From 6b8f66efe1ba754c23326ae6e71f96e05ae132de Mon Sep 17 00:00:00 2001 From: Yineng Zhang Date: Fri, 2 Aug 2024 22:40:52 +0800 Subject: [PATCH] misc: update cuda graph capture exception log (#894) --- python/sglang/srt/model_executor/model_runner.py | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/python/sglang/srt/model_executor/model_runner.py b/python/sglang/srt/model_executor/model_runner.py index 44a4498fa..e65ea7888 100644 --- a/python/sglang/srt/model_executor/model_runner.py +++ b/python/sglang/srt/model_executor/model_runner.py @@ -312,10 +312,12 @@ class ModelRunner: self.cuda_graph_runner.capture(batch_size_list) except RuntimeError as e: raise Exception( - f"Capture cuda graph failed: {e}. Possible solutions:\n" - f"1. disable cuda graph by --disable-cuda-graph\n" - f"2. set --mem-fraction-static to a smaller value\n" - f"Open an issue on GitHub with reproducible scripts if you need help.\n" + f"Capture cuda graph failed: {e}\n" + "Possible solutions:\n" + "1. disable torch compile by not using --enable-torch-compile\n" + "2. disable cuda graph by --disable-cuda-graph\n" + "3. set --mem-fraction-static to a smaller value\n" + "Open an issue on GitHub https://github.com/sgl-project/sglang/issues/new/choose \n" ) @torch.inference_mode()