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
- [Rust Frontend] /derender: reasoning and tool-call parsing (phase 2/3)
- [Rust Frontend] /derender: streaming derender + two-process e2e test (phase 3/3)
- [Performance]: DSD K=0 draft-state sync forward costs 4-11% in a K≡0 workload; resume impact differs sharply between the two tested drafter architectures (MTP vs EAGLE3)
- [RFC]: First-Class, Orchestrator-Agnostic KV Hint Envelope for Agentic Workloads
- [Bug]: INSTALL_KV_CONNECTORS=true in vLLM 0.27.1 Docker image installs LMCache wheel incompatible with shipped PyTorch, causing lmcache.c_ops to fall back to torch baseline
- [Bugfix][DSv4] SM12x FlashInfer sparse MLA kernel block size 64
- [Core][Spec Decode] Opt-in skip of the K=0 draft sync forward (MTP + DFlash, default off)
- [CI] Allow empty final audio streaming chunk
- [Bug]: Long term continuous operation leads to an increase in model output illusion
- [Rust Frontend][RFC] Add Qwen3-ASR WebSocket `/v1/realtime`
- Docs
- Python not yet supported