Task Generation

In this section, we describe how to generate a task in LIBERO. The procedure consists of three steps: 1) Specify a scene from the Scene Creation step; 2) Specify initial state distributions; 3) Specify goals and language instructions.

An example

We present an example of creating a task in LIBERO for intuitive understanding.

The creation of a task given a pre-defined scene can be easily created using a function. A complete example of generating a task can be found in the provided notebook

Layout and Initial State Distributions

Here is an example t odefine the layout, objects, and the initial state dsistributions.

@register_mu(scene_type="kitchen")
class KitchenSceneExample(InitialSceneTemplates):
    def __init__(self):
        fixture_num_info = {
            "kitchen_table": 1,
            "wooden_cabinet": 1,
        }
        object_num_info = {
            "akita_black_bowl": 1,
            "plate": 1,
        }

        super().__init__(
            workspace_name="kitchen_table",
            fixture_num_info=fixture_num_info,
            object_num_info=object_num_info
        )
    def define_regions(self):
        self.regions.update(
            self.get_region_dict(region_centroid_xy=[0.0, -0.30], 
                                 region_name="wooden_cabinet_init_region", 
                                 target_name=self.workspace_name, 
                                 region_half_len=0.01,
                                 yaw_rotation=(np.pi, np.pi))
        )
        self.regions.update(
            self.get_region_dict(region_centroid_xy=[0., 0.0], 
                                 region_name="akita_black_bowl_init_region", 
                                 target_name=self.workspace_name, 
                                 region_half_len=0.025)
        )

        self.regions.update(
            self.get_region_dict(region_centroid_xy=[0.0, 0.25], 
                                 region_name="plate_init_region", 
                                 target_name=self.workspace_name, 
                                 region_half_len=0.025)
        )
        self.xy_region_kwargs_list = get_xy_region_kwargs_list_from_regions_info(self.regions)

    @property
    def init_states(self):
        states = [
            ("On", "akita_black_bowl_1", "kitchen_table_akita_black_bowl_init_region"),
            ("On", "plate_1", "kitchen_table_plate_init_region"),
            ("On", "wooden_cabinet_1", "kitchen_table_wooden_cabinet_init_region")]
        return states

Goal and Language

Once we define the initial state distributions, we can define the task goal and the language specifications. Here is an example how to define them. Notice that objects_of_interest is an optional field, in case you want to modify the LIBERO problem classes to keep track of the internal simulation states of some certrain objects.

# kitchen_scene_example
scene_name = "kitchen_scene_example"
language = "open the top drawer of the cabinet and put the bowl in it"
register_task_info(language,
                    scene_name=scene_name,
                    objects_of_interest=["wooden_cabinet_1", "akita_black_bowl_1"],
                    goal_states=[("Open", "wooden_cabinet_1_top_region"), ("In", "akita_black_bowl_1", "wooden_cabinet_1_top_region")]
)