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.

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

Fix, bf16 optimizer remove dup loop #7054

Merged
merged 2 commits into from
Feb 20, 2025
Merged
Changes from all commits
Commits
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: 3 additions & 3 deletions deepspeed/runtime/bf16_optimizer.py
Original file line number Diff line number Diff line change
Expand Up @@ -472,10 +472,10 @@ def state_dict(self):

# Restore base optimizer fp32 weights bfloat16 weights
def _restore_from_bit16_weights(self):
for i, group in enumerate(self.bf16_groups):
for i, (bf16_partitions,
fp32_partition) in enumerate(zip(self.bf16_partitioned_groups, self.fp32_groups_flat_partition)):
partition_id = dist.get_rank(group=self.real_dp_process_group[i])
for bf16_partitions, fp32_partition in zip(self.bf16_partitioned_groups, self.fp32_groups_flat_partition):
fp32_partition.data.copy_(bf16_partitions[partition_id].data)
fp32_partition.data.copy_(bf16_partitions[partition_id].data)

def refresh_fp32_params(self):
self._restore_from_bit16_weights()
Expand Down