# Reusing Computation in Text-to-Image Diffusion for Efficient Generation of Image Sets

## TL;DR

This paper introduces a training-free method to reduce computational redundancy in text-to-image diffusion models by clustering prompts and sharing early diffusion computations, leading to more efficient image generation.

## Contribution

It presents a novel approach that leverages prompt clustering and diffusion model properties to improve efficiency without additional training, enhancing existing text-to-image pipelines.

## Key findings

- Significantly reduces compute cost for models conditioned on image embeddings.
- Improves image quality while decreasing computational resources.
- Scales effectively with larger prompt sets and integrates seamlessly with existing systems.

## Abstract

Text-to-image diffusion models enable high-quality image generation but are computationally expensive. While prior work optimizes per-inference efficiency, we explore an orthogonal approach: reducing redundancy across correlated prompts. Our method leverages the coarse-to-fine nature of diffusion models, where early denoising steps capture shared structures among similar prompts. We propose a training-free approach that clusters prompts based on semantic similarity and shares computation in early diffusion steps. Experiments show that for models trained conditioned on image embeddings, our approach significantly reduces compute cost while improving image quality. By leveraging UnClip's text-to-image prior, we enhance diffusion step allocation for greater efficiency. Our method seamlessly integrates with existing pipelines, scales with prompt sets, and reduces the environmental and financial burden of large-scale text-to-image generation. Project page: https://ddecatur.github.io/hierarchical-diffusion/