Skip to content

Fix: fused QUproj + RoPE + Quant #3421

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
ghadiaravi13 wants to merge 3 commits into
base: main
Choose a base branch
from
+341 −11
Open

Fix: fused QUproj + RoPE + Quant #3421

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1a9b681
Apply TE PR 3330: backward linear for fused MLA Q up-proj (src only)
ghadiaravi13 Aug 18, 2026
3fb5910
refactor: make the absorbed q_layernorm the only fused Q up-proj path
ghadiaravi13 Aug 22, 2026
1dbb4cd
refactor: simplify comments and documentation in FusedMLAQUpProj classes
ghadiaravi13 Aug 24, 2026
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 transformer_engine/pytorch/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,7 @@
from transformer_engine.pytorch.module import destroy_ub
from transformer_engine.pytorch.module import UserBufferQuantizationMode
from transformer_engine.pytorch.attention import DotProductAttention
from transformer_engine.pytorch.attention import FusedMLAQUpProjRopeQuant
from transformer_engine.pytorch.attention import FusedMLAQUpProjFunction, FusedMLAQUpProjRopeQuant

@greptile-apps greptile-apps Bot Aug 24, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Undocumented public autograd API

FusedMLAQUpProjFunction is now exported from the root and attention namespaces without a corresponding entry in docs/api/pytorch.rst, leaving its extensive argument and shape contract absent from the generated public API reference.

Knowledge Base Used: PyTorch runtime and public API

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

from transformer_engine.pytorch.attention import MultiheadAttention
from transformer_engine.pytorch.attention import InferenceParams
from transformer_engine.pytorch.attention import RotaryPositionEmbedding
Expand Down
3 changes: 2 additions & 1 deletion transformer_engine/pytorch/attention/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,13 +5,14 @@
"""Python interface for attention"""

from .dot_product_attention import DotProductAttention
from .fused_mla_q_uproj import FusedMLAQUpProjRopeQuant
from .fused_mla_q_uproj import FusedMLAQUpProjFunction, FusedMLAQUpProjRopeQuant
from .multi_head_attention import MultiheadAttention
from .inference import InferenceParams
from .rope import RotaryPositionEmbedding

__all__ = [
"DotProductAttention",
"FusedMLAQUpProjFunction",
"FusedMLAQUpProjRopeQuant",
"MultiheadAttention",
"InferenceParams",
Expand Down
Loading
Loading