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 supported3 Subscribers
Add a CodeTriage badge to vllm
Help out
- Issues
- [Core] Default to using per_token quantization for fp8 when cutlass is supported.
- [Misc] add non cuda hf benchmark_througput
- [Bug]: [Usage]: is_xpu should return true when the torch.xpu.is_available is true even w/o IPEX
- [Bug]: Wrong Response with Gemma2 with 8k context length
- [Feature]: DRY Sampling
- [Usage]: How to run VLLM on multiple tpu hosts V4-32
- [Usage]: Standalone Debugging and Measuring the vLLM Engine Backend
- [Misc]: In vllm, I tested that the speed of concurrent server api requests is greater than the speed of offline inference. I would like to ask if there are any performance tests on the official vllm website. Can you tell me? Thank you.
- [Bug]: Wrong "completion_tokens" counts in streaming usage
- [Feature]: OpenAI o1-like Chain-of-thought (CoT) inference workflow
- Docs
- Python not yet supported