Removing the Reward and Message Observation Type, will readd when used

com.unity.ml-agents/Runtime/Sensors/ITypedSensor.cs

         Default = 0,
         // Collected observations contain goal information.
         Goal = 1,
-        // Collected observations contain reward information.
-        Reward = 2,
-        // Collected observations are messages from other agents.
-        Message = 3,
     }
 
 

ml-agents-envs/mlagents_envs/base_env.py

     DEFAULT = 0
     # Observation contains goal information for current task.
     GOAL = 1
-    # Observation contains reward information for current task.
-    REWARD = 2
-    # Observation contains a message from another agent.
-    MESSAGE = 3
 
 
 class ObservationSpec(NamedTuple):