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 supported59 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bug] Empty/missing string content on a developer message returns 500 (and leaks internal _image_indices) instead of a 400
- [Feature request] DeepGEMM MegaMoE: support SM120 (Blackwell workstation) GPUs
- [Proposal][Perf] Wire flashinfer's PcieIpcAllGatherWorkspace into CudaCommunicator.all_gather for latency-bound TP gathers
- [Bug]: /inference/v1/generate usage omits reasoning_tokens in derender mode
- [Test] Let simulate_tool_streaming stream explicit per-chunk token ids
- [Bug]: [CPU] DeepSeek-V4 fails at the first decode step with KeyError: 'c1a'
- [Bugfix][CPU] DeepSeek-V4: return all layer types from the CPU tile scheduler
- [Bugfix] Return 400 for empty DeepSeek V4.1 developer messages
- [Bugfix][Tool Parser] xlam, hunyuan_a13b: keep tool calls whose arguments are not an object
- [Bugfix][Parser] chat_parsing: keep large integers exact for number parameters
- Docs
- Python not yet supported