Useful Plots to Diagnose your Neural Network

As every single network training begins, let's start with the data. Data might give some explanation as to why you're having problems while training. Your data might give insights into why your model isn't behaving as it is expected to. After training a classification model, there might be situations where the output of the model completely or mostly belongs to one class ie, the situation where the model is biased. This is primarily due to an imbalanced dataset.