ROMS-IMLE: A Minimalist Approach to Competitive Single-Step Generative Modelling

2026-07-21

AI Summary

This work challenges the prevailing paradigm that generative models must rely on multi-step progressive denoising, proposing instead an extremely minimalist single-step generation architecture. By abandoning commonly used components in diffusion models—such as multi-step transformations, variational inference, adversarial training, and numerical integration—the method combines an implicit maximum likelihood estimation (IMLE) objective with a moderately sized convolutional network to achieve high-quality image synthesis without iterative refinement. On ImageNet at 256×256 resolution, the model attains a competitive FID score of 2.56 using only a single forward pass, while simultaneously demonstrating both high precision and high recall. These results underscore the viability and potential of ultra-simple single-step generative models.

Abstract

Generative models have undergone many generations of evolution, from VAEs/GANs to diffusion/flow matching. Along the way, the underlying techniques have become more complicated and various beliefs about what drives strong empirical performance have taken hold. Due to the success of diffusion models and flow matching, one of the more common beliefs is the importance of transforming the noise distribution to the data distribution gradually through many small transformations. We ask whether this is truly necessary, and take a minimalist approach to designing a competitive generative model. We start with the bare-bones essentials, namely just a training objective and a model. We purposefully make both simple. For the training objective, we choose Implicit Maximum Likelihood Estimation (IMLE), and eschew more complicated alternatives such as variational inference, adversarial training and numerical integration. For the model, we eschew transformers and instead choose a moderately sized convolutional network. Then we judiciously added elements that are truly essential, which surprisingly do not include iterative denoising. The result is a single-step parameter-efficient generative model that produces high quality samples at fast speed: it achieves an FID of 2.56 on ImageNet 256 and simultaneously attains good precision and recall.

Problem

• generative modelling
• single-step generation
• diffusion models
• flow matching
• model simplicity

Innovation

• single-step generation
• Implicit Maximum Likelihood Estimation
• convolutional generative model
• parameter-efficient
• non-iterative denoising