I am trying to run serve.run in my test method but when the test runs, it throws an error in this part of the code:

serve.run(RayFastApiWrapper.bind(), route_prefix=settings.app_root_url)

FastApiWrapper looks like this:

@serve.deployment(ray_actor_options={"num_gpus": 1, "num_cpus": 8})
@serve.ingress(app)
class RayFastApiWrapper:
    pass

And the module causing the exception seems to be in module ray/_private/accelerators/nvidia_gpu.py

StackTrace:

serve.run(RayFastApiWrapper.bind(), route_prefix=settings.app_root_url)
/home/user/.local/lib/python3.9/site-packages/ray/serve/api.py:578: in run
    handle = _run(
/home/user/.local/lib/python3.9/site-packages/ray/serve/api.py:484: in _run
    client = _private_api.serve_start(
/home/user/.local/lib/python3.9/site-packages/ray/serve/_private/api.py:257: in serve_start
    client = _get_global_client(_health_check_controller=True)
/home/user/.local/lib/python3.9/site-packages/ray/serve/context.py:87: in _get_global_client
    return _connect(raise_if_no_controller_running)
/home/user/.local/lib/python3.9/site-packages/ray/serve/context.py:132: in _connect
    ray.init(namespace=SERVE_NAMESPACE)
/home/user/.local/lib/python3.9/site-packages/ray/_private/client_mode_hook.py:103: in wrapper
    return func(*args, **kwargs)
/home/user/.local/lib/python3.9/site-packages/ray/_private/worker.py:1642: in init
    _global_node = ray._private.node.Node(
/home/user/.local/lib/python3.9/site-packages/ray/_private/node.py:336: in __init__
    self.start_ray_processes()
/home/user/.local/lib/python3.9/site-packages/ray/_private/node.py:1396: in start_ray_processes
    resource_spec = self.get_resource_spec()
/home/user/.local/lib/python3.9/site-packages/ray/_private/node.py:571: in get_resource_spec
    self._resource_spec = ResourceSpec(
/home/user/.local/lib/python3.9/site-packages/ray/_private/resource_spec.py:215: in resolve
    accelerator_manager.get_current_node_accelerator_type()
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

    @staticmethod
    def get_current_node_accelerator_type() -> Optional[str]:
        import ray._private.thirdparty.pynvml as pynvml
    
        try:
            pynvml.nvmlInit()
        except pynvml.NVMLError:
            return None  # pynvml init failed
        device_count = pynvml.nvmlDeviceGetCount()
        cuda_device_type = None
        if device_count > 0:
            handle = pynvml.nvmlDeviceGetHandleByIndex(0)
            device_name = pynvml.nvmlDeviceGetName(handle)
            if isinstance(device_name, bytes):
>               device_name = device_name.decode("utf-8")
E               UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf8 in position 0: invalid start byte

I am using ray 2.23.0 and gpustat 1.0. However, seems ray uses its own fork of pynvml, judging from the code in the repo.