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
- [WIP] Enhance MoE Triton kernel & tuning
- [New Model]: Cogagent
- [Doc]: i want to know. How to run vllms with remote ray cluster
- [Bug]: Tensorizer model loader blocks multi-GPU loading even for HF serialized models
- [BugFix] Fix fp8 quantizer
- [Misc]: int4 support on CPU backend
- [Bug]: Not picking up Neuron on instance (AssertionError: CUDA_HOME is not set)
- [Bug]: Setting best_of and n in SamplingParameters makes the performance of LLaMa3 on GSM8K worse
- [Usage]: In the VLLM background log, only the question asking log can be viewed, but the answer log generated by the large model cannot be viewed. How to display the answer generated by the large model in the log?
- [Bug]: Ray on multi machine cluster fails to detect all nodes.
- Docs
- Python not yet supported