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 supported60 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bug]: Gemma-4 MTP halves prefix-cache hits with `--prefix-cache-retention-interval`
- [Bug]: `raw_logprobs` are computed after the structured-output grammar mask
- [Perf][Kimi] Run one packed prefill convolution
- [Perf][Model] Muse vision encoders: no GPU<->CPU syncs, build Glimmer's vision prep once per call
- [Perf] Capture prefill-sized CUDA graphs by default for Qwen3.5
- [DeepSeek-V4] Warm up DeepGEMM fp8_einsum for the o_proj
- [Tests] Handle Triton async compile cleanup callback
- [Bugfix] Size the default GDN recurrent beta by value heads
- [Bugfix][NixlPush] Supervise the push writer thread (W2)
- [Perf] Launch the Qwen GDN decode and Triton fused-MoE kernels with PDL
- Docs
- Python not yet supported