如题，采用vllm-ascend跑Qwen3-Omni-30B-A3B-Instruct镜像报错，镜像分别使用了v0.10.2rc1和latest版本均报错，报错内容如下。
INFO 09-24 17:02:45 [__init__.py:36] Available plugins for group vllm.platform_plugins:
INFO 09-24 17:02:45 [__init__.py:38] - ascend -> vllm_ascend:register
INFO 09-24 17:02:45 [__init__.py:41] All plugins in this group will be loaded. Set `VLLM_PLUGINS` to control which plugins to load.
INFO 09-24 17:02:45 [__init__.py:207] Platform plugin ascend is activated
WARNING 09-24 17:02:47 [_custom_ops.py:20] Failed to import from vllm._C with ModuleNotFoundError("No module named 'vllm._C'")
WARNING 09-24 17:02:47 [registry.py:483] Model architecture Qwen2VLForConditionalGeneration is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen2_vl:AscendQwen2VLForConditionalGeneration.
WARNING 09-24 17:02:47 [registry.py:483] Model architecture Qwen2_5_VLForConditionalGeneration is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen2_5_vl:AscendQwen2_5_VLForConditionalGeneration.
WARNING 09-24 17:02:47 [registry.py:483] Model architecture DeepseekV2ForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.deepseek_v2:CustomDeepseekV2ForCausalLM.
WARNING 09-24 17:02:47 [registry.py:483] Model architecture DeepseekV3ForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.deepseek_v2:CustomDeepseekV3ForCausalLM.
WARNING 09-24 17:02:48 [registry.py:483] Model architecture DeepSeekMTPModel is already registered, and will be overwritten by the new model class vllm_ascend.models.deepseek_mtp:CustomDeepSeekMTP.
WARNING 09-24 17:02:48 [registry.py:483] Model architecture Qwen3MoeForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen3_moe:CustomQwen3MoeForCausalLM.
WARNING 09-24 17:02:48 [registry.py:483] Model architecture Qwen3NextForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen3_next:Qwen3NextForCausalLM.
INFO 09-24 17:02:48 [importing.py:63] Triton not installed or not compatible; certain GPU-related functions will not be available.
(APIServer pid=660) INFO 09-24 17:02:48 [api_server.py:1896] vLLM API server version 0.10.2
(APIServer pid=660) INFO 09-24 17:02:48 [utils.py:328] non-default args: {'model_tag': '/data/Qwen3-Omni-30B-A3B-Instruct', 'host': '0.0.0.0', 'model': '/data/Qwen3-Omni-30B-A3B-Instruct', 'trust_remote_code': True, 'dtype': 'float16', 'allowed_local_media_path': '/data', 'max_model_len': 43690, 'tensor_parallel_size': 4, 'max_num_batched_tokens': 16384, 'max_num_seqs': 2}
(APIServer pid=660) The argument `trust_remote_code` is to be used with Auto classes. It has no effect here and is ignored.
(APIServer pid=660) Traceback (most recent call last):
(APIServer pid=660) File "/usr/local/python3.11.13/bin/vllm", line 7, in <module>
(APIServer pid=660) sys.exit(main())
(APIServer pid=660) ^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/cli/main.py", line 54, in main
(APIServer pid=660) args.dispatch_function(args)
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/cli/serve.py", line 50, in cmd
(APIServer pid=660) uvloop.run(run_server(args))
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/site-packages/uvloop/__init__.py", line 105, in run
(APIServer pid=660) return runner.run(wrapper())
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/asyncio/runners.py", line 118, in run
(APIServer pid=660) return self._loop.run_until_complete(task)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "uvloop/loop.pyx", line 1518, in uvloop.loop.Loop.run_until_complete
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/site-packages/uvloop/__init__.py", line 61, in wrapper
(APIServer pid=660) return await main
(APIServer pid=660) ^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 1941, in run_server
(APIServer pid=660) await run_server_worker(listen_address, sock, args, **uvicorn_kwargs)
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 1961, in run_server_worker
(APIServer pid=660) async with build_async_engine_client(
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/contextlib.py", line 210, in __aenter__
(APIServer pid=660) return await anext(self.gen)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 179, in build_async_engine_client
(APIServer pid=660) async with build_async_engine_client_from_engine_args(
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/contextlib.py", line 210, in __aenter__
(APIServer pid=660) return await anext(self.gen)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 205, in build_async_engine_client_from_engine_args
(APIServer pid=660) vllm_config = engine_args.create_engine_config(usage_context=usage_context)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/engine/arg_utils.py", line 1119, in create_engine_config
(APIServer pid=660) model_config = self.create_model_config()
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/engine/arg_utils.py", line 963, in create_model_config
(APIServer pid=660) return ModelConfig(
(APIServer pid=660) ^^^^^^^^^^^^
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/site-packages/pydantic/_internal/_dataclasses.py", line 123, in __init__
(APIServer pid=660) s.__pydantic_validator__.validate_python(ArgsKwargs(args, kwargs), self_instance=s)
(APIServer pid=660) pydantic_core._pydantic_core.ValidationError: 1 validation error for ModelConfig
(APIServer pid=660) Value error, The checkpoint you are trying to load has model type `qwen3_omni_moe` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.
(APIServer pid=660)
(APIServer pid=660) You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git` [type=value_error, input_value=ArgsKwargs((), {'model': ...rocessor_plugin': None}), input_type=ArgsKwargs]
(APIServer pid=660) For further information visit https://errors.pydantic.dev/2.11/v/value_error
(APIServer pid=660) [ERROR] 2025-09-24-17:02:48 (PID:660, Device:-1, RankID:-1) ERR99999 UNKNOWN applicaiton exception
点赞
收藏
回复
分享
举报
已解决
如题，采用vllm-ascend跑Qwen3-Omni-30B-A3B-Instruct镜像报错，镜像分别使用了v0.10.2rc1和latest版本均报错，报错内容如下。
INFO 09-24 17:02:45 [__init__.py:36] Available plugins for group vllm.platform_plugins:
INFO 09-24 17:02:45 [__init__.py:38] - ascend -> vllm_ascend:register
INFO 09-24 17:02:45 [__init__.py:41] All plugins in this group will be loaded. Set `VLLM_PLUGINS` to control which plugins to load.
INFO 09-24 17:02:45 [__init__.py:207] Platform plugin ascend is activated
WARNING 09-24 17:02:47 [_custom_ops.py:20] Failed to import from vllm._C with ModuleNotFoundError("No module named 'vllm._C'")
WARNING 09-24 17:02:47 [registry.py:483] Model architecture Qwen2VLForConditionalGeneration is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen2_vl:AscendQwen2VLForConditionalGeneration.
WARNING 09-24 17:02:47 [registry.py:483] Model architecture Qwen2_5_VLForConditionalGeneration is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen2_5_vl:AscendQwen2_5_VLForConditionalGeneration.
WARNING 09-24 17:02:47 [registry.py:483] Model architecture DeepseekV2ForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.deepseek_v2:CustomDeepseekV2ForCausalLM.
WARNING 09-24 17:02:47 [registry.py:483] Model architecture DeepseekV3ForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.deepseek_v2:CustomDeepseekV3ForCausalLM.
WARNING 09-24 17:02:48 [registry.py:483] Model architecture DeepSeekMTPModel is already registered, and will be overwritten by the new model class vllm_ascend.models.deepseek_mtp:CustomDeepSeekMTP.
WARNING 09-24 17:02:48 [registry.py:483] Model architecture Qwen3MoeForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen3_moe:CustomQwen3MoeForCausalLM.
WARNING 09-24 17:02:48 [registry.py:483] Model architecture Qwen3NextForCausalLM is already registered, and will be overwritten by the new model class vllm_ascend.models.qwen3_next:Qwen3NextForCausalLM.
INFO 09-24 17:02:48 [importing.py:63] Triton not installed or not compatible; certain GPU-related functions will not be available.
(APIServer pid=660) INFO 09-24 17:02:48 [api_server.py:1896] vLLM API server version 0.10.2
(APIServer pid=660) INFO 09-24 17:02:48 [utils.py:328] non-default args: {'model_tag': '/data/Qwen3-Omni-30B-A3B-Instruct', 'host': '0.0.0.0', 'model': '/data/Qwen3-Omni-30B-A3B-Instruct', 'trust_remote_code': True, 'dtype': 'float16', 'allowed_local_media_path': '/data', 'max_model_len': 43690, 'tensor_parallel_size': 4, 'max_num_batched_tokens': 16384, 'max_num_seqs': 2}
(APIServer pid=660) The argument `trust_remote_code` is to be used with Auto classes. It has no effect here and is ignored.
(APIServer pid=660) Traceback (most recent call last):
(APIServer pid=660) File "/usr/local/python3.11.13/bin/vllm", line 7, in <module>
(APIServer pid=660) sys.exit(main())
(APIServer pid=660) ^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/cli/main.py", line 54, in main
(APIServer pid=660) args.dispatch_function(args)
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/cli/serve.py", line 50, in cmd
(APIServer pid=660) uvloop.run(run_server(args))
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/site-packages/uvloop/__init__.py", line 105, in run
(APIServer pid=660) return runner.run(wrapper())
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/asyncio/runners.py", line 118, in run
(APIServer pid=660) return self._loop.run_until_complete(task)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "uvloop/loop.pyx", line 1518, in uvloop.loop.Loop.run_until_complete
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/site-packages/uvloop/__init__.py", line 61, in wrapper
(APIServer pid=660) return await main
(APIServer pid=660) ^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 1941, in run_server
(APIServer pid=660) await run_server_worker(listen_address, sock, args, **uvicorn_kwargs)
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 1961, in run_server_worker
(APIServer pid=660) async with build_async_engine_client(
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/contextlib.py", line 210, in __aenter__
(APIServer pid=660) return await anext(self.gen)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 179, in build_async_engine_client
(APIServer pid=660) async with build_async_engine_client_from_engine_args(
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/contextlib.py", line 210, in __aenter__
(APIServer pid=660) return await anext(self.gen)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/entrypoints/openai/api_server.py", line 205, in build_async_engine_client_from_engine_args
(APIServer pid=660) vllm_config = engine_args.create_engine_config(usage_context=usage_context)
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/engine/arg_utils.py", line 1119, in create_engine_config
(APIServer pid=660) model_config = self.create_model_config()
(APIServer pid=660) ^^^^^^^^^^^^^^^^^^^^^^^^^^
(APIServer pid=660) File "/vllm-workspace/vllm/vllm/engine/arg_utils.py", line 963, in create_model_config
(APIServer pid=660) return ModelConfig(
(APIServer pid=660) ^^^^^^^^^^^^
(APIServer pid=660) File "/usr/local/python3.11.13/lib/python3.11/site-packages/pydantic/_internal/_dataclasses.py", line 123, in __init__
(APIServer pid=660) s.__pydantic_validator__.validate_python(ArgsKwargs(args, kwargs), self_instance=s)
(APIServer pid=660) pydantic_core._pydantic_core.ValidationError: 1 validation error for ModelConfig
(APIServer pid=660) Value error, The checkpoint you are trying to load has model type `qwen3_omni_moe` but Transformers does not recognize this architecture. This could be because of an issue with the checkpoint, or because your version of Transformers is out of date.
(APIServer pid=660)
(APIServer pid=660) You can update Transformers with the command `pip install --upgrade transformers`. If this does not work, and the checkpoint is very new, then there may not be a release version that supports this model yet. In this case, you can get the most up-to-date code by installing Transformers from source with the command `pip install git+https://github.com/huggingface/transformers.git` [type=value_error, input_value=ArgsKwargs((), {'model': ...rocessor_plugin': None}), input_type=ArgsKwargs]
(APIServer pid=660) For further information visit https://errors.pydantic.dev/2.11/v/value_error
(APIServer pid=660) [ERROR] 2025-09-24-17:02:48 (PID:660, Device:-1, RankID:-1) ERR99999 UNKNOWN applicaiton exception