Some of the most common issues that arise while training Deep Neural Networks are:

The Vanishing$/$Exploding Gradient Problem

Training is always fast.

The net always under$-$fits the training data.

Training may bis slow.

Overfitting of the training set.

a$),$ d$)$ and e$)$ are true.

a$),$ c$)$ and d$)$ are true.

a$)$ and c$)$ are true.