[Bug]: Dynamo Unsupported due to BasevLLMParameter.torch_function calling disabled super() #25604
Description
Your current environment
The output of python collect_env.py
Your output of `python collect_env.py` here
🐛 Describe the bug
vllm bench throughput --model Qwen/Qwen3-30B-A3B-FP8 --load-format dummy --input-len 1000 --output-len 100 --trust_remote_code --enable-expert-parallel
(EngineCore_DP0 pid=2439401) self.run()
(EngineCore_DP0 pid=2439401) File "/usr/lib/python3.12/multiprocessing/process.py", line 108, in run
(EngineCore_DP0 pid=2439401) self._target(*self._args, **self._kwargs)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/engine/core.py", line 712, in run_engine_core
(EngineCore_DP0 pid=2439401) raise e
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/engine/core.py", line 695, in run_engine_core
(EngineCore_DP0 pid=2439401) engine_core = DPEngineCoreProc(*args, **kwargs)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/engine/core.py", line 965, in __init__
(EngineCore_DP0 pid=2439401) super().__init__(vllm_config, local_client, handshake_address,
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/engine/core.py", line 498, in __init__
(EngineCore_DP0 pid=2439401) super().__init__(vllm_config, executor_class, log_stats,
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/engine/core.py", line 92, in __init__
(EngineCore_DP0 pid=2439401) self._initialize_kv_caches(vllm_config)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/engine/core.py", line 190, in _initialize_kv_caches
(EngineCore_DP0 pid=2439401) self.model_executor.determine_available_memory())
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/executor/abstract.py", line 85, in determine_available_memory
(EngineCore_DP0 pid=2439401) return self.collective_rpc("determine_available_memory")
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/executor/uniproc_executor.py", line 83, in collective_rpc
(EngineCore_DP0 pid=2439401) return [run_method(self.driver_worker, method, args, kwargs)]
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/utils/__init__.py", line 3049, in run_method
(EngineCore_DP0 pid=2439401) return func(*args, **kwargs)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/.venv/lib/python3.12/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
(EngineCore_DP0 pid=2439401) return func(*args, **kwargs)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/worker/gpu_worker.py", line 271, in determine_available_memory
(EngineCore_DP0 pid=2439401) self.model_runner.profile_run()
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/worker/gpu_model_runner.py", line 3292, in profile_run
(EngineCore_DP0 pid=2439401) = self._dummy_run(self.max_num_tokens, is_profile=True)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/.venv/lib/python3.12/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
(EngineCore_DP0 pid=2439401) return func(*args, **kwargs)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/v1/worker/gpu_model_runner.py", line 3069, in _dummy_run
(EngineCore_DP0 pid=2439401) outputs = self.model(
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
(EngineCore_DP0 pid=2439401) return self._call_impl(*args, **kwargs)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/.venv/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1762, in _call_impl
(EngineCore_DP0 pid=2439401) return forward_call(*args, **kwargs)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/models/deepseek_v2.py", line 907, in forward
(EngineCore_DP0 pid=2439401) hidden_states = self.model(input_ids, positions, intermediate_tensors,
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/compilation/decorators.py", line 310, in __call__
(EngineCore_DP0 pid=2439401) output = self.compiled_callable(*args, **kwargs)
(EngineCore_DP0 pid=2439401) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=2439401) File "/home/wentao/.venv/lib/python3.12/site-packages/torch/_dynamo/eval_frame.py", line 659, in _fn
(EngineCore_DP0 pid=2439401) raise e.with_traceback(None) from None
(EngineCore_DP0 pid=2439401) torch._dynamo.exc.Unsupported: non-function or method super: <built-in function _disabled_torch_function_impl>
(EngineCore_DP0 pid=2439401)
(EngineCore_DP0 pid=2439401) from user code:
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/models/deepseek_v2.py", line 783, in forward
(EngineCore_DP0 pid=2439401) hidden_states, residual = layer(positions, hidden_states, residual)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/models/deepseek_v2.py", line 711, in forward
(EngineCore_DP0 pid=2439401) hidden_states = self.mlp(hidden_states)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/models/deepseek_v2.py", line 285, in forward
(EngineCore_DP0 pid=2439401) router_logits, _ = self.gate(hidden_states)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/layers/linear.py", line 381, in forward
(EngineCore_DP0 pid=2439401) output = self.quant_method.apply(self, x, bias)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/layers/linear.py", line 236, in apply
(EngineCore_DP0 pid=2439401) return dispatch_unquantized_gemm()(layer, x, layer.weight, bias)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/layers/utils.py", line 92, in default_unquantized_gemm
(EngineCore_DP0 pid=2439401) return torch.nn.functional.linear(x, weight, bias)
(EngineCore_DP0 pid=2439401) File "/home/wentao/vllm-source/vllm/model_executor/parameter.py", line 119, in __torch_function__
(EngineCore_DP0 pid=2439401) return super().__torch_function__(func, types, args, kwargs)
(EngineCore_DP0 pid=2439401)Before submitting a new issue...
- Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.