How to improve deep learning performance with using normalization and standarization?

Question

I have 8160 dataset, input\'s shape is (1500,1), output\'s shape is (2500,1). input\'s values are between -80 to 3. Outpur values vary between 0 or 1.

I thought that,