[Doc] RLlib Algorithms Documentation: MAML + PyTorch MAML (#9189)

rllib/examples/env/ant_rand_goal.py

@@ -1,9 +1,10 @@
 import numpy as np
 import gym
 from gym.envs.mujoco.mujoco_env import MujocoEnv
+from ray.rllib.env.meta_env import MetaEnv
 
 
-class AntRandGoalEnv(gym.utils.EzPickle, MujocoEnv):
+class AntRandGoalEnv(gym.utils.EzPickle, MujocoEnv, MetaEnv):
     """Ant Environment that randomizes goals as tasks
 
     Goals are randomly sampled 2D positions

rllib/examples/env/halfcheetah_rand_direc.py

@@ -1,9 +1,10 @@
 import numpy as np
 import gym
 from gym.envs.mujoco.mujoco_env import MujocoEnv
+from ray.rllib.env.meta_env import MetaEnv
 
 
-class HalfCheetahRandDirecEnv(MujocoEnv, gym.utils.EzPickle):
+class HalfCheetahRandDirecEnv(MujocoEnv, gym.utils.EzPickle, MetaEnv):
     """HalfCheetah Environment with two diff tasks, moving forwards or backwards
 
     Direction is defined as a scalar: +1.0 (forwards) or -1.0 (backwards)

rllib/examples/env/pendulum_mass.py

@@ -1,9 +1,10 @@
 import numpy as np
 import gym
 from gym.envs.classic_control.pendulum import PendulumEnv
+from ray.rllib.env.meta_env import MetaEnv
 
 
-class PendulumMassEnv(PendulumEnv, gym.utils.EzPickle):
+class PendulumMassEnv(PendulumEnv, gym.utils.EzPickle, MetaEnv):
     """PendulumMassEnv varies the weight of the pendulum
 
     Tasks are defined to be weight uniformly sampled between [0.5,2]