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][Quantization] Floor UE8M0 scales at 1e-10 in per-token-group quant Triton paths
- [Feature]: Opt-in omission of unset vLLM-specific extension fields from /v1/chat/completions responses
- [Bugfix] Surface SamplingParams.verify errors as client errors (HTTP 400)
- [Bug]: tool_call_parser_invocations_total no longer increments for engine-backed parsers (regression after ParserEngine refactor)
- [Bugfix] Include num_speculative_tokens in SpeculativeConfig compute_hash
- [Bugfix][Frontend] Match default MM LoRA for OpenAI input_audio parts
- [Docs] Clarify the cache_salt trust boundary and quantify its cost
- [Bugfix][Spec Decode] Isolate DFlash/DSpark compile cache by depth
- [Bug]: [ROCm][Spec Decode] DFlash2 acceptance length collapses with default ROCM_ATTN draft backend
- [LoRA] Prepare Punica metadata on CPU
- Docs
- Python not yet supported