Hi, thanks for your sharing,and I think your work is focus on efficiency Convolution Compute,Would you test one MoblieNet or ShuffleNet with your method?
Hi, thanks for the comment, RACNN works only for a standard 3x3 convolution and do not work on a depth-wise convolution which is used in MoblieNet or ShuffleNet
Hi, thanks for your sharing,and I think your work is focus on efficiency Convolution Compute,Would you test one MoblieNet or ShuffleNet with your method?