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
- [Feature][KV Connector] Add UMBPStoreConnector for embedded and standalone DRAM
- [Bug]: New watermarking feature (v0.30.0) allows per-request opt-out even when server enforces --watermark-config
- [KV Offload] Request-free CPU->GPU prefetch primitive
- [Frontend] Add watermark_config.enforce to prevent per-request watermark opt-out
- [Test] Cover cumulative parallel sampling outputs
- [Bugfix][ROCm] Gate MXFP8 linear LDS budget on gfx950, not the Triton version string
- [CI Failure]: (DGX) Spark GPQA Eval (GPT-OSS) test_gpqa_corectness
- [Bug][CPU][Spec Decode]: PR #56323 left 6 CPU Triton fallbacks patching module globals that are no longer read — spec decode broken on arm64
- [Bugfix][MiMo] Fix ViT attention-sink semantics and omni wrapper packed_modules_mapping
- [Bug]: An error preventing the compilation of certain code about CUDA kernels soccurred with FlashInfer version 0.6.18, which was automatically installed alongside the precompiled vLLM version 0.29.0
- Docs
- Python not yet supported