Merge branch 'master' into develop-agentprocessor-teammanager

.github/workflows/pytest.yml

       run: python -c "import sys; print(sys.version)"
     - name: Install dependencies
       run: |
-        # pin pip to workaround https://github.com/pypa/pip/issues/9180
-        python -m pip install pip==20.2
+        python -m pip install --upgrade pip
         python -m pip install --upgrade setuptools
         python -m pip install --progress-bar=off -e ./ml-agents-envs
         python -m pip install --progress-bar=off -e ./ml-agents

.pre-commit-config.yaml

         args: [--py3-plus, --py36-plus]
         exclude: >
             (?x)^(
-                .*barracuda.py|
                 .*_pb2.py|
                 .*_pb2_grpc.py
             )$
         args: [--assume-in-merge]
     -   id: check-yaml
         # Won't handle the templating in yamato
-        exclude: \.yamato/*
+        exclude: \.yamato/.*
 
 -   repo: https://github.com/pre-commit/pygrep-hooks
     rev: v1.4.2

.yamato/com.unity.ml-agents-test.yml

     assembly: Unity.ML-Agents
     minCoveragePct: 72
   - name: com.unity.ml-agents.extensions
-    assembly: Unity.ML-Agents.Extensions
+    assembly: Unity.ML-Agents.Extensions*
     minCoveragePct: 75
 ---
 

DevProject/Packages/manifest.json

   "dependencies": {
     "com.unity.2d.sprite": "1.0.0",
     "com.unity.2d.tilemap": "1.0.0",
-    "com.unity.ads": "3.4.9",
+    "com.unity.ads": "3.6.1",
-    "com.unity.ide.vscode": "1.2.1",
+    "com.unity.ide.vscode": "1.2.3",
+    "com.unity.inputsystem": "1.1.0-preview.3",
-    "com.unity.multiplayer-hlapi": "1.0.6",
+    "com.unity.multiplayer-hlapi": "1.0.8",
-    "com.unity.purchasing": "2.1.0",
-    "com.unity.test-framework": "1.1.16",
+    "com.unity.purchasing": "2.2.1",
+    "com.unity.test-framework": "1.1.20",
-    "com.unity.xr.legacyinputhelpers": "2.1.4",
+    "com.unity.xr.legacyinputhelpers": "2.1.7",
     "com.unity.modules.ai": "1.0.0",
     "com.unity.modules.androidjni": "1.0.0",
     "com.unity.modules.animation": "1.0.0",
   "registry": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates",
   "testables": [
     "com.unity.ml-agents",
-    "com.unity.ml-agents.extensions"
+    "com.unity.ml-agents.extensions",
+    "com.unity.inputsystem"
   ]
 }

DevProject/ProjectSettings/ProjectSettings.asset

   androidBlitType: 0
   defaultIsNativeResolution: 1
   macRetinaSupport: 1
-  runInBackground: 0
+  runInBackground: 1
   captureSingleScreen: 0
   muteOtherAudioSources: 0
   Prepare IOS For Recording: 0
   xboxOneMonoLoggingLevel: 0
   xboxOneLoggingLevel: 1
   xboxOneDisableEsram: 0
+  xboxOneEnableTypeOptimization: 0
   xboxOnePresentImmediateThreshold: 0
   switchQueueCommandMemory: 1048576
   switchQueueControlMemory: 16384
   switchNVNOtherPoolsGranularity: 16777216
+  switchNVNMaxPublicTextureIDCount: 0
+  switchNVNMaxPublicSamplerIDCount: 0
+  stadiaPresentMode: 0
+  stadiaTargetFramerate: 0
+  vulkanEnableLateAcquireNextImage: 0
   m_SupportedAspectRatios:
     4:3: 1
     5:4: 1
   useHDRDisplay: 0
   D3DHDRBitDepth: 0
   m_ColorGamuts: 00000000
-  targetPixelDensity: 0
+  targetPixelDensity: 30
   resolutionScalingMode: 0
   androidSupportedAspectRatio: 1
   androidMaxAspectRatio: 2.1
   StripUnusedMeshComponents: 0
   VertexChannelCompressionMask: 4054
   iPhoneSdkVersion: 988
-  iOSTargetOSVersionString: 
+  iOSTargetOSVersionString: 10.0
-  tvOSTargetOSVersionString: 
+  tvOSTargetOSVersionString: 10.0
   uIPrerenderedIcon: 0
   uIRequiresPersistentWiFi: 0
   uIRequiresFullScreen: 1
-  iPhoneSplashScreen: {fileID: 0}
-  iPhoneHighResSplashScreen: {fileID: 0}
-  iPhoneTallHighResSplashScreen: {fileID: 0}
-  iPhone47inSplashScreen: {fileID: 0}
-  iPhone55inPortraitSplashScreen: {fileID: 0}
-  iPhone55inLandscapeSplashScreen: {fileID: 0}
-  iPhone58inPortraitSplashScreen: {fileID: 0}
-  iPhone58inLandscapeSplashScreen: {fileID: 0}
-  iPadPortraitSplashScreen: {fileID: 0}
-  iPadHighResPortraitSplashScreen: {fileID: 0}
-  iPadLandscapeSplashScreen: {fileID: 0}
-  iPadHighResLandscapeSplashScreen: {fileID: 0}
-  iPhone65inPortraitSplashScreen: {fileID: 0}
-  iPhone65inLandscapeSplashScreen: {fileID: 0}
-  iPhone61inPortraitSplashScreen: {fileID: 0}
-  iPhone61inLandscapeSplashScreen: {fileID: 0}
   appleTVSplashScreen: {fileID: 0}
   appleTVSplashScreen2x: {fileID: 0}
   tvOSSmallIconLayers: []
   metalEditorSupport: 1
   metalAPIValidation: 1
   iOSRenderExtraFrameOnPause: 0
+  iosCopyPluginsCodeInsteadOfSymlink: 0
   appleDeveloperTeamID: 
   iOSManualSigningProvisioningProfileID: 
   tvOSManualSigningProvisioningProfileID: 
   ps4ShareFilePath: 
   ps4ShareOverlayImagePath: 
   ps4PrivacyGuardImagePath: 
+  ps4ExtraSceSysFile: 
   ps4NPtitleDatPath: 
   ps4RemotePlayKeyAssignment: -1
   ps4RemotePlayKeyMappingDir: 
   ps4UseResolutionFallback: 0
   ps4ReprojectionSupport: 0
   ps4UseAudio3dBackend: 0
+  ps4UseLowGarlicFragmentationMode: 1
   ps4SocialScreenEnabled: 0
   ps4ScriptOptimizationLevel: 2
   ps4Audio3dVirtualSpeakerCount: 14
   ps4disableAutoHideSplash: 0
   ps4videoRecordingFeaturesUsed: 0
   ps4contentSearchFeaturesUsed: 0
+  ps4CompatibilityPS5: 0
+  ps4GPU800MHz: 1
   ps4attribEyeToEyeDistanceSettingVR: 0
   ps4IncludedModules: []
   ps4attribVROutputEnabled: 0
   additionalIl2CppArgs: 
   scriptingRuntimeVersion: 1
   gcIncremental: 0
+  assemblyVersionValidation: 1
   gcWBarrierValidation: 0
   apiCompatibilityLevelPerPlatform: {}
   m_RenderingPath: 1
   XboxOneCapability: []
   XboxOneGameRating: {}
   XboxOneIsContentPackage: 0
+  XboxOneEnhancedXboxCompatibilityMode: 0
   XboxOneEnableGPUVariability: 1
   XboxOneSockets: {}
   XboxOneSplashScreen: {fileID: 0}
   XboxOneOverrideIdentityName: 
+  XboxOneOverrideIdentityPublisher: 
   vrEditorSettings:
     daydream:
       daydreamIconForeground: {fileID: 0}
   projectName: 
   organizationId: 
   cloudEnabled: 0
-  enableNativePlatformBackendsForNewInputSystem: 0
-  disableOldInputManagerSupport: 0
+  enableNativePlatformBackendsForNewInputSystem: 1
+  disableOldInputManagerSupport: 1
   legacyClampBlendShapeWeights: 0

DevProject/ProjectSettings/ProjectVersion.txt

-m_EditorVersion: 2019.4.7f1
-m_EditorVersionWithRevision: 2019.4.7f1 (e992b1a16e65)
+m_EditorVersion: 2019.4.19f1
+m_EditorVersionWithRevision: 2019.4.19f1 (ca5b14067cec)

Project/Assets/ML-Agents/Examples/Soccer/Scenes/SoccerTwos.unity

   m_Modification:
     m_TransformParent: {fileID: 0}
     m_Modifications:
+    - target: {fileID: 1141134673700168, guid: 54f3340298537426e96a6cc530e2d5d8, type: 3}
+      propertyPath: m_Name
+      value: SoccerFieldTwos
+      objectReference: {fileID: 0}
     - target: {fileID: 4558743310993102, guid: 54f3340298537426e96a6cc530e2d5d8, type: 3}
       propertyPath: m_LocalPosition.x
       value: 0
     - target: {fileID: 4558743310993102, guid: 54f3340298537426e96a6cc530e2d5d8, type: 3}
       propertyPath: m_RootOrder
       value: 4
+      objectReference: {fileID: 0}
+    - target: {fileID: 4558743310993102, guid: 54f3340298537426e96a6cc530e2d5d8, type: 3}
+      propertyPath: m_LocalEulerAnglesHint.x
+      value: 0
+      objectReference: {fileID: 0}
+    - target: {fileID: 4558743310993102, guid: 54f3340298537426e96a6cc530e2d5d8, type: 3}
+      propertyPath: m_LocalEulerAnglesHint.y
+      value: 0
+      objectReference: {fileID: 0}
+    - target: {fileID: 4558743310993102, guid: 54f3340298537426e96a6cc530e2d5d8, type: 3}
+      propertyPath: m_LocalEulerAnglesHint.z
+      value: 0
       objectReference: {fileID: 0}
     m_RemovedComponents: []
   m_SourcePrefab: {fileID: 100100000, guid: 54f3340298537426e96a6cc530e2d5d8, type: 3}
   m_Name: 
   m_EditorClassIdentifier: 
   gravityMultiplier: 1
-  monitorVerticalOffset: 0
+  reuseCollisionCallbacks: 1
 --- !u!114 &1574236051
 MonoBehaviour:
   m_ObjectHideFlags: 0
   blueMaterial: {fileID: 2100000, guid: c9fa44c2c3f8ce74ca39a3355ea42631, type: 2}
   randomizePlayersTeamForTraining: 0
   agentRunSpeed: 2
-  strikerPunish: -0.1
-  strikerReward: 1
-  goaliePunish: -1
-  goalieReward: 0.1
 --- !u!1001 &1606160104
 PrefabInstance:
   m_ObjectHideFlags: 0

Project/Assets/ML-Agents/Examples/Soccer/Scenes/StrikersVsGoalie.unity

   maximumDeltaTime: 0.33333334
   solverIterations: 6
   solverVelocityIterations: 1
+  reuseCollisionCallbacks: 1
 --- !u!114 &1574236051
 MonoBehaviour:
   m_ObjectHideFlags: 0

Project/Assets/ML-Agents/Examples/Sorter/Prefabs/Area.prefab

     VectorActionSpaceType: 0
     hasUpgradedBrainParametersWithActionSpec: 1
   m_Model: {fileID: 11400000, guid: a2b17bcb0df4d448893e800c34d87c4c, type: 3}
-  m_InferenceDevice: 0
+  m_InferenceDevice: 2
   m_BehaviorType: 0
   m_BehaviorName: Sorter
   TeamId: 0

Project/Packages/manifest.json

     "com.unity.collab-proxy": "1.2.15",
     "com.unity.ml-agents": "file:../../com.unity.ml-agents",
     "com.unity.ml-agents.extensions": "file:../../com.unity.ml-agents.extensions",
-    "com.unity.package-manager-ui": "2.0.8",
-    "com.unity.purchasing": "2.0.3",
+    "com.unity.package-manager-ui": "2.0.13",
+    "com.unity.purchasing": "2.2.1",
     "com.unity.textmeshpro": "1.4.1",
     "com.unity.modules.ai": "1.0.0",
     "com.unity.modules.animation": "1.0.0",

Project/ProjectSettings/ProjectVersion.txt

-m_EditorVersion: 2018.4.24f1
+m_EditorVersion: 2018.4.32f1

com.unity.ml-agents.extensions/Documentation~/Grid-Sensor.md

 # Summary
 
-The Grid Sensor combines the generality of data extraction from Raycasts with the image processing power of Convolutional Neural Networks. The Grid Sensor can be used to collect data in the general form of a "Width x Height x Channel" matrix which can be used for training Reinforcement Learning agents or for data analysis.
+The Grid Sensor is an alternative method for collecting observations which combines the generality of data extraction from Raycasts with the image processing power of Convolutional Neural Networks. The Grid Sensor can be used to collect data in the general form of a "Width x Height x Channel" matrix which can be used for training agent policies or for data analysis.
-In MLAgents there are 2 main sensors for observing information that is "physically" around the agent.
+In ML-Agents there are two main sensors for observing information that is "physically" around the agent.
+
-This is simple to implement and provides enough information for most simple games. When few are used, they are computationally fast. However, there are multiple limiting factors:
-* The rays need to be at the same height as the things the agent should observe
-* Objects can remain hidden by line of sight and if the knowledge of those objects is crucial to the success of the agent, then this limitation must be compensated for by the agents networks capacity (i.e., need a bigger brain with memory)
+Raycasts are simple to implement and provides enough information for most simple games. When few are used, they are also computationally lightweight. However, there are multiple limiting factors:
+
+* The rays need to be at the same height as the things the agent should observe.
+* Objects can remain hidden by line of sight and if the knowledge of those objects is crucial to the success of the agent, then this limitation must be compensated for by the agents networks capacity (i.e., need a bigger brain with memory).
-* Typically the length of the raycasts is limited because the agent need not know about objects that are at the other side of the level. Combined with few raycasts for computational efficiency, this means that an agent may not observe objects that fall between these rays and the issue becomes worse as the objects reduce in size.
+* Typically, the length of the raycasts is limited because the agent need not know about objects that are at the other side of the level. Combined with few raycasts for computational efficiency, this means that an agent may not observe objects that fall between these rays and the issue becomes worse as the objects reduce in size.
-The Camera provides the agent with either a grayscale or an RGB image of the game environment. It goes without saying that there non-linear relationships between nearby pixels in an image. It is this intuition that helps form the basis of Convolutional Neural Networks (CNNs) and established the literature of designing networks that take advantage of these relationships between pixels. Following this established literature of CNNs on image based data, the MLAgent's Camera Sensor provides a means by which the agent can include high dimensional inputs (images) into its observation stream.
+The Camera provides the agent with either a grayscale or an RGB image of the game environment. In many cases, what we want to extract from a set of pixels is invariant to the location of those pixels in the image. It is this intuition that helps form the basis of Convolutional Neural Networks (CNNs) and established the literature of designing networks that take advantage of these relationships between pixels. Following this established literature of CNNs on image based data, the ML-Agent's Camera Sensor provides a means by which the agent can include high dimensional inputs (images) into its observation stream.
-* It requires render the scene and thus is computationally slower than alternatives that do not use rendering
-* It has yet been shown that the Camera Sensor can be used on a headless machine which means it is not yet possible (if at all) to train an agent on a headless infrastructure.
+
+* It requires rendering the scene and thus is computationally slower than alternatives that do not use rendering.
-* The RGB of the camera only provides a maximum of 3 channels to the agent.
+* The RGB of the camera only provides a maximum of three channels to the agent.
-An image can be thought of as a matrix of a predefined width (W) and a height (H) and each pixel can be thought of as simply an array of length 3 (in the case of RGB), `[Red, Green, Blue]` holding the different channel information of the color (channel) intensities at that pixel location. Thus an image is just a 3 dimensional matrix of size WxHx3. A Grid Observation can be thought of as a generalization of this setup where in place of a pixel there is a "cell" which is an array of length N representing different channel intensities at that cell position. From a Convolutional Neural Network point of view, the introduction of multiple channels in an "image" isn't a new concept. One such example is using an RGB-Depth image which is used in several robotics applications. The distinction of Grid Observations is what the data within the channels represents. Instead of limiting the channels to color intensities, the channels within a cell of a Grid Observation generalize to any data that can be represented by a single number (float or int).
-
-Before jumping into the details of the Grid Sensor, an important thing to note is the agent performance and qualitatively different behavior over raycasts. Unity MLAgent's comes with a suite of example environments. One in particular, the [Food Collector](https://github.com/Unity-Technologies/ml-agents/tree/release_12_docs/docs/Learning-Environment-Examples.md#food-collector), has been the focus of the Grid Sensor development.
+## Overview
-The Food Collector environment can be described as:
-* Set-up: A multi-agent environment where agents compete to collect food.
-* Goal: The agents must learn to collect as many green food spheres as possible while avoiding red spheres.
-* Agents: The environment contains 5 agents with same Behavior Parameters.
+There are three main phases to the observation process of the Grid Sensor:
-When applying the Grid Sensor to this environment, in place of the Raycast Vector Sensor or the Camera Sensor, a Mean Reward of 40-50 is observed. This performance is on par with what is seen by agents trained with RayCasts but the side-by-side comparison of trained agents, shows a qualitative difference in behavior. A deeper study and interpretation of the qualitative differences between agents trained with Raycasts and Vector Sensors verses Grid Sensors is left to future studies.
-
-<img src="images/gridobs-vs-vectorobs.gif" align="middle" width="3000"/>
-
-## Overview
-
-There are 3 main phases to the Grid Sensor:
 1. **Collection** - data is extracted from observed objects
 2. **Encoding** - the extracted data is encoded into a grid observation
 3. **Communication** - the grid observation is sent to python or used by a trained model
 ## Collection
 
-A Grid Sensor is the Grid Observation analog of a Unity Camera but with some notable differences. The sensor is made up of a grid of identical Box Colliders which designate the "cells" of the grid. The Grid Sensor also has a list of "detectable objects" in the form of Unity gameobject tags. When an object that is tagged as a detectable object is present within a cell's Box Collider, that cell is "activated" and a method on the Grid Sensor extracts data from said object and associates that data with the position of the activated cell. Thus the Grid Sensor is always orthographic:
+A Grid Sensor is the Grid Observation analog of a Unity Camera but with some notable differences. The sensor is made up of a grid of identical Box Colliders which designate the "cells" of the grid. The Grid Sensor also has a list of "detectable objects" in the form of Unity GameObject tags. When an object that is tagged as a detectable object is present within a cell's Box Collider, that cell is "activated" and a method on the Grid Sensor extracts data from said object and associates that data with the position of the activated cell. Thus the Grid Sensor is always orthographic:
 
 <img src="images/persp_ortho_proj.png" width="500">
 <cite><a href="https://www.geofx.com/graphics/nehe-three-js/lessons17-24/lesson21/lesson21.html">geofx.com</a></cite>
-Just like the Raycasts mentioned earlier, the Grid Sensor can extract any kind of data from a detected object and just like the Camera, the Grid Sensor maintains the spacial relationship between nearby cells that allows one to take advantage of the CNN literature. Thus the Grid Sensor tries to take the best of both sensors and combines them to something that is more expressive.
+Just like the Raycasts mentioned earlier, the Grid Sensor can extract any kind of data from a detected object, and just like the Camera, the Grid Sensor maintains the spacial relationship between nearby cells that allows one to take advantage of the computational properties of CNNs. Thus the Grid Sensor tries to take the best of both sensors and combines them to something that is more expressive.
-Lets imagine a scenario where an agent is faced with 2 enemies and there are 2 "equipable" weapons somewhat behind the agent. Lets also keep in mind some important properties of the enemies and weapons that would be useful for the agent to know. For simplicity, lets assume enemies represent their health as a percentage (0-100%). Lets also assume that enemies and weapons are the only 2 kind of objects that the agent would see in the entire game.
+Let's imagine a scenario where an agent is faced with two enemies and there are two "equipable" weapons somewhat behind the agent. It would be helpful for the agent to know the location and properties of both the enemies as well as the equippable items. For simplicity, let's assume enemies represent their health as a percentage (0-100%). Also assume that enemies and weapons are the only two kinds of objects that the agent would see in the entire game.
-<img src="images/gridsensor-example.png" align="middle" width="3000"/>
+<img src="images/gridsensor-example.png" align="middle" width="512"/>
-If a raycast hits an object, not only could we get the distance (normalized by the maximum raycast distance) we would be able to extract its type (enemy vs weapon) and if its an enemy then we could get its health (e.g., .6).
+If a raycast hits an object, not only could we get the distance (normalized by the maximum raycast distance) we would be able to extract its type (enemy vs weapon) and any attribute associate with it (e.g. an enemy's health).
 
 There are many ways in which one could encode this information but one reasonable encoding is this:
 ```
 For example, if the raycast hit nothing then this would be represented by `[0, 0, 0, 1]`.
 If instead the raycast hit an enemy with 60% health that is 50% of the maximum raycast distance, the data would be represented by `[0, 1, .6, .5]`.
 
-The limitations of raycasts which were presented above are easy to visualize in the below image. The agent is unable to see where the weapons are and only sees one of the enemies. Typically in the MLAgents examples, this situation is mitigated by including previous frames of data so that the agent observes changes through time. However, in more complex games, it is not difficult to imagine scenarios where an agent would not be able to observe important information using only Raycasts.
+The limitations of raycasts which were presented above are easy to visualize in the below image. The agent is unable to see where the weapons are and only sees one of the enemies. Typically in the ML-Agents examples, this situation is mitigated by including previous frames of data so that the agent observes changes through time. However, in more complex games, it is not difficult to imagine scenarios where an agent might miss important information using only Raycasts.
-<img src="images/gridsensor-example-raycast.png" align="middle" width="3000"/>
+<img src="images/gridsensor-example-raycast.png" align="middle" width="512"/>
-<img src="images/gridsensor-example-camera.png" align="middle" width="3000"/>
+<img src="images/gridsensor-example-camera.png" align="middle" width="512"/>
 
 #### Grid Sensor
 
 
 Following the same data extraction method presented in the section on raycasts, if a Grid Sensor was used instead of Raycasts or a Camera, then not only would the agent be able to extract the health value of the enemies but it would also be able to encode the relative positions of those objects as is done with Camera. Additionally, as the texture of the objects is not used, this data can be collected without rendering the scene.
 
-<img src="images/gridsensor-example-gridsensor.png" align="middle" width="3000"/>
+In our example, we can collect data in the form of [objectType, health] by overriding `GetObjectData` as the following:
+```csharp
+    protected override float[] GetObjectData(GameObject currentColliderGo, float type_index, float normalized_distance)
+    {
+        float[] channelValues = new float[ChannelDepth.Length]; // ChannelDepth.Length = 2 in this example
+        channelValues[0] = type_index; // this is the observation collected in default implementation
+        if (currentColliderGo.tag == "enemy")
+        {
+            var enemy = currentColliderGo.GetComponent<EnemyClass>();
+            channelValues[1] = enemy.health; // the value may have to be normalized depends on the type of GridSensor encoding you use (see sections below)
+        }
+        return channelValues;
+    }
+```
-At the end of the Collection phase, each cell with an object inside of it has `GetObjectData` called and the returned values (named `channelValues`) is then processed in the Encoding phase which is described in the next section.
+<img src="images/gridsensor-example-gridsensor.png" align="middle" width="512"/>
+
+At the end of the Collection phase, each cell with an object inside of it has `GetObjectData` called and the returned values is then processed in the Encoding phase which is described in the next section.
-The CountingGridSensor builds on the GridSesnor to perform the specific job of counting the number of object types that are based on the different detectable object tags. The encoding and is meant to exploit a key feature of the Grid Sensor. In both the Channel and the Channel Hot DepthTypes, the closest detectable object, in relation to the agent, that lays within a cell is used for encoding the value for that cell. In the CountingGridSensor, the number of each type of object is recorded and then normalized according to a max count, stored in the ChannelDepth.
+The CountingGridSensor builds on the GridSensor to perform the specific job of counting the number of object types that are based on the different detectable object tags. The encoding is meant to exploit a key feature of the GridSensor. In original GridSensor, only the closest detectable object, in relation to the agent, that lies within a cell is used for encoding the value for that cell. In the CountingGridSensor, the number of each type of object is recorded and then normalized according to a max count.
 
 An example of the CountingGridSensor can be found below.
 
 In order to support different ways of representing the data extracted from an object, multiple "depth types" were implemented. Each has pros and cons and, depending on the use-case of the Grid Sensor, one may be more beneficial than the others.
 
-The data stored that is extracted during the *Collection* phase, and stored in `channelValues`, may come from different sources. For instance, going back the Enemy/Weapon example in the previous section, an enemy's health is continuous whereas the object type (enemy or weapon) is categorical data. This distinction is important as categorical data requires a different encoding mechanism than continuous data.
+The stored data that is extracted during the *Collection* phase may come from different sources, and thus be of a different nature. For instance, going back to the Enemy/Weapon example in the previous section, an enemy's health is continuous whereas the object type (enemy or weapon) is categorical data. This distinction is important as categorical data requires a different encoding mechanism than continuous data.
-The Grid Sensor handles this distinction with 4 properties that define how this data is to be encoded:
-* DepthType - Enum signifying the encoding mode: Channel, ChannelHot
-* ObservationPerCell - the total number of values that are in each cell of the grid observation
-* ChannelDepth - int[] describing the range of each data within the `channelValues`
-* ChannelOffset - int[] describing the number of encoded values that come before each data within `channelValues`
+The GridSensor handles this distinction with two user defined properties that define how this data is to be encoded:
-The ChannelDepth and the DepthType are user defined and gives control to the developer to how they can encode their data. The ChannelDepth and ChannelOffset are both initialized and used in different ways depending on the ChannelDepth and the DepthType.
+* DepthType - Enum signifying the encoding mode: Channel, ChannelHot
+* ChannelDepth - `int[]` describing the range of each data and is used differently with different DepthType
-How categorical and continuous data is treated is different between the different DepthTypes as will be explored in the sections below. The sections will use an on-going example similar to example mentioned earlier where, within a cell, the sensor observes: `an enemy with 60% health`. Thus the cell contains 2 kinds of data: categorical data (object type) and the continuous data (health). Additionally, the order of the observed tags is important as it allows one to encode the tag of the observed object by its index within list of observed tags. Note that in the example, the observed tags is defined as ["weapon", "enemy"].
+How categorical and continuous data is treated is different between the different DepthTypes as will be explored in the sections below. The sections will use an on-going example similar to the example mentioned earlier where, within a cell, the sensor observes: `an enemy with 60% health`. Thus the cell contains two kinds of data: categorical data (object type) and the continuous data (health). Additionally, the order of the observed tags is important as it allows one to encode the tag of the observed object by its index within the list of observed tags. Note that in the example, the observed tags is defined as ["weapon", "enemy"].
-The Channel Based Grid Observations is perhaps the simplest in terms of usability and similarity with other machine learning applications. Each grid is of size WxHxC where C is the number of channels. To distinguish between categorical and continuous data, one would use the ChannelDepth array to signify the ranges that the values in the `channelValues` array could take. If one sets ChannelDepth[i] to be 1, it is assumed that the value of `channelValues[i]` is already normalized. Else ChannelDepth[i] represents the total number of possible values that `channelValues[i]` can take.
+The Channel Based Grid Observations represent obsevations in a normalized form with 0 to 1. To distinguish between categorical and continuous data, one would use the ChannelDepth array to signify the ranges that the values in the `channelValues` array could take. If one sets ChannelDepth[i] to be 1, it is assumed that the value of `channelValues[i]` is already normalized. Else ChannelDepth[i] represents the total number of possible values that `channelValues[i]` can take and will be used for normalization.
-Using the example described earlier, if one was using Channel Based Grid Observations, they would have a ChannelDepth = {2, 1} to describe that there are two possible values for the first channel and the 1 represents that the second channel is already normalized.
-As the "enemy" is in the second position of the observed tags, its value can be normalized by:
+For continuous data, you should specify `ChannelDepth[i]` to 1 and the collected data should be already normalized by its min/max range. For discrete data, you should specify `ChannelDepth[i]` to be the total number of possible values, and the collected data should be an integer value within range of `ChannelDepth[i]`.
+
+Using the example described earlier, if one was using Channel Based Grid Observations, they would have a ChannelDepth = {2, 1} to describe that there are two possible values for the first channel (ObjectType) and the 1 represents that the second channel (EnemyHealth) is continuous and should be already normalized.
+
+For ObjectType, "weapon", "enemy" will be represented respectively as:
-num = detectableObjects.IndexOfTag("enemy")/ChannelDepth[0] = 2/2 = 1;
+weapon = DetectableObjects.IndexOfTag("weapon")/ChannelDepth[0] = 1/2 = 0.5;
+enemy = DetectableObjects.IndexOfTag("enemy")/ChannelDepth[0] = 2/2 = 1;
-`[1, .6]`
+`[1, .6]`. If the health in the game is not represented in a normalized form, for example if the health is represented in an integer ranging from -100 to 100, you'll need to manully nomalize it during collection. That is, If you get value 50, you need to normalize it by `50/(100- (-100))=0.25` and collect 0.25 instead of 50.
-The Channel Hot DepthType generalizes the classic OneHot encoding to differentiate combinations of different data. Rather than normalizing the data like in the Channel Based section, each element of `channelValues` is represented by an encoding based on the ChannelDepth. If ChannelDepth[i] = 1, then this represents that `channelValues[i]` is already normalized (between 0-1) and will be used directly within the encoding. However if ChannelDepth[i] is an integer greater than 1, then the value in `channelValues[i]` will be converted into a OneHot encoding based on the following:
+The Channel Hot DepthType generalizes the classic OneHot encoding to differentiate combinations of different data. Rather than normalizing the data like in the Channel Based section, each element of `channelValues` is represented by an encoding based on the ChannelDepth. If ChannelDepth[i] = 1, then this represents that `channelValues[i]` is already normalized (between 0-1) and will be used directly within the encoding which is same as with Channel Based. However if ChannelDepth[i] is an integer greater than 1, then the value in `channelValues[i]` will be converted into a OneHot encoding based on the following:
 
 ```
 float[] arr = new float[ChannelDepth[i] + 1];
 The encoding of each channel is then concatenated together. Clearly using this setup allows the developer to be able to encode values using the classic OneHot encoding. Below are some different variations of the ChannelDepth which create different encodings of the example:
 
 ##### ChannelDepth = {3, 1}
-The first element, 3, signifies that there are 3 possibilities for the first channel and as the "enemy" is 2nd in the detected objects list, the "enemy" in the example is encoded as `[0, 0, 1]` where the first index represents "no object". The second element, 1, signifies that the health is already normalized and, following the table, is used directly. The resulting encoding is thus:
+The first element, 3, signifies that there are three possibilities for the first channel and as the "enemy" is 2nd in the detected objects list, the "enemy" in the example is encoded as `[0, 0, 1]` where the first index represents "no object". The second element, 1, signifies that the health is already normalized and, following the table, is used directly. The resulting encoding is thus:
 ```
 [0, 0, 1, 0.6]
 ```
 
 ### CountingGridSensor
 
-As introduced above, the CountingGridSensor inherits from the GridSensor for the sole purpose of counting the different objects that lay within a cell. In order to normalize the counts so that the grid can be properly encoded as PNG, the ChannelDepth is used to represent the "maximum count" of each type. For the working example, if the ChannelDepth is set as {50, 10}, which represents that the maximum count for objects with the "weapon" and "enemy" tag is 50 and 10, respectively, then the resulting data would be:
+As mentioned above, the CountingGridSensor inherits from the GridSensor for the sole purpose of counting the different objects that lay within a cell. In order to normalize the counts so that the grid can be properly encoded as PNG, the ChannelDepth is used to represent the "maximum count" of each type. For the working example, if the ChannelDepth is set as {50, 10}, which represents that the maximum count for objects with the "weapon" and "enemy" tag is 50 and 10, respectively, then the resulting data would be:
 ```
 encoding = [0 weapons/ 50 weapons, 1 enemy / 10 enemies] = [0, .1]
 ```
-At the end of the Encoding phase, all of the data for a Grid Observation is placed into a float[] referred to as the perception buffer. Now the data is ready to be sent to either the python side for training or to be used by a trained model within Unity. This is where the Grid Sensor takes advantage of 2D textures and the PNG encoding schema to reduce the number of bytes that are being sent.
+At the end of the Encoding phase, all the Grid Observations will be sent to either the python side for training or to be used by a trained model within Unity. Since the data format is similar to images collected by Camera Sensors, Grid Observations also have the CompressionType option to specify whether to send the data directly or send in PNG compressed form for better communication efficiency.
-The 2D texture is a Unity class that encodes the colors of an image. It is used for many ways through out Unity but it has 2 specific methods that the Grid Sensor takes advantage of:
-
-`SetPixels` takes a 2D array of Colors and assigns the color values to the texture.
-
-`EncodeToPNG` returns a byte[] containing the PNG encoding of the colors of the texture.
-
-Together these 2 functions allow one to "push" a WxHx3 normalized array to a PNG byte[]. And indeed, this is how the Camera Sensor in Unity MLAgents sends its data to python. However, the grid sensor can have N channels so there needs to be a more generic way to send the data.
-
-The core idea behind how a Grid Observation is encoded is the following:
-1. split the channels of a Grid Observation into groups of 3
-2. encode each of these groups as a PNG byte[]
-3. concatenate all byte[] and send the combined array to python
-4. reconstruct the Grid Observation by splitting up the array and decoding the sections
-
-Once the bytes are sent to python, they are then decoded and used as a tensor of the correct shape within the mlagents python codebase.
+Once the bytes are sent to Python, they are then decoded and provided as a tensor of the correct shape.

com.unity.ml-agents.extensions/Documentation~/com.unity.ml-agents.extensions.md

  * [Match-3 sensor and actuator](Match3.md)
  * [Grid-based sensor](Grid-Sensor.md)
  * Physics-based sensors
+ * [Input System Package Integration](InputActuatorComponent.md)
 
 ## Installation
 The ML-Agents Extensions package is not currently available in the Package Manager. There are two
 ### Github via Package Manager
 In Unity 2019.4 or later, open the Package Manager, hit the "+" button, and select "Add package from git URL".
 
-![Package Manager git URL](https://github.com/Unity-Technologies/ml-agents/tree/release_12_docs/docs/images/unity_package_manager_git_url.png)
+![Package Manager git URL](https://github.com/Unity-Technologies/ml-agents/blob/release_12_docs/docs/images/unity_package_manager_git_url.png)
+
 In the dialog that appears, enter
  ```
  git+https://github.com/Unity-Technologies/ml-agents.git?path=com.unity.ml-agents.extensions
 This version of the Unity ML-Agents Extensions package is compatible with the
 following versions of the Unity Editor:
 
-- 2018.4 and later
+- If using the `InputActuatorComponent`
+    - 2019.4 or later
+    - install the `com.unity.inputsystem` package version `1.1.0-preview.3` or later.
+- Else 2018.4 and later
-none
+- For the `InputActuatorComponent`
+    - Limited implementation of `InputControls`
+    - No way to customize the action space of the `InputActuatorComponent`
 
 ## Need Help?
 The main [README](https://github.com/Unity-Technologies/ml-agents/tree/release_12_docs/README.md) contains links for contacting the team or getting support.

com.unity.ml-agents.extensions/Runtime/Unity.ML-Agents.Extensions.asmdef

     "name": "Unity.ML-Agents.Extensions",
     "references": [
         "Unity.Barracuda",
-        "Unity.ML-Agents"
+        "Unity.ML-Agents",
+        "Unity.ML-Agents.Extensions.Input"
     ],
     "includePlatforms": [],
     "excludePlatforms": []

com.unity.ml-agents.extensions/package.json

   "unity": "2018.4",
   "description": "A source-only package for new features based on ML-Agents",
   "dependencies": {
-    "com.unity.ml-agents": "1.7.2-preview"
+    "com.unity.ml-agents": "1.8.0-preview"
   }
 }

com.unity.ml-agents/CHANGELOG.md

 ### Minor Changes
 #### com.unity.ml-agents / com.unity.ml-agents.extensions (C#)
 #### ml-agents / ml-agents-envs / gym-unity (Python)
-
+- The `encoding_size` setting for RewardSignals has been deprecated. Please use `network_settings` instead. (#4982)
-
+- An issue that caused `GAIL` to fail for environments where agents can terminate episodes by self-sacrifice has been fixed. (#4971)
 
 ## [1.8.0-preview] - 2021-02-17
 ### Major Changes
 - `InferenceDevice.Burst` was added, indicating that Agent's model will be run using Barracuda's Burst backend.
   This is the default for new Agents, but existing ones that use `InferenceDevice.CPU` should update to
   `InferenceDevice.Burst`. (#4925)
+- Add an InputActuatorComponent to allow the generation of Agent action spaces from an InputActionAsset.
+  Projects wanting to use this feature will need to add the
+  [Input System Package](https://docs.unity3d.com/Packages/com.unity.inputsystem@1.1/manual/index.html)
+  at version 1.1.0-preview.3 or later. (#4881)
 
 #### ml-agents / ml-agents-envs / gym-unity (Python)
 - Tensorboard now logs the Environment Reward as both a scalar and a histogram. (#4878)

com.unity.ml-agents/Documentation~/com.unity.ml-agents.md

 # About ML-Agents package (`com.unity.ml-agents`)
 
-The Unity ML-Agents package contains the C# SDK for the [Unity ML-Agents
+The _ML-Agents_ package contains the primary C# SDK for the [Unity ML-Agents
 Toolkit].
 
 The package allows you to convert any Unity scene to into a learning environment
 instrumenting a Unity scene, setting it up for training, and then embedding the
 trained model back into your Unity scene. The machine learning algorithms that
 orchestrate training are part of the companion [Python package].
+
+Note that we also provide an _ML-Agents Extensions_ package
+(`com.unity.ml-agents.extensions`) that contains early/experimental features
+that you may find useful. This package is only available from the [ML-Agents
+GitHub repo].
 
 ## Package contents
 
 [execution order of event functions]: https://docs.unity3d.com/Manual/ExecutionOrder.html
 [connect with us]: https://github.com/Unity-Technologies/ml-agents#community-and-feedback
 [ml-agents forum]: https://forum.unity.com/forums/ml-agents.453/
+[ML-Agents GitHub repo]: https://github.com/Unity-Technologies/ml-agents/blob/release_12_docs/com.unity.ml-agents.extensions

com.unity.ml-agents/Runtime/Academy.cs

         /// Unity package version of com.unity.ml-agents.
         /// This must match the version string in package.json and is checked in a unit test.
         /// </summary>
-        internal const string k_PackageVersion = "1.7.2-preview";
+        internal const string k_PackageVersion = "1.8.0-preview";
 
         const int k_EditorTrainingPort = 5004;
 
             var port = ReadPortFromArgs();
             if (port > 0)
             {
-                Communicator = new RpcCommunicator(
-                    new CommunicatorInitParameters
-                    {
-                        port = port
-                    }
-                );
+                Communicator = CommunicatorFactory.Create();
             }
 
             if (Communicator != null)
                 bool initSuccessful = false;
                 var communicatorInitParams = new CommunicatorInitParameters
                 {
+                    port = port,
                     unityCommunicationVersion = k_ApiVersion,
                     unityPackageVersion = k_PackageVersion,
                     name = "AcademySingleton",

com.unity.ml-agents/Runtime/Actuators/ActionSpec.cs

 using System;
+using System.Collections.Generic;
-using Unity.MLAgents.Policies;
 using UnityEngine;
 
 namespace Unity.MLAgents.Actuators
         public ActionSpec(int numContinuousActions = 0, int[] discreteBranchSizes = null)
         {
             m_NumContinuousActions = numContinuousActions;
-            BranchSizes = discreteBranchSizes;
+            BranchSizes = discreteBranchSizes ?? Array.Empty<int>();
         }
 
         /// <summary>
                     "ActionSpecs must be all continuous or all discrete."
                 );
             }
+        }
+
+        /// <summary>
+        /// Combines a list of actions specs and allocates a new array of branch sizes if needed.
+        /// </summary>
+        /// <param name="specs">The list of action specs to combine.</param>
+        /// <returns>An ActionSpec which represents the aggregate of the ActionSpecs passed in.</returns>
+        public static ActionSpec Combine(params ActionSpec[] specs)
+        {
+            var numContinuous = 0;
+            var numDiscrete = 0;
+            for (var i = 0; i < specs.Length; i++)
+            {
+                var spec = specs[i];
+                numContinuous += spec.NumContinuousActions;
+                numDiscrete += spec.NumDiscreteActions;
+            }
+
+            if (numDiscrete <= 0)
+            {
+                return MakeContinuous(numContinuous);
+            }
+
+            var branchSizes = new int[numDiscrete];
+            var offset = 0;
+            for (var i = 0; i < specs.Length; i++)
+            {
+                var spec = specs[i];
+                if (spec.BranchSizes.Length == 0)
+                {
+                    continue;
+                }
+                var branchSizesLength = spec.BranchSizes.Length;
+                Array.Copy(spec.BranchSizes,
+                    0,
+                    branchSizes,
+                    offset,
+                    branchSizesLength);
+                offset += branchSizesLength;
+            }
+            return new ActionSpec(numContinuous, branchSizes);
         }
     }
 }

com.unity.ml-agents/Runtime/Actuators/ActuatorManager.cs

 using System.Collections;
 using System.Collections.Generic;
 using UnityEngine;
+using UnityEngine.Profiling;
 
 namespace Unity.MLAgents.Actuators
 {
         /// actions for the IActuators in this list.</param>
         public void UpdateActions(ActionBuffers actions)
         {
+            Profiler.BeginSample("ActuatorManager.UpdateActions");
+            Profiler.EndSample();
         }
 
         static void UpdateActionArray<T>(ActionSegment<T> sourceActionBuffer, ActionSegment<T> destination)
         /// </summary>
         public void ApplyHeuristic(in ActionBuffers actionBuffersOut)
         {
+            Profiler.BeginSample("ActuatorManager.ApplyHeuristic");
             var continuousStart = 0;
             var discreteStart = 0;
             for (var i = 0; i < m_Actuators.Count; i++)
                 continuousStart += numContinuousActions;
                 discreteStart += numDiscreteActions;
             }
+            Profiler.EndSample();
         }
 
         /// <summary>
         /// </summary>
         public void ExecuteActions()
         {
+            Profiler.BeginSample("ActuatorManager.ExecuteActions");
             ReadyActuatorsForExecution();
             var continuousStart = 0;
             var discreteStart = 0;
                 var numContinuousActions = actuator.ActionSpec.NumContinuousActions;
                 var numDiscreteActions = actuator.ActionSpec.NumDiscreteActions;
 
+                if (numContinuousActions == 0 && numDiscreteActions == 0)
+                {
+                    continue;
+                }
+
                 var continuousActions = ActionSegment<float>.Empty;
                 if (numContinuousActions > 0)
                 {
                 continuousStart += numContinuousActions;
                 discreteStart += numDiscreteActions;
             }
+            Profiler.EndSample();
         }
 
         /// <summary>

com.unity.ml-agents/Runtime/Actuators/VectorActuator.cs

+using UnityEngine.Profiling;
+
 namespace Unity.MLAgents.Actuators
 {
     /// <summary>
         /// <inheritdoc />
         public void OnActionReceived(ActionBuffers actionBuffers)
         {
-            ActionBuffers = actionBuffers;
-            m_ActionReceiver.OnActionReceived(ActionBuffers);
+            Profiler.BeginSample("VectorActuator.OnActionReceived");
+            m_ActionBuffers = actionBuffers;
+            m_ActionReceiver.OnActionReceived(m_ActionBuffers);
+            Profiler.EndSample();
+            Profiler.BeginSample("VectorActuator.Heuristic");
+            Profiler.EndSample();
         }
 
         /// <inheritdoc />

com.unity.ml-agents/Runtime/Agent.cs

         /// </summary>
         float[] m_LegacyHeuristicCache;
 
+        /// Currect MultiAgentGroup ID. Default to 0 (meaning no group)
-        internal event Action<Agent> UnregisterFromGroup;
+        /// Delegate for the agent to unregister itself from the MultiAgentGroup without cyclic reference
+        /// between agent and the group
+        internal event Action<Agent> OnAgentDisabled;
 
         /// <summary>
         /// Called when the attached [GameObject] becomes enabled and active.
                 NotifyAgentDone(DoneReason.Disabled);
             }
             m_Brain?.Dispose();
-            UnregisterFromGroup?.Invoke(this);
+            OnAgentDisabled?.Invoke(this);
             m_Initialized = false;
         }
 
 
         internal void SetMultiAgentGroup(IMultiAgentGroup multiAgentGroup)
         {
-            // Unregister from current group if this agent has been assigned one before
-            UnregisterFromGroup?.Invoke(this);
-
-            m_GroupId = multiAgentGroup.GetId();
+            if (multiAgentGroup == null)
+            {
+                m_GroupId = 0;
+            }
+            else
+            {
+                var newGroupId = multiAgentGroup.GetId();
+                if (m_GroupId == 0 || m_GroupId == newGroupId)
+                {
+                    m_GroupId = newGroupId;
+                }
+                else
+                {
+                    throw new UnityAgentsException("Agent is already registered with a group. Unregister it first.");
+                }
+            }
         }
     }
 }

com.unity.ml-agents/Runtime/Agent.deprecated.cs

 using System;
 using UnityEngine;
+using UnityEngine.Profiling;
 
 namespace Unity.MLAgents
 {
         [Obsolete("GetAction has been deprecated, please use GetStoredActionBuffers instead.")]
         public float[] GetAction()
         {
+            Profiler.BeginSample("Agent.GetAction.Deprecated");
             var actionSpec = m_PolicyFactory.BrainParameters.ActionSpec;
             // For continuous and discrete actions together, this shouldn't be called because we can only return one.
             if (actionSpec.NumContinuousActions > 0 && actionSpec.NumDiscreteActions > 0)
             {
                 return storedAction.ContinuousActions.Array;
             }
-            else
-            {
-                return Array.ConvertAll(storedAction.DiscreteActions.Array, x => (float)x);
-            }
+            Profiler.EndSample();
+            return Array.ConvertAll(storedAction.DiscreteActions.Array, x => (float)x);
         }
     }
 }

com.unity.ml-agents/Runtime/Analytics/Events.cs

         public static EventObservationSpec FromSensor(ISensor sensor)
         {
             var shape = sensor.GetObservationShape();
+            var dimProps = (sensor as IDimensionPropertiesSensor)?.GetDimensionProperties();
-                // TODO copy flags when we have them
+                dimInfos[i].Flags = dimProps != null ? (int)dimProps[i] : 0;
             }
 
             var builtInSensorType =

com.unity.ml-agents/Runtime/AssemblyInfo.cs

 [assembly: InternalsVisibleTo("Unity.ML-Agents.Editor.Tests")]
 [assembly: InternalsVisibleTo("Unity.ML-Agents.Editor")]
 [assembly: InternalsVisibleTo("Unity.ML-Agents.Extensions")]
+[assembly: InternalsVisibleTo("Unity.ML-Agents.Extensions.Input")]

com.unity.ml-agents/Runtime/Communicator/RpcCommunicator.cs

-# if UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX
+#if UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX
-#endif
 #if UNITY_EDITOR
 using UnityEditor;
 #endif
         Dictionary<string, ActionSpec> m_UnsentBrainKeys = new Dictionary<string, ActionSpec>();
 
 
-#if UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX
-#endif
-        /// The communicator parameters sent at construction
-        CommunicatorInitParameters m_CommunicatorInitParameters;
-        /// <param name="communicatorInitParameters">Communicator parameters.</param>
-        public RpcCommunicator(CommunicatorInitParameters communicatorInitParameters)
+        public RpcCommunicator()
-            m_CommunicatorInitParameters = communicatorInitParameters;
-        #region Initialization
+#region Initialization
 
         internal static bool CheckCommunicationVersionsAreCompatible(
             string unityCommunicationVersion,
             try
             {
                 initializationInput = Initialize(
+                    initParameters.port,
                     new UnityOutputProto
                     {
                         RlInitializationOutput = academyParameters
             SendCommandEvent(rlInput.Command);
         }
 
-        UnityInputProto Initialize(UnityOutputProto unityOutput, out UnityInputProto unityInput)
+        UnityInputProto Initialize(int port, UnityOutputProto unityOutput, out UnityInputProto unityInput)
-#if UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX
-            var channel = new Channel(
-                "localhost:" + m_CommunicatorInitParameters.port,
-                ChannelCredentials.Insecure);
+            var channel = new Channel($"localhost:{port}", ChannelCredentials.Insecure);
 
             m_Client = new UnityToExternalProto.UnityToExternalProtoClient(channel);
             var result = m_Client.Exchange(WrapMessage(unityOutput, 200));
                 QuitCommandReceived?.Invoke();
             }
             return result.UnityInput;
-#else
-            throw new UnityAgentsException("You cannot perform training on this platform.");
-#endif
-        #endregion
+#endregion
-        #region Destruction
+#region Destruction
 
         /// <summary>
         /// Close the communicator gracefully on both sides of the communication.
-#if UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX
             if (!m_IsOpen)
             {
                 return;
             {
                 // ignored
             }
-#else
-            throw new UnityAgentsException(
-                "You cannot perform training on this platform.");
-#endif
-        #endregion
+#endregion
-        #region Sending Events
+#region Sending Events
 
         void SendCommandEvent(CommandProto command)
         {
             }
         }
 
-        #endregion
+#endregion
-        #region Sending and retreiving data
+#region Sending and retreiving data
 
         public void DecideBatch()
         {
         /// <param name="unityOutput">The UnityOutput to be sent.</param>
         UnityInputProto Exchange(UnityOutputProto unityOutput)
         {
-#if UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX
             if (!m_IsOpen)
             {
                 return null;
                 QuitCommandReceived?.Invoke();
                 return null;
             }
-#else
-            throw new UnityAgentsException(
-                "You cannot perform training on this platform.");
-#endif
         }
 
         /// <summary>
             }
         }
 
-        #endregion
+#endregion
 
 #if UNITY_EDITOR
         /// <summary>
 #endif
     }
 }
+#endif // UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX

com.unity.ml-agents/Runtime/Policies/BehaviorParameters.cs

         [HideInInspector, SerializeField]
         BrainParameters m_BrainParameters = new BrainParameters();
 
+        public delegate void PolicyUpdated(bool isInHeuristicMode);
+
+        internal event PolicyUpdated OnPolicyUpdated;
+
         /// <summary>
         /// The associated <see cref="Policies.BrainParameters"/> for this behavior.
         /// </summary>
             get { return m_BehaviorName + "?team=" + TeamId; }
         }
 
+        void Awake()
+        {
+            OnPolicyUpdated += mode => { };
+        }
+
         internal IPolicy GeneratePolicy(ActionSpec actionSpec, ActuatorManager actuatorManager)
         {
             switch (m_BehaviorType)
             }
         }
 
+        /// <summary>
+        /// Query the behavior parameters in order to see if the Agent is running in Heuristic Mode.
+        /// </summary>
+        /// <returns>true if the Agent is running in Heuristic mode.</returns>
+        public bool IsInHeuristicMode()
+        {
+            if (BehaviorType == BehaviorType.HeuristicOnly)
+            {
+                return true;
+            }
+
+            return BehaviorType == BehaviorType.Default &&
+                ReferenceEquals(Model, null) &&
+                (!Academy.IsInitialized ||
+                    Academy.IsInitialized &&
+                    !Academy.Instance.IsCommunicatorOn);
+        }
+
         internal void UpdateAgentPolicy()
         {
             var agent = GetComponent<Agent>();
             }
             agent.ReloadPolicy();
+            OnPolicyUpdated?.Invoke(IsInHeuristicMode());
         }
     }
 }

com.unity.ml-agents/Tests/Editor/Analytics/InferenceAnalyticsTests.cs

             Assert.AreEqual(2, continuousEvent.ObservationSpecs.Count);
             Assert.AreEqual(3, continuousEvent.ObservationSpecs[0].DimensionInfos.Length);
             Assert.AreEqual(20, continuousEvent.ObservationSpecs[0].DimensionInfos[0].Size);
+            Assert.AreEqual((int)DimensionProperty.TranslationalEquivariance, continuousEvent.ObservationSpecs[0].DimensionInfos[0].Flags);
+            Assert.AreEqual((int)DimensionProperty.None, continuousEvent.ObservationSpecs[0].DimensionInfos[2].Flags);
             Assert.AreEqual("None", continuousEvent.ObservationSpecs[0].CompressionType);
             Assert.AreEqual(Test3DSensor.k_BuiltInSensorType, continuousEvent.ObservationSpecs[0].BuiltInSensorType);
             Assert.AreNotEqual(null, continuousEvent.ModelHash);

com.unity.ml-agents/Tests/Editor/BehaviorParameterTests.cs

 using NUnit.Framework;
+using Unity.Barracuda;
+using UnityEditor;
+using UnityEngine.TestTools;
 
 namespace Unity.MLAgents.Tests
 {
+        const string k_continuousONNXPath = "Packages/com.unity.ml-agents/Tests/Editor/TestModels/continuous2vis8vec2action.onnx";
         public void Heuristic(in ActionBuffers actionsOut)
         {
             // No-op
             {
                 bp.GeneratePolicy(actionSpec, new ActuatorManager());
             });
+        }
+
+        [Test]
+        public void TestIsInHeuristicMode()
+        {
+            var gameObj = new GameObject();
+            var bp = gameObj.AddComponent<BehaviorParameters>();
+            bp.Model = null;
+            gameObj.AddComponent<Agent>();
+            bp.BehaviorType = BehaviorType.HeuristicOnly;
+            Assert.IsTrue(bp.IsInHeuristicMode());
+
+            bp.BehaviorType = BehaviorType.Default;
+            Assert.IsTrue(bp.IsInHeuristicMode());
+
+            bp.Model = ScriptableObject.CreateInstance<NNModel>();
+            Assert.IsFalse(bp.IsInHeuristicMode());
+        }
+
+        [Test]
+        public void TestPolicyUpdateEventFired()
+        {
+            var gameObj = new GameObject();
+            var bp = gameObj.AddComponent<BehaviorParameters>();
+            gameObj.AddComponent<Agent>().LazyInitialize();
+            bp.OnPolicyUpdated += delegate (bool isInHeuristicMode) { Debug.Log($"OnPolicyChanged:{isInHeuristicMode}"); };
+            bp.BehaviorType = BehaviorType.HeuristicOnly;
+            LogAssert.Expect(LogType.Log, $"OnPolicyChanged:{true}");
+
+            bp.BehaviorType = BehaviorType.Default;
+            LogAssert.Expect(LogType.Log, $"OnPolicyChanged:{true}");
+
+            Assert.Throws<UnityAgentsException>(() =>
+            {
+                bp.BehaviorType = BehaviorType.InferenceOnly;
+            });
+
+            bp.Model = AssetDatabase.LoadAssetAtPath<NNModel>(k_continuousONNXPath);
+            LogAssert.Expect(LogType.Log, $"OnPolicyChanged:{false}");
+
+            bp.BehaviorType = BehaviorType.HeuristicOnly;
+            LogAssert.Expect(LogType.Log, $"OnPolicyChanged:{true}");
         }
     }
 }

com.unity.ml-agents/Tests/Editor/ParameterLoaderTest.cs

             return Sensor.GetObservationShape();
         }
     }
-    public class Test3DSensor : ISensor, IBuiltInSensor
+    public class Test3DSensor : ISensor, IBuiltInSensor, IDimensionPropertiesSensor
     {
         int m_Width;
         int m_Height;
         public BuiltInSensorType GetBuiltInSensorType()
         {
             return (BuiltInSensorType)k_BuiltInSensorType;
+        }
+
+        public DimensionProperty[] GetDimensionProperties()
+        {
+            return new[]
+            {
+                DimensionProperty.TranslationalEquivariance,
+                DimensionProperty.TranslationalEquivariance,
+                DimensionProperty.None
+            };
         }
     }
 

com.unity.ml-agents/package.json

 {
   "name": "com.unity.ml-agents",
   "displayName": "ML Agents",
-  "version": "1.7.2-preview",
+  "version": "1.8.0-preview",
-    "com.unity.barracuda": "1.3.0-preview",
+    "com.unity.barracuda": "1.3.1-preview",
     "com.unity.modules.imageconversion": "1.0.0",
     "com.unity.modules.jsonserialize": "1.0.0",
     "com.unity.modules.physics": "1.0.0",

config/imitation/CrawlerStatic.yaml

       gail:
         gamma: 0.99
         strength: 1.0
-        encoding_size: 128
+        network_settings:
+          normalize: true
+          hidden_units: 128
+          num_layers: 2
+          vis_encode_type: simple
         learning_rate: 0.0003
         use_actions: false
         use_vail: false

config/imitation/FoodCollector.yaml

       gail:
         gamma: 0.99
         strength: 0.1
-        encoding_size: 128
+        network_settings:
+          normalize: false
+          hidden_units: 128
+          num_layers: 2
+          vis_encode_type: simple
         learning_rate: 0.0003
         use_actions: false
         use_vail: false

config/imitation/Hallway.yaml

         strength: 1.0
       gail:
         gamma: 0.99
-        strength: 0.1
-        encoding_size: 128
+        strength: 0.01
         learning_rate: 0.0003
         use_actions: false
         use_vail: false

config/imitation/PushBlock.yaml

       num_layers: 2
       vis_encode_type: simple
     reward_signals:
+      extrinsic:
+        gamma: 0.99
+        strength: 1.0
-        strength: 1.0
-        encoding_size: 128
+        strength: 0.01
+        network_settings:
+          normalize: false
+          hidden_units: 128
+          num_layers: 2
+          vis_encode_type: simple
-    max_steps: 15000000
+    max_steps: 1000000
+    behavioral_cloning:
+      demo_path: Project/Assets/ML-Agents/Examples/PushBlock/Demos/ExpertPush.demo
+      steps: 50000
+      strength: 1.0
+      samples_per_update: 0

config/imitation/Pyramids.yaml

       curiosity:
         strength: 0.02
         gamma: 0.99
-        encoding_size: 256
+        network_settings:
+          hidden_units: 256
-        encoding_size: 128
         demo_path: Project/Assets/ML-Agents/Examples/Pyramids/Demos/ExpertPyramid.demo
     behavioral_cloning:
       demo_path: Project/Assets/ML-Agents/Examples/Pyramids/Demos/ExpertPyramid.demo

config/ppo/Pyramids.yaml

       curiosity:
         gamma: 0.99
         strength: 0.02
-        encoding_size: 256
+        network_settings:
+          hidden_units: 256
         learning_rate: 0.0003
     keep_checkpoints: 5
     max_steps: 10000000

config/ppo/PyramidsRND.yaml

       rnd:
         gamma: 0.99
         strength: 0.01
-        encoding_size: 64
+        network_settings:
+          hidden_units: 64
-    framework: pytorch
     threaded: true

config/ppo/VisualPyramids.yaml

       curiosity:
         gamma: 0.99
         strength: 0.01
-        encoding_size: 256
+        network_settings:
+          hidden_units: 256
         learning_rate: 0.0003
     keep_checkpoints: 5
     max_steps: 10000000

config/sac/Pyramids.yaml

       gail:
         gamma: 0.99
         strength: 0.01
-        encoding_size: 128
         learning_rate: 0.0003
         use_actions: true
         use_vail: false

config/sac/VisualPyramids.yaml

       gail:
         gamma: 0.99
         strength: 0.02
-        encoding_size: 128
         learning_rate: 0.0003
         use_actions: true
         use_vail: false

docs/Installation.md

 
 - Unity package ([`com.unity.ml-agents`](../com.unity.ml-agents/)) contains the
   Unity C# SDK that will be integrated into your Unity scene.
+- Unity package
+  ([`com.unity.ml-agents.extensions`](../com.unity.ml-agents.extensions/))
+  contains experimental C#/Unity components that are not yet ready to be part
+  of the base `com.unity.ml-agents` package. `com.unity.ml-agents.extensions`
+  has a direct dependency on `com.unity.ml-agents`.
 - Three Python packages:
   - [`mlagents`](../ml-agents/) contains the machine learning algorithms that
     enables you to train behaviors in your Unity scene. Most users of ML-Agents
 - Install Python (3.6.1 or higher)
 - Clone this repository (Optional)
   - __Note:__ If you do not clone the repository, then you will not be
-  able to access the example environments and training configurations.
-  Additionally, the [Getting Started Guide](Getting-Started.md) assumes that
-  you have cloned the repository.
+  able to access the example environments and training configurations or the
+  `com.unity.ml-agents.extensions` package. Additionally, the
+  [Getting Started Guide](Getting-Started.md) assumes that you have cloned the
+  repository.
+- Install the `com.unity.ml-agents.extensions` Unity package (Optional)
 - Install the `mlagents` Python package
 
 ### Install **Unity 2018.4** or Later
 `com.unity.ml-agents` package
 [directly from the Package Manager registry](https://docs.unity3d.com/Manual/upm-ui-install.html).
 Please make sure you enable 'Preview Packages' in the 'Advanced' dropdown in
-order to find it.
+order to find the latest Preview release of the package.
 
 **NOTE:** If you do not see the ML-Agents package listed in the Package Manager
 please follow the [advanced installation instructions](#advanced-local-installation-for-development) below.
 
 If you are going to follow the examples from our documentation, you can open the
 `Project` folder in Unity and start tinkering immediately.
+
+### Install the `com.unity.ml-agents.extensions` Unity package (Optional)
+
+To install the `com.unity.ml-agents.extensions` package, you need to first
+clone the repo and then complete a local installation similar to what was
+outlined in the previous
+[Advanced: Local Installation for Development](#advanced-local-installation-for-development-1)
+section. Complete installation steps can be found in the
+[package documentation](../com.unity.ml-agents.extensions/Documentation~/com.unity.ml-agents.extensions.md#installation).
 
 ### Install the `mlagents` Python package
 

docs/Learning-Environment-Design-Agents.md

 ```csharp
 public override void WriteDiscreteActionMask(IDiscreteActionMask actionMask)
 {
-    actionMasker.WriteMask(branch, actionIndices)
+    actionMask.WriteMask(branch, actionIndices);
 }
 ```
 
 are masked)
 
 ```csharp
-WriteMask(0, new int[2]{1,2})
+WriteMask(0, new int[2]{1,2});
 ```
 
 Notes:

docs/Learning-Environment-Examples.md

     - Recommended minimum: 6
     - Recommended maximum: 20
 
-# Strikers Vs. Goalie
+## Strikers Vs. Goalie
 
 ![StrikersVsGoalie](images/strikersvsgoalie.png)
 

docs/ML-Agents-Overview.md

 - If you want to help your agents learn (especially with environments that have
   sparse rewards) using pre-recorded demonstrations, you can generally enable
   both GAIL and Behavioral Cloning at low strengths in addition to having an
-  extrinsic reward. An example of this is provided for the Pyramids example
-  environment under `PyramidsLearning` in `config/gail_config.yaml`.
-- If you want to train purely from demonstrations, GAIL and BC _without_ an
-  extrinsic reward signal is the preferred approach. An example of this is
-  provided for the Crawler example environment under `CrawlerStaticLearning` in
-  `config/gail_config.yaml`.
+  extrinsic reward. An example of this is provided for the PushBlock example
+  environment in `config/imitation/PushBlock.yaml`.
+- If you want to train purely from demonstrations with GAIL and BC _without_ an
+  extrinsic reward signal, please see the CrawlerStatic example environment under
+  in `config/imitation/CrawlerStatic.yaml`.
+
+***Note:*** GAIL introduces a [_survivor bias_](https://arxiv.org/pdf/1809.02925.pdf)
+to the learning process. That is, by giving positive rewards based on similarity
+to the expert, the agent is incentivized to remain alive for as long as possible.
+This can directly conflict with goal-oriented tasks like our PushBlock or Pyramids
+example environments where an agent must reach a goal state thus ending the
+episode as quickly as possible. In these cases, we strongly recommend that you
+use a low strength GAIL reward signal and a sparse extrinisic signal when
+the agent achieves the task. This way, the GAIL reward signal will guide the
+agent until it discovers the extrnisic signal and will not overpower it. If the
+agent appears to be ignoring the extrinsic reward signal, you should reduce
+the strength of GAIL.
 
 #### GAIL (Generative Adversarial Imitation Learning)
 

docs/Training-Configuration-File.md

 | :--------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
 | `curiosity -> strength`      | (default = `1.0`) Magnitude of the curiosity reward generated by the intrinsic curiosity module. This should be scaled in order to ensure it is large enough to not be overwhelmed by extrinsic reward signals in the environment. Likewise it should not be too large to overwhelm the extrinsic reward signal. <br><br>Typical range: `0.001` - `0.1` |
 | `curiosity -> gamma`         | (default = `0.99`) Discount factor for future rewards. <br><br>Typical range: `0.8` - `0.995`                                                                                                                                                                                                                                                            |
-| `curiosity -> encoding_size` | (default = `64`) Size of the encoding used by the intrinsic curiosity model. This value should be small enough to encourage the ICM to compress the original observation, but also not too small to prevent it from learning to differentiate between expected and actual observations. <br><br>Typical range: `64` - `256` |
+| `curiosity -> network_settings` | Please see the documentation for `network_settings` under [Common Trainer Configurations](#common-trainer-configurations). The network specs used by the intrinsic curiosity model. The value should of `hidden_units` should be small enough to encourage the ICM to compress the original observation, but also not too small to prevent it from learning to differentiate between expected and actual observations. <br><br>Typical range: `64` - `256` |
 | `curiosity -> learning_rate` | (default = `3e-4`) Learning rate used to update the intrinsic curiosity module. This should typically be decreased if training is unstable, and the curiosity loss is unstable. <br><br>Typical range: `1e-5` - `1e-3`                                                                                                      |
 
 ### GAIL Intrinsic Reward
 | `gail -> strength`      | (default = `1.0`) Factor by which to multiply the raw reward. Note that when using GAIL with an Extrinsic Signal, this value should be set lower if your demonstrations are suboptimal (e.g. from a human), so that a trained agent will focus on receiving extrinsic rewards instead of exactly copying the demonstrations. Keep the strength below about 0.1 in those cases. <br><br>Typical range: `0.01` - `1.0`                                                                              |
 | `gail -> gamma`         | (default = `0.99`) Discount factor for future rewards. <br><br>Typical range: `0.8` - `0.9`                                                                                                                                                                                                                                                                                                                                                                                                        |
 | `gail -> demo_path`     | (Required, no default) The path to your .demo file or directory of .demo files.                                                                                                                                                                                                                                                                                                                                                                                                                        |
-| `gail -> encoding_size` | (default = `64`) Size of the hidden layer used by the discriminator. This value should be small enough to encourage the discriminator to compress the original observation, but also not too small to prevent it from learning to differentiate between demonstrated and actual behavior. Dramatically increasing this size will also negatively affect training times. <br><br>Typical range: `64` - `256`                                                           |
+| `gail -> network_settings` | Please see the documentation for `network_settings` under [Common Trainer Configurations](#common-trainer-configurations). The network specs for the GAIL discriminator. The value of `hidden_units` should be small enough to encourage the discriminator to compress the original observation, but also not too small to prevent it from learning to differentiate between demonstrated and actual behavior. Dramatically increasing this size will also negatively affect training times. <br><br>Typical range: `64` - `256`                                                           |
 | `gail -> learning_rate` | (Optional, default = `3e-4`) Learning rate used to update the discriminator. This should typically be decreased if training is unstable, and the GAIL loss is unstable. <br><br>Typical range: `1e-5` - `1e-3`                                                                                                                                                                                                                                                                  |
 | `gail -> use_actions`   | (default = `false`) Determines whether the discriminator should discriminate based on both observations and actions, or just observations. Set to True if you want the agent to mimic the actions from the demonstrations, and False if you'd rather have the agent visit the same states as in the demonstrations but with possibly different actions. Setting to False is more likely to be stable, especially with imperfect demonstrations, but may learn slower. |
 | `gail -> use_vail`      | (default = `false`) Enables a variational bottleneck within the GAIL discriminator. This forces the discriminator to learn a more general representation and reduces its tendency to be "too good" at discriminating, making learning more stable. However, it does increase training time. Enable this if you notice your imitation learning is unstable, or unable to learn the task at hand.                                                                       |
 | :--------------------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
 | `rnd -> strength`      | (default = `1.0`) Magnitude of the curiosity reward generated by the intrinsic rnd module. This should be scaled in order to ensure it is large enough to not be overwhelmed by extrinsic reward signals in the environment. Likewise it should not be too large to overwhelm the extrinsic reward signal. <br><br>Typical range: `0.001` - `0.01` |
 | `rnd -> gamma`         | (default = `0.99`) Discount factor for future rewards. <br><br>Typical range: `0.8` - `0.995`                                                                                                                                                                                                                                                            |
-| `rnd -> encoding_size` | (default = `64`) Size of the encoding used by the intrinsic RND model. <br><br>Typical range: `64` - `256` |
+| `rnd -> network_settings` | Please see the documentation for `network_settings` under [Common Trainer Configurations](#common-trainer-configurations). The network specs for the RND model. |
 | `curiosity -> learning_rate` | (default = `3e-4`) Learning rate used to update the RND module. This should be large enough for the RND module to quickly learn the state representation, but small enough to allow for stable learning. <br><br>Typical range: `1e-5` - `1e-3`
 
 

gym-unity/README.md

 def main():
     unity_env = UnityEnvironment("./envs/GridWorld")
     env = UnityToGymWrapper(unity_env, 0, uint8_visual=True)
-    logger.configure('./logs') # Çhange to log in a different directory
+    logger.configure('./logs') # Change to log in a different directory
     act = deepq.learn(
         env,
         "cnn", # conv_only is also a good choice for GridWorld

gym-unity/gym_unity/__init__.py

 # Version of the library that will be used to upload to pypi
-__version__ = "0.24.0.dev0"
+__version__ = "0.25.0.dev0"
 
 # Git tag that will be checked to determine whether to trigger upload to pypi
 __release_tag__ = None

gym-unity/gym_unity/envs/__init__.py

         return -float("inf"), float("inf")
 
     @property
-    def spec(self):
-        return None
-
-    @property
     def action_space(self):
         return self._action_space
 

ml-agents-envs/mlagents_envs/__init__.py

 # Version of the library that will be used to upload to pypi
-__version__ = "0.24.0.dev0"
+__version__ = "0.25.0.dev0"
 
 # Git tag that will be checked to determine whether to trigger upload to pypi
 __release_tag__ = None

ml-agents/mlagents/plugins/stats_writer.py

     and evaluates them, and returns the list of all the StatsWriter implementations.
     """
     all_stats_writers: List[StatsWriter] = []
+    if ML_AGENTS_STATS_WRITER not in importlib_metadata.entry_points():
+        logger.warning(
+            f"Unable to find any entry points for {ML_AGENTS_STATS_WRITER}, even the default ones. "
+            "Uninstalling and reinstalling ml-agents via pip should resolve. "
+            "Using default plugins for now."
+        )
+        return get_default_stats_writers(run_options)
+
     entry_points = importlib_metadata.entry_points()[ML_AGENTS_STATS_WRITER]
 
     for entry_point in entry_points:

ml-agents/mlagents/trainers/__init__.py

 # Version of the library that will be used to upload to pypi
-__version__ = "0.24.0.dev0"
+__version__ = "0.25.0.dev0"
 
 # Git tag that will be checked to determine whether to trigger upload to pypi
 __release_tag__ = None

ml-agents/mlagents/trainers/settings.py

 class RewardSignalSettings:
     gamma: float = 0.99
     strength: float = 1.0
+    network_settings: NetworkSettings = attr.ib(factory=NetworkSettings)
 
     @staticmethod
     def structure(d: Mapping, t: type) -> Any:
             enum_key = RewardSignalType(key)
             t = enum_key.to_settings()
             d_final[enum_key] = strict_to_cls(val, t)
+            # Checks to see if user specifying deprecated encoding_size for RewardSignals.
+            # If network_settings is not specified, this updates the default hidden_units
+            # to the value of encoding size. If specified, this ignores encoding size and
+            # uses network_settings values.
+            if "encoding_size" in val:
+                logger.warning(
+                    "'encoding_size' was deprecated for RewardSignals. Please use network_settings."
+                )
+                # If network settings was not specified, use the encoding size. Otherwise, use hidden_units
+                if "network_settings" not in val:
+                    d_final[enum_key].network_settings.hidden_units = val[
+                        "encoding_size"
+                    ]
-    encoding_size: int = 64
+    encoding_size: Optional[int] = None
     use_actions: bool = False
     use_vail: bool = False
     demo_path: str = attr.ib(kw_only=True)
 class CuriositySettings(RewardSignalSettings):
-    encoding_size: int = 64
+    encoding_size: Optional[int] = None
-    encoding_size: int = 64
+    encoding_size: Optional[int] = None
 
 
 # SAMPLERS #############################################################################

ml-agents/mlagents/trainers/torch/components/reward_providers/curiosity_reward_provider.py

 from mlagents.trainers.settings import CuriositySettings
 
 from mlagents_envs.base_env import BehaviorSpec
+from mlagents_envs import logging_util
-from mlagents.trainers.settings import NetworkSettings, EncoderType
+
+logger = logging_util.get_logger(__name__)
 
 
 class ActionPredictionTuple(NamedTuple):
     def __init__(self, specs: BehaviorSpec, settings: CuriositySettings) -> None:
         super().__init__()
         self._action_spec = specs.action_spec
-        state_encoder_settings = NetworkSettings(
-            normalize=False,
-            hidden_units=settings.encoding_size,
-            num_layers=2,
-            vis_encode_type=EncoderType.SIMPLE,
-            memory=None,
-        )
+
+        state_encoder_settings = settings.network_settings
+        if state_encoder_settings.memory is not None:
+            state_encoder_settings.memory = None
+            logger.warning(
+                "memory was specified in network_settings but is not supported by Curiosity. It is being ignored."
+            )
+
         self._state_encoder = NetworkBody(
             specs.observation_specs, state_encoder_settings
         )
         self.inverse_model_action_encoding = torch.nn.Sequential(
-            LinearEncoder(2 * settings.encoding_size, 1, 256)
+            LinearEncoder(2 * state_encoder_settings.hidden_units, 1, 256)
         )
 
         if self._action_spec.continuous_size > 0:
 
         self.forward_model_next_state_prediction = torch.nn.Sequential(
             LinearEncoder(
-                settings.encoding_size + self._action_flattener.flattened_size, 1, 256
+                state_encoder_settings.hidden_units
+                + self._action_flattener.flattened_size,
+                1,
+                256,
-            linear_layer(256, settings.encoding_size),
+            linear_layer(256, state_encoder_settings.hidden_units),
         )
 
     def get_current_state(self, mini_batch: AgentBuffer) -> torch.Tensor:

ml-agents/mlagents/trainers/torch/components/reward_providers/gail_reward_provider.py

 )
 from mlagents.trainers.settings import GAILSettings
 from mlagents_envs.base_env import BehaviorSpec
+from mlagents_envs import logging_util
-from mlagents.trainers.settings import NetworkSettings, EncoderType
+logger = logging_util.get_logger(__name__)
+
-        self._ignore_done = True
+        self._ignore_done = False
         self._discriminator_network = DiscriminatorNetwork(specs, settings)
         self._discriminator_network.to(default_device())
         _, self._demo_buffer = demo_to_buffer(
             )
 
     def update(self, mini_batch: AgentBuffer) -> Dict[str, np.ndarray]:
+
+        self._discriminator_network.encoder.update_normalization(expert_batch)
+
         loss, stats_dict = self._discriminator_network.compute_loss(
             mini_batch, expert_batch
         )
         self._use_vail = settings.use_vail
         self._settings = settings
 
-        encoder_settings = NetworkSettings(
-            normalize=False,
-            hidden_units=settings.encoding_size,
-            num_layers=2,
-            vis_encode_type=EncoderType.SIMPLE,
-            memory=None,
-        )
+        encoder_settings = settings.network_settings
+        if encoder_settings.memory is not None:
+            encoder_settings.memory = None
+            logger.warning(
+                "memory was specified in network_settings but is not supported by GAIL. It is being ignored."
+            )
+
         self._action_flattener = ActionFlattener(specs.action_spec)
         unencoded_size = (
             self._action_flattener.flattened_size + 1 if settings.use_actions else 0
         )
 
-        estimator_input_size = settings.encoding_size
+        estimator_input_size = encoder_settings.hidden_units
         if settings.use_vail:
             estimator_input_size = self.z_size
             self._z_sigma = torch.nn.Parameter(
-                settings.encoding_size,
+                encoder_settings.hidden_units,
                 self.z_size,
                 kernel_init=Initialization.KaimingHeNormal,
                 kernel_gain=0.1,

ml-agents/mlagents/trainers/torch/components/reward_providers/rnd_reward_provider.py

 from mlagents.trainers.settings import RNDSettings
 
 from mlagents_envs.base_env import BehaviorSpec
+from mlagents_envs import logging_util
-from mlagents.trainers.settings import NetworkSettings, EncoderType
+
+logger = logging_util.get_logger(__name__)
 
 
 class RNDRewardProvider(BaseRewardProvider):
 
     def __init__(self, specs: BehaviorSpec, settings: RNDSettings) -> None:
         super().__init__()
-        state_encoder_settings = NetworkSettings(
-            normalize=True,
-            hidden_units=settings.encoding_size,
-            num_layers=3,
-            vis_encode_type=EncoderType.SIMPLE,
-            memory=None,
-        )
+        state_encoder_settings = settings.network_settings
+        if state_encoder_settings.memory is not None:
+            state_encoder_settings.memory = None
+            logger.warning(
+                "memory was specified in network_settings but is not supported by RND. It is being ignored."
+            )
+
         self._encoder = NetworkBody(specs.observation_specs, state_encoder_settings)
 
     def forward(self, mini_batch: AgentBuffer) -> torch.Tensor:

ml-agents/tests/yamato/standalone_build_tests.py

 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("--scene", default=None)
-    parser.add_argument("--build-target", default="mac", choices=["mac", "linux"])
+    parser.add_argument(
+        "--build-target", default="mac", choices=["mac", "linux", "ios", "webgl"]
+    )
     args = parser.parse_args()
     main(args.scene, args.build_target)

ml-agents/tests/yamato/yamato_utils.py

 import subprocess
 import yaml
 from sys import platform
-from typing import List, Optional
+from typing import List, Optional, Mapping
 
 
 def get_unity_executable_path():
     print(f"Running BuildStandalonePlayer via {unity_exe}")
 
     # enum values from https://docs.unity3d.com/2019.4/Documentation/ScriptReference/BuildTarget.html
-    build_target_to_enum = {
+    build_target_to_enum: Mapping[Optional[str], str] = {
+    # Convert the short name to the official enum
+    # Just pass through if it's not on the list.
+    build_target_enum = build_target_to_enum.get(build_target, build_target)
 
     test_args = [
         unity_exe,
         os.makedirs(os.path.dirname(output_path), exist_ok=True)
     if scene_path is not None:
         test_args += ["--mlagents-build-scene-path", scene_path]
-    if build_target is not None:
-        test_args += ["--mlagents-build-target", build_target_to_enum[build_target]]
+    if build_target_enum is not None:
+        test_args += ["--mlagents-build-target", build_target_enum]
     print(f"{' '.join(test_args)} ...")
 
     timeout = 30 * 60  # 30 minutes, just in case

.yamato/standalone-build-webgl-test.yml

+{% capture editor_version %}2020.2{% endcapture %}
+test_webgl_standalone_{{ editor_version }}:
+  name: Test WebGL Standalone {{ editor_version }}
+  agent:
+    type: Unity::VM
+    image: package-ci/ubuntu:stable
+    flavor: b1.large
+  variables:
+    UNITY_VERSION: {{ editor_version }}
+  commands:
+    - |
+      python3 -m venv venv && source venv/bin/activate
+      python -m pip install pyyaml --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple
+      python -m pip install unity-downloader-cli --index-url https://artifactory.prd.it.unity3d.com/artifactory/api/pypi/pypi/simple --upgrade
+      unity-downloader-cli -u {{ editor_version }} -c editor -c WebGL --wait --fast
+      python -u -m ml-agents.tests.yamato.standalone_build_tests --build-target=webgl
+
+  triggers:
+    cancel_old_ci: true
+    recurring:
+      - branch: master
+        frequency: weekly
+  artifacts:
+    logs:
+      paths:
+        - "artifacts/standalone_build.txt"
+    standalonebuild:
+      paths:
+        - "artifacts/testPlayer*/**"
+        - "artifacts/**/UnityPlayer.so"

DevProject/DevProject.sln.DotSettings

+<wpf:ResourceDictionary xml:space="preserve" xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml" xmlns:s="clr-namespace:System;assembly=mscorlib" xmlns:ss="urn:shemas-jetbrains-com:settings-storage-xaml" xmlns:wpf="http://schemas.microsoft.com/winfx/2006/xaml/presentation">
+	<s:String x:Key="/Default/CodeStyle/Naming/CSharpNaming/Abbreviations/=RL/@EntryIndexedValue">RL</s:String></wpf:ResourceDictionary>

DevProject/Packages/packages-lock.json

+{
+  "dependencies": {
+    "com.unity.2d.sprite": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.2d.tilemap": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.ads": {
+      "version": "3.6.1",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.ugui": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.analytics": {
+      "version": "3.3.5",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.ugui": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.barracuda": {
+      "version": "1.3.0-preview",
+      "depth": 1,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.burst": "1.3.4",
+        "com.unity.modules.jsonserialize": "1.0.0",
+        "com.unity.modules.imageconversion": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.burst": {
+      "version": "1.3.4",
+      "depth": 2,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.mathematics": "1.2.1"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.coding": {
+      "version": "0.1.0-preview.13",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "nuget.moq": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.collab-proxy": {
+      "version": "1.2.16",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.ext.nunit": {
+      "version": "1.0.6",
+      "depth": 1,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.ide.rider": {
+      "version": "1.1.4",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.test-framework": "1.1.1"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.ide.vscode": {
+      "version": "1.2.3",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.inputsystem": {
+      "version": "1.1.0-preview.3",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.mathematics": {
+      "version": "1.2.1",
+      "depth": 3,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.ml-agents": {
+      "version": "file:../../com.unity.ml-agents",
+      "depth": 0,
+      "source": "local",
+      "dependencies": {
+        "com.unity.barracuda": "1.3.0-preview",
+        "com.unity.modules.imageconversion": "1.0.0",
+        "com.unity.modules.jsonserialize": "1.0.0",
+        "com.unity.modules.physics": "1.0.0",
+        "com.unity.modules.physics2d": "1.0.0",
+        "com.unity.modules.unityanalytics": "1.0.0"
+      }
+    },
+    "com.unity.ml-agents.extensions": {
+      "version": "file:../../com.unity.ml-agents.extensions",
+      "depth": 0,
+      "source": "local",
+      "dependencies": {
+        "com.unity.ml-agents": "1.7.2-preview"
+      }
+    },
+    "com.unity.multiplayer-hlapi": {
+      "version": "1.0.8",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "nuget.mono-cecil": "0.1.6-preview"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.nuget.mono-cecil": {
+      "version": "0.1.6-preview.2",
+      "depth": 1,
+      "source": "registry",
+      "dependencies": {
+        "nuget.mono-cecil": "0.1.6-preview"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.nuget.newtonsoft-json": {
+      "version": "2.0.0-preview",
+      "depth": 1,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.package-manager-doctools": {
+      "version": "1.7.0-preview",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.package-validation-suite": "0.10.0-preview",
+        "com.unity.nuget.newtonsoft-json": "2.0.0-preview"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.package-validation-suite": {
+      "version": "0.19.0-preview",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.nuget.mono-cecil": "0.1.6-preview.2"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.purchasing": {
+      "version": "2.2.1",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.ugui": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.settings-manager": {
+      "version": "1.0.1",
+      "depth": 1,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.test-framework": {
+      "version": "1.1.20",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.ext.nunit": "1.0.6",
+        "com.unity.modules.imgui": "1.0.0",
+        "com.unity.modules.jsonserialize": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.test-framework.performance": {
+      "version": "2.2.0-preview",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.test-framework": "1.1.0",
+        "com.unity.nuget.newtonsoft-json": "2.0.0-preview"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.testtools.codecoverage": {
+      "version": "1.0.0-pre.3",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.test-framework": "1.0.16",
+        "com.unity.settings-manager": "1.0.1"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.textmeshpro": {
+      "version": "2.0.1",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.ugui": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.timeline": {
+      "version": "1.2.12",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.ugui": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.ui": "1.0.0",
+        "com.unity.modules.imgui": "1.0.0"
+      }
+    },
+    "com.unity.xr.legacyinputhelpers": {
+      "version": "2.1.7",
+      "depth": 0,
+      "source": "registry",
+      "dependencies": {
+        "com.unity.modules.vr": "1.0.0",
+        "com.unity.modules.xr": "1.0.0"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "nuget.castle-core": {
+      "version": "1.0.1",
+      "depth": 2,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "nuget.mono-cecil": {
+      "version": "0.1.6-preview",
+      "depth": 1,
+      "source": "registry",
+      "dependencies": {},
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "nuget.moq": {
+      "version": "1.0.0",
+      "depth": 1,
+      "source": "registry",
+      "dependencies": {
+        "nuget.castle-core": "1.0.1"
+      },
+      "url": "https://artifactory.prd.cds.internal.unity3d.com/artifactory/api/npm/upm-candidates"
+    },
+    "com.unity.modules.ai": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.androidjni": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.animation": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.assetbundle": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.audio": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.cloth": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.physics": "1.0.0"
+      }
+    },
+    "com.unity.modules.director": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.audio": "1.0.0",
+        "com.unity.modules.animation": "1.0.0"
+      }
+    },
+    "com.unity.modules.imageconversion": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.imgui": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.jsonserialize": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.particlesystem": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.physics": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.physics2d": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.screencapture": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.imageconversion": "1.0.0"
+      }
+    },
+    "com.unity.modules.subsystems": {
+      "version": "1.0.0",
+      "depth": 1,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.jsonserialize": "1.0.0"
+      }
+    },
+    "com.unity.modules.terrain": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.terrainphysics": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.physics": "1.0.0",
+        "com.unity.modules.terrain": "1.0.0"
+      }
+    },
+    "com.unity.modules.tilemap": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.physics2d": "1.0.0"
+      }
+    },
+    "com.unity.modules.ui": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.uielements": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.imgui": "1.0.0",
+        "com.unity.modules.jsonserialize": "1.0.0"
+      }
+    },
+    "com.unity.modules.umbra": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.unityanalytics": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.unitywebrequest": "1.0.0",
+        "com.unity.modules.jsonserialize": "1.0.0"
+      }
+    },
+    "com.unity.modules.unitywebrequest": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.unitywebrequestassetbundle": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.assetbundle": "1.0.0",
+        "com.unity.modules.unitywebrequest": "1.0.0"
+      }
+    },
+    "com.unity.modules.unitywebrequestaudio": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.unitywebrequest": "1.0.0",
+        "com.unity.modules.audio": "1.0.0"
+      }
+    },
+    "com.unity.modules.unitywebrequesttexture": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.unitywebrequest": "1.0.0",
+        "com.unity.modules.imageconversion": "1.0.0"
+      }
+    },
+    "com.unity.modules.unitywebrequestwww": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.unitywebrequest": "1.0.0",
+        "com.unity.modules.unitywebrequestassetbundle": "1.0.0",
+        "com.unity.modules.unitywebrequestaudio": "1.0.0",
+        "com.unity.modules.audio": "1.0.0",
+        "com.unity.modules.assetbundle": "1.0.0",
+        "com.unity.modules.imageconversion": "1.0.0"
+      }
+    },
+    "com.unity.modules.vehicles": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.physics": "1.0.0"
+      }
+    },
+    "com.unity.modules.video": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.audio": "1.0.0",
+        "com.unity.modules.ui": "1.0.0",
+        "com.unity.modules.unitywebrequest": "1.0.0"
+      }
+    },
+    "com.unity.modules.vr": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.jsonserialize": "1.0.0",
+        "com.unity.modules.physics": "1.0.0",
+        "com.unity.modules.xr": "1.0.0"
+      }
+    },
+    "com.unity.modules.wind": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {}
+    },
+    "com.unity.modules.xr": {
+      "version": "1.0.0",
+      "depth": 0,
+      "source": "builtin",
+      "dependencies": {
+        "com.unity.modules.physics": "1.0.0",
+        "com.unity.modules.jsonserialize": "1.0.0",
+        "com.unity.modules.subsystems": "1.0.0"
+      }
+    }
+  }
+}

com.unity.ml-agents.extensions/Documentation~/InputActuatorComponent.md

+# Integration of the Input System Package with ML-Agents
+
+## Overview
+One area we are always trying to improve is getting developers up and running with ML-Agents.  With this in mind,
+we have implemented an `InputActuatorComponent`.  This component integrates with the
+[Input System Package](https://docs.unity3d.com/Packages/com.unity.inputsystem@1.1/manual/QuickStartGuide.html)
+to set up an action space for your `Agent` based on an `InputActionAsset` that is referenced by the
+`IInputActionAssetProvider` interface, or the `PlayerInput` component that may be living on your player controlled
+`Agent`.  This means that if you have code outside of your agent that handles input, you will not need to implement
+the Heuristic function in agent as well.  The `InputActuatorComponent` will handle this for you.  You can now train and
+run inference on `Agents` with an action space defined by an `InputActionAsset`.
+
+This implementation includes:
+
+* C# `InputActuatorComponent` you can attach to your Agent.
+* Implement the `IInputActionAssetProvider` in the `Componenet` where you handle player input.
+* An example environment where the input handling code is not in the Heuristic function of the Agent subclass.
+
+### Feedback
+We have only implemented a small subset of `InputControl` types that we thought would cover a large portion of what
+most developers would use.  Please let us know if you want more control types implemented by posting in the [ML-Agents
+forum.](https://forum.unity.com/forums/ml-agents.453/)
+
+We would also like your feedback on the workflow of integrating this into your games.  If you run
+into workflow issues please let us know in the ML-Agents forums, or if you've discovered a bug,
+please file a bug on our GitHub page.
+
+## Getting started
+The C# code for the `InputActuatorComponent` exists inside of the extensions package (com.unity.ml-agents.extensions).  A good first step would be to familiarize with the extensions package by reading the document [here](com.unity.ml-agents.extensions.md).  The second step would be to take a look at how we have implemented the C# code in the example Input Integration scene (located under  ML-Agents-Input-Example/Assets/ML-Agents/Examples/PushBlock/).  Once you have some familiarity, then the next step would be to add the InputActuatorComponent to your player Agent.  The example we have implemented uses C# Events to send information from the Input System.
+
+Additionally, see below for additional technical specifications on the C# code for the InputActuatorComponent.
+
+## Technical specifications for the InputActuatorComponent
+
+### `IInputActionsAssetProvider` Interface
+The `InputActuatorComponent` searches for a `Component` that implements
+`IInputActionAssetProvider` on the `GameObject` they both are attached to.  It is important to note
+that if multiple `Components` on your `GameObject` need to access an `InputActionAsset` to handle events,
+they will need to share the same instance of the `InputActionAsset` that is returned from the
+`IInputActionAssetProvider`.
+
+### `InputActuatorComponent` class
+The `InputActuatorComponent` is the bridge between ML-Agents and the Input System.. It allows ML-Agents to
+* create an `ActionSpec` for your Agent based on an `InputActionAsset` that comes from an
+`IInputActionAssetProvider`.
+* send simulated input from a training process or a neural network
+* let developers keep their input handling code in one place
+
+This is accomplished by adding the `InputActuatorComponenet` to an Agent which already has the PlayerInput component attached.
+
+### Setting up a scene using the `InputActuatorComponent`
+1. Add the `com.unity.inputsystem` version 1.1.0-preview.3 or later to your project via the Package Manager window.
+2. If you have already setup an InputActionAsset skip to Step 3, otherwise follow these sub steps:
+    1. Create an InputActionAsset to allow your Agent to be controlled by the Input System.
+    2. Handle the events from the Input System where you normally would (i.e. a script external to your Agent class).
+3. Add the InputSystemActuatorComponent to the GameObject that has the `PlayerInput` and `Agent` components attached.
+

com.unity.ml-agents.extensions/Editor/Input.meta

+fileFormatVersion: 2
+guid: 1f773a20e85042999e87f5d3c7b55281
+timeCreated: 1613637190

com.unity.ml-agents.extensions/Runtime/Input.meta

+fileFormatVersion: 2
+guid: 1694e881b9ec420ba1c201f0612392d6
+timeCreated: 1610754907

com.unity.ml-agents.extensions/Tests/Runtime/Input.meta

+fileFormatVersion: 2
+guid: 27f8e1ce37d7485f814ce50a37101203
+timeCreated: 1612908869

com.unity.ml-agents/Runtime/Communicator/CommunicatorFactory.cs

+namespace Unity.MLAgents
+{
+    /// <summary>
+    /// Factory class for an ICommunicator instance. This is used to the <see cref="Academy"/> at startup.
+    /// By default, on desktop platforms, an ICommunicator will be created and attempt to connect
+    /// to a trainer. This behavior can be prevented by setting <see cref="CommunicatorFactory.Enabled"/> to false
+    /// *before* the <see cref="Academy"/> is initialized.
+    /// </summary>
+    public static class CommunicatorFactory
+    {
+        static bool s_Enabled = true;
+
+        /// <summary>
+        /// Whether or not an ICommunicator instance will be created when the <see cref="Academy"/> is initialized.
+        /// Changing this has no effect after the <see cref="Academy"/> has already been initialized.
+        /// </summary>
+        public static bool Enabled
+        {
+            get => s_Enabled;
+            set => s_Enabled = value;
+        }
+
+        internal static ICommunicator Create()
+        {
+#if UNITY_EDITOR || UNITY_STANDALONE_WIN || UNITY_STANDALONE_OSX || UNITY_STANDALONE_LINUX
+            if (s_Enabled)
+            {
+                return new RpcCommunicator();
+            }
+#endif
+            // Non-desktop or disabled
+            return null;
+        }
+    }
+}

com.unity.ml-agents/Runtime/Communicator/CommunicatorFactory.cs.meta

+fileFormatVersion: 2
+guid: 0b604cddc07e4484a2cdaba630a971ea
+timeCreated: 1613617949

com.unity.ml-agents/Runtime/SimpleMultiAgentGroup.cs

+using System;
+using System.Linq;
+using System.Collections.Generic;
+
+namespace Unity.MLAgents
+{
+    /// <summary>
+    /// A basic class implementation of MultiAgentGroup.
+    /// </summary>
+    internal class SimpleMultiAgentGroup : IMultiAgentGroup, IDisposable
+    {
+        readonly int m_Id = MultiAgentGroupIdCounter.GetGroupId();
+        HashSet<Agent> m_Agents = new HashSet<Agent>();
+
+
+        public virtual void Dispose()
+        {
+            while (m_Agents.Count > 0)
+            {
+                UnregisterAgent(m_Agents.First());
+            }
+        }
+
+        /// <inheritdoc />
+        public virtual void RegisterAgent(Agent agent)
+        {
+            if (!m_Agents.Contains(agent))
+            {
+                agent.SetMultiAgentGroup(this);
+                m_Agents.Add(agent);
+                agent.OnAgentDisabled += UnregisterAgent;
+            }
+        }
+
+        /// <inheritdoc />
+        public virtual void UnregisterAgent(Agent agent)
+        {
+            if (m_Agents.Contains(agent))
+            {
+                agent.SetMultiAgentGroup(null);
+                m_Agents.Remove(agent);
+                agent.OnAgentDisabled -= UnregisterAgent;
+            }
+        }
+
+        /// <inheritdoc />
+        public int GetId()
+        {
+            return m_Id;
+        }
+
+        /// <summary>
+        /// Get list of all agents currently registered to this MultiAgentGroup.
+        /// </summary>
+        /// <returns>
+        /// List of agents registered to the MultiAgentGroup.
+        /// </returns>
+        public IReadOnlyCollection<Agent> GetRegisteredAgents()
+        {
+            return (IReadOnlyCollection<Agent>)m_Agents;
+        }
+
+        /// <summary>
+        /// Increments the group rewards for all agents in this MultiAgentGroup.
+        /// </summary>
+        /// <remarks>
+        /// This function increases or decreases the group rewards by a given amount for all agents
+        /// in the group. Use <see cref="SetGroupReward(float)"/> to set the group reward assigned
+        /// to the current step with a specific value rather than increasing or decreasing it.
+        ///
+        /// A positive group reward indicates the whole group's accomplishments or desired behaviors.
+        /// Every agent in the group will receive the same group reward no matter whether the
+        /// agent's act directly leads to the reward. Group rewards are meant to reinforce agents
+        /// to act in the group's best interest instead of individual ones.
+        /// Group rewards are treated differently than individual agent rewards during training, so
+        /// calling AddGroupReward() is not equivalent to calling agent.AddReward() on each agent in the group.
+        /// </remarks>
+        /// <param name="reward">Incremental group reward value.</param>
+        public void AddGroupReward(float reward)
+        {
+            foreach (var agent in m_Agents)
+            {
+                agent.AddGroupReward(reward);
+            }
+        }
+
+        /// <summary>
+        /// Set the group rewards for all agents in this MultiAgentGroup.
+        /// </summary>
+        /// <remarks>
+        /// This function replaces any group rewards given during the current step for all agents in the group.
+        /// Use <see cref="AddGroupReward(float)"/> to incrementally change the group reward rather than
+        /// overriding it.
+        ///
+        /// A positive group reward indicates the whole group's accomplishments or desired behaviors.
+        /// Every agent in the group will receive the same group reward no matter whether the
+        /// agent's act directly leads to the reward. Group rewards are meant to reinforce agents
+        /// to act in the group's best interest instead of indivisual ones.
+        /// Group rewards are treated differently than individual agent rewards during training, so
+        /// calling SetGroupReward() is not equivalent to calling agent.SetReward() on each agent in the group.
+        /// </remarks>
+        /// <param name="reward">The new value of the group reward.</param>
+        public void SetGroupReward(float reward)
+        {
+            foreach (var agent in m_Agents)
+            {
+                agent.SetGroupReward(reward);
+            }
+        }
+
+        /// <summary>
+        /// End episodes for all agents in this MultiAgentGroup.
+        /// </summary>
+        /// <remarks>
+        /// This should be used when the episode can no longer continue, such as when the group
+        /// reaches the goal or fails at the task.
+        /// </remarks>
+        public void EndGroupEpisode()
+        {
+            foreach (var agent in m_Agents)
+            {
+                agent.EndEpisode();
+            }
+        }
+
+        /// <summary>
+        /// Indicate that the episode is over but not due to the "fault" of the group.
+        /// This has the same end result as calling <see cref="EndGroupEpisode"/>, but has a
+        /// slightly different effect on training.
+        /// </summary>
+        /// <remarks>
+        /// This should be used when the episode could continue, but has gone on for
+        /// a sufficient number of steps, such as if the environment hits some maximum number of steps.
+        /// </remarks>
+        public void GroupEpisodeInterrupted()
+        {
+            foreach (var agent in m_Agents)
+            {
+                agent.EpisodeInterrupted();
+            }
+        }
+    }
+}

com.unity.ml-agents/Runtime/SimpleMultiAgentGroup.cs.meta

+fileFormatVersion: 2
+guid: 3454e3c3c70964dca93b63ee4b650095
+MonoImporter:
+  externalObjects: {}
+  serializedVersion: 2
+  defaultReferences: []
+  executionOrder: 0
+  icon: {instanceID: 0}
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

com.unity.ml-agents/Tests/Editor/Actuators/ActionSpecTests.cs

+using System;
+using System.Collections.Generic;
+using System.Linq;
+using NUnit.Framework;
+using Unity.MLAgents.Actuators;
+using UnityEngine.TestTools.Constraints;
+using Is = UnityEngine.TestTools.Constraints.Is;
+
+namespace Unity.MLAgents.Tests.Actuators
+{
+    [TestFixture]
+    public class ActionSpecTests
+    {
+        [Test]
+        public void ActionSpecCombineTest()
+        {
+            var as0 = new ActionSpec(3, new[] { 3, 2, 1 });
+            var as1 = new ActionSpec(1, new[] { 35, 122, 1, 3, 8, 3 });
+
+            var as0NumCon = 3;
+            var as0NumDis = as0.NumDiscreteActions;
+            var as1NumCon = 1;
+            var as1NumDis = as1.NumDiscreteActions;
+            var branchSizes = new List<int>();
+            branchSizes.AddRange(as0.BranchSizes);
+            branchSizes.AddRange(as1.BranchSizes);
+
+            var asc = ActionSpec.Combine(as0, as1);
+
+            Assert.AreEqual(as0NumCon + as1NumCon, asc.NumContinuousActions);
+            Assert.AreEqual(as0NumDis + as1NumDis, asc.NumDiscreteActions);
+            Assert.IsTrue(branchSizes.ToArray().SequenceEqual(asc.BranchSizes));
+
+            as0 = new ActionSpec(3);
+            as1 = new ActionSpec(1);
+            asc = ActionSpec.Combine(as0, as1);
+            Assert.IsEmpty(asc.BranchSizes);
+        }
+    }
+}

com.unity.ml-agents/Tests/Editor/Actuators/ActionSpecTests.cs.meta

+fileFormatVersion: 2
+guid: 99d76ec04c944b75bc6b85abfff4ac4e
+timeCreated: 1613680505

com.unity.ml-agents/Tests/Editor/MultiAgentGroupTests.cs

+using Unity.MLAgents;
+using System;
+using System.Reflection;
+using NUnit.Framework;
+using UnityEngine;
+using Unity;
+
+namespace Unity.MLAgents.Tests
+{
+    public class MultiAgentGroupTests
+    {
+        class TestAgent : Agent
+        {
+            internal int _GroupId
+            {
+                get
+                {
+                    return (int)typeof(Agent).GetField("m_GroupId", BindingFlags.Instance | BindingFlags.NonPublic).GetValue(this);
+                }
+            }
+
+            internal float _GroupReward
+            {
+                get
+                {
+                    return (float)typeof(Agent).GetField("m_GroupReward", BindingFlags.Instance | BindingFlags.NonPublic).GetValue(this);
+                }
+            }
+
+            internal Action<Agent> _OnAgentDisabledActions
+            {
+                get
+                {
+                    return (Action<Agent>)typeof(Agent).GetField("OnAgentDisabled", BindingFlags.Instance | BindingFlags.NonPublic).GetValue(this);
+                }
+            }
+        }
+
+        [Test]
+        public void TestRegisteredAgentGroupId()
+        {
+            var agentGo = new GameObject("TestAgent");
+            agentGo.AddComponent<TestAgent>();
+            var agent = agentGo.GetComponent<TestAgent>();
+
+            // test register
+            SimpleMultiAgentGroup agentGroup1 = new SimpleMultiAgentGroup();
+            agentGroup1.RegisterAgent(agent);
+            Assert.AreEqual(agentGroup1.GetId(), agent._GroupId);
+            Assert.IsNotNull(agent._OnAgentDisabledActions);
+
+            // should not be able to registered to multiple groups
+            SimpleMultiAgentGroup agentGroup2 = new SimpleMultiAgentGroup();
+            Assert.Throws<UnityAgentsException>(
+                () => agentGroup2.RegisterAgent(agent));
+            Assert.AreEqual(agentGroup1.GetId(), agent._GroupId);
+
+            // test unregister
+            agentGroup1.UnregisterAgent(agent);
+            Assert.AreEqual(0, agent._GroupId);
+            Assert.IsNull(agent._OnAgentDisabledActions);
+
+            // test register to another group after unregister
+            agentGroup2.RegisterAgent(agent);
+            Assert.AreEqual(agentGroup2.GetId(), agent._GroupId);
+            Assert.IsNotNull(agent._OnAgentDisabledActions);
+        }
+
+        [Test]
+        public void TestRegisterMultipleAgent()
+        {
+            var agentGo1 = new GameObject("TestAgent");
+            agentGo1.AddComponent<TestAgent>();
+            var agent1 = agentGo1.GetComponent<TestAgent>();
+            var agentGo2 = new GameObject("TestAgent");
+            agentGo2.AddComponent<TestAgent>();
+            var agent2 = agentGo2.GetComponent<TestAgent>();
+
+            SimpleMultiAgentGroup agentGroup = new SimpleMultiAgentGroup();
+            agentGroup.RegisterAgent(agent1); // register
+            Assert.AreEqual(agentGroup.GetRegisteredAgents().Count, 1);
+            agentGroup.UnregisterAgent(agent2); // unregister non-member agent
+            Assert.AreEqual(agentGroup.GetRegisteredAgents().Count, 1);
+            agentGroup.UnregisterAgent(agent1); // unregister
+            Assert.AreEqual(agentGroup.GetRegisteredAgents().Count, 0);
+            agentGroup.RegisterAgent(agent1);
+            agentGroup.RegisterAgent(agent1); // duplicated register
+            Assert.AreEqual(agentGroup.GetRegisteredAgents().Count, 1);
+            agentGroup.RegisterAgent(agent2); // register another
+            Assert.AreEqual(agentGroup.GetRegisteredAgents().Count, 2);
+
+            // test add/set group rewards
+            agentGroup.AddGroupReward(0.1f);
+            Assert.AreEqual(0.1f, agent1._GroupReward);
+            agentGroup.AddGroupReward(0.5f);
+            Assert.AreEqual(0.6f, agent1._GroupReward);
+            agentGroup.SetGroupReward(0.3f);
+            Assert.AreEqual(0.3f, agent1._GroupReward);
+            // unregistered agent should not receive group reward
+            agentGroup.UnregisterAgent(agent1);
+            agentGroup.AddGroupReward(0.2f);
+            Assert.AreEqual(0.3f, agent1._GroupReward);
+            Assert.AreEqual(0.5f, agent2._GroupReward);
+
+            // dispose group should automatically unregister all
+            agentGroup.Dispose();
+            Assert.AreEqual(0, agent1._GroupId);
+            Assert.AreEqual(0, agent2._GroupId);
+        }
+
+        [Test]
+        public void TestGroupIdCounter()
+        {
+            SimpleMultiAgentGroup group1 = new SimpleMultiAgentGroup();
+            SimpleMultiAgentGroup group2 = new SimpleMultiAgentGroup();
+            // id should be unique
+            Assert.AreNotEqual(group1.GetId(), group2.GetId());
+        }
+    }
+}

com.unity.ml-agents/Tests/Editor/MultiAgentGroupTests.cs.meta

+fileFormatVersion: 2
+guid: ef0158fde748d478ca5ee3bbe22a4c9e
+MonoImporter:
+  externalObjects: {}
+  serializedVersion: 2
+  defaultReferences: []
+  executionOrder: 0
+  icon: {instanceID: 0}
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

DevProject/ProjectSettings/Packages/com.unity.testtools.codecoverage/Settings.json

+{
+    "m_Name": "Settings",
+    "m_Path": "ProjectSettings/Packages/com.unity.testtools.codecoverage/Settings.json",
+    "m_Dictionary": {
+        "m_DictionaryValues": []
+    }
+}

ML-Agents-Input-Example/.gitignore

+/[Ll]ibrary/
+/Logs/
+/[Tt]emp/
+/[Oo]bj/
+/[Bb]uild/
+/[Bb]uilds/
+/Assets/AssetStoreTools*
+/Assets/Plugins*
+/Assets/Demonstrations*
+/Assets/ML-Agents/Timers*
+/*_timers.json
+
+# Environemnt logfile
+*Project.log
+
+# Visual Studio 2015 cache directory
+/.vs/
+
+# Autogenerated VS/MD/Consulo solution and project files
+/ProjectExportedObj/
+/Project.consulo/
+*.csproj
+*.unityproj
+*.sln
+*.suo
+*.tmp
+*.user
+*.userprefs
+*.pidb
+*.booproj
+*.svd
+*.pdb
+
+# Unity3D generated meta files
+*.pidb.meta
+
+# Unity3D Generated File On Crash Reports
+/sysinfo.txt
+
+# Builds
+*.apk
+*.unitypackage
+*.app
+*.exe
+*.x86_64
+*.x86
+
+# Plugins
+/Assets/VideoRecorder*
+
+# Mac hidden files
+*.DS_Store
+*/.ipynb_checkpoints
+*/.idea
+*.pyc
+*.idea/misc.xml
+*.idea/modules.xml
+*.idea/
+*.iml
+*.cache
+*/build/
+*/dist/
+*.egg-info*
+*.eggs*
+*.gitignore.swp
+
+# VSCode hidden files
+*.vscode/
+
+.DS_Store

ML-Agents-Input-Example/Assets/InputSystem.inputsettings.asset

+%YAML 1.1
+%TAG !u! tag:unity3d.com,2011:
+--- !u!114 &11400000
+MonoBehaviour:
+  m_ObjectHideFlags: 0
+  m_CorrespondingSourceObject: {fileID: 0}
+  m_PrefabInstance: {fileID: 0}
+  m_PrefabAsset: {fileID: 0}
+  m_GameObject: {fileID: 0}
+  m_Enabled: 1
+  m_EditorHideFlags: 0
+  m_Script: {fileID: 11500000, guid: c46f07b5ed07e4e92aa78254188d3d10, type: 3}
+  m_Name: InputSystem.inputsettings
+  m_EditorClassIdentifier: 
+  m_SupportedDevices:
+  - Keyboard
+  - Gamepad
+  m_UpdateMode: 2
+  m_CompensateForScreenOrientation: 0
+  m_FilterNoiseOnCurrent: 0
+  m_DefaultDeadzoneMin: 0.125
+  m_DefaultDeadzoneMax: 0.925
+  m_DefaultButtonPressPoint: 0.5
+  m_ButtonReleaseThreshold: 0.75
+  m_DefaultTapTime: 0.2
+  m_DefaultSlowTapTime: 0.5
+  m_DefaultHoldTime: 0.4
+  m_TapRadius: 5
+  m_MultiTapDelayTime: 0.75
+  m_iOSSettings:
+    m_MotionUsage:
+      m_Enabled: 0
+      m_Description: 

ML-Agents-Input-Example/Assets/InputSystem.inputsettings.asset.meta

+fileFormatVersion: 2
+guid: 778a63fac82734ee1b70a0b7c3e975c0
+NativeFormatImporter:
+  externalObjects: {}
+  mainObjectFileID: 0
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

ML-Agents-Input-Example/Assets/ML-Agents.meta

+fileFormatVersion: 2
+guid: 11630fa83cc8b4194b94352e3e6cdb9d
+folderAsset: yes
+DefaultImporter:
+  externalObjects: {}
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

ML-Agents-Input-Example/Assets/ML-Agents/Examples.meta

+fileFormatVersion: 2
+guid: 6584f096f53dc43eeb32803b91f36c5c
+folderAsset: yes
+DefaultImporter:
+  externalObjects: {}
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

ML-Agents-Input-Example/Assets/ML-Agents/Examples/PushBlock.meta

+fileFormatVersion: 2
+guid: 4197ece8ca6d74a00adc6bafbabda158
+folderAsset: yes
+timeCreated: 1506303336
+licenseType: Pro
+DefaultImporter:
+  externalObjects: {}
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

ML-Agents-Input-Example/Assets/ML-Agents/Examples/PushBlock/Meshes.meta

+fileFormatVersion: 2
+guid: e806d68316eff0e46a3fda372ae42c44
+folderAsset: yes
+DefaultImporter:
+  externalObjects: {}
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

ML-Agents-Input-Example/Assets/ML-Agents/Examples/PushBlock/Meshes/PushBlockCourt.fbx.meta

+fileFormatVersion: 2
+guid: c639386c12f5f7841892163a199dfacc
+ModelImporter:
+  serializedVersion: 22
+  fileIDToRecycleName:
+    100000: GoalArea
+    100002: Ground
+    100004: //RootNode
+    100006: WallsOuter
+    400000: GoalArea
+    400002: Ground
+    400004: //RootNode
+    400006: WallsOuter
+    2100000: rep_WhiteWalls
+    2100002: rep_Floor
+    2100004: rep_Checkers
+    2300000: GoalArea
+    2300002: Ground
+    2300004: WallsOuter
+    3300000: GoalArea
+    3300002: Ground
+    3300004: WallsOuter
+    4300000: WallsOuter
+    4300002: Ground
+    4300004: GoalArea
+  externalObjects:
+  - first:
+      type: UnityEngine:Material
+      assembly: UnityEngine.CoreModule
+      name: rep_Checkers
+    second: {fileID: 2100000, guid: 36c7baa347d68f347a9aa9698aa1bcdd, type: 2}
+  - first:
+      type: UnityEngine:Material
+      assembly: UnityEngine.CoreModule
+      name: rep_Floor
+    second: {fileID: 2100000, guid: bc723809e6ff3174fad3e774cae1aed0, type: 2}
+  - first:
+      type: UnityEngine:Material
+      assembly: UnityEngine.CoreModule
+      name: rep_WhiteWalls
+    second: {fileID: 2100000, guid: 6a39c0407dd85684384bf0277294e9b6, type: 2}
+  materials:
+    importMaterials: 1
+    materialName: 0
+    materialSearch: 1
+    materialLocation: 1
+  animations:
+    legacyGenerateAnimations: 4
+    bakeSimulation: 0
+    resampleCurves: 1
+    optimizeGameObjects: 0
+    motionNodeName: 
+    rigImportErrors: 
+    rigImportWarnings: 
+    animationImportErrors: 
+    animationImportWarnings: 
+    animationRetargetingWarnings: 
+    animationDoRetargetingWarnings: 0
+    importAnimatedCustomProperties: 0
+    animationCompression: 1
+    animationRotationError: 0.5
+    animationPositionError: 0.5
+    animationScaleError: 0.5
+    animationWrapMode: 0
+    extraExposedTransformPaths: []
+    extraUserProperties: []
+    clipAnimations: []
+    isReadable: 1
+  meshes:
+    lODScreenPercentages: []
+    globalScale: 1
+    meshCompression: 0
+    addColliders: 0
+    importVisibility: 1
+    importBlendShapes: 1
+    importCameras: 1
+    importLights: 1
+    swapUVChannels: 0
+    generateSecondaryUV: 1
+    useFileUnits: 1
+    optimizeMeshForGPU: 1
+    keepQuads: 0
+    weldVertices: 1
+    preserveHierarchy: 0
+    indexFormat: 0
+    secondaryUVAngleDistortion: 8
+    secondaryUVAreaDistortion: 15.000001
+    secondaryUVHardAngle: 88
+    secondaryUVPackMargin: 4
+    useFileScale: 1
+  tangentSpace:
+    normalSmoothAngle: 60
+    normalImportMode: 0
+    tangentImportMode: 3
+    normalCalculationMode: 4
+  importAnimation: 1
+  copyAvatar: 0
+  humanDescription:
+    serializedVersion: 2
+    human: []
+    skeleton: []
+    armTwist: 0.5
+    foreArmTwist: 0.5
+    upperLegTwist: 0.5
+    legTwist: 0.5
+    armStretch: 0.05
+    legStretch: 0.05
+    feetSpacing: 0
+    rootMotionBoneName: 
+    rootMotionBoneRotation: {x: 0, y: 0, z: 0, w: 1}
+    hasTranslationDoF: 0
+    hasExtraRoot: 0
+    skeletonHasParents: 1
+  lastHumanDescriptionAvatarSource: {instanceID: 0}
+  animationType: 0
+  humanoidOversampling: 1
+  additionalBone: 0
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

ML-Agents-Input-Example/Assets/ML-Agents/Examples/PushBlock/Prefabs.meta

+fileFormatVersion: 2
+guid: 3d2b9d5547d934200a786212743850c4
+folderAsset: yes
+timeCreated: 1514922259
+licenseType: Free
+DefaultImporter:
+  externalObjects: {}
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: 

ML-Agents-Input-Example/Assets/ML-Agents/Examples/PushBlock/Prefabs/PushBlockArea.prefab.meta

+fileFormatVersion: 2
+guid: 03bcc81e249714a22bb411dddcc5d15e
+timeCreated: 1515023875
+licenseType: Free
+NativeFormatImporter:
+  externalObjects: {}
+  mainObjectFileID: 100100000
+  userData: 
+  assetBundleName: 
+  assetBundleVariant: