提出漂移模型，实现单步生成且质量达顶尖水平。
Generative Modeling via Drifting
- 训练时动态调整样本分布，实现单步推理
- 在256×256图像上FID达1.54（隐空间）和1.61（像素空间）
- 适合追求高效高质生成的场景
生成建模可被表述为学习一个映射函数f，使其前推分布与数据分布一致。该前推过程可在推理时迭代完成，如扩散模型和基于流的方法。本文提出一种新范式——漂移模型（Drifting Models），在训练过程中演化前推分布，并自然支持单步推理。引入漂移场控制样本运动，当分布匹配时达到平衡。这带来一种训练目标，使神经网络优化器可直接演化分布。实验表明，我们的单步生成器在ImageNet 256×256分辨率上达到当前最优表现，隐空间FID为1.54，像素空间FID为1.61。我们希望此工作开启高质量单步生成的新可能。
原文摘要 · Abstract (English)
Generative modeling can be formulated as learning a mapping f such that its pushforward distribution matches the data distribution. The pushforward behavior can be carried out iteratively at inference time, for example in diffusion and flow-based models. In this paper, we propose a new paradigm called Drifting Models, which evolve the pushforward distribution during training and naturally admit one-step inference. We introduce a drifting field that governs the sample movement and achieves equilibrium when the distributions match. This leads to a training objective that allows the neural network optimizer to evolve the distribution. In experiments, our one-step generator achieves state-of-the-art results on ImageNet at 256 x 256 resolution, with an FID of 1.54 in latent space and 1.61 in pixel space. We hope that our work opens up new opportunities for high-quality one-step generation.
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。