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 supported62 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Security] Contain Sentence Transformers module files
- [Frontend] Tokenize the /v1/systemone state once per decision and let big decisions take turns
- [ROCm][Perf][Qwen3.8] Mono decode follow-up: MXFP6 projections, attention front kernel, router through o_proj
- [RFC]: Simulated execution with pluggable timing models (building on #47922)
- [Spec Decode][Qwen3.5] RecoverSSM for Qwen3.5 / Qwen3-Next GDN
- [Bug]: return_sampling_mask=True has two independent request-level preconditions (temperature > 0 and top_k > 0) that are reported one at a time, and only after the engine is already running
- [CPU] Add BF16 fallback for MXFP4 MoE without AMX
- [Transformers backend] add DistilBERT sequence classification via Transformers Backend
- [Frontend] Send ping keepalives while an Anthropic /v1/messages stream waits for its first token
- [FlashInfer] Add opt-in eager/graph autotune measurement
- Docs
- Python not yet supported