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
Add a CodeTriage badge to vllm
Help out
- Issues
- [BugFix] Pick the DP world-group port at bind time via the coordination store
- [Rust Frontend] Use fastokens for tiktoken models
- [Bug]: `SWIGLUOAI_UNINTERLEAVE requires clamp_limit` crash for compressed-tensors W4A4 MXFP4 MoE (blocks MiniMax-M3 MXFP4)
- [Bugfix] Forward SwiGLU clamp/alpha/beta in compressed-tensors W4A4 MXFP4 MoE
- [Benchmark] Add a timeout per socket without cancelling the full run
- [Bug]: DeepSeek-V4 sparse-decode MLA kernel wedges in a spin loop with no timeout and no diagnostic (device-level capture on sm_103; surfaces 300 s later as an unrelated DeepGEMM assert)
- [Frontend] Support repetition_detection as a server-side default
- [Feature][Model] Support stable-window-aware KV reuse for Qwen3-ASR realtime
- [Model] Kimi-K3: all requests degenerate to a repeated token after long-context prefill (NaN logits; packed KDA prefill suspected)
- DeepSeek-V4-Flash-0731: KV cache holds only 150K tokens in 7.7 GiB (56 bytes/token), max_model_len capped at ~121344 on H20 TP=2
- Docs
- Python not yet supported