Training deep learning models requires data… A lot of data! Unfortunately, in most cases data comes messy, and our models are very sensitive towards this. Therefore, we need to be careful while preparing our data to achieve the best results.