Apple's Normalizing Trajectory Models match strong baselines in four image-generation steps
Overview
Apple researchers introduced Normalizing Trajectory Models (NTM), which model each reverse diffusion step as a conditional normalizing flow trained with exact likelihood.
According to the researchers' paper, NTM matches or outperforms strong image generation baselines on text-to-image benchmarks while using only four sampling steps. The authors also state that NTM uniquely retains exact likelihood over the generative trajectory, a property they present as distinguishing it from the baselines. These claims come from the researchers' own report and have not been independently evaluated in the material available here.
Written by AI from the articles below · updated Oct 8, 7:57 PM ET
Check the sources:
Article timeline
Follow the coverage from different perspectives. Times are ET.
- Apple Machine Learning ResearchApple's Normalizing Trajectory Models generate images in four steps with exact likelihood
AIApple researchers introduced Normalizing Trajectory Models (NTM), which model each reverse diffusion step as a conditional normalizing flow trained with exact likelihood. The model matches or outperforms strong image generation baselines on text-to-image benchmarks in just four sampling steps while retaining exact likelihood over the generative trajectory.
Heat trend
Not enough continuous observations to show a trend yet.