diff --git a/fastdeploy/worker/worker_process.py b/fastdeploy/worker/worker_process.py
index 568932a7ee..a81a0ee7d5 100644
--- a/fastdeploy/worker/worker_process.py
+++ b/fastdeploy/worker/worker_process.py
@@ -299,13 +299,6 @@ class PaddleDisWorkerProc:
             suffix=self.parallel_config.local_engine_worker_queue_port,
             create=False,
         )
-        # gpu_cache_lock: file-based lock for mutual exclusion between worker
-        # and CPU transfer when accessing GPU KV cache.
-        self.gpu_cache_lock = IPCLock(
-            name="gpu_cache_lock",
-            suffix=self.parallel_config.local_engine_worker_queue_port,
-            create=False,
-        )
 
     def update_weights_from_tensor(self, mmap_infos):
         """