TaskConfig class
The definition of the task in Habitat.
There are many different Tasks determined by the habitat.task.type config:
- Point navigation : Nav-0
- Image navigation : Nav-0
- Instance image navigation:InstanceImageNav-v1
- Object navigation : ObjectNav-v1
- Rearrangement close drawer: RearrangeCloseDrawerTask-v0
- Rearrangement open drawer: RearrangeOpenDrawerTask-v0
- Rearrangement close fridge : RearrangeCloseFridgeTask-v0
- Rearrangement open fridge : RearrangeOpenFridgeTask-v0
- Rearrangement navigate to object : NavToObjTask-v0
- Rearrangement pick : RearrangePickTask-v0
- Rearrangement place : RearrangePlaceTask-v0
- Rearrangement do nothing : RearrangeEmptyTask-v0
- Rearrangement reach : RearrangeReachTask-v0
- Rearrangement composite tasks : RearrangePddlTask-v0
Special methods
- def __init__(self, physics_target_sps: float = 60.0, reward_measure: typing.Optional[str] = None, success_measure: typing.Optional[str] = None, success_reward: float = 2.5, slack_reward: float = -0.01, end_on_success: bool = False, type: str = 'Nav-v0', lab_sensors: typing.Dict[str, habitat.config.default_structured_configs.LabSensorConfig] = <factory>, measurements: typing.Dict[str, habitat.config.default_structured_configs.MeasurementConfig] = <factory>, rank0_env0_measure_names: typing.List[str] = <factory>, rank0_measure_names: typing.List[str] = <factory>, goal_sensor_uuid: str = 'pointgoal', count_obj_collisions: bool = True, settle_steps: int = 5, constraint_violation_ends_episode: bool = True, constraint_violation_drops_object: bool = False, force_regenerate: bool = False, should_save_to_cache: bool = False, object_in_hand_sample_prob: float = 0.167, min_start_distance: float = 3.0, render_target: bool = True, filter_colliding_states: bool = True, num_spawn_attempts: int = 200, spawn_max_dist_to_obj: float = 2.0, base_angle_noise: float = 0.523599, spawn_max_dist_to_obj_delta: float = 0.02, recep_place_shrink_factor: float = 0.8, ee_sample_factor: float = 0.2, ee_exclude_region: float = 0.0, base_noise: float = 0.05, spawn_region_scale: float = 0.2, joint_max_impulse: float = -1.0, desired_resting_position: typing.List[float] = <factory>, use_marker_t: bool = True, cache_robot_init: bool = False, success_state: float = 0.0, should_enforce_target_within_reach: bool = False, task_spec_base_path: str = 'habitat/task/rearrange/pddl/', task_spec: str = '', pddl_domain_def: str = 'replica_cad', obj_succ_thresh: float = 0.3, enable_safe_drop: bool = False, art_succ_thresh: float = 0.15, robot_at_thresh: float = 2.0, min_distance_start_agents: float = -1.0, actions: typing.Dict[str, habitat.config.default_structured_configs.ActionConfig] = '???') -> None
- def __repr__(self)
Data
- actions: typing.Dict[str, habitat.config.default_structured_configs.ActionConfig] = '???'
- art_succ_thresh: float = 0.15
- base_angle_noise: float = 0.523599
- base_noise: float = 0.05
- cache_robot_init: bool = False
- constraint_violation_drops_object: bool = False
- constraint_violation_ends_episode: bool = True
- count_obj_collisions: bool = True
- ee_exclude_region: float = 0.0
- ee_sample_factor: float = 0.2
- enable_safe_drop: bool = False
- end_on_success: bool = False
- filter_colliding_states: bool = True
- force_regenerate: bool = False
- gfx_replay_dir = 'data/replays'
- goal_sensor_uuid: str = 'pointgoal'
- joint_max_impulse: float = -1.0
- min_distance_start_agents: float = -1.0
- min_start_distance: float = 3.0
- num_spawn_attempts: int = 200
- obj_succ_thresh: float = 0.3
- object_in_hand_sample_prob: float = 0.167
- pddl_domain_def: str = 'replica_cad'
- physics_target_sps: float = 60.0
- recep_place_shrink_factor: float = 0.8
- render_target: bool = True
- reward_measure: typing.Optional[str] = None
- robot_at_thresh: float = 2.0
- settle_steps: int = 5
- should_enforce_target_within_reach: bool = False
- should_save_to_cache: bool = False
- slack_reward: float = -0.01
- spawn_max_dist_to_obj: float = 2.0
- spawn_max_dist_to_obj_delta: float = 0.02
- spawn_region_scale: float = 0.2
- success_measure: typing.Optional[str] = None
- success_reward: float = 2.5
- success_state: float = 0.0
- task_spec: str = ''
- task_spec_base_path: str = 'habitat/task/rearrange/pddl/'
- type: str = 'Nav-v0'
- use_marker_t: bool = True
- lab_sensors: typing.Dict[str, habitat.config.default_structured_configs.LabSensorConfig] = None
- measurements: typing.Dict[str, habitat.config.default_structured_configs.MeasurementConfig] = None
- rank0_env0_measure_names: typing.List[str] = None
- rank0_measure_names: typing.List[str] = None
- desired_resting_position: typing.List[float] = None