Tensorflow-gpu does not recognize my NVIDIA GPU when using Poetry

• I am on the latest Poetry version.

• I have searched the issues of this repo and believe that this is not a duplicate.

• Ubuntu 20.04:

• Poetry 1.0.0:

Issue

Hello! Firstly Nice work of yours!

Well, I’m getting this issue and, for some reason, I couldn’t find anything related in my previous research.

I was building a package for my classification task using tensorflow-gpu and it was working just fine. I’m using tensorflow-gpu on version 2.2.0 and my CUDA version is 10.1, all fine to work with this kind of task and requirements.

Well, for some personal reasons I decide to use poetry to deal with versioning, but just after all the dependencies were installed with poetry tensorflow-gpu stop to recognize my GPU. Even though its all installed and working fine.

Commnad:

tf.config.experimental.list_physical_devices('GPU')

Result:

2020-08-07 22:49:36.619079: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory
2020-08-07 22:49:36.619139: E tensorflow/stream_executor/cuda/cuda_driver.cc:313] failed call to cuInit: UNKNOWN ERROR (303)
2020-08-07 22:49:36.619165: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:169] retrieving CUDA diagnostic information for host: lead-XPS-8930
2020-08-07 22:49:36.619500: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:176] hostname: lead-XPS-8930
2020-08-07 22:49:36.619548: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:200] libcuda reported version is: Not found: was unable to find libcuda.so DSO loaded into this program
2020-08-07 22:49:36.619716: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:204] kernel reported version is: 435.21.0

Anything outside poetry environment runs the command above just fine!

Have no IDEA whats happening. If you guys could give me a hint on how to solve this I’d be greatful! 👍

[joooeey]

See also: my similar issue on the tensorflow issue tracker: https://github.com/tensorflow/tensorflow/issues/48939

[igormcsouza]

Thanks for the advice!