AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kag, Anil, Coskun, Huseyin, Chen, Jierun, Cao, Junli, Menapace, Willi, Siarohin, Aliaksandr, Tulyakov, Sergey, Ren, Jian
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929582272151552
author Kag, Anil
Coskun, Huseyin
Chen, Jierun
Cao, Junli
Menapace, Willi
Siarohin, Aliaksandr
Tulyakov, Sergey
Ren, Jian
author_facet Kag, Anil
Coskun, Huseyin
Chen, Jierun
Cao, Junli
Menapace, Willi
Siarohin, Aliaksandr
Tulyakov, Sergey
Ren, Jian
contents Neural network architecture design requires making many crucial decisions. The common desiderata is that similar decisions, with little modifications, can be reused in a variety of tasks and applications. To satisfy that, architectures must provide promising latency and performance trade-offs, support a variety of tasks, scale efficiently with respect to the amounts of data and compute, leverage available data from other tasks, and efficiently support various hardware. To this end, we introduce AsCAN -- a hybrid architecture, combining both convolutional and transformer blocks. We revisit the key design principles of hybrid architectures and propose a simple and effective \emph{asymmetric} architecture, where the distribution of convolutional and transformer blocks is \emph{asymmetric}, containing more convolutional blocks in the earlier stages, followed by more transformer blocks in later stages. AsCAN supports a variety of tasks: recognition, segmentation, class-conditional image generation, and features a superior trade-off between performance and latency. We then scale the same architecture to solve a large-scale text-to-image task and show state-of-the-art performance compared to the most recent public and commercial models. Notably, even without any computation optimization for transformer blocks, our models still yield faster inference speed than existing works featuring efficient attention mechanisms, highlighting the advantages and the value of our approach.
format Preprint
id arxiv_https___arxiv_org_abs_2411_04967
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation
Kag, Anil
Coskun, Huseyin
Chen, Jierun
Cao, Junli
Menapace, Willi
Siarohin, Aliaksandr
Tulyakov, Sergey
Ren, Jian
Computer Vision and Pattern Recognition
Machine Learning
Neural network architecture design requires making many crucial decisions. The common desiderata is that similar decisions, with little modifications, can be reused in a variety of tasks and applications. To satisfy that, architectures must provide promising latency and performance trade-offs, support a variety of tasks, scale efficiently with respect to the amounts of data and compute, leverage available data from other tasks, and efficiently support various hardware. To this end, we introduce AsCAN -- a hybrid architecture, combining both convolutional and transformer blocks. We revisit the key design principles of hybrid architectures and propose a simple and effective \emph{asymmetric} architecture, where the distribution of convolutional and transformer blocks is \emph{asymmetric}, containing more convolutional blocks in the earlier stages, followed by more transformer blocks in later stages. AsCAN supports a variety of tasks: recognition, segmentation, class-conditional image generation, and features a superior trade-off between performance and latency. We then scale the same architecture to solve a large-scale text-to-image task and show state-of-the-art performance compared to the most recent public and commercial models. Notably, even without any computation optimization for transformer blocks, our models still yield faster inference speed than existing works featuring efficient attention mechanisms, highlighting the advantages and the value of our approach.
title AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2411.04967