[Core] Priority-based scheduling in async engine (#8850)

2024-09-28 00:07:10 +02:00
parent 18e60d7d13
commit bd429f2b75
2 changed files with 24 additions and 3 deletions
--- a/vllm/engine/llm_engine.py
+++ b/vllm/engine/llm_engine.py
@@ -796,7 +796,7 @@ class LLMEngine:
            raise ValueError(f"Got lora_request {lora_request} but LoRA is "
                             "not enabled!")

-        if priority > 0 and not self.scheduler_config.policy == "priority":
+        if priority != 0 and not self.scheduler_config.policy == "priority":
            raise ValueError(f"Got priority {priority} but "
                             "Priority scheduling is not enabled.")