VGG13的参数量
conv1:
3
×
3
×
3
×
64
+
64
=
1
,
792
3\times3\times3\times64+64 = 1,792
3×3×3×64+64=1,792
conv2:
3
×
3
×
64
×
64
+
64
=
36
,
928
3\times3\times64\times64+64 = 36,928
3×3×64×64+64=36,928
conv3:
3
×
3
×
64
×
128
+
128
=
73
,
856
3\times3\times64\times128+128 = 73,856
3×3×64×128+128=73,856
conv4:
3
×
3
×
128
×
128
+
128
=
147
,
584
3\times3\times128\times128+128 = 147,584
3×3×128×128+128=147,584
conv5:
3
×
3
×
128
×
256
+
256
=
295
,
168
3\times3\times128\times256+256 = 295,168
3×3×128×256+256=295,168
conv6:
3
×
3
×
256
×
256
+
256
=
590
,
080
3\times3\times256\times256+256 = 590,080
3×3×256×256+256=590,080
conv7:
3
×
3
×
256
×
512
+
512
=
1
,
180
,
160
3\times3\times256\times512+512 = 1,180,160
3×3×256×512+512=1,180,160
conv8:
3
×
3
×
512
×
512
+
512
=
2
,
359
,
808
3\times3\times512\times512+512 = 2,359,808
3×3×512×512+512=2,359,808
conv9:
3
×
3
×
512
×
512
+
512
=
2
,
359
,
808
3\times3\times512\times512+512 = 2,359,808
3×3×512×512+512=2,359,808
conv10:
3
×
3
×
512
×
512
+
512
=
2
,
359
,
808
3\times3\times512\times512+512 = 2,359,808
3×3×512×512+512=2,359,808
卷积层共有trainable参数的个数为:9,404,992