File size: 9,368 Bytes
b410f90
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
CifNetForImageClassification(
  (resnet): CifNetModel(
    (embedder): CifNetEmbeddings(
      (embedder): CifNetConvLayer(
        (convolution): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
        (normalization): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (activation): SiLU()
      )
      (pooler): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    )
    (encoder): CifNetEncoder(
      (stages): ModuleList(
        (0): CifNetStage(
          (layers): Sequential(
            (0): CifNetBasicLayer(
              (shortcut): Identity()
              (layer): Sequential(
                (0): CifNetConvLayer(
                  (convolution): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
                (1): CifNetConvLayer(
                  (convolution): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
              )
            )
          )
        )
        (1): CifNetStage(
          (layers): Sequential(
            (0): CifNetBasicLayer(
              (shortcut): CifNetShortCut(
                (convolution): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
                (normalization): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
              )
              (layer): Sequential(
                (0): CifNetConvLayer(
                  (convolution): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
                (1): CifNetConvLayer(
                  (convolution): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
              )
            )
          )
        )
        (2): CifNetStage(
          (layers): Sequential(
            (0): CifNetBasicLayer(
              (shortcut): CifNetShortCut(
                (convolution): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
                (normalization): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
              )
              (layer): Sequential(
                (0): CifNetConvLayer(
                  (convolution): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
                (1): CifNetConvLayer(
                  (convolution): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
              )
            )
          )
        )
        (3): CifNetStage(
          (layers): Sequential(
            (0): CifNetBasicLayer(
              (shortcut): CifNetShortCut(
                (convolution): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
                (normalization): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
              )
              (layer): Sequential(
                (0): CifNetConvLayer(
                  (convolution): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
                (1): CifNetConvLayer(
                  (convolution): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
                  (normalization): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
                  (activation): SiLU()
                )
              )
            )
          )
        )
      )
    )
    (pooler): AdaptiveAvgPool2d(output_size=(1, 1))
  )
  (classifier): Sequential(
    (0): Flatten(start_dim=1, end_dim=-1)
    (1): Linear(in_features=512, out_features=10, bias=True)
  )
)
----------------------------------------------------------------
        Layer (type)               Output Shape         Param #
================================================================
            Conv2d-1          [4, 64, 112, 112]           9,408
       BatchNorm2d-2          [4, 64, 112, 112]             128
              SiLU-3          [4, 64, 112, 112]               0
   CifNetConvLayer-4          [4, 64, 112, 112]               0
         MaxPool2d-5            [4, 64, 56, 56]               0
  CifNetEmbeddings-6            [4, 64, 56, 56]               0
            Conv2d-7            [4, 64, 56, 56]          36,864
       BatchNorm2d-8            [4, 64, 56, 56]             128
              SiLU-9            [4, 64, 56, 56]               0
  CifNetConvLayer-10            [4, 64, 56, 56]               0
           Conv2d-11            [4, 64, 56, 56]          36,864
      BatchNorm2d-12            [4, 64, 56, 56]             128
             SiLU-13            [4, 64, 56, 56]               0
  CifNetConvLayer-14            [4, 64, 56, 56]               0
         Identity-15            [4, 64, 56, 56]               0
 CifNetBasicLayer-16            [4, 64, 56, 56]               0
      CifNetStage-17            [4, 64, 56, 56]               0
           Conv2d-18           [4, 128, 28, 28]          73,728
      BatchNorm2d-19           [4, 128, 28, 28]             256
             SiLU-20           [4, 128, 28, 28]               0
  CifNetConvLayer-21           [4, 128, 28, 28]               0
           Conv2d-22           [4, 128, 28, 28]         147,456
      BatchNorm2d-23           [4, 128, 28, 28]             256
             SiLU-24           [4, 128, 28, 28]               0
  CifNetConvLayer-25           [4, 128, 28, 28]               0
           Conv2d-26           [4, 128, 28, 28]           8,192
      BatchNorm2d-27           [4, 128, 28, 28]             256
   CifNetShortCut-28           [4, 128, 28, 28]               0
 CifNetBasicLayer-29           [4, 128, 28, 28]               0
      CifNetStage-30           [4, 128, 28, 28]               0
           Conv2d-31           [4, 256, 14, 14]         294,912
      BatchNorm2d-32           [4, 256, 14, 14]             512
             SiLU-33           [4, 256, 14, 14]               0
  CifNetConvLayer-34           [4, 256, 14, 14]               0
           Conv2d-35           [4, 256, 14, 14]         589,824
      BatchNorm2d-36           [4, 256, 14, 14]             512
             SiLU-37           [4, 256, 14, 14]               0
  CifNetConvLayer-38           [4, 256, 14, 14]               0
           Conv2d-39           [4, 256, 14, 14]          32,768
      BatchNorm2d-40           [4, 256, 14, 14]             512
   CifNetShortCut-41           [4, 256, 14, 14]               0
 CifNetBasicLayer-42           [4, 256, 14, 14]               0
      CifNetStage-43           [4, 256, 14, 14]               0
           Conv2d-44             [4, 512, 7, 7]       1,179,648
      BatchNorm2d-45             [4, 512, 7, 7]           1,024
             SiLU-46             [4, 512, 7, 7]               0
  CifNetConvLayer-47             [4, 512, 7, 7]               0
           Conv2d-48             [4, 512, 7, 7]       2,359,296
      BatchNorm2d-49             [4, 512, 7, 7]           1,024
             SiLU-50             [4, 512, 7, 7]               0
  CifNetConvLayer-51             [4, 512, 7, 7]               0
           Conv2d-52             [4, 512, 7, 7]         131,072
      BatchNorm2d-53             [4, 512, 7, 7]           1,024
   CifNetShortCut-54             [4, 512, 7, 7]               0
 CifNetBasicLayer-55             [4, 512, 7, 7]               0
      CifNetStage-56             [4, 512, 7, 7]               0
    CifNetEncoder-57          [[-1, 512, 7, 7]]               0
AdaptiveAvgPool2d-58             [4, 512, 1, 1]               0
      CifNetModel-59  [[-1, 512, 7, 7], [-1, 512, 1, 1]]               0
          Flatten-60                   [4, 512]               0
           Linear-61                    [4, 10]           5,130
================================================================
Total params: 4,910,922
Trainable params: 4,910,922
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 2.30
Forward/backward pass size (MB): 345.14
Params size (MB): 18.73
Estimated Total Size (MB): 366.17
----------------------------------------------------------------