InferenceEngineException on raspberrypi 'Product of output shape dimensions does not match product of argument shape dimensions'

[pchilders]

When following the instructions on https://docs.openvinotoolkit.org/latest/_docs_install_guides_installing_openvino_raspbian.html I get the following errors.
I was able to test the NCS on Mac OS, so it seems to be related to the raspberry pi build.

pi@raspberrypi:~/build $ python3 openvino_fd_myrad.py 
terminate called after throwing an instance of 'InferenceEngine::details::InferenceEngineException'
  what():  conv4_3_0_norm_mbox_priorbox/naked_not_unsqueezed has zero dimension that is not allowable
Aborted

pi@raspberrypi:~/build $ python3 openvino_fd_myrad.py 
terminate called after throwing an instance of 'InferenceEngine::details::InferenceEngineException'
  what():  Check 'input_shape_product.get_length() == shape_size(m_output_shape)' failed at /teamcity/work/scoring_engine_build/releases_2020_2/ngraph/src/ngraph/op/reshape.cpp:100:
While validating node 'v0::Reshape Reshape_980(PriorBox_754[0]:f32{2,3231918080}) -> (dynamic?)':
Product of output shape dimensions does not match product of argument shape dimensions (output shape: Shape{1, 2, 3231918080}, argument shape: {2,3231918080}).

Any idea what is the error?

[JAVIERJOSE_A_Intel]

Hi (name withheld),

Thanks for reaching out.

We will take a look into your issue and come back to you as soon as possible.

Regards,

Javier A.

[gupta__himant]

Hello,

I have the same error, while trying to run the example on PI.

pi@raspberrypi:~/guardeyelite $ python3 openvino_fd_myriad.py 
terminate called after throwing an instance of 'InferenceEngine::details::InferenceEngineException'
  what():  Check 'input_shape_product.get_length() == shape_size(m_output_shape)' failed at /teamcity/work/scoring_engine_build/releases_2020_2/ngraph/src/ngraph/op/reshape.cpp:100:
While validating node 'v0::Reshape Reshape_971(PriorBox_744[0]:f32{2,4250587136}) -> (dynamic?)':
Product of output shape dimensions does not match product of argument shape dimensions (output shape: Shape{1, 2, 4250587136}, argument shape: {2,4250587136}).

Regards
 

[JAVIERJOSE_A_Intel]

Hi (name withheld), himant,

We are aware of this issue and are able to reproduce it, we are currently investigating the problem and looking for a fix. As a workaround, you can use the OpenVINO™ toolkit 2020.1 version with 2019 R3 version models. This is because there is an incompatibility issue between OpenVINO™ toolkit 2020.1 version and IR v10 format files. Also, this only affects OpenVINO™ toolkit on Raspberry Pi.

Regards,

Javier A.

[la__xl]

I encounter similar issue:

raspberry pi 3B+, 

openvino,  inference_engine_build_2020.3

python 3.5.3

  File "/home/pi/test/baseInferEngine.py", line 80, in initIE
    self.exec_net = self.ie.load_network(network=self.net, device_name=self.device, num_requests=self.num_requests)
  File "ie_api.pyx", line 178, in openvino.inference_engine.ie_api.IECore.load_network
  File "ie_api.pyx", line 187, in openvino.inference_engine.ie_api.IECore.load_network
RuntimeError: Check 'input_shape_product.get_length() == shape_size(m_output_shape)' failed at /home/jenkins/agent/workspace/private-ci/ie/build-linux-debian_9_arm/b/repos/closed-dldt/ngraph/src/ngraph/op/reshape.cpp:100:
While validating node 'v0::Reshape Reshape_17724(PriorBoxClustered_15181[0]:f32{2,4042567040}) -> (dynamic?)':
Product of output shape dimensions does not match product of argument shape dimensions (output shape: Shape{1, 2, 4042567040}, argument shape: {2,4042567040}).

this model with same python code can run successful on windows 10 , python 3.7.7