Plotting the Learning Curve to Analyze the Training Performance of a Neural Network

To detect overfitting and underfitting as well as slow convergence, oscillating, oscillating with divergence

Photo by Maarten Deckers on Unsplash

I’m going to show you an easy method to plot the learning curve with less code in Keras!

Before that, I want to explain what the learning curve is and the objectives of plotting it.

What is the learning curve?

The learning curve can be considered as an alternative for training verbose output that we often see during the training process of a neural network.

The learning curve is created by plotting training and validation errors (losses) or accuracies against the number of epochs.

Objectives of plotting the learning curve

The main objective of plotting the learning curve is to detect underfitting, overfitting and just-right conditions.

• Underfitting: Undefitting happens when the model is too simple. The model will fail to learn essential patterns in the training data. So, it performs poorly on both training and validation…