diff --git a/vllm/config/scheduler.py b/vllm/config/scheduler.py index e041469660..3773a47ec9 100644 --- a/vllm/config/scheduler.py +++ b/vllm/config/scheduler.py @@ -79,7 +79,7 @@ class SchedulerConfig: long_prefill_token_threshold: int = Field(default=0, ge=0) """For chunked prefill, a request is considered long if the prompt is - longer than this number of tokens.""" + longer than this number of tokens. 0 disables the cap (default).""" enable_chunked_prefill: bool = True """If True, prefill requests can be chunked based