SpectralAR: Spectral Autoregressive Visual Generation

TL;DR

SpectralAR introduces a spectral-based autoregressive framework for visual generation that models images from low to high frequency components, achieving causality and efficiency in sequence modeling.

Contribution

It proposes a novel spectral tokenization method and autoregressive generation process that better captures causality in images compared to patch-based methods.

Findings

Achieves 3.02 gFID on ImageNet-1K with 64 tokens

Uses only 310M parameters for high-quality generation

Demonstrates effective coarse-to-fine spectral image reconstruction

Abstract

Autoregressive visual generation has garnered increasing attention due to its scalability and compatibility with other modalities compared with diffusion models. Most existing methods construct visual sequences as spatial patches for autoregressive generation. However, image patches are inherently parallel, contradicting the causal nature of autoregressive modeling. To address this, we propose a Spectral AutoRegressive (SpectralAR) visual generation framework, which realizes causality for visual sequences from the spectral perspective. Specifically, we first transform an image into ordered spectral tokens with Nested Spectral Tokenization, representing lower to higher frequency components. We then perform autoregressive generation in a coarse-to-fine manner with the sequences of spectral tokens. By considering different levels of detail in images, our SpectralAR achieves both sequence causality and token efficiency without bells and whistles. We conduct extensive experiments on ImageNet-1K for image reconstruction and autoregressive generation, and SpectralAR achieves 3.02 gFID with only 64 tokens and 310M parameters. Project page: https://huang-yh.github.io/spectralar/.