DEFT: Efficient Fine-Tuning of Diffusion Models by Learning the Generalised $h$-transform

• URL: http://arxiv.org/abs/2406.01781v3
• Date: Mon, 18 Nov 2024 15:11:11 GMT
• Title: DEFT: Efficient Fine-Tuning of Diffusion Models by Learning the Generalised $h$-transform
• Authors: Alexander Denker, Francisco Vargas, Shreyas Padhy, Kieran Didi, Simon Mathis, Vincent Dutordoir, Riccardo Barbano, Emile Mathieu, Urszula Julia Komorowska, Pietro Lio,
• Abstract summary: We propose DEFT (Doob's h-transform Efficient FineTuning), a new approach for conditional generation that simply fine-tunes a very small network to quickly learn the conditional $h$-transform. On image reconstruction tasks, we achieve speedups of up to 1.6$times$, while having the best perceptual quality on natural images and reconstruction performance on medical images.
• Score: 44.29325094229024
• License:
• Abstract: Generative modelling paradigms based on denoising diffusion processes have emerged as a leading candidate for conditional sampling in inverse problems. In many real-world applications, we often have access to large, expensively trained unconditional diffusion models, which we aim to exploit for improving conditional sampling. Most recent approaches are motivated heuristically and lack a unifying framework, obscuring connections between them. Further, they often suffer from issues such as being very sensitive to hyperparameters, being expensive to train or needing access to weights hidden behind a closed API. In this work, we unify conditional training and sampling using the mathematically well-understood Doob's h-transform. This new perspective allows us to unify many existing methods under a common umbrella. Under this framework, we propose DEFT (Doob's h-transform Efficient FineTuning), a new approach for conditional generation that simply fine-tunes a very small network to quickly learn the conditional $h$-transform, while keeping the larger unconditional network unchanged. DEFT is much faster than existing baselines while achieving state-of-the-art performance across a variety of linear and non-linear benchmarks. On image reconstruction tasks, we achieve speedups of up to 1.6$\times$, while having the best perceptual quality on natural images and reconstruction performance on medical images. Further, we also provide initial experiments on protein motif scaffolding and outperform reconstruction guidance methods.

Related papers

• Conditional diffusions for neural posterior estimation [8.37884129644711]
  We show the effectiveness of conditional diffusions as an alternative to normalizing flows for NPE. Our results show improved stability, superior accuracy, and faster training times, even with simpler, shallower models.
  arXiv  Detail & Related papers  (2024-10-24T19:13:13Z)
• Fast constrained sampling in pre-trained diffusion models [77.21486516041391]
  Diffusion models have dominated the field of large, generative image models. We propose an algorithm for fast-constrained sampling in large pre-trained diffusion models.
  arXiv  Detail & Related papers  (2024-10-24T14:52:38Z)
• One Step Diffusion via Shortcut Models [109.72495454280627]
  We introduce shortcut models, a family of generative models that use a single network and training phase to produce high-quality samples. Shortcut models condition the network on the current noise level and also on the desired step size, allowing the model to skip ahead in the generation process. Compared to distillation, shortcut models reduce complexity to a single network and training phase and additionally allow varying step budgets at inference time.
  arXiv  Detail & Related papers  (2024-10-16T13:34:40Z)
• A Simple Approach to Unifying Diffusion-based Conditional Generation [63.389616350290595]
  We introduce a simple, unified framework to handle diverse conditional generation tasks. Our approach enables versatile capabilities via different inference-time sampling schemes. Our model supports additional capabilities like non-spatially aligned and coarse conditioning.
  arXiv  Detail & Related papers  (2024-10-15T09:41:43Z)
• CoSIGN: Few-Step Guidance of ConSIstency Model to Solve General INverse Problems [3.3969056208620128]
  We propose to push the boundary of inference steps to 1-2 NFEs while still maintaining high reconstruction quality. Our method achieves new state-of-the-art in diffusion-based inverse problem solving.
  arXiv  Detail & Related papers  (2024-07-17T15:57:50Z)
• One-Step Diffusion Distillation via Deep Equilibrium Models [64.11782639697883]
  We introduce a simple yet effective means of distilling diffusion models directly from initial noise to the resulting image. Our method enables fully offline training with just noise/image pairs from the diffusion model. We demonstrate that the DEQ architecture is crucial to this capability, as GET matches a $5times$ larger ViT in terms of FID scores.
  arXiv  Detail & Related papers  (2023-12-12T07:28:40Z)
• Variational Bayes image restoration with compressive autoencoders [4.879530644978008]
  Regularization of inverse problems is of paramount importance in computational imaging. In this work, we first propose to use compressive autoencoders instead of state-of-the-art generative models. As a second contribution, we introduce the Variational Bayes Latent Estimation (VBLE) algorithm.
  arXiv  Detail & Related papers  (2023-11-29T15:49:31Z)
• Steered Diffusion: A Generalized Framework for Plug-and-Play Conditional Image Synthesis [62.07413805483241]
  Steered Diffusion is a framework for zero-shot conditional image generation using a diffusion model trained for unconditional generation. We present experiments using steered diffusion on several tasks including inpainting, colorization, text-guided semantic editing, and image super-resolution.
  arXiv  Detail & Related papers  (2023-09-30T02:03:22Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.