# CLIP

A model trained contrastively on image and caption pairs that embeds both into one space.

## What it is used for

Zero-shot image classification and cross-modal search.