Loading megatron/initialize.py +0 −1 Original line number Diff line number Diff line Loading @@ -180,7 +180,6 @@ def _initialize_distributed(): torch.distributed.init_process_group( backend=args.distributed_backend, world_size=args.world_size, rank=args.rank, init_method=init_method, timeout=timedelta(days=7)) # Set the tensor model-parallel, pipeline model-parallel, and Loading Loading
megatron/initialize.py +0 −1 Original line number Diff line number Diff line Loading @@ -180,7 +180,6 @@ def _initialize_distributed(): torch.distributed.init_process_group( backend=args.distributed_backend, world_size=args.world_size, rank=args.rank, init_method=init_method, timeout=timedelta(days=7)) # Set the tensor model-parallel, pipeline model-parallel, and Loading