WaiT for the Signal: Simple Frequency-Aware Flow-Matching
Krunoslav Lehman Pavasovic, Théophane Vallaeys, Stéphane Mallat, Giulio Biroli, Luke Zettlemoyer, Brian Karrer, Jakob Verbeek
cs.CV, cs.AI, cs.LG, stat.ML
Submitted: 2026-07-30
License: http://creativecommons.org/licenses/by/4.0/
The gist: As image generation models scale to ever higher resolutions, global coherence, local detail, and texture fidelity become critical axes for generation quality.
Terminology
Abstract
As image generation models scale to ever higher resolutions, global coherence, local detail, and texture fidelity become critical axes for generation quality. However, standard flow matching treats all spatial frequencies uniformly, ignoring the natural frequency hierarchy where high-frequency bands become indistinguishable from pure noise far earlier than coarse structures. We introduce WaiT, a Wavelet-aware image Transformer that decomposes generation into coarse and fine bands via lossless wavelets. True to its name, the high-frequency bands wait for the signal: staying pure noise until coarse structure has emerged, then joining the flow for joint refinement. Since standard FID discards fine-grained detail through aggressive downsampling, we introduce a more stringent three-axis evaluation protocol to assess quality at native resolution. On ImageNet 512x512, WaiT achieves a pixel-space FID of 1.43 and is Pareto-optimal across all three axes, reducing sampling compute by up to 50%. With our largest 2B model, we set a new state-of-the-art FID of 1.3 for pixel-space models on ImageNet 512 resolution. Our formulation outperforms even the strongest latent-space models on texture fidelity, and scales seamlessly to high-resolution OpenImages and to video generation, achieving a state-of-the-art FVD of 0.84 on Kinetics-600 with no algorithmic modifications.
Sources
- A Short Note about Kinetics-600
- PixelFlow: Pixel-Space Generative Models with Flow
- DiP: Taming Diffusion Models in Pixel Space
- Diffusion Models Beat GANs on Image Synthesis
- Scaling Rectified Flow Transformers for High-Resolution Image Synthesis
- A Fourier Space Perspective on Diffusion Models
- Photorealistic Video Generation with Diffusion Models
- Unified Latents (UL): How to train your latents
- Simple diffusion: End-to-end diffusion for high resolution images
- Simpler Diffusion (SiD2): 1.5 FID on ImageNet512 with pixel-space diffusion
- Scalable Adaptive Computation for Iterative Generation
- Pyramidal Flow Matching for Efficient Video Generative Modeling
- Understanding Diffusion Objectives as the ELBO with Simple Data Augmentation
- SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformers
- Advanced Unstructured Data Processing for ESG Reports: A Methodology for Structured Transformation and Enhanced Analysis
- Scalable Diffusion Models with Transformers
- Qwen2.5-VL Technical Report
- Towards Accurate Generative Models of Video: A New Metric & Challenges
- PixNerd: Pixel Neural Field Diffusion
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models