Skip to content

[Bugfix][AWQ] Integrate int32 parameter fix from #630 #698

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
yangzhuxinyzx merged 3 commits into from
Sep 26, 2026
Merged

[Bugfix][AWQ] Integrate int32 parameter fix from #630 #698

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4a2b2a7
fix(awq_sm70_moe): register int32 buffer Parameters with requires_gra…
jobordu Sep 14, 2026
8ef32c6
[CI] Format AWQ parameter registration fix
yangzhuxinyzx Sep 26, 2026
8b9ee20
[CI] Integrate AWQ parameter fix with current main
yangzhuxinyzx Sep 26, 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
9 changes: 8 additions & 1 deletion vllm/model_executor/layers/quantization/awq_sm70_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -503,7 +503,14 @@ def _set_parameter(
name: str,
value: torch.Tensor,
) -> None:
param = value if isinstance(value, Parameter) else Parameter(value)
# Required by int32 torch.empty(int32) buffers: the Parameter ctor itself
# refuses or warns when an int tensor is registered with requires_grad=True
# (the default). Pass the flag in directly so the constructor sees it from
# line one. The trailing param.requires_grad_(False) stays as a no-op
# idempotent guard for the already-Parameter path.
param = (
value if isinstance(value, Parameter) else Parameter(value, requires_grad=False)
)
param.requires_grad_(False)
setattr(layer, name, param)

Expand Down
Loading