Compile model and run it without openvino

[Alexandros-Stavrop]

Hello,

I have a model in .onnx form and I have managed to optimize it and compile it using openvino 2021.4, so now I have the file in .blob form.

Is there a way to run this .blob file without needing to install openvino on the deployment device or is the installation required?

Thanks a lot

[Aznie_Intel]

Hi Alexandros-Stavrop,

Thanks for reaching out. Yes, it is possible to run a .blob file without installing the full OpenVINO package. However, you will still need certain OpenVINO runtime libraries with your application. The easiest way is to package the necessary .so (Linux) or .dll (Windows) libraries with your application. Your application must also use the OpenVINO Inference Engine API to load and execute the .blob file.

Regards,

Aznie

[Alexandros-Stavrop]

Hi @Aznie_Intel,

Thank you for your answer. Could you please elaborate on the way I could find the libraries required?

Thanks again.

Alexandros

[Aznie_Intel]

Hi Alexandros-Stavrop,

To help you run your .blob file without installing OpenVINO, could you please provide the following details about your deployment setup?

1. What hardware are you using for deployment?
2. Are you running this on an Intel CPU, an Intel VPU (such as MyriadX or NCS2), an ARM-based device (like a Raspberry Pi or Jetson), or something else? Please specify the exact model if possible.
3. What operating system is installed on the deployment device?
4. Is there anything else we should know about the deployment environment?

Regards,

Aznie

[Aznie_Intel]

Hi Alexandros-Stavrop,

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

Regards,

Aznie