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
- [Bug]: EAGLE3 draft model with ModelOpt NVFP4 quantization fails to load — quant-excluded embed_tokens allocated as packed parameter
- [Bug]: Multimodal requests stall when the long-prefill cap cannot fit an intact item
- [Bugfix][Frontend] Reject NaN logit_bias values instead of banning the token
- [Bugfix][Spec Decode] Keep a full uniform-decode batch reachable by FULL graphs
- [Bugfix][Qwen3-Omni] Pass lm_head prefix for NVFP4 exclusion
- [Bugfix][LoRA] Map lm_head LoRA to batch tokens for prompt logprobs (…
- [MoE] Stage UVA-offloaded expert weights in VRAM for large steps
- [Bugfix][Spec Decode] Keep data_parallel_index in DFlash/DSpark draft configs
- [Bugfix][Qwen4Exp] Fix PLE layerwise reload with CPU offload and TP
- [Bugfix][MoE] Keep per-expert gate/up halves separate under stacked mappers
- Docs
- Python not yet supported