Stable Target Field for Reduced Variance Score Estimation in Diffusion Models

by   Yilun Xu, et al.

Diffusion models generate samples by reversing a fixed forward diffusion process. Despite already providing impressive empirical results, these diffusion models algorithms can be further improved by reducing the variance of the training targets in their denoising score-matching objective. We argue that the source of such variance lies in the handling of intermediate noise-variance scales, where multiple modes in the data affect the direction of reverse paths. We propose to remedy the problem by incorporating a reference batch which we use to calculate weighted conditional scores as more stable training targets. We show that the procedure indeed helps in the challenging intermediate regime by reducing (the trace of) the covariance of training targets. The new stable targets can be seen as trading bias for reduced variance, where the bias vanishes with increasing reference batch size. Empirically, we show that the new objective improves the image quality, stability, and training speed of various popular diffusion models across datasets with both general ODE and SDE solvers. When used in combination with EDM, our method yields a current SOTA FID of 1.90 with 35 network evaluations on the unconditional CIFAR-10 generation task. The code is available at


page 8

page 21

page 22

page 23


PFGM++: Unlocking the Potential of Physics-Inspired Generative Models

We introduce a new family of physics-inspired generative models termed P...

Subspace Diffusion Generative Models

Score-based models generate samples by mapping noise to data (and vice v...

Consistent Diffusion Models: Mitigating Sampling Drift by Learning to be Consistent

Imperfect score-matching leads to a shift between the training and the s...

TrojDiff: Trojan Attacks on Diffusion Models with Diverse Targets

Diffusion models have achieved great success in a range of tasks, such a...

Better Diffusion Models Further Improve Adversarial Training

It has been recognized that the data generated by the denoising diffusio...

On Calibrating Diffusion Probabilistic Models

Recently, diffusion probabilistic models (DPMs) have achieved promising ...

Analytic-DPM: an Analytic Estimate of the Optimal Reverse Variance in Diffusion Probabilistic Models

Diffusion probabilistic models (DPMs) represent a class of powerful gene...

Please sign up or login with your details

Forgot password? Click here to reset