Table 12 Training and validation results for the models.

From: Attention-Enhanced CNNs and transformers for accurate monkeypox and skin disease detection

Trial

Edge loss weight

Learning rate

Mini batch size

Training accuracy (%)

Training loss

Validation accuracy (%)

Validation loss

EfficientNet B0

1

-

0.001

32

98.07

0.0582

94.81

0.1634

2

-

0.001

32

97.32

0.0799

95.01

0.1814

3

0.5 (Canny)

0.001

32

98.60

0.0634

94.74

0.1682

4

0.5 (Canny)

0.001

32

96.87

0.1167

93.28

0.2031

MobileNet V2

1

-

0.001

32

100.0

0.0006

98.67

0.0552

2

0.5 (Sobel)

0.001

32

99.10

0.1753

97.01

0.2493

3

0.1 (Sobel)

0.001

32

100.0

0.0293

99.00

0.0886

4

0.05 (Sobel)

0.001

32

99.90

0.0179

98.47

0.1019

5

0.025 (Sobel)

0.001

32

99.75

0.0152

98.20

0.0750

ResNet-50

1

-

0.001

32

96.44

0.0980

94.75

0.1620

Xception

1

-

0.001

32

99.86

0.0156

99.92

0.0113

Swin Transformer

1

-

0.001

32

99.24

0.4985

98.03

0.4870

2

-

0.001

32

99.63

0.4358

98.88

0.4519

Back to article page