Saved in:
Bibliographic Details
Main Authors: Zhu, Jinchao, Wang, Yuxuan, Tu, Xiaobing, Pan, Siyuan, Wan, Pengfei, Huang, Gao
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2312.15516
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909127614136320
author Zhu, Jinchao
Wang, Yuxuan
Tu, Xiaobing
Pan, Siyuan
Wan, Pengfei
Huang, Gao
author_facet Zhu, Jinchao
Wang, Yuxuan
Tu, Xiaobing
Pan, Siyuan
Wan, Pengfei
Huang, Gao
contents The Stable Diffusion Model (SDM) is a popular and efficient text-to-image (t2i) generation and image-to-image (i2i) generation model. Although there have been some attempts to reduce sampling steps, model distillation, and network quantization, these previous methods generally retain the original network architecture. Billion scale parameters and high computing requirements make the research of model architecture adjustment scarce. In this work, we first explore the computational redundancy part of the network, and then prune the redundancy blocks of the model and maintain the network performance through a progressive incubation strategy. Secondly, in order to maintaining the model performance, we add cross-layer multi-expert conditional convolution (CLME-Condconv) to the block pruning part to inherit the original convolution parameters. Thirdly, we propose a global-regional interactive (GRI) attention to speed up the computationally intensive attention part. Finally, we use semantic-aware supervision (SAS) to align the outputs of the teacher model and student model at the semantic level. Experiments show that this method can effectively train a lightweight model close to the performance of the original SD model, and effectively improve the model speed under limited resources. Experiments show that the proposed method can effectively train a light-weight model close to the performance of the original SD model, and effectively improve the model speed under limited resources. After acceleration, the UNet part of the model is 22% faster and the overall speed is 19% faster.
format Preprint
id arxiv_https___arxiv_org_abs_2312_15516
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle A-SDM: Accelerating Stable Diffusion through Redundancy Removal and Performance Optimization
Zhu, Jinchao
Wang, Yuxuan
Tu, Xiaobing
Pan, Siyuan
Wan, Pengfei
Huang, Gao
Computer Vision and Pattern Recognition
The Stable Diffusion Model (SDM) is a popular and efficient text-to-image (t2i) generation and image-to-image (i2i) generation model. Although there have been some attempts to reduce sampling steps, model distillation, and network quantization, these previous methods generally retain the original network architecture. Billion scale parameters and high computing requirements make the research of model architecture adjustment scarce. In this work, we first explore the computational redundancy part of the network, and then prune the redundancy blocks of the model and maintain the network performance through a progressive incubation strategy. Secondly, in order to maintaining the model performance, we add cross-layer multi-expert conditional convolution (CLME-Condconv) to the block pruning part to inherit the original convolution parameters. Thirdly, we propose a global-regional interactive (GRI) attention to speed up the computationally intensive attention part. Finally, we use semantic-aware supervision (SAS) to align the outputs of the teacher model and student model at the semantic level. Experiments show that this method can effectively train a lightweight model close to the performance of the original SD model, and effectively improve the model speed under limited resources. Experiments show that the proposed method can effectively train a light-weight model close to the performance of the original SD model, and effectively improve the model speed under limited resources. After acceleration, the UNet part of the model is 22% faster and the overall speed is 19% faster.
title A-SDM: Accelerating Stable Diffusion through Redundancy Removal and Performance Optimization
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2312.15516