Shared collection
Software · L5 · Distributed Machine Learning Systems
Systems for training ML models at scale: data/model parallelism, parameter servers, synchronous vs. asynchronous SGD, communication-efficient gradient aggregation, and pipeline parallelism.