Skip to main content

Advertisement

Table 1 Configurations of the ResNet-50

From: Transfer deep convolutional activation-based features for domain adaptation in sensor networks

Config. ResNet-50 Padding
conv_1 \(\left [\begin {array}{ll}7\times 7, & 64\end {array}\right ]\times 1\), stride 2 3
  3×3, max pooling, stride 2 0
conv_2 \(\left [\begin {array}{ll}1\times 1, & 64 \\ 3\times 3, & 64 \\ 1\times 1, & 256\end {array}\right ]\times 3\), stride 2 \(\left [\begin {array}{ll} 0 \\ 1 \\ 0 \end {array}\right ]\)
conv_3 \(\left [\begin {array}{ll} 1\times 1, & 128 \\ 3\times 3, & 128 \\ 1\times 1, & 512 \end {array}\right ]\times 4\), stride 2 \(\left [\begin {array}{ll} 0 \\ 1 \\ 0 \end {array}\right ]\)
conv_4 \(\left [\begin {array}{ll} 1\times 1, & 256 \\ 3\times 3, & 256 \\ 1\times 1, & 1024 \end {array}\right ]\times 6\), stride 2 \(\left [\begin {array}{ll} 0 \\ 1 \\ 0 \end {array}\right ]\)
conv_5 \(\left [\begin {array}{ll} 1\times 1, & 512 \\ 3\times 3, & 512 \\ 1\times 1, & 2048 \end {array}\right ]\times 3\), stride 2 \(\left [\begin {array}{ll} 0 \\ 1 \\ 0 \end {array}\right ]\)
fc Average pooling, 1000-d, softmax
  1. The left part in “[ ]” indicates the size of receipt fields and the right part indicates the number of filter banks. Max pooling is implemented by a 3×3 pixel window. Both the convolution stride and max pooling stride are set to two pixels. The fully connected (FC) layer has 1000 channels