HARDSEA: Hybrid Analog-ReRAM Clustering and Digital-SRAM In-Memory Computing Accelerator for Dynamic Sparse Self-Attention in Transformer-Reference-Cited by-同舟云学术

HARDSEA: Hybrid Analog-ReRAM Clustering and Digital-SRAM In-Memory Computing Accelerator for Dynamic Sparse Self-Attention in Transformer

Published:2024-02 Issue:2 Volume:32 Page:269-282
ISSN:1063-8210
Container-title:IEEE Transactions on Very Large Scale Integration (VLSI) Systems
language:
Short-container-title:IEEE Trans. VLSI Syst.

Author:

Liu Shiwei¹^ORCID,Mu Chen¹,Jiang Hao¹,Wang Yunzhengmao¹^ORCID,Zhang Jinshan¹,Lin Feng¹,Zhou Keji¹^ORCID,Liu Qi¹^ORCID,Chen Chixiao¹^ORCID

Affiliation:

1. State Key Laboratory of Integrated Chips and Systems, Fudan University, Shanghai, China

Funder

Major Project of the Science and Technology Innovation 2030

Strategic Priority Research Program of Chinese Academy of Sciences

National Natural Science Foundation of China

Shanghai Rising-Star Program

Zhangjiang Fudan International Innovation Center

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Subject

Electrical and Electronic Engineering,Hardware and Architecture,Software

Link

Reference60 articles.

1. Attention is all you need;Vaswani

2. BERT: Pre-training of deep bidirectional transformers for language understanding;Devlin;arXiv:1810.04805,2018

3. Language models are unsupervised multitask learners;Radford;OpenAI Blog,2019

4. Language models are few-shot learners;Brown;arXiv:2005.14165,2020

Cited by 1 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