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 supported51 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bug]: moe_wna16_gemm (csrc/libtorch_stable/moe/moe_wna16.cu) is not run-to-run deterministic: split-K partial sums are combined with 16-bit atomicAdd
- Port PR #55738 (GLM-5.3-Flash NoPE MLA prefill) to older vLLM trees + SM120 note
- [ROCm][Model][DCP] Default Kimi-K3's DCP combine to a2a
- [Doc] Add llms.txt so agents and AI tools can find the canonical docs
- [Bugfix][MLA] Keep MXFP8 prefill activations in model dtype
- [RFC]: Windowed Hidden-State Collection for vLLM Rollouts
- [Feature]: TP=2 graph capture + MTP speculative decoding crash on Arc B70 — fix already exists upstream, unmerged
- [Bugfix][Benchmark] Use concurrency for per-user throughput
- [Bugfix][Benchmark] Parse decode metrics as Prometheus samples
- [Quark][ROCm] Add FlyDSL W4A6 (MXFP6 act x MXFP4 weight) dense linear
- Docs
- Python not yet supported