Skip to main content

Table 3 The summary of EfficientNet layered architecture. Architecture

From: Review of deep learning methods for remote sensing satellite images classification: experimental survey and comparative analysis

Block

Units (n)

Operators (F)

Resolutions(H \(\times\) W)

Channels

Block A

12

Conv, BatchNorm, Swish, AdaptiveAvgPool

112 \(\times\) 112, 56 \(\times\) 56, 28 \(\times\) 28, 14 \(\times\) 14, 7 \(\times\) 7

32,96,144,

Block B

4

Conv, BatchNorm, Swish, AdaptiveAvgPool

112 \(\times\) 112, 56 \(\times\) 56, 28 \(\times\) 28, 14 \(\times\) 14, 7 \(\times\) 7

96,144,240

Block C

4

Conv, BatchNorm, Swish

112 \(\times\) 112, 56 \(\times\) 56, 28 \(\times\) 28, 14 \(\times\) 14, 7 \(\times\) 7

96,144,240

Block D

16

Conv, Swish

1 \(\times\) 1

8,4,6

Block E

16

Conv, Sigmoid

1 \(\times\) 1

32,96,144

Block F

16

Conv, BatchNorm

112 \(\times\) 112, 56 \(\times\) 56, 28 \(\times\) 28, 14 \(\times\) 14, 7 \(\times\) 7

24,40,80

Block G

1

Linear, Relu, BatchNorm

7 \(\times\) 7

512