Deep Learning Dictionary - Lightweight Crash Course

Underfitting - Deep Learning Dictionary

A model is said to be underfitting when it's unable to perform well on the data it was trained on, let alone data it hasn't seen before. This is in contrast to we overfitting, which occurs when a trained performs well on the training data, but fails to generalize on data outside of the training set.

We can identify underfitting when the training metrics are poor, meaning that the training accuracy and/or the training loss of the model do not reach acceptable levels.