onrup

Data

Alpaca format

Also called Instruct format.

What is alpaca format?

The Alpaca format stores each example as an instruction, an optional input, and an output. It suits single-turn task data and is simpler than conversation formats, at the cost of being unable to represent multi-turn dialogue.

For classification, extraction and transformation tasks, single-turn structure is a good fit and the simplicity is a virtue: fewer fields means fewer ways to build the dataset wrongly.

It converts cleanly into conversation formats — instruction becomes a user message, output becomes an assistant message — which is why pipelines accept it as a source format even when they train on something else.

Related terms

All terms in the glossary →

Start with the free tier

A magic link creates your account, your tenant and your first API key. No card until you ask for compute.