Retrain Walker (#5911)

* reduce hidden nodes to 256 and retrain 30M steps

* update demo file

Author: Hunter-Unity

Project/Assets/ML-Agents/Examples/Walker/Demos/ExpertWalker.demo

@@ -136,8 +136,13 @@ PrefabInstance:
         type: 3}
       propertyPath: m_Model
       value: 
-      objectReference: {fileID: 5022602860645237092, guid: 71fe1fe4f177841a1b4b85987277495c,
+      objectReference: {fileID: 5022602860645237092, guid: b000e8fcd1edc454d9e6deb84abefbf9,
         type: 3}
+    - target: {fileID: 3599283675615088294, guid: dd62a2a5a838943fba3d2cbb377a75b3,
+        type: 3}
+      propertyPath: m_DeterministicInference
+      value: 0
+      objectReference: {fileID: 0}
     - target: {fileID: 3599283675615088295, guid: dd62a2a5a838943fba3d2cbb377a75b3,
         type: 3}
       propertyPath: target

Project/Assets/ML-Agents/Examples/Walker/Demos/ExpertWalker.demo.meta

@@ -84,8 +84,6 @@ public override void Initialize()
         m_JdController.SetupBodyPart(handR);
 
         m_ResetParams = Academy.Instance.EnvironmentParameters;
-
-        SetResetParameters();
     }
 
     /// <summary>
@@ -107,8 +105,6 @@ public override void OnEpisodeBegin()
         //Set our goal walking speed
         MTargetWalkingSpeed =
             randomizeWalkSpeedEachEpisode ? Random.Range(0.1f, m_maxWalkingSpeed) : MTargetWalkingSpeed;
-
-        SetResetParameters();
     }
 
     /// <summary>
@@ -240,7 +236,10 @@ void FixedUpdate()
 
         // b. Rotation alignment with target direction.
         //This reward will approach 1 if it faces the target direction perfectly and approach zero as it deviates
-        var lookAtTargetReward = (Vector3.Dot(cubeForward, head.forward) + 1) * .5F;
+        var headForward = head.forward;
+        headForward.y = 0;
+        // var lookAtTargetReward = (Vector3.Dot(cubeForward, head.forward) + 1) * .5F;
+        var lookAtTargetReward = (Vector3.Dot(cubeForward, headForward) + 1) * .5F;
 
         //Check for NaNs
         if (float.IsNaN(lookAtTargetReward))
@@ -292,16 +291,4 @@ public void TouchedTarget()
     {
         AddReward(1f);
     }
-
-    public void SetTorsoMass()
-    {
-        m_JdController.bodyPartsDict[chest].rb.mass = m_ResetParams.GetWithDefault("chest_mass", 8);
-        m_JdController.bodyPartsDict[spine].rb.mass = m_ResetParams.GetWithDefault("spine_mass", 8);
-        m_JdController.bodyPartsDict[hips].rb.mass = m_ResetParams.GetWithDefault("hip_mass", 8);
-    }
-
-    public void SetResetParameters()
-    {
-        SetTorsoMass();
-    }
 }

Project/Assets/ML-Agents/Examples/Walker/Prefabs/Platforms/Platform.prefab

@@ -12,7 +12,7 @@ behaviors:
       learning_rate_schedule: linear
     network_settings:
       normalize: true
-      hidden_units: 512
+      hidden_units: 256
       num_layers: 3
       vis_encode_type: simple
     reward_signals: