Neural Networks

Epoch

One complete pass through the entire training dataset during model training.

In-depth explanation

During training, the model typically sees the data multiple times. Each complete iteration through all training examples is one epoch. Training usually requires many epochs for the model to converge. Too few epochs lead to underfitting; too many can cause overfitting. Monitoring validation loss helps determine when to stop training.

Examples

Training for 100 epochs
Early stopping at epoch 50

Related terms

Master Epoch.

Learn how to apply this concept with hands-on projects in our comprehensive AI programs.