neuralmagic / nm-vllm

A high-throughput and memory-efficient inference and serving engine for LLMs
https://nm-vllm.readthedocs.io
Other
251 stars 10 forks source link