shreyashampali/HOnnotate

WARNING: failed to load librasterise.so; rasterisation functions will be unavailable: libtensorflow_framework.so.2: cannot open shared object file: No such file or directory

anjugopinath opened this issue · 2 comments

Anaconda environment:
Python version : 3.6.1
tensorflow version : 1.12.0
tensorflow-gpu version : 1.12.0

I got this error when trying to execute the command : "python objectTrackingSingleFrame.py --seq 'test'"
I got the following error:
PROBLEM : "WARNING: failed to load librasterise.so; rasterisation functions will be unavailable:
libtensorflow_framework.so.2: cannot open shared object file: No such file or directory
"
1.I googled and saw that that one solution is to try adding the path to LD_LIBRARY_PATH
2. The python version in the anaconda virtual environment is 3.6.1
3. I executed "find . -name "libtensorflow_framework.so" -print" in the directory where anaconda is installed. I got the following output:
./anaconda3/pkgs/tensorflow-1.10.0-py36_0/lib/python3.6/site-packages/tensorflow/libtensorflow_framework.so
./anaconda3/pkgs/tensorflow-base-2.3.0-eigen_py38hb57a387_0/lib/python3.8/site-packages/tensorflow/libtensorflow_framework.so.2
./anaconda3/pkgs/tensorflow-base-1.12.0-mkl_py36h3c3e929_0/lib/python3.6/site-packages/tensorflow/libtensorflow_framework.so
./anaconda3/pkgs/tensorflow-base-1.12.0-gpu_py36had579c0_0/lib/python3.6/site-packages/tensorflow/libtensorflow_framework.so
./anaconda3/pkgs/tensorflow-base-2.2.0-gpu_py38h83e3d50_0/lib/python3.8/site-packages/tensorflow/libtensorflow_framework.so.2
./anaconda3/pkgs/tensorflow-base-1.14.0-py37h4531e10_0/lib/python3.7/site-packages/tensorflow/libtensorflow_framework.so.1
./anaconda3/envs/honnotate_tf2/lib/python3.8/site-packages/tensorflow/libtensorflow_framework.so.2
./anaconda3/envs/honnotate/lib/python3.6/site-packages/tensorflow/libtensorflow_framework.so
4. All the instances of "libtensorflow_framework.so.2" are under directories corresponging to python version 3.8 whereas the anaconda environment has python 3.6.

hi,
this seems to be a problem with dirt installation. Can you check if you are able to run dirt test cases?

Hi Shreyas,

Thank You very much for your reply. I tried again with a fresh installation.
But I got the following error:
**I set gcc version to 5.4.0 using export PATH and LD_LIBRARY_PATH commands.
But there are issues with cmake.

$ cmake --version
cmake: /usr/local/gcc-5.4.0/lib64/libstdc++.so.6: version GLIBCXX_3.4.26' not found (required by cmake) cmake: /usr/local/gcc-5.4.0/lib64/libstdc++.so.6: version GLIBCXX_3.4.22' not found (required by cmake)
cmake: /usr/local/gcc-5.4.0/lib64/libstdc++.so.6: version `GLIBCXX_3.4.26' not found (required by /lib64/libjsoncpp.so.22)**

I checked with the admin and he said that I am using a very old version of GCC, and the glibstdc++ in that version does not provide GLIBCXX_3.4.26
The highest version it provides is: GLIBCXX_3.4.21.

I am not sure how to proceed as I installed Tensorflow(1.12) and python(3.5) as suggested in the project website.
These versions are compatible with cuda 9.0 and gcc versions less than or equal to 6, which gives the error.

Thank You,
Anju