Skip to content

Set ddp_find_unused_parameters to False when using distributed training #222

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

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Set ddp_find_unused_parameters to False when using distributed training #222

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
813b323
Set ddp_find_unused_parameters to False when using ddp
aresnow1 Jul 21, 2023
dcb4c14
Fix
aresnow1 Jul 21, 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
10 changes: 9 additions & 1 deletion qlora.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -675,7 +675,15 @@ def train():
set_seed(args.seed)

data_module = make_data_module(tokenizer=tokenizer, args=args)


# When using distributed training, the value of the flag find_unused_parameters passed to
# DistributedDataParallel. Will default to False if gradient checkpointing is used, True otherwise.
if os.environ.get('LOCAL_RANK') is not None:
if training_args.gradient_checkpointing:
training_args.ddp_find_unused_parameters = False
else:
training_args.ddp_find_unused_parameters = True

trainer = Seq2SeqTrainer(
model=model,
tokenizer=tokenizer,
Expand Down