vLLM

ai-ml

High-throughput and memory-efficient inference engine for large language models

Pronunciation

Correct

V-L-L-M

/viː ɛl ɛl ɛm/

The 'v' is pronounced as the letter, followed by LLM spelled out. The project name stands for 'virtually unlimited LLM'.

Source: docs.vllm.ai(official spec)