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
- [Perf][DSV4.1] Enable the fused all-reduce + mHC boundary at TP2
- [Bugfix][Spec Decode] DFlash2: map reduced-vocabulary candidates to target ids
- [Review only][NIXL] Simplify draft head slicing from #58471
- [Bugfix][Model] Apply Nomic dynamic RoPE per input sequence
- [Bug]: CPU arm64 image fails to start on Apple M4: "no support for 'sme' without 'sve2'" (PyTorch Inductor + GCC 15)
- [Model] Add StartLux Instant decisions and platform-safe D3 prefix ranges
- [Perf] Faster BF16x3 router GEMM at large batch sizes on SM100
- [Benchmarks] Add model-free EngineCore IPC microbenchmark
- [Bugfix][Tool Parser] Stream InternLM2 non-plugin action content instead of dropping it
- [Kernel][SM120] Add small-batch block-FP8 projection routes
- Docs
- Python not yet supported