Open DavidNemeskey opened 6 years ago
Specifically check LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib
is present in your .bashrc if you use bash or .zshrc of you use zsh
Should /usr/local/lib be checked for libraries by default? Anyway, I exported it and still no cookie. :(
Actually, in the virtual env:
$ ldd ./lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so
./lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so: /lib/x86_64-linux-gnu/libm.so.6: version `GLIBC_2.23' not found (required by ./lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so)
./lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so: /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.21' not found (required by ./lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so)
./lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so: /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.21' not found (required by /mnt/permanent/Priv/ndavid/venvs/test/./lib/python3.6/site-packages/tensorflow/python/../libtensorflow_framework.so)
linux-vdso.so.1 (0x00007ffcab377000)
/usr/local/lib/libsystat.so (0x00007f087edf1000)
/usr/local/lib/libproch.so (0x00007f087ebef000)
libtensorflow_framework.so => /mnt/permanent/Priv/ndavid/venvs/test/./lib/python3.6/site-packages/tensorflow/python/../libtensorflow_framework.so (0x00007f087da0d000)
libcublas.so.8.0 => /usr/local/cuda/lib64/libcublas.so.8.0 (0x00007f087b05d000)
libcusolver.so.8.0 => /usr/local/cuda/lib64/libcusolver.so.8.0 (0x00007f0877aed000)
libmklml_intel.so => /usr/local/lib/libmklml_intel.so (0x00007f086ecaa000)
libiomp5.so => /usr/local/lib/libiomp5.so (0x00007f086e906000)
libcudart.so.8.0 => /usr/local/cuda/lib64/libcudart.so.8.0 (0x00007f086e6a0000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f086e39f000)
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f086e19b000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f086df7e000)
libgomp.so.1 => /usr/lib/x86_64-linux-gnu/libgomp.so.1 (0x00007f086dd68000)
librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f086db60000)
libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f086d855000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f086d63f000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f086d294000)
/lib64/ld-linux-x86-64.so.2 (0x00007f0889bd2000)
libcuda.so.1 => /usr/lib/x86_64-linux-gnu/libcuda.so.1 (0x00007f086c8b0000)
libcudnn.so.6 => /usr/local/cuda/lib64/libcudnn.so.6 (0x00007f086334e000)
libcufft.so.8.0 => /usr/local/cuda/lib64/libcufft.so.8.0 (0x00007f085a500000)
libcurand.so.8.0 => /usr/local/cuda/lib64/libcurand.so.8.0 (0x00007f0856597000)
libnvidia-fatbinaryloader.so.367.48 => /usr/lib/x86_64-linux-gnu/libnvidia-fatbinaryloader.so.367.48 (0x00007f0856349000)
So it seems that actually, the library was found and linked against. But why the error message then?
What is your OS, and did you follow the instructions in the README when installing mkl?
# If you don't have cmake
sudo apt install cmake
git clone https://github.com/01org/mkl-dnn.git
cd mkl-dnn/scripts && ./prepare_mkl.sh && cd ..
mkdir -p build && cd build && cmake .. && make
sudo make install
echo 'export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib' >> ~/.bashrc
@danqing I followed all instructions. As for LD_LIBRARY_PATH
, see my previous reply.
Hey I'm having this problem too, on plain Ubuntu 18.04.
I faced the same problem.
Executing sudo ldconfig
after all instructions fix the problem with package https://github.com/mind/wheels/releases/download/tf1.4.1-gpu/tensorflow-1.4.1-cp35-cp35m-linux_x86_64.whl
I installed the wheel for Python 3.6 from here. However, when trying to import tensorflow, I see the error below. Even if I install mkl-dnn by hand from the repo, I get the same error.