Baike.dev
All toolsAI codingTrendingOpen sourceNewsSubmit
Log in
Back to tool/Back to issues
#3219·text-generation-inference

Docker Image that works on RTX 5090

Author: shawnrushefskyCreated May 9, 2025Updated Dec 9, 2025

Feature request

Please release a docker image built for cuda 12.8+ that will work on rtx 5090.

Motivation

I work at Salad. Our customers want to run LLMs on RTX 5090. We like to recommend TGI because it’s so simple to set up, but that is not currently true for the latest GPUs.

Your contribution

Last I checked, PyTorch nightly was still needed for these GPUs

Source: huggingface/text-generation-inference

View original on GitHubView discussion on GitHub