Good Seed Makes a Good Crop: Discovering Secret Seeds in Text-to-Image Diffusion Models

TL;DR

This paper systematically studies the impact of random seeds in text-to-image diffusion models, revealing that seed choice significantly affects image quality, composition, and interpretability, and demonstrating the utility of selecting optimal seeds for improved image synthesis.

Contribution

It provides the first large-scale analysis of seed influence in diffusion models, identifying characteristics of effective seeds and showing how seed selection enhances image quality and diversity.

Findings

Best seeds achieve lower FID scores (21.60) compared to worst seeds (31.97)

Classifier can predict seed number with over 99.9% accuracy from generated images

Certain seeds produce consistent visual traits like grayscale or prominent sky regions

Abstract

Recent advances in text-to-image (T2I) diffusion models have facilitated creative and photorealistic image synthesis. By varying the random seeds, we can generate many images for a fixed text prompt. Technically, the seed controls the initial noise and, in multi-step diffusion inference, the noise used for reparameterization at intermediate timesteps in the reverse diffusion process. However, the specific impact of the random seed on the generated images remains relatively unexplored. In this work, we conduct a large-scale scientific study into the impact of random seeds during diffusion inference. Remarkably, we reveal that the best 'golden' seed achieved an impressive FID of 21.60, compared to the worst 'inferior' seed's FID of 31.97. Additionally, a classifier can predict the seed number used to generate an image with over 99.9% accuracy in just a few epochs, establishing that seeds are highly distinguishable based on generated images. Encouraged by these findings, we examined the influence of seeds on interpretable visual dimensions. We find that certain seeds consistently produce grayscale images, prominent sky regions, or image borders. Seeds also affect image composition, including object location, size, and depth. Moreover, by leveraging these 'golden' seeds, we demonstrate improved image generation such as high-fidelity inference and diversified sampling. Our investigation extends to inpainting tasks, where we uncover some seeds that tend to insert unwanted text artifacts. Overall, our extensive analyses highlight the importance of selecting good seeds and offer practical utility for image generation.