Skip to content

GitLab

Commit 50bed026 authored by zhuwenwen's avatar zhuwenwen
Browse files

remove redundant shared_output

parent 42b06117
Hide whitespace changes
Inline Side-by-side
Showing with 0 additions and 5 deletions
vllm/model_executor/layers/fused_moe/layer.py
View file @ 50bed026
......@@ -206,7 +206,6 @@ class FusedMoEMethodBase(QuantizeMethodBase):
layer: torch.nn.Module,
x: torch.Tensor,
router_logits: torch.Tensor,
shared_output: torch.Tensor,
top_k: int,
renormalize: bool,
use_grouped_topk: bool = False,
......@@ -359,7 +358,6 @@ class UnquantizedFusedMoEMethod(FusedMoEMethodBase, CustomOp):
layer: torch.nn.Module,
x: torch.Tensor,
router_logits: torch.Tensor,
shared_output: torch.Tensor,
top_k: int,
renormalize: bool,
use_grouped_topk: bool = False,
......@@ -389,7 +387,6 @@ class UnquantizedFusedMoEMethod(FusedMoEMethodBase, CustomOp):
x=x,
layer=layer,
router_logits=router_logits,
shared_output=shared_output,
top_k=top_k,
renormalize=renormalize,
use_grouped_topk=use_grouped_topk,
......@@ -414,7 +411,6 @@ class UnquantizedFusedMoEMethod(FusedMoEMethodBase, CustomOp):
use_grouped_topk: bool,
top_k: int,
router_logits: torch.Tensor,
shared_output: torch.Tensor,
renormalize: bool,
topk_group: Optional[int] = None,
num_expert_group: Optional[int] = None,
......@@ -1544,7 +1540,6 @@ class FusedMoE(torch.nn.Module):
layer=self,
x=hidden_states,
router_logits=router_logits,
shared_output=shared_output,
top_k=self.top_k,
renormalize=self.renormalize,
use_grouped_topk=self.use_grouped_topk,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment