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 supported59 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Feature][Rust Frontend] Support parallel sampling in gRPC GenerateStream
- [ROCm][Spec Decode] Support draft_lm_head_quantization on RDNA (int4 drafter head, 18% faster MTP decode)
- [Perf][MRV2] Avoid extra passes for temperature-only sampling
- [Bugfix] Make GLM-5.3 sparse MLA batch invariant under VLLM_BATCH_INVARIANT
- [ROCm][CI] Bound DeepSeek V4 MoE output by per-token error
- [Bugfix][Frontend] Reject legacy functions/function_call in Chat Completions
- [Bug]: Ministral-3 / Devstral-Small-2 FP8 with `--config-format hf`: vision tower and projector are built as FP8 and image embeddings are all zeros
- [Recipe] Kimi-K3.1-Pro: add MI325X (gfx942) single-node TP8 with DCP8
- [RFC]: Compute-free prefix KV lookup-and-pin for router-driven peer KV reuse
- [Bugfix][Frontend] Reject Anthropic max_tokens=0 prewarm with an actionable error
- Docs
- Python not yet supported