Failure to initialise Decoder object inside provided Docker file

[DanCorvesor]

Describe the bug I cannot initialise the PyNvDecoder inside the docker container built from the instructions in the README file.

video_file = "lapsleeve.mp4"
gpuID = 0
nvDec = nvc.PyNvDecoder(video_file, gpuID)

RuntimeError: /opt/src/PyNvCodec/src/PyNvCodec.cpp:61
CUDA error: CUDA_ERROR_SYSTEM_DRIVER_MISMATCH
system has unsupported display driver / cuda driver combination

To Reproduce Steps to reproduce the behavior:

1. Build the docker container (as defined in the README)
2. Run the docker container (as described in the README)
3. Try to initialise decoder class on video object locally inside docker container

Expected behavior Class creates successfully

Desktop (please complete the following information):

• OS: Linux Ubuntu 22.04
• Nvidia driver version: 520.56.06
• CUDA Version: 11.2 in container, 11.8 on machine
• Python Version 3.8 (in container)

Additional context I am suspecting this might me to do with the difference in CUDA toolkit versions on the host vs inside the container however I can't find anything obvious documented online to imply this is a problem. If it is the root cause, it would be great if you could update the README to reflect this requirement.

[gedoensmax]

Hi your driver version is lower than the minimal require 525 stated in our README. If you rely on this driver version you will have to downgrade your used Video Codec SDK version by leveraging:

• Setting the respective CMake variable: https://github.com/NVIDIA/VideoProcessingFramework/blob/3ba39fb97434a06d85c0d4c73de049e166c9e162/src/TC/CMakeLists.txt#L102-L104
• or simply replacing files in: https://github.com/NVIDIA/VideoProcessingFramework/tree/3ba39fb97434a06d85c0d4c73de049e166c9e162/src/TC/third_party

[sankovalev]

I ran into the same error with docker, my driver version is 530.30.02