StableRep: Synthetic Images from Text-to-Image Models Make Strong Visual Representation Learners
This addresses the problem of reducing reliance on large-scale real image datasets for visual representation learning, offering a potentially more scalable and efficient approach, though it builds incrementally on existing contrastive learning methods.
The paper tackles learning visual representations from synthetic images generated by text-to-image models, showing that self-supervised training on these images can match or exceed real image performance, and their StableRep method with multi-positive contrastive learning surpasses SimCLR and CLIP using the same prompts and real images, achieving better accuracy with 20M synthetic images than CLIP with 50M real images.
We investigate the potential of learning visual representations using synthetic images generated by text-to-image models. This is a natural question in the light of the excellent performance of such models in generating high-quality images. We consider specifically the Stable Diffusion, one of the leading open source text-to-image models. We show that (1) when the generative model is configured with proper classifier-free guidance scale, training self-supervised methods on synthetic images can match or beat the real image counterpart; (2) by treating the multiple images generated from the same text prompt as positives for each other, we develop a multi-positive contrastive learning method, which we call StableRep. With solely synthetic images, the representations learned by StableRep surpass the performance of representations learned by SimCLR and CLIP using the same set of text prompts and corresponding real images, on large scale datasets. When we further add language supervision, StableRep trained with 20M synthetic images achieves better accuracy than CLIP trained with 50M real images.