DMControlWrapper

torchrl.envs.libs.dm_control.DMControlWrapper(*args, **kwargs)

DeepMind Control lab environment wrapper.

Parameters:

• env (dm_control.suite env) – environment instance

• from_pixels (bool) – if True, the observation

Examples

>>> env = dm_control.suite.load("cheetah", "run")
>>> env = DMControlWrapper(env,
...    from_pixels=True, frame_skip=4)
>>> td = env.rand_step()
>>> print(td)
>>> print(env.available_envs)