← Quit0 / 0 correct
LLMs & AIHard1 of 20

Why do grouped-query attention (GQA) and multi-query attention (MQA) speed up inference?