From 3cf8b8f0d7271867e606df94820510491f1cf398 Mon Sep 17 00:00:00 2001 From: yy-fighting Date: Wed, 8 Jul 2026 07:57:15 +0000 Subject: [PATCH] fix: use --context-length instead of --max-model-len for SGLang server --- experiments/dsv4_h200_sglang_vs_vllm/start_sglang.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/experiments/dsv4_h200_sglang_vs_vllm/start_sglang.sh b/experiments/dsv4_h200_sglang_vs_vllm/start_sglang.sh index 0bc7777..89fc7de 100755 --- a/experiments/dsv4_h200_sglang_vs_vllm/start_sglang.sh +++ b/experiments/dsv4_h200_sglang_vs_vllm/start_sglang.sh @@ -44,7 +44,7 @@ nohup sglang serve \ --model-path "$MODEL_PATH" \ --tp "$TP" \ --moe-runner-backend marlin \ - --max-model-len "$MAX_MODEL_LEN" \ + --context-length "$MAX_MODEL_LEN" \ --max-running-requests "$MAX_RUNNING" \ --mem-fraction-static 0.88 \ --host 0.0.0.0 \