omniglot.py activ = activin.mm( self.w + torch.mul(self.alpha, hebb)) + 1000.0 * inputlabel

Hi,

Everything else here is the same as described as what's in the paper except the part: 1000.0 * inputlabel. Could you explain why this part is here? 