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
- fuse norm into mhc
- [Feature]: Low-overhead asynchronous early-exit checks & KV-cache reclamation for Continuous Batching
- [Bug]: `--cudagraph-metrics` produces no runtime statistics despite active CUDA Graph execution
- [Feature]: Support per-request dynamic LoRA scale
- [Feature]: Map unsupported reasoning_effort values to the nearest level the chat template accepts
- Fix TurboQuant get_kv_cache_shape crash with hybrid models on cache_dtype_str='auto'
- Amd cpu ci minimal reuse
- [Bug]: V1 MTP max_model_len skip still schedules drafts with async scheduling
- [Model][torch.compile] Support torch.compile for SigLIP embeddings
- [Perf] Backfill divergent local hybrid hits from the offloading connector
- Docs
- Python not yet supported