A easy-use classification platform for quick verification and iteration. The current model was designed for simple image classification which was stacked by S-Y modules.
Basically the S-Y module consists of convolution layer which have several addition options including batch normalization, bias, scale factor and drop out. So we can construct a CNN architecture in one line.