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

Having some problem when optimize a gru model

5xx
Beginner
639 Views

Hi, I have some problems when optimize a model. The log shows is as follows: Exception occurred during running replacer "REPLACEMENT_ID" (<class 'extensions.middle.RNNSequenceNormalizeToIE.RNNSequenceNormalize'>): Something bad has happened with graph! Data node "gru_1" has 0 producers.

My version is2021.2.185

Here is my environment :

pyyaml
tornado==5.1.1
python-interface
python-consul
oss2
requests
numpy
gensim
keras==2.2.4
pandas
tensorflow==1.12.0

onnx==1.8.1

keras2onnx==1.7.0

 

0 Kudos
6 Replies
Iffa_Intel
Moderator
621 Views

Hi,

 

Could you provide which model and also demo application that you are using for further investigation?

Plus, your model's topology and framework layer.

 

Sincerely,

Iffa

 

0 Kudos
5xx
Beginner
618 Views

My openVINO version is 2021.2.185 .I have already provided my model struct but the model file cannot provide. I didn't run the demo because the error happened when model optimizing.

0 Kudos
Iffa_Intel
Moderator
608 Views

Since you can't share it, I would like to confirm whether your current model is using dynamic shape for the input?

 

If you are, the dynamic shape for RNN/GRU for ONNX model is yet to be supported. However, our developers have worked on this in Feb 2021 and possible implementation on the next version release of OV - ONNX RNN/GRU enable dynamic input shape by mitruska · Pull Request #4241 · openvinotoolkit/openvino (github.com)

 

Plus, I would like to clarify your model's topology and framework layer to ensure they are supported.

 

Sincerely,

Iffa

 

0 Kudos
5xx
Beginner
604 Views

Yes, the model is using dynamic shape.

0 Kudos
Iffa_Intel
Moderator
598 Views

Hope you are noted regarding the dynamic shape for RNN/GRU for ONNX model is yet to be supported,

as I mentioned in the previous reply.



Sincerely,

Iffa


0 Kudos
Iffa_Intel
Moderator
555 Views

Greetings,


Intel will no longer monitor this thread since we have provided a solution. If you need any additional information from Intel, please submit a new question. 


Sincerely,

Iffa


0 Kudos
Reply