通过截断时间范围训练，让一致性模型更高效生成图像。
Truncated Consistency Models
- 在截断时间范围内训练，专注后期生成而非全程去噪。
- 在CIFAR-10和ImageNet上实现更低的FID，且网络更小。
- 适合追求快速生成且资源有限的图像合成应用。
一致性模型通过直接预测概率流ODE的解来加速扩散模型采样，但其训练需学习所有中间点到终点的映射，任务远超最终的一步生成目标。我们发现该训练方式限制了生成性能。为此，提出在截断时间范围内进行训练，使模型可忽略早期去噪任务，集中精力于生成阶段。设计新的函数参数化与两阶段训练流程，防止训练坍塌。在CIFAR-10和ImageNet $64 imes64$数据集上，该方法以超过2倍更小的网络规模，达到优于iCT-deep等最先进一致性模型的一步与两步生成效果。
原文摘要 · Abstract (English)
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/
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。