BTP-Experiments/BlockCCNN.yaml at master · braceletboy/BTP-Experiments · GitHub

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
##
# @file: BMCCNN.yaml
#
# This file contains the BlockCCNN networks.
#
# @author: Rukmangadh Sai Myana
# @mail: rukman.sai@gmail.com
#


BlockCCNN[0.0]:
  - layer_type: 'conv2d'  # (N, 1, 28, 28) - Input Shape ('N' is batch_size)
    layer_name: 'bnn_conv2d_1'
    in_channels: 1
    out_channels: 8
    kernel_size: 3
    stride: 1
    padding: 1
  - layer_type: 'relu'  # (N, 8, 28, 28)
    layer_name: 'bnn_relu_1'
  - layer_type: 'batch_norm2d'  # (N, 8, 28, 28)
    layer_name: 'bnn_batch_norm2d_1'
    num_features: 8
  - layer_type: 'conv2d'  # (N, 8, 28, 28)
    layer_name: 'bnn_conv2d_2'
    in_channels: 8
    out_channels: 8
    kernel_size: 3
    stride: 1
    padding: 1
  - layer_type: 'relu'  # (N, 8, 28, 28)
    layer_name: 'bnn_relu_2'
  - layer_type: 'batch_norm2d'  # (N, 8, 28, 28)
    layer_name: 'bnn_batch_norm2d_2'
    num_features: 8
  - layer_type: 'conv2d'  # (N, 8, 28, 28)
    layer_name: 'bnn_conv2d_3'
    in_channels: 8
    out_channels: 8
    kernel_size: 5
    stride: 2
    padding: 2
  - layer_type: 'relu'  # (N, 8, 14, 14)
    layer_name: 'bnn_relu_3'
  - layer_type: 'batch_norm2d'  # (N, 8, 14, 14)
    layer_name: 'bnn_batch_norm2d_3'
    num_features: 8
  - layer_type: 'dropout2d'  # (N, 8, 14, 14)
    layer_name: 'bnn_dropout2d_1'
    p: 0.4
  - layer_type: 'conv2d'  # (N, 8, 14, 14)
    layer_name: 'bnn_conv2d_4'
    in_channels: 8
    out_channels: 16
    kernel_size: 3
    stride: 1
    padding: 1
  - layer_type: 'relu'  # (N, 16, 14, 14)
    layer_name: 'bnn_relu_4'
  - layer_type: 'batch_norm2d'  # (N, 16, 14, 14)
    layer_name: 'bnn_batch_norm2d_4'
    num_features: 16
  - layer_type: 'conv2d'  # (N, 16, 14, 14)
    layer_name: 'bnn_conv2d_5'
    in_channels: 16
    out_channels: 16
    kernel_size: 3
    stride: 1
    padding: 1
  - layer_type: 'relu'  # (N, 16, 14, 14)
    layer_name: 'bnn_relu_5'
  - layer_type: 'batch_norm2d'  # (N, 16, 14, 14)
    layer_name: 'bnn_batch_norm2d_5'
    num_features: 16
  - layer_type: 'conv2d'  # (N, 16, 14, 14)
    layer_name: 'bnn_conv2d_6'
    in_channels: 16
    out_channels: 16
    kernel_size: 5
    stride: 2
    padding: 2
  - layer_type: 'relu'  # (N, 16, 7, 7)
    layer_name: 'bnn_relu_6'
  - layer_type: 'batch_norm2d'  # (N, 16, 7, 7)
    layer_name: 'bnn_batch_norm2d_6'
    num_features: 16
  - layer_type: 'dropout2d'  # (N, 16, 7, 7)
    layer_name: 'bnn_dropout2d_2'
    p: 0.4
  - layer_type: 'flatten'  # (N, 16, 7, 7)
    layer_name: 'bnn_flatten_1'
  - layer_type: 'linear'  # (N, 16*7*7)
    layer_name: 'bnn_linear_1'
    in_features: 784
    out_features: 32
  - layer_type: 'relu'  # (N, 32)
    layer_name: 'bnn_relu_7'
  - layer_type: 'batch_norm1d'  # (N, 32)
    layer_name: 'bnn_batch_norm1d_1'
    num_features: 32
  - layer_type: 'dropout'  # (N, 32)
    layer_name: 'bnn_dropout_1'
    p: 0.4
  - layer_type: 'linear'  # (N, 32)
    layer_name: 'bnn_linear_2'
    in_features: 32
    out_features: 5  ## Output is (N, 5) shaped