search

NVIDIA-Merlin / Transformers4Rec

Transformers4Rec is a flexible and efficient library for sequential and session-based recommendation and works with PyTorch.
https://nvidia-merlin.github.io/Transformers4Rec/main
Apache License 2.0
1.08k stars 142 forks source link

Extend LD_LIBRARY_PATH with pytorch directory #717

Closed oliverholworthy closed 1 year ago

oliverholworthy commented 1 year ago

Follow-up to #715 Extends LD_LIBRARY_PATH with pytorch directory instead of overwriting it.

github-actions[bot] commented 1 year ago

Documentation preview

https://nvidia-merlin.github.io/Transformers4Rec/review/pr-717

oliverholworthy commented 1 year ago

This appears to get the test working. One thing that's still a bit unclear to me is that this test runs ok outside the tox environment with LD_LIBRARY_PATH replaced entirely with /opt/tritonserver/backends/pytorch. My guess is that there are other environment variables that help figure out where to find the correct cuda versions to use that make the rest of LD_LIBRARY_PATH not strictly required