A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[Minor] Fix pre-commit error on main (#22579)
Signed-off-by: Isotr0py <2037008807@qq.com>
Isotr0py committed
Aug 10, 2025 at 15:08 UTC
7e8d685775fe9e11c3cea79e84418a9f0bab4a5f
1 file changed
+2
-2
vllm/model_executor/layers/fused_moe/fused_moe.py
+2
-2
@@ -1038,9 +1038,9 @@ def inplace_fused_experts(
1038
w2_zp: Optional[torch.Tensor] = None,
1039
a1_scale: Optional[torch.Tensor] = None,
1040
a2_scale: Optional[torch.Tensor] = None,
1041
- block_shape: Optional[List[int]] = None,
1041
+ block_shape: Optional[List[int]] = None, #noqa: UP006
1042
w1_bias: Optional[torch.Tensor] = None,
1043
- w2_bias: Optional[torch.Tensor] = None) -> None: #noqa: UP006
1043
+ w2_bias: Optional[torch.Tensor] = None) -> None:
1044
fused_experts_impl(hidden_states, w1, w2, topk_weights, topk_ids, True,
1045
activation, is_act_and_mul,
1046
apply_router_weight_on_input, use_fp8_w8a8,