MLfabric accelerates distributed ML training up to 3X by managing communication patterns through ordering, in-network aggregation, and proactive replication of updates.
Synchronous SGD/PS: Here, at each iteration, workers read the latest model and compute a local update using a portion of the mini- batch
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Network-accelerated Distributed Machine Learning Using MLFabric
MLfabric accelerates distributed ML training up to 3X by managing communication patterns through ordering, in-network aggregation, and proactive replication of updates.