.

mujoco_maze/dir_point.py

@@ -78,8 +78,7 @@ class DirPointEnv(AgentModel):
         self.data.site_xpos[-1][2] = 0.1  # Slightly higher in space
 
         next_obs = self._get_obs()
-        reward = -1*min(max(0, distance - 3), 100)*self.action_penalty / 1000
-        return next_obs, reward, False, {}
+        return next_obs, 0.0, False, {}
 
     def _calculate_maze_size(self) -> float:
         """Calculate the effective size of the maze."""