frozenleaves/vllm
1
0
Fork 0
mirror of https://github.com/vllm-project/vllm.git synced 2025-10-20 23:03:52 +08:00
Code Issues Packages Projects Releases Wiki Activity

Intialize io_thread_pool attribute in the beginning. (#18331)

Browse Source 
Signed-off-by: rabi <ramishra@redhat.com>
This commit is contained in:
Rabi Mishra
2025-05-22 08:51:14 +05:30
committed by GitHub
parent 6e0fd34d3c
commit acb54ca8e1
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/v1/executor/multiproc_executor.py
View File

@ -50,6 +50,7 @@ class MultiprocExecutor(Executor):
self.is_failed = False self.is_failed = False
self.shutdown_event = threading.Event() self.shutdown_event = threading.Event()
self.failure_callback: Optional[FailureCallback] = None self.failure_callback: Optional[FailureCallback] = None
self.io_thread_pool: Optional[ThreadPoolExecutor] = None
self.world_size = self.parallel_config.world_size self.world_size = self.parallel_config.world_size
tensor_parallel_size = self.parallel_config.tensor_parallel_size tensor_parallel_size = self.parallel_config.tensor_parallel_size
@ -107,7 +108,6 @@ class MultiprocExecutor(Executor):
# For pipeline parallel, we use a thread pool for asynchronous # For pipeline parallel, we use a thread pool for asynchronous
# execute_model. # execute_model.
self.io_thread_pool: Optional[ThreadPoolExecutor] = None
if self.max_concurrent_batches > 1: if self.max_concurrent_batches > 1:
# Note: must use only 1 IO thread to keep dequeue sequence # Note: must use only 1 IO thread to keep dequeue sequence
# from the response queue # from the response queue