DIGITS: ImportError: libcudart.so.7.0: cannot open shared object file: No such file or directory

I have followed the steps exactly as in your guide using an Amazon AWS GPU instance. When I get to the running of the .\digit-server I get:

Cannot guess value for "caffe_root": caffe binary cannot be found Cannot guess value for "gpu_list": Cannot query GPUs without a valid caffe_root

I also tried to download the .tar but when I run the ./runme file and I get

ImportError: libcudart.so.7.0: cannot open shared object file: No such file or directory

Do you have any suggestions about what I am doing wrong?

About this issue

Original URL
State: closed
Created 9 years ago
Reactions: 2
Comments: 21 (5 by maintainers)

Links to this issue

python - Why can't I import tensorflow - Stack Overflow

Most upvoted comments

For this error ImportError: libcudart.so.7.0: cannot open shared object file: No such file or directory

I have executed this command to solve it sudo ldconfig /usr/local/cuda/lib64

+165

Saneesh on Jul 10, 2015

Two suggestions:

You should use digits-devserver instead of digits-server. The devserver runs a user-level application, so you don’t need root permissions for anything. This will also give you an interactive prompt for setting the configuration options.
Both of your issues seem to be regarding environment variables. You probably need to add these two lines to your .bashrc (but customized to match the installation on your machine, of course):

export CAFFE_ROOT=/home/username/caffe
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

+36

lukeyeager on Jul 10, 2015

I have import error: ImportError: libcudnn.so.7: cannot open shared object file: No such file or directory with CUDA 9.0 Tensorflow 1.5

download https://developer.nvidia.com/rdp/cudnn-download 7.0.5 for cuda 9.0 sudo rm -rf /usr/local/cuda/include/cudnn.h sudo rm -rf /usr/local/cuda/lib64/libcudnn*

cd cd cuda

sudo cp include/cudnn.h /usr/local/cuda/include/ sudo cp lib64/lib* /usr/local/cuda/lib64/ cd /usr/local/cuda/lib64/ sudo chmod +r libcudnn.so.7.0.5 sudo ln -sf libcudnn.so.7.0.5 libcudnn.so.7 sudo ln -sf libcudnn.so.7 libcudnn.so sudo ldconfig It works for me.

+16

famishedrover on Feb 9, 2018

In case someone else runs into this… I got an error (something about libcudnn.so is not a symbolic link or something) when trying ldconfig. My directory looked like this:

screenshot from 2016-10-07 07-38-55

See how the libcudnn.so files are duplicated instead of linked like the others? I deleted the first two, and then made it a chain of symbolic links like so:

sudo rm libcudnn.so
sudo rm libcudnn.so.5
sudo ln -s libcudnn.so.5.0.5 libcudnn.so.5
sudo ln -s libcudnn.so.5 libcudnn.so

I actually came here in search of fixing a TensorFlow problem - this ended up fixing it.

chrisranderson on Oct 7, 2016

@chrisranderson if you installed cuDNN with a deb package instead of with a tarball then you wouldn’t run into this issue.

lukeyeager on Oct 10, 2016

@Saneesh Thanks a lot. When I run sudo ldconfig /usr/local/cuda/lib (not for lib64) it became work.

NickShargan on Sep 8, 2016