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
- [Feature]: Support for 1.58-bit models.
- [Bug]: Incomplete XML Output for Certain Prompts with Temperature=0
- [Performance]: vllm Eagle performance is worse than expected
- [Bug]: The driver_worker gets stuck 100% of the time, when using Medusa with TP > 1
- [Usage]: OpenAI-Compatible Server online Batch inference
- [Performance]: The performance of version 0.6.3 is weaker than that of version 0.6.2 in stress testing.
- [Feature]: Support for Controlled Decoding
- [New Model]: bert-base-chinese
- [WIP] Disaggregated prefilling support X prefill + Y decode
- OOM error :When using four 4500 ada cards to start four lora instances, an error occurs. However, no error occurs when not starting lora on the four 4500 ada cards, and there is no error when starting four lora instances on a single A100 card.
- Docs
- Python not yet supported