将迭代去噪引入流模型，提升图像生成质量与稳定性。
Normalizing Flows with Iterative Denoising
- 结合自回归生成与迭代去噪，保持端到端似然训练。
- 在64/128/256像素下性能媲美扩散模型。
- 适合关注高效生成与可解释性的研究者。
Normalizing Flows（NFs）是一类基于似然的经典生成方法，近年来受到广泛关注。TARFlow等近期工作表明其在图像建模任务中具备优异表现，可作为扩散模型的有力替代。本文提出迭代TARFlow（iTARFlow），通过在采样阶段引入受扩散模型启发的迭代去噪过程，进一步提升流模型性能。训练阶段保持全端到端、基于似然的目标，采样时先进行自回归生成，再执行多轮去噪。大量实验表明，iTARFlow在ImageNet 64、128、256像素分辨率上均达到竞争力水平，展现出强大的生成能力，并推动了流模型的发展边界。此外，我们分析了模型产生的典型伪影，为未来改进提供参考。代码已开源：https://github.com/apple/ml-itarflow。
原文摘要 · Abstract (English)
Normalizing Flows (NFs) are a classical family of likelihood-based methods that have received revived attention. Recent efforts such as TARFlow have shown that NFs are capable of achieving promising performance on image modeling tasks, making them viable alternatives to other methods such as diffusion models. In this work, we further advance the state of Normalizing Flow generative models by introducing iterative TARFlow (iTARFlow). Unlike diffusion models, iTARFlow maintains a fully end-to-end, likelihood-based objective during training. During sampling, it performs autoregressive generation followed by an iterative denoising procedure inspired by diffusion-style methods. Through extensive experiments, we show that iTARFlow achieves competitive performance across ImageNet resolutions of 64, 128, and 256 pixels, demonstrating its potential as a strong generative model and advancing the frontier of Normalizing Flows. In addition, we analyze the characteristic artifacts produced by iTARFlow, offering insights that may shed light on future improvements. Code is available at https://github.com/apple/ml-itarflow.
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。