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 supported4 Subscribers
Add a CodeTriage badge to vllm
Help out
- Issues
- [Misc] Added num_cumulative_preemption metrics
- [Bug]: AssertionError when using automatic prefix caching and prompt_logprobs
- [Performance]: Using vLLM for Llama3.1 405b fp8 on 8xH100 yields poor throughput
- [Bugfix] Fix LongRoPE bug
- [Usage]: Single-node multi-GPU inference
- [Feature]: BitsandBytes quantization with TP>1
- Reshape cache flash kernel to support HND layout
- Add VLLM_LOGGING_INTERVAL_SEC envvar to control logging rate
- [Feature]: Allow partial context in speculative decoding when using draft models with smaller context than target model
- [Bug]: watchdog thread terminated with exception: CUDA error: an illegal memory access was encountered
- Docs
- Python not yet supported