From 3bffe11279c53c04af97434ddac79bb78e38e03c Mon Sep 17 00:00:00 2001 From: 633WHU Date: Tue, 12 Aug 2025 02:05:29 +0800 Subject: [PATCH] Fix chunked prefill size validation for disabled state (#8973) Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> --- python/sglang/srt/server_args.py | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/python/sglang/srt/server_args.py b/python/sglang/srt/server_args.py index 93ceb6797..98a0369e6 100644 --- a/python/sglang/srt/server_args.py +++ b/python/sglang/srt/server_args.py @@ -1999,9 +1999,11 @@ class ServerArgs: ), "enable_mixed_chunk is required for speculative decoding" # Check chunked prefill - assert ( - self.chunked_prefill_size % self.page_size == 0 - ), "chunked_prefill_size must be divisible by page_size" + # Skip validation if chunked prefill is disabled (i.e., size <= 0). + if self.chunked_prefill_size > 0: + assert ( + self.chunked_prefill_size % self.page_size == 0 + ), "chunked_prefill_size must be divisible by page_size" def check_lora_server_args(self): assert self.max_loras_per_batch > 0, "max_loras_per_batch must be positive"