Skip to content

fix: relax padding for fp4 in weight loading #12497

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
wants to merge 3 commits into
base: main
Choose a base branch
from
Open

fix: relax padding for fp4 in weight loading #12497

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4dc3369
more
b8zhong Nov 1, 2025
bab84be
more
b8zhong Nov 1, 2025
8ffc89e
more
b8zhong Nov 1, 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
10 changes: 7 additions & 3 deletions python/sglang/srt/layers/quantization/modelopt_quant.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1438,9 +1438,13 @@ def _slice_scale(w):
("w13", layer.w13_weight_scale),
("w2", layer.w2_weight_scale),
]:
assert (
weight_scale.shape[2] % 16 == 0
), f"Expected {name}_weight_scale.dim(2) to be divisible by 16"
if weight_scale.shape[2] % 4 != 0:
logger.warning(
"NVFP4 %s_weight_scale K' not multiple of 4: shape=%s, group_size=%s",
name,
tuple(weight_scale.shape),
getattr(self.quant_config, "group_size", None),
)
Comment on lines +1442 to +1447
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Nov 1, 2025

Choose a reason for hiding this comment

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

medium

The warning message is a bit cryptic with K'. It would be more helpful for developers if the message explicitly stated which dimension is being referred to and what the consequence is (i.e., that padding will be applied).

                logger.warning(
                    "NVFP4 %s_weight_scale last dim not a multiple of 4 (shape=%s, "
                    "group_size=%s). Padding will be applied.",
                    name,
                    tuple(weight_scale.shape),
                    getattr(self.quant_config, "group_size", None),
                )

assert (
weight_scale.dtype == torch.float8_e4m3fn
), f"{name} Weight Blockscale must be represented as FP8-E4M3"
Expand Down
Loading