update List

vllm/model_executor/layers/fused_moe/fused_moe.py

@@ -1047,7 +1047,7 @@ def try_get_optimal_moe_config(
     dtype: Optional[str],
     M: int,
     is_marlin: bool = False,
-    block_shape: Optional[list[int]] = None,
+    block_shape: Optional[List[int]] = None,
     use_nn_moe: Optional[bool] = False,
 ) -> dict[str, int]:
     from vllm.model_executor.layers.fused_moe import get_config
@@ -1392,7 +1392,7 @@ def fused_experts(
         w2_zp: Optional[torch.Tensor] = None,
         a1_scale: Optional[torch.Tensor] = None,
         a2_scale: Optional[torch.Tensor] = None,
-        block_shape: Optional[list[int]] = None,
+        block_shape: Optional[List[int]] = None,
         allow_deep_gemm: bool = False,
         allow_cutlass_block_scaled_grouped_gemm: bool = False,
         use_nn_moe: Optional[bool] = False) -> torch.Tensor:
@@ -1796,7 +1796,7 @@ class TritonExperts(mk.FusedMoEPermuteExpertsUnpermute):
         use_int8_w8a16: bool = False,
         use_int4_w4a16: bool = False,
         per_act_token_quant: bool = False,
-        block_shape: Optional[list[int]] = None,
+        block_shape: Optional[List[int]] = None,
     ):
         super().__init__(
             FusedMoEQuantConfig.make(
@@ -1981,7 +1981,7 @@ def modular_triton_fused_moe(
     use_int8_w8a16: bool,
     use_int4_w4a16: bool,
     per_act_token_quant: bool,
-    block_shape: Optional[list[int]] = None,
+    block_shape: Optional[List[int]] = None,
 ) -> mk.FusedMoEModularKernel:
     return mk.FusedMoEModularKernel(
         MoEPrepareAndFinalizeNoEP(),