OSError: libcudart.so.9.0: cannot open shared object file: No such file or directory

[hakS07]

when i was trying to prepare voc dataset for training pre trained model https://gluon-cv.mxnet.io/build/examples_datasets/detection_custom.html#derive-from-pascal-voc-format i got this error OSError: libcudart.so.9.0: cannot open shared object file: No such file or directory from gluoncv.data import VOCDetection class VOCLike(VOCDetection): CLASSES = ['iris'] def __init__(self, root, splits, transform=None, index_map=None, preload_label=True): super(VOCLike, self).__init__(root, splits, transform, index_map, preload_label) dataset = VOCLike(root='/home/dell/Bureau/myDataset', splits=((2019, 'train'),)) print(dataset[0][1]) Knowing that I can't use cuda or gpu because i haven't a graphic card nvidia python 3.6 os ubuntu 18 Pillow-5.4.1 gluoncv-0.4.0 scipy-1.2.1 tqdm-4.31.1

[Jerryzcn]

Are you trying to train on CPU?

[hakS07]

yes

[PistonY]

make sure you cuda version is 9.0. And add below to ~/.bashrc.

export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

run

source  ~/.bashrc

Add below to /etc/ld.so.conf.d/cuda.conf

/usr/local/cuda/lib64

run

sudo ldconfig

[hakS07]

@PistonY can i use cuda and i don't have nvidia graphic card????

[Jerryzcn]

what is the command you are using?

[hakS07]

@Jerryzcn i'm not using a cmd command i just try to run a python code https://gluon-cv.mxnet.io/build/examples_datasets/detection_custom.html#derive-from-pascal-voc-format

[Jerryzcn]

which version of mxnet did you install? You might have installed the version that's built with CUDA. Try install the cpu version, or build from source and set CUDA to 0

[hakS07]

@Jerryzcn mxnet '1.4.0' no i have already installed the version compatible with Inux and cpu https://mxnet.apache.org/versions/master/install/index.html?platform=Linux&language=Python&processor=CPU

[xinyu-intel]

@essalahsouad try "import mxnet as mx", does this work well?

[hakS07]

@xinyu-intel yes, it does work

[Jerryzcn]

did you install gluoncv using pip or build from source

[hakS07]

@Jerryzcn I have installed with pip

[Jerryzcn]

Could you try install from source?

[SaidaSaad]

Hello i have the same problem and i am using nvidia rtx 2080

[Jerryzcn]

our default pip install package assume GPU support. So you will probably need to install from source to use non-gpu version. If you encounter this issue while using GPU, it is likely that your CUDA path is not added correctly, or your mxnet version is incompatible with the CUDA version.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.