Efficient Transformer Inference with Statically Structured Sparse Attention-Reference-Cited by-同舟云学术

登录注册会员服务联系我们

Efficient Transformer Inference with Statically Structured Sparse Attention

Published:2023-07-09 Issue: Volume: Page:
ISSN:
Container-title:2023 60th ACM/IEEE Design Automation Conference (DAC)
language:
Short-container-title:

Author:

Dai Steve¹,Genc Hasan²,Venkatesan Rangharajan¹,Khailany Brucek¹

Affiliation:

1. NVIDIA

2. University of California,Berkeley

Publisher

IEEE

Link

http://xplorestaging.ieee.org/ielx7/10247654/10247655/10247993.pdf?arnumber=10247993

Reference27 articles.

1. Linformer: Self-attention with linear complexity;wang,2020

2. Blockwise self-attention for long document understanding;qiu,2019

3. Longformer: The long-document transformer;beltagy,2020

4. Generating long sequences with sparse transformers;child,2019

5. MAGNet: A Modular Accelerator Generator for Neural Networks

Cited by 2 articles. 订阅此论文施引文献订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献

1. ALISA: Accelerating Large Language Model Inference via Sparsity-Aware KV Caching;2024 ACM/IEEE 51st Annual International Symposium on Computer Architecture (ISCA);2024-06-29

2. Training and Serving System of Foundation Models: A Comprehensive Survey;IEEE Open Journal of the Computer Society;2024

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线，采集、加工和组织学术论文而形成的新型学术文献查询和分析系统，可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容，当前同舟云学术共收录了国内外主流学术期刊6万余种，收集的期刊论文及会议论文总量共计约1.5亿篇，并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询！咨询电话：010-8811{复制后删除}0370

www.globalauthorid.com

TOP