CODEWITHSUNDEEP
cudapytorchtorchvision
ah bon
ah bon

Reputation: 10051

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

I'm using Pytorch under Ubuntu 18.04 and trying to import torchvision, but I get an error libcudart.so.9.0: cannot open shared object file: No such file or directory.

Someone could help to fix it? Thanks.

The infos below are detailed error logs:

Traceback (most recent call last):
  File "/home/x/.local/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2882, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-2-6dd351122000>", line 1, in <module>
    import torchvision
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/__init__.py", line 1, in <module>
    from torchvision import models
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/models/__init__.py", line 11, in <module>
    from . import detection
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/models/detection/__init__.py", line 1, in <module>
    from .faster_rcnn import *
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/models/detection/faster_rcnn.py", line 7, in <module>
    from torchvision.ops import misc as misc_nn_ops
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/ops/__init__.py", line 1, in <module>
    from .boxes import nms, box_iou
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/ops/boxes.py", line 2, in <module>
    from torchvision import _C
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
ImportError: libcudart.so.9.0: cannot open shared object file: No such file or directory
import torch
import torch.nn as nn
import torchvision.transforms as transforms
Traceback (most recent call last):
  File "/home/x/.local/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2882, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-4-677acbcfae34>", line 1, in <module>
    import torchvision.transforms as transforms
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/__init__.py", line 1, in <module>
    from torchvision import models
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/models/__init__.py", line 11, in <module>
    from . import detection
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/models/detection/__init__.py", line 1, in <module>
    from .faster_rcnn import *
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/models/detection/faster_rcnn.py", line 7, in <module>
    from torchvision.ops import misc as misc_nn_ops
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/ops/__init__.py", line 1, in <module>
    from .boxes import nms, box_iou
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
  File "/home/x/.local/lib/python3.6/site-packages/torchvision/ops/boxes.py", line 2, in <module>
    from torchvision import _C
  File "/home/x/pycharm-2019.2/helpers/pydev/_pydev_bundle/pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
ImportError: libcudart.so.9.0: cannot open shared object file: No such file or directory

Upvotes: 3

Views: 16625

Answers (5)

Agile Bean
Agile Bean

Reputation: 7171

Here's a generic solution independent from the cuda version.

  1. The goal to set the correct path to the cuda library which is called libcudart.so.<version number>.

  2. We find this path by sudo find / -name 'libcudart.so.<version number>'

  3. We must add the path to the folder (not to the file) containing libcudart.so.<version number> in the .bashrc :

    export LD_LIBRARY_PATH=<path to libcudart.so file>:$LD_LIBRARY_PATH

Upvotes: -1

Gunjan Paul
Gunjan Paul

Reputation: 531

Try the following command

pip3 install torch===1.2.0 torchvision===0.4.0 -f https://download.pytorch.org/whl/torch_stable.html

I am using CUDA Version: 11.6 and this solved my error.

Source: https://github.com/pytorch/pytorch/issues/25173#issuecomment-524802888

Upvotes: -2

Yuan Wenjin
Yuan Wenjin

Reputation: 166

using anaconda, torch version is:

pytorch==1.1.0
torchvision==0.3.0
cudatoolkit=10.0

has the same problem, and fix it by changing torchvision==0.3.0 to torchvision==0.2.2

Upvotes: 2

tsveti_iko
tsveti_iko

Reputation: 8032

  1. Check your CUDA version with one of these:

     nvidia-smi
 nvcc -V

  2. Check your Pytorch CUDA version like this:

     python
 import torch
 torch.version.cuda

  3. The versions should match. Otherwise you need to either upgrade/downgrade your CUDA or do the same but for Pytorch. The error libcudart.so.9.0: cannot open shared object file: No such file or directory means that the versions don't match. You have Pytorch for CUDA 9.0, but your CUDA is another version.

  4. To solve this install Pytorch with the correct CUDA version like this:

     conda install pytorch==1.2.0 torchvision==0.4.0 cudatoolkit=10.0 -c pytorch

NOTE: Replace the Pytorch and the CUDA Toolkit version with the desired ones. Check the available versions here.

  1. If applicable, rebuild your Pytorch:

     rm -rf build
 python setup.py develop

Upvotes: 1

Wasi Ahmad
Wasi Ahmad

Reputation: 37741

If you are using anaconda, the following may fix your problem.

conda install -c anaconda cudatoolkit==9.0

You can also try the followings.

Make sure the CUDA version is 9.0. And add the following 2 lines to ~/.bashrc.

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

Then, run:

source  ~/.bashrc

Add the following lines to /etc/ld.so.conf.d/cuda.conf

/usr/local/cuda/lib64

And run:

sudo ldconfig

Upvotes: 10

Related Questions