09_june_2017

[@gitabcworld] I don't understand the generalization for non-square kernels. Is it just a tiling operation?

[@priba] I don't understand the definition of the Dirac weight tensor defined in (2) why is 1 in l_n=[K_m/2]? Also, can you explain the loss function? why do you want W to have norm 0? and what is L_orig?