Skip to content

[Bugfix] Fix w8a8_int8 import error on NPU #8147

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
merged 1 commit into from
Jul 18, 2025
Merged

[Bugfix] Fix w8a8_int8 import error on NPU #8147

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
2 changes: 2 additions & 0 deletions python/sglang/srt/layers/quantization/w8a8_int8.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -754,6 +754,8 @@ def apply(
x: torch.Tensor,
bias: Optional[torch.Tensor] = None,
) -> torch.Tensor:
from sglang.srt.layers.linear import RowParallelLinear
Copy link
Contributor

Choose a reason for hiding this comment

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

medium

Consider moving the import statement outside of the apply method to reduce runtime overhead. While this resolves the NameError, the import statement in a frequently called method can introduce performance overhead. If circular dependencies prevent a global import, consider importing within the class's __init__ method and storing it as an instance attribute.

class NPU_W8A8LinearMethod:
    def __init__(self):
        from sglang.srt.layers.linear import RowParallelLinear
        self.RowParallelLinear = RowParallelLinear

    def apply(self, layer, ...):
        if isinstance(layer, self.RowParallelLinear):


if isinstance(layer, RowParallelLinear):
tp_rank = get_tensor_model_parallel_rank()
return self.quant_method.apply(layer, x, bias, tp_rank)
Expand Down
Loading