RESEARCH · RESEARCH · #1721
Normalizing Trajectory Models (NTM) introduced as a few-step, likelihood‑based generative method
Researchers (Jiatao Gu et al.) propose Normalizing Trajectory Models (NTM), which model each reverse sampling step as a conditional normalizing flow and are trained with exact trajectory likelihood. NTM combines shallow invertible blocks per step with a deep parallel predictor across the trajectory, can be trained from scratch or initialized from pretrained flow-matching models, and enables a self-distillation step that yields high-quality samples in four sampling steps on text-to-image benchmarks while retaining exact likelihood over the trajectory.
KEY POINTS
- Researchers (Jiatao Gu et al.) propose Normalizing Trajectory Models (NTM), which model each reverse sampling step as a conditional normalizing flow and are trained with exact trajectory likelihood.
- NTM combines shallow invertible blocks per step with a deep parallel predictor across the trajectory, can be trained from scratch or initialized from pretrained flow-matching models, and enables a self-distillation step that yields high-quality samples in four sampling steps on text-to-image benchmarks while retaining exact likelihood over the trajectory.
- NTM matters because it enables fast (few-step) generation while preserving exact likelihood training, combining the speed of distilled samplers with the evaluability of likelihood-based models.
WHY IT MATTERS
NTM matters because it enables fast (few-step) generation while preserving exact likelihood training, combining the speed of distilled samplers with the evaluability of likelihood-based models.