Skip to content

Do not initialize torch.distributed process group if one is already initailized #16487

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Mar 29, 2022
Merged

Do not initialize torch.distributed process group if one is already initailized #16487

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c05de5f
Do not initialize torch process group twice
Yard1 Mar 29, 2022
553151e
Apply suggestions from code review
Yard1 Mar 29, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 7 additions & 1 deletion src/transformers/sagemaker/training_args_sm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,6 +77,11 @@ def __post_init__(self):
@cached_property
def _setup_devices(self) -> "torch.device":
logger.info("PyTorch: setting up devices")
if torch.distributed.is_initialized() and self.local_rank == -1:
logger.warning(
"torch.distributed process group is initialized, but local_rank == -1. "
"In order to use Torch DDP, launch your script with `python -m torch.distributed.launch"
)
if self.no_cuda:
device = torch.device("cpu")
self._n_gpu = 0
Expand Down Expand Up @@ -105,7 +110,8 @@ def _setup_devices(self) -> "torch.device":
else:
# Here, we'll use torch.distributed.
# Initializes the distributed backend which will take care of synchronizing nodes/GPUs
torch.distributed.init_process_group(backend="nccl")
if not torch.distributed.is_initialized():
torch.distributed.init_process_group(backend="nccl")
device = torch.device("cuda", self.local_rank)
self._n_gpu = 1

Expand Down
10 changes: 8 additions & 2 deletions src/transformers/training_args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1022,10 +1022,15 @@ def eval_batch_size(self) -> int:
@torch_required
def _setup_devices(self) -> "torch.device":
logger.info("PyTorch: setting up devices")
if torch.distributed.is_initialized() and self.local_rank == -1:
logger.warning(
"torch.distributed process group is initialized, but local_rank == -1. "
"In order to use Torch DDP, launch your script with `python -m torch.distributed.launch"
)
if self.no_cuda:
device = torch.device("cpu")
self._n_gpu = 0
if self.local_rank != -1:
if self.local_rank != -1 and not torch.distributed.is_initialized():
# Initializes distributed backend for cpu
if self.xpu_backend not in ("mpi", "ccl"):
raise ValueError(
Expand Down Expand Up @@ -1076,7 +1081,8 @@ def _setup_devices(self) -> "torch.device":
else:
# Here, we'll use torch.distributed.
# Initializes the distributed backend which will take care of synchronizing nodes/GPUs
torch.distributed.init_process_group(backend="nccl")
if not torch.distributed.is_initialized():
torch.distributed.init_process_group(backend="nccl")
device = torch.device("cuda", self.local_rank)
self._n_gpu = 1

Expand Down