justinhennessy/100-days-of-code

It also discussed a concept of overfitting, which essentially means, the model matches the training data exactly so losses its ability to "predict" or learning for things it hasn't seen.