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
- [Perf][Core] Restore compact prefix-cache entries after duplicate eviction
- [Bugfix] Keep FlashAttention sparse ahead of FlashInfer SM90 for NoPE-512 MLA
- [Bug]: DFlash2 speculative decoding acceptance permanently collapses to 0% right after prefix cache hit rate goes positive (Qwen3.5 hybrid GDN)
- [Bugfix][Core] Reject malformed logits processor FQCNs
- [Perf][MRV2] Reduce cross-lane reductions in min-p filtering
- [Bugfix][Rust Frontend] Guard every route by default in the API-key middleware
- [Bugfix] Use original context length for mRoPE YaRN correction range
- Fix reduced auxiliary states in NVIDIA DeepSeek V3.2 model
- [Feature]: Reduced draft vocabulary for MTP drafters that share the target lm_head (measured +25-29 % decode)
- [Feature]: [Rust Frontend][Multimodal] Add Gemma 4 image support / multimodal model spec
- Docs
- Python not yet supported