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
- [Feature] IndexCache for DeepSeek-V4 (validated on V4-Flash-0731, including DSpark on to accelerate on top of
- [RFC]: Model Runner V2 Pluggable Design
- [Feature]: DeepSeek V4: option to preserve <think>/</think> markers in chat completion content
- [CI/Build]: source_file_dependencies still point at pre-libtorch_stable csrc paths, so kernel jobs no longer trigger
- [Bugfix][Core] Make EncoderDecoderCacheManager slot release idempotent
- [Bugfix] MiniMax-M3: structured output silently ignored with thinking enabled (scheduler reasoning-end gate never fires)
- [Rocm] Fix AITER_MLA issues for the kimik3-dspark model with Agentic workload
- Add /dev/shm size calculator for tensor parallel deployment
- [Bug][Anthropic] /v1/messages canonicalizes model aliases, causing Claude Code to strip thinking blocks during tool-use loops
- [Bug]: --enable-prefix-caching is a silent 0%-hit no-op on Mamba-2/GDN hybrids (Qwen3_5MoeForConditionalGeneration)
- Docs
- Python not yet supported