AI Foundations
Epoch
One complete pass through the training dataset.
Example
The model reviews the entire training collection five times, completing five epochs.
Why it matters
It gives teams a shared way to describe how much practice the system has had.
You'll hear it
“It has gone through the teaching examples three times.”
Can you control it?
Set by the team
The training team sets a maximum number of epochs or uses stopping rules. People using the finished model do not control this.
Common questions
- Do equal epoch counts mean equal training time?
- No. Larger collections or slower systems can take much longer to complete the same number of passes.
- Are more epochs always better?
- No. Extra passes can waste time or make the system memorize details that do not help later.
- How many epochs does a model need?
- There is no fixed number. Useful progress and independent checks should guide the decision.