TL;DR PerAct is investigated, a language-conditioned behavior-cloning agent for multi-task 6-DoF manipulation that significantly outperforms unstructured image-to-action agents and 3D ConvNet baselines for a wide range of tabletop tasks.
TL;DR PerAct is investigated, a language-conditioned behavior-cloning agent for multi-task 6-DoF manipulation that significantly outperforms unstructured image-to-action agents and 3D ConvNet baselines for a wide range of tabletop tasks.