Annealed Relaxation of Speculative Decoding for Faster Autoregressive Image Generation

TL;DR

This paper introduces COOL-SD, an annealed relaxation method for speculative decoding in autoregressive image generation, providing a theoretical foundation and demonstrating faster image synthesis with maintained or improved quality.

Contribution

It offers the first theoretical analysis of relaxed speculative decoding and proposes COOL-SD, an annealed approach that improves speed and quality in autoregressive image generation.

Findings

COOL-SD achieves faster image generation compared to prior methods.

It maintains or improves image quality at similar or reduced latency.

Experimental results show consistent speed-quality trade-off improvements.

Abstract

Despite significant progress in autoregressive image generation, inference remains slow due to the sequential nature of AR models and the ambiguity of image tokens, even when using speculative decoding. Recent works attempt to address this with relaxed speculative decoding but lack theoretical grounding. In this paper, we establish the theoretical basis of relaxed SD and propose COOL-SD, an annealed relaxation of speculative decoding built on two key insights. The first analyzes the total variation (TV) distance between the target model and relaxed speculative decoding and yields an optimal resampling distribution that minimizes an upper bound of the distance. The second uses perturbation analysis to reveal an annealing behaviour in relaxed speculative decoding, motivating our annealed design. Together, these insights enable COOL-SD to generate images faster with comparable quality, or achieve better quality at similar latency. Experiments validate the effectiveness of COOL-SD, showing consistent improvements over prior methods in speed-quality trade-offs.