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 supported31 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Kernel][DCP][MLA] Pre-reserve packed A2A decode workspace (long-term…
- [XPU][CI] Enable shared loader test
- [DSV4][XPU] Add XPU SYCL inv_rope_bf16 path for DeepSeekV4
- [Feature] Support Eagle3 speculative decoding with Pipeline Parallelism
- [DSV4][XPU] Change DeepSeek-V4 qnorm+rope+kv_insert to SYCL kernel path
- [Bugfix][MoE] Snapshot enforce_eager into FusedMoEConfig for weight reload
- [XPU][DeepSeekV4]Add DeepSeek-V4 fuse_index_q SYCL kernel path
- [Bug]: Sentence-Transformers models saved with sentence-transformers>=5.4.0 are silently mis-pooled — compact `1_Pooling/config.json` (`pooling_mode` string) is not parsed, so vLLM falls back to the architecture-default pooler `
- [Bugfix] Support MTP speculative decoding with pipeline parallelism on the PD producer side
- [KV Connector][Mooncake]Support DeepSeek V4 Mooncake PP/PD shared KV metadata
- Docs
- Python not yet supported