better logging for NaN rewards

Project/Assets/ML-Agents/Examples/Crawler/Scripts/CrawlerAgent.cs

@@ -1,7 +1,9 @@
+using System;
 using UnityEngine;
 using Unity.MLAgents;
 using Unity.MLAgentsExamples;
 using Unity.MLAgents.Sensors;
+using Random = UnityEngine.Random;
 
 [RequireComponent(typeof(JointDriveController))] // Required to set joint forces
 public class CrawlerAgent : Agent
@@ -111,7 +113,7 @@ public override void CollectObservations(VectorSensor sensor)
     {
         //Add body rotation delta relative to orientation cube
         sensor.AddObservation(Quaternion.FromToRotation(body.forward, orientationCube.transform.forward));
-        
+
         //Add pos of target relative to orientation cube
         sensor.AddObservation(orientationCube.transform.InverseTransformPoint(target.transform.position));
 
@@ -211,7 +213,15 @@ void RewardFunctionMovingTowards()
     {
         var movingTowardsDot = Vector3.Dot(orientationCube.transform.forward,
             Vector3.ClampMagnitude(m_JdController.bodyPartsDict[body].rb.velocity, maximumWalkingSpeed));
-        ;
+        if (float.IsNaN(movingTowardsDot))
+        {
+            throw new ArgumentException(
+                "NaN in movingTowardsDot.\n" +
+                $" orientationCube.transform.forward: {orientationCube.transform.forward}\n"+
+                $" body.velocity: {m_JdController.bodyPartsDict[body].rb.velocity}\n"+
+                $" maximumWalkingSpeed: {maximumWalkingSpeed}"
+            );
+        }
         AddReward(0.03f * movingTowardsDot);
     }
 
@@ -220,7 +230,16 @@ void RewardFunctionMovingTowards()
     /// </summary>
     void RewardFunctionFacingTarget()
     {
-        AddReward(0.01f * Vector3.Dot(orientationCube.transform.forward, body.forward));
+        var facingReward = Vector3.Dot(orientationCube.transform.forward, body.forward);
+        if (float.IsNaN(facingReward))
+        {
+            throw new ArgumentException(
+                "NaN in movingTowardsDot.\n" +
+                $" orientationCube.transform.forward: {orientationCube.transform.forward}\n"+
+                $" body.forward: {body.forward}"
+            );
+        }
+        AddReward(0.01f * facingReward);
     }
 
     /// <summary>

Project/Assets/ML-Agents/Examples/Walker/Scripts/WalkerAgent.cs

@@ -1,9 +1,11 @@
+using System;
 using MLAgentsExamples;
 using UnityEngine;
 using Unity.MLAgents;
 using Unity.MLAgentsExamples;
 using Unity.MLAgents.Sensors;
 using BodyPart = Unity.MLAgentsExamples.BodyPart;
+using Random = UnityEngine.Random;
 
 public class WalkerAgent : Agent
 {
@@ -171,11 +173,40 @@ void FixedUpdate()
         // a. Velocity alignment with goal direction.
         var moveTowardsTargetReward = Vector3.Dot(cubeForward,
             Vector3.ClampMagnitude(m_JdController.bodyPartsDict[hips].rb.velocity, maximumWalkingSpeed));
+        if (float.IsNaN(moveTowardsTargetReward))
+        {
+            throw new ArgumentException(
+                "NaN in moveTowardsTargetReward.\n" +
+                $" cubeForward: {cubeForward}\n"+
+                $" hips.velocity: {m_JdController.bodyPartsDict[hips].rb.velocity}\n"+
+                $" maximumWalkingSpeed: {maximumWalkingSpeed}"
+            );
+        }
+
         // b. Rotation alignment with goal direction.
         var lookAtTargetReward = Vector3.Dot(cubeForward, head.forward);
+        if (float.IsNaN(lookAtTargetReward))
+        {
+            throw new ArgumentException(
+                "NaN in lookAtTargetReward.\n" +
+                $" cubeForward: {cubeForward}\n"+
+                $" head.forward: {head.forward}"
+            );
+        }
+
         // c. Encourage head height. //Should normalize to ~1
-        var headHeightOverFeetReward = 
+        var headHeightOverFeetReward =
             ((head.position.y - footL.position.y) + (head.position.y - footR.position.y) / 10);
+        if (float.IsNaN(headHeightOverFeetReward))
+        {
+            throw new ArgumentException(
+                "NaN in headHeightOverFeetReward.\n" +
+                $" head.position: {head.position}\n"+
+                $" footL.position: {footL.position}\n"+
+                $" footR.position: {footR.position}"
+            );
+        }
+
         AddReward(
             + 0.02f * moveTowardsTargetReward
             + 0.02f * lookAtTargetReward