Search papers, labs, and topics across Lattice.
This paper introduces Colored Noise Sampling (CNS), a training-free stochastic solver for diffusion models that dynamically adjusts the injected noise spectrum based on timestep and frequency. CNS addresses the spectral bias of diffusion models by allocating energy towards unresolved frequency bands, leading to more efficient energy transfer and improved sample quality. Experiments across various architectures on ImageNet-256 demonstrate substantial FID score improvements compared to standard ODE and SDE samplers, highlighting the effectiveness of CNS as a plug-and-play inference-time substitution.
Diffusion models get a free lunch: a new sampler that cuts FID scores by up to 20% just by injecting the right kind of noise at each step.
Diffusion models achieve state-of-the-art image synthesis, with their generative trajectories fundamentally exhibiting a spectral bias, resolving low-frequency global structures early and high-frequency fine details later. Conventional stochastic differential equation (SDE) solvers fail to account for this dynamic, naively injecting uniform white noise throughout the entire process and misusing the finite energy budget. In this work, we establish a mathematical framework that reconsiders SDE inference as a targeted, frequency-decoupled energy transfer. Leveraging this framework, we introduce Colored Noise Sampling (CNS), a novel, training-free stochastic solver. Rather than injecting uniform white noise, CNS utilizes a dynamic, timestep- and frequency-dependent schedule that more efficiently allocates injected energy toward structurally unresolved frequency bands. By actively exploiting the model's inherent spectral bias, CNS systematically steers the generated distribution toward the true data manifold. Extensive experiments demonstrate that CNS significantly outperforms standard ODE and SDE baselines as a strictly plug-and-play, inference-time sampler substitution across diverse architectures (SiT, JiT, FLUX). Compared to standard sampling on ImageNet-256, CNS achieves substantial unguided FID reductions, improving from 8.26 to 6.27 on SiT-XL/2, 32.39 to 26.69 on JiT-B/16, and 11.88 to 8.31 on JiT-H/16, while yielding consistent relative FID improvements with Classifier-Free Guidance. Project page is available at https://hadardavidson.github.io/CNS/.
