I am getting Failed to initialize Inference Engine backend (device = GPU) error

hlreicha · ‎03-09-2022

Hi,

This is the error I am receiving:

> Failed to initialize Inference Engine backend (device = GPU): Failed to create plugin /opt/intel/openvino_2021.4.752/deployment_tools/inference_engine/lib/intel64/libclDNNPlugin.so for device GPU
> Please, check your environment
> [CLDNN ERROR]. clGetPlatformIDs error -1001

Hardware:

I7-11700K,
MSI Tomahawk Z690 Motherboard,
32GB 3200Mhz Cl14 Ram,
1TB Nvme SSD,
and 750 watt PSU.
No dedicated GPU, using the iGPU.

Environment:

OS: Ubuntu 20.04 LTS
Running a Python 3.7 Environment in Conda and I am trying to inference an Efficientnet model I optimized through OpenVino with OpenCV on GPU.
I used FP16 during optimization.

Code:

import cv2

model = cv2.dnn.readNet(model_path)

model.setPreferableBackend(cv2.dnn.DNN_BACKEND_INFERENCE_ENGINE)

model.setPreferableTarget(cv2.dnn.DNN_TARGET_OPENCL_FP16 )

Attempted Fixes:

Reinstalled OpenVino
Reinstalled the OS/Environment

Further Notes:

I followed the directions and everything works fine on my I7 TigerLake NUC.
Additionally this runs fine when I try to inference on CPU as well.

Any insights on how I should fix this?

Thanks.

Peh_Intel · ‎03-09-2022

Hi hlreicha,

Thanks for reaching out to us.

Please try to install the GPU driver version 21.29.20389 and see whether it can work. You can rerun the script in /opt/intel/openvino_2021/install_dependencies/ directory with the command below:

sudo -E ./install_NEO_OCL_driver.sh --install_driver 21.29.20389

If you still getting the error: [CLDNN ERROR]. clGetPlatformIDs error -1001, try to check and troubleshoot by referring to Driver issues troubleshooting.

Regards,

Peh

Peh_Intel · ‎03-21-2022

Hi hlreicha,

This thread will no longer be monitored since we have provided solution and suggestion. If you need any additional information from Intel, please submit a new question.

Regards,

Peh