We trained the convolutional networks by minimizing mean square error for 160 iterations, after which no further improvement on both the training and testing sets is observed
We trained the convolutional networks by minimizing meansquare error for 160 iterations, after which no further improvementon both the training and testing sets is observed