As stated in the paper, the proposed convnet-aig can reduce computational cost. But after checking the source codes, I found each layer must be executed. The way you skip a single layer is by multiplying a mask (0 or 1) to the output. If that is the case, which part of the convnet-aig can reduce the computation?
The code below is cited from convnet_aig.py, line 152, where out is the output of a layer, w[:,1] is the mask.
out = self.shortcut(x) + out * w[:,1].unsqueeze(1)
As stated in the paper, the proposed convnet-aig can reduce computational cost. But after checking the source codes, I found each layer must be executed. The way you skip a single layer is by multiplying a mask (0 or 1) to the output. If that is the case, which part of the convnet-aig can reduce the computation?
The code below is cited from convnet_aig.py, line 152, where
outis the output of a layer,w[:,1]is the mask.