Skip to main content

Table 4 Configuration of proposed pruned CNN model

From: PFDI: a precise fruit disease identification model based on context data fusion with faster-CNN in edge computing environment

Layer (type)

Output shape

Param #

Prune__low__magnitude__conv2d

(None, 256, 256, 32)

1762

Prune__low__magnitude__activation

(None, 256, 256, 32)

1

Prune__low__magnitude__conv2__1

(None, 256, 256, 32)

36,930

Prune__low__magnitude__activation

(None, 256, 256, 32)

1

Prune__low__magnitude__max __ pool

(None, 256, 256, 32)

1

Prune__low__magnitude__dropout

(None, 256, 256, 32)

1

Prune__low__magnitude__conv2d_2

(None, 256, 256, 32)

73,794

Prune__low__magnitude__activation

(None, 256, 256, 32)

1

Prune__low__magnitude__con2d__3

(None, 256, 256, 32)

73,794

Prune__low__magnitude__activation

(None, 256, 256, 32)

12

Prune__low__magnitude__max__pool

(None, 256, 256, 32)

1

Prune__low__magnitude__dropout

(None, 256, 256, 32)

1

Prune__low__magnitude__flatten

(None,28,224)

1

Prune__low__magnitude__dense

(None,512)

28,901,890

Prune__low__magnitude__activation

(None,512)

1

Prune__low__magnitude__dropout

(None,512)

1

Prune__low__magnitude__dense__1

(None,5)

5127

Prune__low__magnitude__activation

(None,5)

1