Early Stopping Patience Calculator

Work out early stopping patience instantly with clear inputs, formula shown and shareable results.

About Early Stopping Patience Calculator

Early stopping halts training when the validation metric has not improved by more than a minimum delta for a set number of epochs. Patience must be long enough to survive the plateaus that precede a learning-rate drop, but every epoch after the best checkpoint is compute spent on a model you will discard — which is why patience is tuned against epoch cost, not chosen arbitrarily.

Formula

Early stopping

epochs remaining = patience - epochs since improvement; wasted compute = epochs since improvement x epoch time

Worked example

With the default values below, Early Stopping Patience Calculator returns Epochs remaining before stopping of 3. Change any input above and every figure here is recalculated in your browser.

Inputs used

Maximum epochs configuredminimum 1: 100
Epochs since the last improvementminimum 0: 7
Patience settingminimum 1: 10
Minutes per epochminimum 0.1 min: 12 min

Results

Epochs remaining before stopping: 3
Compute time remaining: 36.0 min
Epochs since the best checkpoint: 7
Recommendation: Continue for up to 3 more epochs out of 100

Step by step

1Values used
2Early stopping
3Epochs remaining before stopping= 3
4Compute time remaining= 36.0 min
5Epochs since the best checkpoint= 7
6Recommendation= Continue for up to 3 more epochs out of 100

Assumptions and limits

The result is derived only from the 4 inputs listed above. Anything not among them is not part of the calculation.
In particular, this calculator does not take inflation, tax, fees or charges or compounding frequency into account.
Figures are rounded for display to 1 decimal place, so totals built from several rounded parts can differ slightly from the unrounded arithmetic.

How these calculators are built and tested is described on the Methodology page. If a figure looks wrong, please report it.

Frequently Asked Questions

How large should patience be?

Longer than the interval between learning-rate reductions, since each drop often produces a fresh improvement. With a step schedule every 30 epochs, patience of 10 will stop too early.

Which checkpoint should be kept?

The one with the best validation metric, not the final epoch. Set restore_best_weights or equivalent, otherwise you ship the worse model.