From 1907d3854ae7dd0e1a5755fafbe93165ae6413a6 Mon Sep 17 00:00:00 2001 From: JianweiZheng <32029023+jwzheng96@users.noreply.github.com> Date: Tue, 30 Jun 2026 20:01:03 +0800 Subject: [PATCH] [Bugfix] Reject negative values for max_logprobs and long_prefill_token_threshold (#44002) Signed-off-by: jwzheng96 Signed-off-by: JianweiZheng <32029023+jwzheng96@users.noreply.github.com> Co-authored-by: Claude Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com> Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> --- vllm/config/scheduler.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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