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 supported46 Subscribers
Add a CodeTriage badge to vllm
Help out
- Issues
- [v1] Allow attention backends to declare custom KV cache page size
- [Kernel] Optimize moeTopK memory access with shared memory and fix indexing bugs
- [Config] Introduce RuntimeDefault fields for runtime initialized config values with correct static types
- [Core] Immediately evict thinking token blocks from prefix cache
- [Kernel] Add fused routing and scatter-reduce decode optimizations for BS<=64 for GPT-OSS
- Add HAProxy to a list of load balancers in documentation
- [Feature] Add Int8/FP8-quantized all-reduce
- [Bug]: Abnormal Scores in Batch Processing of Image-Text Pairs with qwen3-VL-reranker Model
- [RFC]: support n sampling requests to eliminate redundant prefill computation and memory
- [Performance]: Qwen3.5 with mtp is slower than without
- Docs
- Python not yet supported