1 paper · 1 filter
Minghao Li, Alicia Golden, Samuel Hsia +14
The rapid scaling of large language model training requires distributing GPU resources across multiple data center buildings and regions. We refer to such paradigm as "scale-across…