Skip to content

Modify mfsdp default data-parallel-sharding-strategy for dev #3692

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
wplf wants to merge 1 commit into
base: dev
Choose a base branch
from
+1 −1
Open

Modify mfsdp default data-parallel-sharding-strategy for dev #3692

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion megatron/training/arguments.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2477,7 +2477,7 @@ def _add_distributed_args(parser):
group.add_argument('--create-all-gather-group', action='store_true',
help='Create a separate process group for all-gather operations '
'to overlap reduce-scatter and all-gather operations.')
group.add_argument('--data-parallel-sharding-strategy', type=str, default='no_shard',
group.add_argument('--data-parallel-sharding-strategy', type=str, default='optim_grads_params',
choices=['no_shard', 'optim', 'optim_grads', 'optim_grads_params'],
help='Sharding strategy of data parallelism.')
group.add_argument('--outer-dp-sharding-strategy', type=str, default='no_shard',
Expand Down