RuntimeError when loading the network to the GPU

[Wenjing]

The loading works perfectly on CPU, however, when I tried to load the network to the GPU plugin, it outputs a RuntimeError:

    exec_net = ie.load_network(network=net, device_name=args.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: Unsupported number of dimensions: 0

Has anyone encountered such an error before? Looking forward to some help, many thanks!

OpenVINO: 2020.4 Release

python: python 3.7

OS: Ubuntu 16.04 & Ubuntu 18.04

GPUs I've tried on:

Atom E3900 Series Integrated Graphics Controller

Intel Corporation HD Graphics 530

CPUs:

Intel(R) Atom(TM) Processor E3950 @ 1.60GHz

Intel(R) Core(TM) i7-6700T CPU @ 2.80GHz

[Rizal_Intel]

Hi Wenjing,

Just covering the basics first,

have you run the steps for Steps for Intel® Processor Graphics?

Next, try checking if there are layers not in the supported layers for GPU.

Do you mind sharing your IR?

I would like to try running it in the benchmark app to see if it works on other hardware or OS.

Regards,

Rizal

[Rizal_Intel]

Hi Wenjing,

Thank you for your question. If you need any additional information from Intel, please submit a new question as Intel is no longer monitoring this thread.

Regards,

Rizal

[Wenjing]

Hi Rizal,

Sorry for the late reply.

I've run the Steps for Intel® Processor Graphics.

Also, previously, I've successfully loaded other models to the GPU plugin, which mainly consists of convolutional layers and fully connected layers.

I'm willing to share the IR with you via email. Could you please tell me your email address?

Thanks.