Hello, I have installed openvino on my computer and the verification scripts are executed without any errors. Then I ran the image classification sample application with the Caffe* Squeezenet1.1 model using the following commands:
classification_sample_async.exe -i "C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\demo\car.png" -m "C:\Users\Admin\Documents\Intel\OpenVINO\openvino_models\ir\public\squeezenet1.1\FP16\squeezenet1.1.xml" -d GPU
The following error came:
[ INFO ] InferenceEngine: API version ............ 2.1 Build .................. 2020.3.0-3467-15f2c61a-releases/2020/3 Description ....... API [ INFO ] Parsing input parameters [ INFO ] Parsing input parameters [ INFO ] Files were added: 1 [ INFO ] C:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\demo\car.png [ INFO ] Creating Inference Engine GPU clDNNPlugin version ......... 2.1 Build ........... 2020.3.0-3467-15f2c61a-releases/2020/3 [ INFO ] Loading network files [ INFO ] Preparing input blobs [ WARNING ] Image is resized from (787, 259) to (227, 227) [ INFO ] Batch size is 1 [ INFO ] Loading model to the device [ ERROR ] Error has occured for: pooling:pool1 Best_kernel.empty()(true) should be false Cannot find a proper kernel with this arguments
There aren't any errors when I ran the image classification application using CPU (-d CPU). My environment is like this:
CPU: Intel Xeon E3-1515M V5
GPU: Iris(TM) Pro Graphics P580 (Driver version: 184.108.40.20627)
OS: Windows enterprise 2016 LTSB (version 1607)
Openvino version: 2020.3.194
Is there anyone can help with my problem? Thanks in advance!
Ensure that you had carefull followed :https://docs.openvinotoolkit.org/latest/openvino_docs_install_guides_installing_openvino_windows.html#Install-GPU
This should ensure that your GPU is supported & all pre-requisite required for GPU operations are installed.
I'm running Windows 10, i7 with Intel(R) UHD Graphic 620. Check yours and cross check with the tutorial I had attached in the link above.
Thanks for the provided information. According to the hardware components listed under Intel® Processor Graphics section here, the OpenVINO™ toolkit seems to support Intel® Iris® Pro graphics:
- "Intel Xeon processor with Intel Iris Pro graphics and Intel HD Graphics (excluding the e5 family which does not include graphics)".
With this in mind, please try the following and share the results:
- Re-run classification sample using model in FP32 precision instead of FP16
- Simplest way, edit demo_squeezenet_download_convert_run.sh change line #47 to FP32 instead of FP16
- Run the other verification script with GPU (demo_security_barrier_camera.sh) and share results of both FP16 and FP32
- Edit demo_security_barrier_camera.sh change line #140 to FP32 instead of FP16
Intel will no longer monitor this thread since we have provided a solution. If you need any additional information from Intel, please submit a new question