Skip to content

Conversation

@nicholas-happyverse
Copy link
Contributor

The current container image for Triton uses Python 3.12 (see below):
image
Also, using Python 3.12 will sometimes have the "GLIBCXX_3.4.32 not found" error for Custom Python Backends created inside the official Nvidia container image (as mentioned in triton-inference-server/server#7913).

In my personal experience, using 3.12 instead of 3.10 for the conda environment and adding the conda install -c conda-forge libstdcxx-ng=13 -y enabled me to create a custom environment without issues.

Was hoping to update the documentation in case anyone else has issues with it!

@whoisj
Copy link
Contributor

whoisj commented Nov 10, 2025

@nicholas-happyverse
Copy link
Contributor Author

@whoisj Just sent over one, apologies for the large delay. Let me know if there's anything else I can do!

@whoisj
Copy link
Contributor

whoisj commented Dec 8, 2025

@whoisj Just sent over one, apologies for the large delay. Let me know if there's anything else I can do!

Thank you, working on confirming with our legal group.

@whoisj whoisj merged commit 4b15926 into triton-inference-server:main Dec 8, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

2 participants