Can not init Myriad device: NC_ERROR in Raspberry Pi

[ARawa7]

Hi

I have installed OpenVino in Raspberry pi. I have attached Neural stick 2 to raspberry pi and can confirm this using lsusb command. Below is the results:

Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 004: ID 0c45:6366 Microdia 
Bus 001 Device 015: ID 03e7:2485 Intel Movidius MyriadX
Bus 001 Device 002: ID 2109:3431 VIA Labs, Inc. Hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
 

I am running the multi camera multi person python demo. I am running it using the below command:

python3 multi_camera_multi_person_tracking.py -i 0 -m person-detection-retail-0013.xml --m_reid person-reidentification-retail-0031.xml --config config.py -d MYRIAD

I have mentioned MYRIAD to use the as the target device which is basically the Neural stick. But I am getting below error:

INFO:root:Connection  cam 0
1
INFO:root:Creating Inference Engine
INFO:root:Loading network files:
    person-detection-retail-0013.xml
    person-detection-retail-0013.bin
INFO:root:Preparing input blobs
INFO:root:Loading model to the plugin
INFO:root:Creating Inference Engine
INFO:root:Loading network files:
    person-reidentification-retail-0031.xml
    person-reidentification-retail-0031.bin
INFO:root:Preparing input blobs
INFO:root:Loading model to the plugin
Traceback (most recent call last):
  File "multi_camera_multi_person_tracking.py", line 164, in <module>
    main()
  File "multi_camera_multi_person_tracking.py", line 156, in main
    person_recognizer = VectorCNN(args.m_reid, args.device)
  File "/home/pi/Desktop/multi_camera_multi_person_tracking/utils/network_wrappers.py", line 80, in __init__
    self.net = load_ie_model(model_path, device, None, num_reqs=self.max_reqs)
  File "/home/pi/Desktop/multi_camera_multi_person_tracking/utils/ie_tools.py", line 96, in load_ie_model
    exec_net = ie.load_network(network=net, device_name=device, num_requests=num_reqs)
  File "ie_api.pyx", line 85, in openvino.inference_engine.ie_api.IECore.load_network
  File "ie_api.pyx", line 92, in openvino.inference_engine.ie_api.IECore.load_network
RuntimeError: Can not init Myriad device: NC_ERROR
 

Can anyone please help me in getting MYRIAD to run with this python demo.

Thanks

[RandallMan_B_Intel]

Hi AR,

Thank you for reaching out.

Could you please tell us the following:

- Which OS and Openvino version are you using?

- Did you add the USB Rules on the official guide?

- Are you connecting the NCS directly to the host machine or using a powered USB hub?

Regards,

Randall B.

[gilgo]

Hi Randall,

I've encountered the same issue but with a custom model. I want to switch the model running inference on the NCS at regular intervals, the first time initialize the inferenceEngine it works but the second time, the call:

exec_net = self.ie.load_network(network=self.net, device_name=self.device)

 causes the following error:

File "/home/pi/PycharmProjects/network-polaroid-2.3/src/polaroid/communication/stylizer_controller.py", line 16, in swap_style
self.stylizer.set_style(self.style_counter)
File "/home/pi/PycharmProjects/network-polaroid-2.3/src/polaroid/stylizer.py", line 30, in set_style
self.inferenceEngine = InferenceEngine(path_to_model)
File "/home/pi/PycharmProjects/network-polaroid-2.3/src/polaroid/inference_engine.py", line 15, in __init__
self.executable_net = self.get_executable_network(path_to_model)
File "/home/pi/PycharmProjects/network-polaroid-2.3/src/polaroid/inference_engine.py", line 21, in get_executable_network
exec_net = self.ie.load_network(network=self.net, device_name=self.device)
File "ie_api.pyx", line 314, in openvino.inference_engine.ie_api.IECore.load_network
File "ie_api.pyx", line 323, in openvino.inference_engine.ie_api.IECore.load_network
RuntimeError: Can not init Myriad device: NC_ERROR

• I'm using openVINO 2020.4 on both the raspberry pi and on full version for model optimizer.
• I did add the USB Rules, and I am managing to run inference, just not switching out the models
• The NCS is directly connected to the pi but I have a USB hub for other things.

Is it the case that swapping models for inference is more complicated or am I missing something?

Thanks in advance,

gilgo