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
- [Bug]: [RFE] Decode context parallelism: fail-fast with the valid DCP set, or support `--decode-context-parallel-size auto`
- [Bugfix] Show topology-compatible DCP values in configuration errors
- [Bugfix] Include MoE routes in adaptive verification cost profiling
- [Spec Decode] Avoid redundant EAGLE3 auxiliary concatenation with fc_norm
- [Bugfix][Spec Decode] Preserve windowed draft AOT guard after cache promotion
- [Bug]: kv_cache_dtype="turboquant_3bit_nc" crashes on Phi-2 (head_size=80) — supports_head_size() claims "any positive value", but the Hadamard rotation is built at the next power of two (128×128) and matmul'd with an 80-dim key
- [Bugfix][Attention] Reject TORCH_SDPA for decoder attention
- [Quantization][Attention] Enforce power-of-two head size validation in TurboQuant attention
- [Bugfix][GDN] Preserve Torch exp precision in fused prefill
- [Bugfix][Kernel] Honor batched and strided SwiGLU clamp inputs
- Docs
- Python not yet supported