Learning Transferable Visual Models From Natural Language Supervision

by   Alec Radford, et al.

State-of-the-art computer vision systems are trained to predict a fixed set of predetermined object categories. This restricted form of supervision limits their generality and usability since additional labeled data is needed to specify any other visual concept. Learning directly from raw text about images is a promising alternative which leverages a much broader source of supervision. We demonstrate that the simple pre-training task of predicting which caption goes with which image is an efficient and scalable way to learn SOTA image representations from scratch on a dataset of 400 million (image, text) pairs collected from the internet. After pre-training, natural language is used to reference learned visual concepts (or describe new ones) enabling zero-shot transfer of the model to downstream tasks. We study the performance of this approach by benchmarking on over 30 different existing computer vision datasets, spanning tasks such as OCR, action recognition in videos, geo-localization, and many types of fine-grained object classification. The model transfers non-trivially to most tasks and is often competitive with a fully supervised baseline without the need for any dataset specific training. For instance, we match the accuracy of the original ResNet-50 on ImageNet zero-shot without needing to use any of the 1.28 million training examples it was trained on. We release our code and pre-trained model weights at https://github.com/OpenAI/CLIP.


page 13

page 15

page 24

page 42


K-LITE: Learning Transferable Visual Models with External Knowledge

Recent state-of-the-art computer vision systems are trained from natural...

Learning to Prompt for Vision-Language Models

Vision-language pre-training has recently emerged as a promising alterna...

OmDet: Language-Aware Object Detection with Large-scale Vision-Language Multi-dataset Pre-training

Advancing object detection to open-vocabulary and few-shot transfer has ...

EVA: Exploring the Limits of Masked Visual Representation Learning at Scale

We launch EVA, a vision-centric foundation model to explore the limits o...

Exploring CLIP for Assessing the Look and Feel of Images

Measuring the perception of visual content is a long-standing problem in...

An implementation of the "Guess who?" game using CLIP

CLIP (Contrastive Language-Image Pretraining) is an efficient method for...

CLIP goes 3D: Leveraging Prompt Tuning for Language Grounded 3D Recognition

Vision-Language models like CLIP have been widely adopted for various ta...

Please sign up or login with your details

Forgot password? Click here to reset