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 supported62 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Docs] Rewrite the preload (`ipc_cache`) guide and add Docker/Kubernetes usage
- [Distributed] Make --disable-custom-all-reduce fall back to NCCL
- [ROCm][Perf] gfx950 device length aware top-k split policy for k=2048
- [Bugfix][Quantization] Register compressed-tensors block-FP8 weight scale as weight_scale_inv
- [Helm] Add optional ServiceMonitor with metric filtering
- [Bug]: Triton top-p keeps the whole row when most of the vocab has zero fp32 probability
- [Bugfix][Metrics] Initialize offload usage and dev operation gauges at startup
- [Bugfix][Core] Count only held blocks against the sliding window admission cap
- [Benchmark] Fix benchmark_marlin so it runs, and add a ragged-M sweep
- [Bugfix][Metrics] Correct KV cache eviction and idle-time reporting
- Docs
- Python not yet supported