Skip to content

[fix][cpu] Use a SwigluOAI impl which supports interleaved gate-up wei #29273

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
Nov 25, 2025
Merged

[fix][cpu] Use a SwigluOAI impl which supports interleaved gate-up wei #29273

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
29 changes: 8 additions & 21 deletions vllm/model_executor/layers/fused_moe/cpu_fused_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,22 +6,7 @@
from torch.nn import functional as F

from vllm import _custom_ops as ops


def silu_and_mul(x: torch.Tensor) -> torch.Tensor:
d = x.shape[-1] // 2
return F.silu(x[..., :d]) * x[..., d:]


def swigluoai_and_mul(
x: torch.Tensor, alpha: float = 1.702, limit: float = 7.0
) -> torch.Tensor:
d = x.shape[-1] // 2
gate, up = x[..., :d], x[..., d:]
gate = gate.clamp(max=limit)
up = up.clamp(min=-limit, max=limit)
glu = gate * torch.sigmoid(alpha * gate)
return (up + 1) * glu
from vllm.model_executor.layers.activation import SiluAndMul, SwigluOAIAndMul


def grouped_topk(
Expand Down Expand Up @@ -227,6 +212,11 @@ def __init__(self, layer: torch.nn.Module) -> None:
layer.w13_weight = torch.nn.Parameter(torch.empty(0), requires_grad=False)
layer.w2_weight = torch.nn.Parameter(torch.empty(0), requires_grad=False)

self.act_to_impl = {
"silu": SiluAndMul(),
"swigluoai": SwigluOAIAndMul(),
}

def __call__(
self,
layer: torch.nn.Module,
Expand All @@ -246,7 +236,7 @@ def __call__(
apply_router_weight_on_input: bool = False,
activation: str = "silu",
) -> torch.Tensor:
assert activation in {"silu", "swigluoai"}, f"{activation} is not supported."
assert activation in self.act_to_impl, f"{activation} is not supported."
assert not apply_router_weight_on_input
topk_weights, topk_ids = select_experts(
hidden_states=x,
Expand Down Expand Up @@ -283,10 +273,7 @@ def __call__(
tokens_for_this_expert = sorted_tokens[start_idx:end_idx]

gate_up = layer.gate_up_linear[i](tokens_for_this_expert)
if activation == "swigluoai":
gate_up = swigluoai_and_mul(gate_up)
else:
gate_up = silu_and_mul(gate_up)
gate_up = self.act_to_impl[activation].forward_native(gate_up)
expert_out = layer.down_linear[i](gate_up)
outputs.append(expert_out)
start_idx = end_idx
Expand Down