AI Glossary Term

What does “CLIP” mean?

Definition

A multimodal model architecture that learns shared representations between text and images.

Browse the full AI Glossary Explore all free AI guides

How CLIP works

Example

CLIP was trained on pairs of images and captions so that a photo of a dog and the words 'a photo of a dog' land close together in a shared space. You can then classify a new photo by comparing it to text like 'a photo of a cat' versus 'a photo of a dog', without training a new classifier.

Why it matters

CLIP showed that learning from images paired with natural language could produce flexible vision systems. Its ideas underlie many image search and image-generation tools.

Common questions

Who created CLIP?

OpenAI introduced CLIP (Contrastive Language-Image Pre-training) in 2021.

What is CLIP used for?

Image search, zero-shot image classification, and guiding image-generation systems toward text prompts.

Does CLIP understand images like a person?

No. It learns statistical associations between images and text, and it can be fooled or reflect biases in its training data.

Diffusion ModelA generative architecture that learns to reverse noise to synthesize images, audio, or other content.Convolutional Neural Network (CNN)A neural architecture optimized for processing grid-like data such as images.EmbeddingA numeric vector representation that captures semantic meaning of text, images, or other data.Generative AIAI systems that produce new content such as text, images, audio, video, or code.Multimodal ModelA model that can process or generate multiple data types such as text, image, and audio.OCR (Optical Character Recognition)Technology that converts text in images or scans into machine-readable text.

Learn more in our free guides

Gradient Clipping CLIP and Vision-Language Models CLIP Score and Human Preference Metrics

Test yourself

Gradient Clipping Quiz CLIP and Vision-Language Models Quiz CLIP Score and Human Preference Metrics Quiz