(llm311) nvidia@localhost:~$ python -m vllm.entrypoints.openai.api_server \
--model /data/lqy/qwen/Qwen3-VL-8B-Instruct \ --served-model-name Qwen3-VL-8B-Instruct \ --host 0.0.0.0 \ --port 9000 \ --gpu-memory-utilization 0.7 \ --skip-mm-profiling
INFO 11-14 11:34:33 [init.py:216] Automatically detected platform cuda.
(APIServer pid=2490067) INFO 11-14 11:34:34 [api_server.py:1839] vLLM API server version 0.11.0
(APIServer pid=2490067) INFO 11-14 11:34:34 [utils.py:233] non-default args: {‘host’: ‘0.0.0.0’, ‘port’: 9000, ‘model’: ‘/data/lqy/qwen/Qwen3-VL-8B-Instruct’, ‘served_model_name’: [‘Qwen3-VL-8B-Instruct’], ‘gpu_memory_utilization’: 0.7, ‘skip_mm_profiling’: True}
(APIServer pid=2490067) INFO 11-14 11:34:34 [model.py:547] Resolved architecture: Qwen3VLForConditionalGeneration
(APIServer pid=2490067) torch_dtype is deprecated! Use dtype instead!
(APIServer pid=2490067) INFO 11-14 11:34:34 [model.py:1510] Using max model len 262144
(APIServer pid=2490067) INFO 11-14 11:34:35 [scheduler.py:205] Chunked prefill is enabled with max_num_batched_tokens=2048.
INFO 11-14 11:34:39 [init.py:216] Automatically detected platform cuda.
(EngineCore_DP0 pid=2490429) INFO 11-14 11:34:41 [core.py:644] Waiting for init message from front-end.
(EngineCore_DP0 pid=2490429) INFO 11-14 11:34:41 [core.py:77] Initializing a V1 LLM engine (v0.11.0) with config: model=‘/data/lqy/qwen/Qwen3-VL-8B-Instruct’, speculative_config=None, tokenizer=‘/data/lqy/qwen/Qwen3-VL-8B-Instruct’, skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=262144, download_dir=None, load_format=auto, tensor_parallel_size=1, pipeline_parallel_size=1, data_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, structured_outputs_config=StructuredOutputsConfig(backend=‘auto’, disable_fallback=False, disable_any_whitespace=False, disable_additional_properties=False, reasoning_parser=‘’), observability_config=ObservabilityConfig(show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None), seed=0, served_model_name=Qwen3-VL-8B-Instruct, enable_prefix_caching=True, chunked_prefill_enabled=True, pooler_config=None, compilation_config={“level”:3,“debug_dump_path”:“”,“cache_dir”:“”,“backend”:“”,“custom_ops”:,“splitting_ops”:[“vllm.unified_attention”,“vllm.unified_attention_with_output”,“vllm.mamba_mixer2”,“vllm.mamba_mixer”,“vllm.short_conv”,“vllm.linear_attention”,“vllm.plamo2_mamba_mixer”,“vllm.gdn_attention”,“vllm.sparse_attn_indexer”],“use_inductor”:true,“compile_sizes”:,“inductor_compile_config”:{“enable_auto_functionalized_v2”:false},“inductor_passes”:{},“cudagraph_mode”:[2,1],“use_cudagraph”:true,“cudagraph_num_of_warmups”:1,“cudagraph_capture_sizes”:[512,504,496,488,480,472,464,456,448,440,432,424,416,408,400,392,384,376,368,360,352,344,336,328,320,312,304,296,288,280,272,264,256,248,240,232,224,216,208,200,192,184,176,168,160,152,144,136,128,120,112,104,96,88,80,72,64,56,48,40,32,24,16,8,4,2,1],“cudagraph_copy_inputs”:false,“full_cuda_graph”:false,“use_inductor_graph_partition”:false,“pass_config”:{},“max_capture_size”:512,“local_cache_dir”:null}
(EngineCore_DP0 pid=2490429) /data/conda/envs/llm311/lib/python3.11/site-packages/torch/cuda/init.py:326: UserWarning:
(EngineCore_DP0 pid=2490429) NVIDIA Thor with CUDA capability sm_110 is not compatible with the current PyTorch installation.
(EngineCore_DP0 pid=2490429) The current PyTorch install supports CUDA capabilities sm_80 sm_90 sm_100 sm_120.
(EngineCore_DP0 pid=2490429) If you want to use the NVIDIA Thor GPU with PyTorch, please check the instructions at Get Started
(EngineCore_DP0 pid=2490429)
(EngineCore_DP0 pid=2490429) warnings.warn(
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] EngineCore failed to start.
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] Traceback (most recent call last):
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core.py”, line 699, in run_engine_core
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] engine_core = EngineCoreProc(*args, **kwargs)
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core.py”, line 498, in init
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] super().init(vllm_config, executor_class, log_stats,
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core.py”, line 83, in init
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] self.model_executor = executor_class(vllm_config)
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/executor/executor_base.py”, line 54, in init
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] self._init_executor()
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/executor/uniproc_executor.py”, line 54, in _init_executor
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] self.collective_rpc(“init_device”)
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/executor/uniproc_executor.py”, line 83, in collective_rpc
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] return [run_method(self.driver_worker, method, args, kwargs)]
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/utils/init.py”, line 3122, in run_method
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] return func(*args, **kwargs)
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/worker/worker_base.py”, line 259, in init_device
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] self.worker.init_device() # type: ignore
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py”, line 161, in init_device
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] current_platform.set_device(self.device)
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/platforms/cuda.py”, line 83, in set_device
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] _ = torch.zeros(1, device=device)
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] torch.AcceleratorError: CUDA error: no kernel image is available for execution on the device
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] For debugging consider passing CUDA_LAUNCH_BLOCKING=1
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708] Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
(EngineCore_DP0 pid=2490429) ERROR 11-14 11:34:43 [core.py:708]
(EngineCore_DP0 pid=2490429) Process EngineCore_DP0:
(EngineCore_DP0 pid=2490429) Traceback (most recent call last):
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/multiprocessing/process.py”, line 314, in _bootstrap
(EngineCore_DP0 pid=2490429) self.run()
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/multiprocessing/process.py”, line 108, in run
(EngineCore_DP0 pid=2490429) self._target(*self._args, **self._kwargs)
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core.py”, line 712, in run_engine_core
(EngineCore_DP0 pid=2490429) raise e
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core.py”, line 699, in run_engine_core
(EngineCore_DP0 pid=2490429) engine_core = EngineCoreProc(*args, **kwargs)
(EngineCore_DP0 pid=2490429) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core.py”, line 498, in init
(EngineCore_DP0 pid=2490429) super().init(vllm_config, executor_class, log_stats,
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core.py”, line 83, in init
(EngineCore_DP0 pid=2490429) self.model_executor = executor_class(vllm_config)
(EngineCore_DP0 pid=2490429) ^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/executor/executor_base.py”, line 54, in init
(EngineCore_DP0 pid=2490429) self._init_executor()
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/executor/uniproc_executor.py”, line 54, in _init_executor
(EngineCore_DP0 pid=2490429) self.collective_rpc(“init_device”)
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/executor/uniproc_executor.py”, line 83, in collective_rpc
(EngineCore_DP0 pid=2490429) return [run_method(self.driver_worker, method, args, kwargs)]
(EngineCore_DP0 pid=2490429) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/utils/init.py”, line 3122, in run_method
(EngineCore_DP0 pid=2490429) return func(*args, **kwargs)
(EngineCore_DP0 pid=2490429) ^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/worker/worker_base.py”, line 259, in init_device
(EngineCore_DP0 pid=2490429) self.worker.init_device() # type: ignore
(EngineCore_DP0 pid=2490429) ^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py”, line 161, in init_device
(EngineCore_DP0 pid=2490429) current_platform.set_device(self.device)
(EngineCore_DP0 pid=2490429) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/platforms/cuda.py”, line 83, in set_device
(EngineCore_DP0 pid=2490429) _ = torch.zeros(1, device=device)
(EngineCore_DP0 pid=2490429) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2490429) torch.AcceleratorError: CUDA error: no kernel image is available for execution on the device
(EngineCore_DP0 pid=2490429) CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
(EngineCore_DP0 pid=2490429) For debugging consider passing CUDA_LAUNCH_BLOCKING=1
(EngineCore_DP0 pid=2490429) Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
(EngineCore_DP0 pid=2490429)
(APIServer pid=2490067) Traceback (most recent call last):
(APIServer pid=2490067) File “”, line 198, in _run_module_as_main
(APIServer pid=2490067) File “”, line 88, in _run_code
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/entrypoints/openai/api_server.py”, line 1953, in
(APIServer pid=2490067) uvloop.run(run_server(args))
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/uvloop/init.py”, line 92, in run
(APIServer pid=2490067) return runner.run(wrapper())
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/asyncio/runners.py”, line 118, in run
(APIServer pid=2490067) return self._loop.run_until_complete(task)
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “uvloop/loop.pyx”, line 1518, in uvloop.loop.Loop.run_until_complete
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/uvloop/init.py”, line 48, in wrapper
(APIServer pid=2490067) return await main
(APIServer pid=2490067) ^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/entrypoints/openai/api_server.py”, line 1884, in run_server
(APIServer pid=2490067) await run_server_worker(listen_address, sock, args, **uvicorn_kwargs)
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/entrypoints/openai/api_server.py”, line 1902, in run_server_worker
(APIServer pid=2490067) async with build_async_engine_client(
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/contextlib.py”, line 210, in aenter
(APIServer pid=2490067) return await anext(self.gen)
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/entrypoints/openai/api_server.py”, line 180, in build_async_engine_client
(APIServer pid=2490067) async with build_async_engine_client_from_engine_args(
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/contextlib.py”, line 210, in aenter
(APIServer pid=2490067) return await anext(self.gen)
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/entrypoints/openai/api_server.py”, line 225, in build_async_engine_client_from_engine_args
(APIServer pid=2490067) async_llm = AsyncLLM.from_vllm_config(
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/utils/init.py”, line 1572, in inner
(APIServer pid=2490067) return fn(*args, **kwargs)
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/async_llm.py”, line 207, in from_vllm_config
(APIServer pid=2490067) return cls(
(APIServer pid=2490067) ^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/async_llm.py”, line 134, in init
(APIServer pid=2490067) self.engine_core = EngineCoreClient.make_async_mp_client(
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core_client.py”, line 102, in make_async_mp_client
(APIServer pid=2490067) return AsyncMPClient(*client_args)
(APIServer pid=2490067) ^^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core_client.py”, line 769, in init
(APIServer pid=2490067) super().init(
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/core_client.py”, line 448, in init
(APIServer pid=2490067) with launch_core_engines(vllm_config, executor_class,
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/contextlib.py”, line 144, in exit
(APIServer pid=2490067) next(self.gen)
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/utils.py”, line 732, in launch_core_engines
(APIServer pid=2490067) wait_for_engine_startup(
(APIServer pid=2490067) File “/data/conda/envs/llm311/lib/python3.11/site-packages/vllm/v1/engine/utils.py”, line 785, in wait_for_engine_startup
(APIServer pid=2490067) raise RuntimeError("Engine core initialization failed. "
(APIServer pid=2490067) RuntimeError: Engine core initialization failed. See root cause above. Failed core proc(s): {}
(llm311) nvidia@localhost:~$
我在conda通过cudatoolkit下载了cuda12.9，cuda环境下安装torch12.8，python3.11，vllm0.11.0
求救大佬，这个怎么解决。谢谢回答。