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 supported59 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [ROCm][gfx1250] Kimi-K3: run the W_UK/W_UV bmms and the latent up-pro…
- [Performance]: ~1.4 s gpt-oss serving stalls from per-batch-size recompiles of the bundled triton_kernels v3.5.1 routing kernels
- [Bug]: Qwen3.5 RMSNorm JIT-compiles FlashInfer kernels on sm75 (T4)
- [RFC]: Pipeline-parallel equivalence tests for every model that declares `SupportsPP`
- [CI][Test] Add pipeline-parallel equivalence tests for text models
- [Bug]: gemma3_pythonic chat template renders tool call arguments with no separator, so the pythonic parser rejects the whole call
- [Bug]: llama3.2_pythonic and toolace chat templates render string arguments unquoted, so the pythonic parser drops the call
- [Feature]: Deploying vLLM with Ray on SLURM-based university HPC clusters
- [Bug]: NCCL symmetric all-reduce uses the wrong group when TP and ETP differ
- [Bugfix][Distributed] Route NCCL symmetric all-reduce to the calling group
- Docs
- Python not yet supported