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
- [Doc] Add offline distributed continuous batching example
- [Build] Fix for the Wswitch-bool clang warning
- [Performance]: Throughput and Latency degradation with a single LoRA adapter on A100 40 GB
- [New Model]: Support Tencent-Hunyuan-Large
- [CI] Move slow entrypoints tests out of fastcheck
- [Core] Add dynamic chunk size calculation
- [Usage]: Model architectures ['LlavaNextForConditionalGeneration'] are not supported for now
- [Bug]: For speculative decoding with a draft model, the "determine_num_available_blocks" only considers the memory usage of the target model
- [Bug]: last_token_time is equal to arrival_time
- [Bug]: vLLM multi-step scheduling crashes when input prompt is long
- Docs
- Python not yet supported