Table 2 An explanation of the MBConv and Fused-MBConv blocks in the EfficientNetV2L architecture, including important settings and parameters inside these vital parts.

From: A machine learning approach for vocal fold segmentation and disorder classification based on ensemble method

Stage

Operator

Stride

Channels

Layer

0

Conv,3\(\times\)3

2

32

1

1

Fused-MBConv1,3\(\times\)3

1

32

49

2

Fused-MBConv4,3x3

2

64

7

3

Fused-MBConv4,3\(\times\)3

2

96

7

4

MBConv4 SE 0.25,3\(\times\)3

2

192

10

5

MBConv6 SE 0.25,3\(\times\)3

1

224

19

6

MBConv6 SE 0.25,3\(\times\)3

2

384

25

7

MBConv6 SE 0.25,3\(\times\)3

1

640

7

8

Conv1x1 & Pooling & FC

1280

1