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
- [Bugfix][Spec Decode] Implement get_top_tokens() on the DeepSeek V4 MTP drafter
- [Perf][KV Offload] Calibrate the DMA/Triton load cutoff at handler init (opt-in)
- [Feature]: Parser cache for streaming chat derender and stable tool call IDs on retried calls
- [Feature]: Shared per choice `ChatStreamProcessor` for chat serving and streaming derender
- [Feature]: Integer token IDs for logprobs in `/inference/v1/generate` responses (`GenerateLogProbs`)
- [Bug]: GLM-5.3-Flash (glm5_next, qk_rope_head_dim=0 / NoPE MLA) crashes — concat_and_cache_mla requires pe_dim==64
- [Model][DCP] Decode context parallelism for Qwen4Exp QSA
- fix: omit system_fingerprint key when fingerprint-mode=none
- [Bug]: Weight-loading/repacking memory is not reclaimed on discrete GPUs, forcing lower max-num-seqs / max-num-batched-tokens (Kimi K3, B200, fastsafetensors + DeepEP v2)
- [Bugfix][Scheduler] Preserve logprobs across streaming continuations
- Docs
- Python not yet supported