inference/research
verifiedhigh · 0.90updated 2026-10-03

Fast Transformer Decoding: One Write-Head is All You Need

paper

Fields

id
sources/pap-attn-multi-query-attention
URL
https://arxiv.org/abs/1911.02150
Publisher
arXiv
Kind
paper
Published
2019-11-06
Accessed
2026-10-03
notes
arXiv preprint. Widely cited as the origin of MQA; a later venue version exists but was not confirmed here, so venue stays arxiv-preprint.

Notes

arXiv preprint. Widely cited as the origin of MQA; a later venue version exists but was not confirmed here, so venue stays arxiv-preprint.

Cross-references

This record references no other records.