Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net

Xingang PanPing LuoJianping ShiXiaoou Tang

   Papers with code   Abstract  PDF

Convolutional neural networks (CNNs) have achieved great successes in many computer vision problems. Unlike existing works that designed CNN architectures to improve performance on a single task of a single domain and not generalizable, we present IBN-Net, a novel convolutional architecture, which remarkably enhances a CNN's modeling ability on one domain (e.g... (read more)

Benchmarked Models

RANK
MODEL
REPO
CODE RESULT
PAPER RESULT
ε-REPRODUCED
BUILD
1
IBN-ResNeXt-101
(32x4d)
79.1%
--
2
IBN-ResNet-101
78.7%
--
3
IBN-ResNet-50
77.2%
--
4
IBN-DenseNet-169
76.8%
--
5
IBN(b)-ResNet-50
76.4%
--
6
IBN-DenseNet-121
75.5%
--