ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design

Ningning MaXiangyu ZhangHai-Tao ZhengJian Sun

   Papers with code   Abstract  PDF

Datasets, Transforms and Models specific to Computer Vision..

Benchmarked Models

RANK
MODEL
REPO
CODE RESULT
PAPER RESULT
ε-REPRODUCED
BUILD
1
ShuffleNetV2b x2.0
74.4%
--
2
ShuffleNetV2 x2.0
74.1%
--
3
ShuffleNetV2b x1.5
72.7%
--
4
ShuffleNetV2 x1.5
72.5%
--
5
ShuffleNetV2b x1.0
69.4%
--
6
ShuffleNet V2
(1x)
69.0%
69.4%
7
ShuffleNetV2 x1.0
68.6%
--
8
ShuffleNet V2
(0.5x)
60.0%
60.3%
9
ShuffleNetV2b x0.5
59.7%
--
10
ShuffleNetV2 x0.5
59.0%
--