Loss turns into 'nan'

[barisozmen]

I'm experimenting with wrn-16-8 (WideResNet) at this repo. During training, loss suddenly turned into nan. I guess it's a numerical calculation problem.

[barisozmen]

Setting clipnorm to 1. was suggested at https://stackoverflow.com/questions/37232782/nan-loss-when-training-regression-network

It didn't worked..

[lichuanx]

Might some issues from data.