Error when using GPU on YoloV3 sample in R5

[Gao__Jack]

I have met some error when using GPU on YoloV3 sample in R5 like following.

[ INFO ] Parsing input parameters
[ INFO ] Reading input
[ INFO:0] VIDEOIO: Enabled backends(5, sorted by priority): FFMPEG(1000); MSMF(990); DSHOW(980); CV_IMAGES(970); CV_MJPEG(960)
[ INFO ] Loading plugin
[ ERROR ] Cannot find plugin to use :Tried load plugin : clDNNPlugind,  error: Plugin clDNNPlugind cannot be loaded: cannot load plugin: clDNNPlugind from : Cannot load library 'clDNNPlugind.dll': 126 from cwd: C:\Users\JackGao\Documents\Intel\OpenVINO\inference_engine_samples_2015\intel64\Debug
c:\intel\computer_vision_sdk_2018.5.445\deployment_tools\inference_engine\include\details\os/win_shared_object_loader.h:43, skipping

The command I key in is

"object_detection_demo_yolov3_async.exe -i C:\Users\JackGao\Desktop\Work\test_60fps.mkv -m C:/Intel/computer_vision_sdk_2018.5.445/deployment_tools/model_optimizer/yolo_v3.xml -d GPU"

I can use this model on CPU successfully. Can anyone know how to solve this question?

[nikos1]

Hi Jack,

> Cannot load library 'clDNNPlugind.dll': 126

It is an environment setup issue. Please try to find clDNNPlugind.dll and maybe copy to the same folder as your executable. I also got some issues from Visual Studio for debug or release. After that you should be able to load the CLDNN dll.

Let us know if you get any more issues.

Cheers,

Nikos

[Gao__Jack]

I have already put the dll and the exe in the same folder but it still got the same problem.

[nikos1]

Hi Jack,

It seems you are building debug so the sample needs to find the dlls that end with a "d" - like clDNNPlugind.dll ( not clDNNPlugin.dll for release ). 

Could you try to copy from C:\Intel\computer_vision_sdk\inference_engine\bin\intel64\Debug to C:\Users\***YourUsername***\Documents\Intel\OpenVINO\inference_engine_samples_2015\intel64\Debug  or any other folder you build in?

Also could you try release build? Same issue?

What is the model of the CPU btw just to make sure the iGPU is good for -d GPU ?

Cheers,

nikos

[Gao__Jack]

Hi nikos,

I have already copied C:\Intel\computer_vision_sdk\inference_engine\bin\intel64\Debug to C:\Users\***YourUsername***\Documents\Intel\OpenVINO\inference_engine_samples_2015\intel64\Debug and tried release build. All of them are not working but I used this model in R2 SDK is successful.

[Wei__Zhu]

[ ERROR ] Cannot find plugin to use :Tried load plugin : MKLDNNPlugind,  error: Plugin MKLDNNPlugind cannot be loaded: cannot load plugin: MKLDNNPlugind from D:\MyDownloads\Intel\computer_vision_sdk_2018.5.445\deployment_tools\inference_engine\samples\intel64\Debug: Cannot load library 'D:\MyDownloads\Intel\computer_vision_sdk_2018.5.445\deployment_tools\inference_engine\samples\intel64\Debug\MKLDNNPlugind.dll': 126 from cwd: D:\MyDownloads\Intel\computer_vision_sdk_2018.5.445\deployment_tools\inference_engine\samples\classification_sample
d:\mydownloads\intel\computer_vision_sdk_2018.5.445\deployment_tools\inference_engine\include\details\os/win_shared_object_loader.h:43, skipping