Exploding & Vanishing Gradient Problem in Deep Learning | Towards Data Science

How to ensure your neural network doesn’t “die” or “blow-up”