arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2410.14895cs.LGcs.AIcs.CV
收藏

截断一致性模型

Truncated Consistency Models

Carnegie Mellon University(卡内基梅隆大学);NVIDIA(英伟达公司)

Sangyun Lee, Yilun Xu, Tomas Geffner, Giulia Fanti, Karsten Kreis, Arash Vahdat, Weili Nie

AI总结:

针对一致性模型训练中映射所有中间点导致一步生成性能受限的问题,提出截断时间范围的一致性训练,并采用新参数化和两阶段训练,在CIFAR-10和ImageNet上以更小网络取得更优FID。

摘要:

一致性模型最近被引入,通过直接从初始噪声预测概率流常微分方程(PF ODE)的解(即数据)来加速扩散模型的采样。然而,一致性模型的训练需要学习将PF ODE轨迹上的所有中间点映射到其对应的端点。这个任务比一步生成的最终目标更具挑战性,因为后者只关注PF ODE从噪声到数据的映射。我们通过实验发现,这种训练范式限制了一致性模型的一步生成性能。为了解决这个问题,我们将一致性训练推广到截断的时间范围,这使得模型能够忽略较早时间步的去噪任务,并将其能力集中在生成上。我们提出了一种新的一致性函数参数化方法和一个两阶段训练流程,以防止截断时间训练坍缩到平凡解。在CIFAR-10和ImageNet $64\ imes64$数据集上的实验表明,我们的方法在一步和两步FID指标上优于最先进的一致性模型(如iCT-deep),同时使用的网络规模小2倍以上。项目页面:https://truncated-cm.github.io/

英文摘要:

Consistency models have recently been introduced to accelerate sampling from diffusion models by directly predicting the solution (i.e., data) of the probability flow ODE (PF ODE) from initial noise. However, the training of consistency models requires learning to map all intermediate points along PF ODE trajectories to their corresponding endpoints. This task is much more challenging than the ultimate objective of one-step generation, which only concerns the PF ODE's noise-to-data mapping. We empirically find that this training paradigm limits the one-step generation performance of consistency models. To address this issue, we generalize consistency training to the truncated time range, which allows the model to ignore denoising tasks at earlier time steps and focus its capacity on generation. We propose a new parameterization of the consistency function and a two-stage training procedure that prevents the truncated-time training from collapsing to a trivial solution. Experiments on CIFAR-10 and ImageNet $64\times64$ datasets show that our method achieves better one-step and two-step FIDs than the state-of-the-art consistency models such as iCT-deep, using more than 2$\times$ smaller networks. Project page: https://truncated-cm.github.io/

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑