From 2ca835fea74f29564625795fccd0beaefcfaf53b Mon Sep 17 00:00:00 2001 From: root Date: Tue, 25 Aug 2026 10:08:10 +0000 Subject: [PATCH] [fix] timeout error --- qwen3_6_scripts/patch_ops.sh | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/qwen3_6_scripts/patch_ops.sh b/qwen3_6_scripts/patch_ops.sh index 8fa35a2a..3e0b4dec 100755 --- a/qwen3_6_scripts/patch_ops.sh +++ b/qwen3_6_scripts/patch_ops.sh @@ -162,10 +162,7 @@ python3 ./patch_worker_cache_transfer_order.py # when context length is high cp ./paged_attn.py "${VLLM_ROOT}/attention/ops/paged_attn.py" -# --- multimodal: increase image fetch timeout and add retry ------------------ -# Container networks often cannot reach remote image URLs within the default -# 5-second timeout. Raise to 30s and retry 3 times. -cp ./multimodal_utils.py "${VLLM_ROOT}/multimodal/utils.py" + # --- model_runner.py: fix prefix_cache_hit stays True in chunked-prefill chunk 2+ --- # Bug: _compute_for_prefix_cache_hit Case 1 (prefix_cache_len <= context_len)