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, 09:45 UTC
[V1] Fix general plugins not loaded in engine for multiproc (#18326)
Signed-off-by: Yong Hoon Shin <yhshin@meta.com>
Yong Hoon Shin committed
May 21, 2025 at 01:21 UTC
907f935de9f3b6be9d793db32d7b92877419c402
1 file changed
+4
vllm/v1/engine/core.py
+4
@@ -57,6 +57,10 @@ class EngineCore:
57
executor_fail_callback: Optional[Callable] = None):
58
assert vllm_config.model_config.runner_type != "pooling"
59
60
+ # plugins need to be loaded at the engine/scheduler level too
61
+ from vllm.plugins import load_general_plugins
62
+ load_general_plugins()
63
+
64
self.vllm_config = vllm_config
65
logger.info("Initializing a V1 LLM engine (v%s) with config: %s",
66
VLLM_VERSION, vllm_config)