1 paper · 1 filter
Sameera Ramasinghe, Thalaiyasingam Ajanthan, Gil Avraham +2
Scaling models has led to significant advancements in deep learning, but training these models in decentralized settings remains challenging due to communication bottlenecks. While…