Mixture-of-Supernets: Improving Weight-Sharing Supernet Training with Architecture-Routed Mixture-of-Experts.
Ganesh JawaharHaichuan YangYunyang XiongZechun LiuDilin WangFei SunMeng LiAasish PappuBarlas OguzMuhammad Abdul-MageedLaks V. S. LakshmananRaghuraman KrishnamoorthiVikas ChandraPublished in: CoRR (2023)