LAION-5B

5.85 billion image–text pairs filtered from web crawls, used to train large text-to-image and multimodal models.

Papers using LAION-5B (2)

Fully Unsupervised Self-debiasing of Text-to-Image Diffusion Models2025Score Jacobian Chaining: Lifting Pretrained 2D Diffusion Models for 3D Generation2022 · 5 cites