Skip to content

M-FSDP: Cancel erroneous grad accumulation check #3629

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

Merged
shjwudp merged 1 commit into from
Mar 2, 2026
Merged

M-FSDP: Cancel erroneous grad accumulation check #3629

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
6 changes: 0 additions & 6 deletions megatron/core/distributed/fsdp/mcore_fsdp_adapter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,12 +86,6 @@ def __init__(
self.megatron_fsdp_dist_index = self._init_dist_index(pg_collection)

if config.gradient_accumulation_fusion:
assert (
self.megatron_fsdp_dist_index.get_dp_group(is_expert_parallel=True).size() == 1
), (
"Megatron-FSDP with gradient_accumulation_fusion does not support "
"data parallelism when expert parallelism is enabled."
)
assert is_te_min_version("2.10"), (
"Megatron-FSDP with gradient_accumulation_fusion requires "
"Transformer Engine version 2.10 or higher."
Expand Down
Loading