We use a few cookies

Essential cookies keep the library working. With your permission, we'd also like to use analytics to see what's useful and improve the experience. Read our Privacy Policy or Cookie Policy.

Skip to main content
Home/e-Resources/Artificial Intelligence & Machine Learning/Scaling Local Self-Attention for Parameter Efficient Visual Backbones
Conference ProceedingsArtificial Intelligence & Machine Learning

Scaling Local Self-Attention for Parameter Efficient Visual Backbones

Checking your access to this record.

Abstract & Overview

Self-attention has the promise of improving computer vision systems due to parameter-independent scaling of receptive fields and content-dependent interactions, in contrast to parameter-dependent scaling and content-independent interactions of convolutions. Self-attention models have recently been shown to have encouraging improvements on accuracy-parameter trade-offs compared to baseline convolutional models such as ResNet-50. In this work, we develop self-attention models that can outperform not just the canonical baseline models, but even the high-performing convolutional models. We propose two extensions to self-attention that, in conjunction with a more efficient implementation of self-attention, improve the speed, memory usage, and accuracy of these models. We leverage these improvements to develop a new self-attention model family, HaloNets, which reach state-of-the-art accuracies on the parameter-limited setting of the ImageNet classification benchmark. In preliminary transfer learning experiments, we find that HaloNet models outperform much larger models and have better inference performance. On harder tasks such as object detection and instance segmentation, our simple local self-attention and convolutional hybrids show improvements over very strong baselines. These results mark another step in demonstrating the efficacy of self-attention models on settings traditionally dominated by convolutions.1

Cite this record

Ashish Vaswani, Prajit Ramachandran, Aravind Srinivas, Niki Jitendra Parmar, Blake A. Hechtman, Jonathon Shlens (2021). Scaling Local Self-Attention for Parameter Efficient Visual Backbones. arXiv (Cornell University). https://doi.org/10.1109/cvpr46437.2021.01270

Copy in another style

Vaswani, A., Ramachandran, P., Srinivas, A., Parmar, N. J., Hechtman, B. A., & Shlens, J. (2021). Scaling Local Self-Attention for Parameter Efficient Visual Backbones. arXiv (Cornell University). https://doi.org/10.1109/cvpr46437.2021.01270
Author names are used as deposited, so check their order before you submit.

Metadata is harvested from open-access indexes and can be incomplete. Check a citation against the published version before relying on it — see our open access & attribution policy.