In recent years, there has been significant progress in the development of text-to-image generative models. Evaluating the quality of the generative models is one essential step in the development process. Unfortunately, the evaluation process could consume a significant amount of computational resources, making the required periodic evaluation of model performance (e.g., monitoring training progress) impractical. Therefore, we seek to improve the evaluation efficiency by selecting the representative subset of the text-image dataset. We systematically investigate the design choices, including the selection criteria (textural features or image-based metrics) and the selection granularity (prompt-level or set-level). We find that the insights from prior work on subset selection for training data do not generalize to this problem, and we propose FlashEval, an iterative search algorithm tailored to evaluation data selection. We demonstrate the effectiveness of FlashEval on ranking diffusion models with various configurations, including architectures, quantization levels, and sampler schedules on COCO and DiffusionDB datasets. Our searched 50-item subset could achieve comparable evaluation quality to the randomly sampled 500-item subset for COCO annotations on unseen models, achieving a 10x evaluation speedup. We release the condensed subset of these commonly used datasets to help facilitate diffusion algorithm design and evaluation, and open-source FlashEval as a tool for condensing future datasets, accessible at https://github.com/thu-nics/FlashEval.
Fast-GANFIT: Generative Adversarial Network for High Fidelity 3D Face Reconstruction
通过利用深度卷积神经网络(DCNN)的能力,已经做了大量的工作来从单个图像重建3D面部结构。在最近的作品中,纹理特征要么对应于线性纹理空间的分量,要么由自动编码器直接从野外图像中学习。在所有情况下,面部纹理重建的质量仍然不能建模面部纹理与高频细节。在本文中,我们采取了一种完全不同的方法,利用生成对抗网络(GANs)和DCNN的强大功能,从单张图像中重建面部纹理和形状。也就是说,我们利用GANs在\edit{from a large-scale 3D texture dataset}之前训练一个非常强大的面部纹理。然后,我们重新审视原始的3D变形模型(3DMMs)拟合利用非线性优化,以找到最佳的潜在参数,最好地重建测试图像,但在一个新的角度。为了对初始化具有鲁棒性并加快拟合过程,我们提出了一种新的基于自监督回归的方法。我们在真实感和身份保持3D人脸重建中表现出出色的效果,并首次实现了高频细节的面部纹理重建。
When Distillation Breaks Motion Control: Restoring Generative Trajectories for Fast Video Generators
Zhejiang University of Technology(浙江工业大学);University of New South Wales (UNSW Sydney)(新南威尔士大学(悉尼));University of Auckland(奥克兰大学);Zhejiang University(浙江大学)
Generative Adversarial Networks (GAN) Powered Fast Magnetic Resonance Imaging -- Mini Review, Comparison and Perspectives
Imperial College London(伦敦帝国理工学院);Royal Brompton Hospital(皇家布朗普顿医院);Yantai University(烟台大学);University of Cambridge(剑桥大学);Beijing Institute of Technology(北京理工大学)
Fast, Scale-Adaptive, and Uncertainty-Aware Downscaling of Earth System Model Fields with Generative Machine Learning
Technical University Munich(慕尼黑工业大学);Potsdam Institute for Climate Impact Research(波茨坦气候影响研究所);Institute of Atmospheric Physics, Chinese Academy of Sciences(中国科学院大气物理研究所);University of Exeter(埃克塞特大学)
Spontaneous Symmetry Breaking in Generative Diffusion Models
Jheronimus Academy of Data Science(杰罗尼穆斯数据科学学院);Tilburg University(蒂尔堡大学);Radboud University(拉德堡德大学);Donders Institute for Brain, Cognition and Behaviour(唐德斯大脑、认知与行为研究所)
Edge-Enhanced Dual Discriminator Generative Adversarial Network for Fast MRI with Parallel Imaging Using Multi-view Information
Zhejiang Shuren University(浙江树人大学);Imperial College London(伦敦帝国理工学院);Nantong University(南通大学);Yantai University(烟台大学);Peking University School and Hospital of Stomatology(北京大学口腔医学院);Peking University(北京大学);TECNALIA;Basque Research and Technology Alliance(巴斯克研究与技术联盟);University of the Basque Country(巴斯克大学);The First Affiliated Hospital of Shenzhen University Health Science Center(深圳大学医学部第一附属医院);Houston Methodist Cancer Center(休斯顿卫理公会癌症中心);Houston Methodist Hospital(休斯顿卫理公会医院);Weill Cornell Medicine(威尔康奈尔医学院)
GenReg: Deep Generative Method for Fast Point Cloud Registration
Image X Institute, University of Sydney(悉尼大学 Image X 研究所);Chongqing University of Posts and Telecommunications(重庆邮电大学);University of Techonology Sydney(悉尼科技大学);Jiangxi University of Finance and Economics(江西财经大学);Xpeng(小鹏)
Fast refacing of MR images with a generative neural network lowers re-identification risk and preserves volumetric consistency
Lausanne University Hospital (CHUV)(洛桑大学医院);Siemens Healthineers International AG(西门子医疗国际股份公司);Ecole Polytechnique Fédérale de Lausanne (EPFL)(洛桑联邦理工学院);University of Lausanne (UNIL)(洛桑大学)
Wavelet Flow: Fast Training of High Resolution Normalizing Flows
York University(约克大学);Borealis AI(北极光人工智能);Ryerson University(瑞尔森大学);Samsung AI Centre Toronto(三星人工智能中心多伦多分部);Vector Institute(向量研究所)
规范化流是一类概率生成模型,它允许快速密度计算和有效采样,并且在模拟图像等复杂分布时有效。当前方法的一个缺点是其显著的训练成本,有时需要数月的GPU训练时间才能实现最先进的结果。本文介绍了小波流,一个多尺度,规范化的流架构的基础上小波。小波流具有信号尺度的显式表示,其固有地包括较低分辨率信号的模型和较高分辨率信号的条件生成,即,超分辨率。小波流的一个主要优点是能够为高分辨率数据构建生成模型(例如,1024 x 1024图像),这是不切实际的与以前的模型。此外,在标准(低分辨率)基准测试中,Wavelet Flow在每维比特数方面与之前的归一化流具有竞争力,同时训练速度最多快15倍。
Automated Tuning for Diffusion Inverse Problem Solvers without Generative Prior Retraining
Department of Electrical \& Computer Engineering, University of Minnesota, MN, USA;Center for Magnetic Resonance Research, University of Minnesota, MN, USA