arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.17487cs.CV
收藏

为扩散模型学习量化自适应条件

Learning Quantized Adaptive Conditions for Diffusion Models

School of Mathematical Sciences, Peking University(北京大学数学科学学院)

Yuchen Liang, Yuchuan Tian, Lei Yu, Huao Tang, Jie Hu, Xiangzhong Fang, Hanting Chen

AI总结:

本文提出一种利用极轻量级量化编码器生成自适应条件的方法,有效降低了扩散模型ODE轨迹曲率,在仅增加1%参数且无需额外正则化的情况下大幅加速采样并提升图像质量。

摘要:

扩散模型中ODE轨迹的曲率阻碍了其在少数函数评估次数(NFE)下生成高质量图像的能力。本文提出一种新颖有效的方法,通过利用自适应条件来降低轨迹曲率。通过采用极轻量级的量化编码器,我们的方法仅增加1%的训练参数,消除了对额外正则化项的需求,却获得了显著更好的样本质量。该方法在加速ODE采样的同时,保留了SDE技术在下游任务图像编辑方面的能力。大量实验验证了我们的方法能在极有限的采样成本下生成高质量结果。仅用6 NFE,我们在CIFAR-10上取得5.14 FID,在FFHQ 64x64上取得6.91 FID,在AFHQv2上取得3.10 FID。

英文摘要:

The curvature of ODE trajectories in diffusion models hinders their ability to generate high-quality images in a few number of function evaluations (NFE). In this paper, we propose a novel and effective approach to reduce trajectory curvature by utilizing adaptive conditions. By employing a extremely light-weight quantized encoder, our method incurs only an additional 1% of training parameters, eliminates the need for extra regularization terms, yet achieves significantly better sample quality. Our approach accelerates ODE sampling while preserving the downstream task image editing capabilities of SDE techniques. Extensive experiments verify that our method can generate high quality results under extremely limited sampling costs. With only 6 NFE, we achieve 5.14 FID on CIFAR-10, 6.91 FID on FFHQ 64x64 and 3.10 FID on AFHQv2.

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。

↑