On the Scalability of Diffusion-based Text-to-Image Generation

TL;DR

This paper empirically investigates how to effectively scale diffusion-based text-to-image models, analyzing model architecture, training data, and providing scaling laws for performance optimization.

Contribution

It offers a comprehensive analysis of scaling laws for diffusion T2I models, proposing an efficient UNet variant and insights on data quality versus quantity.

Findings

Transformer depth improves alignment more efficiently than width.

An efficient UNet variant is 45% smaller and 28% faster.

Data diversity enhances model performance more than dataset size.

Abstract

Scaling up model and data size has been quite successful for the evolution of LLMs. However, the scaling law for the diffusion based text-to-image (T2I) models is not fully explored. It is also unclear how to efficiently scale the model for better performance at reduced cost. The different training settings and expensive training cost make a fair model comparison extremely difficult. In this work, we empirically study the scaling properties of diffusion based T2I models by performing extensive and rigours ablations on scaling both denoising backbones and training set, including training scaled UNet and Transformer variants ranging from 0.4B to 4B parameters on datasets upto 600M images. For model scaling, we find the location and amount of cross attention distinguishes the performance of existing UNet designs. And increasing the transformer blocks is more parameter-efficient for improving text-image alignment than increasing channel numbers. We then identify an efficient UNet variant, which is 45% smaller and 28% faster than SDXL's UNet. On the data scaling side, we show the quality and diversity of the training set matters more than simply dataset size. Increasing caption density and diversity improves text-image alignment performance and the learning efficiency. Finally, we provide scaling functions to predict the text-image alignment performance as functions of the scale of model size, compute and dataset size.