TReX- Reusing Vision Transformer's Attention for Efficient Xbar-based Computing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Moitra, Abhishek, Bhattacharjee, Abhiroop, Kim, Youngeun, Panda, Priyadarshini
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913477229019136
author Moitra, Abhishek
Bhattacharjee, Abhiroop
Kim, Youngeun
Panda, Priyadarshini
author_facet Moitra, Abhishek
Bhattacharjee, Abhiroop
Kim, Youngeun
Panda, Priyadarshini
contents Due to the high computation overhead of Vision Transformers (ViTs), In-memory Computing architectures are being researched towards energy-efficient deployment in edge-computing scenarios. Prior works have proposed efficient algorithm-hardware co-design and IMC-architectural improvements to improve the energy-efficiency of IMC-implemented ViTs. However, all prior works have neglected the overhead and co-depencence of attention blocks on the accuracy-energy-delay-area of IMC-implemented ViTs. To this end, we propose TReX- an attention-reuse-driven ViT optimization framework that effectively performs attention reuse in ViT models to achieve optimal accuracy-energy-delay-area tradeoffs. TReX optimally chooses the transformer encoders for attention reuse to achieve near iso-accuracy performance while meeting the user-specified delay requirement. Based on our analysis on the Imagenet-1k dataset, we find that TReX achieves 2.3x (2.19x) EDAP reduction and 1.86x (1.79x) TOPS/mm2 improvement with ~1% accuracy drop in case of DeiT-S (LV-ViT-S) ViT models. Additionally, TReX achieves high accuracy at high EDAP reduction compared to state-of-the-art token pruning and weight sharing approaches. On NLP tasks such as CoLA, TReX leads to 2% higher non-ideal accuracy compared to baseline at 1.6x lower EDAP.
format Preprint
id arxiv_https___arxiv_org_abs_2408_12742
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle TReX- Reusing Vision Transformer's Attention for Efficient Xbar-based Computing
Moitra, Abhishek
Bhattacharjee, Abhiroop
Kim, Youngeun
Panda, Priyadarshini
Artificial Intelligence
Hardware Architecture
Due to the high computation overhead of Vision Transformers (ViTs), In-memory Computing architectures are being researched towards energy-efficient deployment in edge-computing scenarios. Prior works have proposed efficient algorithm-hardware co-design and IMC-architectural improvements to improve the energy-efficiency of IMC-implemented ViTs. However, all prior works have neglected the overhead and co-depencence of attention blocks on the accuracy-energy-delay-area of IMC-implemented ViTs. To this end, we propose TReX- an attention-reuse-driven ViT optimization framework that effectively performs attention reuse in ViT models to achieve optimal accuracy-energy-delay-area tradeoffs. TReX optimally chooses the transformer encoders for attention reuse to achieve near iso-accuracy performance while meeting the user-specified delay requirement. Based on our analysis on the Imagenet-1k dataset, we find that TReX achieves 2.3x (2.19x) EDAP reduction and 1.86x (1.79x) TOPS/mm2 improvement with ~1% accuracy drop in case of DeiT-S (LV-ViT-S) ViT models. Additionally, TReX achieves high accuracy at high EDAP reduction compared to state-of-the-art token pruning and weight sharing approaches. On NLP tasks such as CoLA, TReX leads to 2% higher non-ideal accuracy compared to baseline at 1.6x lower EDAP.
title TReX- Reusing Vision Transformer's Attention for Efficient Xbar-based Computing
topic Artificial Intelligence
Hardware Architecture
url https://arxiv.org/abs/2408.12742