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
- [Bug]: Qwen3.8-Flash-Next generates corrupted Thai orthography under vLLM; the same model under llama.cpp at the same QSA budget is clean
- [Bug]: [XPU] gdn_attention asserts on ragged ngram draft lengths (single stream, survives #53077)
- [KV Offload] Scope offload group configs to prefix-cacheable KV cache groups
- [RFC]: The dynamic-SD K decision is a batch-keyed array index — agreeing on its shape before more signals arrive
- [Bugfix][NIXL] Harden handshake listener against bad requests
- [Bugfix] Derive FA version from the worker's own device, not device 0
- [Bugfix][ROCm][Disagg] Bootstrap DP wave 0 from coordinator at engine READY
- [Perf] Qwen3-Next MoE: gate and shared_expert_gate are two separate ReplicatedLinear over the same input
- [Bug]: PLE short-conv batched prefill pads all requests to batch-max query length, causing transient activation OOM on mixed-length workloads(Qwen3.8-Flash-Next)
- [XPU] Route small TP all-reduces to a Level Zero IPC kernel
- Docs
- Python not yet supported