yandex-research/swarm

Official code for "SWARM Parallelism: Training Large Models Can Be Surprisingly Communication-Efficient"

Python

Issues

is there any 'runtime_pb2_grpc' proto file about in this repo?
#3 opened 6 months ago by ddqspace-xyz
0
Backwards fault recovery
#2 opened 10 months ago by NikolayBlagoev
0
Amazing work !
#1 opened 2 years ago by tchaton
2