[MoE Refactor] Remove SharedFusedMoE class (#35782)

Signed-off-by: Bill Nell <bnell@redhat.com>
This commit is contained in:
bnellnm
2026-04-21 18:12:12 -04:00
committed by GitHub
parent 1842447c09
commit 5e584ce9ec
33 changed files with 112 additions and 141 deletions
+1 -2
View File
@@ -31,7 +31,6 @@ from vllm.distributed import (
get_tensor_model_parallel_world_size,
)
from vllm.model_executor.layers.fused_moe import FusedMoE
from vllm.model_executor.layers.fused_moe.shared_fused_moe import SharedFusedMoE
from vllm.model_executor.layers.layernorm import RMSNorm
from vllm.model_executor.layers.linear import ReplicatedLinear
from vllm.model_executor.layers.logits_processor import LogitsProcessor
@@ -130,7 +129,7 @@ class ExaoneMoe(nn.Module):
else:
self.shared_experts = None
self.experts = SharedFusedMoE(
self.experts = FusedMoE(
shared_experts=self.shared_experts,
gate=self.gate,
num_experts=self.n_routed_experts,