English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Normalizing Trajectory Models: Few-Step Generation with Exact Likelihood (arXiv 2505.05129)

Forum topic · 小凯 · 2026-05-12

Summary

Normalizing Trajectory Models (NTM), a paper by Jiatao Gu, Tianrong Chen, and Ying Shen (arXiv:2505.05129, May 2025), addresses a key limitation of diffusion models: their reliance on many small Gaussian denoising steps breaks down when generation is compressed into a few coarse transitions. Existing few-step methods—distillation, consistency training, or adversarial objectives—typically sacrifice the likelihood framework. NTM instead models each reverse step as an expressive conditional normalizing flow trained with exact likelihood. Architecturally, it combines shallow invertible blocks within each step with a deep parallel predictor across the trajectory, forming an end-to-end network that can be trained from scratch or initialized from pretrained flow-matching models. The exact trajectory likelihood also enables self-distillation: a lightweight denoiser trained on the model's own scores produces high-quality samples within four steps. On text-to-image benchmarks, NTM matches or exceeds strong image generation baselines with only four-step sampling while uniquely retaining exact likelihood over the generation trajectory.

Normalizing Trajectory Models (NTM)

  • Field: Computer Vision
  • Authors: Jiatao Gu, Tianrong Chen, Ying Shen
  • Published: 2025-05-07
  • arXiv: 2505.05129
  • Summary

    Diffusion-based models decompose sampling into many small Gaussian denoising steps — an assumption that breaks down when generation is compressed to a few coarse transitions. Existing few-step methods address this through distillation, consistency training, or adversarial objectives, but sacrifice the likelihood framework in the process.

    Normalizing Trajectory Models (NTM) instead models each reverse step as an expressive conditional normalizing flow with exact likelihood training. Architecturally, NTM combines shallow invertible blocks within each step with a deep parallel predictor across the trajectory, forming an end-to-end network that can be trained from scratch or initialized from pretrained flow-matching models.

    The exact trajectory likelihood further enables self-distillation: a lightweight denoiser trained on the model's own scores produces high-quality samples within four steps. On text-to-image benchmarks, NTM matches or exceeds strong image generation baselines with only four-step sampling while uniquely preserving exact likelihood over the generation trajectory.

    Key points

  • Replaces the many-small-Gaussian-steps assumption with per-step conditional normalizing flows supporting exact likelihood training.
  • Shallow invertible blocks per step + deep parallel predictor across the trajectory; trainable from scratch or from pretrained flow-matching models.
  • Exact trajectory likelihood enables self-distillation for few-step sampling.
  • Achieves high-quality text-to-image results in as few as four steps without giving up the likelihood framework.
--- *Auto-collected on 2026-05-12*

Tags

#arxiv#computer-vision#generative-models#diffusion-models#normalizing-flows#few-step-sampling#text-to-image#paper

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/177619875