                reuse=reuse,
+                name="conv_2",
+            )
-        with tf.variable_scope(scope + '/' + 'flat_encoding'):
-            hidden_flat = self.create_vector_observation_encoder(hidden, h_size, activation,
-                                                                 num_layers, scope, reuse)
+        with tf.variable_scope(scope + "/" + "flat_encoding"):
+            hidden_flat = self.create_vector_observation_encoder(
+                hidden, h_size, activation, num_layers, scope, reuse
+            )
         return hidden_flat
 
     @staticmethod
         :return: The action output dimension [batch_size, num_branches] and the concatenated normalized logits
         """
         action_idx = [0] + list(np.cumsum(action_size))
-        branches_logits = [all_logits[:, action_idx[i]:action_idx[i + 1]] for i in range(len(action_size))]
-        branch_masks = [action_masks[:, action_idx[i]:action_idx[i + 1]] for i in range(len(action_size))]
-        raw_probs = [tf.multiply(tf.nn.softmax(branches_logits[k]) + 1.0e-10, branch_masks[k])
-                     for k in range(len(action_size))]
+        branches_logits = [
+            all_logits[:, action_idx[i] : action_idx[i + 1]]
+            for i in range(len(action_size))
+        ]
+        branch_masks = [
+            action_masks[:, action_idx[i] : action_idx[i + 1]]
+            for i in range(len(action_size))
+        ]
+        raw_probs = [
+            tf.multiply(tf.nn.softmax(branches_logits[k]) + 1.0e-10, branch_masks[k])
+            for k in range(len(action_size))
+        ]
-            for k in range(len(action_size))]
-        output = tf.concat([tf.multinomial(tf.log(normalized_probs[k]), 1) for k in range(len(action_size))], axis=1)
-        return output, tf.concat([tf.log(normalized_probs[k] + 1.0e-10) for k in range(len(action_size))], axis=1)
+            for k in range(len(action_size))
+        ]
+        output = tf.concat(
+            [
+                tf.multinomial(tf.log(normalized_probs[k]), 1)
+                for k in range(len(action_size))
+            ],
+            axis=1,
+        )
+        return (
+            output,
+            tf.concat(
+                [
+                    tf.log(normalized_probs[k] + 1.0e-10)
+                    for k in range(len(action_size))
+                ],
+                axis=1,
+            ),
+        )
 
     def create_observation_streams(self, num_streams, h_size, num_layers):
         """
 
         self.visual_in = []
         for i in range(brain.number_visual_observations):
-            visual_input = self.create_visual_input(brain.camera_resolutions[i],
-                                                    name="visual_observation_" + str(i))
+            visual_input = self.create_visual_input(
+                brain.camera_resolutions[i], name="visual_observation_" + str(i)
+            )
             self.visual_in.append(visual_input)
         vector_observation_input = self.create_vector_input()
 
             hidden_state, hidden_visual = None, None
             if self.vis_obs_size > 0:
                 for j in range(brain.number_visual_observations):
-                    encoded_visual = self.create_visual_observation_encoder(self.visual_in[j],
-                                                                            h_size,
-                                                                            activation_fn,
-                                                                            num_layers,
-                                                                            "main_graph_{}_encoder{}"
-                                                                            .format(i, j), False)
+                    encoded_visual = self.create_visual_observation_encoder(
+                        self.visual_in[j],
+                        h_size,
+                        activation_fn,
+                        num_layers,
+                        "main_graph_{}_encoder{}".format(i, j),
+                        False,
+                    )
-                hidden_state = self.create_vector_observation_encoder(vector_observation_input,
-                                                                      h_size, activation_fn,
-                                                                      num_layers,
-                                                                      "main_graph_{}".format(i),
-                                                                      False)
+                hidden_state = self.create_vector_observation_encoder(
+                    vector_observation_input,
+                    h_size,
+                    activation_fn,
+                    num_layers,
+                    "main_graph_{}".format(i),
+                    False,
+                )
             if hidden_state is not None and hidden_visual is not None:
                 final_hidden = tf.concat([hidden_visual, hidden_state], axis=1)
             elif hidden_state is None and hidden_visual is not None:
             else:
-                raise Exception("No valid network configuration possible. "
-                                "There are no states or observations in this brain")
+                raise Exception(
+                    "No valid network configuration possible. "
+                    "There are no states or observations in this brain"
+                )
-    def create_recurrent_encoder(input_state, memory_in, sequence_length, name='lstm'):
+    def create_recurrent_encoder(input_state, memory_in, sequence_length, name="lstm"):
         """
         Builds a recurrent encoder for either state or observations (LSTM).
         :param sequence_length: Length of sequence to unroll.
         _half_point = int(m_size / 2)
         with tf.variable_scope(name):
             rnn_cell = tf.contrib.rnn.BasicLSTMCell(_half_point)
-            lstm_vector_in = tf.contrib.rnn.LSTMStateTuple(memory_in[:, :_half_point],
-                                                           memory_in[:, _half_point:])
-            recurrent_output, lstm_state_out = tf.nn.dynamic_rnn(rnn_cell, lstm_input_state,
-                                                                 initial_state=lstm_vector_in)
+            lstm_vector_in = tf.contrib.rnn.LSTMStateTuple(
+                memory_in[:, :_half_point], memory_in[:, _half_point:]
+            )
+            recurrent_output, lstm_state_out = tf.nn.dynamic_rnn(
+                rnn_cell, lstm_input_state, initial_state=lstm_vector_in
+            )
 
         recurrent_output = tf.reshape(recurrent_output, shape=[-1, _half_point])
         return recurrent_output, tf.concat([lstm_state_out.c, lstm_state_out.h], axis=1)
         hidden_streams = self.create_observation_streams(2, h_size, num_layers)
 
         if self.use_recurrent:
-            self.memory_in = tf.placeholder(shape=[None, self.m_size], dtype=tf.float32,
-                                            name='recurrent_in')
+            self.memory_in = tf.placeholder(
+                shape=[None, self.m_size], dtype=tf.float32, name="recurrent_in"
+            )
-                hidden_streams[0], self.memory_in[:, :_half_point], self.sequence_length,
-                name='lstm_policy')
+                hidden_streams[0],
+                self.memory_in[:, :_half_point],
+                self.sequence_length,
+                name="lstm_policy",
+            )
-                hidden_streams[1], self.memory_in[:, _half_point:], self.sequence_length,
-                name='lstm_value')
-            self.memory_out = tf.concat([memory_policy_out, memory_value_out], axis=1,
-                                        name='recurrent_out')
+                hidden_streams[1],
+                self.memory_in[:, _half_point:],
+                self.sequence_length,
+                name="lstm_value",
+            )
+            self.memory_out = tf.concat(
+                [memory_policy_out, memory_value_out], axis=1, name="recurrent_out"
+            )
-        mu = tf.layers.dense(hidden_policy, self.act_size[0], activation=None,
-                             kernel_initializer=c_layers.variance_scaling_initializer(factor=0.01))
+        mu = tf.layers.dense(
+            hidden_policy,
+            self.act_size[0],
+            activation=None,
+            kernel_initializer=c_layers.variance_scaling_initializer(factor=0.01),
+        )
-        log_sigma_sq = tf.get_variable("log_sigma_squared", [self.act_size[0]], dtype=tf.float32,
-                                       initializer=tf.zeros_initializer())
+        log_sigma_sq = tf.get_variable(
+            "log_sigma_squared",
+            [self.act_size[0]],
+            dtype=tf.float32,
+            initializer=tf.zeros_initializer(),
+        )
-        self.epsilon = tf.placeholder(shape=[None, self.act_size[0]], dtype=tf.float32, name='epsilon')
+        self.epsilon = tf.placeholder(
+            shape=[None, self.act_size[0]], dtype=tf.float32, name="epsilon"
+        )
-        self.output = tf.identity(output_post, name='action')
+        self.output = tf.identity(output_post, name="action")
-        all_probs = - 0.5 * tf.square(tf.stop_gradient(self.output_pre) - mu) / sigma_sq \
-                    - 0.5 * tf.log(2.0 * np.pi) - 0.5 * log_sigma_sq
+        all_probs = (
+            -0.5 * tf.square(tf.stop_gradient(self.output_pre) - mu) / sigma_sq
+            - 0.5 * tf.log(2.0 * np.pi)
+            - 0.5 * log_sigma_sq
+        )
-        self.all_log_probs = tf.identity(all_probs, name='action_probs')
+        self.all_log_probs = tf.identity(all_probs, name="action_probs")
 
         self.entropy = 0.5 * tf.reduce_mean(tf.log(2 * np.pi * np.e) + log_sigma_sq)
 
-        self.all_old_log_probs = tf.placeholder(shape=[None, self.act_size[0]], dtype=tf.float32,
-                                                name='old_probabilities')
+        self.all_old_log_probs = tf.placeholder(
+            shape=[None, self.act_size[0]], dtype=tf.float32, name="old_probabilities"
+        )
-        self.log_probs = tf.reduce_sum((tf.identity(self.all_log_probs)), axis=1, keepdims=True)
-        self.old_log_probs = tf.reduce_sum((tf.identity(self.all_old_log_probs)), axis=1,
-                                           keepdims=True)
+        self.log_probs = tf.reduce_sum(
+            (tf.identity(self.all_log_probs)), axis=1, keepdims=True
+        )
+        self.old_log_probs = tf.reduce_sum(
+            (tf.identity(self.all_old_log_probs)), axis=1, keepdims=True
+        )
 
     def create_dc_actor_critic(self, h_size, num_layers):
         """
         hidden = hidden_streams[0]
 
         if self.use_recurrent:
-            self.prev_action = tf.placeholder(shape=[None, len(self.act_size)], dtype=tf.int32,
-                                              name='prev_action')
-            prev_action_oh = tf.concat([
-                tf.one_hot(self.prev_action[:, i], self.act_size[i]) for i in
-                range(len(self.act_size))], axis=1)
+            self.prev_action = tf.placeholder(
+                shape=[None, len(self.act_size)], dtype=tf.int32, name="prev_action"
+            )
+            prev_action_oh = tf.concat(
+                [
+                    tf.one_hot(self.prev_action[:, i], self.act_size[i])
+                    for i in range(len(self.act_size))
+                ],
+                axis=1,
+            )
-            self.memory_in = tf.placeholder(shape=[None, self.m_size], dtype=tf.float32,
-                                            name='recurrent_in')
-            hidden, memory_out = self.create_recurrent_encoder(hidden, self.memory_in,
-                                                               self.sequence_length)
-            self.memory_out = tf.identity(memory_out, name='recurrent_out')
+            self.memory_in = tf.placeholder(
+                shape=[None, self.m_size], dtype=tf.float32, name="recurrent_in"
+            )
+            hidden, memory_out = self.create_recurrent_encoder(
+                hidden, self.memory_in, self.sequence_length
+            )
+            self.memory_out = tf.identity(memory_out, name="recurrent_out")
-            policy_branches.append(tf.layers.dense(hidden, size, activation=None, use_bias=False,
-                                      kernel_initializer=c_layers.variance_scaling_initializer(factor=0.01)))
+            policy_branches.append(
+                tf.layers.dense(
+                    hidden,
+                    size,
+                    activation=None,
+                    use_bias=False,
+                    kernel_initializer=c_layers.variance_scaling_initializer(
+                        factor=0.01
+                    ),
+                )
+            )
-        self.all_log_probs = tf.concat([branch for branch in policy_branches], axis=1, name="action_probs")
+        self.all_log_probs = tf.concat(
+            [branch for branch in policy_branches], axis=1, name="action_probs"
+        )
-        self.action_masks = tf.placeholder(shape=[None, sum(self.act_size)], dtype=tf.float32, name="action_masks")
+        self.action_masks = tf.placeholder(
+            shape=[None, sum(self.act_size)], dtype=tf.float32, name="action_masks"
+        )
-            self.all_log_probs, self.action_masks, self.act_size)
+            self.all_log_probs, self.action_masks, self.act_size
+        )
-        self.normalized_logits = tf.identity(normalized_logits, name='action')
+        self.normalized_logits = tf.identity(normalized_logits, name="action")
-            shape=[None, len(policy_branches)], dtype=tf.int32, name="action_holder")
-        self.action_oh = tf.concat([
-            tf.one_hot(self.action_holder[:, i], self.act_size[i]) for i in range(len(self.act_size))], axis=1)
+            shape=[None, len(policy_branches)], dtype=tf.int32, name="action_holder"
+        )
+        self.action_oh = tf.concat(
+            [
+                tf.one_hot(self.action_holder[:, i], self.act_size[i])
+                for i in range(len(self.act_size))
+            ],
+            axis=1,
+        )
-            shape=[None, sum(self.act_size)], dtype=tf.float32, name='old_probabilities')
+            shape=[None, sum(self.act_size)], dtype=tf.float32, name="old_probabilities"
+        )
-            self.all_old_log_probs, self.action_masks, self.act_size)
+            self.all_old_log_probs, self.action_masks, self.act_size
+        )
-        self.entropy = tf.reduce_sum((tf.stack([
-            tf.nn.softmax_cross_entropy_with_logits_v2(
-                labels=tf.nn.softmax(self.all_log_probs[:, action_idx[i]:action_idx[i + 1]]),
-                logits=self.all_log_probs[:, action_idx[i]:action_idx[i + 1]])
-            for i in range(len(self.act_size))], axis=1)), axis=1)
+        self.entropy = tf.reduce_sum(
+            (
+                tf.stack(
+                    [
+                        tf.nn.softmax_cross_entropy_with_logits_v2(
+                            labels=tf.nn.softmax(
+                                self.all_log_probs[:, action_idx[i] : action_idx[i + 1]]
+                            ),
+                            logits=self.all_log_probs[
+                                :, action_idx[i] : action_idx[i + 1]
+                            ],
+                        )
+                        for i in range(len(self.act_size))
+                    ],
+                    axis=1,
+                )
+            ),
+            axis=1,
+        )
-        self.log_probs = tf.reduce_sum((tf.stack([
-            -tf.nn.softmax_cross_entropy_with_logits_v2(
-                labels=self.action_oh[:, action_idx[i]:action_idx[i + 1]],
-                logits=normalized_logits[:, action_idx[i]:action_idx[i + 1]]
-            )
-            for i in range(len(self.act_size))], axis=1)), axis=1, keepdims=True)
-        self.old_log_probs = tf.reduce_sum((tf.stack([
-            -tf.nn.softmax_cross_entropy_with_logits_v2(
-                labels=self.action_oh[:, action_idx[i]:action_idx[i + 1]],
-                logits=old_normalized_logits[:, action_idx[i]:action_idx[i + 1]]
-            )
-            for i in range(len(self.act_size))], axis=1)), axis=1, keepdims=True)
+        self.log_probs = tf.reduce_sum(
+            (
+                tf.stack(
+                    [
+                        -tf.nn.softmax_cross_entropy_with_logits_v2(
+                            labels=self.action_oh[:, action_idx[i] : action_idx[i + 1]],
+                            logits=normalized_logits[
+                                :, action_idx[i] : action_idx[i + 1]
+                            ],
+                        )
+                        for i in range(len(self.act_size))
+                    ],
+                    axis=1,
+                )
+            ),
+            axis=1,
+            keepdims=True,
+        )
+        self.old_log_probs = tf.reduce_sum(
+            (
+                tf.stack(
+                    [
+                        -tf.nn.softmax_cross_entropy_with_logits_v2(
+                            labels=self.action_oh[:, action_idx[i] : action_idx[i + 1]],
+                            logits=old_normalized_logits[
+                                :, action_idx[i] : action_idx[i + 1]
+                            ],
+                        )
+                        for i in range(len(self.act_size))
+                    ],
+                    axis=1,
+                )
+            ),
+            axis=1,
+            keepdims=True,
+        )

ml-agents/mlagents/trainers/policy.py

     """
     Related to errors with the Trainer.
     """
+
     pass
 
 
     functions to interact with it to perform evaluate and updating.
     """
-    possible_output_nodes = ['action', 'value_estimate',
-                             'action_probs', 'recurrent_out', 'memory_size',
-                             'version_number', 'is_continuous_control',
-                             'action_output_shape']
+
+    possible_output_nodes = [
+        "action",
+        "value_estimate",
+        "action_probs",
+        "recurrent_out",
+        "memory_size",
+        "version_number",
+        "is_continuous_control",
+        "action_output_shape",
+    ]
 
     def __init__(self, seed, brain, trainer_parameters):
         """
         self.seed = seed
         self.brain = brain
         self.use_recurrent = trainer_parameters["use_recurrent"]
-        self.use_continuous_act = (brain.vector_action_space_type == "continuous")
+        self.use_continuous_act = brain.vector_action_space_type == "continuous"
         self.model_path = trainer_parameters["model_path"]
         self.keep_checkpoints = trainer_parameters.get("keep_checkpoints", 5)
         self.graph = tf.Graph()
             self.m_size = trainer_parameters["memory_size"]
             self.sequence_length = trainer_parameters["sequence_length"]
             if self.m_size == 0:
-                raise UnityPolicyException("The memory size for brain {0} is 0 even "
-                                           "though the trainer uses recurrent."
-                                           .format(brain.brain_name))
+                raise UnityPolicyException(
+                    "The memory size for brain {0} is 0 even "
+                    "though the trainer uses recurrent.".format(brain.brain_name)
+                )
-                raise UnityPolicyException("The memory size for brain {0} is {1} "
-                                           "but it must be divisible by 4."
-                                           .format(brain.brain_name, self.m_size))
+                raise UnityPolicyException(
+                    "The memory size for brain {0} is {1} "
+                    "but it must be divisible by 4.".format(
+                        brain.brain_name, self.m_size
+                    )
+                )
 
     def _initialize_graph(self):
         with self.graph.as_default():
     def _load_graph(self):
         with self.graph.as_default():
             self.saver = tf.train.Saver(max_to_keep=self.keep_checkpoints)
-            logger.info('Loading Model for brain {}'.format(self.brain.brain_name))
+            logger.info("Loading Model for brain {}".format(self.brain.brain_name))
-                logger.info('The model {0} could not be found. Make '
-                            'sure you specified the right '
-                            '--run-id'
-                            .format(self.model_path))
+                logger.info(
+                    "The model {0} could not be found. Make "
+                    "sure you specified the right "
+                    "--run-id".format(self.model_path)
+                )
             self.saver.restore(self.sess, ckpt.model_checkpoint_path)
 
     def evaluate(self, brain_info: BrainInfo):
 
         run_out = self.evaluate(brain_info)
         return ActionInfo(
-            action=run_out.get('action'),
-            memory=run_out.get('memory_out'),
+            action=run_out.get("action"),
+            memory=run_out.get("memory_out"),
-            value=run_out.get('value'),
-            outputs=run_out
+            value=run_out.get("value"),
+            outputs=run_out,
         )
 
     def update(self, mini_batch, num_sequences):
         :return:
         """
         with self.graph.as_default():
-            last_checkpoint = self.model_path + '/model-' + str(steps) + '.cptk'
+            last_checkpoint = self.model_path + "/model-" + str(steps) + ".cptk"
-            tf.train.write_graph(self.graph, self.model_path,
-                                 'raw_graph_def.pb', as_text=False)
+            tf.train.write_graph(
+                self.graph, self.model_path, "raw_graph_def.pb", as_text=False
+            )
 
     def export_model(self):
         """
         with self.graph.as_default():
-            target_nodes = ','.join(self._process_graph())
+            target_nodes = ",".join(self._process_graph())
-                input_graph=self.model_path + '/raw_graph_def.pb',
+                input_graph=self.model_path + "/raw_graph_def.pb",
-                output_graph=(self.model_path + '/frozen_graph_def.pb'),
-                clear_devices=True, initializer_nodes='', input_saver='',
-                restore_op_name='save/restore_all',
-                filename_tensor_name='save/Const:0')
+                output_graph=(self.model_path + "/frozen_graph_def.pb"),
+                clear_devices=True,
+                initializer_nodes="",
+                input_saver="",
+                restore_op_name="save/restore_all",
+                filename_tensor_name="save/Const:0",
+            )
-        tf2bc.convert(self.model_path + '/frozen_graph_def.pb', self.model_path + '.nn')
-        logger.info('Exported ' + self.model_path + '.nn file')
+        tf2bc.convert(self.model_path + "/frozen_graph_def.pb", self.model_path + ".nn")
+        logger.info("Exported " + self.model_path + ".nn file")
 
     def _process_graph(self):
         """
         all_nodes = [x.name for x in self.graph.as_graph_def().node]
         nodes = [x for x in all_nodes if x in self.possible_output_nodes]
-        logger.info('List of nodes to export for brain :' + self.brain.brain_name)
+        logger.info("List of nodes to export for brain :" + self.brain.brain_name)
-            logger.info('\t' + n)
+            logger.info("\t" + n)
         return nodes
 
     @property

ml-agents/mlagents/trainers/ppo/models.py

 
 
 class PPOModel(LearningModel):
-    def __init__(self, brain, lr=1e-4, h_size=128, epsilon=0.2, beta=1e-3, max_step=5e6,
-                 normalize=False, use_recurrent=False, num_layers=2, m_size=None, use_curiosity=False,
-                 curiosity_strength=0.01, curiosity_enc_size=128, seed=0):
+    def __init__(
+        self,
+        brain,
+        lr=1e-4,
+        h_size=128,
+        epsilon=0.2,
+        beta=1e-3,
+        max_step=5e6,
+        normalize=False,
+        use_recurrent=False,
+        num_layers=2,
+        m_size=None,
+        use_curiosity=False,
+        curiosity_strength=0.01,
+        curiosity_enc_size=128,
+        seed=0,
+    ):
         """
         Takes a Unity environment and model-specific hyper-parameters and returns the
         appropriate PPO agent model for the environment.
         self.use_curiosity = use_curiosity
         if num_layers < 1:
             num_layers = 1
-        self.last_reward, self.new_reward, self.update_reward = self.create_reward_encoder()
+        self.last_reward, self.new_reward, self.update_reward = (
+            self.create_reward_encoder()
+        )
         if brain.vector_action_space_type == "continuous":
             self.create_cc_actor_critic(h_size, num_layers)
             self.entropy = tf.ones_like(tf.reshape(self.value, [-1])) * self.entropy
             encoded_state, encoded_next_state = self.create_curiosity_encoders()
             self.create_inverse_model(encoded_state, encoded_next_state)
             self.create_forward_model(encoded_state, encoded_next_state)
-        self.create_ppo_optimizer(self.log_probs, self.old_log_probs, self.value,
-                                  self.entropy, beta, epsilon, lr, max_step)
+        self.create_ppo_optimizer(
+            self.log_probs,
+            self.old_log_probs,
+            self.value,
+            self.entropy,
+            beta,
+            epsilon,
+            lr,
+            max_step,
+        )
-        last_reward = tf.Variable(0, name="last_reward", trainable=False, dtype=tf.float32)
-        new_reward = tf.placeholder(shape=[], dtype=tf.float32, name='new_reward')
+        last_reward = tf.Variable(
+            0, name="last_reward", trainable=False, dtype=tf.float32
+        )
+        new_reward = tf.placeholder(shape=[], dtype=tf.float32, name="new_reward")
         update_reward = tf.assign(last_reward, new_reward)
         return last_reward, new_reward, update_reward
 
             next_visual_encoders = []
             for i in range(self.vis_obs_size):
                 # Create input ops for next (t+1) visual observations.
-                next_visual_input = self.create_visual_input(self.brain.camera_resolutions[i],
-                                                             name="next_visual_observation_" + str(i))
+                next_visual_input = self.create_visual_input(
+                    self.brain.camera_resolutions[i],
+                    name="next_visual_observation_" + str(i),
+                )
-                encoded_visual = self.create_visual_observation_encoder(self.visual_in[i], self.curiosity_enc_size,
-                                                                        self.swish, 1, "stream_{}_visual_obs_encoder"
-                                                                        .format(i), False)
+                encoded_visual = self.create_visual_observation_encoder(
+                    self.visual_in[i],
+                    self.curiosity_enc_size,
+                    self.swish,
+                    1,
+                    "stream_{}_visual_obs_encoder".format(i),
+                    False,
+                )
-                encoded_next_visual = self.create_visual_observation_encoder(self.next_visual_in[i],
-                                                                             self.curiosity_enc_size,
-                                                                             self.swish, 1,
-                                                                             "stream_{}_visual_obs_encoder".format(i),
-                                                                             True)
+                encoded_next_visual = self.create_visual_observation_encoder(
+                    self.next_visual_in[i],
+                    self.curiosity_enc_size,
+                    self.swish,
+                    1,
+                    "stream_{}_visual_obs_encoder".format(i),
+                    True,
+                )
                 visual_encoders.append(encoded_visual)
                 next_visual_encoders.append(encoded_next_visual)
 
         if self.vec_obs_size > 0:
             # Create the encoder ops for current and next vector input. Not that these encoders are siamese.
             # Create input op for next (t+1) vector observation.
-            self.next_vector_in = tf.placeholder(shape=[None, self.vec_obs_size], dtype=tf.float32,
-                                                 name='next_vector_observation')
+            self.next_vector_in = tf.placeholder(
+                shape=[None, self.vec_obs_size],
+                dtype=tf.float32,
+                name="next_vector_observation",
+            )
-            encoded_vector_obs = self.create_vector_observation_encoder(self.vector_in,
-                                                                        self.curiosity_enc_size,
-                                                                        self.swish, 2, "vector_obs_encoder",
-                                                                        False)
-            encoded_next_vector_obs = self.create_vector_observation_encoder(self.next_vector_in,
-                                                                             self.curiosity_enc_size,
-                                                                             self.swish, 2,
-                                                                                 "vector_obs_encoder",
-                                                                             True)
+            encoded_vector_obs = self.create_vector_observation_encoder(
+                self.vector_in,
+                self.curiosity_enc_size,
+                self.swish,
+                2,
+                "vector_obs_encoder",
+                False,
+            )
+            encoded_next_vector_obs = self.create_vector_observation_encoder(
+                self.next_vector_in,
+                self.curiosity_enc_size,
+                self.swish,
+                2,
+                "vector_obs_encoder",
+                True,
+            )
             encoded_state_list.append(encoded_vector_obs)
             encoded_next_state_list.append(encoded_next_vector_obs)
 
         hidden = tf.layers.dense(combined_input, 256, activation=self.swish)
         if self.brain.vector_action_space_type == "continuous":
             pred_action = tf.layers.dense(hidden, self.act_size[0], activation=None)
-            squared_difference = tf.reduce_sum(tf.squared_difference(pred_action, self.selected_actions), axis=1)
-            self.inverse_loss = tf.reduce_mean(tf.dynamic_partition(squared_difference, self.mask, 2)[1])
+            squared_difference = tf.reduce_sum(
+                tf.squared_difference(pred_action, self.selected_actions), axis=1
+            )
+            self.inverse_loss = tf.reduce_mean(
+                tf.dynamic_partition(squared_difference, self.mask, 2)[1]
+            )
-                [tf.layers.dense(hidden, self.act_size[i], activation=tf.nn.softmax)
-                 for i in range(len(self.act_size))], axis=1)
-            cross_entropy = tf.reduce_sum(-tf.log(pred_action + 1e-10) * self.selected_actions, axis=1)
-            self.inverse_loss = tf.reduce_mean(tf.dynamic_partition(cross_entropy, self.mask, 2)[1])
+                [
+                    tf.layers.dense(hidden, self.act_size[i], activation=tf.nn.softmax)
+                    for i in range(len(self.act_size))
+                ],
+                axis=1,
+            )
+            cross_entropy = tf.reduce_sum(
+                -tf.log(pred_action + 1e-10) * self.selected_actions, axis=1
+            )
+            self.inverse_loss = tf.reduce_mean(
+                tf.dynamic_partition(cross_entropy, self.mask, 2)[1]
+            )
 
     def create_forward_model(self, encoded_state, encoded_next_state):
         """
         combined_input = tf.concat([encoded_state, self.selected_actions], axis=1)
         hidden = tf.layers.dense(combined_input, 256, activation=self.swish)
         # We compare against the concatenation of all observation streams, hence `self.vis_obs_size + int(self.vec_obs_size > 0)`.
-        pred_next_state = tf.layers.dense(hidden, self.curiosity_enc_size * (self.vis_obs_size + int(self.vec_obs_size > 0)),
-                                          activation=None)
+        pred_next_state = tf.layers.dense(
+            hidden,
+            self.curiosity_enc_size * (self.vis_obs_size + int(self.vec_obs_size > 0)),
+            activation=None,
+        )
-        squared_difference = 0.5 * tf.reduce_sum(tf.squared_difference(pred_next_state, encoded_next_state), axis=1)
-        self.intrinsic_reward = tf.clip_by_value(self.curiosity_strength * squared_difference, 0, 1)
-        self.forward_loss = tf.reduce_mean(tf.dynamic_partition(squared_difference, self.mask, 2)[1])
+        squared_difference = 0.5 * tf.reduce_sum(
+            tf.squared_difference(pred_next_state, encoded_next_state), axis=1
+        )
+        self.intrinsic_reward = tf.clip_by_value(
+            self.curiosity_strength * squared_difference, 0, 1
+        )
+        self.forward_loss = tf.reduce_mean(
+            tf.dynamic_partition(squared_difference, self.mask, 2)[1]
+        )
-    def create_ppo_optimizer(self, probs, old_probs, value, entropy, beta, epsilon, lr, max_step):
+    def create_ppo_optimizer(
+        self, probs, old_probs, value, entropy, beta, epsilon, lr, max_step
+    ):
         """
         Creates training-specific Tensorflow ops for PPO models.
         :param probs: Current policy probabilities
         :param lr: Learning rate
         :param max_step: Total number of training steps.
         """
-        self.returns_holder = tf.placeholder(shape=[None], dtype=tf.float32, name='discounted_rewards')
-        self.advantage = tf.placeholder(shape=[None, 1], dtype=tf.float32, name='advantages')
-        self.learning_rate = tf.train.polynomial_decay(lr, self.global_step, max_step, 1e-10, power=1.0)
+        self.returns_holder = tf.placeholder(
+            shape=[None], dtype=tf.float32, name="discounted_rewards"
+        )
+        self.advantage = tf.placeholder(
+            shape=[None, 1], dtype=tf.float32, name="advantages"
+        )
+        self.learning_rate = tf.train.polynomial_decay(
+            lr, self.global_step, max_step, 1e-10, power=1.0
+        )
-        self.old_value = tf.placeholder(shape=[None], dtype=tf.float32, name='old_value_estimates')
+        self.old_value = tf.placeholder(
+            shape=[None], dtype=tf.float32, name="old_value_estimates"
+        )
-        decay_epsilon = tf.train.polynomial_decay(epsilon, self.global_step, max_step, 0.1, power=1.0)
-        decay_beta = tf.train.polynomial_decay(beta, self.global_step, max_step, 1e-5, power=1.0)
+        decay_epsilon = tf.train.polynomial_decay(
+            epsilon, self.global_step, max_step, 0.1, power=1.0
+        )
+        decay_beta = tf.train.polynomial_decay(
+            beta, self.global_step, max_step, 1e-5, power=1.0
+        )
-        clipped_value_estimate = self.old_value + tf.clip_by_value(tf.reduce_sum(value, axis=1) - self.old_value,
-                                                                   - decay_epsilon, decay_epsilon)
+        clipped_value_estimate = self.old_value + tf.clip_by_value(
+            tf.reduce_sum(value, axis=1) - self.old_value, -decay_epsilon, decay_epsilon
+        )
-        v_opt_a = tf.squared_difference(self.returns_holder, tf.reduce_sum(value, axis=1))
+        v_opt_a = tf.squared_difference(
+            self.returns_holder, tf.reduce_sum(value, axis=1)
+        )
-        self.value_loss = tf.reduce_mean(tf.dynamic_partition(tf.maximum(v_opt_a, v_opt_b), self.mask, 2)[1])
+        self.value_loss = tf.reduce_mean(
+            tf.dynamic_partition(tf.maximum(v_opt_a, v_opt_b), self.mask, 2)[1]
+        )
 
         # Here we calculate PPO policy loss. In continuous control this is done independently for each action gaussian
         # and then averaged together. This provides significantly better performance than treating the probability
-        p_opt_b = tf.clip_by_value(r_theta, 1.0 - decay_epsilon, 1.0 + decay_epsilon) * self.advantage
-        self.policy_loss = -tf.reduce_mean(tf.dynamic_partition(tf.minimum(p_opt_a, p_opt_b), self.mask, 2)[1])
+        p_opt_b = (
+            tf.clip_by_value(r_theta, 1.0 - decay_epsilon, 1.0 + decay_epsilon)
+            * self.advantage
+        )
+        self.policy_loss = -tf.reduce_mean(
+            tf.dynamic_partition(tf.minimum(p_opt_a, p_opt_b), self.mask, 2)[1]
+        )
-        self.loss = self.policy_loss + 0.5 * self.value_loss - decay_beta * tf.reduce_mean(
-            tf.dynamic_partition(entropy, self.mask, 2)[1])
+        self.loss = (
+            self.policy_loss
+            + 0.5 * self.value_loss
+            - decay_beta
+            * tf.reduce_mean(tf.dynamic_partition(entropy, self.mask, 2)[1])
+        )
 
         if self.use_curiosity:
             self.loss += 10 * (0.2 * self.forward_loss + 0.8 * self.inverse_loss)

ml-agents/mlagents/trainers/ppo/policy.py

         """
         super().__init__(seed, brain, trainer_params)
         self.has_updated = False
-        self.use_curiosity = bool(trainer_params['use_curiosity'])
+        self.use_curiosity = bool(trainer_params["use_curiosity"])
-            self.model = PPOModel(brain,
-                                  lr=float(trainer_params['learning_rate']),
-                                  h_size=int(trainer_params['hidden_units']),
-                                  epsilon=float(trainer_params['epsilon']),
-                                  beta=float(trainer_params['beta']),
-                                  max_step=float(trainer_params['max_steps']),
-                                  normalize=trainer_params['normalize'],
-                                  use_recurrent=trainer_params['use_recurrent'],
-                                  num_layers=int(trainer_params['num_layers']),
-                                  m_size=self.m_size,
-                                  use_curiosity=bool(trainer_params['use_curiosity']),
-                                  curiosity_strength=float(trainer_params['curiosity_strength']),
-                                  curiosity_enc_size=float(trainer_params['curiosity_enc_size']),
-                                  seed=seed)
+            self.model = PPOModel(
+                brain,
+                lr=float(trainer_params["learning_rate"]),
+                h_size=int(trainer_params["hidden_units"]),
+                epsilon=float(trainer_params["epsilon"]),
+                beta=float(trainer_params["beta"]),
+                max_step=float(trainer_params["max_steps"]),
+                normalize=trainer_params["normalize"],
+                use_recurrent=trainer_params["use_recurrent"],
+                num_layers=int(trainer_params["num_layers"]),
+                m_size=self.m_size,
+                use_curiosity=bool(trainer_params["use_curiosity"]),
+                curiosity_strength=float(trainer_params["curiosity_strength"]),
+                curiosity_enc_size=float(trainer_params["curiosity_enc_size"]),
+                seed=seed,
+            )
 
         if load:
             self._load_graph()
-        self.inference_dict = {'action': self.model.output, 'log_probs': self.model.all_log_probs,
-                               'value': self.model.value, 'entropy': self.model.entropy,
-                               'learning_rate': self.model.learning_rate}
+        self.inference_dict = {
+            "action": self.model.output,
+            "log_probs": self.model.all_log_probs,
+            "value": self.model.value,
+            "entropy": self.model.entropy,
+            "learning_rate": self.model.learning_rate,
+        }
-            self.inference_dict['pre_action'] = self.model.output_pre
+            self.inference_dict["pre_action"] = self.model.output_pre
-            self.inference_dict['memory_out'] = self.model.memory_out
-        if is_training and self.use_vec_obs and trainer_params['normalize']:
-            self.inference_dict['update_mean'] = self.model.update_mean
-            self.inference_dict['update_variance'] = self.model.update_variance
+            self.inference_dict["memory_out"] = self.model.memory_out
+        if is_training and self.use_vec_obs and trainer_params["normalize"]:
+            self.inference_dict["update_mean"] = self.model.update_mean
+            self.inference_dict["update_variance"] = self.model.update_variance
-        self.update_dict = {'value_loss': self.model.value_loss,
-                            'policy_loss': self.model.policy_loss,
-                            'update_batch': self.model.update_batch}
+        self.update_dict = {
+            "value_loss": self.model.value_loss,
+            "policy_loss": self.model.policy_loss,
+            "update_batch": self.model.update_batch,
+        }
-            self.update_dict['forward_loss'] = self.model.forward_loss
-            self.update_dict['inverse_loss'] = self.model.inverse_loss
+            self.update_dict["forward_loss"] = self.model.forward_loss
+            self.update_dict["inverse_loss"] = self.model.inverse_loss
 
     def evaluate(self, brain_info):
         """
         """
-        feed_dict = {self.model.batch_size: len(brain_info.vector_observations),
-                     self.model.sequence_length: 1}
+        feed_dict = {
+            self.model.batch_size: len(brain_info.vector_observations),
+            self.model.sequence_length: 1,
+        }
-                feed_dict[self.model.prev_action] = brain_info.previous_vector_actions.reshape(
-                    [-1, len(self.model.act_size)])
+                feed_dict[
+                    self.model.prev_action
+                ] = brain_info.previous_vector_actions.reshape(
+                    [-1, len(self.model.act_size)]
+                )
-                size=(len(brain_info.vector_observations), self.model.act_size[0]))
+                size=(len(brain_info.vector_observations), self.model.act_size[0])
+            )
-            run_out['random_normal_epsilon'] = epsilon
+            run_out["random_normal_epsilon"] = epsilon
         return run_out
 
     def update(self, mini_batch, num_sequences):
         :param mini_batch: Experience batch.
         :return: Output from update process.
         """
-        feed_dict = {self.model.batch_size: num_sequences,
-                     self.model.sequence_length: self.sequence_length,
-                     self.model.mask_input: mini_batch['masks'].flatten(),
-                     self.model.returns_holder: mini_batch['discounted_returns'].flatten(),
-                     self.model.old_value: mini_batch['value_estimates'].flatten(),
-                     self.model.advantage: mini_batch['advantages'].reshape([-1, 1]),
-                     self.model.all_old_log_probs: mini_batch['action_probs'].reshape(
-                         [-1, sum(self.model.act_size)])}
+        feed_dict = {
+            self.model.batch_size: num_sequences,
+            self.model.sequence_length: self.sequence_length,
+            self.model.mask_input: mini_batch["masks"].flatten(),
+            self.model.returns_holder: mini_batch["discounted_returns"].flatten(),
+            self.model.old_value: mini_batch["value_estimates"].flatten(),
+            self.model.advantage: mini_batch["advantages"].reshape([-1, 1]),
+            self.model.all_old_log_probs: mini_batch["action_probs"].reshape(
+                [-1, sum(self.model.act_size)]
+            ),
+        }
-            feed_dict[self.model.output_pre] = mini_batch['actions_pre'].reshape(
-                [-1, self.model.act_size[0]])
-            feed_dict[self.model.epsilon] = mini_batch['random_normal_epsilon'].reshape(
-                [-1, self.model.act_size[0]])
+            feed_dict[self.model.output_pre] = mini_batch["actions_pre"].reshape(
+                [-1, self.model.act_size[0]]
+            )
+            feed_dict[self.model.epsilon] = mini_batch["random_normal_epsilon"].reshape(
+                [-1, self.model.act_size[0]]
+            )
-            feed_dict[self.model.action_holder] = mini_batch['actions'].reshape(
-                [-1, len(self.model.act_size)])
+            feed_dict[self.model.action_holder] = mini_batch["actions"].reshape(
+                [-1, len(self.model.act_size)]
+            )
-                feed_dict[self.model.prev_action] = mini_batch['prev_action'].reshape(
-                    [-1, len(self.model.act_size)])
-            feed_dict[self.model.action_masks] = mini_batch['action_mask'].reshape(
-                [-1, sum(self.brain.vector_action_space_size)])
+                feed_dict[self.model.prev_action] = mini_batch["prev_action"].reshape(
+                    [-1, len(self.model.act_size)]
+                )
+            feed_dict[self.model.action_masks] = mini_batch["action_mask"].reshape(
+                [-1, sum(self.brain.vector_action_space_size)]
+            )
-            feed_dict[self.model.vector_in] = mini_batch['vector_obs'].reshape(
-                [-1, self.vec_obs_size])
+            feed_dict[self.model.vector_in] = mini_batch["vector_obs"].reshape(
+                [-1, self.vec_obs_size]
+            )
-                feed_dict[self.model.next_vector_in] = mini_batch['next_vector_in'].reshape(
-                    [-1, self.vec_obs_size])
+                feed_dict[self.model.next_vector_in] = mini_batch[
+                    "next_vector_in"
+                ].reshape([-1, self.vec_obs_size])
-                _obs = mini_batch['visual_obs%d' % i]
+                _obs = mini_batch["visual_obs%d" % i]
                 if self.sequence_length > 1 and self.use_recurrent:
                     (_batch, _seq, _w, _h, _c) = _obs.shape
                     feed_dict[self.model.visual_in[i]] = _obs.reshape([-1, _w, _h, _c])
                 for i, _ in enumerate(self.model.visual_in):
-                    _obs = mini_batch['next_visual_obs%d' % i]
+                    _obs = mini_batch["next_visual_obs%d" % i]
-                        feed_dict[self.model.next_visual_in[i]] = _obs.reshape([-1, _w, _h, _c])
+                        feed_dict[self.model.next_visual_in[i]] = _obs.reshape(
+                            [-1, _w, _h, _c]
+                        )
-            mem_in = mini_batch['memory'][:, 0, :]
+            mem_in = mini_batch["memory"][:, 0, :]
             feed_dict[self.model.memory_in] = mem_in
         self.has_updated = True
         run_out = self._execute_model(feed_dict, self.update_dict)
             if len(curr_info.agents) == 0:
                 return []
 
-            feed_dict = {self.model.batch_size: len(next_info.vector_observations),
-                         self.model.sequence_length: 1}
+            feed_dict = {
+                self.model.batch_size: len(next_info.vector_observations),
+                self.model.sequence_length: 1,
+            }
-                feed_dict[self.model.selected_actions] = next_info.previous_vector_actions
+                feed_dict[
+                    self.model.selected_actions
+                ] = next_info.previous_vector_actions
-                feed_dict[self.model.next_visual_in[i]] = next_info.visual_observations[i]
+                feed_dict[self.model.next_visual_in[i]] = next_info.visual_observations[
+                    i
+                ]
             if self.use_vec_obs:
                 feed_dict[self.model.vector_in] = curr_info.vector_observations
                 feed_dict[self.model.next_vector_in] = next_info.vector_observations
                 feed_dict[self.model.memory_in] = curr_info.memories
-            intrinsic_rewards = self.sess.run(self.model.intrinsic_reward,
-                                              feed_dict=feed_dict) * float(self.has_updated)
+            intrinsic_rewards = self.sess.run(
+                self.model.intrinsic_reward, feed_dict=feed_dict
+            ) * float(self.has_updated)
             return intrinsic_rewards
         else:
             return None
         """
         feed_dict = {self.model.batch_size: 1, self.model.sequence_length: 1}
         for i in range(len(brain_info.visual_observations)):
-            feed_dict[self.model.visual_in[i]] = [brain_info.visual_observations[i][idx]]
+            feed_dict[self.model.visual_in[i]] = [
+                brain_info.visual_observations[i][idx]
+            ]
         if self.use_vec_obs:
             feed_dict[self.model.vector_in] = [brain_info.vector_observations[idx]]
         if self.use_recurrent:
         if not self.use_continuous_act and self.use_recurrent:
-            feed_dict[self.model.prev_action] = brain_info.previous_vector_actions[idx].reshape(
-                [-1, len(self.model.act_size)])
+            feed_dict[self.model.prev_action] = brain_info.previous_vector_actions[
+                idx
+            ].reshape([-1, len(self.model.act_size)])
         value_estimate = self.sess.run(self.model.value, feed_dict)
         return value_estimate
 
         Updates reward value for policy.
         :param new_reward: New reward to save.
         """
-        self.sess.run(self.model.update_reward,
-                      feed_dict={self.model.new_reward: new_reward})
+        self.sess.run(
+            self.model.update_reward, feed_dict={self.model.new_reward: new_reward}
+        )

ml-agents/mlagents/trainers/ppo/trainer.py

 class PPOTrainer(Trainer):
     """The PPOTrainer is an implementation of the PPO algorithm."""
 
-    def __init__(self, brain, reward_buff_cap, trainer_parameters, training,
-                 load, seed, run_id):
+    def __init__(
+        self, brain, reward_buff_cap, trainer_parameters, training, load, seed, run_id
+    ):
         """
         Responsible for collecting experiences and training PPO model.
         :param trainer_parameters: The parameters for the trainer (dictionary).
         :param run_id: The The identifier of the current run
         """
-        super(PPOTrainer, self).__init__(brain, trainer_parameters,
-                                         training, run_id)
-        self.param_keys = ['batch_size', 'beta', 'buffer_size', 'epsilon', 'gamma', 'hidden_units', 'lambd',
-                           'learning_rate', 'max_steps', 'normalize', 'num_epoch', 'num_layers',
-                           'time_horizon', 'sequence_length', 'summary_freq', 'use_recurrent',
-                           'summary_path', 'memory_size', 'use_curiosity', 'curiosity_strength',
-                           'curiosity_enc_size', 'model_path']
+        super(PPOTrainer, self).__init__(brain, trainer_parameters, training, run_id)
+        self.param_keys = [
+            "batch_size",
+            "beta",
+            "buffer_size",
+            "epsilon",
+            "gamma",
+            "hidden_units",
+            "lambd",
+            "learning_rate",
+            "max_steps",
+            "normalize",
+            "num_epoch",
+            "num_layers",
+            "time_horizon",
+            "sequence_length",
+            "summary_freq",
+            "use_recurrent",
+            "summary_path",
+            "memory_size",
+            "use_curiosity",
+            "curiosity_strength",
+            "curiosity_enc_size",
+            "model_path",
+        ]
-        self.use_curiosity = bool(trainer_parameters['use_curiosity'])
+        self.use_curiosity = bool(trainer_parameters["use_curiosity"])
-        self.policy = PPOPolicy(seed, brain, trainer_parameters,
-                                self.is_training, load)
+        self.policy = PPOPolicy(seed, brain, trainer_parameters, self.is_training, load)
-        stats = {'Environment/Cumulative Reward': [], 'Environment/Episode Length': [],
-                 'Policy/Value Estimate': [], 'Policy/Entropy': [], 'Losses/Value Loss': [],
-                 'Losses/Policy Loss': [], 'Policy/Learning Rate': []}
+        stats = {
+            "Environment/Cumulative Reward": [],
+            "Environment/Episode Length": [],
+            "Policy/Value Estimate": [],
+            "Policy/Entropy": [],
+            "Losses/Value Loss": [],
+            "Losses/Policy Loss": [],
+            "Policy/Learning Rate": [],
+        }
-            stats['Losses/Forward Loss'] = []
-            stats['Losses/Inverse Loss'] = []
-            stats['Policy/Curiosity Reward'] = []
+            stats["Losses/Forward Loss"] = []
+            stats["Losses/Inverse Loss"] = []
+            stats["Policy/Curiosity Reward"] = []
             self.intrinsic_rewards = {}
         self.stats = stats
 
         self.episode_steps = {}
 
     def __str__(self):
-        return '''Hyperparameters for the PPO Trainer of brain {0}: \n{1}'''.format(
-            self.brain_name, '\n'.join(['\t{0}:\t{1}'.format(x, self.trainer_parameters[x]) for x in self.param_keys]))
+        return """Hyperparameters for the PPO Trainer of brain {0}: \n{1}""".format(
+            self.brain_name,
+            "\n".join(
+                [
+                    "\t{0}:\t{1}".format(x, self.trainer_parameters[x])
+                    for x in self.param_keys
+                ]
+            ),
+        )
 
     @property
     def parameters(self):
         Returns the maximum number of steps. Is used to know when the trainer should be stopped.
         :return: The maximum number of steps of the trainer
         """
-        return float(self.trainer_parameters['max_steps'])
+        return float(self.trainer_parameters["max_steps"])
 
     @property
     def get_step(self):
         """
         Increment the step count of the trainer and Updates the last reward
         """
-        if len(self.stats['Environment/Cumulative Reward']) > 0:
-            mean_reward = np.mean(self.stats['Environment/Cumulative Reward'])
+        if len(self.stats["Environment/Cumulative Reward"]) > 0:
+            mean_reward = np.mean(self.stats["Environment/Cumulative Reward"])
             self.policy.update_reward(mean_reward)
         self.policy.increment_step()
         self.step = self.policy.get_current_step()
                 agent_brain_info = next_info
             agent_index = agent_brain_info.agents.index(agent_id)
             for i in range(len(next_info.visual_observations)):
-                visual_observations[i].append(agent_brain_info.visual_observations[i][agent_index])
-            vector_observations.append(agent_brain_info.vector_observations[agent_index])
+                visual_observations[i].append(
+                    agent_brain_info.visual_observations[i][agent_index]
+                )
+            vector_observations.append(
+                agent_brain_info.vector_observations[agent_index]
+            )
             text_observations.append(agent_brain_info.text_observations[agent_index])
             if self.policy.use_recurrent:
                 if len(agent_brain_info.memories) > 0:
             local_dones.append(agent_brain_info.local_done[agent_index])
             max_reacheds.append(agent_brain_info.max_reached[agent_index])
             agents.append(agent_brain_info.agents[agent_index])
-            prev_vector_actions.append(agent_brain_info.previous_vector_actions[agent_index])
-            prev_text_actions.append(agent_brain_info.previous_text_actions[agent_index])
+            prev_vector_actions.append(
+                agent_brain_info.previous_vector_actions[agent_index]
+            )
+            prev_text_actions.append(
+                agent_brain_info.previous_text_actions[agent_index]
+            )
-        curr_info = BrainInfo(visual_observations, vector_observations, text_observations,
-                              memories, rewards, agents, local_dones, prev_vector_actions,
-                              prev_text_actions, max_reacheds, action_masks)
+        curr_info = BrainInfo(
+            visual_observations,
+            vector_observations,
+            text_observations,
+            memories,
+            rewards,
+            agents,
+            local_dones,
+            prev_vector_actions,
+            prev_text_actions,
+            max_reacheds,
+            action_masks,
+        )
-    def add_experiences(self, curr_all_info: AllBrainInfo, next_all_info: AllBrainInfo, take_action_outputs):
+    def add_experiences(
+        self,
+        curr_all_info: AllBrainInfo,
+        next_all_info: AllBrainInfo,
+        take_action_outputs,
+    ):
         """
         Adds experiences to each agent's experience history.
         :param curr_all_info: Dictionary of all current brains and corresponding BrainInfo.
         self.trainer_metrics.start_experience_collection_timer()
         if take_action_outputs:
-            self.stats['Policy/Value Estimate'].append(take_action_outputs['value'].mean())
-            self.stats['Policy/Entropy'].append(take_action_outputs['entropy'].mean())
-            self.stats['Policy/Learning Rate'].append(take_action_outputs['learning_rate'])
+            self.stats["Policy/Value Estimate"].append(
+                take_action_outputs["value"].mean()
+            )
+            self.stats["Policy/Entropy"].append(take_action_outputs["entropy"].mean())
+            self.stats["Policy/Learning Rate"].append(
+                take_action_outputs["learning_rate"]
+            )
 
         curr_info = curr_all_info[self.brain_name]
         next_info = next_all_info[self.brain_name]
-            self.training_buffer[agent_id].last_take_action_outputs = take_action_outputs
+            self.training_buffer[
+                agent_id
+            ].last_take_action_outputs = take_action_outputs
 
         if curr_info.agents != next_info.agents:
             curr_to_use = self.construct_curr_info(next_info)
 
         for agent_id in next_info.agents:
             stored_info = self.training_buffer[agent_id].last_brain_info
-            stored_take_action_outputs = self.training_buffer[agent_id].last_take_action_outputs
+            stored_take_action_outputs = self.training_buffer[
+                agent_id
+            ].last_take_action_outputs
-                        self.training_buffer[agent_id]['visual_obs%d' % i].append(
-                            stored_info.visual_observations[i][idx])
-                        self.training_buffer[agent_id]['next_visual_obs%d' % i].append(
-                            next_info.visual_observations[i][next_idx])
+                        self.training_buffer[agent_id]["visual_obs%d" % i].append(
+                            stored_info.visual_observations[i][idx]
+                        )
+                        self.training_buffer[agent_id]["next_visual_obs%d" % i].append(
+                            next_info.visual_observations[i][next_idx]
+                        )
-                        self.training_buffer[agent_id]['vector_obs'].append(stored_info.vector_observations[idx])
-                        self.training_buffer[agent_id]['next_vector_in'].append(
-                            next_info.vector_observations[next_idx])
+                        self.training_buffer[agent_id]["vector_obs"].append(
+                            stored_info.vector_observations[idx]
+                        )
+                        self.training_buffer[agent_id]["next_vector_in"].append(
+                            next_info.vector_observations[next_idx]
+                        )
-                            stored_info.memories = np.zeros((len(stored_info.agents), self.policy.m_size))
-                        self.training_buffer[agent_id]['memory'].append(stored_info.memories[idx])
-                    actions = stored_take_action_outputs['action']
+                            stored_info.memories = np.zeros(
+                                (len(stored_info.agents), self.policy.m_size)
+                            )
+                        self.training_buffer[agent_id]["memory"].append(
+                            stored_info.memories[idx]
+                        )
+                    actions = stored_take_action_outputs["action"]
-                        actions_pre = stored_take_action_outputs['pre_action']
-                        self.training_buffer[agent_id]['actions_pre'].append(actions_pre[idx])
-                        epsilons = stored_take_action_outputs['random_normal_epsilon']
-                        self.training_buffer[agent_id]['random_normal_epsilon'].append(
-                            epsilons[idx])
+                        actions_pre = stored_take_action_outputs["pre_action"]
+                        self.training_buffer[agent_id]["actions_pre"].append(
+                            actions_pre[idx]
+                        )
+                        epsilons = stored_take_action_outputs["random_normal_epsilon"]
+                        self.training_buffer[agent_id]["random_normal_epsilon"].append(
+                            epsilons[idx]
+                        )
-                        self.training_buffer[agent_id]['action_mask'].append(
-                            stored_info.action_masks[idx], padding_value=1)
-                    a_dist = stored_take_action_outputs['log_probs']
-                    value = stored_take_action_outputs['value']
-                    self.training_buffer[agent_id]['actions'].append(actions[idx])
-                    self.training_buffer[agent_id]['prev_action'].append(stored_info.previous_vector_actions[idx])
-                    self.training_buffer[agent_id]['masks'].append(1.0)
+                        self.training_buffer[agent_id]["action_mask"].append(
+                            stored_info.action_masks[idx], padding_value=1
+                        )
+                    a_dist = stored_take_action_outputs["log_probs"]
+                    value = stored_take_action_outputs["value"]
+                    self.training_buffer[agent_id]["actions"].append(actions[idx])
+                    self.training_buffer[agent_id]["prev_action"].append(
+                        stored_info.previous_vector_actions[idx]
+                    )
+                    self.training_buffer[agent_id]["masks"].append(1.0)
-                        self.training_buffer[agent_id]['rewards'].append(next_info.rewards[next_idx] +
-                                                                         intrinsic_rewards[next_idx])
+                        self.training_buffer[agent_id]["rewards"].append(
+                            next_info.rewards[next_idx] + intrinsic_rewards[next_idx]
+                        )
-                        self.training_buffer[agent_id]['rewards'].append(next_info.rewards[next_idx])
-                    self.training_buffer[agent_id]['action_probs'].append(a_dist[idx])
-                    self.training_buffer[agent_id]['value_estimates'].append(value[idx][0])
+                        self.training_buffer[agent_id]["rewards"].append(
+                            next_info.rewards[next_idx]
+                        )
+                    self.training_buffer[agent_id]["action_probs"].append(a_dist[idx])
+                    self.training_buffer[agent_id]["value_estimates"].append(
+                        value[idx][0]
+                    )
                     if agent_id not in self.cumulative_rewards:
                         self.cumulative_rewards[agent_id] = 0
                     self.cumulative_rewards[agent_id] += next_info.rewards[next_idx]
         self.trainer_metrics.start_experience_collection_timer()
         info = new_info[self.brain_name]
         for l in range(len(info.agents)):
-            agent_actions = self.training_buffer[info.agents[l]]['actions']
-            if ((info.local_done[l] or len(agent_actions) > self.trainer_parameters['time_horizon'])
-                    and len(agent_actions) > 0):
+            agent_actions = self.training_buffer[info.agents[l]]["actions"]
+            if (
+                info.local_done[l]
+                or len(agent_actions) > self.trainer_parameters["time_horizon"]
+            ) and len(agent_actions) > 0:
-                        bootstrapping_info = self.training_buffer[agent_id].last_brain_info
+                        bootstrapping_info = self.training_buffer[
+                            agent_id
+                        ].last_brain_info
                         idx = bootstrapping_info.agents.index(agent_id)
                     else:
                         bootstrapping_info = info
-                self.training_buffer[agent_id]['advantages'].set(
+                self.training_buffer[agent_id]["advantages"].set(
-                        rewards=self.training_buffer[agent_id]['rewards'].get_batch(),
-                        value_estimates=self.training_buffer[agent_id]['value_estimates'].get_batch(),
+                        rewards=self.training_buffer[agent_id]["rewards"].get_batch(),
+                        value_estimates=self.training_buffer[agent_id][
+                            "value_estimates"
+                        ].get_batch(),
-                        gamma=self.trainer_parameters['gamma'],
-                        lambd=self.trainer_parameters['lambd']))
-                self.training_buffer[agent_id]['discounted_returns'].set(
-                    self.training_buffer[agent_id]['advantages'].get_batch()
-                    + self.training_buffer[agent_id]['value_estimates'].get_batch())
+                        gamma=self.trainer_parameters["gamma"],
+                        lambd=self.trainer_parameters["lambd"],
+                    )
+                )
+                self.training_buffer[agent_id]["discounted_returns"].set(
+                    self.training_buffer[agent_id]["advantages"].get_batch()
+                    + self.training_buffer[agent_id]["value_estimates"].get_batch()
+                )
-                self.training_buffer.append_update_buffer(agent_id, batch_size=None,
-                                                          training_length=self.policy.sequence_length)
+                self.training_buffer.append_update_buffer(
+                    agent_id,
+                    batch_size=None,
+                    training_length=self.policy.sequence_length,
+                )
-                    self.cumulative_returns_since_policy_update.append(self.
-                                                                       cumulative_rewards.get(agent_id, 0))
-                    self.stats['Environment/Cumulative Reward'].append(
-                        self.cumulative_rewards.get(agent_id, 0))
-                    self.reward_buffer.appendleft(self.cumulative_rewards.get(agent_id, 0))
-                    self.stats['Environment/Episode Length'].append(
-                        self.episode_steps.get(agent_id, 0))
+                    self.cumulative_returns_since_policy_update.append(
+                        self.cumulative_rewards.get(agent_id, 0)
+                    )
+                    self.stats["Environment/Cumulative Reward"].append(
+                        self.cumulative_rewards.get(agent_id, 0)
+                    )
+                    self.reward_buffer.appendleft(
+                        self.cumulative_rewards.get(agent_id, 0)
+                    )
+                    self.stats["Environment/Episode Length"].append(
+                        self.episode_steps.get(agent_id, 0)
+                    )
-                        self.stats['Policy/Curiosity Reward'].append(
-                            self.intrinsic_rewards.get(agent_id, 0))
+                        self.stats["Policy/Curiosity Reward"].append(
+                            self.intrinsic_rewards.get(agent_id, 0)
+                        )
                         self.intrinsic_rewards[agent_id] = 0
         self.trainer_metrics.end_experience_collection_timer()
 
         Returns whether or not the trainer has enough elements to run update model
         :return: A boolean corresponding to whether or not update_model() can be run
         """
-        size_of_buffer = len(self.training_buffer.update_buffer['actions'])
-        return size_of_buffer > max(int(self.trainer_parameters['buffer_size'] / self.policy.sequence_length), 1)
+        size_of_buffer = len(self.training_buffer.update_buffer["actions"])
+        return size_of_buffer > max(
+            int(self.trainer_parameters["buffer_size"] / self.policy.sequence_length), 1
+        )
 
     def update_policy(self):
         """
-            number_experiences=len(self.training_buffer.update_buffer['actions']),
-            mean_return=float(np.mean(self.cumulative_returns_since_policy_update)))
-        n_sequences = max(int(self.trainer_parameters['batch_size'] / self.policy.sequence_length), 1)
+            number_experiences=len(self.training_buffer.update_buffer["actions"]),
+            mean_return=float(np.mean(self.cumulative_returns_since_policy_update)),
+        )
+        n_sequences = max(
+            int(self.trainer_parameters["batch_size"] / self.policy.sequence_length), 1
+        )
-        advantages = self.training_buffer.update_buffer['advantages'].get_batch()
-        self.training_buffer.update_buffer['advantages'].set(
-            (advantages - advantages.mean()) / (advantages.std() + 1e-10))
-        num_epoch = self.trainer_parameters['num_epoch']
+        advantages = self.training_buffer.update_buffer["advantages"].get_batch()
+        self.training_buffer.update_buffer["advantages"].set(
+            (advantages - advantages.mean()) / (advantages.std() + 1e-10)
+        )
+        num_epoch = self.trainer_parameters["num_epoch"]
-            for l in range(len(self.training_buffer.update_buffer['actions']) // n_sequences):
+            for l in range(
+                len(self.training_buffer.update_buffer["actions"]) // n_sequences
+            ):
-                run_out = self.policy.update(buffer.make_mini_batch(start, end), n_sequences)
-                value_total.append(run_out['value_loss'])
-                policy_total.append(np.abs(run_out['policy_loss']))
+                run_out = self.policy.update(
+                    buffer.make_mini_batch(start, end), n_sequences
+                )
+                value_total.append(run_out["value_loss"])
+                policy_total.append(np.abs(run_out["policy_loss"]))
-                    inverse_total.append(run_out['inverse_loss'])
-                    forward_total.append(run_out['forward_loss'])
-        self.stats['Losses/Value Loss'].append(np.mean(value_total))
-        self.stats['Losses/Policy Loss'].append(np.mean(policy_total))
+                    inverse_total.append(run_out["inverse_loss"])
+                    forward_total.append(run_out["forward_loss"])
+        self.stats["Losses/Value Loss"].append(np.mean(value_total))
+        self.stats["Losses/Policy Loss"].append(np.mean(policy_total))
-            self.stats['Losses/Forward Loss'].append(np.mean(forward_total))
-            self.stats['Losses/Inverse Loss'].append(np.mean(inverse_total))
+            self.stats["Losses/Forward Loss"].append(np.mean(forward_total))
+            self.stats["Losses/Inverse Loss"].append(np.mean(inverse_total))
+
 
 def discount_rewards(r, gamma=0.99, value_next=0.0):
     """

ml-agents/mlagents/trainers/tests/test_barracuda_converter.py

 
 def test_barracuda_converter():
     path_prefix = os.path.dirname(os.path.abspath(__file__))
-    tmpfile = os.path.join(tempfile._get_default_tempdir(), next(tempfile._get_candidate_names()) + '.nn')
+    tmpfile = os.path.join(
+        tempfile._get_default_tempdir(), next(tempfile._get_candidate_names()) + ".nn"
+    )
-    tf2bc.convert(path_prefix+'/BasicLearning.pb', tmpfile)
+    tf2bc.convert(path_prefix + "/BasicLearning.pb", tmpfile)
-    assert (os.path.isfile(tmpfile))
+    assert os.path.isfile(tmpfile)
-    # 100 bytes is high enough to prove that conversion was successful 
-    assert (os.path.getsize(tmpfile) > 100)
+    # 100 bytes is high enough to prove that conversion was successful
+    assert os.path.getsize(tmpfile) > 100
 
     # cleanup
     os.remove(tmpfile)

ml-agents/mlagents/trainers/tests/test_bc.py

 @pytest.fixture
 def dummy_config():
     return yaml.load(
-        '''
+        """
             hidden_units: 128
             learning_rate: 3.0e-4
             num_layers: 2
-            ''')
+            """
+    )
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-        discrete_action=False, visual_inputs=0)
-    env = UnityEnvironment(' ')
+        discrete_action=False, visual_inputs=0
+    )
+    env = UnityEnvironment(" ")
-    trainer_parameters['model_path'] = model_path
-    trainer_parameters['keep_checkpoints'] = 3
+    trainer_parameters["model_path"] = model_path
+    trainer_parameters["keep_checkpoints"] = 3
-    assert run_out['action'].shape == (3, 2)
+    assert run_out["action"].shape == (3, 2)
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=False, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=False, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]])}
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=True, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=True, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            feed_dict = {model.batch_size: 2,
-                         model.dropout_rate: 1.0,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.action_masks: np.ones([2, 2])}
+            feed_dict = {
+                model.batch_size: 2,
+                model.dropout_rate: 1.0,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.action_masks: np.ones([2, 2]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=True, visual_inputs=2)
-            env = UnityEnvironment(' ')
+                discrete_action=True, visual_inputs=2
+            )
+            env = UnityEnvironment(" ")
-            feed_dict = {model.batch_size: 2,
-                         model.dropout_rate: 1.0,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.visual_in[1]: np.ones([2, 40, 30, 3]),
-                         model.action_masks: np.ones([2, 2])}
+            feed_dict = {
+                model.batch_size: 2,
+                model.dropout_rate: 1.0,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.visual_in[1]: np.ones([2, 40, 30, 3]),
+                model.action_masks: np.ones([2, 2]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=False, visual_inputs=2)
-            env = UnityEnvironment(' ')
+                discrete_action=False, visual_inputs=2
+            )
+            env = UnityEnvironment(" ")
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.visual_in[1]: np.ones([2, 40, 30, 3])}
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.visual_in[1]: np.ones([2, 40, 30, 3]),
+            }
-if __name__ == '__main__':
+if __name__ == "__main__":
     pytest.main()

ml-agents/mlagents/trainers/tests/test_buffer.py

     b = Buffer()
     for fake_agent_id in range(4):
         for step in range(9):
-            b[fake_agent_id]['vector_observation'].append(
-                [100 * fake_agent_id + 10 * step + 1,
-                 100 * fake_agent_id + 10 * step + 2,
-                 100 * fake_agent_id + 10 * step + 3]
+            b[fake_agent_id]["vector_observation"].append(
+                [
+                    100 * fake_agent_id + 10 * step + 1,
+                    100 * fake_agent_id + 10 * step + 2,
+                    100 * fake_agent_id + 10 * step + 3,
+                ]
+            )
+            b[fake_agent_id]["action"].append(
+                [
+                    100 * fake_agent_id + 10 * step + 4,
+                    100 * fake_agent_id + 10 * step + 5,
+                ]
-            b[fake_agent_id]['action'].append([100 * fake_agent_id + 10 * step + 4,
-                                               100 * fake_agent_id + 10 * step + 5])
-    a = b[1]['vector_observation'].get_batch(batch_size=2, training_length=1, sequential=True)
+    a = b[1]["vector_observation"].get_batch(
+        batch_size=2, training_length=1, sequential=True
+    )
-    a = b[2]['vector_observation'].get_batch(batch_size=2, training_length=3, sequential=True)
-    assert_array(a, np.array([
-        [[231, 232, 233], [241, 242, 243], [251, 252, 253]],
-        [[261, 262, 263], [271, 272, 273], [281, 282, 283]]
-    ]))
-    a = b[2]['vector_observation'].get_batch(batch_size=2, training_length=3, sequential=False)
-    assert_array(a, np.array([
-        [[251, 252, 253], [261, 262, 263], [271, 272, 273]],
-        [[261, 262, 263], [271, 272, 273], [281, 282, 283]]
-    ]))
+    a = b[2]["vector_observation"].get_batch(
+        batch_size=2, training_length=3, sequential=True
+    )
+    assert_array(
+        a,
+        np.array(
+            [
+                [[231, 232, 233], [241, 242, 243], [251, 252, 253]],
+                [[261, 262, 263], [271, 272, 273], [281, 282, 283]],
+            ]
+        ),
+    )
+    a = b[2]["vector_observation"].get_batch(
+        batch_size=2, training_length=3, sequential=False
+    )
+    assert_array(
+        a,
+        np.array(
+            [
+                [[251, 252, 253], [261, 262, 263], [271, 272, 273]],
+                [[261, 262, 263], [271, 272, 273], [281, 282, 283]],
+            ]
+        ),
+    )
-    assert len(b.update_buffer['action']) == 10
-    assert np.array(b.update_buffer['action']).shape == (10, 2, 2)
+    assert len(b.update_buffer["action"]) == 10
+    assert np.array(b.update_buffer["action"]).shape == (10, 2, 2)
-    assert c['action'].shape == (1, 2, 2)
+    assert c["action"].shape == (1, 2, 2)

ml-agents/mlagents/trainers/tests/test_curriculum.py

 from mlagents.trainers import Curriculum
 
 
-dummy_curriculum_json_str = '''
+dummy_curriculum_json_str = """
     {
         "measure" : "reward",
         "thresholds" : [10, 20, 50],
             "param3" : [0.2, 0.3, 0.7, 0.9]
         }
     }
-    '''
+    """
-bad_curriculum_json_str = '''
+bad_curriculum_json_str = """
     {
         "measure" : "reward",
         "thresholds" : [10, 20, 50],
             "param3" : [0.2, 0.3, 0.7, 0.9]
         }
     }
-    '''
+    """
+
-    return 'TestBrain.json'
+    return "TestBrain.json"
 
 
 @pytest.fixture
 
-@patch('builtins.open', new_callable=mock_open, read_data=dummy_curriculum_json_str)
+@patch("builtins.open", new_callable=mock_open, read_data=dummy_curriculum_json_str)
-    assert curriculum._brain_name == 'TestBrain'
+    assert curriculum._brain_name == "TestBrain"
-    assert curriculum.measure == 'reward'
+    assert curriculum.measure == "reward"
-@patch('builtins.open', new_callable=mock_open, read_data=bad_curriculum_json_str)
-def test_init_curriculum_bad_curriculum_raises_error(mock_file, location, default_reset_parameters):
+@patch("builtins.open", new_callable=mock_open, read_data=bad_curriculum_json_str)
+def test_init_curriculum_bad_curriculum_raises_error(
+    mock_file, location, default_reset_parameters
+):
-@patch('builtins.open', new_callable=mock_open, read_data=dummy_curriculum_json_str)
+@patch("builtins.open", new_callable=mock_open, read_data=dummy_curriculum_json_str)
 def test_increment_lesson(mock_file, location, default_reset_parameters):
     curriculum = Curriculum(location, default_reset_parameters)
     assert curriculum.lesson_num == 0
     assert curriculum.lesson_num == 3
 
 
-@patch('builtins.open', new_callable=mock_open, read_data=dummy_curriculum_json_str)
+@patch("builtins.open", new_callable=mock_open, read_data=dummy_curriculum_json_str)
-    curriculum = Curriculum('TestBrain.json', {"param1": 1, "param2": 1, "param3": 1})
+    curriculum = Curriculum("TestBrain.json", {"param1": 1, "param2": 1, "param3": 1})
-    assert curriculum.get_config() == {'param1': 0.3, 'param2': 20, 'param3': 0.7}
+    assert curriculum.get_config() == {"param1": 0.3, "param2": 20, "param3": 0.7}
     assert curriculum.get_config(0) == {"param1": 0.7, "param2": 100, "param3": 0.2}

ml-agents/mlagents/trainers/tests/test_demo_loader.py

 
 def test_load_demo():
     path_prefix = os.path.dirname(os.path.abspath(__file__))
-    brain_parameters, brain_infos, total_expected = load_demonstration(path_prefix+'/test.demo')
-    assert (brain_parameters.brain_name == "Ball3DBrain")
-    assert (brain_parameters.vector_observation_space_size == 8)
-    assert (len(brain_infos) == total_expected)
+    brain_parameters, brain_infos, total_expected = load_demonstration(
+        path_prefix + "/test.demo"
+    )
+    assert brain_parameters.brain_name == "Ball3DBrain"
+    assert brain_parameters.vector_observation_space_size == 8
+    assert len(brain_infos) == total_expected
-    assert (len(demo_buffer.update_buffer['actions']) == total_expected - 1)
+    assert len(demo_buffer.update_buffer["actions"]) == total_expected - 1

ml-agents/mlagents/trainers/tests/test_learn.py

 @pytest.fixture
 def basic_options():
     return {
-        '--docker-target-name': 'None',
-        '--env': 'None',
-        '--run-id': 'ppo',
-        '--load': False,
-        '--train': False,
-        '--save-freq': '50000',
-        '--keep-checkpoints': '5',
-        '--base-port': '5005',
-        '--num-envs': '1',
-        '--curriculum': 'None',
-        '--lesson': '0',
-        '--slow': False,
-        '--no-graphics': False,
-        '<trainer-config-path>': 'basic_path',
-        '--debug': False,
+        "--docker-target-name": "None",
+        "--env": "None",
+        "--run-id": "ppo",
+        "--load": False,
+        "--train": False,
+        "--save-freq": "50000",
+        "--keep-checkpoints": "5",
+        "--base-port": "5005",
+        "--num-envs": "1",
+        "--curriculum": "None",
+        "--lesson": "0",
+        "--slow": False,
+        "--no-graphics": False,
+        "<trainer-config-path>": "basic_path",
+        "--debug": False,
-@patch('mlagents.trainers.learn.SubprocessUnityEnvironment')
-@patch('mlagents.trainers.learn.create_environment_factory')
-@patch('mlagents.trainers.learn.load_config')
+@patch("mlagents.trainers.learn.SubprocessUnityEnvironment")
+@patch("mlagents.trainers.learn.create_environment_factory")
+@patch("mlagents.trainers.learn.load_config")
-    mock_env.academy_name = 'TestAcademyName'
+    mock_env.academy_name = "TestAcademyName"
     create_environment_factory.return_value = mock_env
     trainer_config_mock = MagicMock()
     load_config.return_value = trainer_config_mock
         with patch.object(TrainerController, "start_learning", MagicMock()):
             learn.run_training(0, 0, basic_options(), MagicMock())
             mock_init.assert_called_once_with(
-                './models/ppo-0',
-                './summaries',
-                'ppo-0',
+                "./models/ppo-0",
+                "./summaries",
+                "ppo-0",
                 50000,
                 None,
                 False,
                 subproc_env_mock.return_value.external_brains,
                 0,
-                True
+                True,
-@patch('mlagents.trainers.learn.SubprocessUnityEnvironment')
-@patch('mlagents.trainers.learn.create_environment_factory')
-@patch('mlagents.trainers.learn.load_config')
+@patch("mlagents.trainers.learn.SubprocessUnityEnvironment")
+@patch("mlagents.trainers.learn.create_environment_factory")
+@patch("mlagents.trainers.learn.load_config")
-    mock_env.academy_name = 'TestAcademyName'
+    mock_env.academy_name = "TestAcademyName"
-    options_with_docker_target['--docker-target-name'] = 'dockertarget'
+    options_with_docker_target["--docker-target-name"] = "dockertarget"
 
     mock_init = MagicMock(return_value=None)
     with patch.object(TrainerController, "__init__", mock_init):
-            assert(mock_init.call_args[0][0] == '/dockertarget/models/ppo-0')
-            assert(mock_init.call_args[0][1] == '/dockertarget/summaries')
+            assert mock_init.call_args[0][0] == "/dockertarget/models/ppo-0"
+            assert mock_init.call_args[0][1] == "/dockertarget/summaries"

ml-agents/mlagents/trainers/tests/test_meta_curriculum.py

     """This class allows us to test MetaCurriculum objects without calling
     MetaCurriculum's __init__ function.
     """
+
     def __init__(self, brains_to_curriculums):
         self._brains_to_curriculums = brains_to_curriculums
 
-    return {'param1' : 1, 'param2' : 2, 'param3' : 3}
+    return {"param1": 1, "param2": 2, "param3": 3}
-    return {'param4' : 4, 'param5' : 5, 'param6' : 6}
+    return {"param4": 4, "param5": 5, "param6": 6}
-    return {'Brain1' : 0.2, 'Brain2' : 0.3}
+    return {"Brain1": 0.2, "Brain2": 0.3}
-    return {'Brain1' : 7, 'Brain2' : 8}
+    return {"Brain1": 7, "Brain2": 8}
-@patch('mlagents.trainers.Curriculum.get_config', return_value={})
-@patch('mlagents.trainers.Curriculum.__init__', return_value=None)
-@patch('os.listdir', return_value=['Brain1.json', 'Brain2.json'])
-def test_init_meta_curriculum_happy_path(listdir, mock_curriculum_init,
-                                         mock_curriculum_get_config,
-                                         default_reset_parameters):
-    meta_curriculum = MetaCurriculum('test/', default_reset_parameters)
+@patch("mlagents.trainers.Curriculum.get_config", return_value={})
+@patch("mlagents.trainers.Curriculum.__init__", return_value=None)
+@patch("os.listdir", return_value=["Brain1.json", "Brain2.json"])
+def test_init_meta_curriculum_happy_path(
+    listdir, mock_curriculum_init, mock_curriculum_get_config, default_reset_parameters
+):
+    meta_curriculum = MetaCurriculum("test/", default_reset_parameters)
-    assert 'Brain1' in meta_curriculum.brains_to_curriculums
-    assert 'Brain2' in meta_curriculum.brains_to_curriculums
+    assert "Brain1" in meta_curriculum.brains_to_curriculums
+    assert "Brain2" in meta_curriculum.brains_to_curriculums
-    calls = [call('test/Brain1.json', default_reset_parameters),
-             call('test/Brain2.json', default_reset_parameters)]
+    calls = [
+        call("test/Brain1.json", default_reset_parameters),
+        call("test/Brain2.json", default_reset_parameters),
+    ]
-@patch('os.listdir', side_effect=NotADirectoryError())
+@patch("os.listdir", side_effect=NotADirectoryError())
-        MetaCurriculum('test/', default_reset_parameters)
+        MetaCurriculum("test/", default_reset_parameters)
-@patch('mlagents.trainers.Curriculum')
-@patch('mlagents.trainers.Curriculum')
+@patch("mlagents.trainers.Curriculum")
+@patch("mlagents.trainers.Curriculum")
-    meta_curriculum = MetaCurriculumTest({'Brain1' : curriculum_a,
-                                          'Brain2' : curriculum_b})
+    meta_curriculum = MetaCurriculumTest(
+        {"Brain1": curriculum_a, "Brain2": curriculum_b}
+    )
-    meta_curriculum.lesson_nums = {'Brain1' : 1, 'Brain2' : 3}
+    meta_curriculum.lesson_nums = {"Brain1": 1, "Brain2": 3}
-@patch('mlagents.trainers.Curriculum')
-@patch('mlagents.trainers.Curriculum')
+@patch("mlagents.trainers.Curriculum")
+@patch("mlagents.trainers.Curriculum")
-    meta_curriculum = MetaCurriculumTest({'Brain1' : curriculum_a,
-                                          'Brain2' : curriculum_b})
+    meta_curriculum = MetaCurriculumTest(
+        {"Brain1": curriculum_a, "Brain2": curriculum_b}
+    )
 
     meta_curriculum.increment_lessons(measure_vals)
 
 
-@patch('mlagents.trainers.Curriculum')
-@patch('mlagents.trainers.Curriculum')
-def test_increment_lessons_with_reward_buff_sizes(curriculum_a, curriculum_b,
-                                                  measure_vals,
-                                                  reward_buff_sizes):
+@patch("mlagents.trainers.Curriculum")
+@patch("mlagents.trainers.Curriculum")
+def test_increment_lessons_with_reward_buff_sizes(
+    curriculum_a, curriculum_b, measure_vals, reward_buff_sizes
+):
-    meta_curriculum = MetaCurriculumTest({'Brain1' : curriculum_a,
-                                          'Brain2' : curriculum_b})
+    meta_curriculum = MetaCurriculumTest(
+        {"Brain1": curriculum_a, "Brain2": curriculum_b}
+    )
-    meta_curriculum.increment_lessons(measure_vals,
-                                      reward_buff_sizes=reward_buff_sizes)
+    meta_curriculum.increment_lessons(measure_vals, reward_buff_sizes=reward_buff_sizes)
-@patch('mlagents.trainers.Curriculum')
-@patch('mlagents.trainers.Curriculum')
+@patch("mlagents.trainers.Curriculum")
+@patch("mlagents.trainers.Curriculum")
-    meta_curriculum = MetaCurriculumTest({'Brain1' : curriculum_a,
-                                          'Brain2' : curriculum_b})
+    meta_curriculum = MetaCurriculumTest(
+        {"Brain1": curriculum_a, "Brain2": curriculum_b}
+    )
 
     meta_curriculum.set_all_curriculums_to_lesson_num(2)
 
 
-@patch('mlagents.trainers.Curriculum')
-@patch('mlagents.trainers.Curriculum')
-def test_get_config(curriculum_a, curriculum_b, default_reset_parameters,
-                    more_reset_parameters):
+@patch("mlagents.trainers.Curriculum")
+@patch("mlagents.trainers.Curriculum")
+def test_get_config(
+    curriculum_a, curriculum_b, default_reset_parameters, more_reset_parameters
+):
-    meta_curriculum = MetaCurriculumTest({'Brain1' : curriculum_a,
-                                          'Brain2' : curriculum_b})
+    meta_curriculum = MetaCurriculumTest(
+        {"Brain1": curriculum_a, "Brain2": curriculum_b}
+    )
 
     assert meta_curriculum.get_config() == default_reset_parameters
 

ml-agents/mlagents/trainers/tests/test_policy.py

 from mlagents.trainers.policy import *
 from unittest.mock import MagicMock
 
+
+
-    return {
-        "use_recurrent": False,
-        "model_path": "my/path"
-    }
+    return {"use_recurrent": False, "model_path": "my/path"}
 
 
 def test_take_action_returns_empty_with_no_agents():
     result = policy.get_action(no_agent_brain_info)
-    assert(result == ActionInfo([], [], [], None, None))
+    assert result == ActionInfo([], [], [], None, None)
 
 
 def test_take_action_returns_nones_on_missing_values():
-    brain_info_with_agents = BrainInfo([], [], [], agents=['an-agent-id'])
+    brain_info_with_agents = BrainInfo([], [], [], agents=["an-agent-id"])
-    assert(result == ActionInfo(None, None, None, None, {}))
+    assert result == ActionInfo(None, None, None, None, {})
 
 
 def test_take_action_returns_action_info_when_available():
-        'action': np.array([1.0]),
-        'memory_out': np.array([2.5]),
-        'value': np.array([1.1])
+        "action": np.array([1.0]),
+        "memory_out": np.array([2.5]),
+        "value": np.array([1.1]),
-    brain_info_with_agents = BrainInfo([], [], [], agents=['an-agent-id'])
+    brain_info_with_agents = BrainInfo([], [], [], agents=["an-agent-id"])
-        policy_eval_out['action'],
-        policy_eval_out['memory_out'],
+        policy_eval_out["action"],
+        policy_eval_out["memory_out"],
-        policy_eval_out['value'],
-        policy_eval_out
+        policy_eval_out["value"],
+        policy_eval_out,
-    assert (result == expected)
+    assert result == expected

ml-agents/mlagents/trainers/tests/test_ppo.py

 @pytest.fixture
 def dummy_config():
     return yaml.load(
-        '''
+        """
         trainer: ppo
         batch_size: 32
         beta: 5.0e-3
         use_curiosity: false
         curiosity_strength: 0.0
         curiosity_enc_size: 1
-        ''')
+        """
+    )
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-        discrete_action=False, visual_inputs=0)
-    env = UnityEnvironment(' ')
+        discrete_action=False, visual_inputs=0
+    )
+    env = UnityEnvironment(" ")
-    trainer_parameters['model_path'] = model_path
-    trainer_parameters['keep_checkpoints'] = 3
-    policy = PPOPolicy(0, env.brains[env.brain_names[0]], trainer_parameters, False, False)
+    trainer_parameters["model_path"] = model_path
+    trainer_parameters["keep_checkpoints"] = 3
+    policy = PPOPolicy(
+        0, env.brains[env.brain_names[0]], trainer_parameters, False, False
+    )
-    assert run_out['action'].shape == (3, 2)
+    assert run_out["action"].shape == (3, 2)
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=False, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=False, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.log_probs, model.value, model.entropy,
-                        model.learning_rate]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]],),
-                         model.epsilon: np.array([[0, 1], [2, 3]])}
+            run_list = [
+                model.output,
+                model.log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.epsilon: np.array([[0, 1], [2, 3]]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=False, visual_inputs=2)
-            env = UnityEnvironment(' ')
+                discrete_action=False, visual_inputs=2
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.log_probs, model.value, model.entropy,
-                        model.learning_rate]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.visual_in[1]: np.ones([2, 40, 30, 3]),
-                         model.epsilon: np.array([[0, 1], [2, 3]])}
+            run_list = [
+                model.output,
+                model.log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.visual_in[1]: np.ones([2, 40, 30, 3]),
+                model.epsilon: np.array([[0, 1], [2, 3]]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=True, visual_inputs=2)
-            env = UnityEnvironment(' ')
+                discrete_action=True, visual_inputs=2
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.visual_in[1]: np.ones([2, 40, 30, 3]),
-                         model.action_masks: np.ones([2, 2],)
-                         }
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.visual_in[1]: np.ones([2, 40, 30, 3]),
+                model.action_masks: np.ones([2, 2]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=True, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=True, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.action_masks: np.ones([2, 2])}
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.action_masks: np.ones([2, 2]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=True, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=True, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            model = PPOModel(env.brains["RealFakeBrain"], use_recurrent=True, m_size=memory_size)
+            model = PPOModel(
+                env.brains["RealFakeBrain"], use_recurrent=True, m_size=memory_size
+            )
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate, model.memory_out]
-            feed_dict = {model.batch_size: 1,
-                         model.sequence_length: 2,
-                         model.prev_action: [[0], [0]],
-                         model.memory_in: np.zeros((1, memory_size)),
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.action_masks: np.ones([1, 2])}
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+                model.memory_out,
+            ]
+            feed_dict = {
+                model.batch_size: 1,
+                model.sequence_length: 2,
+                model.prev_action: [[0], [0]],
+                model.memory_in: np.zeros((1, memory_size)),
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.action_masks: np.ones([1, 2]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=False, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=False, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            model = PPOModel(env.brains["RealFakeBrain"], use_recurrent=True, m_size=memory_size)
+            model = PPOModel(
+                env.brains["RealFakeBrain"], use_recurrent=True, m_size=memory_size
+            )
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate, model.memory_out]
-            feed_dict = {model.batch_size: 1,
-                         model.sequence_length: 2,
-                         model.memory_in: np.zeros((1, memory_size)),
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.epsilon: np.array([[0, 1]])}
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+                model.memory_out,
+            ]
+            feed_dict = {
+                model.batch_size: 1,
+                model.sequence_length: 2,
+                model.memory_in: np.zeros((1, memory_size)),
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.epsilon: np.array([[0, 1]]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=True, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=True, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate, model.intrinsic_reward]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.next_vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                         [3, 4, 5, 3, 4, 5]]),
-                         model.action_holder: [[0], [0]],
-                         model.action_masks: np.ones([2,2])}
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+                model.intrinsic_reward,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.next_vector_in: np.array(
+                    [[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]
+                ),
+                model.action_holder: [[0], [0]],
+                model.action_masks: np.ones([2, 2]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=False, visual_inputs=0)
-            env = UnityEnvironment(' ')
+                discrete_action=False, visual_inputs=0
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate, model.intrinsic_reward]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.next_vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                         [3, 4, 5, 3, 4, 5]]),
-                         model.output: [[0.0, 0.0], [0.0, 0.0]],
-                         model.epsilon: np.array([[0, 1], [2, 3]])}
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+                model.intrinsic_reward,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.next_vector_in: np.array(
+                    [[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]
+                ),
+                model.output: [[0.0, 0.0], [0.0, 0.0]],
+                model.epsilon: np.array([[0, 1], [2, 3]]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=True, visual_inputs=2)
-            env = UnityEnvironment(' ')
+                discrete_action=True, visual_inputs=2
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate, model.intrinsic_reward]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.next_vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                         [3, 4, 5, 3, 4, 5]]),
-                         model.action_holder: [[0], [0]],
-                         model.visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.visual_in[1]: np.ones([2, 40, 30, 3]),
-                         model.next_visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.next_visual_in[1]: np.ones([2, 40, 30, 3]),
-                         model.action_masks: np.ones([2, 2])
-                         }
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+                model.intrinsic_reward,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.next_vector_in: np.array(
+                    [[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]
+                ),
+                model.action_holder: [[0], [0]],
+                model.visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.visual_in[1]: np.ones([2, 40, 30, 3]),
+                model.next_visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.next_visual_in[1]: np.ones([2, 40, 30, 3]),
+                model.action_masks: np.ones([2, 2]),
+            }
-@mock.patch('mlagents.envs.UnityEnvironment.executable_launcher')
-@mock.patch('mlagents.envs.UnityEnvironment.get_communicator')
+@mock.patch("mlagents.envs.UnityEnvironment.executable_launcher")
+@mock.patch("mlagents.envs.UnityEnvironment.get_communicator")
-                discrete_action=False, visual_inputs=2)
-            env = UnityEnvironment(' ')
+                discrete_action=False, visual_inputs=2
+            )
+            env = UnityEnvironment(" ")
-            run_list = [model.output, model.all_log_probs, model.value, model.entropy,
-                        model.learning_rate, model.intrinsic_reward]
-            feed_dict = {model.batch_size: 2,
-                         model.sequence_length: 1,
-                         model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                    [3, 4, 5, 3, 4, 5]]),
-                         model.next_vector_in: np.array([[1, 2, 3, 1, 2, 3],
-                                                         [3, 4, 5, 3, 4, 5]]),
-                         model.output: [[0.0, 0.0], [0.0, 0.0]],
-                         model.visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.visual_in[1]: np.ones([2, 40, 30, 3]),
-                         model.next_visual_in[0]: np.ones([2, 40, 30, 3]),
-                         model.next_visual_in[1]: np.ones([2, 40, 30, 3]),
-                         model.epsilon: np.array([[0, 1], [2, 3]])
-                         }
+            run_list = [
+                model.output,
+                model.all_log_probs,
+                model.value,
+                model.entropy,
+                model.learning_rate,
+                model.intrinsic_reward,
+            ]
+            feed_dict = {
+                model.batch_size: 2,
+                model.sequence_length: 1,
+                model.vector_in: np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
+                model.next_vector_in: np.array(
+                    [[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]
+                ),
+                model.output: [[0.0, 0.0], [0.0, 0.0]],
+                model.visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.visual_in[1]: np.ones([2, 40, 30, 3]),
+                model.next_visual_in[0]: np.ones([2, 40, 30, 3]),
+                model.next_visual_in[1]: np.ones([2, 40, 30, 3]),
+                model.epsilon: np.array([[0, 1], [2, 3]]),
+            }
             sess.run(run_list, feed_dict=feed_dict)
             env.close()
 
     np.testing.assert_array_almost_equal(returns, np.array([0.729, 0.81, 0.9, 1.0]))
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     pytest.main()

ml-agents/mlagents/trainers/tests/test_trainer_controller.py

 @pytest.fixture
 def dummy_config():
     return yaml.load(
-        '''
+        """
         default:
             trainer: ppo
             batch_size: 32
             use_curiosity: false
             curiosity_strength: 0.0
             curiosity_enc_size: 1
-        ''')
+        """
+    )
-        '''
+        """
         default:
             trainer: online_bc
             brain_to_imitate: ExpertBrain
             use_curiosity: false
             curiosity_strength: 0.0
             curiosity_enc_size: 1
-        ''')
+        """
+    )
-        '''
+        """
-            demo_path: '''
-        + os.path.dirname(os.path.abspath(__file__)) + '''/test.demo
+            demo_path: """
+        + os.path.dirname(os.path.abspath(__file__))
+        + """/test.demo
             batches_per_epoch: 16
             batch_size: 32
             beta: 5.0e-3
             use_curiosity: false
             curiosity_strength: 0.0
             curiosity_enc_size: 1
-        ''')
+        """
+    )
+
-    base['testbrain'] = {}
-    base['testbrain']['normalize'] = False
+    base["testbrain"] = {}
+    base["testbrain"]["normalize"] = False
+
-        '''
+        """
         default:
             trainer: incorrect_trainer
             brain_to_imitate: ExpertBrain
             summary_freq: 1000
             use_recurrent: false
             memory_size: 8
-        ''')
+        """
+    )
+
-        model_path='test_model_path',
-        summaries_dir='test_summaries_dir',
-        run_id='test_run_id',
+        model_path="test_model_path",
+        summaries_dir="test_summaries_dir",
+        run_id="test_run_id",
         save_freq=100,
         meta_curriculum=None,
         load=True,
-        external_brains={'testbrain': brain_info},
+        external_brains={"testbrain": brain_info},
-        fast_simulation=True
+        fast_simulation=True,
-@patch('numpy.random.seed')
-@patch('tensorflow.set_random_seed')
+
+@patch("numpy.random.seed")
+@patch("tensorflow.set_random_seed")
-    TrainerController('', '', '1', 1, None, True, False, False, None, {}, seed, True)
+    TrainerController("", "", "1", 1, None, True, False, False, None, {}, seed, True)
-def assert_bc_trainer_constructed(trainer_cls, input_config, tc, expected_brain_info, expected_config):
+def assert_bc_trainer_constructed(
+    trainer_cls, input_config, tc, expected_brain_info, expected_config
+):
-        assert(brain == expected_brain_info)
-        assert(trainer_params == expected_config)
-        assert(training == tc.train_model)
-        assert(load == tc.load_model)
-        assert(seed == tc.seed)
-        assert(run_id == tc.run_id)
+        assert brain == expected_brain_info
+        assert trainer_params == expected_config
+        assert training == tc.train_model
+        assert load == tc.load_model
+        assert seed == tc.seed
+        assert run_id == tc.run_id
-        assert('testbrain' in tc.trainers)
-        assert(isinstance(tc.trainers['testbrain'], trainer_cls))
+        assert "testbrain" in tc.trainers
+        assert isinstance(tc.trainers["testbrain"], trainer_cls)
-def assert_ppo_trainer_constructed(input_config, tc, expected_brain_info,
-                                   expected_config, expected_reward_buff_cap=0):
-    def mock_constructor(self, brain, reward_buff_cap, trainer_parameters, training, load, seed, run_id):
-        self.trainer_metrics = TrainerMetrics('', '')
-        assert(brain == expected_brain_info)
-        assert(trainer_parameters == expected_config)
-        assert(reward_buff_cap == expected_reward_buff_cap)
-        assert(training == tc.train_model)
-        assert(load == tc.load_model)
-        assert(seed == tc.seed)
-        assert(run_id == tc.run_id)
+def assert_ppo_trainer_constructed(
+    input_config, tc, expected_brain_info, expected_config, expected_reward_buff_cap=0
+):
+    def mock_constructor(
+        self, brain, reward_buff_cap, trainer_parameters, training, load, seed, run_id
+    ):
+        self.trainer_metrics = TrainerMetrics("", "")
+        assert brain == expected_brain_info
+        assert trainer_parameters == expected_config
+        assert reward_buff_cap == expected_reward_buff_cap
+        assert training == tc.train_model
+        assert load == tc.load_model
+        assert seed == tc.seed
+        assert run_id == tc.run_id
-        assert('testbrain' in tc.trainers)
-        assert(isinstance(tc.trainers['testbrain'], PPOTrainer))
+        assert "testbrain" in tc.trainers
+        assert isinstance(tc.trainers["testbrain"], PPOTrainer)
-@patch('mlagents.envs.BrainInfo')
+@patch("mlagents.envs.BrainInfo")
-    expected_config = full_config['default']
-    expected_config['summary_path'] = tc.summaries_dir + '/test_run_id_testbrain'
-    expected_config['model_path'] = tc.model_path + '/testbrain'
-    expected_config['keep_checkpoints'] = tc.keep_checkpoints
+    expected_config = full_config["default"]
+    expected_config["summary_path"] = tc.summaries_dir + "/test_run_id_testbrain"
+    expected_config["model_path"] = tc.model_path + "/testbrain"
+    expected_config["keep_checkpoints"] = tc.keep_checkpoints
-    assert_bc_trainer_constructed(OfflineBCTrainer, full_config, tc, brain_info_mock, expected_config)
+    assert_bc_trainer_constructed(
+        OfflineBCTrainer, full_config, tc, brain_info_mock, expected_config
+    )
-@patch('mlagents.envs.BrainInfo')
+@patch("mlagents.envs.BrainInfo")
-    expected_config = full_config['default']
-    expected_config['summary_path'] = tc.summaries_dir + '/test_run_id_testbrain'
-    expected_config['model_path'] = tc.model_path + '/testbrain'
-    expected_config['keep_checkpoints'] = tc.keep_checkpoints
+    expected_config = full_config["default"]
+    expected_config["summary_path"] = tc.summaries_dir + "/test_run_id_testbrain"
+    expected_config["model_path"] = tc.model_path + "/testbrain"
+    expected_config["keep_checkpoints"] = tc.keep_checkpoints
-    expected_config['normalize'] = False
+    expected_config["normalize"] = False
-    assert_bc_trainer_constructed(OfflineBCTrainer, full_config, tc, brain_info_mock, expected_config)
+    assert_bc_trainer_constructed(
+        OfflineBCTrainer, full_config, tc, brain_info_mock, expected_config
+    )
-@patch('mlagents.envs.BrainInfo')
+@patch("mlagents.envs.BrainInfo")
-    expected_config = full_config['default']
-    expected_config['summary_path'] = tc.summaries_dir + '/test_run_id_testbrain'
-    expected_config['model_path'] = tc.model_path + '/testbrain'
-    expected_config['keep_checkpoints'] = tc.keep_checkpoints
+    expected_config = full_config["default"]
+    expected_config["summary_path"] = tc.summaries_dir + "/test_run_id_testbrain"
+    expected_config["model_path"] = tc.model_path + "/testbrain"
+    expected_config["keep_checkpoints"] = tc.keep_checkpoints
-    assert_bc_trainer_constructed(OnlineBCTrainer, full_config, tc, brain_info_mock, expected_config)
+    assert_bc_trainer_constructed(
+        OnlineBCTrainer, full_config, tc, brain_info_mock, expected_config
+    )
-@patch('mlagents.envs.BrainInfo')
+@patch("mlagents.envs.BrainInfo")
-    expected_config = full_config['default']
-    expected_config['summary_path'] = tc.summaries_dir + '/test_run_id_testbrain'
-    expected_config['model_path'] = tc.model_path + '/testbrain'
-    expected_config['keep_checkpoints'] = tc.keep_checkpoints
+    expected_config = full_config["default"]
+    expected_config["summary_path"] = tc.summaries_dir + "/test_run_id_testbrain"
+    expected_config["model_path"] = tc.model_path + "/testbrain"
+    expected_config["keep_checkpoints"] = tc.keep_checkpoints
-@patch('mlagents.envs.BrainInfo')
+@patch("mlagents.envs.BrainInfo")
 def test_initialize_invalid_trainer_raises_exception(BrainInfoMock):
     brain_info_mock = BrainInfoMock()
     tc = basic_trainer_controller(brain_info_mock)
         tc.initialize_trainers(bad_config)
-        assert(1 == 0, "Initialize trainers with bad config did not raise an exception.")
+        assert (
+            1 == 0,
+            "Initialize trainers with bad config did not raise an exception.",
+        )
     except UnityEnvironmentException:
         pass
 
     trainer_mock.get_step = 0
     trainer_mock.get_max_steps = 5
-    trainer_mock.parameters = {'some': 'parameter'}
+    trainer_mock.parameters = {"some": "parameter"}
-    tc.trainers = {'testbrain': trainer_mock}
+    tc.trainers = {"testbrain": trainer_mock}
-        tc.trainers['testbrain'].get_step += 1
-        if tc.trainers['testbrain'].get_step > 10:
+        tc.trainers["testbrain"].get_step += 1
+        if tc.trainers["testbrain"].get_step > 10:
+
     tc.take_step.side_effect = take_step_sideeffect
 
     tc._export_graph = MagicMock()
 
-@patch('tensorflow.reset_default_graph')
+@patch("tensorflow.reset_default_graph")
 def test_start_learning_trains_forever_if_no_train_model(tf_reset_graph):
     tc, trainer_mock = trainer_controller_with_start_learning_mocks()
     tc.train_model = False
     tf_reset_graph.assert_called_once()
     tc.initialize_trainers.assert_called_once_with(trainer_config)
     env_mock.reset.assert_called_once()
-    assert (tc.take_step.call_count == 11)
+    assert tc.take_step.call_count == 11
-@patch('tensorflow.reset_default_graph')
+@patch("tensorflow.reset_default_graph")
 def test_start_learning_trains_until_max_steps_then_saves(tf_reset_graph):
     tc, trainer_mock = trainer_controller_with_start_learning_mocks()
     trainer_config = dummy_config()
     tf_reset_graph.assert_called_once()
     tc.initialize_trainers.assert_called_once_with(trainer_config)
     env_mock.reset.assert_called_once()
-    assert(tc.take_step.call_count == trainer_mock.get_max_steps + 1)
+    assert tc.take_step.call_count == trainer_mock.get_max_steps + 1
     env_mock.close.assert_called_once()
     tc._save_model.assert_called_once_with(steps=6)
 
     tc.lesson = 5
 
     tc.start_learning(env_mock, trainer_config)
-    meta_curriculum_mock.set_all_curriculums_to_lesson_num.assert_called_once_with(tc.lesson)
+    meta_curriculum_mock.set_all_curriculums_to_lesson_num.assert_called_once_with(
+        tc.lesson
+    )
 
 
 def trainer_controller_with_take_step_mocks():
-    trainer_mock.parameters = {'some': 'parameter'}
+    trainer_mock.parameters = {"some": "parameter"}
-    tc.trainers = {'testbrain': trainer_mock}
+    tc.trainers = {"testbrain": trainer_mock}
 
     return tc, trainer_mock
 
     env_mock.reset = MagicMock(return_value=brain_info_mock)
     env_mock.global_done = True
 
-    trainer_mock.get_action = MagicMock(return_value = ActionInfo(None, None, None, None, None))
+    trainer_mock.get_action = MagicMock(
+        return_value=ActionInfo(None, None, None, None, None)
+    )
 
     tc.take_step(env_mock, brain_info_mock)
     env_mock.reset.assert_called_once()
     env_mock.global_done = False
 
     action_output_mock = ActionInfo(
-        'action',
-        'memory',
-        'actiontext',
-        'value',
-        {'some': 'output'}
+        "action", "memory", "actiontext", "value", {"some": "output"}
     )
     trainer_mock.get_action = MagicMock(return_value=action_output_mock)
 
     env_mock.step.assert_called_once_with(
-        vector_action={'testbrain': action_output_mock.action},
-        memory={'testbrain': action_output_mock.memory},
-        text_action={'testbrain': action_output_mock.text},
-        value={'testbrain': action_output_mock.value}
+        vector_action={"testbrain": action_output_mock.action},
+        memory={"testbrain": action_output_mock.memory},
+        text_action={"testbrain": action_output_mock.text},
+        value={"testbrain": action_output_mock.value},
-    trainer_mock.process_experiences.assert_called_once_with(curr_info_mock, env_step_output_mock)
+    trainer_mock.process_experiences.assert_called_once_with(
+        curr_info_mock, env_step_output_mock
+    )
     trainer_mock.update_policy.assert_called_once()
     trainer_mock.write_summary.assert_called_once()
     trainer_mock.increment_step_and_update_last_reward.assert_called_once()

ml-agents/mlagents/trainers/tests/test_trainer_metrics.py

 import unittest.mock as mock
 from mlagents.trainers import TrainerMetrics
 
-class TestTrainerMetrics:
+class TestTrainerMetrics:
-        field_names = ['Brain name', 'Time to update policy',
-                       'Time since start of training',
-                       'Time for last experience collection',
-                       'Number of experiences used for training', 'Mean return']
+        field_names = [
+            "Brain name",
+            "Time to update policy",
+            "Time since start of training",
+            "Time for last experience collection",
+            "Number of experiences used for training",
+            "Mean return",
+        ]
+
-    @mock.patch('mlagents.trainers.trainer_metrics.time', mock.MagicMock(return_value=42))
+    @mock.patch(
+        "mlagents.trainers.trainer_metrics.time", mock.MagicMock(return_value=42)
+    )
-        mock_path = 'fake'
-        mock_brain_name = 'fake'
-        trainer_metrics = TrainerMetrics(path=mock_path,
-                                         brain_name=mock_brain_name)
+        mock_path = "fake"
+        mock_brain_name = "fake"
+        trainer_metrics = TrainerMetrics(path=mock_path, brain_name=mock_brain_name)
-    @mock.patch('mlagents.trainers.trainer_metrics.time', mock.MagicMock(return_value=42))
+    @mock.patch(
+        "mlagents.trainers.trainer_metrics.time", mock.MagicMock(return_value=42)
+    )
-        mock_path = 'fake'
-        mock_brain_name = 'fake'
-        fake_buffer_length =  350
+        mock_path = "fake"
+        mock_brain_name = "fake"
+        fake_buffer_length = 350
-        trainer_metrics = TrainerMetrics(path=mock_path,
-                                         brain_name=mock_brain_name)
+        trainer_metrics = TrainerMetrics(path=mock_path, brain_name=mock_brain_name)
-        trainer_metrics.start_policy_update_timer(number_experiences=fake_buffer_length,
-                                                  mean_return=fake_mean_return)
+        trainer_metrics.start_policy_update_timer(
+            number_experiences=fake_buffer_length, mean_return=fake_mean_return
+        )
-        fake_row = [mock_brain_name, 0,0, 0, 350, '0.300']
+        fake_row = [mock_brain_name, 0, 0, 0, 350, "0.300"]
-

ml-agents/mlagents/trainers/trainer.py

 
 LOGGER = logging.getLogger("mlagents.trainers")
 
+
+
     pass
 
 
         self.brain_name = brain.brain_name
         self.run_id = run_id
         self.trainer_parameters = trainer_parameters
-        self.summary_path = trainer_parameters['summary_path']
+        self.summary_path = trainer_parameters["summary_path"]
-        self.trainer_metrics = TrainerMetrics(path=self.summary_path + '.csv',
-                                              brain_name=self.brain_name)
+        self.trainer_metrics = TrainerMetrics(
+            path=self.summary_path + ".csv", brain_name=self.brain_name
+        )
-        return '''{} Trainer'''.format(self.__class__)
+        return """{} Trainer""".format(self.__class__)
 
     def check_param_keys(self):
         for k in self.param_keys:
-                    "brain {2}.".format(k, self.__class__, self.brain_name))
+                    "brain {2}.".format(k, self.__class__, self.brain_name)
+                )
 
     @property
     def parameters(self):
-        raise UnityTrainerException(
-            "The parameters property was not implemented.")
+        raise UnityTrainerException("The parameters property was not implemented.")
 
     @property
     def graph_scope(self):
-        raise UnityTrainerException(
-            "The graph_scope property was not implemented.")
+        raise UnityTrainerException("The graph_scope property was not implemented.")
 
     @property
     def get_max_steps(self):
         """
-        raise UnityTrainerException(
-            "The get_max_steps property was not implemented.")
+        raise UnityTrainerException("The get_max_steps property was not implemented.")
 
     @property
     def get_step(self):
         """
-        raise UnityTrainerException(
-            "The get_step property was not implemented.")
+        raise UnityTrainerException("The get_step property was not implemented.")
 
     @property
     def get_last_reward(self):
         """
-        raise UnityTrainerException(
-            "The get_last_reward property was not implemented.")
+        raise UnityTrainerException("The get_last_reward property was not implemented.")
 
     def increment_step_and_update_last_reward(self):
         """
-            "The increment_step_and_update_last_reward method was not implemented.")
+            "The increment_step_and_update_last_reward method was not implemented."
+        )
 
     def get_action(self, curr_info: BrainInfo) -> ActionInfo:
         """
         self.trainer_metrics.end_experience_collection_timer()
         return action
 
-    def add_experiences(self, curr_info: AllBrainInfo, next_info: AllBrainInfo,
-                        take_action_outputs):
+    def add_experiences(
+        self, curr_info: AllBrainInfo, next_info: AllBrainInfo, take_action_outputs
+    ):
         """
         Adds experiences to each agent's experience history.
         :param curr_info: Current AllBrainInfo.
-        raise UnityTrainerException(
-            "The add_experiences method was not implemented.")
+        raise UnityTrainerException("The add_experiences method was not implemented.")
 
     def process_experiences(self, current_info: AllBrainInfo, next_info: AllBrainInfo):
         """
         :param next_info: Dictionary of all next-step brains and corresponding BrainInfo.
         """
         raise UnityTrainerException(
-            "The process_experiences method was not implemented.")
+            "The process_experiences method was not implemented."
+        )
 
     def end_episode(self):
         """
-        raise UnityTrainerException(
-            "The end_episode method was not implemented.")
+        raise UnityTrainerException("The end_episode method was not implemented.")
 
     def is_ready_update(self):
         """
-        raise UnityTrainerException(
-            "The is_ready_update method was not implemented.")
+        raise UnityTrainerException("The is_ready_update method was not implemented.")
-        raise UnityTrainerException(
-            "The update_model method was not implemented.")
+        raise UnityTrainerException("The update_model method was not implemented.")
 
     def save_model(self):
         """
         :param lesson_num: Current lesson number in curriculum.
         :param global_step: The number of steps the simulation has been going for
         """
-        if global_step % self.trainer_parameters['summary_freq'] == 0 and global_step != 0:
-            is_training = "Training." if self.is_training and self.get_step <= self.get_max_steps else "Not Training."
-            if len(self.stats['Environment/Cumulative Reward']) > 0:
-                mean_reward = np.mean(
-                    self.stats['Environment/Cumulative Reward'])
-                LOGGER.info(" {}: {}: Step: {}. "
-                            "Time Elapsed: {:0.3f} s "
-                            "Mean "
-                            "Reward: {"
-                            ":0.3f}. Std of Reward: {:0.3f}. {}"
-                            .format(self.run_id, self.brain_name,
-                                    min(self.get_step, self.get_max_steps),
-                                    delta_train_start,
-                                    mean_reward, np.std(
-                                        self.stats['Environment/Cumulative Reward']),
-                                    is_training))
+        if (
+            global_step % self.trainer_parameters["summary_freq"] == 0
+            and global_step != 0
+        ):
+            is_training = (
+                "Training."
+                if self.is_training and self.get_step <= self.get_max_steps
+                else "Not Training."
+            )
+            if len(self.stats["Environment/Cumulative Reward"]) > 0:
+                mean_reward = np.mean(self.stats["Environment/Cumulative Reward"])
+                LOGGER.info(
+                    " {}: {}: Step: {}. "
+                    "Time Elapsed: {:0.3f} s "
+                    "Mean "
+                    "Reward: {"
+                    ":0.3f}. Std of Reward: {:0.3f}. {}".format(
+                        self.run_id,
+                        self.brain_name,
+                        min(self.get_step, self.get_max_steps),
+                        delta_train_start,
+                        mean_reward,
+                        np.std(self.stats["Environment/Cumulative Reward"]),
+                        is_training,
+                    )
+                )
-                LOGGER.info(" {}: {}: Step: {}. No episode was completed since last summary. {}"
-                            .format(self.run_id, self.brain_name, self.get_step, is_training))
+                LOGGER.info(
+                    " {}: {}: Step: {}. No episode was completed since last summary. {}".format(
+                        self.run_id, self.brain_name, self.get_step, is_training
+                    )
+                )
-                    summary.value.add(tag='{}'.format(
-                        key), simple_value=stat_mean)
+                    summary.value.add(tag="{}".format(key), simple_value=stat_mean)
-            summary.value.add(tag='Environment/Lesson', simple_value=lesson_num)
+            summary.value.add(tag="Environment/Lesson", simple_value=lesson_num)
             self.summary_writer.add_summary(summary, self.get_step)
             self.summary_writer.flush()
 
         """
         try:
             with tf.Session() as sess:
-                s_op = tf.summary.text(key, tf.convert_to_tensor(
-                    ([[str(x), str(input_dict[x])] for x in input_dict])))
+                s_op = tf.summary.text(
+                    key,
+                    tf.convert_to_tensor(
+                        ([[str(x), str(input_dict[x])] for x in input_dict])
+                    ),
+                )
-                "Cannot write text summary for Tensorboard. Tensorflow version must be r1.2 or above.")
+                "Cannot write text summary for Tensorboard. Tensorflow version must be r1.2 or above."
+            )
             pass

ml-agents/mlagents/trainers/trainer_controller.py

 
 
 class TrainerController(object):
-    def __init__(self,
-                 model_path: str,
-                 summaries_dir: str,
-                 run_id: str,
-                 save_freq: int,
-                 meta_curriculum: Optional[MetaCurriculum],
-                 load: bool,
-                 train: bool,
-                 keep_checkpoints: int,
-                 lesson: Optional[int],
-                 external_brains: Dict[str, BrainParameters],
-                 training_seed: int,
-                 fast_simulation: bool):
+    def __init__(
+        self,
+        model_path: str,
+        summaries_dir: str,
+        run_id: str,
+        save_freq: int,
+        meta_curriculum: Optional[MetaCurriculum],
+        load: bool,
+        train: bool,
+        keep_checkpoints: int,
+        lesson: Optional[int],
+        external_brains: Dict[str, BrainParameters],
+        training_seed: int,
+        fast_simulation: bool,
+    ):
         """
         :param model_path: Path to save the model.
         :param summaries_dir: Folder to save training summaries.
         self.summaries_dir = summaries_dir
         self.external_brains = external_brains
         self.external_brain_names = external_brains.keys()
-        self.logger = logging.getLogger('mlagents.envs')
+        self.logger = logging.getLogger("mlagents.envs")
         self.run_id = run_id
         self.save_freq = save_freq
         self.lesson = lesson
     def _get_measure_vals(self):
         if self.meta_curriculum:
             brain_names_to_measure_vals = {}
-            for brain_name, curriculum \
-                in self.meta_curriculum.brains_to_curriculums.items():
-                if curriculum.measure == 'progress':
-                    measure_val = (self.trainers[brain_name].get_step /
-                        self.trainers[brain_name].get_max_steps)
+            for (
+                brain_name,
+                curriculum,
+            ) in self.meta_curriculum.brains_to_curriculums.items():
+                if curriculum.measure == "progress":
+                    measure_val = (
+                        self.trainers[brain_name].get_step
+                        / self.trainers[brain_name].get_max_steps
+                    )
-                elif curriculum.measure == 'reward':
-                    measure_val = np.mean(self.trainers[brain_name]
-                                          .reward_buffer)
+                elif curriculum.measure == "reward":
+                    measure_val = np.mean(self.trainers[brain_name].reward_buffer)
                     brain_names_to_measure_vals[brain_name] = measure_val
             return brain_names_to_measure_vals
         else:
         """
         for brain_name in self.trainers.keys():
             self.trainers[brain_name].save_model()
-        self.logger.info('Saved Model')
+        self.logger.info("Saved Model")
-        self.logger.info('Learning was interrupted. Please wait '
-                         'while the graph is generated.')
+        self.logger.info(
+            "Learning was interrupted. Please wait " "while the graph is generated."
+        )
         self._save_model(steps)
 
     def _write_training_metrics(self):
         """
         trainer_parameters_dict = {}
         for brain_name in self.external_brains:
-            trainer_parameters = trainer_config['default'].copy()
-            trainer_parameters['summary_path'] = '{basedir}/{name}'.format(
-                basedir=self.summaries_dir,
-                name=str(self.run_id) + '_' + brain_name)
-            trainer_parameters['model_path'] = '{basedir}/{name}'.format(
-                basedir=self.model_path,
-                name=brain_name)
-            trainer_parameters['keep_checkpoints'] = self.keep_checkpoints
+            trainer_parameters = trainer_config["default"].copy()
+            trainer_parameters["summary_path"] = "{basedir}/{name}".format(
+                basedir=self.summaries_dir, name=str(self.run_id) + "_" + brain_name
+            )
+            trainer_parameters["model_path"] = "{basedir}/{name}".format(
+                basedir=self.model_path, name=brain_name
+            )
+            trainer_parameters["keep_checkpoints"] = self.keep_checkpoints
             if brain_name in trainer_config:
                 _brain_key = brain_name
                 while not isinstance(trainer_config[_brain_key], dict):
             trainer_parameters_dict[brain_name] = trainer_parameters.copy()
         for brain_name in self.external_brains:
-            if trainer_parameters_dict[brain_name]['trainer'] == 'offline_bc':
+            if trainer_parameters_dict[brain_name]["trainer"] == "offline_bc":
-                    trainer_parameters_dict[brain_name], self.train_model,
-                    self.load_model, self.seed, self.run_id)
-            elif trainer_parameters_dict[brain_name]['trainer'] == 'online_bc':
+                    trainer_parameters_dict[brain_name],
+                    self.train_model,
+                    self.load_model,
+                    self.seed,
+                    self.run_id,
+                )
+            elif trainer_parameters_dict[brain_name]["trainer"] == "online_bc":
-                    trainer_parameters_dict[brain_name], self.train_model,
-                    self.load_model, self.seed, self.run_id)
-            elif trainer_parameters_dict[brain_name]['trainer'] == 'ppo':
+                    trainer_parameters_dict[brain_name],
+                    self.train_model,
+                    self.load_model,
+                    self.seed,
+                    self.run_id,
+                )
+            elif trainer_parameters_dict[brain_name]["trainer"] == "ppo":
-                    self.meta_curriculum
-                        .brains_to_curriculums[brain_name]
-                        .min_lesson_length if self.meta_curriculum else 0,
+                    self.meta_curriculum.brains_to_curriculums[
+                        brain_name
+                    ].min_lesson_length
+                    if self.meta_curriculum
+                    else 0,
-                    self.train_model, self.load_model, self.seed,
-                    self.run_id)
-                self.trainer_metrics[brain_name] = self.trainers[brain_name].trainer_metrics
+                    self.train_model,
+                    self.load_model,
+                    self.seed,
+                    self.run_id,
+                )
+                self.trainer_metrics[brain_name] = self.trainers[
+                    brain_name
+                ].trainer_metrics
-                raise UnityEnvironmentException('The trainer config contains '
-                                                'an unknown trainer type for '
-                                                'brain {}'
-                                                .format(brain_name))
+                raise UnityEnvironmentException(
+                    "The trainer config contains "
+                    "an unknown trainer type for "
+                    "brain {}".format(brain_name)
+                )
 
     @staticmethod
     def _create_model_path(model_path):
         except Exception:
-            raise UnityEnvironmentException('The folder {} containing the '
-                                            'generated model could not be '
-                                            'accessed. Please make sure the '
-                                            'permissions are set correctly.'
-                                            .format(model_path))
+            raise UnityEnvironmentException(
+                "The folder {} containing the "
+                "generated model could not be "
+                "accessed. Please make sure the "
+                "permissions are set correctly.".format(model_path)
+            )
 
     def _reset_env(self, env: BaseUnityEnvironment):
         """Resets the environment.
             environment.
         """
         if self.meta_curriculum is not None:
-            return env.reset(train_mode=self.fast_simulation, config=self.meta_curriculum.get_config())
+            return env.reset(
+                train_mode=self.fast_simulation,
+                config=self.meta_curriculum.get_config(),
+            )
         else:
             return env.reset(train_mode=self.fast_simulation)
 
 
         if self.train_model:
             for brain_name, trainer in self.trainers.items():
-                trainer.write_tensorboard_text('Hyperparameters',
-                                               trainer.parameters)
+                trainer.write_tensorboard_text("Hyperparameters", trainer.parameters)
-            while any([t.get_step <= t.get_max_steps \
-                       for k, t in self.trainers.items()]) \
-                  or not self.train_model:
+            while (
+                any([t.get_step <= t.get_max_steps for k, t in self.trainers.items()])
+                or not self.train_model
+            ):
-                if self.global_step % self.save_freq == 0 and self.global_step != 0 \
-                        and self.train_model:
+                if (
+                    self.global_step % self.save_freq == 0
+                    and self.global_step != 0
+                    and self.train_model
+                ):
                     # Save Tensorflow model
                     self._save_model(steps=self.global_step)
                 curr_info = new_info
     def take_step(self, env: BaseUnityEnvironment, curr_info: AllBrainInfo):
         if self.meta_curriculum:
             # Get the sizes of the reward buffers.
-            reward_buff_sizes = {k: len(t.reward_buffer)
-                                 for (k, t) in self.trainers.items()}
+            reward_buff_sizes = {
+                k: len(t.reward_buffer) for (k, t) in self.trainers.items()
+            }
-            lessons_incremented = \
-                self.meta_curriculum.increment_lessons(
-                    self._get_measure_vals(),
-                    reward_buff_sizes=reward_buff_sizes)
+            lessons_incremented = self.meta_curriculum.increment_lessons(
+                self._get_measure_vals(), reward_buff_sizes=reward_buff_sizes
+            )
-        if (self.meta_curriculum
-                and any(lessons_incremented.values())):
+        if self.meta_curriculum and any(lessons_incremented.values()):
             curr_info = self._reset_env(env)
             for brain_name, trainer in self.trainers.items():
                 trainer.end_episode()
             vector_action=take_action_vector,
             memory=take_action_memories,
             text_action=take_action_text,
-            value=take_action_value
+            value=take_action_value,
-            trainer.add_experiences(curr_info, new_info,
-                                    take_action_outputs[brain_name])
+            trainer.add_experiences(
+                curr_info, new_info, take_action_outputs[brain_name]
+            )
-            if trainer.is_ready_update() and self.train_model \
-                    and trainer.get_step <= trainer.get_max_steps:
+            if (
+                trainer.is_ready_update()
+                and self.train_model
+                and trainer.get_step <= trainer.get_max_steps
+            ):
                 # Perform gradient descent with experience buffer
 
                 trainer.update_policy()
                 trainer.write_summary(
                     self.global_step,
-                    delta_train_start, lesson_num=self.meta_curriculum
-                        .brains_to_curriculums[brain_name]
-                        .lesson_num)
+                    delta_train_start,
+                    lesson_num=self.meta_curriculum.brains_to_curriculums[
+                        brain_name
+                    ].lesson_num,
+                )
-            if self.train_model \
-                    and trainer.get_step <= trainer.get_max_steps:
+            if self.train_model and trainer.get_step <= trainer.get_max_steps:
                 trainer.increment_step_and_update_last_reward()
         return new_info

ml-agents/mlagents/trainers/trainer_metrics.py

 from time import time
 
 LOGGER = logging.getLogger("mlagents.trainers")
-FIELD_NAMES = ['Brain name', 'Time to update policy',
-               'Time since start of training', 'Time for last experience collection',
-               'Number of experiences used for training', 'Mean return']
+FIELD_NAMES = [
+    "Brain name",
+    "Time to update policy",
+    "Time since start of training",
+    "Time for last experience collection",
+    "Number of experiences used for training",
+    "Mean return",
+]
+
 
 class TrainerMetrics:
     """
+
     def __init__(self, path: str, brain_name: str):
         """
         :str path: Fully qualified path where CSV is stored.
 
     def _add_row(self, delta_train_start):
         row = [self.brain_name]
-        row.extend(format(c, '.3f') if isinstance(c, float) else c
-                   for c in [self.delta_policy_update, delta_train_start,
-                             self.delta_last_experience_collection,
-                             self.last_buffer_length, self.last_mean_return])
+        row.extend(
+            format(c, ".3f") if isinstance(c, float) else c
+            for c in [
+                self.delta_policy_update,
+                delta_train_start,
+                self.delta_last_experience_collection,
+                self.last_buffer_length,
+                self.last_mean_return,
+            ]
+        )
         self.delta_last_experience_collection = None
         self.rows.append(row)
 
         else:
             self.delta_policy_update = 0
         delta_train_start = time() - self.time_training_start
-        LOGGER.debug(" Policy Update Training Metrics for {}: "
-                     "\n\t\tTime to update Policy: {:0.3f} s \n"
-                     "\t\tTime elapsed since training: {:0.3f} s \n"
-                     "\t\tTime for experience collection: {:0.3f} s \n"
-                     "\t\tBuffer Length: {} \n"
-                     "\t\tReturns : {:0.3f}\n"
-                     .format(self.brain_name, self.delta_policy_update,
-                             delta_train_start, self.delta_last_experience_collection,
-                             self.last_buffer_length, self.last_mean_return))
+        LOGGER.debug(
+            " Policy Update Training Metrics for {}: "
+            "\n\t\tTime to update Policy: {:0.3f} s \n"
+            "\t\tTime elapsed since training: {:0.3f} s \n"
+            "\t\tTime for experience collection: {:0.3f} s \n"
+            "\t\tBuffer Length: {} \n"
+            "\t\tReturns : {:0.3f}\n".format(
+                self.brain_name,
+                self.delta_policy_update,
+                delta_train_start,
+                self.delta_last_experience_collection,
+                self.last_buffer_length,
+                self.last_mean_return,
+            )
+        )
         self._add_row(delta_train_start)
 
     def write_training_metrics(self):
-        with open(self.path, 'w') as file:
+        with open(self.path, "w") as file:
             writer = csv.writer(file)
             writer.writerow(FIELD_NAMES)
             for row in self.rows:

ml-agents/setup.py

 here = path.abspath(path.dirname(__file__))
 
 # Get the long description from the README file
-with open(path.join(here, 'README.md'), encoding='utf-8') as f:
+with open(path.join(here, "README.md"), encoding="utf-8") as f:
-    name='mlagents',
-    version='0.8.1',
-    description='Unity Machine Learning Agents',
+    name="mlagents",
+    version="0.8.1",
+    description="Unity Machine Learning Agents",
-    long_description_content_type='text/markdown',
-    url='https://github.com/Unity-Technologies/ml-agents',
-    author='Unity Technologies',
-    author_email='ML-Agents@unity3d.com',
-
+    long_description_content_type="text/markdown",
+    url="https://github.com/Unity-Technologies/ml-agents",
+    author="Unity Technologies",
+    author_email="ML-Agents@unity3d.com",
-        'Intended Audience :: Developers',
-        'Topic :: Scientific/Engineering :: Artificial Intelligence',
-        'License :: OSI Approved :: Apache Software License',
-        'Programming Language :: Python :: 3.6'
+        "Intended Audience :: Developers",
+        "Topic :: Scientific/Engineering :: Artificial Intelligence",
+        "License :: OSI Approved :: Apache Software License",
+        "Programming Language :: Python :: 3.6",
-
-    packages=['mlagents.trainers', 'mlagents.trainers.bc', 'mlagents.trainers.ppo'],  # Required
+    packages=["mlagents.trainers"],  # Required
-
-        'mlagents_envs==0.8.1',
-        'tensorflow>=1.7,<1.8',
-        'Pillow>=4.2.1',
-        'matplotlib',
-        'numpy>=1.13.3,<=1.14.5',
-        'jupyter',
-        'pytest>=3.2.2,<4.0.0',
-        'docopt',
-        'pyyaml',
-        'protobuf>=3.6,<3.7',
-        'grpcio>=1.11.0,<1.12.0',
-        'pypiwin32==223;platform_system=="Windows"'],
-
+        "mlagents_envs==0.8.1",
+        "tensorflow>=1.7,<1.8",
+        "Pillow>=4.2.1",
+        "matplotlib",
+        "numpy>=1.13.3,<=1.14.5",
+        "jupyter",
+        "pytest>=3.2.2,<4.0.0",
+        "docopt",
+        "pyyaml",
+        "protobuf>=3.6,<3.7",
+        "grpcio>=1.11.0,<1.12.0",
+        'pypiwin32==223;platform_system=="Windows"',
+    ],
-
-    entry_points={
-        'console_scripts': [
-            'mlagents-learn=mlagents.trainers.learn:main',
-        ],
-    },
+    entry_points={"console_scripts": ["mlagents-learn=mlagents.trainers.learn:main"]},
 )