ov::busy error on a project inspired by open_model_zoo

[caracole77]

Hello ! 

I'm currently  learning Openvino  in C++. I tried to start by creating a program inspired by open_model_zoo's  interactive_face_detection (https://github.com/openvinotoolkit/open_model_zoo/tree/master/demos/interactive_face_detection_demo/cpp). The interactive_face_dection program works perfectly on my machine. I tried to have the same pipeline for the ov frame processing (using the wrapper present in the common directory), and I'm only using AgeGenderDetection and FaceDetection in async mode (with start_async, wait). I'm using the exact same functions in detectors.cpp/hpp, and  I'm facing the error "ov::busy error" and a progressive fps drop until freeze.  It's hard to debug this crash, I don't know where to start investigating.  

[Zulkifli_Intel]

Hi caracole77,

Thank you for reaching out.

Based on class Busy : public ov::Exception, the error is thrown when calling the

InferRequest methods while the request is busy with compute operation.

To investigate this further, can you share the all related files with us?

Regards,

Zul

[caracole77]

Hello,

Thank you for your time!

Here the concerned files. I used the same designs as OMZ for the model, and the loading, took some liberties. But somehow something is wrong. It's a mess because of the debug and my level in programming, I'm really sorry.

Cordially

Caracole

[Zulkifli_Intel]

Hi caracole77,

Thank you for the attached files. Can you also share the cmakefile and the steps you used to build the program?

Regards,

Zul.

[caracole77]

Hi,

To build the program I installed all the dependencies (opencv, openvino) and then I create the build directory where I run the CMake.

Thank you for your help and time,

Caracole

[Zulkifli_Intel]

Hi caracole77,

I wasn't able to reproduce the issue on my end. May I know if is there any particular reason you want to create this custom code? For your info, it is recommended to use the demos and samples from OMZ to run with the pre-trained models as the demos and samples are tested and validated by Intel, ensuring reliable execution with various models. While the custom code might introduce bugs or inefficiencies.

Regards,

Zul

[Zulkifli_Intel]

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.