From 944f1ea0ecaf6db72f7fdaf39f3668573c9a5039 Mon Sep 17 00:00:00 2001 From: feng397 Date: Mon, 22 Sep 2025 15:43:12 +0800 Subject: [PATCH] fix capture_bs when speculative decoding enabled (#10730) --- python/sglang/srt/model_executor/cuda_graph_runner.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/python/sglang/srt/model_executor/cuda_graph_runner.py b/python/sglang/srt/model_executor/cuda_graph_runner.py index 460776995..0a1cededd 100644 --- a/python/sglang/srt/model_executor/cuda_graph_runner.py +++ b/python/sglang/srt/model_executor/cuda_graph_runner.py @@ -179,7 +179,7 @@ def get_batch_sizes_to_capture(model_runner: ModelRunner): capture_bs = ( list(range(1, 9)) + list(range(10, 33, 2)) - + list(range(40, 64, 8)) + + list(range(40, 65, 8)) + list(range(80, 161, 16)) )