HIRI-ViT: Scaling Vision Transformer with High Resolution Inputs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yao, Ting, Li, Yehao, Pan, Yingwei, Mei, Tao
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910371175989248
author Yao, Ting
Li, Yehao
Pan, Yingwei
Mei, Tao
author_facet Yao, Ting
Li, Yehao
Pan, Yingwei
Mei, Tao
contents The hybrid deep models of Vision Transformer (ViT) and Convolution Neural Network (CNN) have emerged as a powerful class of backbones for vision tasks. Scaling up the input resolution of such hybrid backbones naturally strengthes model capacity, but inevitably suffers from heavy computational cost that scales quadratically. Instead, we present a new hybrid backbone with HIgh-Resolution Inputs (namely HIRI-ViT), that upgrades prevalent four-stage ViT to five-stage ViT tailored for high-resolution inputs. HIRI-ViT is built upon the seminal idea of decomposing the typical CNN operations into two parallel CNN branches in a cost-efficient manner. One high-resolution branch directly takes primary high-resolution features as inputs, but uses less convolution operations. The other low-resolution branch first performs down-sampling and then utilizes more convolution operations over such low-resolution features. Experiments on both recognition task (ImageNet-1K dataset) and dense prediction tasks (COCO and ADE20K datasets) demonstrate the superiority of HIRI-ViT. More remarkably, under comparable computational cost ($\sim$5.0 GFLOPs), HIRI-ViT achieves to-date the best published Top-1 accuracy of 84.3% on ImageNet with 448$\times$448 inputs, which absolutely improves 83.4% of iFormer-S by 0.9% with 224$\times$224 inputs.
format Preprint
id arxiv_https___arxiv_org_abs_2403_11999
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle HIRI-ViT: Scaling Vision Transformer with High Resolution Inputs
Yao, Ting
Li, Yehao
Pan, Yingwei
Mei, Tao
Computer Vision and Pattern Recognition
Multimedia
The hybrid deep models of Vision Transformer (ViT) and Convolution Neural Network (CNN) have emerged as a powerful class of backbones for vision tasks. Scaling up the input resolution of such hybrid backbones naturally strengthes model capacity, but inevitably suffers from heavy computational cost that scales quadratically. Instead, we present a new hybrid backbone with HIgh-Resolution Inputs (namely HIRI-ViT), that upgrades prevalent four-stage ViT to five-stage ViT tailored for high-resolution inputs. HIRI-ViT is built upon the seminal idea of decomposing the typical CNN operations into two parallel CNN branches in a cost-efficient manner. One high-resolution branch directly takes primary high-resolution features as inputs, but uses less convolution operations. The other low-resolution branch first performs down-sampling and then utilizes more convolution operations over such low-resolution features. Experiments on both recognition task (ImageNet-1K dataset) and dense prediction tasks (COCO and ADE20K datasets) demonstrate the superiority of HIRI-ViT. More remarkably, under comparable computational cost ($\sim$5.0 GFLOPs), HIRI-ViT achieves to-date the best published Top-1 accuracy of 84.3% on ImageNet with 448$\times$448 inputs, which absolutely improves 83.4% of iFormer-S by 0.9% with 224$\times$224 inputs.
title HIRI-ViT: Scaling Vision Transformer with High Resolution Inputs
topic Computer Vision and Pattern Recognition
Multimedia
url https://arxiv.org/abs/2403.11999