mjlab.tasks

mjlab.tasks#

Task registry system for managing environment registration and creation.

Functions

mjlab.tasks.registry.register_mjlab_task(task_id: str, env_cfg: ManagerBasedRlEnvCfg, play_env_cfg: ManagerBasedRlEnvCfg, rl_cfg: RslRlOnPolicyRunnerCfg, runner_cls: type | None = None) → None[source]#

Register an environment task.

Parameters:

  • task_id – Unique task identifier (e.g., “Mjlab-Velocity-Rough-Unitree-Go1”).

  • env_cfg – Environment configuration used for training.

  • play_env_cfg – Environment configuration in “play” mode.

  • rl_cfg – RL runner configuration.

  • runner_cls – Optional custom runner class. If None, uses OnPolicyRunner.

mjlab.tasks.registry.list_tasks() → list[str][source]#

List all registered task IDs.

mjlab.tasks.registry.load_env_cfg(task_name: str, play: bool = False) → ManagerBasedRlEnvCfg[source]#

Load environment configuration for a task.

Returns a deep copy to prevent mutation of the registered config.

mjlab.tasks.registry.load_rl_cfg(task_name: str) → RslRlOnPolicyRunnerCfg[source]#

Load RL configuration for a task.

Returns a deep copy to prevent mutation of the registered config.

mjlab.tasks.registry.load_runner_cls(task_name: str) → type | None[source]#

Load the runner class for a task.

If None, the default OnPolicyRunner will be used.

Contents