May be None for completely random seeding. Sets the random seeds for all environments, based on a given seed.Įach individual environment will still get its own seed, by incrementing the given seed. Observation abstract seed ( seed = None ) ¶ If step_async is still doing work, that work willīe cancelled and step_wait() should not be called
Observations, or a tuple of observation arrays. Reset all the environments and return an array of Mode ( str) – the rendering type Return type Name of module whose attribute is being shadowed, if any. Name ( str) – name of attribute to check forĪlready_found ( bool) – whether this attribute has already been found in a wrapper Sequence getattr_depth_check ( name, already_found ) ¶Ĭheck if an attribute reference is being hidden in a recursive call to _getattr_ Parameters Return RGB images from each environment Return type List of values of ‘attr_name’ in all environments get_images ( ) ¶ Indices ( Union]) – Indices of envs to get attribute from ParametersĪttr_name ( str) – The name of the attribute whose value to return Return attribute from vectorized environment. List of items returned by the environment’s method call abstract get_attr ( attr_name, indices = None ) ¶ Method_name ( str) – The name of the environment method to invoke. abstract env_method ( method_name, * method_args, indices = None, ** method_kwargs ) ¶Ĭall instance methods of vectorized environments. True if the env is wrapped, False otherwise, for each env queried. Method_kwargs – Any keyword arguments to provide in the call Method_args – Any positional arguments to provide in the call Indices ( Union]) – Indices of envs whose method to call Method_name – The name of the environment method to invoke. None abstract env_is_wrapped ( wrapper_class, indices = None ) ¶Ĭheck if environments are wrapped with a given wrapper. Observation_space ( Space) – the observation spaceĬlean up the environment’s resources. Num_envs ( int) – the number of environments VecEnv ( num_envs, observation_space, action_space ) ¶Īn abstract asynchronous, vectorized environment. step_wait () return obs, reward, done, info env = DummyVecEnv () # Wrap the VecEnv env = VecExtractDictObs ( env, key = "observation" ) VecEnv ¶ class stable_env. step_async ( actions ) def step_wait ( self ) -> VecEnvStepReturn : obs, reward, done, info = self. reset () return obs def step_async ( self, actions : np. _init_ ( venv = venv, observation_space = venv. Similar to Gym's FilterObservation wrapper: :param venv: The vectorized environment :param key: The key of the dictionary observation """ def _init_ ( self, venv : VecEnv, key : str ): self. Import numpy as np from stable_env.base_vec_env import VecEnv, VecEnvStepReturn, VecEnvWrapper class VecExtractDictObs ( VecEnvWrapper ): """ A vectorized wrapper for filtering a specific key from dictionary observations.
0 kommentar(er)
