# What is SimCLR? Definition and meaning

A self-supervised framework that maximizes agreement between two augmented views of the same image using a contrastive loss.

## Definition

**SimCLR**: A self-supervised framework that maximizes agreement between two augmented views of the same image using a contrastive loss.

In short: Contrastive learning of image representations.

## When it is used

Pre-training visual encoders without labels.

## Learn more

The full guide to **[SimCLR](/tools/simclr)** explains how it works, when to use it and how it compares with the alternatives, with examples. The [glossary](/glossary) lists every term on the site from A to Z.

Previous term: [Silver layer](/glossary/silver-layer) · Next term: [SimPO](/glossary/simpo)
