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 supported51 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [ROCm][RDNA3] Fix W4A16 split-K accuracy and determinism
- [NIXL] Feed decoder pre-tokenized prompt ids and split multi-prompt requests in toy proxy
- [Bug]: Qwen4Exp (Qwen3.8-Flash-Next) hard-refuses PP>1 for any PLE checkpoint, and its MTP draft head branches on the global `is_first_rank`
- [Bug]: `resolve_kv_cache_layout` asserts every worker reports identical KV layouts — false for hybrid models under pipeline parallelism
- [Bug]: PP: a KV-cache group projected empty still emits `KVCacheTensor`s for other ranks' layers, causing a bare `StopIteration` in `allocate_kv_cache`
- [Bug]: `--load-format dummy` OOMs on large FP8 parameters — `initialize_single_dummy_weight` materializes a full-size fp16 temporary
- [Kernel][Perf] Triton unified attention: window-align the SWA/chunked KV-tile loop start
- [Bug]: FlashInfer TRTLLM MoE Backend stuck under DP/EP (Kimi K2 base, DeepSeek v3 base)
- [Security] Add hash collision defense for prefix cache with non-crypto hashes
- [Bugfix][Spec Decode] Fix DFlash draft KV cache dtype isolation
- Docs
- Python not yet supported