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

Dist-optim fixes #392

Closed
wants to merge 4 commits into from
Closed

Dist-optim fixes #392

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
2f6ddc3
check if using mcore dist opt
akoumpa Nov 8, 2024
5f95d42
pad to mult
akoumpa Nov 8, 2024
6ac0099
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Nov 8, 2024
6ddddd2
Update megatron_gpt_dpo_model.py
akoumpa Nov 8, 2024
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
6 changes: 4 additions & 2 deletions nemo_aligner/utils/train_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -110,7 +110,8 @@ def grad_reductions(ptl_model):
# synchronize asynchronous grad reductions
# note: not necessary, but reduces performance degradation
# from multiple simultaneous NCCL calls
ptl_model._optimizer._finish_bucket_grad_sync()
if not ptl_model.use_mcore_dist_optim:
ptl_model._optimizer._finish_bucket_grad_sync()
elif ptl_model.megatron_amp_O2:
# when using pipeline parallelism grads must be all-reduced after the pipeline (not asynchronously)
if ptl_model.cfg.get("pipeline_model_parallel_size", 1) > 1 or ptl_model.cfg.get("sequence_parallel", False):
Expand Down Expand Up @@ -164,7 +165,8 @@ def clip_gradients(ptl_model, clip_val):
return

if ptl_model.with_distributed_adam:
grad_norm = clip_grad_norm_distributed_optimizer(ptl_model._optimizer, clip_val)
if not ptl_model.use_mcore_dist_optim:
grad_norm = clip_grad_norm_distributed_optimizer(ptl_model._optimizer, clip_val)
else:
if ptl_model.megatron_amp_O2:
# grep fp32 master parameters for gradient clipping
Expand Down
Loading