Table 4.

Block structure of ResNet and ResNeXt (cardinality = 32). ResNeXt incorporates a group convolution with a group size of 32 in the repeatable residual block structure.

Stage	ResNet-50	ResNeXt-50(32 × 4d)
conv1	7 × 7, 64, stride 2	7 × 7, 64, stride 2
conv2	3 × 3 max pool, stride 2	3 × 3 max pool, stride 2
	$\left[\begin{array}{c}\begin{array}{cc}1\times 1,& 64\\ 3\times 3,& 64\end{array}\\ \begin{array}{cc}1\times 1,& 256\end{array}\end{array}\right]\times 3$	$\left[\begin{array}{c}1\times 1,128\\ 3\times 3,128\\ 1\times 1,256\end{array},C=32\right]\times 3$
conv3	$\left[\begin{array}{c}\begin{array}{cc}1\times 1,& 128\\ 3\times 3,& 128\end{array}\\ \begin{array}{cc}1\times 1,& 512\end{array}\end{array}\right]\times 4$	$\left[\begin{array}{c}\begin{array}{cc}1\times 1,& 256\\ 3\times 3,& 256\end{array}\\ \begin{array}{cc}1\times 1,& 512\end{array}\end{array},,,C,=,32\right]\times 4$
Conv4	$\left[\begin{array}{c}\begin{array}{cc}1\times 1,& 256\\ 3\times 3,& 256\end{array}\\ \begin{array}{cc}1\times 1,& 1024\end{array}\end{array}\right]\times 6$	$\left[\begin{array}{c}\begin{array}{cc}1\times 1,& 512\\ 3\times 3,& 512\end{array}\\ \begin{array}{cc}1\times 1,& 1024\end{array}\end{array},,,C,=,32\right]\times 6$
Conv5	$\left[\begin{array}{c}\begin{array}{cc}1\times 1,& 512\\ 3\times 3,& 512\end{array}\\ \begin{array}{cc}1\times 1,& 2048\end{array}\end{array}\right]\times 3$	$\left[\begin{array}{c}\begin{array}{cc}1\times 1,& 1024\\ 3\times 3,& 1024\end{array}\\ \begin{array}{cc}1\times 1,& 2048\end{array}\end{array},,,C,=,32\right]\times 6$