3D卷积实现keras group conv(Interleaved Group Convolutions for Deep Neural Networks)
x0=Input((10,10,64))x=Reshape((10,10,2,32),input_shape=(10,10,64))(x0)交换位置拿到上下两层x=Permute((1,2,4,3))(x)x1=Lambda(lambdaz:K.expand_dims(z[:,:,:,:,0],axis=-1))(x)x2=Lambda(lambdaz:K.expand_dims(z[:,:,:,