让扩散模型自动学习去噪时机，提升图像生成速度与质量
Learning When to Denoise: Optimizing Asynchronous Schedules for Latent Diffusion
- 通过可学习的异步去噪调度，优化多视角图像表示的去噪顺序
- 200轮训练达FID 1.05，仅需原模型4分之一训练量即超越大模型表现
- 适合追求高效训练与高质量生成的图像生成研究者
多表示扩散模型可通过去噪图像的互补视图提升视觉合成效果，但其性能高度依赖于各表示去噪时机的异步调度。本文提出学习该调度的方法，将多空间流匹配形式化，并采用调度修正目标，确保调度变化时局部加噪时间权重保持不变。我们设计了一类构造上凸且单调的灵活参数化调度，通过快速联合探针实现学习，额外计算开销低于1%。在ImageNet 256x256上，所学调度显著提升收敛速度与最终质量。使用800轮训练的基线模型（SFD-XL）FID为1.04，而本方法仅200轮即达FID 1.05，训练量仅为四分之一；训练至600轮进一步降至FID 1.02，优于10亿参数的SFD-XXL模型（FID 1.04），且使用更小模型。在无引导设置下，200轮模型达FID 2.37，已优于800轮基线（FID 2.54），600轮时进一步降至FID 2.14。代码已开源。
原文摘要 · Abstract (English)
Multi-representation diffusion models can improve visual synthesis by denoising complementary views of an image, but their performance depends critically on the asynchronous schedule that determines when each representation is denoised. We propose to learn this schedule. Our method formulates asynchronous flow matching over multiple representation spaces and uses a schedule-corrected objective that keeps each representation's local noising-time weights fixed as the schedule changes. We instantiate the schedule with a flexible parametric class that is convex and monotone by construction, and learn it using a fast joint probe with less than 1% additional training compute. On ImageNet 256x256, the learned schedule substantially improves both convergence speed and final quality under a matched 675M-parameter XL backbone. With AutoGuidance, our 200-epoch model reaches FID 1.05, matching the 800-epoch SFD-XL baseline with 4x less training. Training to 600 epochs further improves to FID 1.02, outperforming the 1B-parameter SFD-XXL result of FID 1.04 while using a smaller model. In the unguided setting, our 200-epoch model reaches FID 2.37, already below the best 800-epoch SFD-XL result (2.54) at 4x less training, and improves to FID 2.14 at 600 epochs. Code is available at https://github.com/bsq532087/LWD
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。