通过识别生成早期的收敛差异,动态分配计算资源,实现扩散模型加速。
Jano: Adaptive Diffusion Generation with Early-stage Convergence Awareness
- 基于早期去噪阶段的区域收敛特性,智能识别不同区域计算需求。
- 平均提速2.0倍,最高达2.4倍,生成质量保持不变。
- 无需训练,适配主流扩散模型,适合大规模内容生成场景。
扩散模型在生成式AI中取得显著成功,但其计算效率仍是重大挑战,尤其对需要密集全注意力计算的扩散Transformer(DiTs)。现有加速方法多采用与内容无关的均匀优化策略,我们观察到生成内容中不同区域在去噪过程中呈现异构收敛模式。本文提出Jano,一个无需训练的框架,利用该洞察实现区域感知的高效生成。Jano引入一种早期阶段复杂度识别算法,可在初始去噪步骤中准确判断各区域的收敛需求,并结合自适应令牌调度运行时,优化计算资源分配。在先进模型上的全面评估表明,Jano实现了显著加速(平均2.0倍,最高2.4倍),同时保持生成质量。本工作挑战了传统统一处理假设,为大规模内容生成提供实用加速方案。代码已开源:https://github.com/chen-yy20/Jano。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。