Normalization is the strategy to get your data with a certain range

Purpose of normalization boils down to 2 things

  1. Bias prevention, no single feature dominates
  2. Speeds up convergence in training