The NVIDIA driver on your system is too old

 

The NVIDIA driver on your system is too old

Error code:

import torch


def _check_driver():
    if not hasattr(torch._C, '_cuda_isDriverSufficient'):
        raise AssertionError("Torch not compiled with CUDA enabled")
    if not torch._C._cuda_isDriverSufficient():
        if torch._C._cuda_getDriverVersion() == 0:
            # found no NVIDIA driver on the system
            raise AssertionError("""
Found no NVIDIA driver on your system. Please check that you
have an NVIDIA GPU and installed a driver from
http://www.nvidia.com/Download/index.aspx""")
        else:
            # TODO: directly link to the alternative bin that needs install
            raise AssertionError("""
The NVIDIA driver on your system is too old (found version {}).
Please update your GPU driver by downloading and installing a new
version from the URL: http://www.nvidia.com/Download/index.aspx
Alternatively, go to: https://pytorch.org to install
a PyTorch version that has been compiled with your version
of the CUDA driver.""".format(str(torch._C._cuda_getDriverVersion())))


if __name__ == '__main__':

    _check_driver()

 

My solution:

The P100 driver is not easy to find. I installed a new cuda and installed only the driver. It was solved.

 

https://www.zhihu.com/question/309583980

 

The reason is that the pytorch version I installed requires cudav version 9.2, and the v9.0 version installed on my computer is not compatible

 

Therefore, the solution is to install cudav 9.2, the problem is solved

Reason analysis: The graphics card driver of the system that has just been reinstalled is provided by Windows 10 itself. It may be an older version and you need to install the latest driver.

Solution: Re-download the corresponding graphics card driver on the NVIDIA official website and install it (note: it is the driver , not CUDA)

Solution

The first method is to update the graphics card driver.
(1) nvidia-smi -L check your own graphics card model.
(2) Then download the corresponding driver in https://www.nvidia.cn/Download/index.aspx?lang=cn .
(3) Install the driver.

The second method is to install cuda 7.5.
(1) First of all , download the corresponding cuda 7.5 installation file in https://developer.nvidia.com/cuda-toolkit-archive according to the machine environment. I downloaded the sh file.
(2) Then install cuda 7.5.
(3) Finally, sudo pip install mxnet-cu75 == 1.2.1 to install the mxnet version of cuda 7.5.