Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Swin Transformer V2: Scaling Up Capacity and Resolution #516

Open
tkuri opened this issue Oct 22, 2022 · 0 comments
Open

Swin Transformer V2: Scaling Up Capacity and Resolution #516

tkuri opened this issue Oct 22, 2022 · 0 comments
Labels
Conference: CVPR Conference on Computer Vision and Pattern Recognition Subject: Backbone Year: 2022

Comments

@tkuri
Copy link
Owner

tkuri commented Oct 22, 2022

論文概要

Swin Transformerを30億パラメータまで拡張し1,536×1,536の解像度の画像を学習可能に。様々なベンチマークでSOTA。学習における不安定性を解決するためにモデルを改良(Layer Normの順番、Cosine Attentionの導入等)。更にGPUのメモリ消費量を大幅に削減する実装方法を提案。

bib_20220908 00
https://openaccess.thecvf.com/content/CVPR2022/html/Liu_Swin_Transformer_V2_Scaling_Up_Capacity_and_Resolution_CVPR_2022_paper.html

Code

https://github.com/microsoft/Swin-Transformer

@tkuri tkuri added Conference: CVPR Conference on Computer Vision and Pattern Recognition Year: 2022 Subject: Backbone labels Oct 22, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Conference: CVPR Conference on Computer Vision and Pattern Recognition Subject: Backbone Year: 2022
Projects
None yet
Development

No branches or pull requests

1 participant