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.
6257 Discussions

Conversion of onnx model with legacy IR version

MicheleBolognini
Beginner
655 Views

Hi,

I am trying to convert a model from .onnx format into OpenVINO, for use with an old software framework.
I know the Model Optimizer should be used for this, but if I try

mo --input_model model.onnx

then the new model gets generated with IR version 11. Unfortunately, my old framework does not support newer IR versions. Is there a way of forcing the conversion with an older IR version?

 

Thanks.

Labels (2)
0 Kudos
3 Replies
Hairul_Intel
Moderator
624 Views

Hi MicheleBolognini,

Thank you for reaching out to us.

 

For your information, OpenVINO™ Toolkit 2022.1 release does provide backward compatibility support for reading IR v10 model. However, it is not possible to convert model into older IR v10 using the new OpenVINO™ Toolkit 2022.1 release.

 

In order to convert a model into legacy IR v10 model, you'd need to use Model Optimizer from OpenVINO™ Toolkit 2021.4.2 release. You can follow the guide on Converting a ONNX* Model for further information.

 

Here is the result for converting a resnet50 ONNX model into IR v10 model using OpenVINO™ Toolkit 2021.4.2:

result.png

 

 

 

Regards,

Hairul

 

0 Kudos
MicheleBolognini
Beginner
616 Views

Thank you @Hairul_Intel for the clarification.

I noticed that, though I had installed the 2020.3 Release, the Model Optimizer version was 2022.1.something, probably a carryover from a previous installation.
I will thoroughly clean up the installed version and make sure I install and older release before I try again.

Thanks
Michele

0 Kudos
Hairul_Intel
Moderator
583 Views

Hi Michele,

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

 

 

Regards,

Hairul


0 Kudos
Reply