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 supported59 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bugfix] Exclude FlashInfer autotune peak from KV cache sizing
- [Bugfix] Use 64-bit row offsets in the block-FP8 Triton GEMM
- [Rust Frontend] Read client IPs from trusted proxies
- [ROCm][Test] Cover the MXFP4 W4A16 intermediate pad that reaches AITER
- [ROCm][gfx1250] Kimi-K3 FP8 MLA: aiter FP8 MHA + MLA Prefill/Decode
- [Bugfix] Build the dequant identity in the layer's activation dtype
- [ROCm][MoE] Forward the quant-config SwiGLU clamp to AITER fused MoE
- [ROCm][Perf][DSpark] Stack the DeepSeek-V4/V4.1 context WKV projections on ROCm
- [Bugfix] Refuse TP splits that cut compressed-tensors FP4 scale groups
- [ROCm][gfx1250] Run per-token FP8 linear layers on aiter's gemm_a8w8
- Docs
- Python not yet supported