# What is Tokenizer training? Definition and meaning

Fitting a vocabulary from data so that inputs are represented efficiently.

## Definition

**Tokenizer training**: Fitting a vocabulary from data so that inputs are represented efficiently.

In short: Learning how raw input is split into tokens.

## When it is used

Preparing text or structured inputs for a language or foundation model.

## Learn more

The full guide to **[Tokenizer training](/tools/tokenizer-training)** 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: [Thales](/glossary/thales) · Next term: [Torchtune](/glossary/torchtune)
