Jano: Adaptive Diffusion Generation with Early-stage Convergence Awareness

TL;DR

Jano is a training-free, adaptive diffusion framework that identifies regional convergence needs early in the denoising process, enabling significant acceleration while maintaining quality.

Contribution

Jano introduces a novel early-stage convergence recognition algorithm and adaptive token scheduling for efficient, region-aware diffusion generation without additional training.

Findings

Achieves up to 2.4x speedup in diffusion generation.

Maintains high quality of generated content.

Validates effectiveness across state-of-the-art models.

Abstract

Diffusion models have achieved remarkable success in generative AI, yet their computational efficiency remains a significant challenge, particularly for Diffusion Transformers (DiTs) requiring intensive full-attention computation. While existing acceleration approaches focus on content-agnostic uniform optimization strategies, we observe that different regions in generated content exhibit heterogeneous convergence patterns during the denoising process. We present Jano, a training-free framework that leverages this insight for efficient region-aware generation. Jano introduces an early-stage complexity recognition algorithm that accurately identifies regional convergence requirements within initial denoising steps, coupled with an adaptive token scheduling runtime that optimizes computational resource allocation. Through comprehensive evaluation on state-of-the-art models, Jano achieves substantial acceleration (average 2.0 times speedup, up to 2.4 times) while preserving generation quality. Our work challenges conventional uniform processing assumptions and provides a practical solution for accelerating large-scale content generation. The source code of our implementation is available at https://github.com/chen-yy20/Jano.