Squeeze-and-Excitation Networks

Jie HuLi ShenSamuel AlbanieGang SunEnhua Wu

   Papers with code   Abstract  PDF

The central building block of convolutional neural networks (CNNs) is the convolution operator, which enables networks to construct informative features by fusing both spatial and channel-wise information within local receptive fields at each layer. A broad range of prior research has investigated the spatial component of this relationship, seeking to strengthen the representational power of a CNN by enhancing the quality of spatial encodings throughout its feature hierarchy... (read more)

Benchmarked Models

RANK
MODEL
REPO
CODE RESULT
PAPER RESULT
ε-REPRODUCED
BUILD
1
SENet-154
81.4%
82.7%
3
SE-ResNeXt-101 32x4d
80.2%
--
4
SE-ResNeXt-101
(32x4d)
80.0%
--
5
SE-ResNet-50b
79.2%
--
7
SE-ResNeXt-50
(32x4d)
79.0%
--
9
SE-ResNet-152
78.5%
--
10
SE-ResNet-BC-38b
78.4%
--
11
SE-PreResNet-BC-38b
78.4%
--
13
SE-ResNet-101
78.1%
--
14
SENet-28
78.1%
--
16
SE-ResNet-50
77.5%
--
17
SE-ResNeXt-26 32x4d
77.1%
--
18
SE-PreResNet-BC-26b
76.8%
--
19
SE-ResNet-BC-26b
76.4%
--
21
SENet-16
74.4%
--
22
SE-ResNet-26
74.3%
--
23
SE-PreResNet-18
71.9%
--
24
SE-ResNet-18
71.8%
--
26
SE-ResNet-10
66.1%
--
27
SE-PreResNet-10
66.0%
--