Merge branch 'development' into dev-logfile

.gitignore

 /unity-environment/[Bb]uild/
 /unity-environment/[Bb]uilds/
 /unity-environment/[Pp]ackages/
+/unity-environment/[Uu]nity[Pp]ackage[Mm]anager/
 /unity-environment/Assets/AssetStoreTools*
 
 # Tensorflow Model Info

docs/Example-Environments.md

     * Observations: None
 * Reset Parameters: One, corresponding to goal size.
 
-## Spider
+## Crawler
-![Spider](../images/spider.png)
+![Crawler](../images/crawler.png)
-* Set-up: A spider with 4 arms and 4 forearms.
+* Set-up: A creature with 4 arms and 4 forearms.
 * Goal: The agents must move its body along the x axis without falling.
 * Agents: The environment contains 3 agent linked to a single brain.
 * Agent Reward Function (independent): 

python/ppo.py

 
 Options:
   --help                     Show this message.
-  --curriculum=<file>        Curriculum json file for environment [default: None]
+  --batch-size=<n>           How many experiences per gradient descent update step [default: 64].
+  --beta=<n>                 Strength of entropy regularization [default: 2e-3].
+  --buffer-size=<n>          How large the experience buffer should be before gradient descent [default: 2048].
+  --curriculum=<file>        Curriculum json file for environment [default: None].
+  --epsilon=<n>              Acceptable threshold around ratio of old and new policy probabilities [default: 0.2].
+  --gamma=<n>                Reward discount rate [default: 0.995].
+  --hidden-units=<n>         Number of units in hidden layer [default: 64].
+  --keep-checkpoints=<n>     How many model checkpoints to keep [default: 5].
+  --lambd=<n>                Lambda parameter for GAE [default: 0.95].
+  --learning-rate=<rate>     Model learning rate [default: 3e-4].
+  --load                     Whether to load the model or randomly initialize [default: False].
+  --num-epoch=<n>            Number of gradient descent steps per batch of experiences [default: 5].
-  --load                     Whether to load the model or randomly initialize [default: False].
-  --train                    Whether to train model, or only run inference [default: False].
-  --summary-freq=<n>         Frequency at which to save training statistics [default: 10000].
-  --gamma=<n>                Reward discount rate [default: 0.995].
-  --lambd=<n>                Lambda parameter for GAE [default: 0.95].
+  --summary-freq=<n>         Frequency at which to save training statistics [default: 10000].
-  --beta=<n>                 Strength of entropy regularization [default: 1e-3].
-  --num-epoch=<n>            Number of gradient descent steps per batch of experiences [default: 5].
-  --epsilon=<n>              Acceptable threshold around ratio of old and new policy probabilities [default: 0.2].
-  --buffer-size=<n>          How large the experience buffer should be before gradient descent [default: 2048].
-  --learning-rate=<rate>     Model learning rate [default: 3e-4].
-  --hidden-units=<n>         Number of units in hidden layer [default: 64].
-  --batch-size=<n>           How many experiences per gradient descent update step [default: 64].
-  --keep-checkpoints=<n>     How many model checkpoints to keep [default: 5].
-  --worker-id=<n>            Number to add to communication port (5005). Used for asynchronous agent scenarios [default: 0].
+  --train                    Whether to train model, or only run inference [default: False].
+  --worker-id=<n>            Number to add to communication port (5005). Used for multi-environment [default: 0].
 '''
 
 options = docopt(_USAGE)
     if steps != 0 and train_model:
         save_model(sess, model_path=model_path, steps=steps, saver=saver)
 env.close()
-export_graph(model_path, env_name)
+graph_name = (env_name.strip()
+      .replace('.app', '').replace('.exe', '').replace('.x86_64', '').replace('.x86', ''))
+graph_name = os.path.basename(os.path.normpath(graph_name))
+export_graph(model_path, graph_name)

python/ppo/models.py

 
 class PPOModel(object):
     def create_global_steps(self):
+        """Creates TF ops to track and increment global training step."""
+        """Creates TF ops to track and increment recent average cumulative reward."""
         self.last_reward = tf.Variable(0, name="last_reward", trainable=False, dtype=tf.float32)
         self.new_reward = tf.placeholder(shape=[], dtype=tf.float32, name='new_reward')
         self.update_reward = tf.assign(self.last_reward, self.new_reward)

python/ppo/trainer.py

 class Trainer(object):
     def __init__(self, ppo_model, sess, info, is_continuous, use_observations, use_states, training):
         """
-        Responsible for collecting experinces and training PPO model.
+        Responsible for collecting experiences and training PPO model.
         :param ppo_model: Tensorflow graph defining model.
         :param sess: Tensorflow session.
         :param info: Environment BrainInfo object.
         self.use_states = use_states
 
     def running_average(self, data, steps, running_mean, running_variance):
+        """
+        Computes new running mean and variances.
+        :param data: New piece of data.
+        :param steps: Total number of data so far.
+        :param running_mean: TF op corresponding to stored running mean.
+        :param running_variance: TF op corresponding to stored running variance.
+        :return: New mean and variance values.
+        """
         mean, var = self.sess.run([running_mean, running_variance])
         current_x = np.mean(data, axis=0)
         new_mean = mean + (current_x - mean) / (steps + 1)
                     history['episode_steps'] = 0
 
     def reset_buffers(self, brain_info=None, total=False):
+        """
+        Resets either all training buffers or local training buffers
+        :param brain_info: The BrainInfo object containing agent ids.
+        :param total: Whether to completely clear buffer.
+        """
         self.training_buffer = vectorize_history(empty_local_history({}))
         if not total:
             for key in self.history_dict:

python/test_unityagents.py

 def test_initialialization():
     with mock.patch('subprocess.Popen') as mock_subproc_popen:
         with mock.patch('socket.socket') as mock_socket:
-            mock_socket.return_value.accept.return_value = (mock_socket, 0)
-            mock_socket.recv.return_value.decode.return_value = dummy_start
-            env = UnityEnvironment(' ')
-            with pytest.raises(UnityActionException):
-                env.step([0])
-            assert env.brain_names[0] == 'RealFakeBrain'
-            env.close()
+            with mock.patch('glob.glob') as mock_glob:
+                mock_glob.return_value = ['FakeLaunchPath']
+                mock_socket.return_value.accept.return_value = (mock_socket, 0)
+                mock_socket.recv.return_value.decode.return_value = dummy_start
+                env = UnityEnvironment(' ')
+                with pytest.raises(UnityActionException):
+                    env.step([0])
+                assert env.brain_names[0] == 'RealFakeBrain'
+                env.close()
-            mock_socket.return_value.accept.return_value = (mock_socket, 0)
-            mock_socket.recv.return_value.decode.return_value = dummy_start
-            env = UnityEnvironment(' ')
-            brain = env.brains['RealFakeBrain']
-            mock_socket.recv.side_effect = dummy_reset 
-            brain_info = env.reset()
-            env.close()
-            assert not env.global_done
-            assert isinstance(brain_info, dict)
-            assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
-            assert isinstance(brain_info['RealFakeBrain'].observations, list) 
-            assert isinstance(brain_info['RealFakeBrain'].states, np.ndarray)
-            assert len(brain_info['RealFakeBrain'].observations) == brain.number_observations
-            assert brain_info['RealFakeBrain'].states.shape[0] == len(brain_info['RealFakeBrain'].agents)
-            assert brain_info['RealFakeBrain'].states.shape[1] == brain.state_space_size 
+            with mock.patch('glob.glob') as mock_glob:
+                mock_glob.return_value = ['FakeLaunchPath']
+                mock_socket.return_value.accept.return_value = (mock_socket, 0)
+                mock_socket.recv.return_value.decode.return_value = dummy_start
+                env = UnityEnvironment(' ')
+                brain = env.brains['RealFakeBrain']
+                mock_socket.recv.side_effect = dummy_reset 
+                brain_info = env.reset()
+                env.close()
+                assert not env.global_done
+                assert isinstance(brain_info, dict)
+                assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
+                assert isinstance(brain_info['RealFakeBrain'].observations, list) 
+                assert isinstance(brain_info['RealFakeBrain'].states, np.ndarray)
+                assert len(brain_info['RealFakeBrain'].observations) == brain.number_observations
+                assert brain_info['RealFakeBrain'].states.shape[0] == len(brain_info['RealFakeBrain'].agents)
+                assert brain_info['RealFakeBrain'].states.shape[1] == brain.state_space_size 
-            mock_socket.return_value.accept.return_value = (mock_socket, 0)
-            mock_socket.recv.return_value.decode.return_value = dummy_start
-            env = UnityEnvironment(' ')
-            brain = env.brains['RealFakeBrain']
-            mock_socket.recv.side_effect = dummy_reset 
-            brain_info = env.reset()
-            mock_socket.recv.side_effect = dummy_step
-            brain_info = env.step([0] * brain.action_space_size * len(brain_info['RealFakeBrain'].agents))
-            with pytest.raises(UnityActionException):
-                env.step([0])
-            brain_info = env.step([0] * brain.action_space_size * len(brain_info['RealFakeBrain'].agents))
-            with pytest.raises(UnityActionException):
-                env.step([0] * brain.action_space_size * len(brain_info['RealFakeBrain'].agents))
-            env.close()
-            assert env.global_done
-            assert isinstance(brain_info, dict)
-            assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
-            assert isinstance(brain_info['RealFakeBrain'].observations, list) 
-            assert isinstance(brain_info['RealFakeBrain'].states, np.ndarray)
-            assert len(brain_info['RealFakeBrain'].observations) == brain.number_observations
-            assert brain_info['RealFakeBrain'].states.shape[0] == len(brain_info['RealFakeBrain'].agents)
-            assert brain_info['RealFakeBrain'].states.shape[1] == brain.state_space_size
-            assert not brain_info['RealFakeBrain'].local_done[0]
-            assert brain_info['RealFakeBrain'].local_done[2]
+            with mock.patch('glob.glob') as mock_glob:
+                mock_glob.return_value = ['FakeLaunchPath']
+                mock_socket.return_value.accept.return_value = (mock_socket, 0)
+                mock_socket.recv.return_value.decode.return_value = dummy_start
+                env = UnityEnvironment(' ')
+                brain = env.brains['RealFakeBrain']
+                mock_socket.recv.side_effect = dummy_reset 
+                brain_info = env.reset()
+                mock_socket.recv.side_effect = dummy_step
+                brain_info = env.step([0] * brain.action_space_size * len(brain_info['RealFakeBrain'].agents))
+                with pytest.raises(UnityActionException):
+                    env.step([0])
+                brain_info = env.step([0] * brain.action_space_size * len(brain_info['RealFakeBrain'].agents))
+                with pytest.raises(UnityActionException):
+                    env.step([0] * brain.action_space_size * len(brain_info['RealFakeBrain'].agents))
+                env.close()
+                assert env.global_done
+                assert isinstance(brain_info, dict)
+                assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
+                assert isinstance(brain_info['RealFakeBrain'].observations, list) 
+                assert isinstance(brain_info['RealFakeBrain'].states, np.ndarray)
+                assert len(brain_info['RealFakeBrain'].observations) == brain.number_observations
+                assert brain_info['RealFakeBrain'].states.shape[0] == len(brain_info['RealFakeBrain'].agents)
+                assert brain_info['RealFakeBrain'].states.shape[1] == brain.state_space_size
+                assert not brain_info['RealFakeBrain'].local_done[0]
+                assert brain_info['RealFakeBrain'].local_done[2]
 
 
 
         with mock.patch('socket.socket') as mock_socket:
-            mock_socket.return_value.accept.return_value = (mock_socket, 0)
-            mock_socket.recv.return_value.decode.return_value = dummy_start
-            env = UnityEnvironment(' ')
-            assert env._loaded
-            env.close()
-            assert not env._loaded
-            mock_socket.close.assert_called_once()
+            with mock.patch('glob.glob') as mock_glob:
+                mock_glob.return_value = ['FakeLaunchPath']
+                mock_socket.return_value.accept.return_value = (mock_socket, 0)
+                mock_socket.recv.return_value.decode.return_value = dummy_start
+                env = UnityEnvironment(' ')
+                assert env._loaded
+                env.close()
+                assert not env._loaded
+                mock_socket.close.assert_called_once()
 
 
 

python/unityagents/curriculum.py

 
 class Curriculum(object):
     def __init__(self, location, default_reset_parameters):
+        """
+        Initializes a Curriculum object.
+        :param location: Path to JSON defining curriculum.
+        :param default_reset_parameters: Set of reset parameters for environment.
+        """
         self.lesson_number = 0
         self.lesson_length = 0
         self.measure_type = None
         self.lesson_number = max(0, min(value, self.max_lesson_number))
 
     def get_lesson(self, progress):
+        """
+        Returns reset parameters which correspond to current lesson.
+        :param progress: Measure of progress (either reward or percentage steps completed).
+        :return: Dictionary containing reset parameters.
+        """
         if self.data is None or progress is None:
             return {}
         if self.data["signal_smoothing"]:

python/unityagents/environment.py

 
 class UnityEnvironment(object):
     def __init__(self, file_name, worker_id=0,
-                 base_port=5005, curriculum = None):
+                 base_port=5005, curriculum=None):
         """
         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.
                                "or use a different worker number.".format(str(worker_id)))
 
         cwd = os.getcwd()
-        try:
-            true_filename = os.path.basename(os.path.normpath(file_name))
-            launch_string = ""
-            if platform == "linux" or platform == "linux2":
-                candidates = glob.glob(os.path.join(cwd, file_name) + '.x86_64')
-                if len(candidates) == 0:
-                    candidates = glob.glob(os.path.join(cwd, file_name) + '.x86')
-                if len(candidates) > 0:
-                    launch_string = candidates[0]
-                else:
-                    raise UnityEnvironmentException("Couldn't launch new environment. Provided filename "
-                                                    "does not match any environments in {}. ".format(cwd))
-            elif platform == 'darwin':
-                launch_string = os.path.join(cwd, file_name + '.app', 'Contents', 'MacOS', true_filename)
-            elif platform == 'win32':
-                launch_string = os.path.join(cwd, file_name + '.exe')
+        file_name = (file_name.strip()
+                     .replace('.app', '').replace('.exe', '').replace('.x86_64', '').replace('.x86', ''))
+        true_filename = os.path.basename(os.path.normpath(file_name))
+        launch_string = None
+        if platform == "linux" or platform == "linux2":
+            candidates = glob.glob(os.path.join(cwd, file_name) + '.x86_64')
+            if len(candidates) == 0:
+                candidates = glob.glob(os.path.join(cwd, file_name) + '.x86')
+            if len(candidates) == 0:
+                candidates = glob.glob(file_name + '.x86_64')
+            if len(candidates) == 0:
+                candidates = glob.glob(file_name + '.x86')
+            if len(candidates) > 0:
+                launch_string = candidates[0]
+        elif platform == 'darwin':
+            candidates = glob.glob(os.path.join(cwd, file_name + '.app', 'Contents', 'MacOS', true_filename))
+            if len(candidates) == 0:
+                candidates = glob.glob(os.path.join(file_name + '.app', 'Contents', 'MacOS', true_filename))
+            if len(candidates) > 0:
+                launch_string = candidates[0]
+        elif platform == 'win32':
+            candidates = glob.glob(os.path.join(cwd, file_name + '.exe'))
+            if len(candidates) == 0:
+                candidates = glob.glob(file_name + '.exe')
+            if len(candidates) > 0:
+                launch_string = candidates[0]
+        if launch_string is None:
+            self.close()
+            raise UnityEnvironmentException("Couldn't launch the {0} environment. "
+                                            "Provided filename does not match any environments."
+                                            .format(true_filename))
+        else:
-        except os.error:
-            self.close()
-            raise UnityEnvironmentException("Couldn't launch new environment. "
-                "Provided filename does not match any \environments in {}."
-                .format(cwd))
-        self._socket.settimeout(30) 
+        self._socket.settimeout(30)
         try:
             try:
                 self._socket.listen(1)
                 p = json.loads(p)
             except socket.timeout as e:
                 raise UnityEnvironmentException(
-                "The Unity environment took too long to respond. Make sure {} does not need user interaction to launch "
-                "and that the Academy and the external Brain(s) are attached to objects in the Scene.".format(
-                    str(file_name)))
-        
+                    "The Unity environment took too long to respond. Make sure {} does not need user interaction to "
+                    "launch and that the Academy and the external Brain(s) are attached to objects in the Scene."
+                    .format(str(file_name)))
+
             if "apiNumber" not in p:
                 self._unity_api = "API-1"
             else:
                     "The API number is not compatible between Unity and python. Python API : {0}, Unity API : "
-                    "{1}.".format(self._python_api, self._unity_api))
+                    "{1}.\nPlease go to https://github.com/Unity-Technologies/ml-agents to download the latest version "
+                    "of ML-Agents.".format(self._python_api, self._unity_api))
 
             self._data = {}
             self._global_done = None
             logger.info("\n'{}' started successfully!".format(self._academy_name))
             if (self._num_external_brains == 0):
                 logger.warning(" No External Brains found in the Unity Environment. "
-                    "You will not be able to pass actions to your agent(s).")
+                               "You will not be able to pass actions to your agent(s).")
         except UnityEnvironmentException:
             proc1.kill()
             self.close()
                                               "\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 _recv_bytes(self):
         s = self._conn.recv(self._buffer_size)
         config = self._curriculum.get_lesson(progress) if config is None else config
         if old_lesson != self._curriculum.get_lesson_number():
             logger.info("\nLesson changed. Now in Lesson {0} : \t{1}"
-                .format(self._curriculum.get_lesson_number(),
-                    ', '.join([str(x)+' -> '+str(config[x]) for x in config])))
+                        .format(self._curriculum.get_lesson_number(),
+                                ', '.join([str(x) + ' -> ' + str(config[x]) for x in config])))
-                .format(self._curriculum.get_lesson_number(),
-                    ', '.join([str(x)+' -> '+str(config[x]) for x in config])))
+                        .format(self._curriculum.get_lesson_number(),
+                                ', '.join([str(x) + ' -> ' + str(config[x]) for x in config])))
         if self._loaded:
             self._conn.send(b"RESET")
             self._conn.recv(self._buffer_size)
                 raise UnityActionException("Brain {0} has an invalid state. "
                                            "Expecting {1} {2} state but received {3}."
                                            .format(b, n_agent if self._brains[b].state_space_type == "discrete"
-                                            else str(self._brains[b].state_space_size * n_agent),
-                                            self._brains[b].state_space_type,
-                                            len(state_dict["states"])))
+                else str(self._brains[b].state_space_size * n_agent),
+                                                   self._brains[b].state_space_type,
+                                                   len(state_dict["states"])))
-            if n_agent > 0 :
-                actions =  np.array(state_dict["actions"]).reshape((n_agent, -1))
-            else :
+            if n_agent > 0:
+                actions = np.array(state_dict["actions"]).reshape((n_agent, -1))
+            else:
                 actions = np.array([])
 
             observations = []
         arr = [float(x) for x in arr]
         return arr
 
-    def step(self, action = None, memory=None, value=None):
+    def step(self, action=None, memory=None, value=None):
         """
         Provides the environment with an action, moves the environment dynamics forward accordingly, and returns
         observation, state, and reward information to the agent.
                         "and actions as values".format(self._num_brains))
                 else:
                     raise UnityActionException(
-                        "There are no external brains in the environment, " 
+                        "There are no external brains in the environment, "
-                    
+
             if isinstance(memory, (int, np.int_, float, np.float_, list, np.ndarray)):
                 if self._num_external_brains == 1:
                     memory = {self._external_brain_names[0]: memory}
                         "and memories as values".format(self._num_brains))
                 else:
                     raise UnityActionException(
-                        "There are no external brains in the environment, " 
+                        "There are no external brains in the environment, "
-                elif self._num_external_brains > 1:  
+                elif self._num_external_brains > 1:
-                        "There are no external brains in the environment, " 
+                        "There are no external brains in the environment, "
                         "step cannot take a value input")
 
             for brain_name in list(action.keys()) + list(memory.keys()) + list(value.keys()):
-                        "in the environment". format(brain_name))
+                        "in the environment".format(brain_name))
 
             for b in self._external_brain_names:
                 n_agent = len(self._data[b].agents)
                     raise UnityActionException(
                         "There was a mismatch between the provided memory and environment's expectation: "
                         "The brain {0} expected {1} memories but was given {2}"
-                        .format(b, self._brains[b].memory_space_size * n_agent, len(memory[b])))
+                            .format(b, self._brains[b].memory_space_size * n_agent, len(memory[b])))
                 if not ((self._brains[b].action_space_type == "discrete" and len(action[b]) == n_agent) or
                             (self._brains[b].action_space_type == "continuous" and len(
                                 action[b]) == self._brains[b].action_space_size * n_agent)):
-                        .format(b, n_agent if self._brains[b].action_space_type == "discrete" else
+                            .format(b, n_agent if self._brains[b].action_space_type == "discrete" else
-                        str(action[b])))
+                                    str(action[b])))
             self._conn.send(b"STEP")
             self._send_action(action, memory, value)
             return self._get_state()
             self._socket.close()
             self._loaded = False
         else:
-            raise UnityEnvironmentException("No Unity environment is loaded.")
+            raise UnityEnvironmentException("No Unity environment is loaded.")

unity-environment/Assets/ML-Agents/Examples/3DBall/Prefabs/Game.prefab

   - component: {fileID: 65551894134645910}
   - component: {fileID: 23487775825466554}
   - component: {fileID: 114980646877373948}
-  - component: {fileID: 114290313258162170}
   m_Layer: 0
   m_Name: Platform
   m_TagString: Untagged
   m_Enabled: 1
   m_CastShadows: 0
   m_ReceiveShadows: 0
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   serializedVersion: 2
   m_Size: {x: 1, y: 1, z: 1}
   m_Center: {x: 0, y: 0, z: 0}
---- !u!114 &114290313258162170
-MonoBehaviour:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  m_GameObject: {fileID: 1914042422505674}
-  m_Enabled: 0
-  m_EditorHideFlags: 0
-  m_Script: {fileID: 11500000, guid: e040eaa8759024abbbb14994dc4c55ee, type: 3}
-  m_Name: 
-  m_EditorClassIdentifier: 
-  fixedPosition: 1
-  verticalOffset: 10
-  DisplayBrainName: 1
-  DisplayBrainType: 1
-  DisplayFrameCount: 1
-  DisplayCurrentReward: 0
-  DisplayMaxReward: 1
-  DisplayState: 0
-  DisplayAction: 0
 --- !u!114 &114980646877373948
 MonoBehaviour:
   m_ObjectHideFlags: 1
   reward: 0
   done: 0
   value: 0
-  CummulativeReward: 0
+  CumulativeReward: 0
   stepCounter: 0
   agentStoredAction: []
   memory: []

unity-environment/Assets/ML-Agents/Examples/3DBall/Scene.unity

     m_PVRDirectSampleCount: 32
     m_PVRSampleCount: 500
     m_PVRBounces: 2
-    m_PVRFiltering: 0
+    m_PVRFilterTypeDirect: 0
+    m_PVRFilterTypeIndirect: 0
+    m_PVRFilterTypeAO: 0
-    m_PVRFilteringAtrousColorSigma: 1
-    m_PVRFilteringAtrousNormalSigma: 1
-    m_PVRFilteringAtrousPositionSigma: 1
+    m_PVRFilteringAtrousPositionSigmaDirect: 0.5
+    m_PVRFilteringAtrousPositionSigmaIndirect: 2
+    m_PVRFilteringAtrousPositionSigmaAO: 1
   m_LightingDataAsset: {fileID: 0}
   m_UseShadowmask: 0
 --- !u!196 &4
     manualTileSize: 0
     tileSize: 256
     accuratePlacement: 0
+    debug:
+      m_Flags: 0
---- !u!114 &83032124
+--- !u!114 &1195891
 MonoBehaviour:
   m_ObjectHideFlags: 0
   m_PrefabParentObject: {fileID: 0}
   m_EditorHideFlags: 0
-  m_Script: {fileID: 11500000, guid: 35813a1be64e144f887d7d5f15b963fa, type: 3}
-  m_Name: (Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)
+  m_Script: {fileID: 11500000, guid: 8b23992c8eb17439887f5e944bf04a40, type: 3}
+  m_Name: (Clone)(Clone)(Clone)(Clone)
+  broadcast: 1
+  graphModel: {fileID: 4900000, guid: e28cc81d8dc98464b952e295ae9850fc, type: 3}
+  graphScope: 
+  graphPlaceholders:
+  - name: epsilon
+    valueType: 1
+    minValue: 0
+    maxValue: 0
+  BatchSizePlaceholderName: batch_size
+  StatePlacholderName: state
+  RecurrentInPlaceholderName: recurrent_in
+  RecurrentOutPlaceholderName: recurrent_out
+  ObservationPlaceholderName: []
+  ActionPlaceholderName: action
   brain: {fileID: 667765197}
 --- !u!1001 &119733639
 Prefab:
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
     stateSpaceType: 1
   brainType: 3
   CoreBrains:
-  - {fileID: 1939754508}
-  - {fileID: 1376879964}
-  - {fileID: 83032124}
-  - {fileID: 1950067154}
-  instanceID: 10298
+  - {fileID: 1235501299}
+  - {fileID: 878319284}
+  - {fileID: 859680324}
+  - {fileID: 1195891}
+  instanceID: 19744
 --- !u!1001 &764818074
 Prefab:
   m_ObjectHideFlags: 0
     m_RemovedComponents: []
   m_ParentPrefab: {fileID: 100100000, guid: ff026d63a00abdc48ad6ddcff89aba04, type: 2}
   m_IsPrefabParent: 0
+--- !u!114 &859680324
+MonoBehaviour:
+  m_ObjectHideFlags: 0
+  m_PrefabParentObject: {fileID: 0}
+  m_PrefabInternal: {fileID: 0}
+  m_GameObject: {fileID: 0}
+  m_Enabled: 1
+  m_EditorHideFlags: 0
+  m_Script: {fileID: 11500000, guid: 35813a1be64e144f887d7d5f15b963fa, type: 3}
+  m_Name: (Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)
+  m_EditorClassIdentifier: 
+  brain: {fileID: 667765197}
+--- !u!114 &878319284
+MonoBehaviour:
+  m_ObjectHideFlags: 0
+  m_PrefabParentObject: {fileID: 0}
+  m_PrefabInternal: {fileID: 0}
+  m_GameObject: {fileID: 0}
+  m_Enabled: 1
+  m_EditorHideFlags: 0
+  m_Script: {fileID: 11500000, guid: 943466ab374444748a364f9d6c3e2fe2, type: 3}
+  m_Name: (Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)
+  m_EditorClassIdentifier: 
+  broadcast: 1
+  brain: {fileID: 0}
+--- !u!114 &1235501299
+MonoBehaviour:
+  m_ObjectHideFlags: 0
+  m_PrefabParentObject: {fileID: 0}
+  m_PrefabInternal: {fileID: 0}
+  m_GameObject: {fileID: 0}
+  m_Enabled: 1
+  m_EditorHideFlags: 0
+  m_Script: {fileID: 11500000, guid: 41e9bda8f3cf1492fa74926a530f6f70, type: 3}
+  m_Name: (Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)
+  m_EditorClassIdentifier: 
+  broadcast: 1
+  continuousPlayerActions:
+  - key: 276
+    index: 0
+    value: 1
+  - key: 275
+    index: 0
+    value: -1
+  - key: 273
+    index: 1
+    value: 1
+  - key: 274
+    index: 1
+    value: -1
+  discretePlayerActions: []
+  defaultAction: -1
+  brain: {fileID: 667765197}
 --- !u!1001 &1318922267
 Prefab:
   m_ObjectHideFlags: 0
     m_RemovedComponents: []
   m_ParentPrefab: {fileID: 100100000, guid: ff026d63a00abdc48ad6ddcff89aba04, type: 2}
   m_IsPrefabParent: 0
---- !u!114 &1376879964
-MonoBehaviour:
-  m_ObjectHideFlags: 0
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 0}
-  m_GameObject: {fileID: 0}
-  m_Enabled: 1
-  m_EditorHideFlags: 0
-  m_Script: {fileID: 11500000, guid: 943466ab374444748a364f9d6c3e2fe2, type: 3}
-  m_Name: (Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)
-  m_EditorClassIdentifier: 
-  brain: {fileID: 0}
 --- !u!1 &1397918840
 GameObject:
   m_ObjectHideFlags: 0
   m_OcclusionCulling: 1
   m_StereoConvergence: 10
   m_StereoSeparation: 0.022
-  m_StereoMirrorMode: 0
 --- !u!4 &1397918845
 Transform:
   m_ObjectHideFlags: 0
   maxSteps: 0
   frameToSkip: 4
   waitTime: 0
+  isInference: 0
   trainingConfiguration:
     width: 128
     height: 72
   done: 0
   episodeCount: 0
   currentStep: 0
-  isInference: 0
-  windowResize: 1
 --- !u!1 &1746325439
 GameObject:
   m_ObjectHideFlags: 0
     m_RemovedComponents: []
   m_ParentPrefab: {fileID: 100100000, guid: ff026d63a00abdc48ad6ddcff89aba04, type: 2}
   m_IsPrefabParent: 0
---- !u!114 &1939754508
-MonoBehaviour:
-  m_ObjectHideFlags: 0
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 0}
-  m_GameObject: {fileID: 0}
-  m_Enabled: 1
-  m_EditorHideFlags: 0
-  m_Script: {fileID: 11500000, guid: 41e9bda8f3cf1492fa74926a530f6f70, type: 3}
-  m_Name: (Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)(Clone)
-  m_EditorClassIdentifier: 
-  continuousPlayerActions:
-  - key: 276
-    index: 0
-    value: 1
-  - key: 275
-    index: 0
-    value: -1
-  - key: 273
-    index: 1
-    value: 1
-  - key: 274
-    index: 1
-    value: -1
-  discretePlayerActions: []
-  defaultAction: -1
-  brain: {fileID: 667765197}
---- !u!114 &1950067154
-MonoBehaviour:
-  m_ObjectHideFlags: 0
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 0}
-  m_GameObject: {fileID: 0}
-  m_Enabled: 1
-  m_EditorHideFlags: 0
-  m_Script: {fileID: 11500000, guid: 8b23992c8eb17439887f5e944bf04a40, type: 3}
-  m_Name: (Clone)(Clone)(Clone)
-  m_EditorClassIdentifier: 
-  graphModel: {fileID: 4900000, guid: e28cc81d8dc98464b952e295ae9850fc, type: 3}
-  graphScope: 
-  graphPlaceholders:
-  - name: epsilon
-    valueType: 1
-    minValue: 0
-    maxValue: 0
-  BatchSizePlaceholderName: batch_size
-  StatePlacholderName: state
-  RecurrentInPlaceholderName: recurrent_in
-  RecurrentOutPlaceholderName: recurrent_out
-  ObservationPlaceholderName: []
-  ActionPlaceholderName: action
-  brain: {fileID: 667765197}
 --- !u!1 &2063801192
 GameObject:
   m_ObjectHideFlags: 0

unity-environment/Assets/ML-Agents/Examples/Area/Prefabs/PushArea.prefab

   m_NavMeshLayer: 0
   m_StaticEditorFlags: 0
   m_IsActive: 1
---- !u!1 &1433578436929790
-GameObject:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  serializedVersion: 5
-  m_Component:
-  - component: {fileID: 4379575143948170}
-  - component: {fileID: 20019178239241812}
-  m_Layer: 0
-  m_Name: AgentCam
-  m_TagString: MainCamera
-  m_Icon: {fileID: 0}
-  m_NavMeshLayer: 0
-  m_StaticEditorFlags: 1
-  m_IsActive: 0
 --- !u!1 &1493483966197088
 GameObject:
   m_ObjectHideFlags: 0
   - component: {fileID: 65667507680723650}
   - component: {fileID: 23372235118339718}
   - component: {fileID: 54979596905483200}
-  - component: {fileID: 114772202904937800}
   - component: {fileID: 114371909629876940}
   m_Layer: 0
   m_Name: Agent
   m_Father: {fileID: 4797372522419518}
   m_RootOrder: 1
   m_LocalEulerAnglesHint: {x: 0, y: 0, z: 0}
---- !u!4 &4379575143948170
-Transform:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  m_GameObject: {fileID: 1433578436929790}
-  m_LocalRotation: {x: 0.38268343, y: -0, z: -0, w: 0.92387956}
-  m_LocalPosition: {x: 0, y: 9, z: -7}
-  m_LocalScale: {x: 1, y: 1, z: 1}
-  m_Children: []
-  m_Father: {fileID: 4777271908265906}
-  m_RootOrder: 0
-  m_LocalEulerAnglesHint: {x: 45, y: 0, z: 0}
 --- !u!4 &4466531474194842
 Transform:
   m_ObjectHideFlags: 1
   m_LocalRotation: {x: -0, y: -0, z: -0, w: 1}
   m_LocalPosition: {x: 0, y: 1, z: -9}
   m_LocalScale: {x: 1, y: 1, z: 1}
-  m_Children:
-  - {fileID: 4379575143948170}
+  m_Children: []
   m_Father: {fileID: 4797372522419518}
   m_RootOrder: 0
   m_LocalEulerAnglesHint: {x: 0, y: 0, z: 0}
   m_Father: {fileID: 0}
   m_RootOrder: 0
   m_LocalEulerAnglesHint: {x: 0, y: 0, z: 0}
---- !u!20 &20019178239241812
-Camera:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  m_GameObject: {fileID: 1433578436929790}
-  m_Enabled: 1
-  serializedVersion: 2
-  m_ClearFlags: 2
-  m_BackGroundColor: {r: 0, g: 0, b: 0, a: 0}
-  m_NormalizedViewPortRect:
-    serializedVersion: 2
-    x: 0
-    y: 0
-    width: 1
-    height: 1
-  near clip plane: 0.3
-  far clip plane: 1000
-  field of view: 60
-  orthographic: 1
-  orthographic size: 5
-  m_Depth: 0
-  m_CullingMask:
-    serializedVersion: 2
-    m_Bits: 4294967295
-  m_RenderingPath: -1
-  m_TargetTexture: {fileID: 0}
-  m_TargetDisplay: 0
-  m_TargetEye: 3
-  m_HDR: 1
-  m_AllowMSAA: 1
-  m_ForceIntoRT: 0
-  m_OcclusionCulling: 1
-  m_StereoConvergence: 10
-  m_StereoSeparation: 0.022
-  m_StereoMirrorMode: 0
 --- !u!23 &23361251118430556
 MeshRenderer:
   m_ObjectHideFlags: 1
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   reward: 0
   done: 0
   value: 0
-  CummulativeReward: 0
+  CumulativeReward: 0
   stepCounter: 0
   agentStoredAction: []
   memory: []
   block: {fileID: 1493483966197088}
---- !u!114 &114772202904937800
-MonoBehaviour:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  m_GameObject: {fileID: 1893190605547158}
-  m_Enabled: 0
-  m_EditorHideFlags: 0
-  m_Script: {fileID: 11500000, guid: e040eaa8759024abbbb14994dc4c55ee, type: 3}
-  m_Name: 
-  m_EditorClassIdentifier: 
-  fixedPosition: 0
-  verticalOffset: 2
-  DisplayBrainName: 0
-  DisplayBrainType: 0
-  DisplayFrameCount: 0
-  DisplayCurrentReward: 0
-  DisplayMaxReward: 0
-  DisplayState: 0
-  DisplayAction: 0
 --- !u!114 &114824876818673492
 MonoBehaviour:
   m_ObjectHideFlags: 1

unity-environment/Assets/ML-Agents/Examples/Area/Prefabs/WallArea.prefab

   m_NavMeshLayer: 0
   m_StaticEditorFlags: 0
   m_IsActive: 1
---- !u!1 &1353235869903194
-GameObject:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  serializedVersion: 5
-  m_Component:
-  - component: {fileID: 4812106314987956}
-  - component: {fileID: 20444604189257518}
-  m_Layer: 0
-  m_Name: AgentCam
-  m_TagString: MainCamera
-  m_Icon: {fileID: 0}
-  m_NavMeshLayer: 0
-  m_StaticEditorFlags: 0
-  m_IsActive: 0
 --- !u!1 &1506931635139370
 GameObject:
   m_ObjectHideFlags: 0
   - component: {fileID: 23613202787206598}
   - component: {fileID: 54973601004961626}
   - component: {fileID: 114200736018184412}
-  - component: {fileID: 114280768661186092}
   m_Layer: 0
   m_Name: Agent
   m_TagString: Untagged
   m_LocalRotation: {x: -0, y: -0, z: -0, w: 1}
   m_LocalPosition: {x: 0, y: 1, z: -9}
   m_LocalScale: {x: 1, y: 1, z: 1}
-  m_Children:
-  - {fileID: 4812106314987956}
+  m_Children: []
   m_Father: {fileID: 4601298475788644}
   m_RootOrder: 0
   m_LocalEulerAnglesHint: {x: 0, y: 0, z: 0}
   m_PrefabInternal: {fileID: 100100000}
   m_GameObject: {fileID: 1718829686919056}
   m_LocalRotation: {x: -0, y: -0, z: -0, w: 1}
-  m_LocalPosition: {x: -15, y: 0, z: 0}
+  m_LocalPosition: {x: -10, y: 0, z: 0}
   m_LocalScale: {x: 1, y: 1, z: 1}
   m_Children:
   - {fileID: 4182184265281318}
   m_PrefabInternal: {fileID: 100100000}
   m_GameObject: {fileID: 1960504220667788}
   m_LocalRotation: {x: -0, y: -0, z: -0, w: 1}
-  m_LocalPosition: {x: 0, y: 0.5, z: 0}
+  m_LocalPosition: {x: 0, y: 0.25, z: 0}
   m_LocalScale: {x: 2.2, y: 1.1, z: 2.2}
   m_Children:
   - {fileID: 4743003964451344}
   m_Father: {fileID: 4291466584800356}
   m_RootOrder: 0
   m_LocalEulerAnglesHint: {x: 0, y: 0, z: 0}
---- !u!4 &4812106314987956
-Transform:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  m_GameObject: {fileID: 1353235869903194}
-  m_LocalRotation: {x: 0.38268343, y: -0, z: -0, w: 0.92387956}
-  m_LocalPosition: {x: 0, y: 9, z: -7}
-  m_LocalScale: {x: 1, y: 1, z: 1}
-  m_Children: []
-  m_Father: {fileID: 4182184265281318}
-  m_RootOrder: 0
-  m_LocalEulerAnglesHint: {x: 45, y: 0, z: 0}
 --- !u!4 &4826682057551308
 Transform:
   m_ObjectHideFlags: 1
   m_Father: {fileID: 4601298475788644}
   m_RootOrder: 4
   m_LocalEulerAnglesHint: {x: 0, y: 0, z: 0}
---- !u!20 &20444604189257518
-Camera:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  m_GameObject: {fileID: 1353235869903194}
-  m_Enabled: 1
-  serializedVersion: 2
-  m_ClearFlags: 2
-  m_BackGroundColor: {r: 0, g: 0, b: 0, a: 0}
-  m_NormalizedViewPortRect:
-    serializedVersion: 2
-    x: 0
-    y: 0
-    width: 1
-    height: 1
-  near clip plane: 0.3
-  far clip plane: 1000
-  field of view: 60
-  orthographic: 1
-  orthographic size: 5
-  m_Depth: 0
-  m_CullingMask:
-    serializedVersion: 2
-    m_Bits: 4294967295
-  m_RenderingPath: -1
-  m_TargetTexture: {fileID: 0}
-  m_TargetDisplay: 0
-  m_TargetEye: 3
-  m_HDR: 1
-  m_AllowMSAA: 1
-  m_ForceIntoRT: 0
-  m_OcclusionCulling: 1
-  m_StereoConvergence: 10
-  m_StereoSeparation: 0.022
-  m_StereoMirrorMode: 0
 --- !u!23 &23415401456686744
 MeshRenderer:
   m_ObjectHideFlags: 1
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   m_Enabled: 1
   m_CastShadows: 1
   m_ReceiveShadows: 1
+  m_DynamicOccludee: 1
   m_MotionVectors: 1
   m_LightProbeUsage: 1
   m_ReflectionProbeUsage: 1
   m_PreserveUVs: 1
   m_IgnoreNormalsForChartDetection: 0
   m_ImportantGI: 0
+  m_StitchLightmapSeams: 0
   m_SelectedEditorRenderState: 3
   m_MinimumChartSize: 4
   m_AutoUVMaxDistance: 0.5
   reward: 0
   done: 0
   value: 0
-  CummulativeReward: 0
+  CumulativeReward: 0
+  area: {fileID: 1718829686919056}
-  area: {fileID: 1718829686919056}
   wall: {fileID: 1506931635139370}
 --- !u!114 &114215545026793406
 MonoBehaviour:
   academy: {fileID: 0}
   block: {fileID: 1312874949203312}
   goalHolder: {fileID: 1960504220667788}
---- !u!114 &114280768661186092
-MonoBehaviour:
-  m_ObjectHideFlags: 1
-  m_PrefabParentObject: {fileID: 0}
-  m_PrefabInternal: {fileID: 100100000}
-  m_GameObject: {fileID: 1986507356043702}
-  m_Enabled: 0
-  m_EditorHideFlags: 0
-  m_Script: {fileID: 11500000, guid: e040eaa8759024abbbb14994dc4c55ee, type: 3}
-  m_Name: 
-  m_EditorClassIdentifier: 
-  fixedPosition: 0
-  verticalOffset: 2
-  DisplayBrainName: 0
-  DisplayBrainType: 0
-  DisplayFrameCount: 0
-  DisplayCurrentReward: 0
-  DisplayMaxReward: 0
-  DisplayState: 0
-  DisplayAction: 0

unity-environment/Assets/ML-Agents/Examples/Area/Scripts/AreaAgent.cs

 		state.Add(velocity.x);
 		state.Add(velocity.y);
 		state.Add(velocity.z);
+		return state;
+	}
+    public void MoveAgent(int movement) {
+        float directionX = 0;
+        float directionZ = 0;
+        float directionY = 0;
+        if (movement == 1) { directionX = -1; }
+        if (movement == 2) { directionX = 1; }
+        if (movement == 3) { directionZ = -1; }
+        if (movement == 4) { directionZ = 1; }
+        if (movement == 5 && GetComponent<Rigidbody>().velocity.y <= 0) { directionY = 1; }
-		return state;
-	}
+        float edge = 0.499f;
+        float rayDepth = 0.51f;
+
+        Vector3 fwd = transform.TransformDirection(Vector3.down);
+        if (!Physics.Raycast(transform.position, fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(edge, 0f, 0f), fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(-edge, 0f, 0f), fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(0.0f, 0f, edge), fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(0.0f, 0f, -edge), fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(edge, 0f, edge), fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(-edge, 0f, edge), fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(edge, 0f, -edge), fwd, rayDepth) &&
+            !Physics.Raycast(transform.position + new Vector3(-edge, 0f, -edge), fwd, rayDepth))
+        {
+            directionY = 0f;
+            directionX = directionX / 5f;
+            directionZ = directionZ / 5f;
+        }
+
+        gameObject.GetComponent<Rigidbody>().AddForce(new Vector3(directionX * 40f, directionY * 300f, directionZ * 40f));
+        if (GetComponent<Rigidbody>().velocity.sqrMagnitude > 25f)
+        {
+            GetComponent<Rigidbody>().velocity *= 0.95f;
+        }
+    }
-		float movement = act[0];
-		float directionX = 0;
-		float directionZ = 0;
-		float directionY = 0;
-		if (movement == 1) { directionX = -1; }
-		if (movement == 2) { directionX = 1; }
-		if (movement == 3) { directionZ = -1; }
-		if (movement == 4) { directionZ = 1; }
-		if (movement == 5 && GetComponent<Rigidbody>().velocity.y <= 0) { directionY = 1; }
+        int movement = Mathf.FloorToInt(act[0]);
+        MoveAgent(movement);
-
-		Vector3 fwd = transform.TransformDirection(Vector3.down);
-		if (!Physics.Raycast(transform.position, fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(0.499f, 0f, 0f), fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(-0.499f, 0f, 0f), fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(0.0f, 0f, 0.499f), fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(0.0f, 0f, -0.499f), fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(0.499f, 0f, 0.499f), fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(-0.499f, 0f, 0.499f), fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(0.499f, 0f, -0.499f), fwd, 0.51f) &&
-			!Physics.Raycast(transform.position + new Vector3(-0.499f, 0f, -0.499f), fwd, 0.51f))
-		{
-			directionY = 0f;
-			directionX = directionX / 5f;
-			directionZ = directionZ / 5f;
-		}
-
-		gameObject.GetComponent<Rigidbody>().AddForce(new Vector3(directionX * 40f, directionY * 300f, directionZ * 40f));
-		if (GetComponent<Rigidbody>().velocity.sqrMagnitude > 25f)
-		{
-			GetComponent<Rigidbody>().velocity *= 0.95f;
-		}
-
-		if (gameObject.transform.position.y < 0.0f || Mathf.Abs(gameObject.transform.position.x - area.transform.position.x) > 8f || Mathf.Abs(gameObject.transform.position.z + 5 - area.transform.position.z) > 8)
+		if (gameObject.transform.position.y < 0.0f || Mathf.Abs(gameObject.transform.position.x - area.transform.position.x) > 8f || 
+            Mathf.Abs(gameObject.transform.position.z + 5 - area.transform.position.z) > 8)
 		{
 			done = true;
 			reward = -1f;

unity-environment/Assets/ML-Agents/Examples/Area/Scripts/Push/PushAgent.cs

 	public override void AgentStep(float[] act)
 	{
         reward = -0.01f;
-		float movement = act[0];
-		float directionX = 0;
-		float directionZ = 0;
-		float directionY = 0;
-		if (movement == 1) { directionX = -1; }
-		if (movement == 2) { directionX = 1; }
-		if (movement == 3) { directionZ