From 27e327b415cfd009ce1cf747b3fc0a1bbe7ee15b Mon Sep 17 00:00:00 2001 From: pansicheng Date: Tue, 3 Jun 2025 13:48:52 +0800 Subject: [PATCH] fix new_page_count_next_decode (#6671) --- python/sglang/srt/managers/schedule_batch.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/python/sglang/srt/managers/schedule_batch.py b/python/sglang/srt/managers/schedule_batch.py index 6788debb2..2839cc9d2 100644 --- a/python/sglang/srt/managers/schedule_batch.py +++ b/python/sglang/srt/managers/schedule_batch.py @@ -1333,7 +1333,9 @@ class ScheduleBatch(ScheduleBatchDisaggregationDecodeMixin): page_size = self.token_to_kv_pool_allocator.page_size if page_size == 1: return len(self.reqs) - return sum(1 for req in self.reqs if req.seqlen % page_size == 0) + # In the decoding phase, the length of a request's KV cache should be + # the total length of the request minus 1 + return sum(1 for req in self.reqs if (req.seqlen - 1) % page_size == 0) def check_decode_mem(self, buf_multiplier=1): tokens_required = (