hSrSSKJr SSKJr SSKrSSKrSSK J r SSK J r SSK JrJrJr SSKJr S /r"S S \\R(R*5rg) zhA collection of stateful observation wrappers. * ``NormalizeObservation`` - Normalize the observations ) annotations)AnyN)ObsType)warn) AutoresetMode VectorEnvVectorObservationWrapper)RunningMeanStdNormalizeObservationc^\rSrSrSrS S Sjjr\S Sj5r\RSSj5rSSS.SU4Sjjjr SS jr S r U=r $)r aThis wrapper will normalize observations s.t. each coordinate is centered with unit variance. The property `_update_running_mean` allows to freeze/continue the running mean calculation of the observation statistics. If `True` (default), the `RunningMeanStd` will get updated every step and reset call. If `False`, the calculated statistics are used but not updated anymore; this may be used during evaluation. Note: The normalization depends on past trajectories and observations will not be normalized correctly if the wrapper was newly instantiated or the policy was changed recently. Example without the normalize reward wrapper: >>> import gymnasium as gym >>> envs = gym.make_vec("CartPole-v1", num_envs=3, vectorization_mode="sync") >>> obs, info = envs.reset(seed=123) >>> _ = envs.action_space.seed(123) >>> for _ in range(100): ... obs, *_ = envs.step(envs.action_space.sample()) >>> np.mean(obs) np.float32(0.024251968) >>> np.std(obs) np.float32(0.62259156) >>> envs.close() Example with the normalize reward wrapper: >>> import gymnasium as gym >>> envs = gym.make_vec("CartPole-v1", num_envs=3, vectorization_mode="sync") >>> envs = NormalizeObservation(envs) >>> obs, info = envs.reset(seed=123) >>> _ = envs.action_space.seed(123) >>> for _ in range(100): ... obs, *_ = envs.step(envs.action_space.sample()) >>> np.mean(obs) np.float32(-0.2359734) >>> np.std(obs) np.float32(1.1938739) >>> envs.close() c[RRRXS9 [R"X5 SUR R ;a[US35 O.UR R S[R1;de[URRURRS9UlX lSUlg)zThis wrapper will normalize observations s.t. each coordinate is centered with unit variance. Args: env (Env): The environment to apply the wrapper epsilon: A stability parameter that is used when scaling the observations. )epsilonautoreset_modez is missing `autoreset_mode` data. Assuming that the vector environment it follows the `NextStep` autoreset api or autoreset is disabled. Read https://farama.org/Vector-Autoreset-Mode for more details.)shapedtypeTN)gymutilsRecordConstructorArgs__init__r envmetadatarr NEXT_STEPr single_observation_spacerrobs_rmsr_update_running_mean)selfrrs h/home/james-whalen/.local/lib/python3.13/site-packages/gymnasium/wrappers/vector/stateful_observation.pyrNormalizeObservation.__init__As ''000G ))$4 488#4#4 4 &bc 88$$%56=;R;R:SS SS%//55//55   $(!cUR$)zWProperty to freeze/continue the running mean calculation of the observation statistics.r)rs rupdate_running_mean(NormalizeObservation.update_running_meanYs(((r cXlg)z`Sets the property to freeze/continue the running mean calculation of the observation statistics.Nr")rsettings rr#r$^s %,!r Nseedoptionscp>Ub&SU;d [R"US5(de[TU] XS9$)zVReset function for `NormalizeObservationWrapper` which is disabled for partial resets. reset_maskr')npallsuperreset)rr(r) __class__s rr/NormalizeObservation.resetcsC O7*vvgl+,,  -w}$}88r cUR(aURRU5 XRR- [R "URR UR-5- $)zDefines the vector observation normalization function. Args: observations: A vector observation from the environment Returns: the normalized observation )rrupdatemeanr,sqrtvarr)r observationss rr7!NormalizeObservation.observationsqsZ  $ $ LL   -||000BGG LL  t|| +5   r )rrr)g:0yE>)rrrfloat)returnbool)r&r;)r(zint | list[int] | Noner)zdict[str, Any] | Noner:ztuple[ObsType, dict[str, Any]])r7rr:r) __name__ __module__ __qualname____firstlineno____doc__rpropertyr#setterr/r7__static_attributes__ __classcell__)r0s@rr r sv$L)0)), ,(,)- 9% 9' 9 ( 9 9   r )r@ __future__rtypingrnumpyr, gymnasiumrgymnasium.corergymnasium.loggerrgymnasium.vector.vector_envrrr gymnasium.wrappers.utilsr __all__rrr r rrOsR #"! 4 " "d 3SYY5T5Td r