small cleanup

Project/Assets/ML-Agents/Examples/Hallway/Scripts/HallwayCollabAgent.cs

     public bool isSpotter = true;
     int m_Message = 0;
 
+
-        m_HallwaySettings = FindObjectOfType<HallwaySettings>();
-        m_AgentRb = GetComponent<Rigidbody>();
-        m_GroundRenderer = ground.GetComponent<Renderer>();
-        m_GroundMaterial = m_GroundRenderer.material;
+        base.Initialize();
         if (isSpotter)
         {
             var teamManager = new HallwayTeamManager();

Project/Assets/ML-Agents/Examples/Hallway/Scripts/HallwayTeamManager.cs

 using System.Collections.Generic;
-using UnityEngine;
-using Unity.MLAgents.Actuators;
 using Unity.MLAgents.Sensors;
 
 public class HallwayTeamManager : BaseTeamManager
 
     public override void OnAgentDone(Agent agent, Agent.DoneReason doneReason, List<ISensor> sensors)
     {
-        // Possible implementation - save reference to Agent's IPolicy so that we can repeatedly
-        // call IPolicy.RequestDecision on behalf of the Agent after it's dead
-        // If so, we'll need dummy sensor impls with the same shape as the originals.
         agent.SendDoneToTrainer();
     }
 

com.unity.ml-agents.extensions/Runtime/Teams/BaseTeamManager.cs

 {
     public class BaseTeamManager : ITeamManager
     {
-        string m_Id = System.Guid.NewGuid().ToString();
+        readonly string m_Id = System.Guid.NewGuid().ToString();
+
         public virtual void RegisterAgent(Agent agent)
         {
             throw new System.NotImplementedException();

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

             {
                 SendDoneToTrainer();
             }
-            ResetSensors();
 
             // We also have to write any to any DemonstationStores so that they get the "done" flag.
             foreach (var demoWriter in DemonstrationWriters)
         {
             // We request a decision so Python knows the Agent is done immediately
             m_Brain?.RequestDecision(m_Info, sensors);
+            ResetSensors();
         }
 
         /// <summary>

ml-agents/mlagents/trainers/agent_processor.py

         self.experience_buffers: Dict[str, List[AgentExperience]] = defaultdict(list)
         self.last_experience: Dict[str, AgentExperience] = {}
         self.last_step_result: Dict[str, Tuple[DecisionStep, int]] = {}
-        # current_group_obs is used to collect the last seen obs of all the agents, and assemble the next_collab_obs.
+        # current_group_obs is used to collect the last seen obs of all the agents in the same group,
+        # and assemble the next_collab_obs.
         self.current_group_obs: Dict[str, Dict[str, List[np.ndarray]]] = defaultdict(
             lambda: defaultdict(list)
         )
                 next_obs = step.obs
                 next_collab_obs = []
                 for _id, _exp in self.current_group_obs[step.team_manager_id].items():
-                    if _id == global_id:
-                        continue
-                    else:
+                    if _id != global_id:
                         next_collab_obs.append(_exp)
 
                 trajectory = Trajectory(