42 lines
1.3 KiB
Python
42 lines
1.3 KiB
Python
import warnings
|
|
|
|
import gym
|
|
|
|
from tianshou.env import ShmemVectorEnv, VectorEnvNormObs
|
|
|
|
try:
|
|
import envpool
|
|
except ImportError:
|
|
envpool = None
|
|
|
|
|
|
def make_mujoco_env(task, seed, training_num, test_num, obs_norm):
|
|
"""Wrapper function for Mujoco env.
|
|
|
|
If EnvPool is installed, it will automatically switch to EnvPool's Mujoco env.
|
|
|
|
:return: a tuple of (single env, training envs, test envs).
|
|
"""
|
|
if envpool is not None:
|
|
train_envs = env = envpool.make_gym(task, num_envs=training_num, seed=seed)
|
|
test_envs = envpool.make_gym(task, num_envs=test_num, seed=seed)
|
|
else:
|
|
warnings.warn(
|
|
"Recommend using envpool (pip install envpool) "
|
|
"to run Mujoco environments more efficiently."
|
|
)
|
|
env = gym.make(task)
|
|
train_envs = ShmemVectorEnv(
|
|
[lambda: gym.make(task) for _ in range(training_num)]
|
|
)
|
|
test_envs = ShmemVectorEnv([lambda: gym.make(task) for _ in range(test_num)])
|
|
env.seed(seed)
|
|
train_envs.seed(seed)
|
|
test_envs.seed(seed)
|
|
if obs_norm:
|
|
# obs norm wrapper
|
|
train_envs = VectorEnvNormObs(train_envs)
|
|
test_envs = VectorEnvNormObs(test_envs, update_obs_rms=False)
|
|
test_envs.set_obs_rms(train_envs.get_obs_rms())
|
|
return env, train_envs, test_envs
|