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] Ray Backend V1
- [Bug]: I cannot able to load the model on TESLA T4 GPU in Full precision
- [Performance]: Any up-to-date and convincing benchmark for chosing the fastest engine ?
- [Bug]: falconmambaForCasualLM- tiiuae/falcon-mamba-7b-instruct
- [Bug]: [Regression Issue] The output from Qwen2 VL are different between vLLM v0.6.3-post1 and vLLM v0.6.1-post2
- [Usage]: How to use `llava-hf/llava-1.5-7b-hf` with bitsandbytes quantization in `vllm serve`?
- [Bugfix] Fix crash in multi-step mode when set skip_tokenizer_init=True
- [Usage]: Are there any batch size requirements for offline batch inference? For example, is 10,000 okay?
- [Performance]: FP8 performance worse than FP16 for Qwen2-VL-2B-Instruct
- [Doc]: batch example file has outdated parameters
- Docs
- Python not yet supported