Noise Estimation for Generative Diffusion Models
- URL: http://arxiv.org/abs/2104.02600v1
- Date: Tue, 6 Apr 2021 15:46:16 GMT
- Title: Noise Estimation for Generative Diffusion Models
- Authors: Robin San-Roman, Eliya Nachmani, Lior Wolf
- Abstract summary: In this work, we present a simple and versatile learning scheme that can adjust the noise parameters for any given number of steps.
Our approach comes at a negligible computation cost.
- Score: 91.22679787578438
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Generative diffusion models have emerged as leading models in speech and
image generation. However, in order to perform well with a small number of
denoising steps, a costly tuning of the set of noise parameters is needed. In
this work, we present a simple and versatile learning scheme that can
step-by-step adjust those noise parameters, for any given number of steps,
while the previous work needs to retune for each number separately.
Furthermore, without modifying the weights of the diffusion model, we are able
to significantly improve the synthesis results, for a small number of steps.
Our approach comes at a negligible computation cost.
Related papers
- Bayesian Inference of General Noise Model Parameters from Surface Code's Syndrome Statistics [0.0]
We propose general noise model Bayesian inference methods that integrate the surface code's tensor network simulator.
For stationary noise, where the noise parameters are constant and do not change, we propose a method based on the Markov chain Monte Carlo.
For time-varying noise, which is a more realistic situation, we introduce another method based on the sequential Monte Carlo.
arXiv Detail & Related papers (2024-06-13T10:26:04Z) - One Noise to Rule Them All: Learning a Unified Model of Spatially-Varying Noise Patterns [33.293193191683145]
We present a single generative model which can learn to generate multiple types of noise as well as blend between them.
We also present an application of our model to improving inverse procedural material design.
arXiv Detail & Related papers (2024-04-25T02:23:11Z) - Blue noise for diffusion models [50.99852321110366]
We introduce a novel and general class of diffusion models taking correlated noise within and across images into account.
Our framework allows introducing correlation across images within a single mini-batch to improve gradient flow.
We perform both qualitative and quantitative evaluations on a variety of datasets using our method.
arXiv Detail & Related papers (2024-02-07T14:59:25Z) - Not All Steps are Equal: Efficient Generation with Progressive Diffusion
Models [62.155612146799314]
We propose a novel two-stage training strategy termed Step-Adaptive Training.
In the initial stage, a base denoising model is trained to encompass all timesteps.
We partition the timesteps into distinct groups, fine-tuning the model within each group to achieve specialized denoising capabilities.
arXiv Detail & Related papers (2023-12-20T03:32:58Z) - One More Step: A Versatile Plug-and-Play Module for Rectifying Diffusion
Schedule Flaws and Enhancing Low-Frequency Controls [77.42510898755037]
One More Step (OMS) is a compact network that incorporates an additional simple yet effective step during inference.
OMS elevates image fidelity and harmonizes the dichotomy between training and inference, while preserving original model parameters.
Once trained, various pre-trained diffusion models with the same latent domain can share the same OMS module.
arXiv Detail & Related papers (2023-11-27T12:02:42Z) - Boosting Fast and High-Quality Speech Synthesis with Linear Diffusion [85.54515118077825]
This paper proposes a linear diffusion model (LinDiff) based on an ordinary differential equation to simultaneously reach fast inference and high sample quality.
To reduce computational complexity, LinDiff employs a patch-based processing approach that partitions the input signal into small patches.
Our model can synthesize speech of a quality comparable to that of autoregressive models with faster synthesis speed.
arXiv Detail & Related papers (2023-06-09T07:02:43Z) - Realistic Noise Synthesis with Diffusion Models [68.48859665320828]
Deep image denoising models often rely on large amount of training data for the high quality performance.
We propose a novel method that synthesizes realistic noise using diffusion models, namely Realistic Noise Synthesize Diffusor (RNSD)
RNSD can incorporate guided multiscale content, such as more realistic noise with spatial correlations can be generated at multiple frequencies.
arXiv Detail & Related papers (2023-05-23T12:56:01Z) - On tuning consistent annealed sampling for denoising score matching [17.10144603522758]
Score-based generative models provide state-of-the-art quality for image and audio synthesis.
In this note, we first overview three common sampling schemes for models trained with denoising score matching.
arXiv Detail & Related papers (2021-04-08T12:19:10Z)
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.