ResNeSt：Split-Attention网络，提升ResNet性能

需积分: 32 89 浏览量更新于2024-07-16 收藏 549KB PDF 举报

"ResNeSt Split-Attention Networks是ResNet的一种变形，它通过引入Split-Attention块提高了模型的性能，并且在保持ResNet结构的基础上，适用于各种下游任务，如对象检测和语义分割，而无需增加额外的计算成本。" 在计算机视觉领域，ResNet（残差网络）因其深度学习模型的简单性和模块化结构，长期以来一直是许多任务如图像分类、目标检测和语义分割的首选骨干网络。然而，随着技术的进步，研究者们一直在寻找能进一步提升性能的方法。ResNeSt（Split-Attention Networks）就是这样的一个创新，它由Hang Zhang等人提出，旨在增强ResNet的注意力机制。 Split-Attention块是ResNeSt的核心，其设计目的是允许特征图组之间的注意力交互。传统的自注意力机制通常会计算所有位置的全局依赖，这在计算上可能非常昂贵。相比之下，Split-Attention块将特征图分成多个分组，每个分组内部进行注意力计算，然后将这些分组的结果合并，这样既实现了注意力的分散，又降低了计算复杂性。通过以ResNet风格堆叠这些Split-Attention块，ResNeSt网络得以构建。这种网络保留了ResNet的基本结构，可以直接用于下游任务，而且由于其优化的设计，不增加额外的计算负担。这使得ResNeSt在保持效率的同时，增强了模型的表达能力和适应性。实验结果显示，ResNeSt模型在性能上优于其他具有相似模型复杂度的网络。例如，ResNeSt-50模型在使用单个224x224的图像裁剪尺寸时，能够在ImageNet数据集上达到81.13%的Top-1准确率，显示出其在图像分类任务上的强大能力。此外，ResNeSt的优秀表现也体现在目标检测和语义分割等应用中，表明Split-Attention机制对于提升这些任务的性能同样有效。这使得ResNeSt成为研究人员和开发者在处理复杂视觉问题时的一个有力工具，特别是在需要平衡性能和计算效率的情况下。 ResNeSt Split-Attention Networks通过引入分组注意力机制，成功地增强了ResNet的性能，同时保持了易于使用和计算效率高的优点，是当前计算机视觉领域中值得探索和应用的网络架构之一。

ResNeSt: Split-Attention Networks 5

3.1 Split-Attention Block

Our Split-Attention block is a computational unit, consisting feature-map group

and split attention operations. Figure 1 (Right) depicts an overview of a Split-

Attention Block.

Feature-map Group. As in ResNeXt blocks [61], the input feature-map can

be divided into several groups along the channel dimension, and the number of

feature-map groups is given by a cardinality hyperparameter K. We refer to the

resulting feature-map groups as cardinal groups. We introduce a new radix hy-

perparameter R that dictates the number of splits within a cardinal group. Then

the block input X are split into G = KR groups along the channel dimension

X = {X

, X

, ...X

} as shown in Figure 1. We apply a series of transformations

, F

, ...F

} to each individual group, then the intermediate representation

of each group can be represented as U

= F

), for i ∈ {1, 2, ...G}.

Input 1 Input r

…

Input 2

Global pooling

Dense c’ + BN + ReLU

Dense c

r-Softmax

Dense c

x x

(h, w, c)

(c, )

(c’, )

(c, )

(h, w, c)

Fig. 2: Split-Attention within a

cardinal group. For easy visu-

alization in the ﬁgure, we use

c = C/K in this ﬁgure.

Split Attention in Cardinal Groups. Fol-

lowing [30, 38], a combined representation for

each cardinal group can be obtained by fusing

via an element-wise summation across mul-

tiple splits. The representation for k-th car-

dinal group is

i=R(k−1)+1

∈

H×W×C/K

for k ∈ 1, 2, ...K. Global contex-

tual information with embedded channel-wise

statistics can be gathered with global aver-

age pooling across spatial dimensions s

∈

C/K

[29,38]. Here the c-th component is cal-

culated as:

H × W

i=1

j=1

(i, j). (1)

A weighted fusion of the cardinal group

representation V

∈ R

H×W×C/K

is aggre-

gated using channel-wise soft attention, where

each feature-map channel is produced using a

weighted combination over splits. The c-th channel is calculated as:

i=1

(c)U

R(k−1)+r

, (2)

where a







exp(G

))

j=0

exp(G

))

if R > 1,

1+exp(−G

))

if R = 1,

(3)

and mapping G

determines the weight of each split for the c-th channel based

on the global context representation s

剩余21页未读，继续阅读

佑林杉

粉丝: 10
资源: 28

ResNeSt：Split-Attention网络，提升ResNet性能

ResNeSt:需求

ResNeSt_Split-Attention_Network_ResNeSt.zip

Python库 | CTH_sentence_split-0.0.11.tar.gz

Python库 | CTH_sentence_split-0.0.16-py3-none-any.whl

PyPI 官网下载 | s3_upload_split-0.1.6.dev4-py3-none-any.whl

Python库 | pymysql_split_tool-0.5.6.tar.gz

F_SPLIT-字符函数.sql

resnest: split-attention networks

data_key = image_file.split('/')[-1].split('.')[0] label_key = label_file.split('/')[-1].split('.')[0]

最新资源