[Misc] local import code clean (vllm-project#23420)

Signed-off-by: Andy Xie <[email protected]>

Author: andyxning

vllm/v1/worker/gpu_worker.py

@@ -292,7 +292,6 @@ def initialize_from_config(self, kv_cache_config: KVCacheConfig) -> None:
             allocator = CuMemAllocator.get_instance()
             context = allocator.use_memory_pool(tag="kv_cache")
         else:
-            from contextlib import nullcontext
             context = nullcontext()
         with context:
             self.model_runner.initialize_kv_cache(kv_cache_config)

vllm/worker/worker.py

@@ -3,6 +3,7 @@
 """A GPU worker class."""
 import gc
 import os
+from contextlib import nullcontext
 from typing import Dict, List, Optional, Set, Tuple, Type, Union
 
 import torch
@@ -206,7 +207,6 @@ def load_model(self):
                 "used for one instance per process.")
             context = allocator.use_memory_pool(tag="weights")
         else:
-            from contextlib import nullcontext
             context = nullcontext()
         with context:
             self.model_runner.load_model()
@@ -330,7 +330,6 @@ def initialize_cache(self, num_gpu_blocks: int,
             allocator = CuMemAllocator.get_instance()
             context = allocator.use_memory_pool(tag="kv_cache")
         else:
-            from contextlib import nullcontext
             context = nullcontext()
         with context:
             self._init_cache_engine()