English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present an efficient approach for Masked Image Modeling (MIM) withhierarchical Vision Transformers (ViTs), allowing the hierarchical ViTs todiscard masked patches and operate only on the visible ones. Our approachconsists of three key designs. First, for window attention, we propose a GroupWindow Attention scheme following the Divide-and-Conquer strategy. To mitigatethe quadratic complexity of the self-attention w.r.t. the number of patches,group attention encourages a uniform partition that visible patches within eachlocal window of arbitrary size can be grouped with equal size, where maskedself-attention is then performed within each group. Second, we further improvethe grouping strategy via the Dynamic Programming algorithm to minimize theoverall computation cost of the attention on the grouped patches. Third, as forthe convolution layers, we convert them to the Sparse Convolution that worksseamlessly with the sparse data, i.e., the visible patches in MIM. As a result,MIM can now work on most, if not all, hierarchical ViTs in a green andefficient way. For example, we can train the hierarchical ViTs, e.g., SwinTransformer and Twins Transformer, about 2.7$\times$ faster and reduce the GPUmemory usage by 70%, while still enjoying competitive performance on ImageNetclassification and the superiority on downstream COCO object detectionbenchmarks. Code and pre-trained models have been made publicly available athttps://github.com/LayneH/GreenMIM.

Green Hierarchical Vision Transformer for Masked Image Modeling