ReCycle: Resilient Training of Large DNNs using Pipeline Adaptation.
Swapnil Gandhi, Mark Zhao, Athinagoras Skiadopoulos, Christos Kozyrakis
Browse the full SOSP paper archive.
Swapnil Gandhi, Mark Zhao, Athinagoras Skiadopoulos, Christos Kozyrakis
Browse the full SOSP paper archive.