【算法工程】vllm突然报错No platform detected, vLLM is running on UnspecifiedPlatform

1. 问题背景

        前一秒vllm还在正常运行服务，关闭了服务，想重新启动，突然就报No platform detected, vLLM is running on UnspecifiedPlatform：

Traceback (most recent call last):
  File "/home/research/miniconda3/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap
    self.run()
  File "/home/research/miniconda3/lib/python3.10/multiprocessing/process.py", line 108, in run
    self._target(*self._args, **self._kwargs)
  File "/home/research/miniconda3/lib/python3.10/site-packages/vllm/engine/multiprocessing/engine.py", line 391, in run_mp_engine
    raise e
  File "/home/research/miniconda3/lib/python3.10/site-packages/vllm/engine/multiprocessing/engine.py", line 380, in run_mp_engine
    engine = MQLLMEngine.from_engine_args(engine_args=engine_args,
  File "/home/research/miniconda3/lib/python3.10/site-packages/vllm/engine/multiprocessing/engine.py", line 118, in from_engine_args
    engine_config = engine_args.create_engine_config(usage_context)
  File "/home/research/miniconda3/lib/python3.10/site-packages/vllm/engine/arg_utils.py", line 1074, in create_engine_config
    device_config = DeviceConfig(device=self.device)
  File "/home/research/miniconda3/lib/python3.10/site-packages/vllm/config.py", line 1626, in __init__
    raise RuntimeError("Failed to infer device type")
RuntimeError: Failed to infer device type

        尝试了主流的大模型，无非就是让重新安装。然后还是直接到vllm的开源项目，找对应的issue，发现这是个常见的问题。解决方案也很简单。

        安装：

pip install pynvml==12.0.0

2. 参考材料

【1】https://github.com/vllm-project/vllm/issues/12847

【2】https://github.com/vllm-project/vllm/pull/12809