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 supported62 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bug][ROCm]: v0.30.0 release is missing #57252, so GLM-5.3 (GlmMoeDsa) can't start on AMD
- [ROCm][Perf][Tracking Issue]: Hy4-Preview
- [CPU][MoE] Add AVX2 fallback for symmetric GPTQ INT4 experts
- [Core] Add --release-weight-page-cache to drop checkpoint pages after load
- [Performance][ROCm] V2 runner PP side streams make gfx1201 (RDNA4) forward ~2× slower; main-stream option fixes it
- [Bug]: When using the V4 parser in vLLM to parse outputs from DeepSeek V4.1 Flash, the leading extra space in V4.1 DSML tags breaks state machine matching and parameter regex, filters arg deltas, and prevents streaming return of tool call arguments.
- [Frontend][Rust][Multimodal] Add Gemma 4 image support / multimodal model spec
- [Hybrid GDN] Explicit --block-size misaligned with mamba page makes SimpleCPU KV offload silently write-only; startup should reject or warn
- [Bug]: Rust frontend: top_logprobs repeats the sampled token; prompt + max_tokens over max_model_len silently accepted
- [Bugfix] Join Anthropic system blocks with a separator instead of ""
- Docs
- Python not yet supported