lint

Signed-off-by: Bill Nell <[email protected]>

Author: bnellnm

vllm/model_executor/layers/fused_moe/cutlass_moe.py

@@ -7,12 +7,9 @@
 
 import vllm.model_executor.layers.fused_moe.modular_kernel as mk
 from vllm import _custom_ops as ops
-import vllm.model_executor.layers.fused_moe.modular_kernel as mk
 from vllm.model_executor.layers.fused_moe.dispatch_combine import (
-    StandardDispatchCombine
-)
-from vllm.model_executor.layers.fused_moe.utils import (_resize_cache,
-                                                        _fp8_perm)
+    StandardDispatchCombine)
+from vllm.model_executor.layers.fused_moe.utils import _fp8_perm, _resize_cache
 from vllm.scalar_type import scalar_types
 
 

vllm/model_executor/layers/fused_moe/fused_batched_moe.py

@@ -390,6 +390,7 @@ class BatchedDispatchCombine(mk.FusedMoEQuantizeDispatchCombine):
     expert batched format, i.e. E x max_num_tokens x K.  This is the format
     that the PPLX dispatch/combine kernels use.
     """
+
     def __init__(self, max_num_tokens: Optional[int], world_size: int,
                  dp_size: int, rank: int):
         super().__init__()
@@ -487,6 +488,7 @@ class BatchedExperts(mk.FusedMoEPermuteExpertsUnpermute):
     i.e. E x max_num_tokens x K.  This is the format that the pplx
     dispatch/combine kernels use.
     """
+
     def __init__(
         self,
         world_size: int,
@@ -593,6 +595,7 @@ class BatchedTritonExperts(mk.FusedMoEPermuteExpertsUnpermute):
     i.e. E x max_num_tokens x K.  This is the format that the pplx
     dispatch/combine kernels use.
     """
+
     def __init__(
         self,
         max_num_tokens: Optional[int] = None,