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
- fix(watermarking): align DualKeyGumbelWatermarkDetector alpha default with generation default
- [Core] Pick hybrid KV cache group size by worst-case padding bytes; add --min-kv-cache-group-layers
- [CI][ROCm] Match the AMD steps to the mirror declarations they carry
- feat: use safetensors prefetch api
- [Bugfix][Core] Do not double-count CUDAGraph memory in --kv-cache-memory suggestion
- [Model] Extend device-side mm normalization to DeepSeekV4 Vision / DeepSeekV41
- [Bug]: `/v1/responses` output items are always `status: "completed"`, even inside an incomplete response
- [Bugfix][OpenAI] Suppress hidden reasoning metadata in batch responses
- [Bugfix][Frontend] Fix Responses item status and Harmony event lifecycle
- FlashInfer autotune for trtllm_fp4_block_scale_moe wedges forever on SM103 (GB300): trtllm_gemm cubin is sm_100a-only, no PTX, and the autotuner has no timeout
- Docs
- Python not yet supported