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
- Yi-34B-200K have empty output under default config (max_position_embedding=20000)
- Refactor openai completion api w.r.t Prefix Cache
- avoid decode_sequence when only input prompt_token_ids
- When using tp =2 to inference model in greedy mode and vllm 0.3 , results are random
- Nvidia-H20 with nvcr.io/nvidia/pytorch:23.12-py3，CUBLAS Error！
- AutoQuant: Automatic quantization model for INT8/INT4 inference, INT4 runs faster than AWQ and FP16.
- Multi-GPU Support Failures with AMD MI210
- The service results based on vllm qwen7B are inconsistent with the original qwen results, and the accuracy will drop significantly
- Fix: Echo without asking for new tokens or logprobs in OpenAI Completions API
- Failed to find C compiler. Please specify via CC environment variable
- Docs
- Python not yet supported