Open sourcevLLM project
vLLM 0.33
A new release of the open-source engine for serving AI models. It adds support for many new models, upgrades to PyTorch 2.14 and ships images only for CUDA 13.0.
What’s new: For teams running their own models. CUDA 12.9 builds are dropped, so check your setup before upgrading; a new image targets AWS EFA networking.
927 commits from 389 contributors