Skip to content

[https://nvbugs/5676748][fix] Cherry-pick #9336: Fix mismatched nvfp4 gemm sf shape. #9437

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
hyukn merged 2 commits into from
Nov 26, 2025
Merged

[https://nvbugs/5676748][fix] Cherry-pick #9336: Fix mismatched nvfp4 gemm sf shape. #9437

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
819eb39
[https://nvbugs/5676748][fix] Fix mismatched nvfp4 gemm sf shape. (#9…
hyukn Nov 24, 2025
287fed2
update pre-commit ignore scope
hyukn Nov 26, 2025
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 .pre-commit-config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,7 +66,7 @@ repos:
additional_dependencies:
- tomli
# add ignore words list
args: ["-L", "Mor,ans,thirdparty"]
args: ["-L", "Mor,ans,thirdparty", "--skip", "security_scanning/*"]
- repo: https://github.com/astral-sh/ruff-pre-commit
rev: v0.9.4
hooks:
Expand Down
6 changes: 3 additions & 3 deletions tensorrt_llm/_torch/attention_backend/trtllm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -300,9 +300,9 @@ def create_output(self, q: torch.Tensor, out_dtype: torch.dtype):
size_per_token // num_nvfp4_elements_per_container),
dtype=torch.uint8)
# Create a sf (scaling factors) tensor for NVFP4 (use INT8 as the container dtype).
output_sf = q.new_empty(compute_swizzled_sf_shape(
num_tokens, size_per_token // scaling_vector_size),
dtype=torch.uint8)
padded_row, padded_col = compute_swizzled_sf_shape(
num_tokens, size_per_token // scaling_vector_size)
output_sf = q.new_empty(padded_row * padded_col, dtype=torch.uint8)
else:
output = q.new_empty((num_tokens, self.num_heads * v_head_size),
dtype=out_dtype)
Expand Down