Skip to main content
Journal of Big Data

Table 4 Number of experiments per activation and dataset

From: Evaluation of maxout activations in deep learning across several big data domains

Dataset

LR

M21

M31

M32

M61

R

R2X

R3X

R6X

SL

T

CIFAR-10

5

5

5

5

5

5

5

5

5

5

5

CIFAR-100

2

2

2

2

2

2

2

2

2

2

2

F-MNIST

5

5

5

5

5

5

5

5

5

5

5

MNIST

5

5

5

5

5

5

5

5

5

5

5

LFW

2

2

2

2

2

2

2

2

2

2

2

MS-Celeb

2

2

2

2

0

2

2

0

0

2

2

Amazon1M

2

2

2

2

2

2

2

0

0

1

2

Amazon4M

2

2

2

1

1

2

2

0

0

0

2

Sent140

2

2

2

2

2

2

2

0

0

2

2

Yelp500K

2

2

2

2

2

2

2

0

0

0

2

Yelp1M

2

2

2

2

2

2

2

0

0

1

2

Med Part B

5

5

5

5

5

5

5

0

0

5

5

Med Part D

5

5

5

5

5

5

5

0

0

5

5

DMEPOS

5

5

5

5

5

5

5

0

0

5

5

Combined CMS

5

5

5

5

5

5

5

0

0

5

5

GSC

2

2

2

2

2

2

2

0

0

2

2

IRMAS

2

2

2

2

2

2

2

0

0

2

2

IDMT-SMT-Audio-Effects

2

2

2

2

2

2

2

0

0

2

2

  1. LReLU (LR), Maxout 2-1 (M21), Maxout 3-1 (M31), Maxout 3-2 (M32), Maxout 6-1 (M61), ReLU (R), ReLU2x (R2X), ReLU3x (R3X), ReLU6x (R6X), SeLU (SL), Tanh (T)
Back to article page