EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Fix chunked prefill size validation for disabled state (#8973)

Browse Source 
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
This commit is contained in:
633WHU
2025-08-12 02:05:29 +08:00
committed by GitHub
parent 44426e54be
commit 3bffe11279
1 changed files with 5 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
python/sglang/srt/server_args.py
View File

@@ -1999,9 +1999,11 @@ class ServerArgs:
), "enable_mixed_chunk is required for speculative decoding" ), "enable_mixed_chunk is required for speculative decoding"
# Check chunked prefill # Check chunked prefill
assert ( # Skip validation if chunked prefill is disabled (i.e., size <= 0).
self.chunked_prefill_size % self.page_size == 0 if self.chunked_prefill_size > 0:
), "chunked_prefill_size must be divisible by page_size" assert (
self.chunked_prefill_size % self.page_size == 0
), "chunked_prefill_size must be divisible by page_size"
def check_lora_server_args(self): def check_lora_server_args(self):
assert self.max_loras_per_batch > 0, "max_loras_per_batch must be positive" assert self.max_loras_per_batch > 0, "max_loras_per_batch must be positive"