Title: GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints
Authors: Joshua Ainslie, James Lee-Thorp, Michiel de Jong, Yury Zemlyanskiy, Federico LebrĂłn, Sumit Sanghai
Published: 22nd May 2023 (Monday) @ 17:16:38
Link: http://arxiv.org/abs/2305.13245v1
Abstract
Multi-query attention (MQA), which only uses a single key-value head, drastically speeds up decoder inference. However, MQA can lead to quality degradation, and moreover it may not be desirable to train a separate model just for faster inference. We (1) propose a recipe for uptraining existing multi-head language model checkpoints into models with MQA using 5% of original pre-training compute, and (2) introduce grouped-query attention (GQA), a generalization of multi-query attention which uses an intermediate (more than one, less than number of query heads) number of key-value heads. We show that uptrained GQA achieves quality close to multi-head attention with comparable speed to MQA.