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
- [Kernel] Let FlashInfer select the Kimi-K3 fused KDA decode kernel (backend="auto")
- [Bug]: FlashInfer CUTLASS MoE OOM during serving after successful memory profiling (NVFP4, DP+EP)
- [Perf][Model][Multimodal] Batch Gemma 4 vision pooler and fuse vision/audio RMSNorm and 2D RoPE
- [Compile] Defunctionalize rms_norm_per_block_quant
- [Benchmark] Add SLO attainment and output-token goodput reporting
- [KV Connector][NIXL] Reject asymmetric DCP transfers with interleave != block_size
- [KV Connector][NIXL] Validate DCP interleave size in handshake
- [Bugfix][Scheduler] Do not schedule ahead of an uncapped lone prefill chunk
- [Bugfix] Reserve sparse MLA indexer metadata before sizing the KV cache
- [Performance]: RMSNorm and int8 quantization are never fused for W8A8 int8 models, and rms_norm_dynamic_per_token_quant is up to 2x slower than the unfused pair at prefill sizes (sm_89)
- Docs
- Python not yet supported