Project Supernova is to build a common machine learning inference service framework by enabling machine learning inference accelerators across edge endpoint devices, edge systems and cloud, with or without hardware accelerators.
