Early Stopping

Automatically stop training when the model stops improving.

Early stopping monitors validation loss during training and stops when improvement stalls. This prevents wasted computation and can prevent overfitting.

How it works

1. After each epoch, validation loss is calculated

2. If loss doesn't improve for 'patience' epochs, training stops

3. The best model (lowest validation loss) is saved

Patience

The number of epochs to wait for improvement before stopping. Higher patience allows for temporary plateaus but may waste time.

Recommendations

• Keep early stopping enabled (default)

• Patience of 10 works well for most cases

• Increase patience to 15-20 if training seems to stop too early