Accelerating Diffusion Models with One-to-Many Knowledge Distillation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Linfeng, Ma, Kaisheng
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912060435070976
author Zhang, Linfeng
Ma, Kaisheng
author_facet Zhang, Linfeng
Ma, Kaisheng
contents Significant advancements in image generation have been made with diffusion models. Nevertheless, when contrasted with previous generative models, diffusion models face substantial computational overhead, leading to failure in real-time generation. Recent approaches have aimed to accelerate diffusion models by reducing the number of sampling steps through improved sampling techniques or step distillation. However, the methods to diminish the computational cost for each timestep remain a relatively unexplored area. Observing the fact that diffusion models exhibit varying input distributions and feature distributions at different timesteps, we introduce one-to-many knowledge distillation (O2MKD), which distills a single teacher diffusion model into multiple student diffusion models, where each student diffusion model is trained to learn the teacher's knowledge for a subset of continuous timesteps. Experiments on CIFAR10, LSUN Church, CelebA-HQ with DDPM and COCO30K with Stable Diffusion show that O2MKD can be applied to previous knowledge distillation and fast sampling methods to achieve significant acceleration. Codes will be released in Github.
format Preprint
id arxiv_https___arxiv_org_abs_2410_04191
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Accelerating Diffusion Models with One-to-Many Knowledge Distillation
Zhang, Linfeng
Ma, Kaisheng
Computer Vision and Pattern Recognition
Artificial Intelligence
Significant advancements in image generation have been made with diffusion models. Nevertheless, when contrasted with previous generative models, diffusion models face substantial computational overhead, leading to failure in real-time generation. Recent approaches have aimed to accelerate diffusion models by reducing the number of sampling steps through improved sampling techniques or step distillation. However, the methods to diminish the computational cost for each timestep remain a relatively unexplored area. Observing the fact that diffusion models exhibit varying input distributions and feature distributions at different timesteps, we introduce one-to-many knowledge distillation (O2MKD), which distills a single teacher diffusion model into multiple student diffusion models, where each student diffusion model is trained to learn the teacher's knowledge for a subset of continuous timesteps. Experiments on CIFAR10, LSUN Church, CelebA-HQ with DDPM and COCO30K with Stable Diffusion show that O2MKD can be applied to previous knowledge distillation and fast sampling methods to achieve significant acceleration. Codes will be released in Github.
title Accelerating Diffusion Models with One-to-Many Knowledge Distillation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2410.04191