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
- [Usage]: How to use ROPE scaling for llama3.1 and gemma2?
- [Installation]: can't get the cu118 version of vllm 0.6.3 by https://github.com/vllm-project/vllm/releases/download/v0.6.3/vllm-0.6.3+cu118-cp310-cp310-manylinux1_x86_64.whl
- Support softcap in ROCm Flash Attention
- [Core] Implement disagg prefill by StatelessProcessGroup
- Turn on V1 for H200 build
- [Bug]: Model does not split in multiple Gpus instead it occupy same memory on each GPU
- [Feature]: Support for Registering Model-Specific Default Sampling Parameters
- [Frontend] Add Command-R and Llama-3 chat template
- [V1] VLM prefix caching: Add hashing of images
- [core] overhaul memory profiling and fix backward compatibility
- Docs
- Python not yet supported