Using Knowledge Graphs to harvest datasets for efficient CLIP model training

TL;DR

This paper presents a method using knowledge graphs and smart web search strategies to efficiently harvest datasets, enabling training of high-quality CLIP models with substantially less data and time, especially for domain-specific applications.

Contribution

The authors introduce a novel approach combining knowledge graphs with web search to create datasets for training CLIP models more efficiently, including the new EntityNet dataset.

Findings

A CLIP model for living organisms trained with only 10M images.

EntityNet dataset contains 33M images and 46M text descriptions.

Efficient training reduces data and time requirements significantly.

Abstract

Training high-quality CLIP models typically requires enormous datasets, which limits the development of domain-specific models -- especially in areas that even the largest CLIP models do not cover well -- and drives up training costs. This poses challenges for scientific research that needs fine-grained control over the training procedure of CLIP models. In this work, we show that by employing smart web search strategies enhanced with knowledge graphs, a robust CLIP model can be trained from scratch with considerably less data. Specifically, we demonstrate that an expert foundation model for living organisms can be built using just 10M images. Moreover, we introduce EntityNet, a dataset comprising 33M images paired with 46M text descriptions, which enables the training of a generic CLIP model in significantly reduced time.