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
- [Bugfix][Rust Frontend][Parser] Preserve DeepSeek DSML JSON types without parameter schemas
- [RFC]: Truncate mode for `thinking_token_budget`
- [Bug]: thinking_token_budget ignored by Model Runner V2 with Qwen3.8 NVFP4 + MTP
- [Bug]: GLM-5.3 accuracy collapses on ROCm (GSM8K 91.6% -> 14.9%) after #53155 forces MRV1
- [Feature]: Tokenized Prompt as Input for Chat and Messages Endpoints when serving
- [Bugfix][CPU][Spec Decode] Write expand_kernel output at caller dtype in CPU shim
- Update CuteDSL fused MoE API for 16-row interleave and low latency kernel support
- [Bugfix] Account for PCP when validating DCP for GQA/MQA models
- [Bugfix][Frontend] Preserve model text when tool calling is off
- [Kernel][Quantization] Cut transient memory in NVFP4 Marlin scale-factor computation
- Docs
- Python not yet supported