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][Perf][DSv4.1] Mono: tuned K2 FFN seam at TP4 / TP8
- [Feat][BI][3/N] add batch-invariant DeepSeek V4 support (mHC)
- [Bugfix][DSA] Disable native padded decode on ragged spec-decode batches (#64855)
- [Bugfix][ROCm] Fix AITER FA extend_forward zero-chunk and drafter KV workspace (#64680)
- [Bugfix][Compile] Support distinct attention geometries in RopeKVCacheFusionPass (#64679)
- [Perf][Attention] Cache TritonAttentionImpl K/V cache views per tensor (#64572)
- [Security] Gate NIXL remote-decode pulls on bidirectional transfer
- [Perf][DSv4.1] Overlap mHC coefficients behind the fused all-reduce on SM90
- [Perf][DSv4.1] DeepSeek-V4.1-Flash specialized decode on Hopper (H200 TP4)
- [Perf][LoRA] Compute kernel metadata on the CPU to remove per-step GPU syncs
- Docs
- Python not yet supported