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
- [ROCm] MXFP8 block32 GEMM: add TP1 tuned configs for DeepSeek-V4.1-Flash DP8+EP8
- [Feature]: KV-pool-saturated admission retries inflate decode TPOT ~2.3x: request automatic admission backoff
- [DSv4.1][PCP] Support adaptive DSpark verification
- [Bugfix][CuMem] Retry the host backup copy in wake_up
- [Bugfix][Frontend] Reject apply_softmax=false in /generative_scoring under --logprobs-mode *_logits
- [Bugfix][Model] Preserve fp32 second_per_grid_ts and compute Qwen2.5-VL temporal M-RoPE IDs in fp32
- [Mooncake] Move KV through a transfer backend and support TENT
- [Bug]: vllm/vllm-openai:unlimited-ocr image cannot serve Unlimited-OCR model
- [Metrics][KV Cache] Add per-type KV cache usage for hybrid models
- [Perf][DSv4.1] Prefetch CPU-offloaded Engram rows for eager prefill
- Docs
- Python not yet supported