Kaleido Diffusion: Improving Conditional Diffusion Models with Autoregressive Latent Modeling

TL;DR

Kaleido Diffusion introduces autoregressive latent priors to enhance diversity in conditional diffusion model outputs, effectively controlling image generation while maintaining high quality.

Contribution

The paper proposes Kaleido, a novel method integrating autoregressive latent modeling to improve diversity and control in diffusion-based image generation from text.

Findings

Increased diversity of generated images with high quality.

Effective control of image outputs guided by latent variables.

Compatibility with various discrete latent representations.

Abstract

Diffusion models have emerged as a powerful tool for generating high-quality images from textual descriptions. Despite their successes, these models often exhibit limited diversity in the sampled images, particularly when sampling with a high classifier-free guidance weight. To address this issue, we present Kaleido, a novel approach that enhances the diversity of samples by incorporating autoregressive latent priors. Kaleido integrates an autoregressive language model that encodes the original caption and generates latent variables, serving as abstract and intermediary representations for guiding and facilitating the image generation process. In this paper, we explore a variety of discrete latent representations, including textual descriptions, detection bounding boxes, object blobs, and visual tokens. These representations diversify and enrich the input conditions to the diffusion models, enabling more diverse outputs. Our experimental results demonstrate that Kaleido effectively broadens the diversity of the generated image samples from a given textual description while maintaining high image quality. Furthermore, we show that Kaleido adheres closely to the guidance provided by the generated latent variables, demonstrating its capability to effectively control and direct the image generation process.