These are chat archives for arita37/tensorflow

17th
Jun 2017
Syed Ather Rizvi
@Ather23
Jun 17 2017 19:42
@akki6843 The batch size will help your model converge faster as you are feeding in less data per epoch. My hunch is that your model will converge quickly to a high loss value
Syed Ather Rizvi
@Ather23
Jun 17 2017 20:19
@segevmalool you can create a diagonal matrix using numpy and convert it to a tensor object.