Make ORI_IND optional and adjust the size of maze for swimmer

mujoco_maze/agent_model.py

@@ -11,7 +11,7 @@ from gym.utils import EzPickle
 class AgentModel(ABC, MujocoEnv, EzPickle):
     FILE: str
     MANUAL_COLLISION: bool
-    ORI_IND: int
+    ORI_IND: Optional[int] = None
     RADIUS: Optional[float] = None
 
     def __init__(self, file_path: str, frame_skip: int) -> None:

mujoco_maze/maze_task.py

@@ -331,7 +331,7 @@ class SubGoalTRoom(GoalRewardTRoom):
 
 class NoRewardRoom(MazeTask):
     REWARD_THRESHOLD: float = 0.0
-    MAZE_SIZE_SCALING: Scaling = Scaling(4.0, 4.0, 4.0)
+    MAZE_SIZE_SCALING: Scaling = Scaling(4.0, 4.0, 1.0)
 
     def reward(self, obs: np.ndarray) -> float:
         return 0.0

mujoco_maze/swimmer.py

@@ -37,7 +37,6 @@ class SwimmerEnv(AgentModel):
     def step(self, action: np.ndarray) -> Tuple[np.ndarray, float, bool, dict]:
         xy_pos_before = self.sim.data.qpos[:2].copy()
         self.do_simulation(action, self.frame_skip)
-
         forward_reward = self._forward_reward(xy_pos_before)
         ctrl_cost = self._ctrl_cost_weight * np.sum(np.square(action))
         return (