Fix precommit

wenscarl · wenscarl · commit 7cd508da8bd4 · 2025-04-15T20:00:03.000Z
Signed-off-by: shuw &lt;shuw@nvidia.com&gt;
diff --git a/vllm/attention/backends/flashinfer.py b/vllm/attention/backends/flashinfer.py
@@ -24,9 +24,6 @@
         BatchPrefillWithPagedKVCacheWrapper = None
     FLASHINFER_WORKSPACE_BUFFER_SIZE = 0
 
-FLASHINFER_KV_CACHE_LAYOUT: str = os.getenv("FLASHINFER_KV_CACHE_LAYOUT",
-                                            "NHD").upper()
-
 import torch
 
 import vllm.envs as envs
@@ -49,6 +46,9 @@
     from vllm.worker.model_runner import (ModelInputForGPUBuilder,
                                           ModelInputForGPUWithSamplingMetadata)
 
+FLASHINFER_KV_CACHE_LAYOUT: str = os.getenv("FLASHINFER_KV_CACHE_LAYOUT",
+                                            "NHD").upper()
+
 
 class FlashInferBackend(AttentionBackend):
 
