CUDA driver version is insufficient for CUDA runtime version

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Sorry if this isn’t the right place to post this, I was referred here from the NVIDIA Geforce forums.

GTX 1070. Windows 10. I’m running DeepFaceLabCUDA10.1AVX_build_05_10_2019 and get the following error:

tensorflow.python.framework.errors_impl.InternalError: cudaGetDevice() failed. Status: CUDA driver version is insufficient for CUDA runtime version

Running

nvcc --version

in the Windows terminal outputs

Cuda compilation tools, release 10.1, V10.1.168

It seems to me that the CUDA versions match, so I am confused. Thanks for the help.

EDIT: I reposted this as https://devtalk.nvidia.com/default/topic/1051999/cuda-setup-and-installation/cuda-driver-version-is-insufficient-for-cuda-runtime-version/ in the CUDA section instead. I’m not sure how to delete this topic.

2

what is your nvidia GPU driver version?

3

Oh, it is 397.64. I will update that right now to 430.64.

EDIT: Of course! How could I have not thought of that? I’m sorry for even wasting the time, I thought it was an issue with the CUDA versions.