模型单独跑预测时是正常的，但是开多线程执行预测时出现如下错误：

Resources released

Traceback (most recent call last):

  File "/home/HwHiAiUser/samples/notebooks/11_ambulance_detection/frame_processor.py", line 50, in process_frames

    results = self.model.predict(frame_data["frame"])

  File "/home/HwHiAiUser/samples/notebooks/11_ambulance_detection/yolo/yolov8_acl.py", line 301, in predict

    self._data_from_host_to_device(input_data)

  File "/home/HwHiAiUser/samples/notebooks/11_ambulance_detection/yolo/yolov8_acl.py", line 182, in _data_from_host_to_device

    self._data_interaction(images, ACL_MEMCPY_HOST_TO_DEVICE)

  File "/home/HwHiAiUser/samples/notebooks/11_ambulance_detection/yolo/yolov8_acl.py", line 147, in _data_interaction

    self._check_ret("acl.rt.memcpy", ret)

  File "/home/HwHiAiUser/samples/notebooks/11_ambulance_detection/yolo/yolov8_acl.py", line 353, in _check_ret

    raise RuntimeError(f"{message} failed, ret={ret}")

RuntimeError: acl.rt.memcpy failed, ret=107002

python-BaseException

ACL resources were released successfully.

Traceback (most recent call last):

  File "/home/HwHiAiUser/samples/notebooks/11_ambulance_detection/ambulance_detection.py", line 108, in <module>

    ret = acl.util.stop_thread(frame_processor_thread)

RuntimeError: exception=RuntimeError,msg=thread call failed;