Denoising Diffusion Probabilistic Model (DDPM) is able to make flexible conditional image generation from prior noise to real data, by introducing an independent noise-aware classifier to provide conditional gradient guidance at each time step of denoising process. However, due to the ability of classifier to easily discriminate an incompletely generated image only with high-level structure, the gradient, which is a kind of class information guidance, tends to vanish early, leading to the collapse from conditional generation process into the unconditional process. To address this problem, we propose two simple but effective approaches from two perspectives. For sampling procedure, we introduce the entropy of predicted distribution as the measure of guidance vanishing level and propose an entropy-aware scaling method to adaptively recover the conditional semantic guidance. For training stage, we propose the entropy-aware optimization objectives to alleviate the overconfident prediction for noisy data.On ImageNet1000 256x256, with our proposed sampling scheme and trained classifier, the pretrained conditional and unconditional DDPM model can achieve 10.89% (4.59 to 4.09) and 43.5% (12 to 6.78) FID improvement respectively.
翻译:DDPM(DDPM)能够从先前的噪音到真实数据灵活而有条件的图像生成,从先前的噪音到真实数据,方法是引入一个独立的噪音分级器,在拆除过程的每个阶段都提供有条件的梯度指导。然而,由于分类器能够很容易地将不完全生成的图像与高层次结构区分开来,梯度(这是一种阶级信息指导)往往会过早消失,导致从有条件生成过程向无条件过程的崩溃。为了解决这一问题,我们从两个角度提出了两种简单而有效的方法。对于取样程序,我们引入了预测分布的聚合物,作为指南消失水平的测量标准,并提出了一种对导值的测量方法,以适应性地恢复有条件的语义指导。对于培训阶段,我们提出了对恒度的优化目标,以缓解对扰动数据过于自信的预测。关于图像Net1000 256x256,我们提议的取样计划和经过培训的分类方法,预先培训的有条件和无条件的DDPM模型可以分别实现10.89%(4.59%至4.09)和43.5%(12至6.78)的改进。