Beginner · Fundamentals

Iteration

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. One update step of the model's parameters, processing a single batch.

Technical Definition

One update step of the model's parameters, processing a single batch.

How it works

An iteration = one forward pass, one loss computation, one backward pass, one parameter update. Number of iterations per epoch = dataset size / batch size. Training progress is often plotted per iteration to inspect short-term loss dynamics.

Related Concepts

• Training — The process of adjusting a model's parameters so it learns patterns from labeled or unlabeled data.
• Epoch — One full pass through the entire training dataset during model training.
• Batch — A group of training examples processed together in a single forward and backward pass.