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 supported30 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [KV offload] Parallel-agnostic fs-tier cache for single full-attention group
- CUDA device-side assert with nvidia/Qwen3.6-35B-A3B-NVFP4 + MTP on nightly f91fb2f
- [Bugfix] Canonicalize FP8 weight layout to (K, N) at the source
- [Opt] Optimize rotary embedding cache length
- [Installation]: ROCm install silently falls back to CUDA abi3 wheel on Python < 3.12 (no ROCm abi3 / cp311 / cp313 wheels)
- [Bug]: thinking_token_budget forces reasoning_end tokens into the middle of tool-call arguments — ThinkingBudgetStateHolder doesn't treat <tool_call> as implicit reasoning end (Qwen3.5+)
- add cute helion kernel
- [Bug]: huge `/metrics` payload for Nemotron 3 Ultra NVFP4 breaks Prometheus
- [Test][Quantization] Restrict test_machete_mm gate to Hopper family
- [Spec][PP] Support MTP speculative decoding under pipeline parallelism (PP>1)
- Docs
- Python not yet supported