Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6392 Discussions

Error AssertionFailed: !ieDims.empty() when apply pre-train model (Reinforcement Learning) on NCS2

LAndr23
Beginner
474 Views

I have a problem when I try to run my model Reinforcement Learning (RL) with LSTM on VPU (NCS2). It cannot load my network to the device (plugin = IEPlugin(device="MYRIAD")). But if I change device=CPU or GPU, It’ll work well.

 

This is my code I face the issue :

 

from openvino.inference_engine import IENetwork, IEPlugin

plugin = IEPlugin(device="MYRIAD")

net = IENetwork(model=model_xml, weights=model_bin)

exec_net = plugin.load(network=net)

 

This is the error I got :

Traceback (most recent call last):

 File "NCS2_UTL_test.py", line 68, in <module>

  exec_net = plugin.load(network=net)

 File "ie_api.pyx", line 1205, in openvino.inference_engine.ie_api.IEPlugin.load

 File "ie_api.pyx", line 1215, in openvino.inference_engine.ie_api.IEPlugin.load

RuntimeError: AssertionFailed: !ieDims.empty()

 

So, can you give me the reason and how can I fix it if I want to run it on VPU?

Thank you

 

0 Kudos
4 Replies
RandallMan_B_Intel
374 Views

Hi Andrew,

 

It seems like the layer is not supported by MYRIAD plugin; you can check that here. However, check this thread can be helpful.

 

Regards,

Randall B.

0 Kudos
LAndr23
Beginner
374 Views

Hi Randall B,

I tried to use the thread but it's still not working. You montioned the layer is not supported by MYRIAD, so how can I show exactly what's the name of layer is not supported?

Thanks.

0 Kudos
LAndr23
Beginner
374 Views

Additionly, This's my error I got when trying to through the xml replacing : 

 

<output>

                <port id="1" precision="I64"/>

</output>

To

 

            <output>

                <port id="1" precision="I64">

                    <dim>1</dim>

                </port>

            </output>

 

 

Traceback (most recent call last):

 File "NCS2_UTL_test.py", line 59, in <module>

  net = IENetwork(model=model_xml, weights=model_bin)

 File "ie_api.pyx", line 980, in openvino.inference_engine.ie_api.IENetwork.__cinit__

RuntimeError: Check 'axis_ps.rank().is_static() && (size_t)axis_ps.rank() == 0' failed at ..\ngraph\src\ngraph\op\fused\split.cpp:133:

While validating node 'Split[Split_148](Reshape_146: float{1,248}, Constant_147: int64_t{1}) -> (??)':

The 'axis' input is expected to be a scalar. Got: {1}

 

This error occurs both running on CPU device and VPU when changing the xml.

And I can still run on the CPU if I don't change the xml file according to your instructions.

 

Thanks

0 Kudos
RandallMan_B_Intel
374 Views

Hi Andrew,

 

Unfortunately, the LSTM is not supported by the MYRIAD plugin (NCS2).

Let me know if you have additional questions.

 

Regards,

Randall B.

0 Kudos
Reply