Except for the speed, does it perform like the original caffe convolution?
jiangyangzhou opened this issue · 1 comments
jiangyangzhou commented
I know it can run faster than original conv, but I find all the params are available in original caffe convolution, so if I just replace DepthwiseConvolution with Convolution, do they have same performance?
Or if depthwiseConvolution does something else?
Thanks very much
yonghenglh6 commented
The param's meaning is same as original conv because mycode inherits the cpu part from it.
There is only one difference that depthwise-conv 'group' param must be set to input channel.