Dual Path Networks

Yunpeng ChenJianan LiHuaxin XiaoXiaojie JinShuicheng YanJiashi Feng

   Papers with code   Abstract  PDF

In this work, we present a simple, highly efficient and modularized Dual Path Network (DPN) for image classification which presents a new topology of connection paths internally. By revealing the equivalence of the state-of-the-art Residual Network (ResNet) and Densely Convolutional Network (DenseNet) within the HORNN framework, we find that ResNet enables feature re-usage while DenseNet enables new features exploration which are both important for learning good representations... (read more)

Benchmarked Models

RANK
MODEL
REPO
CODE RESULT
PAPER RESULT
ε-REPRODUCED
BUILD
1
DPN-107
(320x320, Mean-Max Poo…)
81.8%
--
2
DPN-131
(320x320, Mean-Max Poo…)
81.4%
81.5%
3
DPN-92
(320x320, Mean-Max Poo…)
81.3%
81.0%
4
DPN-98
(320x320, Mean-Max Poo…)
81.2%
81.3%
7
DPN-131
(224x224)
79.8%
80.1%
10
DPN-68b
(320x320, Mean-Max Poo…)
79.4%
--
12
DPN-68
(320x320, Mean-Max Poo…)
78.5%
78.5%
15
DPN-68
(224x224)
76.3%
76.4%