vllm
https://github.com/vllm-project/vllm
Python
A high-throughput and memory-efficient inference and serving engine for LLMs
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported61 Subscribers
Add a CodeTriage badge to vllm
Help out
- Issues
- [Rust Frontend] Add request observers to vllm-llm
- [Rust Frontend] Register request observers through ServeExtensions
- [Bugfix][Tool Parser] Preserve quoted markers in Apertus arguments
- [ROCm][Perf] Default AITER fused shared experts on for GLM-5.x MXFP4
- [ROCm][Perf] Skip DSA indexer scoring for short dense-MHA prefills
- [XPU][CI] Remove Docker's privileged mode to prevent security risks in intel GPU CI
- [Perf][GDN] Capture the GDN core in breakable CUDA graphs of mixed batches (CUDA graph 2/2)
- [Bug]: pythonic and llama4_pythonic tool parsers break on a raw newline inside a string argument
- [Bug]: `vllm serve --uds <path>` cannot be restarted on the same path: the socket file is never removed, so the second start fails with `[Errno 98] Address already in use`
- [AMD][DSV4.1Flash] Quantize MoE activations to MXFP8 before the MoRI dispatch
- Docs
- Python not yet supported