Failure due to generic standard exception. Error: No kernel with name: DLA_CHECKSUM_ found on board

[Wang__Ting]

Hello,

I was trying to run the yolov3 model with FPGA(Openvino Starter Kit) with the api (object_detection_yolov3_async.py) provided by openvino toolkit 2020_3, and the error comes as the image (Question.png) showed. Can anyone tell me how to solve the problem?

Regards,

Ting

[Iffa_Intel]

Greetings,

Please help to carefully setup according to your hardware:

1. Configuration Guide for Intel® Distribution of OpenVINO™ toolkit 2020.3 and the Intel® Programmable Acceleration Card with Intel® Arria® 10 FPGA GX on CentOS or Ubuntu*: https://docs.openvinotoolkit.org/2020.3/_docs_install_guides_PAC_Configure.html
2. Configuration Guide for the Intel® Distribution of OpenVINO™ toolkit 2020.3 and the Intel® Vision Accelerator Design with an Intel® Arria® 10 FPGA SG2 (IEI's Mustang-F100-A10) on Linux*: https://docs.openvinotoolkit.org/2020.3/_docs_install_guides_VisionAcceleratorFPGA_Configure.html

And probably you can try to get the FPGA Plugin : https://docs.openvinotoolkit.org/2020.3/_docs_IE_DG_supported_plugins_FPGA.html

Sincerely,

Iffa

[Wang__Ting]

Hello,

I was using FPGA GT instead of FPGA GX version, and it seems that the configuration guide provides the package with FPGA GX version. What should I do to correctly set up the environment for my FPGA?

Regards,

Ting

[Max_L_Intel]

Hi @Wang__Ting 

Do you see the same error if you try ./demo_squeezenet_download_convert_run.sh verification script with -d FPGA parameter?

However, we have to inform you that OpenVINO toolkit officially supports PAC based on Arria 10 GX only, so we cannot guarantee this should work correctly with Cyclone V GT FPGA.

[Wang__Ting]

Hello,

I've tried the demo you mentioned, and it didn't work with the same error either... Do you know how to solve the problem?

Regards,

Ting

[Max_L_Intel · Accepted solution]

Hi @Wang__Ting 

As mentioned above, current version of OpenVINO toolkit does not officially support FPGA GT.
We apologize for the inconvenience.

[Max_L_Intel]

If you need any additional information from Intel, please submit a new question as this thread is no longer being monitored.

Thank you for your question.