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
- [Core] Support token-range cache salt regions in block hashing
- [Bugfix] Restore vllm.transformers_utils.tokenizer shim for lm-format-enforcer
- Accept x-api-key header as auth fallback
- [Quantizaiton][Feature] Pluggable KV Cache Data Types
- [Feature]: Detect and break exact repeating loops inside reasoning sections (spec-decode-aware ThinkingLoopBreaker)
- [Bugfix] Skip Dynamic NTK scaling when served length is below trained length
- [KV Offload] Pack disk backend slots into a single flat buffer for O_DIRECT alignment
- [Bug]: Qwen3.8-27B-FP8 hangs indefinitely at startup during CUDA-graph capture on Ampere (RTX A5000, TP=4) — fixed by --enforce-eager
- [Bugfix][Kernel] Fix no-op num_blocks division in get_moe_wna16_block_config
- [Bugfix][ROCm] Restore KV-offload Triton fast path via mapped device …
- Docs
- Python not yet supported