From b0d597363af80b918539a52fbe86c78870e48079 Mon Sep 17 00:00:00 2001 From: muh-engine Date: Thu, 6 Aug 2026 01:41:49 +0000 Subject: [PATCH] [BUGFIX] qwen3_6_scripts/model_runner.py: fix max_decode_seq_len (deployment version) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit CRITICAL: patch_ops.sh deploys qwen3_6_scripts/ files, NOT vllm/ files. Previous bugfix only fixed vllm/worker/model_runner.py but the DEPLOYED version (qwen3_6_scripts/model_runner.py) still had the bug. Fix: max_decode_seq_len=max_encoder_seq_len → max_decode_seq_len=max_decode_seq_len This ensures CUDA graph capture correctly checks actual decode sequence length, not the encoder length (which is 0 for decoder-only Qwen3.6). Discovery from reading CCCL adjacent_difference custom_policy_hub test: the test showed that custom policy hubs OVERRIDE defaults. Our project has the same pattern: qwen3_6_scripts/ overrides vllm/ via patch_ops.sh. Therefore ALL fixes must go to qwen3_6_scripts/ to survive deployment. CCCL file: cub/test/catch2_test_device_adjacent_difference_custom_policy_hub.cu --- qwen3_6_scripts/model_runner.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/qwen3_6_scripts/model_runner.py b/qwen3_6_scripts/model_runner.py index 3c4f65a3..e4eba0d1 100644 --- a/qwen3_6_scripts/model_runner.py +++ b/qwen3_6_scripts/model_runner.py @@ -837,7 +837,7 @@ class ModelInputForGPUBuilder(ModelRunnerInputBuilderBase[ModelInputForGPU]): cuda_graph_pad_size = self._get_cuda_graph_pad_size( num_seqs=len(seq_lens), - max_decode_seq_len=max_encoder_seq_len, + max_decode_seq_len=max_decode_seq_len, max_encoder_seq_len=max_encoder_seq_len) batch_size = len(input_tokens)