Depthwise and Separable convolutions in Pytorch?

2

for Depthwise / Separable, you can use Conv’s groups parameter.

http://pytorch.org/docs/master/nn.html#conv2d

If groups = nInputPlane, then it is Depthwise. If groups = nInputPlane, kernel=(K, 1), (and before is a Conv2d layer with groups=1 and kernel=(1, K)), then it is separable.

24 Likes