FCOS No CUDA runtime is found, using CUDA_HOME=’/usr/local/cuda-10.0′

FCOS appears No CUDA runtime is found, using CUDA_HOME=’/usr/local/cuda-10.0′

  • appear below error </ li>
    • reason for the error </ li>
    • view version </ li>
    • solution (cuda10.0 and torch did not match the 1.2.0) </ li> </ ul>

    )

    appears with the following error

    AssertionError:
    The NVIDIA driver on your system is too old (found version 10000).
    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.

    error cause

    cuda version does not match the torch version
    . On my machine, the pytorch version is too new, while the cuda version is too old to match.
    is usually the torch version that does not fit

    view version

    NVCC -v

    # or pip3 list
    PIP list

    cuda10.0 with torch 1.3.1 mismatch </ p>

    solution (cuda10.0 and torch 1.2.0 match only)

    uninstall the original torch1.3.1
    pip3 uninstall torch
    # or
    PIP uninstall torch

    reshipment torch1.2.0
    pip3 install torch 1.2.0 torchvision </ mark> 0.4.0 -f https://download.pytorch.org/whl/torch_stable.html
    PIP install torch 1.2.0 torchvision </ mark> 0.4.0 – f https://download.pytorch.org/whl/torch_stable.html

    after checking whether torch1.2.0 has been installed successfully,
    pip3 list
    # or
    PIP list

    the final version of the cuda with torch version match go to website check can
    see pytorch website https://pytorch.org/get-started/previous-versions/ to see such a </ p>

Read More: