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.
6404 Discussions

Difference between 'Debug' and 'Release' versions of inference model

Sirisha_K_Intel
Employee
815 Views

Hi, can someone explain the difference between 'debug' and 'release' inference examples available in openVINO? Is there more logging with the debug version? From some examples that I ran with both, I see no difference.
This leads to my second question - If there is an error encountered while running an inference, is there a way to enable debug to get more logs to understand the error?

0 Kudos
2 Replies
Shubha_R_Intel
Employee
815 Views

Dear Kolli, Sirisha,

The difference between Debug and Release in OpenVino would be similar to Debug and Release on any other application for WIndows or Linux. One of the main differences is Debug Symbols (in Debug) so that when you step through code you will see actual symbol names for functions.  Also, optimization is often turned off in compilation for Debug so that the source code exactly matches line by line to the binary being debugged. Try stepping through the debugger in a Release build, it won't work well !

To answer your question, yes there is a way to get deeper into errors.  Build a Debug Version of dldt github Inference Engine. WIthin dldt you will find not only full Inference Engine source code but also full source code for the majority of our hardware plugins under thirdparty .

If you build the IE as Debug and the plugin also as Debug you should be able to step in very deep using a debugger to figure out what is going on. For Windows we recommend Visual Studio 2017 and for Linux, Eclipse.

Hope it helps,

Thanks,

Shubha

 

0 Kudos
Sirisha_K_Intel
Employee
815 Views

Thank you Shubha for your response. I will give it a try.

Sirisha

0 Kudos
Reply