class
SPLSPL (Success weighted by Path Length)
ref: On Evaluation of Embodied Agents - Anderson et. al https://arxiv.org/pdf/1807.06757.pdf The measure depends on Distance to Goal measure and Success measure to improve computational performance for sophisticated goal areas.
Methods
- def get_metric(self)
- def reset_metric(self, episode, task, *args: typing.Any, **kwargs: typing.Any)
- def update_metric(self, episode, task: core.embodied_task.EmbodiedTask, *args: typing.Any, **kwargs: typing.Any)
Special methods
- def __init__(self, sim: core.simulator.Simulator, config: DictConfig, *args: typing.Any, **kwargs: typing.Any)
Data
Method documentation
def habitat. tasks. nav. nav. SPL. get_metric(self)
Returns | the current metric for Measure . |
---|