Also: nets too small (not enough layers) gradients not flowing (residual connections) layer outputs not normalized training algorithms and procedures not op

Also: nets too small (not enough layers) gradients not flowing (residual connections) layer outputs not normalized training algorithms and procedures not optimal (Adam, warm-up, etc)