biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[BugFix] Fix invalid flashinfer_fused_moe_blockscale_fp8 op registration (#32855)

Browse Source 
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
This commit is contained in:
Fadi Arafeh
2026-01-22 22:27:40 +00:00
committed by GitHub
parent d08b356ee0
commit fc56f4a071
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/layers/fused_moe/flashinfer_trtllm_moe.py
View File

@@ -131,7 +131,7 @@ def flashinfer_fused_moe_blockscale_fp8(
expert_offset: int, expert_offset: int,
local_num_experts: int, local_num_experts: int,
block_shape: list[int], block_shape: list[int],
routing_method_type: int = RoutingMethodType.DeepSeekV3, routing_method_type: int = int(RoutingMethodType.DeepSeekV3),
routed_scaling: float | None = 1.0, routed_scaling: float | None = 1.0,
) -> torch.Tensor: ) -> torch.Tensor:
from vllm.utils.flashinfer import flashinfer_trtllm_fp8_block_scale_moe from vllm.utils.flashinfer import flashinfer_trtllm_fp8_block_scale_moe