opentau.envs.factory

This module contains factory methods to create environments based on their configuration.

Functions

make_env_config(env_type, **kwargs)

Factory method to create an environment config based on the env_type.

make_envs(cfg, train_cfg[, n_envs, ...])

Makes a nested collection of gym vector environment according to the config.
opentau.envs.factory.make_env_config(env_type: str, **kwargs) → EnvConfig[source]

Factory method to create an environment config based on the env_type. Supports ‘libero’ and ‘robocasa’.

opentau.envs.factory.make_envs(cfg: EnvConfig, train_cfg: TrainPipelineConfig, n_envs: int = 1, use_async_envs: bool = False) → dict[str, dict[int, VectorEnv]][source]

Makes a nested collection of gym vector environment according to the config.

Parameters:

  • cfg (EnvConfig) – the config of the environment to instantiate.

  • n_envs (int, optional) – The number of parallelized env to return. Defaults to 1.

  • use_async_envs (bool, optional) – Whether to return an AsyncVectorEnv or a SyncVectorEnv. Defaults to False.

Raises:

  • ValueError – if n_envs < 1

  • ModuleNotFoundError – If the requested env package is not installed

Returns:

A mapping from suite name to indexed vectorized environments. - For multi-task benchmarks (e.g., LIBERO): one entry per suite, and one vec env per task_id. - For single-task environments: a single suite entry (cfg.type) with task_id=0.

Return type:

dict[str, dict[int, gym.vector.VectorEnv]]