Skip to content
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

Jump to bottom

Enable power users to bypass device_map="auto" training block #1881

Merged
merged 5 commits into from
Aug 24, 2023
Merged

Enable power users to bypass device_map="auto" training block #1881

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
f8ba548
Enable TP greedy env var
muellerzr Aug 23, 2023
d08a2da
Right env setting
muellerzr Aug 23, 2023
d2bd6a4
Use true, not false
muellerzr Aug 23, 2023
3c560bc
Design nit
muellerzr Aug 23, 2023
9b8cc1e
ACCELERATE_BYPASS_DEVICE_MAP
muellerzr Aug 24, 2023
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
17 changes: 15 additions & 2 deletions src/accelerate/accelerator.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1193,10 +1193,12 @@ def prepare(self, *args, device_placement=None):
)

for obj in args:
# TODO: Look at enabling native TP training directly with a proper config
if (
isinstance(obj, torch.nn.Module)
and self.verify_device_map(obj)
and self.distributed_type != DistributedType.NO
and os.environ.get("ACCELERATE_BYPASS_DEVICE_MAP", "false") != "true"
):
raise ValueError(
"You can't train a model that has been loaded with `device_map='auto'` in any distributed mode."
Expand Down Expand Up @@ -1328,7 +1330,12 @@ def prepare_model(self, model: torch.nn.Module, device_placement: bool = None, e
device_placement = self.device_placement and self.distributed_type != DistributedType.FSDP
self._models.append(model)

if self.verify_device_map(model) and self.distributed_type != DistributedType.NO:
# TODO: Look at enabling native TP training directly with a proper config
if (
self.verify_device_map(model)
and self.distributed_type != DistributedType.NO
and os.environ.get("ACCELERATE_BYPASS_DEVICE_MAP", "false") != "true"
):
raise ValueError(
"You can't train a model that has been loaded with `device_map='auto'` in any distributed mode."
" Please rerun your script specifying `--num_processes=1` or by launching with `python {{myscript.py}}`."
Expand Down Expand Up @@ -1401,8 +1408,14 @@ def prepare_model(self, model: torch.nn.Module, device_placement: bool = None, e
):
if any(p.requires_grad for p in model.parameters()):
kwargs = self.ddp_handler.to_kwargs() if self.ddp_handler is not None else {}
# TODO: Look at enabling native TP training directly with a proper config
if os.environ.get("ACCELERATE_BYPASS_DEVICE_MAP", "false") != "true":
device_ids, output_device = [self.local_process_index], self.local_process_index
else:
device_ids, output_device = None, None

model = torch.nn.parallel.DistributedDataParallel(
model, device_ids=[self.local_process_index], output_device=self.local_process_index, **kwargs
model, device_ids=device_ids, output_device=output_device, **kwargs
)
elif self.distributed_type == DistributedType.FSDP:
from torch.distributed.fsdp.fully_sharded_data_parallel import FullyShardedDataParallel as FSDP
Expand Down
Loading