A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 04:45 UTC
[Misc] Clean up useless code (#19889)
Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
wangxiyuan committed
Jun 21, 2025 at 05:09 UTC
e773a9e1c2c175e193b383ed497ad6fcb73cdfe5
1 file changed
-11
vllm/config.py
-11
@@ -1900,17 +1900,6 @@ class ParallelConfig:
1900
os.environ["VLLM_ENABLE_V1_MULTIPROCESSING"] = "0"
1901
logger.info("Disabling V1 multiprocessing for external launcher.")
1902
1903
- ray_only_devices: list[str] = []
1904
- from vllm.platforms import current_platform
1905
- if (current_platform.device_type in ray_only_devices
1906
- and self.world_size > 1):
1907
- if self.distributed_executor_backend is None:
1908
- self.distributed_executor_backend = "ray"
1909
- if self.distributed_executor_backend != "ray":
1910
- raise ValueError(
1911
- f"{current_platform.device_type.upper()} backend only "
1912
- "supports Ray for distributed inference.")
1913
-
1903
if self.distributed_executor_backend is None and self.world_size > 1:
1904
# We use multiprocessing by default if world_size fits on the
1905
# current node and we aren't in a ray placement group.